microsatellites as a molecular tool to assess population genetics...
TRANSCRIPT
Faculteit Bio-ingenieurswetenschappen
Academiejaar 2015 – 2016
Microsatellites as a molecular tool to assess population genetics in declining Brazilian bumblebee and stingless
bee species
Laura Golsteyn Promotoren: Prof. dr. ir. Guy Smagghe & dr. Ivan Meeus Tutor: dr. Kevin Maebe
Masterproef voorgedragen tot het behalen van de graad van Master in de bio-ingenieurswetenschappen: Landbouwkunde
1
Preface To bee or not to be… That’s the question. This definitely accounted for me during this master thesis!
But it is also an important issue for the rest of the world. For that reason I was really determined to
do a study concerning bees and their conservation. That is how I ended up in the bumble-bee lab of
Professor Smagghe and Ivan Meeus. I would like to thank my promotors for the opportunity to
conduct my thesis at their research group. Without Kevin I would never have accomplished this
thesis, he taught me a lot and guided me throughout the whole process. Furthermore, I thank my
Brazilian supervisors Betina and Patricia for welcoming me so warmly in Brazil and offering their
helping hand both personally and scientifically during my stay there. Finally, I am very grateful to be
surrounded by such great people as my family and friends, they supported and encouraged me
through the whole journey. Thank you!
2
Contents
Preface ..................................................................................................................................................... 1
List of abbreviations ................................................................................................................................ 4
Abstract ................................................................................................................................................... 5
Samenvatting ........................................................................................................................................... 7
Introduction ............................................................................................................................................. 9
1. Literature study ............................................................................................................................. 11
1.1 Bumblebee species of Southern Brazil .................................................................................. 11
1.2 Bumblebee decline: general .................................................................................................. 14
1.2.1 Introduction ................................................................................................................... 14
1.2.2 The role of genetics ....................................................................................................... 14
1.2.3 Previous research .......................................................................................................... 15
1.3 Bumblebee decline: the case of Southern Brazil ................................................................... 15
1.3.1 Introduction ................................................................................................................... 15
1.3.2 Local extinction of Bombus bellicosus ........................................................................... 16
1.4 Stingless bees and the genus Melipona ................................................................................ 17
1.4.1 Introduction ................................................................................................................... 17
1.4.2 Stingless bee reproduction ............................................................................................ 18
1.4.3 Stingless bee pollination ................................................................................................ 18
1.4.4 Melipona species of Southern Brazil ............................................................................. 19
1.5 Meliponiculture and its genetic consequences ..................................................................... 20
1.5.1 Introduction ................................................................................................................... 20
1.5.2 Melipona decline ........................................................................................................... 20
1.5.3 Meliponiculture for conservation.................................................................................. 21
1.5.4 Genetic consequences of artificial maintenance of Melipona colonies ........................ 22
1.6 Microsatellites as a molecular tool for population genetic research .................................... 23
1.6.1 Microsatellites ............................................................................................................... 23
1.6.2 Population genetic parameters ..................................................................................... 28
2. Material and methods ................................................................................................................... 30
2.1 Collection of samples ............................................................................................................ 30
2.2 DNA-extraction, primer selection, PCR, capillary electrophoresis ........................................ 32
2.2.1 DNA-extraction .............................................................................................................. 32
2.2.2 Selection MS primers ..................................................................................................... 32
3
2.2.3 Amplification & visualization ......................................................................................... 33
2.2.4 Genotyping .................................................................................................................... 34
2.3 Data analysis .......................................................................................................................... 35
2.3.1 Genetic diversity ............................................................................................................ 37
2.3.2 Population structure ...................................................................................................... 37
3. Results ........................................................................................................................................... 38
3.1 Melipona ................................................................................................................................ 38
3.1.1 Primer selection ............................................................................................................. 38
3.1.2 Dataset preparation ...................................................................................................... 39
3.2 Bombus .................................................................................................................................. 40
3.2.1 Dataset preparation ...................................................................................................... 40
3.2.2 Estimation of parameters for genetic diversity and inbreeding ................................... 43
3.2.3 Estimation of population structure ............................................................................... 51
4. Discussion ...................................................................................................................................... 53
4.1 Collection of Brazilian wild bees ............................................................................................ 53
4.2 Initialisation Melipona project .............................................................................................. 54
4.3 Genetic parameters of bumblebees in Rio Grande do Sul .................................................... 55
5. Conclusion & future perspectives ................................................................................................. 58
References ............................................................................................................................................. 59
Supplementary data .............................................................................................................................. 71
4
List of abbreviations
AFLP Amplified Fragment Length Polymorphism
AR Allelic richness
bp Base pairs
Dest Jost’ D
DNA Deoxyribonucleic acid
F1 First filial generation
F2 Second filial generation
FIS Inbreeding coefficient
FST Genetic differentiation : gentic structure values
HE Expected Heterozygosity
HO Observed Heterozygosity
HWE Hardy-Weinberg Equilibrium
MS Microsatellite
n Number
Ne Effective population size
nt Nucleotide
p Probability
PCR Polymerase Chain Reaction
RNA Ribonucleic acid
SE Standard Error
SSR Simple Sequence Repeat
Ta Annealing temperature
Taq Thermus aquaticus DNA polymerase
VNTR Variable Number Tandem Repeats
5
Abstract Worldwide there is a decline in pollinators going on. Also in South-Brazil, the bumblebee species
Bombus bellicosus has been reported to be extinct in the most north-eastern range of its distribution.
In literature, several drivers are described to explain the phenomenon of pollinator decline. Also
genetic factors can play a role, as low genetic variability can make bees more vulnerable to changes
in the environment. In order to understand whether low genetic diversity was historically inherent to
the endangered species B. bellicosus, or was caused by its recent decline, its population genetics
were studied in this research. A comparison was made with two other South-Brazilian bumblebee
species B. pauloensis and B. morio, which appear to be still abundant in this region and might thus be
less sensitive to the drivers of bee loss.
The second objective of this thesis was to start up a project regarding another genus of Neo-tropical
bees. Melipona stingless bees are important native pollinators in South Brazil. Unfortunately wild
populations of these species are disappearing. On the other hand bee keepers are trying to conserve
the bees by meliponiculture. To maintain managed populations viable, bee keepers introduce new
genetic material continuously. However, when colonies from distant populations, like from other
states, are illegally introduced, this could lead to a gene flow which would not be possible without
human intervention. To uncover whether such an exchange of genetic material truly took place, the
allelic profile of the Melipona species of Rio Grande do Sul can be studied through time and
compared to populations of the same species from possible introducing states. When a change in
genetic profile of Rio Grande do Sul populations can be observed through time, it could possibly be
attributed to the introduction of hives from other states.
To gather the necessary data for this research, a trip to Porto Alegre in the South-Brazilian state Rio
Grande do Sul was undertaken. Pin-mounted specimens of the Bombus and Melipona species were
sampled from several museum collections. Additionally, entomological field work was done to collect
individuals of wild bee populations. Unfortunately, this year’s extreme weather conditions in South-
East Brazil prevented the observation of bees in general, and mainly of the rare bumblebees species.
For the bumblebee species, tissue samples of populations from 8 locations and 4 time periods,
covering 5 decades, were finally collected. For the Melipona species: M. bicolor schencki, M.
quadrifasciata quadrifasciata and M. torrida also several populations from the states Rio Grande do
Sul, Santa Catarina and São Paulo were sampled.
To genetically study the collected populations of Brazilian wild bees, the microsatellite marker
technology was used. For the Bombus species a set of microsatellite primers, previously established
for European bumblebee species, was applied to the Brazilian species. For Melipona a new set of
microsatellite primers were assembled from literature and tested. The first step of this technology is
extraction of DNA from the tissue samples. Next the microsatellite loci were amplified and
fluorescently labelled by the polymerase chain reaction and visualised with capillary electrophoresis.
From the obtained profile of fragment lengths, the alleles of each microsatellite locus could be
determined from each individual.
Finally, from the genotyping data several population genetic parameters could be estimated for the
three Bombus species. Genetic diversity, inbreeding and population structuring were determined for
all populations and compared between time periods and species. Some conclusions regarding the
genetic status of the bumblebee species in Rio Grande do Sul could cautiously be made. Since the
6
1950s, the genetic diversity of B. pauloensis has gradually decreased, while such a trend could not be
observed for B. morio. In general, and overall studied time periods, no difference in genetic
variability of the two presumably stable Brazilian bumblebee species could be observed. For the
declining species B. bellicosus, insufficient data could be assembled to draw any conclusion, but its
absence on the other hand is also an indication for its disappearance and thus loss of genetic
diversity.
7
Samenvatting Wereldwijd wordt er tegenwoordig een achteruitgang van bestuivers waargenomen. Ook in het
zuiden van Brazilië, is de hommelsoort B. bellicosus gerapporteerd als uitgestorven in het meest
Noord-Oostelijke gedeelte van zijn distributie. In de literatuur worden verschillende oorzaken voor
de achteruitgang van bestuivers beschreven. Ook genetische factoren kunnen hierin een rol spelen,
zo zal een lage genetische diversiteit populaties altijd gevoelig maken voor veranderingen in de
omgeving. Om in te kunnen schatten of een lage genetische diversiteit al van oudsher aanwezig was
in de bedreigde soort B. bellicosus, of dat die eerder veroorzaakt is door de recente achteruitgang
van de soort, werd in dit onderzoek de populatie genetica bestudeerd. Tegelijk werd ook een
vergelijking gemaakt met twee andere hommelsoorten uit Zuid-Brazilië, namelijk B. pauloensis en B.
morio. Deze soorten lijken nog steeds abundant aanwezig te zijn in deze regio en dus minder gevoelig
te zijn voor de oorzaken van het verlies aan bijenpopulaties.
De tweede doelstelling binnen deze thesis was het opstarten van een project rond een ander genus
van wilde bijen uit de Neotropische regio. Melipona ‘stingless bees’ zijn een groep van belangrijke
inheemse bestuivers in het zuiden van Brazilië. Helaas gaan ook veel populaties van deze bijen er op
achteruit in het wild of verdwijnen ze in sommige gevallen zelfs. Om dit tegen te gaan, proberen
imkers deze bijensoorten in stand te houden door de teelt, ook wel ‘meliponiculture’ genaamd. Om
de gedomesticeerde populaties sterk te houden, trachten de bijenhouders voortdurend nieuw
genetisch materiaal te introduceren. Echter, wanneer kolonies van verder gelegen gebieden, zoals
naburige staten, illegaal worden geïntroduceerd, zou dit kunnen leiden tot uitwisseling van genetisch
materiaal die zonder menselijke tussenkomst nooit mogelijk zou zijn geweest. Om te ontdekken of
zo’n uitwisseling van genetisch materiaal werkelijk heeft plaats gevonden, kan het genetische profiel
van de Melipona soorten uit Rio Grande do Sul bestudeerd worden doorheen de tijd. Tegelijk kan dat
genetisch profiel vergeleken worden met die van populaties van dezelfde soorten, afkomstig van de
staten waaruit de introductie hoogstwaarschijnlijk heeft plaatsgevonden. Indien een wijziging in het
genetische profiel van de populaties uit Rio Grande do Sul kan worden waargenomen, zou die
mogelijks veroorzaakt kunnen zijn door de introductie van genetisch materiaal uit die naburige
staten.
Om de nodige data voor dit onderzoek te verzamelen, werd er naar Porto Alegre, een stad in de Zuid-
Braziliaanse staat Rio Grande do Sul, afgereisd. Daar werden uit museum collecties stalen van
vastgespelde specimens van de Bombus en Melipona soorten genomen. Daarnaast werd er ook
veldwerk uitgevoerd, in de hoop ook de huidige wilde populaties te kunnen bemonsteren. Dankzij
het extreem vochtige weer dit jaar in Zuid-Oost Brazilië, was het moeilijk tot zelfs onmogelijk om
bijen en hommels te observeren, laat staan ze te verzamelen. Voor de hommelsoorten kon
uiteindelijk weefsel verzameld worden voor specimens van populaties afkomstig uit 8 verschillende
locaties en 4 verschillende tijdsperiodes (die samen ongeveer 50 jaar beslaan). Voor de Melipona
soorten M. bicolor schencki, M. quadrifasciata quadrifasciata and M. torrida werden ook
verschillende populaties van de staten Rio Grande do Sul, Santa Catarina en São Paulo bemonsterd.
Om de genetische status van de verzamelde populaties van Braziliaanse wilde bijen te bestuderen,
werd gebruik gemaakt van de microsatelliet marker technologie. Voor de hommelsoorten werd een
combinatie van microsatelliet primers, oorspronkelijk ontwikkeld voor Europese hommelsoorten,
succesvol toegepast op de Braziliaanse soorten. Voor de Melipona soorten moest eerst en vooral een
goede set primers samengesteld en uitgetest worden. Na DNA-extractie werden de microsatelliet loci
8
geamplificeerd en fluorescent gelabeld via de Polymerase Chain Reaction en vervolgens
gevisualiseerd met capillaire elektroforese. Het resultaat hiervan was een profiel voor de
verschillende geamplificeerde fragmenten, waaruit de lengte van de verschillende allelen van elk
locus kon afgeleid worden.
Tenslotte konden na het genotyperen van alle individuen, verschillende populatie genetische
parameters geschat worden voor de hommelsoorten. Genetische diversiteit, inteelt en populatie
structuring werden bepaald voor alle populaties en vergeleken tussen tijdsperiodes en soorten. Een
aantal preliminaire conclusies konden na dit onderzoek getrokken worden voor de hommelsoorten
van Rio Grande do Sul. Sinds de jaren ’50, is de genetische diversiteit van B. pauloensis geleidelijk
gedaald, maar een dergelijke trend kon niet geobserveerd worden voor B. morio. Anderzijds was het
verschil in genetische diversiteit tussen de twee als stabiel voorgestelde soorten over het algemeen
niet verschillend. Voor de bedreigde soort B. bellicosus is conclusies trekken moeilijk, aangezien
slechts geringe data kon worden verzameld. Anderzijds kan dit laatste echter ook een indicatie zijn
voor het verdwijnen van deze soort en is het een duidelijk verlies van genetische diversiteit.
9
Introduction ‘If the bee disappeared from the surface of the earth, man would have no more than four years left to
live. No more bees, no more pollination, no more plants, no more animals, no more man.’ Whether
Albert Einstein ever stated this, or Charles Darwin or the Belgian Maurice Maeterlinck, and whether
this statement is correct, remains unclear. However, the importance of bees and their current
worldwide declines cannot be denied.
Bumblebees are essential pollinators all over the world in natural and managed ecosystems. These
fluffy animals mainly occur in the temperate areas of North America and Eurasia, although some
species also inhabit South-America. In Brazil, seven bumblebee species are known, of which B. morio,
B. pauloensis and B. bellicosus. For this last one there is evidence of local extinction. The Brazilian bee
fauna is endangered due to a continuous process of urbanization and agricultural development. In
the Neotropical region, the loss of habitat due to deforestation and the introduction of exotic species
comprise the main threats.
Other native bees of South America that suffer from bee decline, are the stingless bees of the genus
Melipona. They represent one of the most important insects for pollination in natural and cultivated
areas in the Neotropical region. Some of these Melipona species are endangered in Brazil. For the
state Rio Grande do Sul, M. bicolor schenki and M. quadrifasciata quadrifasciata are classified as ‘in
danger’ and M. obscurior as ‘vulnerable’ (Marques et al., 2002). A measure to prevent local
extinction is meliponiculture, which also is an encouraging economic practice supposed to be in
alignment with sustainable development.
Pollination of wild flowers and agricultural crops by insects is an important ecosystem service.
Unfortunately, just like the given examples, many bee species are facing the threat of population
decline (Williams & Osborne, 2009). Different potential drivers have been suggested to cause this
phenomenon, such as: habitat degradation and fragmentation, agricultural intensification,
unselective use of agrochemicals and pesticides, and introduction/spread of exotic species. They all
result in bee populations becoming smaller and isolated, which finally increases their risk to
disappear. Genetic drift and reduced gene flow cause a lower genetic diversity within these small
populations. Furthermore, they become more vulnerable to inbreeding and consequently inbreeding
depression. The thereby caused further diminished genetic diversity makes populations less able to
adapt to current and future changes in the environment, which can lead to extinction ultimately.
In order to protect biodiversity and avoid the extinction of such endangered bee species,
conservation genetics are of great importance. Population genetics investigate genetic processes that
take place in populations. Microsatellites are a very useful molecular tool to address such population
genetic questions, because of their exceptional polymorphism in natural populations. Microsatellites
are DNA tracts consisting of simple tandemly repeated fragments. They suffer from higher rates of
mutation compared to the rest of the genome. Thus, at the same locus, the population as a whole
contains several alleles, each with a different number of repeats. Furthermore these DNA regions can
easily be identified and therefore microsatellites are popular markers for research similar to what is
conducted in this thesis.
In the following sections, first a summary of the available relevant literature is given. Then an
overview of the used materials and methods explains how this research was conducted and
10
thereafter the results are displayed. Finally some observations, reflections and tentative conclusions
are reported including some perspectives for the future.
11
1. Literature study
1.1 Bumblebee species of Southern Brazil Bombus is a genus of the family Apidae distributed worldwide (approximately 250 known species
(Goulson et al., 2008)), principally in temperate areas of North America and Eurasia (Maggi et al.,
2011) as it is a typical Holarctic group (Sakagami & Laroca, 1971). On the other hand, for the
Neotropical region, 42 species occupy a variety of habitats, ranging from sea level to about 4,400 m
in the Andes (Abrahamovich & Díaz, 2002). In South America, most of the Bombus species are
distributed along the Andes and in temperate regions, with only a few species recorded in the warm
lowlands (Santos Júnior et al., 2015).
Seven species of bumblebees are known to occur in Brazil, six of them belonging to the subgenus
Fervidobombus Skorikov, 1922 (Santos Júnior et al., 2015) or Thoracobombus Dalla Torre, 1880
(Williams et al., 2008), depending on which classification one adopts (Santos Júnior et al., 2015), and
one to Robustobombus (Martins & Melo, 2010). Five of these species from the Fervidobombus
subgenus (B. bellicosus Smith (1879); B. brasiliensis Lepeletier (1836); B. brevivillus Franklin (1913); B.
pauloensis Friese (1913); and B. transversalis Olivier (1789)) seem to be very closely related, while
the sixth one, B. morio Swederus (1787) belongs to a distinctive clade in the same subgenus (Santos
Júnior et al., 2015).
The low number of Bombus species found in Brazil, despite their wide distribution, is very contrasting
to the high number found in other world regions as the Palaearctic (120 species), Orient (108
species), and Japan (23 species) (Williams, 1998). The low number of species may be a consequence
of ecological and environmental features, among other factors (Françoso et al., 2015). The tropical
climate may facilitate dispersion allowing gene flow even in large areas, which would mean that the
species integrity can be maintained (Françoso et al., 2015). However, the low number of Bombus
species presenting large geographic distribution in Brazil may be unreal and consequently cryptic
species may be still unrecognized (Françoso et al., 2015).
Brazilian species present high intra-specific variability, and parallelism in colour patterns is observed
in sympatric species (Françoso et al., 2015). Françoso et al. (2015) elaborated a key for identification
of Brazilian Bombus species, including a new species endemic to Brazil, B. (Thoracobombus)
applanatus Oliveira, Françoso & Arias, sp. nov..
Originally some bionomic differences between the (sub)tropical lowland species and the northern
temperate species were assumed. Ihering (1903) allocated to the (sub)tropical lowland species
ecological characteristics like perennial colonies with pleometrotic association and establishment of
new polygynous colonies by swarming (Sakagami & Laroca, 1971). Later, Sakagami & Laroca (1971)
considered an annual cycle, as for north temperate species. In southern Brazil, B. bellicosus has
annual colonies (in which new queens hibernate through winter) (De Paula & Melo, 2015). In
contrast B. pauloensis colonies may persist for more than a year (Sakagami & Moure, 1962), as noted
by continuous appearance of workers, in the state of São Paulo (Garófalo, 1974). Caste and sex
emergence is seasonal. In both B. pauloensis and B. bellicosus, queens are mainly collected from
August to November, indicating activities of colony founding and foraging in spring before the
appearance of workers. Queens collected during autumn from March to May represent newly
emerged individuals, as suggested by their relatively fresh wings. Appearance of males coincides well
12
with the new queens. After mating, the new queens become quiescent and move to small cavities
toward the end of the summer when the maternal colony dies (Sakagami, 1976). The period of
dormancy of the queen can last 6 to 9 months. It ends due to rising temperatures in the spring, when
queens begin to establish their new colony individually (Beekman et al., 1998). Nests of all Brazilian
Bombus (Fervidobombus) species are built in cavities in the soil and covered with a layer of plant
detritus (Moure & Sakagami, 1962). Workers are most active in the months between the two periods
of queen activity. Differences in longevity were also observed in the workers of temperate and
tropical Bombus species (Silva-Matos & Garófalo, 2000). The active period of tropical bumblebee
species can be prolonged, as some colonies can survive winter, while in northern areas the
temperate species have normally only a short annual life cycle.
Most of the Northern Hemisphere species are polylectic (i.e. visiting a wide range of plants and
flowers). Moure & Sakagami (1962) suggested a similar behaviour for the Brazilian species
(Abrahamovich et al., 2001). By studying Bombus species and their associated flora in Argentina,
Abrahamovich et al. (2001) suggest that B. pauloensis visits the widest range of plants and that the
most visited plants by Brazilian bumblebee species were Fabaceae, Asteraceae and Solanaceae. B.
pauloensis has a longer proboscis, possibly obtaining nectar easier and faster than B. bellicosus,
which has a shorter proboscis (Arbulo et al., 2011).
In southern Brazil (Moure & Sakagami), where climate is predominantly moist subtropical (Cfa in
Köppen classification), three species of bumblebees, B. bellicosus Smith, B. morio Swederus and B.
pauloensis Friese, are found (De Paula & Melo, 2015). In terms of habitats, Bombus morio and
Bombus pauloensis are generalists (Laroca & Orth, 2002). Both species can be found from Rio Grande
do Sul, in the south, to southern Bahia, Goiás and Mato Grosso, in the north. According to Maggi et
al. (2011) B. pauloensis is the most abundant and dispersive species, presenting a higher climatic and
altitudinal tolerance and a broader distribution in regard to the other bumblebee species.
1. Bombus pauloensis Friese, 1913
Bombus pauloensis Friese (1913), subgenus Fervidobombus, formerly named B. atratus Franklin
(1913) (Moure & Melo, 2012), is widespread in South America, with a distribution from north-
western South America to south-eastern Brazil, Uruguay, and central Argentina (Colombia,
Venezuela, Ecuador, Peru, Bolivia, Paraguay, Uruguay, Argentina, and found in the Brazilian
States: Bahía, Espirito Santo, Goiás, Mato Grosso, Mato Grosso do Sul, Minas Gerais, Paraná, Rio
de Janeiro, Rio Grande do Sul, Rondônia, and Santa Catarina) (Abrahamovich et al., 2004).
In Brazil, this species is occurring in sympatry with B. morio and B. bellicosus (Moure & Sakagami,
1962; Françoso & Arias, 2012). 50 years ago, B. bellicosus outnumbered B. pauloensis at some
places, but these days B. pauloensis is generally the most abundant species (Sakagami & Laroca,
1971). B. pauloensis has a black body with yellow stripes, but is polytypic with the colour varying
from completely black to distinctly yellow banded (Sakagami & Laroca, 1971). This body pattern
presents a colour gradation that follows the species latitudinal gradient of distribution, whereas
in the south they show yellow stripes and become darker in the north (Moure & Sakagami, 1962;
Françoso & Arias, 2012). B. pauloensis is larger in average than B. bellicosus (13.6 mm and 10.6
mm, respectively) (Cortopassi- Laurino et al., 2003).
13
This species is considered an important pollinator of Brazilian native flora and also shows
potential as pollinator for economic crops (Camillo & Garófalo, 1989). B. pauloensis is associated
with plant species from 29 families in Argentina (Abrahamovich et al., 2001). The largest number
of visited species corresponds to the families Asteraceae and Fabaceae (Abrahamovich et al.,
2001).
Particular for B. pauloensis new queens can supersede the foundress, and new colonies may be
initiated by swarming in the same way as has been described for honeybees (Goulson, 2010). If
the queen disappears or is removed from her colony, she will be succeeded by a mated worker,
the false queen, which produces both female and male offspring and maintains colony
development until the reproductive phase when new queens are produced (da Silva-Matos &
Garófalo 2000). Silva-Matos & Garófalo (1995) showed that a new colony can be started from a
group of queenless workers in the presence of some brood (da Silva-Matos & Garófalo 2000). In
this case, the colony development is guaranteed by the appearance of a false queen, who takes
over the queen's role. This is in contradiction with what is known for workers of all bumblebee
species of temperate areas. Indeed, in these species, workers do not have the capacity to mate
or produce fertilized offspring (Goulson, 2010). Although one worker can also take over the role
of the queen, when the queen is removed or killed, she will become a so called ‘pseudo-queen’,
an unmated worker which will produce only unfertilized eggs (Goulson, 2010).
2. Bombus morio Swederus, 1787
Bombus morio is present in Brazil, Bolivia, Colombia, Ecuador, Paraguay, Peru, Uruguay,
Venezuela and northern Argentina (Abrahamovich et al., 2004). In Brazil, B. morio is distributed
in a large area in the south and southeast (Moure & Sakagami, 1962). This species prefers
warmer areas with trees (Sakagami & Laroca, 1971). Besides its ecological importance, B. morio
can be a good model for biogeography and evolutionary studies due to its high dependence on
forest (Françoso et al., 2012). B. morio has a high dispersal ability and the capacity to survive in
urban environments (Francisco et al., 2015). In Argentina B. morio is mainly associated with plant
species from Asteraceae, Fabaceae and Solanaceae (Abrahamovich et al., 2001). B. morio is
completely black (Françoso et al., 2012). More information on the biology of this species is not
accessibly available in literature.
3. Bombus bellicosus Smith, 1879
Bombus bellicosus is present in Uruguay, southern Brazil and most of Argentina (Abrahamovich
et al., 2001). This species reaches its northern limit in the Brazilian state, Parana, where it was
relatively abundant until the early 1980s, but is now assumed to be locally disappeared there. B.
bellicosus has been found in sympatry with B. pauloensis and B. morio in most regions of its
distribution, having the most restricted distribution of the 3 species (Moure & Sakagami, 1962).
Originally B. bellicosus is found in areas covered with open native grass fields and shrub
vegetation (Martins & Melo, 2010). In Argentina B. bellicosus mainly visits plant species
belonging to the families Asteraceae and Fabaceae (Abrahamovich et al., 2001).
The biology of B. bellicosus is also poorly known. B. bellicosus is a surface ground-nesting bee,
which requires only soil and plant detritus to build its nest (Varela 1992 a, b). Varela (1992 a, b)
reported a wax layer covering the brood chamber after observations on a nest of B. bellicosus,
14
found in Uruguay. The production of this wax cover, probably an adaptation against low
temperatures, is present in Bombus species inhabiting cooler climates. This unique trait among
tropical and subtropical Bombus species, may link B. bellicosus to temperate and sub-temperate
climates (Sakagami et al. 1967).
1.2 Bumblebee decline: general
1.2.1 Introduction
Pollination of wild flowers and agricultural crops by insects is an important ecosystem service.
Bumblebees are essential pollinators all over the world in natural and managed ecosystems. Contrary
to honeybees, they are capable of buzz pollination and they are considered more efficient pollinators
than honeybees (Apis mellifera) in greenhouses (Velthuis & van Doorn, 2006; Goulson, 2010).
However, currently pollinators are undergoing declines worldwide (Potts et al., 2010). Unfortunately
also many bumblebee species are facing the threat of population decline (Williams & Osborne, 2009).
Different potential drivers have been suggested to cause this phenomenon:
Change in land-use together with the loss and fragmentation of habitats by urbanization and
agricultural development, leading to decreased resource (food, nesting …) availability and
diversity (Biesmeijer et al ., 2006; Goulson et al., 2008);
Use of pesticides (Rortais et al., 2005);
Introduction of non-native species and the associated spread off pathogens (Neumann &
Carreck, 2010);
Climate change (Williams et al., 2007).
1.2.2 The role of genetics
Due to these different drivers, bumblebee populations become smaller and isolated. These
populations will finally risk to disappear due to the impact of genetic causes. Indeed, genetic drift
causes a reduced genetic diversity within these populations and small populations are more
vulnerable to inbreeding and consequently inbreeding depression (Frankham, 2005; Zayed, 2009;
Goulson, 2010). The thereby caused lower genetic diversity makes populations less able to adapt to
current and future changes in the environment, which can lead to extinction ultimately (Frankham,
2005; Zayed, 2009). Furthermore in fragmented populations dispersal or gene flow is limited, which
makes it rather impossible for extinct patches to be recolonized, and an increase of genetic diversity
from neighbouring populations is inhibited (Goulson, 2010).
The rate of genetic drift, is determined by the effective population size, Ne. As Ne depends on the
number of egg-laying queens and their mates from each individual colony (the colonies reproductive
success) and not the total number of individuals within a colony, this parameter will be low in social
insects. For bumblebee species, Ne is even very low, because colonies usually have only one queen
and are mostly monoandrous (Estoup et al., 1995; Schmid-Hempel & Schmid-Hempel, 2000). For
tropical Bombus species, only one study (Garofalo et al., 1986) described the polyandrous mating
behaviour of B. pauloensis with up to three matings per queen. However, since these observations
were conducted under highly artificial experimental conditions, it remains unclear whether this also
occurs under natural conditions (Huth-Schwarz et al., 2010).
Gene flow depends greatly on dispersal ability. The dispersal ability of bumblebees differs between
species, with the difference varying according to the study. For populations of species with a more
15
limited dispersal rate, repopulation of a local extinction will be less likely. These species will also be
more vulnerable to inbreeding and inbreeding depression. Within the South-Brazilian bumblebee
species, Bombus morio is known to have a high dispersal rate and this for both males and females
(Francisco et al., 2015). It is likely that B. morio has a strong flight capability (Moure & Sakagami,
1962), but the dispersal range of the reproductives is currently unknown (Francisco et al., 2015).
In Hymenoptera fertilized eggs will develop into females and unfertilized eggs usually in haploid
males. However, bumblebee sex is actually determined by a single sex-determining locus (Cook &
Crozier, 1995). Unfertilized haploid eggs are hemizygous and will all develop in males. Diploid eggs,
which are heterozygous at the sex-determining locus will develop in females, while bees homozygous
at this locus will develop in diploid males (Duchateau et al., 1994). A decrease in polymorphism of
this sex determination locus can lead to a greater chance of producing sterile diploid or triploid
males. Inbreeding can cause this decrease in polymorphism, due to an increase in the probability of
matched-pair matings at the sex locus. The presence of diploid males affects the colony negatively,
as half of the workers will be replaced by diploid males. These males don’t contribute to colony tasks
and have a low fertility (Duchateau & Marien, 1995). Hence, inbreeding can result in some negative
population effects, in general called inbreeding depression, because of the expression of deleterious
recessive alleles (Frankham, 2005; Zayed, 2009). Although haplo-diploid organisms like bumblebees,
are considered to be less sensitive to genetic pauperization than diploid species, because deleterious
alleles are purged from the population in the haploid males (Sorati et al., 1996; Packer & Owen,
2001).
1.2.3 Previous research
Several population genetic studies on bumblebee species in Europe (e.g. Charman et al., 2010),
North-America (Cameron et al., 2011, Lozier et al., 2011) and Japan (Takahashi et al., 2008) have
been published. Most of this research concluded that the genetic diversity parameters observed in
populations of declining bumblebee species were lower than those in populations of more stable
bumblebee species. However, all these studies were based on recent bumblebee specimens. Only in
2009, Lozier & Cameron compared the genetic variation between recent and historical populations in
North-America of the declining and stable bumblebee species, B. pensylvanicus and B. impatiens
respectively. Recently, Maebe et al. (2015) observed in populations of declining bumblebee species
in the Netherlands a significantly lower genetic diversity before their major declines than in co-
occurring stable species. So is the low genetic diversity caused by population decline and/ or recent
bottlenecks or is it inherent to historical, pre-decline differences in genetic variation among species?
1.3 Bumblebee decline: the case of Southern Brazil
1.3.1 Introduction
With the exception of a report of the local extinction of a bumblebee species in South Brazil (Martins
& Melo, 2010), population genetic studies on the South-American bumblebee species (B. pauloensis,
B. morio and B. bellicosus) have barely been carried out.
In the Neotropical region, the loss of habitat, due to deforestation or agricultural development, and
the introduction of exotic species comprise the main threats to the bee fauna (Freitas et al., 2009).
The high diversity of the Brazilian bee fauna, with about 1,700 species (Melo, 2007), is endangered
due to a continuous process of urbanization and agricultural development (Martins et al., 2013).
Although the decline of European and North-American bees and bumblebees is documented widely,
16
not much information is available on changes in the status of bumblebee populations in tropical and
subtropical regions. B. bellicosus is one of the rare cases of reported threatened bumblebees in South
America (Martins et al., 2015). Also B. dahlbomii Guérin-Meneville, 1835 was reported as being
replaced by the invasive species B. terrestris (Linnaeus, 1758) in Chile and Argentina (Morales et al.,
2013).
1.3.2 Local extinction of Bombus bellicosus
Martins & Melo (2010) presented evidence for the local extinction of B. bellicosus in the
northeastern range of its distribution, in the Brazilian state Parana. The northern limit of B. bellicosus
in Brazil is found in the so called Parana’s 2nd plateau. The 2nd plateau has large tracts of flatlands,
which were originally covered with open grasslands. B. bellicosus was also found in the areas covered
by native grasslands near the city of Curitiba, in Parana’s 1st plateau. In a survey of bees from a grass
field site at the 1st plateau, conducted in the years of 1962-1963, B. bellicosus was the most
abundant native bee species (Sakagami et al., 1967). Twenty years later (1981-1982), a second survey
by Bortoli & Laroca (1990) in the same site revealed a slight decline in the abundance of this
bumblebee, then the third most abundant native species (Martins & Melo, 2010). A third survey, in
2004-2005 (Melo et al., 2006), failed to recover any specimen of B. bellicosus at this site. The last
specimens collected in the 1st plateau are those reported by Bortoli and Laroca (1990) (Martins &
Melo, 2010). B. bellicosus has also disappeared from Parana’s 2nd plateau. In a survey conducted
during 2 years (2002 - 2004) in sites of native grass fields, only B. pauloensis and B. morio were found
on flowers (Gonçalves & Melo, 2005; Gonçalves et al., 2009). The most recent record of the species
in Parana’s 2nd plateau was of a queen collected in February 1980.
According to Martins & Melo (2010), several causes might be involved in the local extinction of B.
bellicosus in Parana. Habitat conversion has probably caused the strongest impact. Indeed, the native
grass fields have been extensively destroyed during the process of urbanization and the areas
covered with grass fields have almost disappeared. The human population in the area grew, and no
conservation reserves remained available in this region. In the 2nd plateau, the region was occupied
with extensive cattle ranching, using the native grasses for grazing, for most of the last 200 years.
More recently, most of the native grass fields have been converted into agricultural land.
The whole region has also received intense pollution from pesticides used for agriculture. If pollution
has played a major role in the disappearance of B. bellicosus from Parana, one should expect it to
have also impacted the other bumblebee species in the area. However, populations of the two
sympatric species, B. morio and B. pauloensis, apparently have not been impacted so strongly as B.
bellicosus in Parana.
Compared to other bumblebee species, B. bellicosus does seem able to thrive in agricultural
landscapes. B. bellicosus is known to be polylectic, using more than 60 plant species from 18 families
as nectar and pollen sources (Martins & Melo, 2010: Bortoli & Laroca, 1990). This species thus
probably does not require any special resource as food source, nor for their nesting as described by
Varela (1992a, b) (Martins & Melo, 2010). Therefore, habitat and resource limitation may be not the
major causes of extinction of this species in Parana.
Global warming, the third putative cause for the extinction of B. bellicosus in Parana, is currently
considered one of the most important factors involved in species extinction (Parmesan & Yohe,
2003), and range and phenology changes (Parmesan, 2006). Considering that this species reaches its
17
northernmost limit in Parana (see before), it is likely that recent changes of temperature might have
strongly affected these populations. Martins et al., 2015 show that the suitable climatic conditions
for B. bellicosus will retreat southwards. Populations living at the edge of the species distribution are
considered to have a high chance of extinction (Doherty et al., 2003), in particular due to their lower
densities. In the case of B. bellicosus, the historical data indicate that the species was locally
abundant up to the early 1980s, suggesting that density alone would not make it more prone to local
extinction. Current data indicates that a trend of accelerated global warming started in the second
half of the 1970s (Hansen et al., 2006). It is interesting to note that this pattern coincides with the
last known records of the species in the region being from the period of 1980-1982. Furthermore,
additional evidence for a strong role of global warming in affecting B. bellicosus derives from the
historical increase in relative abundance of the two sympatric Bombus species in Parana. Since these
species are more widely distributed, both reaching more northern latitudes than B. bellicosus, they
should be better able to tolerate increases in temperature. A wax cover inside the nests is usually
related to Bombus species inhabiting cooler climates (Martins et al., 2015). This cover enables the
maintenance of higher temperatures inside the nest and may be deleterious for species like B.
bellicosus under future warmer climates.
Considering the available evidence, it is not possible to conclude a single cause for the extinction of
B. bellicosus in Parana. Future studies should aim to obtain a more detailed mapping of the
populations in Brazil and an evaluation of their current conservation status. Although more precise
data is lacking, Martins & Melo (2010) argue for adding B. bellicosus to future editions of the Brazilian
list of threatened animal species.
In this thesis the genetic diversity of B. bellicosus, of which the distribution is declining, will be
compared trough time with the genetic diversity of 2 more stable bumblebee species (B. morio en B.
pauloensis). Microsatellites will be used to genotype specimens.
1.4 Stingless bees and the genus Melipona
1.4.1 Introduction
Melipona (Hymenoptera: Apidae) is a strictly Neotropical genus of stingless bees. It comprises more
than sixty species in Brazil (Camargo & Pedro, 2013). They represent one of the most important
insects for pollination in natural and cultivated areas in this region (Heard, 1999; Slaa et al., 2006).
Stingless bees are eusocial insects. All species produce honey, which is stored in large egg-shaped
pots and has a high water content (Ferreira Junior et al., 2010). This honey has been appreciated by
humans since ancient times. In all Latin American countries, stingless bee honey is used both as
medicine and a sweetener (Cortopassi-Laurino et al., 2006). Plant resins are used by the stingless
bees for construction and hygienization of their nests (Ferreira Junior et al., 2010; Roubik, 2006). The
resins can also be mixed with mud collected by the bees, forming batumen or geopropolis, which is
applied to seal cracks (Ferreira Junior et al., 2010), giving greater protection to the nests. The mixture
of propolis with wax, secreted by the workers from their abdominal glands, forms cerumen, the
material that they use to build brood combs, food pots and the envelope that protects the combs
(Ferreira Junior et al., 2010). The cerumen of stingless bee nests can also be mixed with plant resin
(Cortopassi-Laurino et al., 2006).
18
1.4.2 Stingless bee reproduction
Stingless bees make their nests mainly in tree trunk cavities in forests, but some species have
subterranean nests (Cortopassi-Laurino et al., 2006). The bees enter and leave their nests via an
opening through which only one bee can pass at a time (Ferreira Junior et al., 2010). The entrance of
the nests of some stingless bee species, e.g. Trigona, allows more than one bee to pass. Colonies of
stingless bees are perennial (Heard, 1999) and typically headed by one single-mated queen (Peters et
al., 1999). The main exception to this pattern is found in M. bicolor, which has been discovered to
exhibit facultative polygyny. In this species, several queens may coexist and share reproduction
inside the colony for considerable periods of time (Alves et al., 2011b; Velthuis et al., 2006). Besides,
some anecdotal reports of temporary, transient episodes of polygyny are described for a few other
stingless bee species (e.g. in M. scutellaris) (Carvalho-Zilse & Kerr, 2004), which are usually associated
with queen replacement events (supersedure) (Alves et al., 2011b). Alves et al. (2011b) reported also
a new case of occasional polygyny in the stingless bee M. quadrifasciata. Thus polygyny might not be
so uncommon in Melipona, given the high levels of queen (over)production in this genus (Alves et al.,
2011b; Santos-Filho, 2006), caused by larval caste self-determination (Bourke & Ratnieks, 1999;
Wenseleers et al., 2004). This overproduction of new queens in great excess of colony needs is in
contrast to honeybees (Wenseleers et al., 2011b). If several such queens can simultaneously seize
the chance to start reproducing, this will evidently provide them with large individual fitness
benefits. Nevertheless, it is unlikely to raise the productivity of the colony as a whole, given that in
stingless bees, the cell building rate and not queen fecundity limits total colony productivity
(Velthuis, 2006). After hatching, the new queens are usually killed or excluded if a vigorous queen is
already present in the nest. The excess production of queens exerts much greater selection on
queens to seek alternative reproductive options. New queens of Melipona bees, can found colonies
not only via swarming or supersedure (substitution of the old queen), but also by infiltrating and
taking over other unrelated nests (Wenseleers et al., 2011b). This is a phenomenon of social
parasitism where upon death of the mother queen, colonies are invaded by unrelated queens that fly
in from unrelated hives nearby (Van Oystaeyen et al., 2013).
Unlike honeybees, which seal their brood cells only at pupae stage, stingless bees seal their mass
provisioned brood cells at immediately after egg deposition (Alves et al., 2011a). As a result, adult
stingless bee workers are unable to eliminate diploid males at an early stage (Alves et al., 2011a). For
M. quadrifasciata it has been shown that queens, which had made a matched mating, were killed by
the workers within 6-30 days after the first diploid males began to emerge (Alves et al., 2011a).
1.4.3 Stingless bee pollination
Nowadays stingless bee keeping is mainly a non-commercial small-scale business, although a few
large-scale stingless beekeepers exist. Stingless bees and Melipona in particular, have an important
ecological role as pollinators of many wild plant species but also seem good candidates for future
alternatives in commercial pollination of greenhouse crops (Slaa et al., 2006; Nunes-Silva et al., 2013;
Bomfim et al., 2014). Melipona bees are capable of buzz pollination, ejecting pollen grains by
vibration of the pollen-bearing anthers of flowers that erupt pollen through pores (Buchmann, 1983;
Heard, 1999). As they are pollinators of a wide variety of wild plants that have flowers with poricidal
anthers, they are of great importance for the conservation of natural biodiversity in the tropics (Alves
et al., 2011a). Several studies have reported on the pollination effectiveness of M. quadrifasciata in
tomato grown under enclosed conditions in Brazil (dos Santos et al., 2009; Bartelli et al., 2014). These
stingless bees are easily domesticated (harmless) and show various useful behavioral traits such as
19
recruitment of foragers (Nieh, 2004), perenniality, high flower constancy, great diet-breadth
(polylectic), and easy adaptation to new plant species (Ramalho et al., 1994; Heard, 1999; Slaa et al.,
2006). Stingless bees can also recruit nestmates to interesting food resources, with a wide range of
strategies (Nieh et al., 2003a). Melipona species may possess functionally referential communication,
the ability to transform environmental information into specific, abstract coded signals (Hrncir et al.,
2000). Melipona bees are the only animals, other than honeybees, in which such potential spatial
coding has been reported (Nieh & Roubik 1998; Aguilar & Bricenõ, 2002). M. bicolor can also
communicate height when the food source is at the forest canopy level, which is extremely
important as there the major food sources occur in the tropics (Nieh et al., 2003b). It is unclear
whether location-specific recruitment provides an adaptive benefit by significantly contributing to
foraging in Melipona colonies (Nieh et al., 2003a).
1.4.4 Melipona species of Southern Brazil
Only three species of Melipona occur in the southern region of the Atlantic forest, including the
inland forests of the Parana river basin (Melo, 2013). The southern populations of all three species
exhibit differences from their northern counterparts and have traditionally been treated as: M.
quadrifasciata quadrifasciata Lepeletier, 1836, M. bicolor schencki Gribodo, 1893 and M. marginata
obscurior Moure. Camargo & Pedro (2013) changed Moure’s subspecies of M. marginata to species
status, but maintained the other two taxa as subspecies.
1. Melipona (Melipona) quadrifasciata quadrifasciata Lepeletier, 1836 (Camargo & Pedro,
2013)
Common names for this species are "Tumbihkihrasd", "Mandasdi", "coeirupú", "coirepú",
"mandaçaia-grande", "mandaçaia". Melipona (Melipona) quadrifasciata Lepeletier, 1863 is
widely distributed in southern and south-eastern Brazil, namely the Atlantic Forest (Camargo
& Pedro, 2013). The Atlantic Forest is a biome which extends along the Atlantic coast of
Brazil from Rio Grande do Norte state in the North to Rio Grande do Sul state in the South,
and inland as far as Paraguay and the Misiones province of Argentina. In this region M.
quadrifasciata is bred and especially valued in tomato production (dos Santos et al., 2009).
Additionally, its propolis has medicinal properties (Mercês et al., 2013). Nesting sites are tree
trunks or in the ground. To seal openings in the nest, these bees collect clay (Camargo &
Pedro, 2013). A typical M. quadrifasciata colony contains up to 300-400 bees, of which only a
fraction are foragers (Nieh, 2004).
M. (Melipona) quadrifasciata quadrifasciata is a subspecies of M. quadrifasciata Lepeletier,
1863. It can be distinguished from the other subspecies M. quadrifasciata anthidioides by the
yellow metasomal tergite stripes at the 3rd to the 6th segment which are continuous in M. q.
quadrifasciata (Batalho-Filho et al., 2009; Nascimento et al., 2010). The morphology alone is
not an exclusive means for correctly identifying the different subspecies of M. quadrifasciata
(Tavares et al., 2013). M. q. quadrifasciata is found in regions with colder climates, in the
states of Paraná, Santa Catarina and Rio Grande do Sul (Moretto & Arias, 2005), but it is also
found in altitudes above 1500 m in São Paulo, Rio de Janeiro and Minas Gerais (Moretto &
Arias, 2005).
20
2. Melipona (Eomelipona) bicolor schencki Gribodo, 1893 (Camargo & Pedro, 2013)
Common names for this species are "eira-aviyú", "eirû", "guarupú", "guaraipo", "guaráipo",
"pé-de-pau". This subspecies can be found in the southern region of Brazil and in cool regions
at high elevations in the southeastern part of the country (Camargo & Pero, 2013). They nest
subterranean among the roots at bases of trees or in cavities of trees close to the ground. M.
bicolor is special among the stingless bees because of its facultative polygyny, which is rare
among eusocial bees. M. bicolor schencki is currently considered a threatened stingless bee
in southern Brazil (Ferreira et al., 2013: IBAMA. (2003) Lista das espécies de fauna ameaçada
de extinção.)
3. Melipona marginata obscurior Moure, 1971
This Melipona species is found in southern Brazil (Mato Grosso, Paraná, Rio Grande do Sul,
Santa Catarina, São Paulo), Argentina (Misiones) and Paraguay (Caaguazú). M. torrida,
proposed by Friese (1916) as a variety of M. marginata Lepeletier, 1836, was described
based on three workers putatively collected in Costa Rica. But M. marginata var. torrida
Friese, 1916 never had its identity properly recognized. It was argued by Melo (2013) that M.
torrida was based on mislabeled specimens and corresponds to M. marginata obscurior
Moure, 1971. Whether M. torrida should be given a species status separate from M.
marginata needs further investigation. Little literature on the biology of this species is
available.
1.5 Meliponiculture and its genetic consequences
1.5.1 Introduction
Despite their ecological importance as pollinators of native species in Brazil (Kerr et al., 2001;
Imperatriz-Fonseca et al., 2006; Tavares et al., 2013), stingless bee populations have decreased in the
Neotropics, similarly to other bee populations, due to habitat degradation and fragmentation,
agricultural intensification, unselective use of agrochemicals and pesticides, and introduction/spread
of exotic species (Freitas et al., 2009).
Use and management of non-Apis bees for crop pollination is important because the ability of honey
bees to pollinate is threatened. Reasons for this limitation are factors such as Africanization,
diseases, parasites, low efficiency on some crop species, climatic limitations, economic pressures and
not naturally occurring of honey bees in some places (Heard, 1999). For Southern-Brazil specifically,
Africanization, climate and not naturally occurring of honeybees are nevertheless not really
restrictions.
1.5.2 Melipona decline
Due to habitat loss caused by deforestation, some Melipona species are endangered in several
Brazilian States. For the state Rio Grande do Sul, M. bicolor schenki and M. quadrifasciata
quadrifasciata are classified as ‘in danger’ in a list of animal species threatened with extinction and
M. obscurior as ‘vulnerable’ (Marques et al., 2002). The extreme fragmentation of the Atlantic Forest
negatively impacts the bee species by interruption of gene flow and decrease in genetic diversity
(Freiria et al., 2012). Consequently these endangered species are more sensitive to environmental
changes. Adaptation to these new environments depends on the level of genetic variation which
21
exists among the members of a population. Without the presence of the necessary genetic variation,
populations are more prone to extinction.
Koser et al. (2014), found through the use of polymorphic microsatellite primers low genetic
variability in populations of M. quadrifasciata and other stingless bee species at a meliponary. If
natural populations of these species have a similar low genetic variability, they might be endangered.
However, little is known about the genetic variability status of hives and wild colonies of these
species, although such knowledge is essential for development of conservation strategies and
rational exploitation of native species (Cortopassi-Laurino et al., 2006).
Low genetic variability in other native stingless bees has been documented in several other studies
using molecular markers, like microsatellites. An explanation for this could be the species’ natural
biology (Koser et al., 2014). Most stingless bee queens are monandric (mating with single male
during nuptial flight) (Peters et al., 1999; Palmer et al., 2002). Nevertheless, the absence of polyandry
doesn’t automatically imply low colony-level genetic diversity. Stingless bee species like Trigona
spinipes have probably developed alternative mechanisms to increase intra-colonial genetic diversity,
like frequent queen supersedure, temporal polygyny or rare invasion by foreign queens (Jaffé et al.,
2014).
Stingless bee queens nest near maternal nests (Koser et al., 2014). Colonies and their female
reproductives do not migrate and show only reduced dispersal (a few hundred meters at the most)
during swarming (Quezada-Euán et al., 2007). For other stingless bee species, males are the
dispersing sex. Low dispersion increases genetic drift and inbreeding within sub-populations (Hartl &
Clark, 2007). Especially the maintenance of alleles of the complementary sex determination system is
a concern in small populations, because inbreeding can lead to the production of diploid males.
1.5.3 Meliponiculture for conservation
A way to prevent local extinction is maintaining hives. Meliponiculture for crop pollination and for
honey extraction is an encouraging economic practice, supposed to be in alignment with sustainable
development (Cortopassi-Laurino et al., 2006). The economic value of Melipona has encouraged the
practice of exchanging queens among commercial breeders (Moretto & Arias, 2005). This practice
can promote contact between subspecies and lead to gene flow and development of secondary
hybrid populations (Moretto & Arias, 2005). It would be useful to determine the maternal origin of
hybrids colonies, which may occur naturally or due to colony transportation among breeders. The
development of reliable genetic markers is of significant importance in studies where the
identification of (i) maternal origin in commercial populations or (ii) subspecies in natural
populations, is required (Moretto & Arias, 2005).
The low genetic variability found for stingless bee species at a meliponary (Koser et al., 2014) could
also be related to the artificial maintenance of hives, through founder events and bottlenecks.
However, due to strong care over the nests, populations can successfully be maintained for many
years (Alves et al., 2011a). Indeed, Alves et al. (2011a) showed that despite imposing a severe
bottleneck by setting up a M. scutellaris population from only 2 founder colonies and breeding it
outside its natural occurrence, the genetically impoverished population could be successfully bred for
at least 10 years.
22
Hives maintained at a meliponary, should be originally extracted from geographically close nature
and not have been divided artificially. In reality however, artificial division is a common practice and
for now the only way to multiply colonies of Melipona. For other stingless bee species, new colonies
can be captured when swarming, but it is a slow process. Managed populations of local bees can be
considered a reservoir of genetic diversity if they can interbreed with wild populations (Alves et al.,
2011a). By human intervention transport of hives of distant populations could cause gene flow,
which would be hardly possible in a natural way. Introduction of bees from distant populations or
another species that could hybridize with local populations might cause outbreeding depression.
1.5.4 Genetic consequences of artificial maintenance of Melipona colonies
Significant outbreeding depression is sometimes observed in crosses between distant populations of
the same species (Dobzhansky, 1948; Waser & Price, 1989). Decline in fitness under outcrossing is
usually attributed to a breakup of co-adapted gene complexes or favourable epistatic relationships
evolved in isolated populations (Mayr, 1963). Outcrossing enhancement in the F1 can be caused by
dominance of favourable genes isolated in the two parental populations, or by the existence of
favourable additive x additive epistatic interactions between genes from different sources (Lynch,
1991). The advantages of outcrossing revealed in the F1 individuals can be completely reversed in the
following generation (Lynch, 1991). Since F2 individuals have only half the between-source
heterozygosity as F1 individuals, even stronger depression in second (F2) and later generation
‘‘hybrid’’ progeny is predicted, unless there are highly favourable between-source dominance x
dominance epistasis effects (Lynch, 1991). Dominance of favourable genes increases rapidly to a
maximum with physical distance between mates. The break-up of co-adapted gene complexes has
consequences that build up more slowly with distance, but eventually they become more substantial
than the advantages of dominance (Lynch, 1991). Waser et al. (2000) calls this disruption of allelic co-
adaptation within or across gene loci in distant crosses the ‘physiological’’ explanation for
outbreeding depression’.
Outbreeding depression in progeny fitness may not only arise from disruption of allelic co-
adaptation, but also from disruption of local adaptation, or a combination of these ‘‘environmental’’
and ‘‘physiological’’ mechanisms (Waser et al., 2000). Thus differences in outbreeding depression
among populations or individuals can often be related to environmental variation. The expression of
outbreeding depression will consequently vary among situations. This is not surprising given likely
spatial and temporal variation in gene flow and selection regimes, different population histories, and
different parental and progeny environments (Waser et al., 2000). Populations and metapopulations
of a given species can be expected to vary in their degree of spatial genetic differentiation, similar to
what can be seen for samples drawn from a single population in different years. Whatever the
mechanism, outbreeding depression represents a partial crossing barrier. This barrier is postzygotic,
because outbreeding depression (by analogy to inbreeding depression) refers to the performance of
progeny (Waser et al., 2000).
High levels of genetic differentiation between several geographic populations can indicate the
presence of a cryptic species (Tavares et al., 2007). Quezada-Euán et al. (2007) recommended that
movement of colonies between two distinct regions should be reconsidered given that the two
regions may harbour two cryptic species. Knowing the marked differentiation between populations
from different regions that may even represent different species, caution needs to be taken in the
movement of colonies between locations.
23
1.6 Microsatellites as a molecular tool for population genetic research
1.6.1 Microsatellites
Population genetics investigate the connection between demographic features of populations and
the distribution of molecular variants within them (Sunnucks, 2000). Examining genetic markers with
appropriate rates of change, can obtain information about populations and evolutionary processes.
Population genetic analysis can extract useful information from the genetic data. Conservation
genetics are of primary importance for protection of biodiversity by avoiding the extinction of most
endangered species (Oliveira et al., 2006). The application to conservation genetics of molecular
techniques has made the examination of the genetics of species in danger of extinction possible
(Oliveira et al., 2006). The central topic in conservation genetics is loss of genetic variability, because
small populations occurring in fragmented areas can suffer from inbreeding depression leading to the
loss of heterozygosity, genetic diversity, fitness and adaptability (Oliveira et al., 2006).
Microsatellite (MS) analysis is very useful in population genetics (Oliveira et al., 2006). The high
sensitivity of PCR-based MS analysis is of great benefit for research based on samples with limited
DNA amounts (e.g. social insect) or degraded DNA (e.g. museum material). Microsatellite markers
can provide relevant information for investigating genetic processes that take place in populations,
such as patterns of gene flow, incidence of genetic drift and population genetic structure (Oliveira et
al., 2006).
Nevertheless, the application of MS to population genetic questions requires a detailed
understanding of the microsatellite DNA fragments and their evolution.
1.6.1.1 Introduction
MS are simple tandemly repeated DNA fragments that can be found at high frequency in the
eukaryotic and prokaryotic genome (Toth et al., 2000; Selkoe & Toonen, 2006). MS sequences are
non-coding but can be located in both coding and non-coding regions (Miah et al., 2013). Coding
regions can thus contain MS tracts (Li et al., 2002). MS suffer from higher rates of mutation
compared to the rest of the genome (Jarne & Lagoda, 1996). The mutation rate has been estimated
to be between 10-6 and 10-2 per generation, which is significantly higher than base substitution rates
(Schlötterer, 2000; Miah et al., 2013). Because of their exceptional polymorphism in natural
populations, MS prove to be very versatile molecular tools to solve biological problems and have
been one of the most widely applied classes of molecular markers used in genetic studies (Oliveira et
al., 2006).
In MS, short simple DNA sequences of 1 to 6 base pairs long, are repeated multiple times into arrays.
The minimum total array size is suggested to be 8 nucleotides (Chambers & MacAvoy, 2000).
Microsatellite loci typically vary in length between 5 and 40 repeats (Selkoe & Toonen, 2006) and the
tracts can be up to approximately 10² nucleotides long (Chambers & MacAvoy, 2000). Microsatellite
loci are represented by the general formula -(N1N2N3…Nx)n- for which the unit repeat size x: 2-6 nt
and the number of repeat units n such that: x*n>8 nt (Chambers & MacAvoy, 2000). An overarching
term for nucleotide repeat arrays in genomes is variable number of tandem repeats (VNTR)
(Chambers & MacAvoy, 2000). Microsatellites are also known as short tandem repeats (STR), simple
sequence repeats (SSR) or simple sequence length polymorphisms (SSLP) (Chambers & MacAvoy,
2000).
24
Classification of MS can be based on their size, type of the repeated unit or position within the
genome (Miah et al., 2013). According to the number of nucleotides per repeat unit, MS can be
classified as mono-, di-, tri-, tetra-, penta- or hexa-nucleotide repeat (Miah et al., 2013). In most
species, the majority of microsatellites are dinucleotide repeats (Schug et al., 1998; Selkoe & Toonen,
2006). The most common choices for molecular genetic studies are dinucleotide, trinucleotide and
tetranucleotide repeats (Selkoe & Toonen, 2006).
Microsatellites are also classified according to the type of repeat sequence (Oliveira et al., 2006).
Depending on the arrangement of nucleotides within the repeat motifs, the terms perfect, imperfect
and compound are used by Weber (1990) and terms simple perfect, simple imperfect, compound
perfect and compound imperfect by Wang et al. (2009). In pure (perfect) microsatellites only one
single repeated motif is found within the tandem array of the given microsatellite locus (Jarne &
Lagoda, 1996; Chambers & MacAvoy, 2000). Compound (composite) and complex microsatellites
contain respectively two or more types of repeat motifs in various configurations within their arrays
at a given locus (Chambers & MacAvoy, 2000). The term interrupted (imperfect) is used to indicate
that there are one or more single non-repeated units that don’t match the motif sequence internal to
the array (Jarne & Lagoda, 1996; Chambers & MacAvoy, 2000). Finally the classification can also be
based on the location of the MS in the genome: nuclear, chloroplastic, mitochondrial (Miah et al.,
2013: Kalia et al., 2011).
Microsatellite markers have a broad range of applications, due to the fact that they are co-dominant,
multi-allelic, highly reproducible, have high resolution and are based on the polymerase chain
reaction (Oliveira et al., 2006). Originally microsatellite analysis was designed for research in humans,
but it has become a powerful tool for research on animals and plants too (Oliveira et al., 2006). They
have been used extensively in the construction of genetic maps, in research on human genetic
diseases and in studies of population genetics, for genotyping and paternity analysis (Oliveira et al.,
2006). Their use in population genetic studies and forensic science is justified by their high specificity
and discriminating power (Selkoe & Toonen, 2006; Oliveira et al., 2006). At the same locus, the
population as a whole contains several alleles, each with a different number of repeats (Oliveira et
al., 2006).
1.6.1.2 The biology of MS
Distribution
For many species microsatellites have been characterized, but the distribution of microsatellites in
nature is not completely clear yet (Chambers & MacAvoy, 2000). Without doubt, MS distribution
within and between genomes is not random (Li et al., 2002). Microsatellites are longer and more
common in vertebrates than in invertebrates. Hence, among vertebrates, cold blooded species have
longer repeat arrays (Chambers & MacAvoy, 2000). The chromosomal locations of microsatellites are
well known in several organisms and they are rather uniformly distributed across intergenic areas
(except telomeres) and they are usually rare in coding regions (Chambers & MacAvoy, 2000). The
lower frequency of MS in coding regions could be attributed to negative selection against frameshift
mutations in coding regions (Metzgar et al., 2000). In contrast to di- and tetranucleotide MS,
trinucleotide and hexanucleotide repeats do appear frequently in coding regions because they do not
cause a frameshift (Selkoe & Toonen, 2006).
25
Genesis
Microsatellites originate from regions of cryptic simplicity (Tautz et al., 1986), in which variants of
simple repetitive DNA sequence motifs are already over represented (Tautz et al., 1986). So ancestral
sequences of microsatellites are cryptically simple sequences (Chambers & MacAvoy, 2000). They are
short (<8 nt long) runs of repetitive nucleotides (Chambers & MacAvoy, 2000) that can give rise to
microsatellite arrays when they mutate (Chambers & MacAvoy, 2000).
Other authors have different approaches for explaining microsatellite genesis. Schlötterer (2000)
mentions ‘proto’-MS, as from which MS originate. These short ‘proto’-MS are generated by chance
(Levinson & Gutman, 1987), by random point mutations (Jarne et al., 1998). A minimum number of
repeats is required before the proto-MS can be further extended (Rose and Falush, 1998). Once the
threshold length has been reached, DNA slippage becomes active and the role of point mutations
becomes negligible (Rose & Falush, 1998). The threshold length would result from the minimum
length of the repeated sequence allowing stable misalignment and therefore DNA slippage. Other
reports say no minimum number of repeats is required and slippage occurs even at very short MS,
although at a reduced rate (Pupko & Graur, 1999; Noor et al., 2001; Sokol & Williams, 2005).
Microsatellites can thus be generated by random point mutations followed by rare slippage events
that extend short proto-MS (Schlötterer, 2000).
More recent studies suggested that a third molecular mechanism, indel slippage, may contribute to
the creation of very short MS (Dieringer & Schlotterer, 2003). In contrast to DNA slippage, which
occurs only in tandem repeats, indel slippage is expected to happen at a constant rate and at random
genomic positions (Dieringer & Schlotterer, 2003). Indel slippage can be explained through
nonhomologous end joining repair (Leclercq et al., 2010). A completely different mechanism for the
generation of MS has been proposed, namely retrotransposition (Nadir et al., 1996), but it remains
unclear whether this is involved.
Processes underlying MS instability
The exact mutational dynamics of the MS regions are not well understood yet, but several mutation
mechanisms have been suggested.
The main mutation mechanism of MS is DNA slippage (Eisen, 1999). During DNA replication the
nascent and template strand possibly realign out of register. If DNA synthesis continues on this
molecule, the repeat number of the MS is altered (Schlötterer, 2000). In reality however,
exonucleolytic proofreading and the mismatch repair system correct many of these mutations.
Microsatellites are finally the unrepaired products of slipped strand mispairing during DNA synthesis
(Chambers & MacAvoy, 2000). The mutation rate depends greatly on the efficiency of the mismatch
repair system and proofreading.
MS length could potentially also be changed by unequal crossing over during recombination (Li et al.,
2002). Recombination events are believed to be a minor source of MS variability (Schlötterer, 2000;
Leclercq et al., 2010). Nevertheless, some studies point to the important role of nonreciprocal
recombination (gene conversion) in destabilization of tandem repeats (Jakupciak & Wells, 2000;
Richard & Pâques, 2000). MS stability could also be affected by an interaction of replication slippage
and recombination, namely replication slippage during recombination-dependent DNA repair (Li et
al., 2002).
26
Primarily the number of repeats and thus the length of the repeat string is changed by these
mutations (Selkoe & Toonen, 2006). Both the gain or loss of one or more repeat units within the
microsatellite array can occur (Chambers & MacAvoy, 2000). When the gain of unit mutations occurs
more frequently than, or is repaired less frequently than, the loss of unit mutations, the size of a
microsatellite array will increase over time (Chambers & MacAvoy, 2000). The mutation/repair
process is supposed to be upwardly biased (Chambers & MacAvoy, 2000). When hosts don’t lack
efficient mismatch repair systems, microsatellite arrays containing longer units evolve faster than
those containing shorter units, particularly in microorganisms (Eisen, 1999; Chambers & MacAvoy,
2000). Leclercq et al. (2010) state that longer MS are more variable than shorter ones because the
mutation rate is correlated to the number of repeats (Schlötterer, 2000). This can mostly be
attributed to the mismatch repair systems, that are relatively inefficient for repair of larger
mismatched segments (Chambers & MacAvoy, 2000). Another view that favors expansion in
mutation processes, assumes heterozygotes whose microsatellite alleles show the greatest
difference in array size, to be more sensitive to mutations (Chambers & MacAvoy, 2000).
For short MS (<15-20 bp long), the mechanisms responsible for MS dynamics remain poorly
understood. A threshold above which slippage becomes effective has been suggested by most
studies. Others also considered the mechanism of indel slippage on small repeats. Leclercq et al.
(2010) found no evidence of a minimum threshold because tandem duplications/deletions acted on
short MS. Their study showed that tandem duplication rate increases exponentially as a function of
STR length, beginning with the shortest possible length of (period+1) nucleotides. This is the shortest
size at which the polymerase can be misled at replication. Another result of their study was the
confirmation at wide genomic scale of the occurrence of indel slippage.
Theoretical models
Some theoretical mutation models have been developed to apply to MS data which have been
described to correctly determine population genetic parameters from them:
- Infinite alleles model (Wright, 1931; Kimura & Crow, 1964): each mutation
randomly creates a new allele. Mutations alter the number of repeats in
microsatellite loci, with any number. Proximity in terms of number of repeats
does not indicate greater phylogenetic relationship.
- Stepwise mutation model (Kimura & Ohta, 1978; Slatkin, 1995): when mutated, a
microsatellite locus gains or loses a repeat, with equal probability if model is
symmetrical). Two alleles differing by only one repeat are more related than
alleles differing by several repeats. This model is preferred when estimating
relationships between individuals and population structure, but not when
homoplasy (two alleles are identical by state but not by descent) is present.
- Two phase model (Di Rienzo et al., 1994): extension of the SM-model that states
that most mutational events result an increase or decrease of one repeat unit,
though alternations of a large number of repeat also occur.
- K-alleles model (Crow & Kimura, 1970): this model assumes that if there are
exactly k possible allelic states in a given locus, then the probability of a given
allele mutating into any of the other K-1 alleles is µ⁄(k-1), where µ is the mutation
rate.
27
These models are used in the statistical analyses of genetic variation and to derive the expected
number of alleles in a population from the observed heterozygosity (Oliveira et al., 2006).
Unfortunately, they all have some disadvantages (Oliveira et al., 2006).
1.6.1.3 Detection of MS
Usually microsatellites can be identified by their flanking region of DNA. Primers are designed to bind
to these flanking regions and initiate the amplification of the MS loci by PCR (Selkoe & Toonen,
2006). The primers for the flanking regions are the result of a microsatellite marker isolation process.
Alleles of different length (repeat numbers) begin to develop in a population as mutations slowly
alter particular microsatellite arrays (Chambers & MacAvoy, 2000). The mutation rates of many
microsatellites are high, which generates the high levels of allelic diversity (Selkoe & Toonen, 2006)
or polymorphism (Leclercq et al., 2010). Consequently, the alleles can be distinguished by
electrophoresis according to their size (Selkoe & Toonen, 2006). Unlabeled primer sequences can be
delivered relatively quickly and cheap (Selkoe & Toonen, 2006). Fluorescently labelled primers are
much more expensive (Selkoe & Toonen, 2006).
In general the sequences of these flanking regions are highly conserved across individuals of the
same species and sometimes of different species. Primers developed for microsatellite loci of species
in a certain genus or family can be used for cross-species amplification (Selkoe & Toonen 2006). The
rate in which this cross-species amplification works determines the MS transferability (Oliveira et al.,
2006). Unfortunately the success rate of these primers decreases with the genetic distance between
the species of interest and the origin species (Wright et al., 2004 ). In addition, allelic diversity often
also decreases when primers are used in non-source species (Ellegren et al., 1997; Neff & Gross,
2001; Wright et al., 2004). Other factors for successful transfer are the size and complexity of the
genome concerned and whether or not the MS belongs to a coding region (Oliveira et al., 2006).
Attempting the amplification of existing primers from related species will generally be less expensive
and time-consuming than isolating new primers (Squirrell et al., 2003).
The most common choice of MS for molecular genetic studies are dinucleotide, trinucleotide and
tetranucleotide repeats (Selkoe & Toonen, 2006). Tri- and tetra-nucleotide repeat motifs generally
yield fewer stutter bands (Miah et al., 2013). However, di-nucleotide motif repeats are more
common in most species than tri- or tetranucleotide motifs and are therefore easier to use in
combinatorial screens.
1.6.1.4 Advantages and weaknesses of MS
The primary advantage of MS as genetic markers is that they are inherited in a Mendelian fashion as
co-dominant markers (Miah et al., 2013). Co-dominance is beneficial because it means that the 2
alleles of a locus can both be identified and analyzed (Sunnucks, 2000). Furthermore high
polymorphism rates, high abundance and a broad distribution throughout the genome have made
MS one of the most popular genetic markers (Miah et al., 2013).
As mentioned before, the high mutation rates of microsatellite markers result in high allelic diversity.
In species for which populations are small or recently bottlenecked, markers with high mutation
rates like MS, will be very variant and likely to be very informative (Hedrick, 1999). When trying to
detect changes in the recent past (10-100 generations), it is most interesting to select loci with high
mutational rates and thus allelic diversity (Selkoe & Toonen, 2006). For studies that utilize population
allele frequency estimates, the numerous alleles of high diversity microsatellites will give more
28
statistical power to distinguish populations (Kalinowski, 2002; Wilson & Rannala, 2003). Comparing
populations and individuals is more precise and statistically powerful when multiple samples of the
genome are used, by combining the results from multiple loci (Selkoe & Toonen, 2006). A multilocus
microsatellite study usually has a higher resolution and power than other multilocus techniques, like
AFLP and allozymes (Sunnucks, 2000).
Practically MS also have several advantages. Sample preparation is easy and fast, and cheap DNA
extraction methods are generally sufficient (Selkoe & Toonen, 2006). Only a small amount of sample
tissue is needed, since the MS markers are amplified by PCR. DNA on itself is quite stable, so storage
of the tissue is quite simple. If some DNA degradation does occur, the MS can still be amplified by
PCR, because MS are usually short in length (Taberlet et al., 1999). So MS can be easily used as
markers within experiments with ancient DNA (Taberlet et al., 1999). Cross-contamination by non-
target organisms is not likely, because MS are mostly species-specific (more than techniques that use
universal primers, like AFLP) (Selkoe & Toonen, 2006).
The drawbacks of MS-based methods are mainly the high development costs and technical
challenges during the construction of enriched libraries and species-specific primers (Miah et al.,
2013).
Homoplasy can represent a great inconvenience in the reliability of results concluded with MS
markers. Homoplasy indicates that 2 alleles of a locus can be identical in state (size), but not
necessarily identical by descent (Estoup et al., 2002). Homoplasy keeps allelic diversity hidden and
makes identification of each allele by its size ambiguous. In some cases size homoplasy is molecularly
accessible, when the electromorph hides different sequences (interruptions, varying flanking region)
(Estoup et al., 2002). Homoplasy forms a problem when undetectable. Marker homoplasy is most
problematic for applications in which populations are distantly related, or even distinct species, but
the bias is expected to be small (Selkoe & Toonen, 2006).
As said before, the flanking primer regions of MS are highly conserved, so they are sometimes not
suitable across broader taxonomic groups. Marker isolation is thus usually species-specific, so
primers need to be developed for each species.
Complex mutational processes are involved in the evolution of MS. The mutation models are
developed to approach these processes in theory but cannot completely represent reality
unfortunately. Some statistics based on estimates of allele frequencies, rely explicitly on these
mutation models. Metrics based on a certain mutation model, can be sensitive to violations of the
model. Therefore more complex and realistic mutation models must be included in statistical
packages.
1.6.2 Population genetic parameters
In population genetic studies, microsatellites are used to estimate the genetic diversity, inbreeding
levels and the genetic structure of (sub)populations (e.g. Selkoe & Toonen, 2006; Zayed, 2009).
Generally the main subject is genetic diversity. It defines the variation in genetic material present
within a population of a certain species. A population or species with greater genetic diversity will
have more chances to withstand changes in the environment.
The population genetic diversity can be expressed by 3 parameters:
29
- expected heterozygosity He, calculated based on allele frequencies;
- observed heterozygosity Ho;
- allelic richness AR.
An individual is homozygous when both alleles of one locus are identical. When both alleles are
different, the individual is called heterozygous for that particular locus. The observed heterozygosity
(Ho) represents the number of heterozygotes within a population. The expected heterozygosity (He)
on the other hand is the expected amount of heterozygotes within a population following the Hardy-
Weinberg equilibrium of random mating. He values range between 0 and 1, with a value of 1 meaning
complete heterozygote population and a value of 0 indicating a completely homozygous population.
These extremities are theoretical, usually He has a value slightly greater than 0.5. The relationship
((He - Ho)/ He) between observed and expected heterozygosity determines the inbreeding coefficient
(Fis), which ranges between -1 and 1. When the value of the inbreeding coefficient reaches 1, there is
an excess of homozygous individuals present in that population, possibly caused by inbreeding.
When the value reaches -1, there is an excess of heterozygotes present in that population, showing
outbreeding. This is mostly a sign of an artificially composed population (Widmer & Schmid-Hempel,
1999; Schmid-Hempel et al., 2007). The allelic richness indicates the number of alleles within a
population for a certain locus. The sampling size plays an important role, since a greater sampling of
a genetically poor population can deliver more alleles than a small sampling of a genetic rich
population. In this case a comparison between the two populations is meaningless. For this reason
the allelic richness should always be corrected for population size.
Population genetic studies also often calculate F-statistics to determine the population structuring by
calculation of FST (Weir & Cockerham, 1984; Nei, 1987). For microsatellites this parameter ranges
from 0 to 1, with zero representing no differentiation and a value of FST = 1, means fixation of
different alleles between the populations and thus population structuring (Meirmans & Hedrick,
2011). Recently, the use and accuracy of FST - values were under debate (Jost, 2008; Whitlock, 2011;
and as reviewed in Meirmans & Hedrick, 2011). Indeed, due to its dependency on within-population
diversity, FST -values are not always trustworthy. Therefore, a new estimated parameter (Dest) was
described based on the effective number of alleles (Jost, 2008). Currently, both parameters are
estimated and used together in population genetic studies (Meirmans & Hedrick, 2011; Cameron et
al., 2011; Lozier et al., 2011).
Other ways to assess population genetics can be: (i) searching for evidence of population bottlenecks
(sharp reduction in the size of a population due to environmental events), (ii) assess the effective
population size, and (iii) investigate the magnitude and directionality of gene flow between
populations (e.g. Selkoe & Toonen, 2006; Zayed, 2009).
30
2. Material and methods
2.1 Collection of samples To collect the necessary data for this research, I travelled to Porto Alegre in the Southern-Brazilian
state Rio Grande do Sul. I stayed there for 2 months during the spring/summer (October-December).
In order to collect historical and recent DNA material from bees, I worked in the lab of Prof. Betina
Blochtein at Pontifícia Universidade Católica do Rio Grande do Sul (PUCRS) and performed fieldwork
across the state of Rio Grande do Sul, always under guidance of a collaborator of the lab.
Material was gathered from the entomological (Hymenoptera) collection of the Museum of Science
and Technology at PUCRS. Specimens from several locations were sampled, and selected taking into
consideration their distribution. For bumblebees (Bombus) three species were of interest: B.
bellicosus, B. pauloensis and B. morio. According to the availability in the collection, the objective was
to extract DNA of 25 specimens of each species from each sampling location. From the pin-mounted
specimens, one middle leg was extracted and cut longitudinal to expose the internal tissue for later
DNA extraction. Tools for cutting (forceps and knife) were sterilised with alcohol under a flame. The
amputated legs were stored in microcentrifuge tubes (Eppendorf) and frozen at -20°C. Specimens
were also sampled from other bumblebee collections: from the entomological collection at Museu de
Ciências Naturais da Fundação Zoobotânica do Rio Grande do Sul (curator: Aline Barcellos), from the
entomological collection at UNISC in Santa Cruz do Sul (curator: Andreas Kôhler) and a private
collection of Mardiore Pinheiro (collaborator of Betina Blochtein). Each pin-mounted bee was tagged
with a collection number and through a database the according information could be assembled
(determination of species, collection date, location, ...).
To make a comparison of bumblebee-populations in time, choices of sampling locations for current
specimens were deduced from the locations that were most sampled in the past (overview of
sampling locations and data + number of collected bees). The collection of foraging bees was done by
observation of flowers throughout the day between 9 am and 5 pm. During these hours of daylight,
bees are most active. An entomological net or a glass jar was used to capture the bees from the
flowers. Afterwards these bees were transferred to plastic tubes. During the sampling time, the
tubes with collected bees were kept in ice to sedate the bees. Back in the lab, a middle leg was
extracted, cut and stored in the freezer (-20°C) and the rest of the body was transferred to 90%
ethanol and stored in the fridge. Given the context of scarce and declining species, nonlethal
sampling (e.g. Holehouse et al., 2003) would be preferred. Nevertheless, whole specimens were
collected for an associated study on pathogens in Brazilian bees. At every visited location there was
also aimed to collect 20 honeybees and 20 stingless bees for this project.
For Melipona three species were sampled from the entomological collection at PUCRS and the
entomological collection at Museu de Ciências Naturais da Fundação Zoobotânica do Rio Grande do
Sul. The selected species were: M. quadrifasciata quadrifasciata, M. bicolor schencki, M. marginata
obscurior. Specimens from a broad range of locations and time periods were sampled. The purpose
was to get DNA from approximately 25 specimens of these three species, from a period before and
after the hypothesized introduction of hives from other states. Specimens were selected from both
locations spread over Rio Grande do Sul and locations of neighbouring states which expected to
deliver the introduced hives (like Santa Catarina, and São Paulo). The legs were sampled by the same
method as the legs of the bumblebee specimens. For Melipona sampling, a stingless bee keeper in
31
Rolante was visited additionally. Sampling was done directly from the hives. With a sucking system
Melipona bees were taken from their nest, after opening the hive boxes. Hence, also some bees were
collected in a tube at the entrance of their nest.
Table 1: Summary of sampling locations and dates. Species name and number of collected specimens are also indicated
Location Date Species collected bee (nr)
Viamão FEPAGRO
29/10/2015 B. morio (1) (not collected, only observed)
15/11/2015
Plebeia emerina (3)
Mourella caerulea (15)
Trigona spinipes (6)
Apis mellifera (17)
Bombus morio (1)
Candiota Chácara Hortec
30/11/2015 No bumblebees observed
Guaíba Fazenda Matzenbacher
30/10/2015 Apis mellifera (20)
11/12/2015 Trigona spinipes (8)
Bombus morio (1)
Rolante Fazenda Girlei dos Passos
6/12/2015
Apis mellifera (20)
Melipona quadrifasciata quadrifasciata (20)
Melipona bicolor schencki (19)
Melipona marginata obscurior (20)
Santa Cruz do Sul Cinturão Verde
17/12/2015 Bombus morio (5)
Apis mellifera (8)
São Francisco de Paula Pró-Mata
8/11/2015
Apis mellifera (2)
Melipona bicolor schencki (7)
Melipona quadrifasciata quadrifasciata (8)
Trigona spinipes (1)
Augochloropsis (1)
15/12/2015
Apis mellifera (10)
Melipona bicolor schencki (2)
Bombus pauloensis (1)
Another occupation during the stay in Porto Alegre was photographing bumblebee wings through a
microscopic lens. For the three bumblebee species occurring in Rio Grande do Sul, both right and left
wing of a number of approximately 20 specimens were photographed. These pictures will be
processed by a collaborating group from the Naturalis Biodiversity Center (Leiden, The Netherlands),
which is developing a technique for species recognition through analysis of veins in wing images.
Diploid males and consequently inbreeding could also be detected by this technique (Gerard et al.,
2015).
32
2.2 DNA-extraction, primer selection, PCR, capillary electrophoresis
2.2.1 DNA-extraction
The first step to extract genomic DNA was cutting the legs longitudinal to expose the internal tissues,
which are surrounded by an exoskeleton. DNA of the bumblebee specimens was then extracted
according to the Chelex DNA extraction protocol for normal samples (based on Walsh et al., 1991).
This protocol consisted of: adding 200 µl 5% Chelex (InstaGeneTM Matrix, BioRad) and 10µl of
proteinase K to the samples. This suspension was then incubated while shaking at 55°C for 2 hours
and thereafter for 15 min at 97°C. Afterwards, these extractions were stored at -20°C.
Proteïnase K was added to set the nucleic acids (DNA and RNA) free from the tissues by breaking
down the cell membranes and proteins. At 97°C Chelex will bind all polar cell components and not
the nucleic acids, which are not polar. By centrifugation of this suspension, the nucleic acids, which
stay in the supernatant, could be separated from the pellet with cell components. This extraction
method was successful for recent specimens (samples collected after the year 2005). Unfortunately
quality of DNA extracted with this method was poor for older (museum) specimens (before 2005). To
improve the quality Amicon Ulta-0.5 Centrifugal Filter Devices (Millipore) were tested to purify and
concentrate the DNA. The effect of this application was variable. Another option for improving the
DNA quality was to repeat the extraction on the pellet, in case the first extraction wasn’t intense
enough to set the DNA free from the cells. In this case, the second attempt of extraction DNA is
optionally done according to the Chelex DNA extraction protocol for museum samples. Nevertheless,
this resulted in variable outcomes and was not executed any further.
Extractions on a number of both old and recent Melipona test samples were done according to the
overnight Chelex DNA extraction protocol for museum samples, since experience with the Bombus
museum samples showed insufficient extraction using the short extraction method. The Chelex DNA
extraction protocol for museum samples follows the same procedure as the one for normal samples,
but adds more 5% Chelex (InstaGeneTM Matrix, BioRad) (300 µl) and proteinase K (20 µl) and the
incubation is prolonged: overnight (±16 hours) at 37°C and 1 hour at 97°C.
2.2.2 Selection MS primers
Secondly, a selection of MS loci to genotype the Brazilian bumblebees was made. The used
microsatellites were developed for European bumblebee species, so they needed to be validated for
their reliable use in the different Brazilian bumblebee spp. The selection was done by testing the
primers on both old and recent bumblebee specimens of the 3 species collected at several locations.
The primers of the following 16 polymorph MS loci gave reliable signals in previous research on
bumblebee museum samples (Maebe et al., 2015): BL13, BT02, BT23, BT24, BL02, BT04, BT05, BT08,
and BT10 (Reber-Funk et al., 2006), B100, B11, B126, and B132 (Estoup et al., 1993), and 0294, 0304,
and 0810 (Stolle et al., 2009). These primers were originally developed for B. terrestris and B.
lucorum. These MS loci have a range lower or around 200 bp to avoid the chance of null alleles
(Wandeler et al., 2007). All 16 MS were retained for the MS analysis on the Brazilian bumblebees.
For the three Melipona species, 32 primers previously described in literature were selected to test on
the Melipona samples. 19 MS loci (Mbi11, Mbi28, Mbi32, Mbi33, Mbi85, Mbi85, Mbi201, Mbi215,
Mbi218, Mbi219, Mbi221, Mbi232, Mbi233, Mbi254, Mbi256, Mbi259, Mbi278, Mbi278, Mbi305, and
Mbi522) were developed by Peters et al. (1998) for stingless bees and obtained from M. bicolor
bicolor. They investigated the utility of their presented markers by genotyping M. bicolor and M.
33
quadrifasciata adult workers. These 19 loci were polymorphic in M. bicolor and yielded scoreable
PCR products for M. quadrifasciata. Only 8 of those 19 were polymorphic in M. quadrifasciata. From
the 10 MS primers identified and characterized specifically for M. mondury Smith, 1863 by Lopes et
al. (2009), the 6 MS primers (Mmo03, Mmo08, Mmo10, Mmo11, Mmo20, and Mmo22) that
amplified in both M. quadrifasciata and M. bicolor were chosen. From the 8 MS primers designed for
M. rufiventris by Lopes et al. (2009), the 7 MS primers (Mru03, Mru05, Mru06, Mru08, Mru09,
Mru12, and Mru14) that allowed cross-species amplification in both M. quadrifasciata and M. bicolor
were chosen. All these primers were tested on some samples of the three Melipona species from
different time periods and locations.
2.2.3 Amplification & visualization
Next the microsatellites were amplified by PCR reaction. The first step of a PCR reaction is
denaturation at 95°C in which the double DNA strand gets separated into 2 single DNA strands. In the
next step, the MS primers will bind to the DNA template at a certain annealing temperature specific
to each primer (Ta). In the third step the temperature is raised to 72°C, the optimal temperature for
DNA extension. At this temperature, the added Taq-polymerase will build a complementary DNA
strand starting from the annealed primer. By repeating the steps of this cycle several times, the DNA
region flanked by the forward and reverse primer will increase exponentially.
The microsatellites were amplified by PCR in 10µ volumes using the Type-it QIAGEN PCR kit. Each
reaction contained 1.33 µl template DNA, Type-it Multiplex PCR Master Mix (2x, Qiagen) and the
forward and reverse primer of four microsatellite primers for each of four multiplex mixes. The
multiplex PCR technique allows several primers to be amplified simultaneously. The forward primer
of each microsatellite locus was 5’-end labelled with a fluorescent label for visualisation of the
amplicon by capillary electrophoresis. Four fluorescent dyes are used (FAM, PET, NED, and VIC). The
four primers in one multiplex mix reaction can be distinguished because of their label. Four
multiplexes for the 16 primers were established previously (Maebe et al., 2016) (Table 2), so labelled
forward primers could be purchased. For older (collected before 2000) samples with poor DNA
quality, the number of repeated expansion cycles was increased from 30 (for recent samples) to 40
cycles, to guarantee sufficient amplification.
To separately validate and optimise the new Melipona primers, a fluorescently labelled M13 primer
was used in a nested PCR method (Schuelke, 2000). For the different MS primers that should be
tested, the sequence of the forward primer was extended with a specific sequence, the M13-tail. The
M13 forward primer has this specific tail as its sequence and is labelled fluorescently. During PCR,
after incorporation of the MS forward primers, the fluorescent M13 forward primer anneals to these
primers. This method allows the visualisation of the microsatellite fragments without the cost of
labelling all the MS forward primers. This method is nevertheless more time consuming. After
validation of the primers, a suggestion for a multiplex PCR combination of the selected primers was
proposed. First the compatibility of the primers to be amplified simultaneously in groups of 4 was
tested. While validating the Melipona primers, it was established that the number of repeated
expansion cycles in the PCR reaction preferably should be increased from 30 to 40 cycles, to
guarantee sufficient amplification.
The PCR amplification products were visualised with capillary electrophoresis. This technique
separates DNA fragments according to their size and measures the intensity of the fluorescence. PCR
34
products were run by the Genetic Service Units of the University Hospital Gent on a ABI3730xl
sequencer (Applied Biosystems) using an internal size standard (Genescan 500 LIZ). This internal
standard is used to size DNA fragments with length between 35 and 500 base pairs
Table 2: Multiplexes for Bombus
Multiplex 1
MS locus
Dye Sequence (5'-3')
Forward primer Reverse primer
BL13 PET CGAATGTTGGGATTTTCGTG GCGAGTACGTGTACGTGTTCTATG
BT02 NED AGGAACCGAGCGATAGAACCAC GCTTTGCCTTTCCATCTTGCTG
BT23 FAM GCAACAGAAAATCGTCGGTAGTG GCGGCAATAAAGCAATCGG
BT24 VIC TCTTTCCGTTTTCCCCCTG CACCCACTTACATACATACACGCTC
Multiplex 2
MS locus
Dye Sequence (5'-3')
Forward primer Reverse primer
BL02 NED GAACAGTGAGAGCGAGGAACAGAG TTGCCACGTATATCCGAGCGAACC
BT04 FAM GAGAGAGATCGAATGGTGAGAGC TGAGCACGTTCTTTCGTTCAC
BT08 PET AGAACCTCCGTATCCCTTCG AGCCTACCCAGTGCTGAAAC
BT10 VIC TCTTGCTATCCACCACCCGC GGACAGAAGCATAGACGCACCG
Multiplex 3
MS locus
Dye Sequence (5'-3')
Forward primer Reverse primer
B100 FAM CGTCCTCGTATCGGGCTAAC CGTGGAAACGTCGTGACG
B11 NED GCAACGAAACTCGAAATCG GTTCATCCAAGTTTCATCCG
B126 PET GCTTGCTGGTGAATTGTGC CGATTCTCTCGTGTACTCC
B1232 VIC GAAATTCGTGCGGAGGG CAGAGAACTACCTAGTGCTACGC
Multiplex 4
MS locus
Dye Sequence (5'-3')
Forward primer Reverse primer
0294 FAM AGTACGATAAAGCCAGGAAAG TGTATGCCTATTGTACGAGTGT
0304 NED GTATGAGTGAGTGATGTGCAAG CCCTTCATCTCTGAACAATATC
0810 PET TTAACAAATCCGAATTTAAAGG GATAGTGGTTGCTTGTCATCTT
BT05 VIC TTTCCTATGCCGAACGTCACC CCCAGATAAAAGACCGCCTCTAGTC
2.2.4 Genotyping
The MS fragments were examined manually using the Peak Scanner Software v 1.0 (Applied
Biosystems). This software identifies peaks and fragment sizes for capillary electrophoresis assays.
The alleles of the MS are graphically represented by peaks (Figure 2). The length of the different
alleles within one MS locus can be scored using the internal standard (Figure 1), reading from the x-
axis the fragments length in base pairs. The length of the alleles of most MS varies ranges between
100 and 300 base pairs. The height of the peak indicates the intensity of fluorescence.
In a professional research situation, to ensure data quality, genotyping should be replicated for a
number of random individuals. A genotypic error rate could be estimated from this.
35
Figure 1: Example of sizing standard 500 LIZ
Figure 2: Example of Peakscanner output for 1 heterozygous locus and 3 homozygous loci
2.3 Data analysis Unfortunately the dataset of bumblebees was not perfect for the proposed objective. Nevertheless,
the collected data was divided in four time periods (1: 1946-1959, 2: 1991-1994, 3: 1999-2004, and 4:
2010-2015). Seven locations were assumed as sampled throughout these periods (Figure 3): Cambará
do Sul +Torres (1), Candiota (2), Esmeralda (3), Guarani das Missões (4), Osório + Capão da Canoa (5),
Porto Alegre + Viamão+ Guaíba (6), and Santa Cruz do Sul (7) (Figure 3). For each population, the
intention was to genotype at least 10, up to maximum 25 specimens. Some populations exist of
specimens sampled from different locations. As these locations are in close proximity and within the
same bio-geographical regions, it was assumed that they represent the same population. This helped
to increase the number of specimens available per population and thus the power of the genetic
36
analyses. Samples from 1 population (Sao Francisco de Paula, Pro-Mata + Nova Petropolis) did not fit
within the fixed time periods, but were used as an extra data point. This location is special anyway,
since it is an area of conservation. For B. bellicosus only data from one location and one point in time
wa available (samples from Candiota, 1999). At that point in space and time, genetics of a population
of B. bellicosus can be compared with the genetics of a population of B. pauloensis and B. morio.
Figure 3: Map of Bombus sampling locations in the state of Rio Grande do Sul
In order to study the genetic status of these populations through time, genetic diversity, inbreeding
and population structure within and between the populations were assessed using several
parameters. But first the dataset had to be adjusted by some validation steps to meet certain criteria.
As the microsatellites used here were developed from B. terrestris and B. lucorum, we needed to
validate if they could be used in a reliable manner in the different Brazilian bumblebee species.
Genotyped individuals which could not be scored in a reliable manner for a minimum of MS loci,
were not included. This also applied to the MS loci which created to much missing data.
To detect the presence of multiple sisters from the same colony in a population, Colony 1.2 (Wang,
2004) was applied (employing corrections for genotyping errors, 5% per locus). To exclude problems
using Colony 1.2 on populations with low genetic variability (Ashley et al., 2008), the data should in
theory also be checked with the program Kinalyzer (Ashley et al., 2009), but this was omitted in this
thesis. Because sisters have a negative influence on the estimates of genetic parameters like allelic
richness (Wang, 2004), further analyses were made after removal of the identified sisters and
selecting 1 random individual (with the best scoring).
The allele frequencies of the different MS per population of each species have to be divided
according the Hardy-Weinberg principle. Under a Hardy-Weinberg equilibrium (HWE) the genotype
frequencies are expected to remain constant over generations in the absence of other evolutionary
influences. Meeting this equilibrium assumption is necessary because the calculation of population
37
genetic parameters is largely based on this criteria. Testing for genotype frequencies against the
HWE expectations was done with GenAlEx 6.501 (Peakall & Smouse, 2006).
When significant deviations from HWE are found, this could be indicative for excess homozygosity in
a population or presence of null alleles at certain loci in a population. Mainly in old museum
specimens, null alleles can reach high levels (Wandeler et al., 2007, Strange et al., 2009). Null alleles
are alleles of MS loci that fail to amplify during PCR. They are usually the result of a mutation in the
MS flanking region, to which the primers should anneal. The program MICROCHECKER 2.2.3 (van
Oosterhout et al., 2004) was used to check for evidence of null alleles and their frequencies.
Genotypic linkage disequilibrium between pairs of loci across populations of a species was tested
with randomization methods implemented in FSTAT 2.9.3 (Goudet, 2001). Linkage disequilibrium
implies a non-random association between alleles of different loci. In that case the statistical
associations between alleles at different loci are different from what would be expected if alleles
were independently and randomly sampled according to their individual allele frequencies. The
different MS loci used in this research were previously selected from different linkage groups, which
are representing different bumblebee chromosomes, by which linkage disequilibrium is unexpected.
2.3.1 Genetic diversity
For the adjusted dataset the genetic diversity could be estimated by calculating the allelic richness
(AR) and the expected and observed heterozygosity (HE & HO) for each MS locus. These parameters
were determined using the GenAIEx 6.501 software (Peakall en Smouse, 2006). HE in each population
was estimated with Nei’s unbiased expected heterozygosity (Nei, 1978). This statistic is unbiased by
sample size and appears not to be affected by null alleles (Chapuis et al., 2008). A paired Student’s t-
test in SPSS 23.0 could be performed to examine whether the mean genetic diversity and allelic
richness significantly differed between two populations within a species. An ANOVA with
Repeated Measures Factors was used to examine whether the means of genetic diversity
differed significantly between several populations of the species.
An estimation of the inbreeding coefficient was also made with the GenAIEx 6.501 add-in for
Microsoft Excel (Peakall & Smouse, 2006). A one sample t-test was executed in SPSS 23.0 to test
whether the inbreeding coefficients of each population significantly differed from 0. When a
population violated the normality assumption, a nonparametric test was applied (Wilcoxon Signed
Rank Test).
2.3.2 Population structure
Pairwise differentiation values (FST) among the different populations were calculated both in
GenAIEx 6.501 as in FSTAT 2.9.3 using 1000 permutations (Goudet, 2001). Jost’s D was also
estimated (Dest; Jost, 2008), using the online software SMOGD v1.2.5 (Crawford, 2010)
(http://www.ngcrawford.com/django/jost/). This recently developed statistic provides a true
measure of differentiation for highly variable markers, such as microsatellites.
38
3. Results
3.1 Melipona
3.1.1 Primer selection
Most amplifications with the Mbi primers were successful, while the Mru and Mmo primers did not
seem to produce any results on our species of interest, contrary to what was stated in literature. The
following primers proved to work best on the three species simultaneously: Mbi28, Mbi88, Mbi215,
Mbi233, Mbi201, Mbi278, Mbi305, Mbi218 and also Mbi32.
Primers for the MS loci Mbi28, Mbi88, Mbi215 and Mbi233 have the same annealing temperature
(57.5°C), and so do primers for Mbi201, Mbi218, Mbi278 and Mbi305 (60°C). So each group of four
primers could be combined in a multiplex PCR, respectively. However, first the chance of primer
dimer formation should be excluded. Primer dimers are a potential by-product in PCR. Primer dimers
are primers that have hybridized to each other, because of strings of complementary bases in their
sequences. When this occurs at the start of the polymerase chain reaction, the primer dimer will be
amplified. This will lead to competition for PCR reagents resulting in inhibition of target DNA
sequence amplification.
All combinations of forward and reverse primers for the four selected loci of each of the desired
multiplex combinations were tested with OligoAnalyser 3.1. OligoAnalyser examines the possibility of
an oligo to form duplexes by annealing to other sequences. A Delta G (kcal/mole) is calculated by
taking into account the longest stretch of complementary bases. In this way, the chance of
problematic primer dimer formation can be estimated. For the first combination (Mbi28, Mbi88,
Mbi215 and Mbi233), the Mbi215 primer seemed to result in slightly too negative Delta G values
when combined in a reaction with some of the other primers. Therefore this MS locus was replaced
by locus Mbi32, which also amplified well during the primer selection test and has a similar annealing
temperature as the other primers in that combination (57.5°C). After this validation, the following 2
multiplex combinations (Table 1) were concluded to be used in further research. Of course, the first
step in further research will be validation of these multiplexes. Only after proving their functionality,
they can be used for genotyping Melipona species.
Table 1: Proposal of Melipona multiplexes
Multiplex 1
MS locus Size Ta (°C) Forward primer Reverse primer
Mbi28 108 57.5 TTTTATCGCTCCTATCCTCC AATCCAACAGGACGGTGT
Mbi32 154 57.5 CTTTATCCGGTGCGTCGAA GAAGGCATTCCGGGTTGTT
Mbi88 94 57.5 GCCGCCGTACGTTCTGA GCGCTCGAGCAGCGTT
Mbi233 119 57.5 ACGAGCACGGGCCATAA GATCCATCGACCGCTTCTT
Multiplex 2
MS locus Size Ta (°C) Forward primer Reverse primer
Mbi201 152 60 GTTTAATCGCCCAAAGAGGC GTTGACGAGAAGGAGCACG
Mbi218 131 60 CTCGACTTAATTTCCATCGGC GCAATTTCAATCGCGACC
Mbi278 113 60 GTTCGTGTTTCGTGGTGAAT GTTGCGAGAACTCTGACGAT
Mbi305 106 60 GATCCGCTGCGCGAGAC GGACGAGGCTGAGGCATG
39
3.1.2 Dataset preparation
Table 2 summarizes the specimens of which tissue has been gathered for MS genotyping.
Nevertheless, as can be concluded from this table, it will be necessary to sample some more
populations in order to approach the Melipona research question in the future.
Table 2: Overview of Melipona specimens
Species State Location Year of
collection Number of specimens
M. bicolor schencki
Rio Grande do Sul
São Francisco de Paula
1995 14
1997 25
1999 18
2009 15
2012-2013 12
Riozinho 2009 17
Rolante 2015 19
M. quadrifasciata quadrifasciata
Rio Grande do Sul São Francisco de Paula 1997-1998 25
Rolante 2015 20
Santa Catarina São Martinho 1984 25
São Paulo São Paulo 1988 6
Ribeirão Preto 1989 5
M. torrida
Rio Grande do Sul
Canela 1984 8
Plantalto Parque Nonoai 1985 10
Osório 1992-1993 10
São Francisco de Paula
1995 23
1997 25
1999 24
Bom Jesus 2009 14
Vacaria 2009 8
Rolante 2015 20
São Paulo Caxias do Sul Cap. São
Paulo 1984 10
Santa Catarina São Martinho 1984 15
40
3.2 Bombus
3.2.1 Dataset preparation
3.2.1.1 Definition of populations
First of all, every Bombus population is coded, in order to have a clear overview throughout the next
sections. A population is named after the location and within which time period it was sampled
(Table 3 & Table 4).
Table 3: Coding of time periods
Time period Code
1946-1959 1
1991-1994 2
1999-2004 3
2010-2015 4
Table 4: Coding of locations
Location Code
Porto Alegre + Viamão + Guaíba POA
Cambará do Sul + Torres CATO
Osório + Capão da Canoa OSCA
Nova Petropolis + São Francisco de Paula NPSFdP
Candiota CAN
Santa Cruz do Sul SCdS
Esmeralda ESM
Guarani das Missões GdM
As noted before, the specimens from the location Nova Petropolis + São Francisco de Paula date
from the years 1996-1997, which aren’t included in any of the time periods, but are retained as an
extra information point.
3.2.1.2 Detection of sisters
For every population of each species, Colony 1.2 generated an output, which summarizes the several
detected groups within this population. Group detection is based on the estimation of possible
mother and father genotypes for all specimens of the population. Each group represents a colony to
which genetic similar individuals will be assigned to. Thus, individuals assigned to the same group are
considered to be sisters originating from the same colony. When sisters are detected, only one
individual from that colony can be retained in the dataset for further analysis. This will be the
individual with the least missing data, or when there are several well scored individuals, one is
randomly selected.
In Table 5, the column ‘n without sisters’ shows the number of individuals retained after exclusion of
sisters.
41
Table 5: Overview of the number of individuals per population
B. pauloensis
Population n1 n without sisters n final
POA1 13 13 10
POA2 8 7 7
CATO2 6 6 5
OSCA2 9 9 6
NPSFdP 25 18 18
CAN3 19 6 6
POA3 6 6 6
SCdS8 8 8 6
ESM4 22 18 17
GdM4 33 29 28
POA4 13 12 12
B. morio
Population n1 n without sisters n final
POA1 5 5 5
CATO2 6 6 6
POA2 3 3 0
OSCA2 9 9 8
NPSFdP 19 16 16
SCdS3 20 17 17
POA3 15 5 5
POA4 12 11 11
GdM4 6 5 5
SCdS4 5 5 5
B. bellicosus
Population n1 n without sisters n final
CAN3 50 23 22
3.2.1.3 Linkage disequilibrium
On the resulting dataset, and thus without the detected sisters, a test for linkage disequilibrium was
executed with FSTAT 2.9.3. This software generates an output displaying p-values for genotypic
disequilibrium between all locus pairs across all populations of a species. The p-values for genotypic
disequilibrium (based on 12000 permutations) were never smaller than the adjusted p-value for a 5%
nominal level (0.000417). This indicates that there was no evidence of significant linkage
disequilibrium for any of the locus pairs in the three species.
3.2.1.4 Exclusion of loci or individuals from the data analysis
To obtain the final sets of genotyping data for all three bumblebee species, MS loci that exhibited
excessive amounts of missing data, and/or specimens which could not be scored in a reliable manner
for a minimum number of loci were excluded.
The MS loci BT08 and 0294 didn’t amplify or were impossible to score in many B. morio individuals.
Therefore the genotypes of these two loci were excluded, also for B. pauloensis and B. bellicosus, to
42
keep comparison of diversity based on MS loci between the species possible. So, in the final dataset
genotypes of 14 MS loci were present.
Specimens with missing data for more than 5 out of 14 loci, were also excluded. A lot of missing data
could indicate for a lower DNA quality or amplification problems for a specimen, which makes
outcomes less reliable. A sensitivity analysis of HE for B. pauloensis was performed (Table 2 in
Supplementary data). As less missing data is allowed, the dataset will shrink and some population
sizes reduce drastically. Mainly the old populations become negligibly small and would have to be left
out of the further analysis. In turn this would mean that the comparison of genetic diversity through
time is not possible anymore. Here, the threshold is an acceptance of missing data for 4 loci, as this is
also the number of MS loci within a primer mix. On the other hand, while obviously the diversity
parameter HE mainly decreases with the number of individuals, it seems to stay relatively stable as
long the number of individuals doesn’t become too small.
For B. morio individuals, a less strict maximum number of missing data was allowed, to ensure
retention of populations. Only individuals with more than 50% missing data were excluded. However,
this indulgence could lead to obstacles in data analysis or less trustworthy results later on.
Finally, populations containing less than 5 individuals (B. morio population POA2) were also dropped
out of analysis. The last column of Table 5 contains the final number of individuals within each
population.
3.2.1.5 Hardy-Weinberg equilibrium
Testing for genotype frequencies against Hardy-Weinberg equilibrium (HWE) expectations displayed
several loci with heterozygote deficits. Since this could be indicative for the presence of null alleles,
the null allele frequencies were determined at every locus of every population deviating from the
HWE. However for most small populations null alleles cannot be calculated, as MICROCHECKER 2.2.3
is not able to estimate the null allele frequencies when there are insufficient samples in a population.
When the deviations from HWE cannot be attributed to the presence of null alleles, they could be
interpreted as an indication of inbreeding. This was the case for some loci of B. pauloensis
populations POA1, NPSFdP, POA3, and SCdS3, for some loci of B. morio populations SCdS3 and
NPSFdP, and for the B. bellicosus population CAN3 (Table 1 in Supplementary data). When estimating
the inbreeding coefficient later on, it can be determined if significant inbreeding is actually present in
these populations.
Furthermore, many loci were monomorphic, implying that for these loci only one allele was present
in that population. In this case, the deviation from the HWE cannot be calculated and no conclusion
can be drawn regarding inbreeding.
43
3.2.2 Estimation of parameters for genetic diversity and inbreeding
3.2.2.1 B. pauloensis
Table 6: N = number of individuals per population, AR = allelic richness, HO = observed heterosygosity, HE = expected heterozygosity, FIS = inbreeding coefficient, SE = standard error, * = significantly different from (p < 0.05)
Species Population n AR SE HO SE HE SE FIS SE
B. pauloensis
POA1 10 3.243 0.476 0.379 0.063 0.617 0.076 0.270* 0.086
POA2 7 2.312 0.403 0.498 0.095 0.477 0.081 -0.158 0.095
CATO2 5 2.328 0.345 0.281 0.045 0.519 0.075 0.278* 0.097
OSCA2 6 3.075 0.561 0.498 0.085 0.580 0.080 0.070 0.060
NPSFdP 18 2.950 0.707 0.437 0.074 0.510 0.074 0.127 0.075
CAN3 6 2.588 0.566 0.393 0.085 0.478 0.082 0.061 0.097
POA3 6 2.206 0.475 0.351 0.099 0.399 0.082 0.126 0.131
SCdS3 6 2.038 0.324 0.214 0.062 0.386 0.090 0.380* 0.103
ESM4 17 2.531 0.560 0.407 0.110 0.385 0.092 0.057 0.107
GdM4 28 2.932 0.813 0.347 0.090 0.374 0.097 0.034 0.030
POA4 12 2.552 0.695 0.350 0.102 0.349 0.097 -0.014 0.070
mean
2.614 0.166 0.378 0.026 0.461 0.026 0.113 0.029
As can be deduced from Table 6, only three B. pauloensis populations show an inbreeding coefficient
significantly different from zero: POA1, CATO2 and SCdS3. Although the POA1 population shows signs
of inbreeding, it is still characterized by relatively high diversity parameters (AR and HE). As HO is much
smaller than HE, it is this difference which is causing the inbreeding coefficient to be significant.
Although the HE of CATO2 and SCdS3 are not particularly small, they have very small HO values,
resulting in inbreeding. The particularly high inbreeding coefficient of SCdS3 could be a sign of a
severe isolation of this population, driving it towards inbreeding. Negative FIS values could indicate
outbreeding.
In order to detect a change in genetic diversity of B. pauloensis trough time, the mean HE and AR of
the four time periods were compared. For a particular time period the HE or AR value of a certain
locus is the mean of the HE or AR values of that locus over all the populations within that period. This
way a value of HE or AR for each of the 14 MS loci can be obtained for each time period.
Table 7: Mean HE of B. pauloensis populations within time periods
Measure: HE
Timeperiod Mean SE 95% Confidence Interval
Lower Bound Upper Bound
1946-1959 0.617 0.076 0.453 0.781
1991-1994 0.525 0.074 0.367 0.684
1999-2004 0.421 0.078 0.253 0.589
2010-2015 0.369 0.094 0.165 0.573
The mean HE seems to be diminishing through time (Table 7). Whether there is a significant overall
difference between HE means of the different time periods and between which time periods exactly
44
this difference is significant, was determined with an ANOVA Repeated Measures test. A significant
overall difference between the HE means over the different time periods was observed (F = 3.983, df
= 3, p = 0.03 ). The significant difference was situated between time periods 1 and 3 (p = 0.03),
between time periods 1 and 4 (p =0.,015), T2 and T3 (p = 0.011), and also between T2 and T4 (p =
0.013) (with no adjustment of significance for multiple comparisons). When applying an adjustment
for multiple comparisons (Bonferroni), the differences between the old and more recent time
periods were no longer significant.
To rule out the possibility that a certain location biases the estimated genetic diversity, and thus
could cause the observed significant difference throughout time, it is necessary to check whether the
evolution seen for the time periods also occurs within one location. For this purpose also the HE
means of population POA1, POA2, POA3 and POA4 are compared.
Table 8: Mean HE of B. pauloensis populations from location POA
Measure: HE
Population Mean SE 95% Confidence Interval
Lower Bound Upper Bound
POA1 0.617 0.076 0.453 0.781
POA2 0.477 0.081 0.302 0.652
POA3 0.399 0.082 0.221 0.576
POA4 0.349 0.097 0.139 0.559
Here the mean HE also seems to be diminishing through time (Table 8). There is a significant overall
difference between HE means of the different time periods present (F = 3.695, df = 3, p = 0.046). The
significant difference was situated between time periods 1 and 3 (p = 0.024), T1 and T4 (p = 0,011),
and T2 and T4 (p = 0.035) (when not adjusting the significance for multiple comparisons). When
applying an adjustment for multiple comparisons (Bonferroni), the differences between the old and
more recent populations were no longer significant.
A similar procedure is carried out for the parameter allelic richness AR.
Table 9: Mean AR of B. pauloensis populations within time periods
Measure: AR
Timeperiod Mean SE 95% Confidence Interval
Lower Bound Upper Bound
1946-1959 3.243 0.476 2.214 4.272
1991-1994 2.572 0.389 1.730 3.413
1999-2004 2.277 0.419 1.373 3.181
2010-2015 2.672 0.686 1.190 4.154
The mean AR seems to be decreasing through time, but increases again in the last time period (Table
9). There was a significant overall difference between AR of the different time periods present
(F(3,11) = 5.003, p = 0.020). The significant difference was situated between time periods 1 and 3 (p =
45
0.023) (when not adjusting the significance for multiple comparisons). When applying an adjustment
for multiple comparisons (Bonferroni), this difference was no longer significant.
To rule out the possibility that a certain location with an outlying genetic diversity, causes the
significant difference between the two time periods, it is necessary to check whether the decreasing
trend is also visible through time within one location. For this purpose the AR means of population
POA1, POA2, POA3 and POA4 are compared.
Table 10: Mean AR of B. pauloensis populations from location POA
Measure: AR
Population Mean SE 95% Confidence Interval
Lower Bound Upper Bound
POA1 3.243 0.476 2.214 4.272
POA2 2.312 0.403 1.440 3.183
POA3 2.206 0.475 1.179 3.233
POA4 2.552 0.695 1.050 4.055
Here the mean AR also seems to be decreasing through time and to increase again in the last time
period (Table 10). Nevertheless, there was no significant overall difference between AR means of the
different time periods present (F(3,11) = 1.934, p = 0.183).
It is also informative to explore the difference in genetic diversity between the various locations
within a time period. All time periods comprise more than 2 locations, so they are compared with an
ANOVA Repeated Measures (Table11 & Table 12).
Table 11: Significance of difference in mean HE between pairs of locations within the same time period
Timeperiod Overall difference in mean HE Pairs of compared locations
1991-1994 ns (F(2,12) = 2.424, p = 0.131) POA2 & CATO2 POA2 & OSCA2 CATO2 & OSCA2
1999-2004 ns (F(2,12) = 1.672, p = 0.229) CAN3 & POA3 CAN3 & SCdS3 POA3 & SCdS3
2010-2015 ns (F(2,12) = 1.003, p = 0.396) ESM4 & GdM4 ESM4 & POA4 GdM4 & POA4
Table 12: Significance of difference in mean AR between pairs of locations within the same time period
Timeperiod Overall difference in mean AR Pairs of compared locations
1991-1994 ns (F(2,12) = 2.233, p = 0.150) POA2 & CATO2 POA2 & OSCA2 CATO2 & OSCA2
1999-2004 ns (F(2,12) = 1.241, p = 0.324) CAN3 & POA3 CAN3 & SCdS3 POA3 & SCdS3
2010-2015 ns (F(2,12) = 3.202, p = 0.077) ESM4 & GdM4 ESM4 & POA4 GdM4 & POA4
These results justify the fact that populations from different locations are grouped within one time
period.
Furthermore, as will be presented later on in the section about estimation of population structuring,
no time period demonstrated significant structuring between its locations. This also indicates that
locations could be pooled.
46
3.2.2.2 B. morio
Table 13: N = number of individuals per population, AR = allelic richness, HO = observed heterozygosity, HE = expected heterozygosity, FIS = inbreeding coefficient, SE = standard error, * = significantly different from 0
Species Population n AR SE HO SE HE SE FIS SE
B. morio
POA1 5 2.886 0.498 0.390 0.085 0.593 0.088 0.235 0.118
CATO2 6 2.758 0.289 0.425 0.082 0.633 0.069 0.266* 0.106
OSCA2 8 3.791 0.542 0.519 0.095 0.708 0.065 0.242 0.122
NPSFdP 16 3.930 0.700 0.469 0.069 0.645 0.062 0.267* 0.070
SCdS3 17 4.623 0.967 0.370 0.063 0.628 0.076 0.400* 0.068
POA3 5 3.589 0.558 0.507 0.079 0.682 0.077 0.147 0.099
POA4 11 3.851 0.754 0.559 0.094 0.566 0.092 -0.024 0.044
GdM4 5 3.140 0.601 0.514 0.098 0.530 0.101 -0.101 0.052
SCdS4 5 2.866 0.640 0.519 0.099 0.530 0.094 -0.147 0.088
Mean
3.493 0.213 0.475 0.028 0.613 0.027 0.151 0.033
From Table 13 it can be concluded that three populations of B. morio have an inbreeding coefficient
(FIS) significantly differing from zero. Although the SCdS3 population shows signs of inbreeding, it is
still characterized by relatively high diversity parameters (AR and HE). On the other hand the observed
heterozygosity of that population is very low. The difference between the high HE and low HO causes
inbreeding. For time period 4, the populations seem to be rather in a state of outbreeding. However
this observation is not significant. The inbreeding coefficient of the SCdS3 population of B. morio is
high, as it is also high for the SCdS3 population of B. pauloensis, which indicates that these
populations were isolated, driving them to inbreeding.
In order to detect a change in genetic diversity of B. morio through time, the mean HE and AR (over all
populations within time period) of the four time periods were compared.
Table 14: Mean HE of B. morio populations within time periods
Measure: HE
Timeperiod Mean SE 95% Confidence Interval
Lower Bound Upper Bound
1946-1959 0.593 0.088 0.403 0.782
1991-1994 0.670 0.064 0.532 0.808
1999-2004 0.655 0.073 0.498 0.812
2010-2015 0.542 0.091 0.344 0.740
The mean HE seems to be diminishing through time since time period 2 (Table 14). Nevertheless,
there is no significant overall difference between HE over the different time periods present (F(3,11)
= 1.901, p = 0.188). Time period 1 doesn’t fit within the overall trend of decrease and has a lower
mean HE than time period 2. This could be attributed to the fact that this time period comprises only
one single population of just 5 individuals.
47
Therefore the ANOVA Repeated Measures was repeated by leaving time period 1 from the analysis.
This did not influence the result as still no significant overall difference between HE means of the
different time periods was present (F(2,12) = 2.670, p = 0.110).
Although there is no significant difference throughout time, the HE means of B. morio populations
POA1, POA3 and POA4 are also compared, to explore whether the diminishing trend since time
period 2 is also visible or significant through time within one location (POA). Although the mean HE
seems to have decreased between period 3 and 4, no significant overall difference between HE over
the different time periods is present (F(2,12) = 2.568, p = 0.118) (Table 15).
Table 15: Mean HE of B. morio populations from location POA
Measure: HE
Population Mean SE 95% Confidence Interval
Lower Bound Upper Bound
POA1 0.593 0.088 0.403 0.782
POA3 0.682 0.077 0.515 0.849
POA4 0.566 0.092 0.368 0.763
A similar procedure is carried out for the parameter allelic richness AR.
Table 16: Mean AR of B. morio populations within time periods
Measure: AR
Time period Mean SE 95% Confidence Interval
Lower Bound Upper Bound
1946-1959 2.886 0.498 1.810 3.961
1991-1994 3.275 0.390 2.432 4.118
1999-2004 4.106 0.746 2.495 5.717
2010-2015 3.286 0.627 1.932 4.640
The mean AR seems to be increasing between time period 1 and 3, but drops again in time period 4
(Table 16). Whether there is a significant overall difference between AR over the different time
periods and between which time periods exactly this difference is significant, was determined with
an ANOVA repeated Measures test (with no adjustment for multiple comparisons).This test
concluded that there is no real significant overall difference between AR means over the different
time periods (F(3,11) = 2.879, p = 0.084).
Although difference throughout time is not significant, the AR means of B. morio populations POA1,
POA3 and POA4 are compared, to explore whether the increasing trend between periods 1 and 3 is
also visible but not significant within one location (POA). The mean AR seems to have increased
between time period 1 and 4, without a drop for the most recent population (POA4) (Table 17).
However, the difference between AR over the different time periods is not significant (F(2,12) =
1.089, p = 0.368).
48
Table 17: Mean AR of B. morio populations from location POA
Measure: AR
Population Mean Std. Error 95% Confidence Interval
Lower Bound Upper Bound
POA1 2.886 0.498 1.810 3.961
POA3 3.589 0.558 2.383 4.795
POA4 3.851 0.754 2.223 5.480
It is also informative to explore a difference in genetic diversity between the various locations within
a time period (Table 18 & Table 19). When a time period comprises more than 2 locations, they are
compared with an ANOVA Repeated Measures. When it includes only 2 locations, a paired t-test is
used (or Wilcoxon Signed Rank Test when normality assumptions are violated).
Table 18: : Significance of difference in mean HE between pairs of locations within the same time period
Timeperiod Overall difference in mean HE Pairs of compared locations
1991-1994 ns (Z = -1.852, p = 0.064) OSCA2 & CATO2
1999-2004 ns (Z = -0.973, p = 0.331) POA3 & SCdS3
2010-2015 ns (F(2,12) = 0.944, p = 0.416) SCdS4 & GdM4 SCdS4 & POA4 GdM4 & POA4
These results justify the fact that populations from different locations are grouped within one time
period.
Table 19: Significance of difference in mean AR between pairs of locations within the same time period
Timeperiod Overall difference in mean AR Pairs of compared locations
1991-1994 sign (t(13) = 2.709, p = 0.018) OSCA2 & CATO2*
1999-2004 ns (Z = -1.726, p = 0.084) POA3 & SCdS3
2010-2015 sign (F(2,12) = 4.233, p = 0.041) SCdS4 & GdM4 SCdS4 & POA4 GdM4 & POA4*
These results show that it might not be justified to group locations within one time period, since their
mean allelic richness differs significantly.
3.2.2.3 B. bellicosus
Table 20: N = number of individuals per population, AR = allelic richness, HO = observed heterozygosity, HE = expected heterozygosity, FIS = inbreeding coefficient, SE = standard error, * = significantly different from 0
Species Population N AR SE HO SE HE SE FIS SE
B. bellicosus CAN3 22 2.818 0.598 0.351 0.064 0.474 0.077 0.268* 0.075
Within the one population of B. bellicosus which was sampled, there was significant inbreeding
present (t(13) = 3.573, p = 0.003) (Table 20).
49
3.2.2.4 Comparison between the three bumblebee species
Within each time period the diversity of the different species were here compared (Table 21).
Significance of possible differences in mean HE or AR between species was obtained with a Mann
Whitney U test.
Table 21: Comparison of genetic diversity between species, ns = not significant, * = significant
Time period Compared species Mean HE Sign difference Mean AR Sign difference
1946-1959 B. pauloensis 0.617
ns 3.243
ns B. morio 0.593 2.886
1991-1994 B. pauloensis 0.525
ns 2.572
ns B. morio 0.670 3.275
1999-2004
B. pauloensis 0.421 *
2.277 *
B. morio 0.655 4.106
B. bellicosus 0.474 ns
2.818 ns
B. pauloensis 0.421 2.277
B.bellicosus 0.474 ns
2.818 ns
B. morio 0.655 4.106
2010-2015 B. pauloensis 0.369
ns 2.672
ns B. morio 0.542 3.286
The mean HE and AR of B. morio are generally higher than the HE and AR of the other species, only in
time period 1, B. pauloensis has a higher HE and AR. These higher HE and AR of B. morio are only in one
case significant, in time period 3, B. morio has a significantly higher HE and AR than B. pauloensis (Z = -
2.205, p = 0.027 and Z = -2.205, p = 0.027 respectively). B. bellicosus was only sampled in time period
3 and has a higher HE than B. pauloensis but lower than B. morio. B. bellicosus was only sampled from
one location (CAN3), together with B. pauloensis. The mean HE of B. bellicosus (0,474) at location
CAN in time period 3 is not significantly different from the mean HE of the CAN3 population of B.
pauloensis (0,478) (t(26) = 0.040, p = 0.968). On the other hand, as noted before, the B. bellicosus
population showed significant inbreeding, while B. pauloensis didn’t. At this location, both species
could not be observed during the recent sampling attempts for this research.
3.2.2.5 Conclusion
The presumption of inbreeding deduced from HWE deviations combined with no evidence for high
null allele frequencies, was confirmed by a significant inbreeding coefficient in B. pauloensis
populations POA1 and SCdS3, in B. morio populations SCdS3 and NPSFdP and in the B. bellicosus
population CAN3.
B. pauloensis showed a clear decreasing trend in HE (Figure 1). This trend was visible both within one
location and within the pooled locations. Because the populations within a time period seemed to
have comparable HE values, it could be assumed as justified to pool the locations sampled within a
time period. The allelic richness also declined within a period of 50 years, although it seems to
slightly increased again recently (Figure 2). This trend was less substantial for the particular location
POA. The data from the population NPSFdP (1996-1997) fit more or less within the trends which
were generally observed for B. pauloensis.
50
The HE of B. morio seems to have declined since the 90s (Figure 1), but this could not be statistically
confirmed. The HE of the first time period (1946-1959) was lower than that of the 90s. Period 1 is
spread over more than 10 years and comprises only one location with 5 individuals, so possibly the
HE value calculated is not very indicative anyway. On the other hand, the allelic richness of B. morio
appears to have increased at the POA location, but also over time in the pooled locations (Figure 2).
Nonetheless, this was not proved statistically. The locations within some time periods showed
significant differences in mean AR, so possibly it actually was not correct to pool them in the first
place. Also for B. morio the data from the population NPSFdP (1996-1997) fit more or less within the
trends which were generally observed. Significant inbreeding was found in the NPSFdP population of
B. morio.
When comparing the three bumblebee species, it appears to be sound to say that B. morio is
generally more genetically variable (higher AR and HE) than the other two species (Figure 1 & Figure
2). Nevertheless this was only once statistically approved, namely in time period 3. Only in time
period 1 B. pauloesis displayed higher HE and AR values than B. morio, but this could possibly be
attributed to the fact that the number of B. pauloensis individuals (10) is here twice the number of B.
morio individuals (5).
Finally, B. bellicosus was unfortunately only sampled in time period 3. This will make any conclusion
regarding this species rather inconvenient. In general for time period 3, B. bellicosus has a higher HE
and AR than B. pauloensis but lower than B. morio. B. bellicosus was actually sampled from only one
location (CAN3), together with B. pauloensis. The mean HE and AR of B. bellicosus at that location
were higher than the genetic diversity parameters of the B. pauloensis population, but the difference
was not significant. On the other hand, as displayed before, the B. bellicosus population shows
significant inbreeding, while B. pauloensis does not. At this location, both species could not be
observed during the sampling attempts for this research.
Figure 1: Evolution of mean HE of the three bumblebee species
0
0,1
0,2
0,3
0,4
0,5
0,6
0,7
1946-1959 1991-1994 1999-2004 2010-2015
Me
an H
E
Time period
B. pauloensis
B.morio
B. bellicosus
51
Figure 2: Evolution of mean AR of the three bumblebee species
3.2.3 Estimation of population structure
3.2.3.1 B. pauloensis
Table 22 presents the pairwise population FST values estimated by GenAlEx 6.501 under the diagonal
and the harmonic mean of Dest above the diagonal. Both methods have a different way of estimating
the structuring parameter, resulting in different values above and under the diagonal. More
important is to notice that the values show a similar trend.
Table 22: Pairwise population Fst values estimated with GenAlEx 6.501 for B. pauloensis
POA1 POA2 CATO2 OSCA2 NPSFdP CAN3 POA3 SCdS3 ESM4 GdM4 POA4
- 0.054 0.138 0.002 0.033 0.013 0.077 0.303 0.260 0.233 0.162 POA1
0.114 - 0.009 0.004 0.001 0.015 0.000 0.013 0.012 0.012 0.007 POA2
0.126 0.067 - 0.005 0.027 0.023 0.039 0.093 0.048 0.064 0.061 CATO2
0.089 0.051 0.085 - 0.009 0.012 0.002 0.031 0.015 0.002 0.001 OSCA2
0.088 0.027 0.061 0.026 - 0.002 0.015 0.038 0.042 0.028 0.029 NPSFdP
0.116 0.042 0.077 0.050 0.027 - 0.002 0.005 0.006 0.001 0.005 CAN3
0.174 0.061 0.099 0.097 0.061 0.065 - 0.007 0.010 0.014 0.001 POA3
0.201 0.076 0.122 0.090 0.073 0.068 0.073 - 0.005 0.013 0.007 SCdS3
0.190 0.061 0.089 0.081 0.058 0.060 0.051 0.040 - 0.000 0.000 ESM4
0.189 0.059 0.094 0.082 0.056 0.058 0.053 0.036 0.012 - 0.000 GdM4
0.198 0.066 0.109 0.092 0.065 0.063 0.061 0.037 0.021 0.014 - POA4
The pairwise differentiation values (FST) between the different populations were also estimated using
1000 permutations in FSTAT 2.9.3, but they are not displayed here (Table 3 in Supplementary data).
FSTAT 2.9.3 simultaneously generates an output file with p-values (obtained after 1100
permutations), indicating whether the FST values in the matrix are significantly different from zero.
0,00
0,50
1,00
1,50
2,00
2,50
3,00
3,50
4,00
4,50
1946-1959 1991-1994 1999-2004 2010-2015
Me
an A
R
Time period
B. pauloensis
B. morio
B. bellicosus
52
The indicative adjusted nominal level (5%) for multiple comparisons was 0.0009. For B. pauloensis
there was no significant structuring found between any of the population pairs.
Nonetheless, it should be noted that Dest reaches quite high values for some particular pairs of
populations (POA1xSCdS3, ESM4xPOA1, GdM4xPOA1). Since this parameter is an estimator of actual
differentiation, one can assume that these high values indicate true structuring between the old
POA1 population and the more recent populations SCdS3, ESM4 and GdM4.
3.2.3.2 B. morio
Table 23 presents the pairwise population FST values estimated by GenAlEx 6.501 under the diagonal
and the harmonic mean of Dest above the diagonal.
Table 23: Pairwise population Fst values estimated with GenAlEx 6.501 for B. morio
SCdS4 GdM4 POA4 POA3 SCdS3 NPSFdP OSCA2 CATO2 POA1
- 0.000 0.001 0.000 0.010 0.077 0.216 0.168 0.475 SCdS4
0.076 - 0.009 0.000 0.017 0.048 0.135 0.167 0.314 GdM4
0.052 0.033 - 0.030 0.033 0.047 0.161 0.175 0.333 POA4
0.092 0.072 0.054 - 0.022 0.053 0.017 0.024 0.171 POA3
0.066 0.061 0.034 0.052 - 0.055 0.230 0.174 0.414 SCdS3
0.120 0.103 0.081 0.063 0.052 - 0.046 0.019 0.150 NPSFdP
0.195 0.174 0.150 0.103 0.114 0.050 - 0.029 0.008 OSCA2
0.190 0.173 0.148 0.105 0.119 0.058 0.069 - 0.094 CATO2
0.283 0.259 0.236 0.187 0.194 0.112 0.076 0.117 - POA1
The pairwise differentiation values (FST) between the different populations were also estimated using
1000 permutations in FSTAT 2.9.3, but are not displayed here (Table 4 in Supplementary data).
Unfortunately, for populations with too few (effective) individuals FSTAT appears to be unable to
generate the p-value matrix with significances. Therefore populations containing less than 6
individuals were left out of analysis and the pairwise differentiation values (FST) (estimated using
1000 permutations in FSTAT 2.9.3) were generated again for this reduced number of populations
(Table 24). For these 5 populations of B. morio there was no significant structuring found between
any of the population pairs (indicative adjusted nominal level (5%) for multiple comparisons: 0.005).
Dest reaches quite high values for some particular pairs of populations (OSCA2xSCdS4, POA1xSCdS4,
POA1xGdM4, POA1xPOA4, POA1xSCdS3, and OSCA2xSCdS3) (Table 23). Since this parameter is an
estimator of actual differentiation, one can assume that these high values indicate true structuring
between the older populations and the more recent populations.
Table 24: Pairwise differentiation values (FST) under the diagonal, * indicating value significantly differing from 0
POA4 SCdS3 NPSFdP OSCA2 CATO2
-
POA4
0.012 -
SCdS3
0.091 0.052 -
NPSFdP
0.178 0.120 0.018 -
OSCA2
0.181 0.119 0.020 0.004 - CATO2
53
4. Discussion The objective of this thesis was to study the genetic variability of the South-Brazilian bumblebee
species B. bellicosus, of which local extinction has been reported in literature, and to compare it with
the genetic diversity found in two more stable bumblebee species occurring in the same region, B.
pauloensis and B. morio. Secondly a project regarding Melipona stingless bees in Southern Brazil was
started up. The aim of this project is to study the genetic profiles of three threatened Melipona
species, in order to uncover possible gene flow between several South-Brazilian states, artificially
imposed by meliponiculture. For both questions, microsatellite markers were used to genotype the
populations.
4.1 Collection of Brazilian wild bees First of all, to collect the necessary data for this research, I personally travelled to Porto Alegre in the
Southern-Brazilian state Rio Grande do Sul, from mid October till mid December. During this period,
it is spring and early summer in Southern Brazil. At this time of the year, bumblebee queens normally
get out of dormancy, start to found their nests, and produce progeny. Unfortunately this year, the
season had shifted. During my stay in Southern Brazil almost no bumblebees were seen in the wild,
only at the end of my stay some started to appear. After speaking to some Southern Brazilian bee
specialists, it seems that the absence of bumblebees has been noticed since several years. This
phenomenon becomes visible by a reduction in their abundance, due to less observations during the
last years. Their reduction could be part of the general, worldwide decline as discussed before, but
specifically for this year there should be an explanation for their slow establishment. Only during the
month April, the Brazilian colleagues were able to collect some more workers.
The weather conditions end 2015 were more extreme than usual in Southern Brazil. The region
encountered a lot of rain and floods. These could be possibly caused by the meteorological
phenomenon El Niño, which is going on in 2015 and 2016. El Niño refers to the periodic warming of
the eastern equatorial Pacific Ocean that increases sea surface temperatures above average
(https://wunderground.atavist.com/el-nino-forecast). El Niño conditions can last for a year or two
and develop concurrently with atmospheric changes leading to a variety of global effects. El Niño
develops about every 2 to 6 years. The eastern tropical Pacific continually oscillates between warmer
and cooler conditions. An El Niño event is thus not abnormal, but strong El Niños are more unusual
and infrequent, while super El Niños are even less common. This El Niño is classified “super”, and
joins only two other super El Niño events in the past 65 years: 1982-1983 and 1997-1998. La Niña is
its counterpart, characterized by the abnormal cooling of the eastern Pacific Ocean.
El Niño pushes west winds and warm water across the tropical Pacific, toward South America. This
ocean-atmosphere circulation of El Niño can boost the amount of rain that is produced in some
regions. A classic El Niño event will intensify during the northern fall (southern spring), peak in the
winter (southern summer) and decrease in the spring (southern fall).
During the spring months (September-November) of 2015, frequent storm systems have brought
flooding rain to parts of Argentina, Uruguay and southeast Brazil. This pattern continued throughout
the summer (http://www.accuweather.com/en/weather-news/south-america-summer-forecast-
2015-2016/53158136). Rainfall averaged 100-250 percent of normal from August through the middle
of October. Slow-moving cold fronts were the primary weather feature from northern Argentina
through Uruguay, southeast Brazil and Paraguay, and caused heavy rainfall and occasional severe
54
thunderstorms with damaging winds and hail. Frontal boundaries often stalled across this region
resulting in several days of locally heavy rainfall. Warm air provided the fuel for strong
thunderstorms as the fronts initially pushed northward.
Porto Alegre was one of the cities at risk for occasional severe weather events during the summer
months. Flooding events were common resulting in numerous impacts throughout the region of
Southern Brazil. Nearly 200 cities in states the Rio Grande do Sul, Santa Catarina, Parana and Mato
Grosso do Sul have declared states of emergency due to flooding from rains
(http://www.reuters.com/article/us-brazil-farming-weather-idUSKCN0V62SS).
Now how could these altered weather conditions have affected the bumblebee abundance during
the spring of 2015 in the state of Rio Grande do Sul? Excess rainfall is harmful for vegetation and can
result in a lack of available flowering plants. Although plants need water, their roots also absorb
oxygen from the soil. When soil is filled with water for an extended period, plants can suffer and
produce less flowers, which are necessary food resources for bumblebees. Generally, extremely wet
weather also keeps pollinators away from the flowers. It has been suggested by Peat & Goulson
(2005) that weather strongly determines whether bees collect pollen, with pollen mainly being
collected on warmer, windy days with low humidity. According to them, this pattern may have a
common explanation. Pollen is likely to be difficult to gather when there are water droplets on the
flower or on the bee, such as morning dew or following rainfall. Dry conditions favour anther
dehiscence, meaning that more pollen is likely to be available. Given that pollen is a vital resource
without which bumblebees can rear no offspring, the availability of suitable dry conditions for pollen
collection may be a critical factor affecting the success of bumblebee nests. On the other hand, rates
of foraging for collecting nectar are often positively correlated with humidity, perhaps because
nectar secretion rates are higher or evaporation of nectar lower at high humidity. Low humidity may
place plants under water stress, so that nectar production is reduced but its sugar concentration will
increase by higher evaporation. In that case foraging rate is lower but the sugar concentration of the
nectar is greater. Sanderson et al. (2015) also concluded that forager activity was negatively
associated with rainfall, humidity and wind-speed.
4.2 Initialisation Melipona project In this thesis, a first attempt was made to initiating a study to assess if gene flow within Melipona
populations is caused by stingless bee keeping and /or by trade of colonies between different states.
As a result, a proposal of several effective MS markers was presented here together with a nearly
complete set of tissue samples which could be used to answer this research question in the near
future.
The next step in further research will be the validation of the proposed MS multiplexes. Only after
proving their functionality, they should be used for extensive genotyping in the Melipona species.
Hence, other polymorphic MS markers could complement the evaluated MS loci in this thesis.
Testing of more available Melipona primers, such as the 9 MS primers for M. quadrifasciata
developed by Tavares et al., (2013) could be performed to investigate the possibility of these MS to
be used in additional multiplexes. These multiplexes could then increase the power of genetic
diversity parameter estimations within the proposed population genetic research on the Melipona
species of Southern Brazil.
55
Furthermore, it will be necessary to sample some more populations in order to approach the
Melipona research question in the future. For M. bicolor schenki, specimens from populations from
Rio Grande do Sul neighbouring states like Santa Catarina will be needed. Extra sampling should
include both old data from museum collections and more recent specimens from the wild, in the
case the latter has not been carried out lately for this species in these states. For Rio Grande do Sul it
would also be preferable to collect more data from before the 90s, both for M. bicolor schencki and
M. quadrifasciata quadrfasciata.
Intense beekeeping activities, including the exchange of genetic material among beekeepers,
possibly causes a homogenization of populations. This is of course done with the best intentions of
keeping the populations of these species viable, but also implies a gene flow that in natural
circumstances would not occur. Whether this took place and in which degree, will be determined in
the future.
4.3 Genetic parameters of bumblebees in Rio Grande do Sul First of all, 14 of the 16 microsatellite markers (developed for European bumblebee species) were
successfully introduced in genetic population research on Brazilian bumblebees. Genotyping of older
populations sometimes involved difficulties, but this should probably rather be attributed to
insufficient DNA quality or ineffective DNA extraction.
In general bumblebees are rather rare in Southern Brazil. Such rare species are a real inconvenience
for population genetic research based on MS markers, since the latter requires samples of preferably
25 individuals per population for several locations and periods in time. But this is usually a problem in
conservational genetics.
The objective of this thesis was to link the disappearance of a locally extinct Brazilian bumblebee to a
low inherent genetic diversity. Unfortunately sampling of both museum collections and recent
populations of this species was problematic. Regarding the first statement, the latter is of course
explainable. First of all, this species was not abundantly present in the sampled museum collections.
Actually the entomological (Hymenoptera) collection of the Museum of Science and Technology at
PUCRS, only contained numerous B. bellicosus specimens of one location, namely Candiota (years
1998-1999). Furthermore, from São Francisco de Paula (1995, 1996, and 1997) there were only 3
specimens available, and singular specimens from Cambará do Sul (1991), Osório (1992), Santana de
Livramento (1993) and Esmeralda (2011). Following the underrepresentation of this species in the
museum collection, it could be presumed that this species never appeared really abundantly in Rio
Grande do Sul. Hence, this species was collected in Esmeralda in 2011, though in a very small number
(1). This indicates that the species is not yet extinct in the state Rio Grande do Sul. Nevertheless, it
was not observed during sampling attempts for this research, while B. morio and B. pauloensis were.
Although the research question concerning B. bellicosus could not be answered, it was still
interesting to study the genetics of the two other species: B. pauloensis and B. morio. Both species
were represented in the museum collections in higher numbers. They were also frequently sampled
since 2010, so populations of these species remained stable. However, this year they could not be
observed in large numbers, possibly because of the shifted weather conditions as explained before.
During the sampling for this research, mainly B. morio was observed and collected, while only one B.
pauloensis individual was observed. Whether this indicates that B. morio can better withstand the
extremely wet weather is not clear and too preliminary taken the limited sampling into account.
56
One of the things that can be cautiously concluded from the results of the data analysis in this thesis,
is that the genetic variability within the species B. pauloensis seems to have gradually decreased
since 1950. This indicates an ongoing decline in these populations which is in line with the worldwide
phenomenon. For B. morio such a clear trend could not be detected. Most parameters presented a
decrease in genetic diversity, with the exception of an observed increase in AR. Hence, the two
species did not differ significantly regarding their genetic variability in most time periods; not in the
1950s and not in more recent years. Two assumptions could be made from this observation. First, if
the genetic variability of B. pauloensis decreased significantly, as the results showed, it might also
have decreased for B. morio. Second, both species probably have similar chances of withstanding and
adapting to changes in the environment. On the other hand, when simply looking at the diversity
parameter values, B. morio seems to be more variable and this was significant in the beginning of the
years 2000. Nevertheless, since B. morio specimens were more difficult to score (several small peaks
possibly because of primer dimer formation, stuttering, and such inconveniences in the Peakscanner
profile), this could have led to misinterpretation of alleles towards a greater variability.
B. bellicosus was sampled for only one location, together with B. pauloensis. The collection of these
specimens was done in 1999 by a long-term and intensive sampling during 2 years. A comparison of
these two species showed no significant different genetic diversity, but the B. bellicosus population
showed signs of inbreeding. During the recent sampling attempts of this study, both species could no
longer be detected at this location. Possibly this could be explained by the quite recent conversion of
this particular location, from horticultural land, with a wide variety of crops, to an orchard. At the
moment of sampling for this research, the parcel was fallow land, with few flowering plants.
It is expected that populations of declining species become rare and isolated (Goulson et al., 2008).
As a consequence, populations of declining species exhibit a loss of genetic diversity (drop in
heterozygosity and allelic richness) due to an increased genetic drift and loss of gene flow, while for
stable populations such changes are less likely to occur (Goulson et al., 2008). Although B. pauloensis
is supposed to be a stable species, the observed loss in genetic diversity here could be taken into
account in future conservational strategies. If the genetic diversity would further decrease and
become too low, this species might also become sensitive to local extinction. This might be also the
result of inbreeding and inbreeding depression, vulnerability to changes and stressors in the
environment, and susceptibility to diseases and pathogens (Whitehorn et al., 2009; Cameron et al.,
2011). Determination of genetic diversity could thus be a useful tool for predicting decline of
bumblebee species.
Unfortunately, nothing can be concluded in this thesis about the locally extinct and declining species
B. bellicosus. According to Martins & Melo (2010) the extinction of this species in Parana (a more
northern, neighbouring state of Rio Grande do Sul) occurred somewhere between 1980 (last known
specimens collected) and 2000 (no more specimens recovered). Therefore it would have been very
informative to compare the genetic diversity of the declining species B. bellicosus with that of the
presumably stable species in both time period 1 (1946-1959) and the more recent time periods. It
could have given a decisive answer to the question whether the declining species in Southern Brazil
had a lower genetic diversity before its decline and local extinction, when compared to the more
stable species. If there would have been observed a difference in genetic variation between the
declining and stable bumblebee species for recent populations, it could then be studied whether
this was due to a recent reduction in genetic diversity or would have been already present in the
57
years before the decline. It also could have confirmed the hypothesis that bumblebee species with a
low genetic diversity were the first to decline when drivers for bumblebee decline started to act
(Maebe et al., 2016). If a historically low genetic diversity was detected, it could be explained by the
rarity of the B. bellicosus species, as small bumblebee species can have a reduced genetic diversity as
a result of higher genetic drift (Frankham, 2005; Zayed, 2009).
Some populations showed a significant inbreeding coefficient. The inbreeding coefficients observed
within these populations (for B. pauloensis populations: POA1 and SCdS3; for B. morio populations:
SCdS3 and NPSFdP; and the B. bellicosus population: CAN3), could not be explained by the
occurrence of null alleles, as the null allele frequencies were low. In these populations it seems to be
likely that the significant positive inbreeding coefficients represents high levels of inbreeding. It
should be noted that inbreeding was occasionally present in all time periods, but remarkably not in
the most recent populations, dating from the years after 2010 (for which the inbreeding coefficients
were rather small or even negative). However, this is probably more a location effect as both CAT02
and SCdS3 population showed inbreeding for both B. pauloensis and B. morio which could thus be
indicative for isolated populations in the fragmented environment. Finally, inbreeding does not
always directly affect a species success. Only when it causes inbreeding depression it will have a
serious effect on a population. Some populations could be robust against effects of low genetic
diversity and inbreeding, but this will probably only occur in very favourable environments (Schmid-
Hempel et al., 2007). Inbreeding will always play an indirect role in the decline of a species when it
causes diploid male production and further reduces HE, which decreases the species’ capacity to
adjust to environmental changes.
Comparison of the different time periods revealed some cases of high genetic differentiation (Dest)
between populations of different time periods, mainly when the oldest populations were compared
with the most recent. However, within time periods, genetic differentiation between the locations
was negligible, implying that gene flow for B. pauloensis and B. morio was not significantly restricted
within the state of Rio Grande do Sul.
Finally, similar studies on the Neo-tropical bumblebees of South-East Brazil are not or only very
limitedly available. Francisco et al. (2015) evaluated the genetic structure and diversity of B. morio
populations on the mainland of South East Brazil and on nearby islands. The state Rio Grande do Sul
was not included in this study, but neighbouring states like Santa Catarina were. They used 12 MS
loci, designed specifically for B. morio (Françoso et al., 2012) and two B. terrestris loci. The methods,
techniques and procedure of this research were very similar to those used in this thesis. Although
comparing the results of studies which used different MS markers is troublesome, as MS have other
levels of polymorphism due to different mutation rates, it could be informative to compare the
results. Francisco et al. (2015) found high levels of diversity in most populations, which were similar
on islands and the mainland. They also noticed that B. morio, in general, shows limited genetic
divergence between distant populations. Their data suggested that long-term isolation is not
affecting the population viabilities of this species and they attribute it to the high dispersal ability of
B. morio and its capacity to survive in urban environments and highly fragmented landscapes. The
values of the genetic diversity parameters obtained in their research, were quite similar as those
obtained in this study although they reached slightly higher levels , which could be due to the higher
number of individuals (per population) used in their study (659 bees, 24 populations).
58
5. Conclusion & future perspectives There are definitely several possibilities to improve and extend the research carried out in this thesis.
First of all the dataset could be enlarged by further sampling. However, as stated before, some of the
species of interest might be very rare, imposing difficulties on obtaining sufficient data. Access to a
greater number of museum collections, could extend the set of old data. However receiving access to
very precious old collections and tissue samples is never obvious, since they are protected with great
care. All of this could take away the problem of deficient data and could improve the reliability of the
conclusions drawn from results of data analysis. Furthermore, it is important to ensure and improve
the quality of DNA extracted from museum samples, and to obtain a good set of MS markers. The
used MS loci should be polymorph and efficient for the concerned species. Therefore it could be a
good possibility to introduce MS loci, specifically designed for B. morio and B. pauloensis (Françoso &
Arias, 2012; Françoso et al., 2012). Nevertheless, as summarised before, some things were
accomplished and observed in this research. Making hard conclusions is not yet possible, but some
indications for certain trends were shown.
It is clear that the population genetics of neo-tropical bumblebees are still largely undiscovered.
Therefore it will be necessary to undertake more research concerning this topic (e.g. Françoso,
2015). The gap of knowledge is less substantial for other genera of neo-tropical bees. Mainly stingless
bees have been studied more intensely, probably because of their greater abundance and relevance
in this region (Francisco et al., 2006). The highly fragmented Brazilian Atlantic Forest is a main
concern for conservation in this region, because it causes restriction of gene flow, although in
general the fragmented landscapes have maintained high levels of population genetic diversity.
For conservation of bumblebee species in the tropical and subtropical regions, it will be important in
the future to assemble more information on changes in the status of populations. For now practically
no such information is available (Martins & Melo, 2010). Monitoring of bumblebee species should be
performed more frequently and species like B. bellicosus should be added to future editions of the
Brazilian list of threatened animal species.
Studies of population genetics are very valuable for conservation of precious pollinators like
bumblebees. However, to preserve bumblebee diversity also the current drivers of bumblebee
decline must be tackled, to ensure that the species with both low and even high genetic diversity
species do not face extinction. It is therefore recommended that conservation strategies conserve
and create more suitable habitats for sustaining bumblebee populations.
59
References Abrahamovich, A. H., & Díaz, N. B. (2002). Bumble bees of the Neotropical Region (Hymenoptera:
Apidae). Biota Colombiana, 3(2), 199–214.
Abrahamovich, A. H., Díaz, N. B., Morrone, J. J. (2004). Distributional Patterns of the Neotropical and Andean species of the genus Bombus (Hymenoptera: Apifae). Acta Zoológica Mexicana (n.s.) 20(1): 99-117.
Abrahamovich, A. H., Tellería, M. C., & Díaz, N. B. (2001). Bombus species and their associated flora in Argentina. Bee World, 82 (2), 76–87.
Ingrid Aguilar, Daniel Briceño. Sounds in Melipona costaricensis (Apidae: Meliponini): effect of sugar concentration and nectar source distance. Apidologie, Springer Verlag, 33 (4),375-388.
Alves, D. A., Imperatriz-Fonseca, V. L., Francoy, T. M., Santos-Filho, P. S., Billen, J., & Wenseleers, T. (2011a). Successful maintenance of a stingless bee population despite a severe genetic bottleneck. Conservation Genetics, 12(3), 647–658.
Alves, D. A., Menezes, C., Imperatriz-F, L. V., & Wensellers, T. (2011b). First discovery of a rare polygyne colony in the stingless bee Melipona quadrifasciata (Apidae, Meliponini). Apidologie, 42, 211–213.
Arbulo, N, Santos, E, Salvarrey, S, & Invernizzi, C. (2011). Proboscis length and resource utilization in two ruguayan bumblebees: Bombus atratus Franklin and Bombus bellicosus Smith (Hymenoptera: Apidae). Neotropical Entomology, 40(4), 483–488.
Arias, M. C., Arnoux, E., Bell, J. J., Bernadou, A., Bino, G., Blatrix, R., … Zhu, L. (2012). Permanent Genetic Resources added to Molecular Ecology Resources Database 1 December 2011 - 31 January 2012. Molecular Ecology Resources, 12(3), 570–572.
Ashley, M.V., Berger-Wolf, T.Y., Caballero, I.C., Chaovalitwongse, W., Das Gupta, B., et al. (2008) Full siblings reconstruction in wild populations from microsatellite genetic markers. Computational Biology: New Research. Nova Science Publishers, Hauppauge, New York.
Ashley, M.V., Caballero; I.C., Chaovalitwongse, W., Das Gupta, B., Govindan, P., et al. (2009) Kinalyzer, a computer program for reconstructing sibling groups. Mol. Ecol. Resour. 9, 1127-1131.
Bartelli, B. F., & Nogueira-Ferreira, F. H. (2014). Pollination services provided by Melipona quadrifasciata Lepeletier (Hymenoptera: Meliponini) in greenhouses with Solanum lycopersicum L. (Solanaceae). Sociobiology, 61(4), 510–516.
Batalha-Filho, H., Melo, G. a. R., Waldschmidt, A. M., Campos, L. a. O., & Fernandes-Salomão, T. M. (2009). Geographic distribution and spatial differentiation in the color pattern of abdominal stripes of the Neotropical stingless bee Melipona quadrifasciata (Hymenoptera: Apidae). Zoologia (Curitiba, Impresso), 26(2), 213–219.
Beekman, M., Van Stratum, P., & Lingeman, R. (2011). Diapause survival and post-diapause performance in bumblebee queens (Bombus terrestris). Entomologia Experimentalis et Applicata 89, 207–214.
Biesmeijer, J. C., Roberts, S. P. M., Reemer, M., Ohlemüller, R., Edwards, M., Peeters, T., & Settele, J. (2006). Parallel declines in pollinators and insect-pollinated plants in Britain and the Netherlands. Science, 313(5785), 351-354.
60
Bomfim, I. G. A., De M. Bezerra, A. D., Nunes, A. C., De Aragão, F. A. S., & Freitas, B. M. (2014). Adaptive and foraging behavior of two stingless bee species (Apidae: Meliponini) in greenhouse mini watermelon pollination. Sociobiology, 61(4), 502–509.
Bortoli C, Laroca S (1990). Estudo biocenotico em Apoidea (Hymenoptera) de uma area restrita em Sao José dos Pinhais (PR, Sul do Brasil), com notas comparativas. Dusenia 15, 1–112.
Bourke, A. F. G., & Ratnieks, F. L. W. (1999). Kin conflict over caste determination in social Hymenoptera. Behavioral Ecology and Sociobiology, 46(5), 287–297.
Buchmann, S. L., Jones, C. E., & Little, R. J. (1983). "Buzz pollination in angiosperms." Handbook of experimental pollination biology 73-113.
Camargo, J. M., & Pedro, S. R. M. (2013) Meliponini Lepeletier, 1836. In Moure, J. S., Urban, D. & Melo, G. A. R. (Orgs). Catalogue of Bees (Hymenoptera, Apoidea) in the Neotropical Region - online version. Available at http://www.moure.cria.org.br/catalogue. Accessed May/20/2016
Cameron, S. A., Lozier, J. D., Strange, J. P., Koch, J. B., Cordes, N., Solter, L. F., … Robinson, G. E. (2011). Patterns of widespread decline in North American bumble bees. Proceedings of the National Academy of Sciences of the United States of America, 108(2), 662–667.
Camillo, E., & Garófalo, C. A. (1989). Analysis of the niche of two sympatric species of Bombus (Hymenoptera, Apidae) in southeastern Brazil. Journal of Tropical Ecology, 5(01), 81–92.
Carvalho-Zilse, G. A., & Kerr, W. E. (2004). Substituição natural de rainhas fisogástricas e distância de vôo dos machos em Tiuba (Melipona compressipes fasciculata Smith, 1854) e Uruçu (Melipona scutellaris Latreille, 1811) (Apidae, Meliponini). Acta Amazonica, 34(4), 649–652.
Chambers, G. K., & Macavoy, E. S. (2000). Microsatellites : consensus and controversy. Comparative Biochemistry and Physiology Part B 126, 455–476.
Chapuis, M.-P., Lecoq, M., Michalakis, Y., Loiseau, A., Sword, G.A., et al. (2008) Do outbreaks affect genetic population structure? A worldwide survey in Locusta migratoria, a pest plagued by microsatellite null alleles. Mol. Ecol. 17, 3640-3653.
Charman, T. G., Sears, J., Green, R. E., & Bourke, A. F. G. (2010). Conservation genetics, foraging distance and nest density of the scarce Great Yellow Bumblebee (Bombus distinguendus). Molecular Ecology, 19(13), 2661–2674.
Cook, J. M., & Crozier, R. H. (1995). Sex determination and population biology in the hymenoptera. Trends in Ecology & Evolution, 10(7), 281–286.
Cortopassi-laurino, M., Imperatriz-fonseca, V. L., Ward, D., Dollin, A., Heard, T., Aguilar, I., … I, V. L. (2006). Global meliponiculture : challenges and opportunities. Apidologie, 37 (2), pp.275-292.
Cortopassi-Laurino, M., Knoll, F. R. N., & Imperatriz-Fonseca, V. L. I. (2003). Nicho trófico e abundância de Bombus morio e Bombus atratus em diferentes biomas brasileiros. Apoidea Neotropica, (Tabela I), 285–295.
Crow, J.F., & Kimura, M. (1970). An Introduction to Population Genetics Theory. Harper and Row, New York.
da Silva-Matos, E. V., & Garófalo, C. A. (2000). Worker life tables, survivorship, and longevity in colonies of Bombus (Fervidobombus) atratus (Hymenoptera: Apidae). Revista de Biología Tropical, 48(2-3), 657–664.
61
de Paula, G. A. R., & Melo, G. A. R. (2015). Inferring Sex and Caste Seasonality Patterns in Three Species of Bumblebees from Southern Brazil Using Biological Collections. Neotropical Entomology, 44(1), 10–20.
Di Rienzo, A., Peterson, A. C., Garza, J. C., Valdes, A. M., Slatkin, M., & Freimer, N. B. (1994). Mutational processes of simple-sequence repeat loci in human populations. Proceedings of the National Academy of Sciences, 91(8), 3166–3170.
Dieringer, D., & Schlötterer, C. (2003). Two Distinct Modes of Microsatellite Mutation Processes : Evidence From the Complete Genomic Sequences of Nine Species Two Distinct Modes of Microsatellite Mutation Processes : Evidence From the Complete Genomic Sequences of Nine Species, 2242–2251.
Dobzhansky, T. (1948). Genetics of Natural Populations. Xviii. Experiments on Chromosomes of Drosophila Pseudoobscura from Different Geographic Regions. Genetics, 33(6), 588–602.
Doherty, P. F., Boulinier, T., & Nichols, J. D. (2003). Local extinction and turnover rates at the edge and interior of speciesʼ ranges. Ann. Zool. Fennici, 40(April), 145–153.
dos Santos, S. A., Roselino, A. C., Hrncir, M., & Bego, L. R. (2009). Pollination of tomatoes by the stingless bee Melipona quadrifasciata and the honey bee Apis mellifera (Hymenoptera, Apidae). Genetics and Molecular Research : GMR, 8(2), 751–757.
Duchateau, M.J., Hishiba, H., Velthuis, H.H.W. (1994) Diploid males in the bumble bee Bombus terrestris. Entomol. Exp. Appl. 71, 263-269.
Duchateau, M.J., Marien, J. (1995) Sexual biology of haploid and diploid males in the bumble bee Bombus terrestris. Insectes Soc. 42, 255-266.
Eisen, J.A. (1999). Mechanistic basis for microsatellilte instability. In: Goldstein, D.B., & Schlö tterer, C. (Eds.), Microsatellites: Evolution and Applications. Oxford University Press, Oxford, 34–48.
Ellegren, H., Moore, S., Robinson, N., Byrne, K., Ward, W., & Sheldon, B. C. (1997). Microsatellite evolution--a reciprocal study of repeat lengths at homologous loci in cattle and sheep. Molecular Biology and Evolution, 14(8), 854–860.
Estoup, A., Solignac, M., Harry, M., Cornuet, J.-M. (1993) Characterization of (GT)n and (CT)n microsatellites in two insect species Apis mellifera and Bombus terrestris. Nucleic Acids Res. 21, 1427-1431.
Estoup, A., Tailliez, C., Cornuet, J. M., & Solignac, M. (1995). Size homoplasy and mutational processes of interrupted microsatellites in two bee species, Apis mellifera and Bombus terrestris (Apidae). Molecular Biology and Evolution, 12(6), 1074-1084.
Estoup, A., Jarne, P., Cornuet, J. (2002). Homoplasy and mutation model at microsatellite loci and their consequences for population genetics analysis. Molecular Ecology 11, 1591–1604.
Ferreira Junior, N. T., Blochtein, B., & Moraes, J. F. De. (2010). Seasonal flight and resource collection patterns of colonies of the stingless bee Melipona bicolor schencki Gribodo (Apidae, Meliponini) in an Araucaria forest area in southern Brazil. Revista Brasileira de Entomologia, 54(4), 630–636.
Ferreira, N. T., Blochtein, B., & Serrão, J. E. (2013). Seasonal production and spatial distribution of Melipona bicolor schencki (Apidae; Meliponini) castes in brood combs in southern Brazil. Apidologie, 44(2), 176–187.
62
Francisco, F. de O., Brito, R. M., & Arias, M. C. (2006). Alelle number and heterozigosity for microsatellite loci in different stingless bee species (Hymenoptera: Apidae, Meliponini). Neotropical Entomology, 35(5), 638–643.
Francisco, F. O., Santiago, L. R., Mizusawa, Y. M., Oldroyd, B. P., & Arias, M. C. (2015). Genetic structure of island and mainland populations of a Neotropical bumble bee species. bioRxiv, 027813.
Françoso, E., & Arias, M. C. (2012). Characterization of microsatellite loci for Bombus pauloensis (Hymenoptera, Apidae, Bombini). Molecular Ecology Resources
Françoso, E., Francisco, F. O., Arias, M. C. (2012). Microsatellite loci for Bombus morio (Hymenoptera, Apidae, Bombini). Molecular Ecology Resources
Françoso, E. (2011). Phylogeography of Bombus morio and B. pauloensis (Hymenoptera, Apidae), 28.
Françoso, E., Oliveira, F. F. de, & Arias, M. C. (2015). An integrative approach identifies a new species of bumblebee (Hymenoptera: Apidae: Bombini) from northeastern Brazil. Apidologie, 103, 1–15.
Frankham, R. (2005). Genetics and extinction. Biological Conservation, 126(2), 131–140.
Freiria, G. A., Ruim, J. B., De Souza, R. F., & Sofia, S. H. (2012). Population structure and genetic diversity of the orchid bee Eufriesea violacea (Hymenoptera, Apidae, Euglossini) from Atlantic Forest remnants in southern and southeastern Brazil. Apidologie, 43(4), 392–402.
Freitas, B. M., Imperatriz-Fonseca, V. L., Medina, L. M., Kleinert, A. D. M. P., Galetto, L., Nates-Parra, G., & Quezada-Euán, J. J. G. (2009). Diversity, threats and conservation of native bees in the Neotropics. Apidologie, 40(3), 332–346.
Garofalo, C.A. (1974). Aspectos evolutivos da biologia da reproducao em abelhas (Hymenoptera, Apoidea). Dissertation thesis, Universidade de Sao Paulo, Ribeirao Preto, Brazil.
Garofalo, C.A., Zucchi, R., Muccillo, G. (1986). Reproductive studies of a neotropical bumblebee, Bombus atratus (Hymenoptera, Apidae). Bras. J. Genet. 9, 231–243.
Gerard, M., Michez, D., Fournier, D., Maebe, K., Smagghe, G., Biesmeijer, J. C., & De Meulemeester, T. (2015). Discrimination of haploid and diploid males of Bombus terrestris (Hymenoptera; Apidae) based on wing shape. Apidologie, 46(5), 644–653.
Gonçalves, R. B., & Melo, G. a. R. (2005). A comunidade de abelhas (Hymenoptera, Apidae s. l.) em uma área restrita de campo natural no Parque Estadual de Vila Velha, Paraná: diversidade, fenologia e fontes florais de alimento. Revista Brasileira de Entomologia, 49(1990), 557–571.
Gonçalves, R. B., Melo, G. a. R., & Aguiar, a. J. C. (2009). A assembléia de abelhas (Hymenoptera, Apidae) de uma área restrita de campos naturais do Parque Estadual de Vila Velha, Paraná e comparações com áreas de campos e cerrado. Papéis Avulsos de Zoologia, 49(14), 163–181.
Goulson, D., Lye, G. C., & Darvill, B. (2008). Decline and conservation of bumble bees. Annual Review of Entomology, 53, 191–208.
Goulson, D. (2010) Bumblebees, behaviour, ecology and conservation. Oxford University Press, Oxford, p. 336.
Hansen, J., Sato, M., Ruedy, R., Lo, K., Lea, D. W., & Medina-Elizade, M. (2006). Global temperature change. Proceedings of the National Academy of Sciences of the United States of America,
63
103(39), 14288–14293.
Hartl, D.L., & Clark, A.G. (2007). Principles of Population Genetics. Sunderland: Sinauer Associates, 545 p.
Heard, T. A. (1999). The Role of Stingless Bees in Crop Pollination. Annual Review of Entomology, 44(1), 183–206.
Hedrick, P. W. (1999). Perspective : Highly Variable Loci and Their Interpretation in Evolution and Conservation. Evolution 53 (2) , 313-318.
Holehouse, K. A., Hammond, R. L., & Bourke, A. F. G. (2003). Non-lethal sampling of DNA from bumble bees for conservation genetics. Insectes Sociaux, 50(3), 277–285.
Hrncir, M., Jarau, S., Zucchi, R., & Barth, F. G. (2000). Recruitment behavior in stingless bees, Melipona scutellaris and M. quadrifasciata . II. Possible mechanisms of communication. Apidologie, 31(1), 93–113.
Huth-Schwarz, A., León, A., Vandame, R., Moritz, R. F. A., & Kraus, F. B. (2011). Mating frequency and genetic colony structure of the neotropical bumblebee Bombus wilmattae (Hymenoptera: Apidae). Apidologie, 42(4), 519–525.
Imperatriz-fonseca, V. L., Saraiva, A. M., & Jong, D. De. (2006). Bees as pollinators in Brazil. Ribeirão Preto.
Jaffé, R., Pioker-Hara, F. C., Dos Santos, C. F., Santiago, L. R., Alves, D. A., De M. P. Kleinert, A., … Imperatriz-Fonseca, V. L. (2014). Monogamy in large bee societies: A stingless paradox. Naturwissenschaften, 101(3), 261–264.
Jakupciak, J. P., & Wells, R. D. (2000). Gene conversion (recombination) mediates expansions of CTG??CAG repeats. Journal of Biological Chemistry, 275(51), 40003–40013.
Jarne, P., David, P., & Viard, F. (1998). Microsatellites, transposable elements and the X chromosome. Molecular Biology and Evolution, 15(1), 28–34.
Jarne, P., & Lagoda, P. J. L. (1996). Microsatellites, from molecules to populations and back. TREE 1(10).
Jost, L. (2008). GST and its relatives do not measure differentiation. Molecular Ecology, 17(18), 4015–4026.
Kalia, R. K., Rai, M. K., Kalia, S., Singh, R., & Dhawan, A. K. (2011). Microsatellite markers: An overview of the recent progress in plants. Euphytica, 177(3), 309–334.
Kalinowski, S. T. (2002). How many alleles per locus should be used to estimate genetic distances? Heredity, 88(1), 62–65.
Kerr, W. E., Carvalho, G. a, Silva, a C., & Assis, M. G. P. (2001). Aspectos pouco mencionados da biodiversidade amazônica. Mensagem Doce, 80, 45–60.
Kimura, M., & Crow, J. F. (1964). the Number of Alleles That Can Be Maintained in a Finite Population. Genetics, 49, 725–738.
Kimura, M., & Ohta, T. (1978). Stepwise mutation model and distribution of allelic frequencies in a finite population*. Proc. Nati. Acad. Scd. USA Genetics, 75(6), 2868–2872.
64
Koser, J. R., Francisco, F. O., & Moretto, G. (2014). Genetic variability of stingless bees Melipona mondury Smith and Melipona quadrifasciata Lepeletier (Hymenoptera: Apidae) from a meliponary. Sociobiology, 61(3), 313–317.
Laroca, S., & Orth, a. I. (2002). Melissocoenology: historical perspective, method of sampling, and recommendations to the “Program of conservation and sustainable use of pollinators, with emphasis on bees”(ONU). IN: Kevan P & Imperatriz Fonseca VL (eds) - Pollinating Bees - The Conservation Link Between Agriculture and Nature - Ministry of Environment/Brasilia, 217–225.
Leclercq, S. B., Rivals, E., & Jarne, P. (2010). DNA slippage occurs at microsatellite loci without minimal threshold length in humans: A comparative genomic approach. Genome Biology and Evolution, 2(1), 325–335.
Levinson, G., & Gutman, G. A. (1987). High frequencies of short frameshifts in poly-CA/TG tandem repeats borne by bacteriophage M13 in Escherichia coli K-12. Nucleic Acids Research, 15(13), 5323–5338.
Li, Y. C, Korol, A. B., Fahima, T., Beiles, A., Nevo, E. (2002). Microsatellites : genomic distribution , putative functions and mutational mechanisms : a review. Molecular ecology 11, 2453–2465.
Lozier, J. D., & Cameron, S. A. (2009). Comparative genetic analyses of historical and contemporary collections highlight contrasting demographic histories for the bumble bees Bombus pensylvanicus and B. impatiens in Illinois. Molecular Ecology, 18(9), 1875–1886.
Lozier, J. D., Strange, J. P., Stewart, I. J., & Cameron, S. A. (2011). Patterns of range-wide genetic variation in six North American bumble bee (Apidae: Bombus) species. Molecular Ecology, 20(23), 4870–4888.
Lynch, M. (1991). The Genetic Interpretation of Inbreeding Depression and Outbreeding Depression. Society, 45(3), 622–629.
Maebe, K., Meeus, I., Ganne, M., De Meulemeester, T., Biesmeijer, K., & Smagghe, G. (2015). Microsatellite Analysis of Museum Specimens Reveals Historical Differences in Genetic Diversity between Declining and More Stable Bombus Species. Plos One, 10(6), e0127870.
Maebe, K., Meeus, I., Vray, S., Claeys, T., Dekoninck, W., Boevé, J.-L., Rasmont, P., Smagghe, G. (2016) Genetic diversity of restricted wild bumblebees was already low a century ago. Molecular Ecology, under revision.
Maggi, M., Lucia, M., & Abrahamovich, A. H. (2011). Study of the acarofauna of native bumblebee species (Bombus) from Argentina. Apidologie, 42(3), 280–292.
Martins, A. C., Goncalves, R. B., & Melo, G. a R. (2013). Changes in wild bee fauna of a grassland in Brazil reveal negative effects associated with growing urbanization during the last 40 years. Zoologia, 30(2), 157–176.
Martins, A. C., & Melo, G. A. R. (2010). Has the bumblebee Bombus bellicosus gone extinct in the northern portion of its distribution range in Brazil? Journal of Insect Conservation, 14(2), 207–210.
Martins, A. C., Silva, D. P., De Marco, P., & Melo, G. A. R. (2015). Species conservation under future climate change: the case of Bombus bellicosus, a potentially threatened South American bumblebee species. Journal of Insect Conservation, 19(1), 33–43.
Marques, A. A. B. et al. Lista de Referência da Fauna Ameaçada de Extinção no Rio Grande do Sul.
65
Decreto no 41.672, de 11 junho de 2002. Porto Alegre: FZB/MCT–PUCRS/PANGEA, 2002. 52p. (Publicações Avulsas FZB, 11)
Mayr, E. (1963). Animal Species and Evolution. Belk- nap Press, Cambridge, MA.
Meirmans, P. G., & Hedrick, P. W. (2011). Assessing population structure: FST and related measures. Molecular Ecology Resources, 11(1), 5–18.
Melo, G.A.R. (2007). Introductory Remarks, p. V-XI. In: J.S. Moure, D. Urban & G.A.R. Melo (Eds). Catalogue of Bees (Hymenoptera, Apoidea) in the Neotropical Region. Curitiba, Sociedade Brasileira de Entomologia.
Melo, G. a R. (2013). On the identity of Melipona torrida Friese (Hymenoptera, Apidae). Revista Brasileira de Entomologia, 57(1916), 1–5.
Melo, G. a R., Martins, A. C., & Gonçalves, R. B. (2006). Alterações de Longo Prazo na Estrutura de Assembléias de Abelhas: Conhecimento Atual e Perspectivas. Anais Do VII Encontro Sobre Abelhas, 150–155.
Melorose, J., Perroy, R., & Careas, S. (2015). No Title No Title. Statewide Agricultural Land Use Baseline 2015, 1(iii).
Mercês, M. D., Peralta, E. D., Trovatti, A. P., & Lucchese, A. M. (2013). Antimicrobial activity of honey from five species of Brazilian stingless bees. Ciência Rural, Santa Maria, 43, 672–675.
Metzgar, D., Bytof, J., & Wills, C. (2000). Selection against frameshift mutations limits microsatellite expansion in coding DNA. Genome Research, 10(1), 72–80.
Miah, G., Rafii, M. Y., Ismail, M. R., Puteh, A. B., Rahim, H. A., Islam, N. K., & Latif, M. A. (2013). A review of microsatellite markers and their applications in rice breeding programs to improve blast disease resistance. International Journal of Molecular Sciences, 14(11), 22499–22528.
Morales, C. L., Arbetman, M. P., Cameron, S. A., & Aizen, M. A. (2013). Rapid ecological replacement of a native bumble bee by invasive species. Frontiers in Ecology and the Environment, 11(10), 529–534.
Moretto, G and Arias, M. C. (2005). Systematics, Morphology and Physiology. Detection of Mitochondrial DNA Restriction Site Differences Between the Subspecies of Melipona quadrifasciata Lepeletier ( Hymenoptera : Apidae : Meliponini ). Neotropical Entomology, 34(3), 381–385.
Mountain, R., & Butte, C. (2000). Outbreeding Depression Varies Among Cohorts of Ipomopsis Aggregata Planted in Nature, 54(2), 485–491.
Moure, J. S., & Melo, G. A. R. (2012) Bombini Latreille, 1802. In Moure, J. S., Urban, D. & Melo, G. A. R. (Orgs). Catalogue of Bees (Hymenoptera, Apoidea) in the Neotropical Region - online version. Available at http://www.moure.cria.org.br/catalogue. Accessed May/20/2016
Moure JS, Sakagami SF (1962) As mamangabas sociais do Brasil (Bombus Latreille) (Hymenoptera, Apoidea). Stud. Entomol 5, 65–194.
Murray, T. E., Fitzpatrick, Ú., Brown, M. J. F., & Paxton, R. J. (2008). Cryptic species diversity in a widespread bumble bee complex revealed using mitochondrial DNA RFLPs. Conservation Genetics, 9(3), 653–666.
Nadir, E., Margalit, H., Gallily, T., & Ben-Sasson, S. a. (1996). Microsatellite spreading in the human
66
genome: evolutionary mechanisms and structural implications. Proceedings of the National Academy of Sciences of the United States of America, 93(13), 6470–6475.
Nascimento, M. A., Batalha-Filho, H., Waldschmidt, A. M., Tavares, M. G., Campos, L. A. O., & Salomão, T. M. F. (2010). Variation and genetic structure of Melipona quadrifasciata Lepeletier (Hymenoptera, Apidae) populations based on ISSR pattern. Genetics and Molecular Biology, 33(2), 394–397.
Neff, B. D., & Gross, M. R. (2001). Microsatellite Evolution in Vertebrates: Inference From Ac Dinucleotide Repeats. Evolution, 55(9), 1717–1733.
Nei, M. (1978). Estimation of average heterozygosity and genetic distance from a small number of individuals. Genetics, 89(3), 583–590.
Neumann, P., & Carreck, N. L. (2010). Honey bee colony losses. Journal of Apicultural Research, 49(1), 1-6.
Nieh, J. C., Contrera, F. a. L., Ramírez, S., & Imperatriz–Fonseca, V. L. (2003a). Variation in the ability to communicate three-dimensional resource location by stingless bees from different habitats. Animal Behaviour, 66(6), 1129–1139.
Nieh, J. C., Contrera, F. A. L., Rangel, J., & Imperatriz-Fonseca, V. L. (2003b). Effect of food location and quality on recruitment sounds and success in two stingless bees, Melipona mandacaia and Melipona bicolor. Behavioral Ecology and Sociobiology, 55(1), 87–94.
Nieh, J. C. (2004).Recruitment communication in stingless bees (Hymenoptera, Apidae, Meliponini.
Apidologie 35,159–182.
Nieh, J. C., & Roubik, D. W. (1998). Potential mechanisms for the communication of height and distance by a stingless bee, Melipona panamica. Behavioral Ecology and Sociobiology, 43(6), 387–399.
Noor, M. a, Kliman, R. M., & Machado, C. a. (2001). Evolutionary history of microsatellites in the obscura group of Drosophila. Molecular Biology and Evolution, 18(4), 551–6.
Nunes-Silva, P., Hrncir, M., Da Silva, C. I., Roldão, Y. S., & Imperatriz-Fonseca, V. L. (2013). Stingless bees, Melipona fasciculata, as efficient pollinators of eggplant (Solanum melongena) in greenhouses. Apidologie, 44(5), 537–546.
Lopes, D. M., Silva, F. O., Fernandes-Salomao, M. T., Campos, L. A. O., & Tavares, M. G. (2010). Scientific note A scientific note on the characterization of microsatellite loci for Melipona mondury ( Hymenoptera : Apidae ). Apidologie 41, 138–140.
Oliveira, E. J., Pádua, J. G., Zucchi, M. I., Vencovsky, R., Lúcia, M., & Vieira, C. (2006). Origin, evolution and genome distribution of microsatellites, Genetics and Molecular Biology, 29 (2), 294-307.
Packer, L., & Owen, R. (2001). Population Genetic Aspects of Pollinator Decline. Conservation Ecology, 5(1):4.
Palmer, K. A., Oldroyd, B. P., Quezada-Eu??n, J. J. G., Paxton, R. J., & May-Itza, W. D. J. (2002). Paternity frequency and maternity of males in some stingless bee species. Molecular Ecology, 11(10), 2107–2113.
Parmesan, C. (2006). Ecological and Evolutionary Responses to Recent Climate Change. Annual Review of Ecology Evolution and Systematics, 37(May), 637–669.
67
Parmesan, C., & Yohe, G. (2003). A globally coherent fingerprint of climate change impacts across natural systems. Nature, 421(6918), 37–42.
Peakall, R., Smouse, F. (2006) GENALEX 6: Genetic Analysis in Excel. Population Genetic Software for Teaching and Research. Australian National University, Canberra, Australia.
Peat, J., & Goulson, D. (2005). Effects of experience and weather on foraging rate and pollen versus nectar collection in the bumblebee, Bombus terrestris. Behavioral Ecology and Sociobiology, 58(2), 152–156.
Peters, J. M., Queller, D. C., Imperatriz-fonseca, V. L., Roubik, D. W., & Strassmann, J. E. (1999). Mate number , kin selection and social con ¯ icts in stingless bees and honeybees, (November 1998).
Peters, J. M., Queller, D. C., Imperatriz-fonseca, V. L., Roubik, D. W., & Strassmann, J. E. (1997). Microsatellite loci for stingless bees. Molecular Ecology 7,783-792.
Potts, S. G., Biesmeijer, J. C., Kremen, C., Neumann, P., Schweiger, O., & Kunin, W. E. (2010). Global pollinator declines: Trends, impacts and drivers. Trends in Ecology and Evolution, 25(6), 345–353.
Pupko, T., & Graur, D. (1999). Evolution of microsatellites in the yeast Saccharomyces cerevisiae: role of length and number of repeated units. Journal of Molecular Evolution, 48(3), 313–316.
Quezada-Euan, J. J. G., Paxton, R. J., Palmer, K. A., May-Itza, W. de J., Tay, W. T., & Oldroyd, B. P. (2007). Morphological and Molecular Characters Reveal Differentiation in a Neotropical Social Bee, Melipona Beecheii (Apidae : Meliponini). Apidologie, 38(3), 247–258.
Ramalho, M., Giannini, T. C., Malagodi-Braga, K. S., & Imperatriz-Fonseca, V. L. (1994). Pollen harvest by stingless bee foragers (Hymenoptera, Apidae, Meliponinae). Grana, 33(May), 239–244.
Reis, A. M. M., Braga, A. C., Lemes, M. R., Gribel, R., & Collevatti, R. G. (2009). Development and characterization of microsatellite markers for the Brazil nut tree Bertholletia excelsa Humb. & Bonpl. (Lecythidaceae). Molecular Ecology Resources, 9(3), 920–923. .
Reber-Funk, C., Schmidt-Hempel, R., Schmid-Hempel, P. (2006) Microsatellite loci for Bombus spp. Mol. Ecol. Notes 6, 83-86.
Richard, G. F., & Paques, F. (2000). Mini- and microsatellite expansions: the recombination connection. EMBO Reports, 1(2), 122–126.
Rortais, A., Arnold, G., Halm, M. P., & Touffet-Briens, F. (2005). Modes of honeybees exposure to systemic insecticides: estimated amounts of contaminated pollen and nectar consumed by different categories of bees. Apidologie, 36(1), 71-83.
Rose, O., & Falush, D. (1998). Letter to the Editor A Threshold Size for Microsatellite Expansion. Molecular Biology and Evolution, 15 (5),613–615.
Roubik, D. W., Moreno, J. E., Vergara, C., & Wittmann, D. (1986). Sporadic food competition with the African honey bee : projected impact on neotropical social bees. Journal of Tropical Ecology 2(2), 97–111.
Roubik, D. W. (2006). Stingless bee nesting biology. Apidology, 37(2), 124–143.
Sakagami, S. F., Laroca, S., & Moure, J. S. (1967) Wild bee biocenotics in São José dos Pinhais (PR), South Brazil. Preliminary Report. J Fac Sci Hokkaido Univ Series VI Zool 16:253–291
68
Sakagami, S. F., & Laroca, S. (1971). Relative abundance, phenology and flower visits of Apid bees in Eastern Parana, Southern Brazil (Hymenoptera, Apidae). Kentydi, 39 (3) :217-230.
Sakagami, S. F. (1976). Specific differences in the bionomic characters of bumblebees a comparative review. Jour. Fac. Sct. Hokkaldo Univ. Ser. VI, 20 (3), 390–447.
Sanderson, R. A., Goffe, L. A., & Leifert, C. (2015). Time-series models to quantify short-term effects of meteorological conditions on bumblebee forager activity in agricultural landscapes. Agricultural and Forest Entomology, 17(3), 270–276.
Santos, J. E., Santos, F. R., & Silveira, F. A. (2015). Hitting an unintended target: Phylogeography of Bombus brasiliensis lepeletier, 1836 and the first new Brazilian bumblebee species in a century (Hymenoptera: Apidae). PLoS ONE, 10(5), 1–21.
Santos-Filho, P. de S., Alves, D. D. A., Eterovic, A., Imperatriz-Fonseca, V. L., & Kleinert, A. de M. P. (2006). Numerical investment in sex and caste by stingless bees (Apidae: Meliponini): a comparative analysis. Apidologie, 37(2), 207–221.
Schlotterer, C. (2000). Evolutionary dynamics of microsatellite DNA. Chromosoma, 109(6), 365–371.
Schmid-Hempel, R., & Schmid-Hempel, P. (2000). Female mating frequencies in Bombus spp. from Central Europe. Insectes Sociaux, 47(1), 36-41.
Schmid-Hempel, P., Schmid-Hempel, R., Brunner, P. C., Seeman, O. D., & Allen, G. R. (2007). Invasion success of the bumblebee, Bombus terrestris, despite a drastic genetic bottleneck. Heredity, 99(4), 414–422.
Schug, M. D., Wetterstrand, K. A., Gaudette, M. S., Lim, R. H., Hutter, C. M., & Aquadro, C. F. (1998). The distribution and frequency of microsatellite loci in Drosophila melanogaster. Molecular Ecology, 7(1), 57–70.
Selkoe, K. A., & Toonen, R. J. (2006). Microsatellites for ecologists: A practical guide to using and evaluating microsatellite markers. Ecology Letters, 9(5), 615–629.
Slaa, , ChavesSampaio, K., Hofstede, F. E., Slaa, E. J., S, L. A., S, L. A. (2006). Stingless bees in applied pollination : practice and perspectives
Slaa, E. J., Chaves, L. A. S., Malagodi-Braga, K S., Hofstede, F. E. (2006) Stingless bees in applied pollination: practice and perspectives. Apidologie, 37 (2), 293-315.
Slatkin, M. (1995). A measure of population subdivision based on microsatellite allele frequencies. Genetics, 139(1), 457–462.
Sokol, K. A., & Williams, C. G. (2005). Evolution of a triplet repeat in a conifer. Genome, 48(3), 417–426.
Sorati, M., Newman, M., Hovman, A.A. (1996) Inbreeding and incompatibility in Trichogramma brassicae: evidence and implications for quality control. Entomol. Expe. Appl. 78, 283-290.
Squirrell, J., Hollingsworth, P. M., Woodhead, M., Russell, J., Lowe, A. J., Gibby, M., & Powell, W. (2003). How much effort is required to isolate nuclear microsatellites from plants? Molecular Ecology, 12(6), 1339–1348.
E., Rohde, M., Vautrin, D., Solignac, M., Schmid-Hempel, P., Schmid-Hempel, R., Moritz, R.F.A. (2009) Novel microsatellite DNA loci for Bombus terrestris (Linnaeus, 1758). Mol. Ecol. Resour. 9, 1345-1352.
69
Strange, J.P., Knoblett, J., Griswold, T. (2009) DNA amplification from pin-mounted bumble bees (Bombus) in a museum collection: effects of fragment size and specimen age on successful PCR. Apidologie 40, 134-139.
Sunnucks, P. (2000). Efficient genetic markers for population biology. Tree, 15(5), 199–203.
Taberlet, P., Waits, L.P., Luikart, G. (1999).Noninvasive genetic sampling: look before you leap. TREE 14 (8).
Takahashi, J., Ayabe, T., Mitsuhata, M., Shimizu, I., & Ono, M. (2008). Diploid male production in a rare and locally distributed bumblebee, Bombus florilegus (Hymenoptera, Apidae). Insectes Sociaux, 55(1), 43–50.
Tautz, D., Trick, M., & Dover, G. a. (1986). Cryptic simplicity in DNA is a major source of genetic variation. Nature, 322(6080), 652–656.
Tavares, M. G., Dias, L. A. D. S., Borges, A. A., Lopes, D. M., Busse, A. H. P., Costa, R. G., … Campos, L. A. D. O. (2007). Genetic divergence between populations of the stingless bee uruçu amarela (Melipona rufiventris group, Hymenoptera, Meliponini): is there a new Melipona species in the Brazilian state of Minas Gerais? Genetics and Molecular Biology, 30(3), 667–675.
Tavares, M. G., Pietrani, N. T., Durvale, M. de C., Resende, H. C., & Campos, L. A. de O. (2013). Genetic divergence between Melipona quadrifasciata Lepeletier (Hymenoptera, Apidae) populations. Genetics and Molecular Biology, 36(1), 111–117.
Tóth, G., Gáspári, Z., Jurka, J., Toth, G., Tóth, G., Gáspári, Z., & Jurka, J. (2000). Microsatellites in different eukaryotic genomes: survey and analysis. Genome Research, 10(7), 967–981.
Van Oosterhout, C., Hutchinson, W.F., Wills, D.P.M., Shipley, P. (2004). MICROCHECKER: software for identifying and correcting genotyping errors. Mol. Ecol. Notes 4, 535-538.
Van Oystaeyen, A., Araujo Alves, D., Caliari Oliveira, R., Lima do Nascimento, D., Santos do Nascimento, F., Billen, J., & Wenseleers, T. (2013). Sneaky queens in Melipona bees selectively detect and infiltrate queenless colonies. Animal Behaviour, 86(3), 603–609.
Varela, G. (1992a). Nota preliminar sobre los componentes de un nido de Bombus bellicosus Smith, 1879 (Hymenoptera, Apoidea). Bol Soc Zool Urug 7, 55–56.
Varela, G. (1992b). Nota preliminar sobre la fenologia del nido de Bombus bellicosus Smith, 1879 (Hymenoptera, Apoidea). Bol Soc Zool Urug 7, 53–54.
Velthuis, H. H. W., & Doorn, A. Van. (2006a). A century of advances in bumblebee domestication and the economic and environmental aspects of its commercialization for pollination. Apidologie, 37, 421–451.
Velthuis, H. H. W., Vries, H. de, & Imperatriz-Fonseca, V. L. (2006b). The polygyny of Melipona bicolor : scramble competition among queens. Apidologie, 37(2), 222–239.
Walsh, S.P., Metzger, D.A., Higuchi, R. (1991) Chelex 100 as a medium for simple extraction of DNA for PCR-based typing from forensic material. BioTechniques 10, 506-513.
Wandeler, P., Hoeck, P.E.A., Keller, L.F. (2007) Back to the future: museum specimens in population genetics. Trends Ecol. Evol. 22, 634-642.
Wang, M. L., Barkley, N. a., & Jenkins, T. M. (2009). Microsatellite markers in plants and insects. Part I: applications of biotechnology. Genes, Genomes And, Genomics, 3, 54–67.
70
Waser, N. M., Price, M.V., & Shaw, R.G. (2000). Outbreeding depression varies among cohorts of Ipomopsis aggregata planted in nature. Evolution, 54(2), 485–491
Weber, J. L. (1990). Informativeness of human (dC-dA)n.(dG-dT)n polymorphisms. Genomics, 7(4), 524–530.
Weir, B. S., & Cockerham, C. C. (1984). Estimating F-Statistics for the Analysis of Population Structure Evolution, 38(6), 1358–1370.
Wenseleers, T., Alves, D. A., Francoy, T. M., Billen, J., & Imperatriz-Fonseca, V. L. (2011). Intraspecific queen parasitism in a highly eusocial bee. Biology Letters, 7(2), 173–176.
Wenseleers, T., Hart, A. G., Ratnieks, F. L. W., & Quezada-Eu??n, J. J. G. (2004). Queen execution and caste conflict in the stingless bee Melipona beecheii. Ethology, 110(9), 725–736.
Whitehorn, P.R., Tinsley, M.C., Brown, M.J.F., Darvill, B., Goulson, D. (2009) Impacts of inbreeding on bumblebee colony fitness under field conditions. BMC Evol. Biol. 9, 152.
Whitlock, M. C. (2011). G’ST and D do not replace FST. Molecular Ecology, 20(6), 1083–1091.
Widmer, a., & Schmid-Hempel, P. (1999). The population genetic structure of a large temperate pollinator species, Bombus pascuorum (Scopoli) (Hymenoptera: Apidae). Molecular Ecology, 8, 387–398.
Williams, P. H. (1998). An annotated checklist of bumble bees with an analysis of patterns of description (Hymenoptera: Apidae, Bombini). Bulletin of The Natural History Museum Entomology , 67(1), 79-152.
Williams, P. H., Araújo, M. B., & Rasmont, P. (2007). Can vulnerability among British bumblebee (Bombus) species be explained by niche position and breadth?. Biological Conservation, 138(3), 493-505.
Williams, P. H., Cameron, S. a., Hines, H. M., Cederberg, B., & Rasmont, P. (2008). A simplified subgeneric classification of the bumblebees (genus Bombus). Apidologie, 39, 1–29.
Williams, P. H. P. H. P. H., & Osborne, J. L. (2009). Bumblebee vulnerability and conservation world-wide. Apidologie, 40(3), 367–387.
Wilson A., G., & Rannala, B. (2003). Bayesian inference of recent migration rates using multilocus genotypes. Genetics, 163(3), 1177–1191.
Wright, S. (1931). Evolution in mendelian populations. Bulletin of Mathematical Biology, 52(1-2), 241–295.
Wright, T. F., Johns, P. M., Walters, J. R., Lerner, A. P., Swallow, J. G., & Wilkinson, G. S. (2004). Microsatellite variation among divergent populations of stalk-eyed flies, genus Cyrtodiopsis. Genetical Research, 84(1), 27–40.
Zayed, A. (2009). Bee genetics and conservation. Apidologie, 40(3), 237–262.
71
Supplementary data
Table 1: Deviation from HWE (* P<0.05, ** P<0.01, *** P<0.001) and significance of null allele frequency
Pop Locus Sign HWE Sign null allele freq
B. pauloensis
POA1 BL13 * no
POA1 BT23 * yes
POA1 0810 * no
NPSFdP BT23 *** yes
NPSFdP B11 *** yes
NPSFdP B126 * yes
NPSFdP 0304 *** no
NPSFdP BT05 *** no
CAN3 B126 * yes
CAN3 0304 ** yes
POA3 B100 ** no
POA3 0304 * no
POA3 0810 * no
POA3 BT05 ** no
SCdS3 BL13 * no
SCdS3 BT23 * no
ESM4 BT02 *** no
ESM4 0810 ** no
GdM4 BL02 ** no
GdM4 B132 * yes
POA4 BL02 * no
B. morio
POA4 0304 *** no
POA3 BL13 * yes
POA3 BT23 * yes
SCdS3 BL13 *** yes
SCdS3 BT02 * no
SCdS3 BT24 *** yes
SCdS3 BL02 * yes
SCdS3 B100 * yes
SCdS3 B11 *** yes
SCdS3 B126 ** yes
SCdS3 B132 *** yes
SCdS3 0304 *** no
SCdS3 0810 ** yes
SCdS3 BT05 * no
72
NPSFdP BL13 ** yes
NPSFdP BT02 *** yes
NPSFdP BT23 * no
NPSFdP BT24 ** no
NPSFdP BL02 ** no
NPSFdP BT04 * no
NPSFdP B100 ** yes
NPSFdP B132 ** yes
NPSFdP 0304 * no
B. bellicosus
CAN3 BT02 *** no
CAN3 BT23 *** yes
CAN3 BT10 *** yes
CAN3 B11 ** yes
CAN3 B132 ** no
CAN3 0304 *** yes
CAN3 0810 *** no
CAN3 BT05 *** no
Table 2: Sensitivity analysis HE
5/14 4/14 3/14 2/14 1/14 0/14
Location N HE N HE N HE N HE N HE N HE
POA1 10 0.617 7 0.624 4 0.614 4 0.614 3 0.590 2 0.429
POA2 7 0.477 6 0.451 3 0.457 3 0.457 3 0.457 2 0.405
CATO2 5 0.519 5 0.519 4 0.559 3 0.481 3 0.481 2 0.488
OSCA2 6 0.580 6 0.580 5 0.580 5 0.580 5 0.580 4 0.579
NPSFdP 18 0.510 18 0.510 14 0.506 13 0.477 13 0.477 10 0.465
CAN3 6 0.478 6 0.478 5 0.446 5 0.446 5 0.446 5 0.446
POA3 6 0.399 6 0.399 6 0.399 4 0.408 4 0.408 4 0.408
SCdS3 6 0.386 6 0.386 6 0.386 6 0.386 6 0.386 4 0.388
ESM4 17 0.385 17 0.385 17 0.385 16 0.373 16 0.373 16 0.373
GdM4 28 0.374 28 0.374 28 0.374 28 0.374 28 0.374 28 0.374
POA4 12 0.349 11 0.353 11 0.353 11 0.353 11 0.353 10 0.337
Mean 0.461
0.460
0.460
0.450
0.448
0.427
73
Table 3: Pairwise differentiation values (FST) under the diagonal and the harmonic mean of Dest above the diagonal, * indicating value significantly differing from 0
POA1 POA2 CATO2 OSCA2 NPSFdP CAN3 POA3 SCdS3 ESM4 GdM4 POA4
- 0,054 0,138 0,002 0,033 0,013 0,077 0,303 0,260 0,233 0,162 POA1
0,114 - 0,009 -0,004 -0,001 -0,015 0,000 0,013 0,012 0,012 0,007 POA2
0,091 0,008 - 0,005 0,027 0,023 0,039 0,093 0,048 0,064 0,061 CATO2
0,066 -0,008 0,028 - -0,009 -0,012 0,002 0,031 0,015 0,002 0,001 OSCA2
0,108 -0,026 0,048 -0,020 - -0,002 0,015 0,038 0,042 0,028 0,029 NPSFdP
0,116 -0,028 0,016 -0,011 -0,017 - 0,002 0,005 0,006 0,001 0,005 CAN3
0,209 0,020 0,072 0,074 0,053 0,041 - 0,007 0,010 0,014 0,001 POA3
0,245 0,034 0,101 0,051 0,062 0,025 0,024 - 0,005 0,013 0,007 SCdS3
0,286 0,061 0,119 0,099 0,077 0,070 0,048 0,016 - 0,000 0,000 ESM4
0,297 0,054 0,131 0,101 0,074 0,064 0,047 0,025 -0,002 - 0,000 GdM4
0,279 0,054 0,134 0,099 0,070 0,060 0,054 0,012 -0,003 -0,006 - POA4
Table 4: Pairwise differentiation values (FST) under the diagonal and the harmonic mean of Dest above the diagonal, * indicating value significantly differing from 0
SCdS4 GdM4 POA4 POA3 SCdS3 NPSFdP OSCA2 CATO2 POA1 - 0,000 0,001 0,000 0,010 0,077 0,216 0,168 0,475 SCdS4
0,036 - -0,009 0,000 0,017 0,048 0,135 0,167 0,314 GdM4
0,021 -0,014 - 0,030 0,033 0,047 0,161 0,175 0,333 POA4
0,031 0,006 0,023 - 0,022 0,053 0,017 0,024 0,171 POA3
0,017 0,013 0,012 0,001 - 0,055 0,230 0,174 0,414 SCdS3
0,110 0,091 0,091 0,027 0,052 - 0,046 0,019 0,150 NPSFdP
0,198 0,183 0,178 0,053 0,120 0,018 - 0,029 0,008 OSCA2
0,189 0,174 0,181 0,041 0,119 0,020 -0,004 - 0,094 CATO2
0,301 0,285 0,277 0,151 0,217 0,099 -0,005 0,053 - POA1