frequencies of mutagen-induced coincident mitotic recombination at unlinked loci in saccharomyces...

A

dtafatfytM©

Kg

1

mrhol[

f

0

Mutation Research 616 (2007) 119–132

Frequencies of mutagen-induced coincident mitotic recombinationat unlinked loci in Saccharomyces cerevisiae

Kathryn M. Freeman, George R. Hoffmann ∗Department of Biology, College of the Holy Cross, One College Street, Worcester, MA 01610-2395, USA

Received 27 June 2006; received in revised form 8 August 2006Available online 6 December 2006

bstract

Frequencies of coincident genetic events were measured in strain D7 of Saccharomyces cerevisiae. This diploid strain permits theetection of mitotic gene conversion involving the trp5-12 and trp5-27 alleles, mitotic crossing-over and gene conversion leadingo the expression of the ade2-40 and ade2-119 alleles as red and pink colonies, and reversion of the ilv1-92 allele. The three genesre on different chromosomes, and one might expect that coincident (simultaneous) genetic alterations at two loci would occur atrequencies predicted by those of the single alterations acting as independent events. Contrary to this expectation, we observed thatde2 recombinants induced by bleomycin, �-propiolactone, and ultraviolet radiation occur more frequently among trp5 convertantshan among total colonies. This excess among trp5 recombinants indicates that double recombinants are more common than expectedor independent events. No similar enrichment was found among Ilv+ revertants. The possibility of an artifact in which haploideasts that mimic mitotic recombinants are generated by a low frequency of cryptic meiosis has been excluded. Several hypotheses
hat can explain the elevated incidence of coincident mitotic recombination have been evaluated, but the cause remains uncertain.
ost evidence suggests that the excess is ascribable to a subset of the population being in a recombination-prone state.2006 Elsevier B.V. All rights reserved.

eywords: Mitotic recombination; Gene conversion; Coincident recombination; Bleomycin; �-Propiolactone; Ultraviolet light; Yeast; Simultaneous
enetic events
. Introduction

The yeast Saccharomyces cerevisiae is a eukaryoticicroorganism that lends itself to the study of allelic

ecombination, occurring between related sequences onomologous chromosomes in a diploid cell, as well as
f ectopic recombination between sites of homologyocated at any position in either a haploid or diploid cell1]. Mitotic recombination in a diploid cell may be recip-
∗ Corresponding author. Tel.: +1 508 793 3416;ax: +1 508 793 2696.

E-mail address: [email protected] (G.R. Hoffmann).

027-5107/$ – see front matter © 2006 Elsevier B.V. All rights reserved.doi:10.1016/j.mrfmmm.2006.11.014

rocal or nonreciprocal [1,2]. Reciprocal recombination(crossing-over) entails the exchange of segments of chro-matids between homologous chromosomes, whereasnonreciprocal recombination (gene conversion) is thechange of an allele into that on the homologous chromo-some by the unidirectional transfer of a DNA sequencebetween nonsister chromatids [1,3–6]. Gene conver-sion is responsible for the great majority of events ofintragenic recombination [1]. Mechanisms of mitoticrecombination have been covered in several excellent
reviews [4–6].
Recombinational processes in mitotic cells maintainthe integrity of the genome by repairing breaks in DNAand chromosomes and by circumventing the effects of

mailto:[email protected]

dx.doi.org/10.1016/j.mrfmmm.2006.11.014

Mutatio
120 K.M. Freeman, G.R. Hoffmann /
chemical adducts in DNA. They are also implicated inthe formation of some chromosomal rearrangements [1].Because it can create homozygosity in a previously het-erozygous cell, mitotic recombination is implicated inhuman somatic cell disease, most notably cancer [7–9].The principal initiating lesion for mitotic recombinationis a double-strand break in DNA [1,8]. Single-strandbreaks also induce recombination, but they are thoughtto be processed into double-strand breaks before beingsubject to the enzymes that complete the process [1,8].Other DNA structures that can trigger mitotic recombi-nation apparently include single-stranded regions arisingat sites of stalled replication [10]. Many chemical muta-gens and radiation have been shown to induce bothreciprocal and nonreciprocal mitotic recombination inyeast [11]. DNA damage may induce recombinationby providing additional substrates (i.e., lesions) for theinitiation of the recombinational process or by activat-ing genes encoding enzymes required for recombination[1]. Split-dose experiments have failed to provide clearevidence for the latter, in that most results have beenconsistent with additivity. Even though more evidencesupports the former mechanism than the latter, the twomodes of induction need not be mutually exclusive [1].

The diploid strain D7 of S. cerevisiae was con-structed by F.K. Zimmermann for the efficient detectionof mitotic gene conversion at the trp5 locus, mitoticcrossing over and gene conversion at ade2, andpoint mutations causing reversion of the ilv1-92 allele[3,12,13]. The genotype of D7 is shown in Table 1. D7is heteroallelic at trp5 (trp5-12/trp5-27), requires tryp-tophan supplementation for growth, and gives rise totryptophan-independent (Trp+) recombinants by mitoticgene conversion [12]. The ade2-40 and ade2-119 alle-les exhibit interallelic complementation, making D7
an adenine prototroph that produces typical whitecolonies. Haploid or homozygous yeasts that expressade2-40 produce red colonies, resulting from a block-age in the adenine pathway, whereas expression of the
Table 1Strains of Saccharomyces cerevisiae

Strain Alternative designation Source

D7 – a

FZ1 ITO-3A a

FZ2 ITO-6C a

STX20-1C ATCC#208008 b

P49 ATCC#208009 b

X2144-S19 ATCC#204590 b

X2144-S22 ATCC#204591 b

a F.K. Zimmermann, Technische Hochschule, Darmstadt, Germany.b American Type Culture Collection/Yeast Genetic Stock Center (http://ww

n Research 616 (2007) 119–132

leaky ade2-119 allele causes a pink phenotype. Mitoticcrossing-over between the centromere and the ade2 locusgives rise to sectored pink and red twin-spot colonies [3].Gene conversion at ade2 can produce red colonies, pinkcolonies, and red/white or pink/white sectored colonies,but not red/pink twin spots. A small fraction of redor pink altered colonies may arise from other geneticevents, including point mutation, deletion, and chromo-some loss, but these are minor contributors to the totalfollowing exposure to recombinagens. The ilv1-92 locusis homozygous and permits the detection of mutations byselecting for isoleucine prototrophy (Ilv+). Revertantsarise by true back mutation and suppressor mutations[3,14]. The D7 assay is well characterized and responsiveto diverse mutagens and carcinogens [11].

One would expect mutations or recombinationalevents involving unlinked genes to be independent ofone another, such that the frequency of a genetic eventinvolving one gene is not influenced by that involvingother genes, other than as reflected in rates of mutationor recombination for the organism as a whole. Thus, ifmutation occurs in gene A at a frequency of 10−6 andin gene B at 10−7, one would expect the frequency ofcoincident mutation in both A and B to be 10−13. Themutagenesis literature contains few analyses of coinci-dent genetic events, perhaps because they occur at lowfrequencies and their measurement is prone to artifactsthat can mimic the rare events of interest. Yeast strainD7 is well suited to investigating this subject, in thatthe indicator genes trp5, ilv1 and ade2 are on differentchromosomes, and the spontaneous frequencies of therecombinational events are at least 100 times higher thantypical mutation frequencies. Moreover, mutagen treat-ments can substantially increase the recombinant andrevertant frequencies, making it possible to detect coin-
cident genetic events at easily measurable frequencies.
Most of our experiments were conducted withbleomycin (BLM), which is a potent mutagen andrecombinagen in strain D7. BLM is a glycopeptide

Genotype

MATa/� ade2-40/ade2-119 trp5-12/trp5-27 ilv1-92/ilv1-92MATa ade2-119 trp5-b ilv1-92 MAL SUCMAT� ade2-40 trp5-a ilv1-92 MAL SUCMATa lys2 gal2MAT� lys2 gal2MATa met13 leu1 trp5-48 cyh2 aro2 lys5 ade5 galMAT� met13 leu1 trp5-48 cyh2 aro2 lys5 ade5 gal

w.atcc.org).

http://www.atcc.org/

Mutatio

acrttssa

atabadbohethdotarii

2

2

(w[pHllmA(

2

msdcca

K.M. Freeman, G.R. Hoffmann /

ntibiotic from Streptomyces and is used as a cancerhemotherapy drug. BLM is activated through a processequiring a reduced transition metal, oxygen, and an elec-ron source. Activated BLM abstracts a hydrogen fromhe 4′ position of deoxyribose in DNA, forming site-pecific radicals that are processed into single and doubletrand breaks [15–18]. BLM is mutagenic, clastogenic,nd recombinagenic in diverse assays [15,16,19,20].

In this paper we report that the frequency of ade2ltered colonies induced in strain D7 is higher amongrp5 convertants than in the population as a whole ormong Ilv+ revertants. This finding indicates that dou-le recombinants are more common than expected undern assumption of independent events. Earlier studiesemonstrating elevated frequencies of coincident recom-ination are reviewed, and hypotheses that can explainur results in D7 are proposed and explored. Theseypotheses fall into three categories: (1) the apparentxcess of double recombinants is an artifact, such thathe excess of pink and red colonies is attributable toaploids arising by cryptic sporulation rather than toiploid mitotic recombinants; (2) the excess reflectsverdispersion specifically applicable to BLM, similaro that reported for BLM-induced chromosomal dam-ge in mammalian cells [21]; (3) the excess of doubleecombinants is ascribable to a subpopulation that isn a recombination-prone state or hypersensitive to thenduction of mitotic recombination.

. Materials and methods

.1. Yeast strains

Strains of S. cerevisiae are listed in Table 1. Strain D7MATa/� ade2-40/ade 2-119 trp5-12/trp5-27 ilv1-92/ilv1-92)as used to measure mitotic recombination and point mutations

3,12,13]. D7 and the haploid strains FZ1 and FZ2 that com-rise D7 were obtained from F.K. Zimmermann (Technischeochschule, Darmstadt, Germany). Strains STX20-1C (MATa

ys2 gal2), P49 (MATα lys2 gal2), XS144-S19 (MATa met13eu1 trp5-48 cyh2 aro2 lys5 ade5 gal), and XS144-S22 (MAT�

et13 leu1 trp5-48 cyh2 aro2 lys5 ade5 gal) were from themerican Type Culture Collection/Yeast Genetic Stock Center

http://www.atcc.org).

.2. Media and solutions

The rich medium for nonselective growth of yeast andutagenic treatment during growth was YEPD, which con-

ists of 1% Difco yeast extract, 2% Difco peptone, and 2%-glucose [12]. Cell densities (i.e., numbers of colony-formingells) were measured on minimal medium containing 2% glu-ose, 0.67% Difco Yeast Nitrogen Base without amino acids,nd 2% Difco Bacto-agar, supplemented with adenine sul-

n Research 616 (2007) 119–132 121

fate (5 mg/l), isoleucine (60 mg/l), and tryptophan (10 mg/l)[12]. This supplemented minimal medium (YMAIT) has alower adenine concentration than standard supplementation(10 mg/l) to encourage strong red/pink color developmentby ade2 recombinants. Tryptophan was omitted from theminimal medium to select for Trp+ convertants (YMAI), andisoleucine was omitted to select for Ilv+ revertants (YMAT).Minimal medium supplemented only with adenine (YMA) wasused to select against isoleucine auxotrophs and tryptophanauxotrophs in mating-type tests.

Presporulation medium was 50 mM potassium phthalate(pH 5) containing 0.67% Yeast Nitrogen Base without aminoacids, 0.1% Difco Yeast Extract, 1% potassium acetate, andthe supplements needed by D7: adenine sulfate (10 mg/l),isoleucine (60 mg/l) and tryptophan (10 mg/l). Sporulationmedium was a 1% potassium acetate solution supplemented forthe auxotrophic requirements. Mutagenic treatments of non-growing yeast were in 0.05 M phosphate buffer, pH7. Yeastdilutions were in saline (0.9% sodium chloride).

2.3. Chemicals

Bleomycin sulfate (BLM; Chemical Abstracts Service Reg-istry Number (CAS) 9041-93-4) was stored frozen at −20 ◦Cin 250 �l aliquots at 4 mg/ml and thawed immediately beforeuse. �-Propiolactone (�PL; CAS 57-57-8) was stored as a pureliquid at −20 ◦C and diluted as needed immediately before use.BLM and �PL were from Sigma Chemical Company, St. Louis,MO.

2.4. Microbiological methods

Single colonies were isolated from YEPD or YMAIT, inoc-ulated into 125 ml Erlenmeyer flasks containing 50 ml YEPDor 20 mm culture tubes containing 5 ml YEPD, and grown16–24 h in a shaker to approximately 2 × 108 cells/ml, or 12 hto obtain exponential cultures with about 1 × 108 cells/ml,as determined by hemocytometer counts and by numbers ofcolony-forming cells. To ensure that cultures were in station-ary phase, a 48-h incubation was used. Growth phase is relatedto the cell cycle, in that the size of a bud gives the approximateposition in the cell cycle, all cell-cycle stages are representedin mid-exponential phase [22], and stationary phase cells exitthe cycle in G0 [22,23]. Cells arrest as unbudded cells whennutrients are depleted [22].

Spontaneous frequencies of trp5 convertants (recom-binants) and ilv1 revertants were determined by platingapproximately 2 × 106 cells on YMAI and 2 × 107 cells onYMAT, respectively. Cell densities (i.e., numbers of colony-forming cells) were measured by plating approximately 200cells on YMAIT. YMAIT and YMAI plates were incubated
60 h, and YMAT plates 72 h at 28 ◦C. Spontaneous convertantfrequencies ranged from 0.4 to 2.0 × 10−5, and spontaneousrevertant frequencies from 0.1 to 1.1 × 10−6. These values areconsistent with our historical controls and published literatureon strain D7. Cultures were stored at 4 ◦C until use.
http://www.atcc.org/

Mutatio
2.5. Mutagenesis

Chemical mutagenesis included treatment of nondividingcells in 50 mM phosphate buffer, pH7, with BLM or �PL andtreatment of dividing cells with BLM in growth medium. ForBLM treatment of nondividing cells, yeast from a culture lessthan 3 weeks old, whose spontaneous convertant and rever-tant frequencies had been determined, were centrifuged andresuspended in phosphate buffer at 2 × 108 cells/ml. Dilutionsof BLM were made into 1 ml volumes of phosphate bufferat twice the desired BLM concentration. To initiate the treat-ment, the BLM solutions were mixed with 1 ml of yeast, giving1 × 108 cells/ml at the proper concentration. Treatment tubeswere incubated in a shaker at 28 ◦C for 2 h. To treat divid-ing cells, BLM was diluted into 1 ml aliquots of YEPD. Thetubes were inoculated with 20 �l of yeast from a culture withknown spontaneous convertant and revertant frequencies andincubated in a shaker for 16 h at 28 ◦C. For treatment with�PL, yeast from a characterized culture was centrifuged andresuspended in phosphate buffer at 2 × 108 cells/ml. Treatmentwas initiated by adding 5–20 �l �PL to 1 ml yeast suspension,and the treatment tubes were incubated in a shaker at 28 ◦Cfor 2 h. Treatments with BLM or �PL were terminated by1:10 dilution with cold saline, centrifugation, and resuspensionin saline at 2 × 108 cells/ml. Ultraviolet light (UV) treatmentswere performed by exposing cells on the surface of platingmedia to UV-C radiation from a 15 W germicidal lamp for spec-ified times at a distance of 45 cm. Cell densities for irradiationwere approximately 200 cells per plate on YMAIT, 2 × 106

on YMAI, and 1 × 107 on YMAT. Plates were protected fromlight to prevent photorepair and incubated at 28 ◦C.

In split-dose experiments, a 2-h treatment of stationary-phase cells with BLM in phosphate buffer was followed by4 or 16 h incubation in rich medium without BLM and then asecond BLM treatment in buffer. In priming-dose experiments,a dividing yeast culture was treated for 16 h with a minimallymutagenic dose of BLM in YEPD (0.005 and 0.02 �g/ml), fol-lowed by a 2-h treatment with a challenging dose of BLMin phosphate buffer (0, 0.0975, and 1.56 �g/ml). In liquid-holding experiments, 2-h treatments with BLM in phosphatebuffer were followed by 20 h liquid holding in aerated phos-phate buffer in a shaker at 28 ◦C and a second BLM treatmentunder the same conditions as the first treatment.

2.6. Detection of mitotic recombinants and revertants

Yeast suspensions were diluted in 0.9% saline to cell den-sities appropriate to the selection of convertants and revertants(see table footnotes) and approximately 200 cells per plateon nonselective medium. Plates were incubated 60 h at 28 ◦Cbefore counting total colonies and convertants, and for 72 h
for revertants. Toxicity was expressed as a relative platingefficiency (RPE) on nonselective medium (colony count rel-ative to that of the control). For treatments in buffer, RPEis a measure of survival, whereas for treatments in growthmedium it reflects both killing and growth inhibition. Con-
n Research 616 (2007) 119–132

vertant and revertant frequencies were derived from colonycounts on selective media and those on nonselective media.After counting colonies, plates were refrigerated overnight andthen incubated 3–4 days at 28 ◦C (i.e., 7 days total incubation)before recording numbers of ade2 altered colonies. This proce-dure optimizes color development of ade2 recombinants. Pink,red, and sectored colonies (pink/red; pink/white; red/white; andpink/red/white) were tabulated separately and summed to getan overall frequency of ade2 altered colonies.

2.7. Haploid detection

Haploid strains (Table 1) were used in mating-type tests todetermine whether the diploid D7 gave rise to haploid deriva-tives under our assay conditions. Approximately 2 × 107 cellsof a detector strain and 100 �l of fresh YEPD (insufficient forvisible growth) were spread on minimal medium supplementedonly with adenine to form a confluent lawn of nongrowingcells. D7 cultures to be tested for the presence of haploidswere spread on the lawns at densities ranging from 8 × 105 to4 × 107 cells per lawn. Most haploids derived from D7, beingTrp−, were detected on lawns of STX20-1C and P49. Rare Trp+

haploids were detected on lawns of the trp5-48 strains XS144-S19 and XS144-S22. All tests included positive and negativecontrols (mating and nonmating, respectively). The controlsfor lawns of STX20-1C and P49 were 200 cells of strains FZ1and FZ2 in reciprocal crosses, and those for lawns of XS144-S19 and XS144-S22 were reciprocal crosses with 200 cells ofSTX20-1C and P49.

3. Results

The results in Table 2 show dose-dependent toxic-ity, induction of gene conversion at the trp5 locus, andreversion of the ilv1-92 allele caused by the treatment ofgrowing yeast with BLM in rich medium for 16 h. Con-vertant and revertant frequencies are significantly greaterthan those in the control at all dosages tested. The tablealso shows the induction of red and/or pink ade2 alteredcolonies. The increase in altered colonies, which ariseprimarily by gene conversion, is dose-dependent, andit is detectable among all colonies on YMAIT (nonse-lective medium), convertants on YMAI, and revertantson YMAT. The frequencies of ade2 recombinants arehigher among Trp+ convertants than among all surviv-ing cells. The Trp+ altered colonies represent doublerecombinants, and their frequency, calculated from trp5convertant frequencies and numbers of altered colonies(Table 2), increases from 2.7 × 10−6 at 0.39 �g/ml to6.8 × 10−5 at 3.125 �g/ml. The excess of double recom-
binants among trp5 convertants is highly significant atevery BLM dosage except 3.125 �g/ml, for which thetrend is in the same direction but statistical power islacking owing to toxicity. Unlike the result for conver-

K.M. Freeman, G.R. Hoffmann / MutatioTa

ble

2In

duct

ion

bybl

eom

ycin

ofge

neco

nver

sion

atth

etr

p5lo

cus,

reve

rsio

nof

the

ilv1

-92

alle

le,a

ndad

e2al

tere

dco

loni

esm

easu

red

amon

gal

lcol

onie

s,T

rp+

conv

erta

nts,

and

Ilv+

reve

rtan

tsin

grow

ing

cells

ofS.

cere

visi

aest

rain

D7

BL

M(�

g/m

l)a

Rel

ativ

epl

atin

gef

ficie

ncy

Trp

+co

nver

tant

spe

rpl

atea

Ilv+

reve

rtan

tspe

rpl

atea

Con

vert

ants

per

105

cells

bR

ever

tant

spe

r10

6ce

llsb

ade2

alte

red

colo

nies

amon

g

All

cells

onY

MA

ITT

rp+

onY

MA

IcIl

v+on

YM

AT

c

∑%

∑%

∑%

01.

007.

301.

800.

61±

0.04

0.15

±0.

041/

2400

0.04

20/

146

<0.

680/

36<

2.8

0.39

1.20

134.

1049

.30

9.31

±0.

35*

3.42

±0.

22*

17/2

878

0.59

177

/268

22.

870**

**8/

986

0.81

1N

S0.

780.

4313

3.90

43.6

725

.75

±0.

67**

8.40

±0.

63*

11/1

038

1.06

086

/267

83.

211**

**6/

437

1.37

3N

S1.

560.

1692

.10

65.1

147

.97

±2.

90**

33.9

1±

1.55

**9/

765

1.17

681

/184

24.

397**

**13

/586

2.21

8N

S3.

125

0.01

14.8

513

.70

123.

75±

7.82

**11

4.17

±12

.24**

2/69

2.89

915

/297

5.50

5NS

5/27

41.

825

NS

aT

reat

men

tat2

8◦ C

for

16h

inY

EPD

.The

rew

ere

20re

plic

ate

plat

esin

the

mea

sure

men

tsof

surv

ival

and

gene

conv

ersi

onan

d10

repl

icat

epl

ates

inm

easu

rem

ents

ofre

vers

ion.

Plat

ing

cell

dens

ities

wer

e2

×10

7ce

llspe

rpl

ate

inth

ese

lect

ion

for

reve

rtan

tsan

d2×

106

cells

per

plat

ein

the

sele

ctio

nfo

rco

nver

tant

s.b

Freq

uenc

ies

are

mea

ns±

S.E

.M.T

hesi

gnifi

canc

eof

diff

eren

ces

from

the

untr

eate

dco

ntro

lwas

dete

rmin

edby

AN

OV

Aw

itha

Dun

nett

mul

tiple

com

pari

sons

test

(*p

<0.

05;**

p<

0.01

).c

The

sign

ifica

nce

ofdi

ffer

ence

sin

freq

uenc

ies

ofad

e2al

tere

dco

loni

esfr

omth

ose

amon

gal

lcel

lson

nons

elec

tive

med

ium

(YM

AIT

)was

dete

rmin

edby

Fish

er’s

exac

ttes

t(**

**p

<0.

0001

;NS:

nons

igni

fican

t).

n Research 616 (2007) 119–132 123

tants, there is no significant excess of ade2 recombinantsamong ilv1 revertants.

Mating-type tests were used to explore whether someof the red and pink colonies ascribed to mitotic recom-bination might actually be haploid yeast. We measuredfrequencies of haploids in D7 by plating D7 on lawnsof strains P49 (MAT�), X2144-S22 (MAT�), STX20-1C (MATa) and XS144-S19 (MATa). Genotypes are inTable 1. Spontaneously arising haploids were detectedin D7 at a frequency of 7.37 ± 0.31 × 10−5 in eightmeasurements. Positive control cells of the mating typeopposite to that of the lawn were detected with an effi-ciency over 65% (66.4–85.8% with a mean of 75.0%),whereas negative controls of the same mating type pro-duced no colonies.

Most haploids arising from D7 are expected to beTrp−, carrying either the trp5-12 or trp5-27 allele. OnlyTrp+ haploids, which may arise by intragenic meioticrecombination, would be a potential source of artifactin our experiments, so we estimated their frequency bytwo means: a calculation combining the frequencies justdescribed with a measurement of frequencies of Trp+

meiotic products; and a large-scale plating of D7 on alawn of Trp− strain XS144-S19 on medium without tryp-tophan. The frequency of Trp+ cells among the meioticproducts in deliberately sporulated cultures of D7 was7.8 × 10−6. This frequency, when combined with ourmeasured frequency of cryptic sporulation in diploid cul-tures (7.4 × 10−5), suggests that the frequency of crypticTrp+ haploids arising in D7 should be approximately5.8 × 10−10. Plating on a lawn of XS144-S19 (MATa)indicated the frequency of MAT� Trp+ haploids aris-ing spontaneously in D7 to be 3.185 × 10−10. Assumingthat MATa and MAT� haploids occur at equal frequency,the total frequency of Trp+ haploids is 6.37 × 10−10,corresponding closely to the estimate of 5.8 × 10−10.Thus, both determinations indicate that Trp+ haploidsare too rare to explain the observed excess of doublerecombinants.

Given that coincident recombination was measured inmutagen-treated cultures, we determined whether muta-genesis increases the frequency of Trp− or Trp+ haploids,as might be expected if it induced meiosis or increasedthe frequency of intragenic meiotic recombination. Thefrequencies of haploids were measured in yeast treatedfor 2 h with BLM in phosphate buffer. The treatmentsinduced Trp+ convertants (15-fold and 35-fold over aspontaneous frequency of 0.57 × 10−5 at BLM dosages
of 6.25 and 25 �g/ml, respectively) and Ilv+ revertants(17-fold and 59-fold over a spontaneous frequency of0.31 × 10−6) as expected. The average frequencies ofhaploids in duplicate measurements were 5.92 × 10−5 at

Mutation Research 616 (2007) 119–132

y�

-pro

piol

acto

neof

gene

conv

ersi

onat

the

trp5

locu

s,re

vers

ion

ofth

eil

v1-9

2al

lele

,and

ade2

alte

red

colo

nies

mea

sure

dam

ong

alls

urvi

vors

,Trp

+co

nver

tant

s,an

dIl

v+re

vert

ants

ine

stra

inD

7

aSu

rviv

al(%

)T

rp+

conv

erta

nts

per

plat

eaIl

v+re

vert

ants

per

plat

eaC

onve

rtan

tspe

r10

5ce

llsb

Rev

erta

nts

per

106

cells

bad

e2al

tere

dco

loni

esam

ong

All

cells

onY

MA

ITT

rp+

onY

MA

IcIl

v+on

YM

AT

c

∑%

∑%

∑%

100

29.8

03.

881.

42±

0.06

0.18

±0.

031

1/42

130.

024

1/59

60.

168

0/38

<2.

687

143.

9523

1.90

7.88

±0.

17**

12.6

9±

0.46

**7/

3656

0.19

134

/287

91.

181**

*2/

2319

0.08

6N

S73

132.

1562

.50

85.5

9±

1.67

**80

.96

±4.

80**

65/3

086

2.10

612

3/26

434.

654**

*13

/625

2.08

0N

S

enta

t28

◦ Cfo

r2h

inph

osph

ate

buff

er,p

H7.

The

rew

ere

20re

plic

ate

plat

esin

the

mea

sure

men

tsof

surv

ival

,gen

eco

nver

sion

,and

reve

rsio

n.Pl

atin

gce

llde

nsiti

esw

ere

2×

107,2

×10

7

6at

�PL

dosa

ges

of0,

1,an

d4

mM

,res

pect

ivel

y,in

the

sele

ctio

nfo

rre

vert

ants

.The

rew

ere

2×

106,2

×10

6,a

nd2

×10

5ce

llspe

rpl

ate

inth

ese

lect

ion

for

conv

erta

nts.

ncie

sar

em

eans

±S.

E.M

.The

sign

ifica

nce

ofdi

ffer

ence

sfr

omth

eun

trea

ted

cont

rolw

asde

term

ined

byA

NO

VA

with

aD

unne

ttm

ultip

leco

mpa

riso

nste

st(**

p<

0.01

).an

ceof

diff

eren

ces

infr

eque

ncie

sof

ade2

alte

red

colo

nies

from

thos

eam

ong

alls

urvi

vors

dete

rmin

edby

Fish

er’s

exac

ttes

t(**

*p

<0.

001;

NS:

nons

igni

fican

t).


6.25 �g/ml, 6.40 × 10−5 at 25 �g/ml, and 5.83 × 10−5 inthe untreated controls. The overall average, 6.05 ± 0.18,was consistent with the 7.37 × 10−5 measured in theearlier series of experiments. Trp+ haploids were notdetected at any dosage (frequency <7 × 10−8). Thus,there is no increase in haploids associated with BLMtreatment. Taken together, these experiments eliminatehaploid yeast as a possible source of artifact.

To evaluate whether the elevated frequency of coin-cident genetic events is unique to BLM treatments, wemeasured frequencies of ade2 altered colonies amongtrp5 convertants and ilv1 revertants after treatments with�PL and UV. Table 3 shows the dose-dependent induc-tion of Trp+ convertants and Ilv+ revertants by �PL andthe distribution of ade2 altered colonies among all sur-viving yeast, trp5 convertants, and ilv1 revertants. Thereis a highly significant excess of double recombinants atboth �PL doses, whereas the frequency of ade2 recombi-nants among ilv1 revertants does not differ significantlyfrom that among all survivors. Table 4 shows the induc-tion of gene conversion and point mutation in D7 byUV. Convertant frequencies, revertant frequencies andkilling all increase with dose, as does the percentageof ade2 recombinants among all surviving cells, Trp+

convertants, and Ilv+ revertants. At all doses, the propor-tion of ade2 altered colonies among trp5 recombinants ishigher than that among all survivors. Fisher’s exact testshows that this excess of double recombinants is highlysignificant at the highest dose.

Table 5 presents data for BLM treatments of non-dividing cells in buffer, comparing exponential-phasecells (12 h culture in YEPD) with stationary-phase cells(48 h). A higher BLM dosage is required for effectivemutagenesis in yeast suspensions in buffer (Table 5)than in dividing yeast in medium (Table 2). Never-theless, both the stationary-phase and exponential cellsshow a dose-dependent induction of gene conversion attrp5 and reversion of ilv1-92. The differences betweenexponential phase and stationary phase were modest butstatistically significant, with exponential cells showinghigher frequencies of convertants and revertants at bothBLM dosages (6.25 and 25 �g/ml). Table 5 also showssignificantly higher frequencies of ade2 altered coloniesamong Trp+ convertants than among all surviving cellsor among Ilv+ revertants in both the exponential- andstationary-phase cultures. There are no significant dif-ferences in percentages of ade2 recombinants betweenrevertants and total survivors. Extending the culture time
to 96 h, although suboptimal for the assay, showed thatthe enrichment for coincident recombination persists.Thus, the excess of double recombinants applies to non-dividing cells harvested in either exponential phase or Ta
ble

3In

duct

ion

bS.

cere

visi

a

�PL

(mM

)

0 1 4

aT

reat

man

d1

×10

bFr

eque

cSi

gnifi

c

K.M

.Freem

an,G.R

.Hoffm

ann/M

utationR

esearch616

(2007)119–132

125

Table 4Induction by ultraviolet light of gene conversion at the trp5 locus, reversion of the ilv1-92 allele, and ade2 altered colonies measured among all survivors, Trp+ convertants, and Ilv+ revertants inS. cerevisiae strain D7

UV (s)a Survival (%) Trp+ convertantsper platea

Ilv+ revertantsper platea

Convertantsper 105 cellsb

Revertants per106 cellsb

ade2 altered colonies among

all cells on YMAIT Trp+ on YMAIc Ilv+ on YMATc

∑%

∑%

∑%

0 100 18.75 14.45 1.35 ± 0.092 1.04 ± 0.076 4/2770 0.144 0/375 <0.27 0/289 <0.355 93 148.90 81.55 10.74 ± 0.22*** 6.34 ± 0.20*** 5/2573 0.194 6/2829 0.212 NS 2/1631 0.123 NS

10 85 485.30 144.00 35.01 ± 1.17*** 12.12 ± 0.30*** 26/2350 1.106 127/9706 1.308 NS 19/2880 0.660 NS20 68 952.65 365.50 68.73 ± 2.02*** 39.05 ± 1.18*** 34/1871 1.817 784/19,053 4.115**** 121/7310 1.655 NS

a UV-C irradiation of yeast on the surface of plating media at 22 ◦C from a distance of 45 cm. The cell density at all dosages was approximately 200 cells per plate on nonselective medium(YMAIT), 2 × 106 cells per plate in the selection for Trp+ convertants on YMAI, and 2 × 107 cells per plate in the selection for Ilv+ revertants on YMAT. Irradiation was followed by incubationin the dark at 28 ◦C.

b Frequencies are means ± S.E.M. The significance of differences from the untreated control was determined by ANOVA with a Dunnett multiple comparisons test (***p < 0.001).c Significance of differences in frequencies of ade2 altered colonies from those among all survivors determined by Fisher’s exact test (****p < 0.0001; NS: nonsignificant).

Table 5Induction by bleomycin of trp5 gene conversion, reversion of ilv1-92, and ade2 altered colonies measured among all survivors, Trp+ convertants, and Ilv+ revertants in exponential- andstationary-phase cells of S. cerevisiae strain D7BLM (�g/ml)a Survival (%) Trp+ convertants per

plateaIlv+ revertantsper platea

Convertants per105 cellsb

Revertants per106 cellsb

ade2 altered colonies amongAll cells on YMAIT Trp+ on YMAIc Ilv+ on YMATc

∑%

∑%

∑%

Exponential-phase (12-h growth)0 100 145.11 5.10 2.03 ± 0.047 0.43 ± 0.042 1/2144 0.047 2/2902 0.069 0/61 <1.66.25 104 206.84 115.00 46.46 ± 1.13 15.50 ± 0.46 53/2227 2.380 209/3930 5.318**** 18/1150 1.565 NS25 66 134.00 81.20 70.68 ± 2.21 34.26 ± 1.27 22/711 3.094 111/1206 9.204**** 18/812 2.217 NS

Stationary-phase (48-h growth)d

0 100 141.90 2.50 1.68 ± 0.034 NS 0.15 ± 0.020 NS 0/2415 <0.041 2/2838 0.070 0/25 <4.06.25 111 242.10 118.10 43.06 ± 0.61** 12.60 ± 0.31* 61/2810 2.171 207/4600 4.500**** 20/1181 1.693 NS25 86 180.60 92.00 62.28 ± 1.35*** 25.38 ± 1.20*** 24/1087 2.208 110/1806 6.091**** 18/920 1.957 NSa Treatment at 28 ◦C for 2 h in phosphate buffer, pH7. Plating cell densities were 2 × 107, 1 × 107, and 5 × 106 at BLM dosages of 0, 6.25, and 25 �g/ml, respectively, in the selection for

revertants and 1 × 107, 6 × 105, and 4 × 105 in the selection for convertants. There were 20 replicate plates in measurements of survival and gene conversion in the control and BLM treatment at6.25 �g/ml. There were 10 replicate plates for all reversion tests and all treatments with BLM at 25 �g/ml.

b Frequencies are means ± S.E.M. Convertant and revertant frequencies in stationary-phase cells were compared to the corresponding treatment in exponential-phase cells by ANOVA withBonferroni comparisons (NS: nonsignificant; *p < 0.05; **p < 0.01; ***p < 0.001). The frequencies in all the BLM treatments are significantly greater than those in the untreated controls (**p < 0.01;ANOVA with a Dunnett multiple comparisons test).

c Significance of differences in frequencies of ade2 altered colonies from those among all surviving cells determined by Fisher’s exact test (****p < 0.0001; NS: nonsignificant).d Cultures older than 48 h are suboptimal, but the enrichment for double recombinants persists. BLM treatment (25 �g/ml) of a 96-h culture gave 1.75% ade2 altered colonies (29/1656) among

all cells, 5.36% (43/802****) among Trp+ convertants, and 2.40% (17/707NS) among Ilv+ revertants.

Mutatio
stationary phase (Table 5), treatments of growing cellsin medium (Table 2), and aging cultures in which essen-tially all cells are in G0 (Table 5).

We based our experiments on total altered colonies(red, pink, and sectored) that arise by reciprocal andnonreciprocal recombination rather than on the sec-tored red/pink twin-spot colonies that indicate reciprocalrecombination (mitotic crossing-over) between the cen-tromere and ade2 in order to have sufficient statisticalpower to measure coincident recombination at low fre-quencies. However, all our experiments also showed anelevation in the frequency of ade2 twin-spot coloniesamong trp5 recombinants relative to that among all sur-vivors when we pooled data from different dosages, andit rose to the level of statistical significance in half theexperiments.

Split-dose experiments were used in attempts toinduce a recombination-prone state by means of DNAdamage. When cells were exposed to two successiveBLM treatments (0, 0.39 or 12.5 �g/ml for 2 h in phos-phate buffer, pH 7) separated by an intervening 16-hgrowth period in YEPD, convertant and revertant fre-quencies were compatible with additivity. A split-doseexperiment involving 2-h BLM treatments in buffer(0, 0.39 or 12.5 �g/ml) separated by a 4-h metabolismperiod in rich medium similarly produced frequenciesconsistent with additivity. Treating yeast with BLM for2 h in phosphate buffer at 0.1 and 1.56 �g/ml after theywere grown for 16 h in low doses of BLM (0.005 and0.02 �g/ml) gave recombinagenic and mutagenic effectsthat were additive or perhaps slightly less than addi-tive. When an initial 2-h BLM treatment of nondividingcells in phosphate buffer was followed by 20-h liquidholding and a second BLM treatment, the induction ofconvertants and revertants from the initial treatment wasconsistent with past experiments, and repair during liq-uid holding reduced their frequencies; a second 2-h BLMtreatment after liquid holding elicited a response equal toor slightly less than the first treatment. The interpretationof experiments involving successive treatments is com-plicated because the two treatments cannot be assumedto be identical in physiological conditions (e.g., oxygentensions; metabolic state) or proportions of cycling cells.Nevertheless, these experiments formed a consistent pat-tern, and they provide no evidence that BLM induces arecombination-prone state or increases the susceptibilityof D7 to the induction of mitotic recombination.

4. Discussion

Although represented by relatively few studies,reports that coincident genetic alterations at unlinked loci

n Research 616 (2007) 119–132

occur at frequencies greater than expected for indepen-dent events extend back over 40 years. Fogel and Hurst[24] reported in 1963 that UV-induced ad6 convertantsexhibited a 5–10-fold higher frequency of homozygosityassociated with mitotic recombination involving vari-ous other loci than did nonconvertants. In the sameyear, Wilkie and Lewis [25] showed that the frequencyof mitotic recombination at 10 loci in UV-irradiatedyeast was higher in those that were recombinant atanother locus than in total colonies. In a follow-up study,Hurst and Fogel [26] selected histidine prototrophs ina heteroallelic diploid and monitored their frequencyof homozygosity for an unlinked gene and four linkedgenes for which the strain had been heterozygous. Thedata were consistent with an elevated frequency ofcoincident recombination for both linked and unlinkedgenes. Montelone et al. [27] later determined that mitoticrecombinants for the unlinked lys2, tyr1 and ura3 lociand the linked trp5 locus are more common among ade5recombinants than among nonrecombinants. In humanlymphoblasts, elevated frequencies of coincident geneticevents were detected by screening thymidine kinasemutations induced by X-rays or ethyl methanesulfonatewith probes for variable number tandem repeats [28].Such studies suggest that processes common to phy-logenetically diverse eukaryotes influence proneness tomutation or recombination involving unlinked loci, evenin cells or organisms that are genetically homogeneouswith respect to known genes that influence recombina-tion, mutagenesis, and repair.

Fogel and Hurst [24] suggested that the increase incoincident recombination that they observed in yeastdepends on some general “precondition, process, orevent,” perhaps cell-cycle stage, and they proposed thata subpopulation prone to recombination may explainthe simultaneous recombinational events. Elevated fre-quencies of coincident recombination in yeast showsimilarities to bacterial transformation, in which higher-than-expected frequencies of cotransformation fordistant markers result from recombination-competencein part of the population [29]. Heterogeneity in recom-bination competence has also been advanced as anexplanation for coincident mitotic recombination inyeast [25,27,30].

Henaut and Luzzati [31] reported 10-fold enhance-ment of intragenic recombination in histidine-starvedyeast, suggested that histidine starvation may stimu-late chromosome pairing, and hypothesized that certain
subpopulations of yeast are more prone to undergorecombination. More recent evidence suggests thatchromosome pairing is not a limiting step in mitoticrecombination. Allelic and ectopic recombination occur

Mutatio

aihnMgctim

qgmgmidrtbdbDpartriatsuairi

etDtcardtaacc


t similar frequencies, and this implies that homologys searched throughout the genome, not exclusively onomologs [1]. Nevertheless, recombination-pronenesseed not depend on a chromosome-pairing mechanism.inet et al. [32] similarly concluded that there is hetero-

eneity in the tendency to undergo recombination amongells of Schizosaccharomyces pombe, and they proposedhat a fraction of the population is in a “parameiotic state”n which recombination frequencies are more typical of

eiosis than of mitosis.Several of these studies showed that elevated fre-

uencies of coincident recombination apply to linkedenes as well as unlinked genes [26,27]. Results forultiple recombinational events involving six linked

enes induced in newly formed diploids by X-ray treat-ent of one of the haploid parents support the same

nterpretation [33]. Subsequent analyses delineated twoistinct kinds of coincident recombination. That occur-ing between linked genes is distance-dependent, inhat the extent of enrichment for coincident recom-ination decreases with map distance [34,35]. Theistance-dependent phenomenon has been interpreted aseing a consequence of a long region of heteroduplexNA [27,34,35] or the concerted formation of multi-le recombinational intermediates along the length ofchromosome [35,36]. The other kind of coincident

ecombination is distance-independent [35]. Its exis-ence is supported by elevated frequencies of coincidentecombination among unlinked genes and by compar-sons of the recombinational loss of a plasmid insertiont linked and unlinked sites [35]. The extent of eleva-ion in coincident recombination for linked genes canometimes exceed 1000-fold [34,36], whereas that fornlinked genes is much smaller [35]. Linked genesre affected both by distance-dependent and distance-ndependent mechanisms, whereas only the latter areelevant for unlinked genes, including those that we stud-ed in strain D7.

We measured the frequency of coincident geneticvents for both mitotic recombination and point muta-ions induced by chemical mutagens and UV in strain7, a yeast assay that has found extensive use in genetic

oxicology. Frequencies of red and pink ade2 alteredolonies among Trp+ convertants were higher than thosemong all survivors or among Ilv+ revertants. If mitoticecombination at trp5 and ade2 behaved as indepen-ent events, frequencies of ade2 recombinants amongrp5 convertants should be given by the product of the
de2 recombinant frequency on nonselective mediumnd the trp5 convertant frequency. The actual frequen-ies are higher that those expected. The coefficient ofoincidence (observed/expected frequency of coincident
n Research 616 (2007) 119–132 127

genetic events) ranges from 2 to 5 for ade2 and trp5,whereas for ade2 and ilv1 it does not differ significantlyfrom the value of 1 expected for independent events.Thus, the elevated incidence of coincident genetic eventsseems restricted to double recombination rather thanencompassing point mutations. Although there is vari-ation in the coefficient of coincidence, the excess oftrp5 -ade2 double recombinants is highly reproduciblein many experiments. The coincidence coefficients arealso compatible with the finding of Golin and Tampe [35]that coincident recombination for the unlinked LEU1 andTYR1 genes is, at most, 4 to 8 times more frequent thanexpected for independent events.

An alternative explanation of our results that neededto be excluded is that the observation is an artifact owingto haploid yeasts arising under the experimental condi-tions and acting as phenocopies of mitotic recombinants.The conditions under which sporulation occurs in diploidyeast are well defined, and sporulation is not expectedunder typical growth conditions in young cultures peri-odically reisolated from single colonies. However, even alow frequency of haploids could be relevant if they mimicdouble recombinants. The developer of strain D7, F.K.Zimmermann, warned of cryptic sporulation as a poten-tial source of artifact in D7 [3,11]. To minimize this risk,conditions that trigger meiosis, such as acetate mediumand nitrogen starvation [3,11,37], were avoided; experi-ments were performed with fresh, characterized cultures;and recombinant and revertant frequencies were moni-tored to ensure that they were consistent with historicalcontrols.

We tested for the presence of haploid yeast and foundthe frequency of cryptic haploids in D7 to be about7.4 × 10−5. Known haploid controls were detected underthe same conditions with an efficiency of 75%. We didnot attempt to measure a/a or �/� diploids that mayarise by mating-type conversion. Such diploids wouldnot have the Trp+ Ade− phenotype that would mimiccoincident recombination. However, they may mate withhaploid strains [30] and be counted as haploids in ourmating type tests. If so, they would cause an overesti-mation of haploid frequencies, not an underestimation.Thus, they could not lead to an erroneous conclusionthat the frequency of haploids is too low to be a sourceof artifact.

Most meiotic products derived from D7 are Trp−,but a low frequency of Trp+ haploids may arise byintragenic meiotic recombination between trp5-12 and
trp5-27. Given that only Trp+ haploids would be a poten-tial source of artifact in our study, we though it prudent tomeasure their frequency, especially since meiotic recom-bination rates in yeast are higher than those in many

Mutatio
other eukaryotes and some chromosomal regions arehotspots for recombination [4]. The frequency of Trp+

haploids estimated from the frequency of cryptic sporu-lation and the proportion of Trp+ haploids in sporulatedD7 was 5.8 × 10−10. A direct measurement of Trp+ hap-loids in D7 gave a similar value: 6.4 × 10−10. Both linesof evidence indicate that the frequency of Trp+ hap-loids is much too low to explain the excess of doublerecombinants that we observed. Moreover, measure-ments after mutagenic treatment showed no increase inhaploids associated with BLM exposure. These experi-ments eliminate sporulation as a source of artifact in ourmeasurements of coincident recombination.

An excess of double recombinants could be explainedby heterogeneity among cells in susceptibility or inactual dosages of mutagen delivered to the genetic target.For example, consider typical yeast and a strongly-affected subpopulation that comprise 90% and 10% ofthe population and have induced frequencies of singlerecombination of 10−3 and 10−2, respectively. The sin-gle recombinant frequency in the population as a wholewould be (0.9 × 10−3) + (0.1 × 10−2) = 1.90 × 10−3.The expected frequency of double recombinants for thetotal population would be (1.90 × 10−3)2 = 3.61 × 10−6,but the observed frequency of double recombinantswould be (0.9 × 10−6) + (0.1 × 10−4) = 1.09 × 10−5.Thus, the observed frequency of double recombinantsin this example would be about three times higher thanthat expected on the basis of the single recombinantfrequencies.

We initially hypothesized that the excess of doublerecombinants induced by BLM in strain D7 (Table 2) isa manifestation in yeast of the overdispersion of BLM-induced genetic damage in other assays [15], such asthe finding that numbers of BLM-induced micronu-clei per cell do not conform to Poisson expectations inhuman lymphocytes [21,38]. If so, the overdispersionmay be a consequence of heterogeneity in permeabil-ity to BLM [15], such that cells exhibiting preferentialuptake would experience an intracellular dose of BLMhigher than the nominal dose. To test whether the excessof double recombinants is a unique attribute of BLM-induced damage, we quantified ade2 altered coloniesamong all survivors, Trp+ convertants and Ilv+ rever-tants after irradiation with UV (Table 4) or treatmentwith the electrophile �PL (Table 3). Elevated coincidentrecombination was observed in both cases.

The result with �PL shows that the phenomenon is
not unique to BLM, which causes DNA strand breaksby an indirect mechanism, but extends to a chemicalmutagen that acts through direct electrophilic attack,forming 2-carboxyethyl adducts in DNA [39,40]. Results
n Research 616 (2007) 119–132

with UV show that it is not restricted to chemicalmutagens whose passage through membranes can be alimiting factor. Data published by Zimmermann [13] onrecombinagenic effects of captafol in D7 also show anenrichment for double recombinants. Calculations basedon his data reveal a significantly elevated frequency ofade2 recombinants among trp5 convertants for severalindividual doses and a very highly significant difference(Fisher’s exact test; p < 0.0001***) for the experiment asa whole (255 ade2 recombinants/13,670 = 1.86% amongtrp5 convertants and 51/13,829 = 0.37% among all cells).Thus, data on BLM, �PL, UV, and captafol indicate thatthe phenomenon cannot be ascribed to a specific mech-anism of mutagenesis or to intercellular differences inpermeability.

To explore whether the enrichment for coincidentgenetic events is true of point mutations or applies onlyto recombinational events, we quantified ade2 alteredcolonies among Ilv+ revertants. The ilv1-92 allele issuppressible, and genetic evidence indicates that Ilv+

revertants arise both by base-pair substitutions and byframeshift mutations [3,14]. Data for treatments with�PL (Table 3), UV (Table 4), and BLM either dur-ing growth (Table 2) or in buffer (Table 5) all show asignificant enrichment for ade2 altered colonies amongTrp+ convertants but not among Ilv+ revertants. Thesefindings do not preclude a weak enrichment for ade2recombinants among revertants, but the evidence for itis minimal. The highly reproducible, statistically signifi-cant elevation in coincident genetic events encompassescoincident recombination but not point mutations. Thisresult provides further evidence that the phenomenoncannot be ascribed to a difference among cells in perme-ability to BLM or to haploid or aneuploid yeast, as theseshould affect convertants and revertants equally.

Having eliminated cryptic sporulation, uniqueattributes of BLM, and permeability to chemical muta-gens as explanations for the elevated incidence ofcoincident recombination, we explored whether it mightbe ascribable to a subpopulation in a recombination-prone or recombinagen-sensitive state. As noted in earlystudies, any environmental or cellular property thatmakes recombination more likely in a subset of the cellswould favor coincident recombination [24,41]. The sus-ceptible subpopulation could be a small fraction of thecells that is highly recombination-prone or a larger frac-tion that is slightly so. Factors that seem likely candidatesfor contributing to a sensitivity difference include cell
cycle stage or an inducible recombination-prone state,perhaps resulting from heterogeneity in the activationof recombinational repair in cells that have experiencedDNA damage. Local oxygen tension, while important in

Mutatio

ycbaafm

tiGtfa(bisausirrd

rgtMempmlbniobBaacubcosnef


east mutagenesis with BLM [15,42], is unlikely to beritical in the elevated frequencies of coincident recom-ination, as the phenomenon was observed in growingnd nongrowing cells, 2- and 16-h treatments, stationary-nd exponential-phase cells, and irradiation on the sur-ace of media, thus spanning a range of euoxic andarginally hypoxic conditions.Cell cycle stage and differences between exponen-

ial and stationary-phase cultures contribute to variationn susceptibility to the induction of genetic alterations.rowing cells tend to be more sensitive to most mutagens

han nongrowing cells in buffer [8,43]. The requirementor higher concentrations of BLM to induce genetic alter-tions in nondividing cells (Table 5) than in growing cellsTable 2) is a reflection of this difference. Mitotic recom-ination can be triggered by errors in DNA synthesis, butt also occurs in G1 and G2 [1,4]. While it is clear thatpontaneous and induced recombination occurs in cellsrrested at many points in the cell cycle [1], there is stillncertainty about the extent of recombinational repair atpecific stages of the cycle [4,36]. If a particular stages more susceptible to recombination, even if it is a nar-ow window of susceptibility, the fraction of cells in thisecombination-prone stage could generate an excess ofouble recombinants.

We explored effects of growth phase on coincidentecombination by analyzing 16-h treatments of yeast inrowth medium (Table 2) and 2-h treatments of cul-ures of different ages in phosphate buffer (Table 5).

utagenic treatment of fresh cultures in growth mediumxposes cells at all stages of the cell cycle. Buffer treat-ents, on the other hand, expose some cells at particular

oints in the cell cycle and others in G0. In buffer treat-ents of aging cultures (e.g., 48 or 96 h), most cells are

ikely to be in G0. A significant excess of double recom-inants was detected both in dividing cells (Table 2) andondividing cells (Table 5) treated with BLM, includ-ng those harvested from exponential cultures (12 h) andlder stationary cultures (48 h). The enrichment for dou-le recombinants persisted in a 96-h culture treated withLM in buffer at 25 �g/ml (Table 5). Even a cultureged at 4 ◦C for 3 months before treatment in buffert 25 �g/ml showed 2.2% ade2 altered colonies amongonvertants and 1.1% among total cells. Although these of older cultures is suboptimal owing to lower via-ility and the risk of sporulation, it let us extend ouromparison to yeast that no longer have buds and areverwhelmingly in G . Older cultures (e.g., 48, 96 h)
0how the same phenomenon as younger cultures in expo-ential or stationary phase and cells treated over thentire cell cycle during growth. Thus, the enrichmentor double recombinants occurs in G0 (Table 5), as well
n Research 616 (2007) 119–132 129

as in actively dividing cells (Table 2). The data supportthe conclusion that the enrichment for double recom-binants is not an obvious correlate of cell-cycle stage.The potency of induction is greatest in the youngestcultures, but the elevation in coincident recombinationoccurs in cultures of all ages. Although modest influ-ences of cell cycle stage have not been excluded, theyare unlikely to explain the excess of double recom-binants under diverse experimental conditions. If theexcess is ascribable to a portion of the population beingin a recombination-prone state, that state is unlikely tobe merely a window of susceptibility in the normal cellcycle.

Murray and Gottschling demonstrated a roughly 100-fold increase in mitotic recombination associated withcellular aging in yeast [44]. A mother cell can produceonly a limited number of daughter cells by budding,thus defining a finite lifespan of about 30 divisions[45]. Mother cells heterozygous for MET15 producedfew recombinant daughter cells in the first 25 divi-sions, after which there was an abrupt increase in thespontaneous frequency of gene conversion [44,45]. Oldmother cells are apparently defective in cell-cycle arrestafter the induction of double strand breaks. It has beenpostulated that an accumulation of damaged proteinscompromises genomic integrity by interfering with dam-age detection [44]. It is possible that older mother cellsin our study were more recombination-prone, but coin-cident recombination in cultures of all ages, includingyoung exponential cultures that have undergone rel-atively few divisions, suggests either that coincidentrecombination in D7 is not directly correlated with agingor that the influence of cellular aging is overshadowedby the processing of the high levels of DNA damage inmutagen-exposed cells.

Fabre and Roman proposed that coincident recom-bination may reflect the induction of a recombination-competent state by DNA damage [30,41]. The existenceof inducible responses to damage lends support to theirhypothesis. Transcription of RAD54, which is requiredfor double-strand break repair and recombination in bothmeiosis and mitosis, increases as much as six-fold afterDNA damage [46]. Activation of this and other genesmay increase recombination proneness. DNA lesionstrigger the formation of recombinational intermediatesthat lead to mitotic crossing-over and mitotic gene con-version [4]. Perhaps these lesions or intermediates alsotrigger a recombination-prone state, affecting not only
the vicinity of the lesions but other sites throughout thegenome. Such induction may increase the number ofrecombination-prone cells or the degree of proneness ofthe competent cells.

Mutatio
No process truly parallel to the bacterial SOSresponse has been found in yeast or other eukary-otes [46], but there are indications of damage-inducedchanges in responsiveness to mutagens. Increases inboth mutability and survival after UV exposure werereported in split-dose experiments in which yeast hadbeen given an initial exposure to a low dose of UV[47]. Fabre and Roman [30] suggested that the major-ity of vegetative cells are repressed for recombinationbut that irradiation causes the release of a factor neces-sary for recombination, thereby increasing the numberof cells prone to recombination. Some studies of UV-induced mitotic recombination support this hypothesis,in that UV dose–response curves have nonlinear regionsthat may be related to UV-inducible components of therecombination response [41,48]. Despite these hints ofan inducible state of proneness to genetic alterations,other studies have been unable to demonstrate its exis-tence. Split-dose irradiation of yeast led Schenk et al.[49] to conclude that no synergistic interaction betweenthe two UV doses is evident in mutation frequencies.They cautioned, however, that split-dose experiments areproblematic, owing to long incubation times and differ-ences in the physiological state of the cells receiving thefirst and second mutagenic exposures.

Our attempts to trigger a recombination-prone statehave been unsuccessful. Experiments using split dosesof BLM separated by liquid holding, metabolism, and/orgrowth showed appropriate genotoxic effects and repairduring liquid holding, but there was no heightenedsensitivity to subsequent mutagenic treatments. Priorexposure to low dosages of BLM similarly did notincrease the potency of recombinagenic effects in follow-up treatments. Our results for BLM are consistent withthose of Simon and Moore [50] for UV, in that prior expo-sure to UV did not enhance the induction of interallelicrecombination or mutation by subsequent UV exposures.The inability to induce a recombination-prone state doesnot preclude its existence, but it suggests that, if it exists,it is not triggered by DNA damage in a straightforwardway analogous to characterized inducible responses.

The induction of ade recombinants, trp recombinants,and Ilv+ revertants by BLM, �PL, and UV was dose-dependent, but the extent of the enrichment in doublerecombinants did not increase regularly with dose. Thislack of dose-dependence of coincident recombinationis consistent with the early data of Hurst and Fogel onUV-induced recombinants [26]. On this basis, we sus-
pect that mutagenesis with BLM, �PL or UV enabled usto measure double recombinational events by increas-ing the frequencies of all mitotic recombination, not byinducing a phenomenon that would not be observed at
n Research 616 (2007) 119–132

lower doses or even spontaneously if there were suf-ficient statistical power to measure it at low absolutefrequencies. This finding suggests that focusing on thepossibility of a mutagen-induced recombination-pronestate may be misleading, and that endogenous factorsthat may trigger a recombination-prone state in a subsetof the population deserve greater consideration.

We speculate that the existence of a recombination-prone subpopulation may be associated with transcrip-tional activity. Associations between transcription andrecombination have been demonstrated in systems asdiverse as E. coli and mammalian cells [8]. Transcriptionhas been linked to increased levels of recombination inyeast, but it is not known whether the increase is a directeffect or a secondary consequence of chromatin restruc-turing [4,8]. Transcription may stimulate the pairingof homologous sequences or affect the recombinationalmechanism [8], and the unwinding of transcribed DNAmay increase its exposure to recombinagens. Severallines of evidence suggest that the increase in recombi-nation is not uniform but is influenced by the specificpolymerase, promoter, and gene [8]. Evidence that theinsertion of rRNA promoter regions into the yeastgenome stimulates gene conversion and crossing-overfurther links transcription and recombination [4]. RNA-polymerase activity enhances recombination in yeast,not only at the site where transcription is occurring,but also more widely in the genome [8]. As cells exitstationary phase, over 2500 genes are transcribed in aperiod of a few minutes [23], and this period may con-fer recombination proneness as G0 cells enter the cellcycle. Since changes in chromatin structure and tran-scription alter levels of recombination with specificity,they fit the observation of an elevated incidence of coin-cident recombination without a concomitant increase incoincident point mutation. Those cells that are mosttranscriptionally active or have undergone the requi-site changes in chromatin structure may represent arecombination-prone subpopulation, as conjectured tounderlie the elevated frequencies of coincident recombi-nation.

In conclusion, our results show an elevated frequencyof ade2 recombinants among Trp+ convertants relative tothat among all surviving cells or among Ilv+ revertants inyeast strain D7. The enrichment is not mutagen-specific,in that it is observed among recombinants induced byBLM, �PL, or UV. We have experimentally ruled outthe possibility that the enrichment is an artifact caused
by sporulation. A likely explanation is the existence of asubpopulation in a recombination-prone state. The evi-dence does not favor a major role for cell-cycle stageor oxygen tensions in the recombination proneness, and

Mutatio

aDtqtoseccaeen

A

nRmsftTohekhtupd

R

[

[

[

[

[

[

[

[

[

[

[

[

[

[

[

[25] D. Wilkie, D. Lewis, The effect of ultraviolet light on recombi-


ttempts to induce the recombination-prone state byNA damage have been unsuccessful. It is conceivable

hat recombination proneness may arise as a conse-uence of changes in chromatin structure or increasedranscription. Our data show that elevated frequenciesf coincident mitotic recombination at unlinked loci aretatistically significant and highly reproducible in manyxperiments in a standard genotoxicity assay, but theirause remains obscure. A full appreciation of the impli-ations of the phenomenon for the distribution of geneticlterations in cells and populations awaits studies thatxtend the understanding of coincident recombinationalvents from the phenomenological level to the mecha-istic.

cknowledgments

We thank Malcolm Lippert, Heinrich Malling, Ken-eth Prestwich, Joseph Quaranta, Matt Ronan, andobert Schiestl for valuable discussions, F.K. Zimmer-ann for strains, and Darlene Colonna for excellent

ecretarial assistance. We are pleased to submit this workor inclusion in the Special Issue of Mutation Researchhat is dedicated to the memory of Anthony V. Carrano.ony Carrano was a model and inspiration for many, notnly for his distinguished scientific work, but also foris generous and exceptionally skilled service to the sci-ntific community. All those who had the pleasure ofnowing him also appreciated his character, judgment,umor, and kindness. For one of us (GRH), the chanceo contribute to this issue brings fond memories of a val-ed colleague and friend, and for the other (KMF), it is arivilege to submit her first scientific paper to a volumeedicated to someone who has contributed so much.

eferences

[1] M. Kupiec, Damage-induced recombination in the yeast Saccha-romyces cerevisiae, Mutat. Res. 451 (2000) 91–105.

[2] B.O. Krogh, L.S. Symington, Recombination proteins in yeast,Annu. Rev. Genet. 38 (2004) 233–271.

[3] F.K. Zimmermann, R. Kern, H. Rasenberger, A yeast strain for thesimultaneous detection of induced mitotic crossing over, mitoticgene conversion and reverse mutation, Mutat. Res. 28 (1975)381–388.

[4] T.D. Petes, R.E. Malone, L.S. Symington, Recombination inyeast, in: J.R. Broach, J.R. Pringle, E.W. Jones (Eds.), TheMolecular and Cellular Biology of the Yeast Saccharomyces, vol.1, Genome Dynamics, Protein Synthesis, and Energetics, ColdSpring Harbor Laboratory Press, 1991, pp. 407–521.

[5] F. Paques, J.E. Haber, Multiple pathways of recombination
induced by double-strand breaks in Saccharomyces cerevisiae,Microbiol. Mol. Biol. Rev. 63 (1999) 349–404.
[6] Y. Aylon, M. Kupiec, New insights into the mechanism of homol-ogous recombination in yeast, Mutat. Res. 566 (2004) 231–248.

[

n Research 616 (2007) 119–132 131

[7] G.R. Hoffmann, Induction of genetic recombination: conse-quences and model systems, Environ. Mol. Mutagen. 23 (suppl.24) (1994) 59–66.

[8] A. Aguilera, S. Chavez, F. Malagon, Mitotic recombinationin yeast: elements controlling its incidence, Yeast 16 (2000)731–754.

[9] A.J.R. Bishop, R.H. Schiestl, Role of homologous recombinationin carcinogenesis, Exp. Mol. Pathol. 74 (2003) 94–105.

10] F. Fabre, A. Chan, W.-D. Heyer, S. Gangloff, Alternate pathwaysinvolving Sgs1/Top3, Mus81/ Mms4, and Srs2 prevent formationof toxic recombination intermediates from single-stranded gapscreated by DNA replication, Proc. Natl. Acad. Sci. U.S.A. 99(2002) 16887–16892.

11] F.K. Zimmermann, R.C. von Borstel, E.S. von Halle, J.M. Parry,D. Siebert, G. Zetterberg, R. Barale, N. Loprieno, Testing ofchemicals for genetic activity with Saccharomyces cerevisiae: areport of the U.S. Environmental Protection Agency Gene-ToxProgram, Mutat. Res. 133 (1984) 199–244.

12] F.K. Zimmermann, Procedures used in the induction of mitoticrecombination and mutation in the yeast Saccharomyces cere-visiae, Mutat. Res. 31 (1975) 71–86.

13] F.K. Zimmermann, Tests for recombinagens in fungi, Mutat. Res.284 (1992) 147–158.

14] E. Gundelach, Suppressor studies on ilv1 mutants of Saccha-romyces cerevisiae, Mutat. Res. 20 (1973) 25–33.

15] L.F. Povirk, M.J.F. Austin, Genotoxicity of bleomycin, Mutat.Res. 257 (1991) 127–143.

16] L.F. Povirk, DNA damage and mutagenesis by radiomimeticDNA-cleaving agents: bleomycin, neocarzinostatin and otherenediynes, Mutat. Res. 355 (1996) 71–89.

17] R.M. Burger, Cleavage of nucleic acids by bleomycin, Chem. Rev.98 (1998) 1153–1169.

18] J. Chen, J. Stubbe, Bleomycins: towards better therapeutics, Nat.Rev. Cancer 5 (2005) 102–112.

19] G.R. Hoffmann, A.M. Sayer, L.G. Littlefield, Potentiation ofbleomycin by the aminothiol WR-1065 in assays for chromoso-mal damage in G0 human lymphocytes, Mutat. Res. 307 (1994)273–283.

20] G.R. Hoffmann, J. Buccola, M.S. Merz, L.G. Littlefield,Structure-activity analysis of the potentiation by aminothiols ofthe chromosome-damaging effect of bleomycin in G0 humanlymphocytes, Environ. Mol. Mutagen. 37 (2001) 117–127.

21] G.R. Hoffmann, S.P. Colyer, L.G. Littlefield, Induction ofmicronuclei by bleomycin in G0 human lymphocytes. I. Doseresponse and distribution, Environ. Mol. Mutagen. 21 (1993)130–135.

22] D. Burke, D. Dawson, T. Stearns, Methods in Yeast Genetics:A Cold Spring Harbor Laboratory Course Manual, Cold SpringHarbor Laboratory Press, 2000.

23] M. Radonjic, J.-C. Andrau, P. Lijnzaad, P. Kemmeren, T.T.J.P.Kockelkorn, D. van Leenen, N.L. van Berkum, F.C.P. Hois-tege, Genome-wide analyses reveal RNA polymerase II locatedupstream of genes poised for rapid response upon S. cerevisiaestationary phase exit, Mol. Cell 18 (2005) 171–183.

24] S. Fogel, D.D. Hurst, Coincidence relations between gene con-version and mitotic recombination in Saccharomyces, Genetics48 (1963) 321–328.

nation in yeast, Genetics 48 (1963) 1701–1716.26] D.D. Hurst, S. Fogel, Mitotic recombination and heteroallelic

repair in Saccharomyces cerevisiae, Genetics 50 (1964) 435–458.

Mutatio

[

[

[

[

[

[

[

[

[

[

[

[

[

[

[

[

[

[

[

[

[

[


27] B.A. Montelone, S. Prakash, L. Prakash, Spontaneous mitoticrecombination in mms8-1, an allele of the CDC9 gene of Saccha-romyces cerevisiae, J. Bacteriol. 147 (1981) 517–525.

28] C.-Y. Li, D.W. Yandell, J.B. Little, Evidence for coincident muta-tions in human lymphoblast clones selected for functional loss ofa thymidine kinase gene, Mol. Carcinogen. 5 (1992) 270–277.

29] E.W. Nester, B.A.D. Stocker, Biosynthetic latency in early stagesof deoxyribonucleic acid transformation in Bacillus subtilis, J.Bacteriol. 86 (1963) 785–796.

30] F. Fabre, H. Roman, Genetic evidence for inducibility of recom-bination competence in yeast, Proc. Natl. Acad. Sci. U.S.A. 74(1977) 1667–1671.

31] A. Henaut, M. Luzzati, Controle de l’aptitude a recombiner pen-dant la phase vegetative chez Saccharomyces cerevisiae, Mol.Gen. Genet. 116 (1972) 26–34.

32] M. Minet, A.-M. Grossenbacher-Grunder, P. Thuriaux, The originof a centromere effect on mitotic recombination: A study in thefission yeast Schizosaccharomyces pombe, Curr. Genet. 2 (1980)53–60.

33] D.A. Campbell, The induction of mitotic gene conversion byX-irradiation of haploid Saccharomyces cerevisiae, Genetics 74(1973) 243–258.

34] J.E. Golin, M.S. Esposito, Coincident gene conversion dur-ing mitosis in Saccharomyces, Genetics 107 (1984) 355–365.

35] J.E. Golin, H. Tampe, Coincident recombination during mitosisin Saccharomyces: distance-dependent and -independent compo-nents, Genetics 119 (1988) 541–547.

36] J.E. Golin, S.C. Falco, J.P. Margolskee, Coincident gene conver-sion events in yeast that involve a large insertion, Genetics 114(1986) 1081–1094.

37] Z. Olempska-Beer, Current methods for Saccharomyces cere-visiae. II. Sporulation, Anal. Biochem. 164 (1987) 278–286.

38] G.R. Hoffmann, S.P. Colyer, L.G. Littlefield, Induction of
micronuclei by bleomycin in G0 human lymphocytes. II. Poten-tiation by radioprotectors, Environ. Mol. Mutagen. 21 (1993)136–143.
39] D.J. Brusick, The genetic properties of beta-propiolactone, Mutat.Res. 39 (1977) 241–256.

[

[

n Research 616 (2007) 119–132

40] G.R. Hoffmann, R.A. Shorter, J.L. Quaranta, P.D. McMaster, Twomechanisms of antimutagenicity of the aminothiols cysteamineand WR-1065 in Saccharomyces cerevisiae, Toxicol. In Vitro 13(1999) 1–9.

41] F. Fabre, Mitotic transmission of induced recombinational abil-ity in yeast, in: E.C. Friedberg, B.A. Bridges (Eds.), CellularResponses to DNA Damage, Alan R. Liss, Inc., New York, 1983,pp. 379–384.

42] G.R. Hoffmann, J.L. Quaranta, R.A. Shorter, L.G. Littlefield,Modulation of bleomycin-induced mitotic recombination in yeastby the aminothiols cysteamine and WR-1065, Mol. Gen. Genet.249 (1995) 366–374.

43] F.K. Zimmermann, Basic principles and methods of genotoxicitytesting in the yeast Saccharomyces cerevisiae, in: B.J. Kilbey, M.Legator, W. Nichols, C. Ramel (Eds.), Handbook of Mutagenic-ity, 2nd ed., Elsevier Scientific Publishers, Amsterdam, 1984, pp.215–238.

44] M.A. McMurray, D.E. Gottschling, An age-induced switch to ahyper-recombinational state, Science 301 (2003) 1908–1911.

45] M.A. McMurray, D.E. Gottschling, Aging and genetic instabilityin yeast, Curr. Opin. Microbiol. 7 (2004) 673–679.

46] E.C. Friedberg, W. Siede, A.J. Cooper, Cellular responses to DNAdamage in yeast, in: J.R. Broach, J.R. Pringle, E.W. Jones (Eds.),The Molecular and Cellular Biology of the Yeast Saccharomyces,vol. 1, Genome Dynamics, Protein Synthesis, and Energetics,Cold Spring Harbor Laboratory Press, 1991, pp. 147–192.

47] F. Eckardt, E. Moustacchi, R.H. Haynes, On the inducibility oferror-prone repair in yeast, in: P.C. Hanawalt, E.C. Friedberg,C.F. Fox (Eds.), DNA Repair Mechanisms, Academic Press, NewYork, 1978, pp. 421–423.

48] B.A. Kunz, F. Eckardt, R.H. Haynes, Analysis of nonlinearities infrequency curves for UV-induced mitotic recombination in wild-type and excision-repair-deficient strains of yeast, Mutat. Res.151 (1985) 235–242.

49] K. Schenk, F. Zolzer, J. Kiefer, Mutation induction in haploidyeast after split-dose radiation exposure. I. Fractionated UV-irradiation, Radiat. Environ. Biophys. 28 (1989) 101–111.

50] J.R. Simon, P.D. Moore, Induction of homologous recombinationin Saccharomyces cerevisiae, Mol. Gen. Genet. 214 (1988) 37–41.

frequencies of mutagen-induced coincident mitotic recombination at unlinked loci in saccharomyces...

Documents