P h t PathoZogy (1998) 47, 601-608

Genetic and pathogenic variation of Xanthomonas axonopodis pv. manihotis in Venezuela

V/rd ie r ” *+ , S. Restrepob, G. Mosquerab, M. C. Duqueb, A. Gerstl‘ and R. Laberry‘ a Institut Français de Recherche Scientifique pour le Développement en Cooperation, Laboratoire de Phytopathologie Tropicale. 6P 5045, 34032 Montpellier, France; Centro lnternacional de Agricultura Tropical, AA 6713, Cali, Colombia; and ‘Agropecuaria Mandioca CA, AA 394, Maturin, Monagas, Venezuela

DNA polymorphism and variation in virulence of Xanthomoizas axoizopodis pv. inanihotis (Xanz), the causal agent of cassava bacterial blight, were studied within a pathogen population from Venezuela. Collections were made in several fields at different sites within an edaphoclimatic zone where cassava is a major crop. DNA polymorphism was assessed by RFLP analysis, using an X a m plasmidic DNA sequence (pthB) as a probe to determine the relatedn&.s of 91 Venezuelan isolates. A high degree of polymorphism existed among the isolates, whether collected from the same or different fields. Based on a multiple correspondence analysis, the Xam population was distributed -into eight clusters and no correlation was observed between genetic diversity and geographic origin. One set of haplotype strains representing the range of variability detected in Venezuela was further characterized by another RFLP analysis using two repetitive genomic probes (pBS6 and pBS8) to establish the usefulness of these probes and their complementarity with the pthB probe. Variation for virulence was observed in the Xanz Venezuelan collection by inoculating a set of cassava cultivars with 28 isolates of the pathogen, each representing a haplotype. Understanding the genetic and pathogenic variation in the pathogen population is useful for designing cassava bacterial blight management strategies.

Introduction

Cassava (Manihot esculenta) is a starchy root and a major tropical food crop in Africa, Latin America and Asia. In Venezuela, cassava is grown over 37 O00 ha, for a total annual production of 285 O00 t. It evolved rapidly from a subsistence to a market-orientated crop during the past few years. Venezuelan farmers are targeting the cassava-processing industries as a viable and attractive marltet for their production.

Yield losses occur through diseases and pests. One significant foliar and stem disease is the cassava bacterial blight (CBB), caused by Xarzthomonas axonopodis pv. nzanibotis (Xanz). CBB is widespread and was first described in South America in 1920 and in Africa in the 1970s (Maraite et al., 1981). Yield losses ranging from 12 to 100% were reported (Lozano & Sequeira, 1974); and the disease is also a major constraint for the production of healthy planting material. In Venezuela, the disease was first reported in the eastern and central regions (Gonzales, 1973; Trujillo et al., 1980).

* To whom correspondence should be addressed.

t E-mail: v.verdier@cgnet.com

Accepted 16 March 1998.

’ Q 1998 BSPP - / i

X a m can cause a wide combination of symptoms, including angular leaf lesions, blight, wilt, stem exu- dates, stem canker and dieback (Maraite, 1993). Losses from this disease can be reduced by a combination of cultural practices and host resistance (Lozano, 1986; Boher & Verdier, 1994). Resistance to CBB has been developed mostly from M. esculenta and M. glaziouii, and is thought to be polygenic and additively inherited (Hahn, 1979). However, genes governing resistance have not yet been identified.

During the past decade, advanced molecular techni- ques were developed for ,identifying and characterizing plant pathogenic bacteria. Restriction fragment length polymorphism (RFLP), using DNA probes, was used to differentiate strains within or among Xaizthoinoizas campestris pathovars and to conduct epidemiological studies (Ardales et al., 1996; Vera Cruz et al., 1996; Gagnevin et al., 1997). The diversity of Xanz popula- tions was investigated for virulence (Elango & Lozano, 1981; Grousson et al., 1990; Verdier et al., 1993), physiological traits (Maraite et al., 1981), protein patterns (Vauterin et aZ., 1991) and DNA polymorph- isms (Berthier et al., 1993; Verdier et al., 1993). By using a universal probe (16 + 23s rRNA probe from Escher- ichia coli), five major genetic groups were detected among several isolates of the pathogen collected around

Fonds Documentaire ORSTOM Ex: A

601

!' . 602 y Verdier et al.

v the world (Verdier et al., 1993). The polymorphisms revealed by a DNA plasmid probe also permitted clarification of the phylogenetic relationships among Xanz collections and showed a higher level of diversity within South American than within African populations of the pathogen (Verdier et al., 1993). Recently, the distribution of pathogen diversity among agroeco- systems was studied in Colombia, and geographic differentiation of the pathogen was demonstrated (Restrepo & Verdier, 1997).

Wide variations in aggressiveness have been reported among isolates from various countries and from different areas within a given country (Grousson et al., 1990; Maraite, 1993). However, no clear-cut interactions were established between cassava cultivars and isolates. The existence of physiological races among Xam populations is still under investigation.

The objectives of the present study were (1) to characterize genetic diversity in the Xanthonzoizas axoizopodis pv. manihotis population at different sites located in three Venezuelan States, (2) to study the structure of Xanz populations on a regional scale by comparing the results obtained in three States, (3) to determine relationships between haplotypic and patho- typic analyses, and (4) to assess the possible existence of specific isolate-genotype interactions.

Materials and methods

Collection of infected leaf samples

Eight localities were selected from three States: Monagas (MN), Anzoátegui (AZ) and Bolívar (BV) (Fig. 1). Severe symptoms of bacterial blight were observed in all the sites visited except near Ciudad Guyana, Bolívar, where the disease was observed in only one field, at Santa Rosa.

Collections were made at eight \ SI *the rainy season in November 1995: Agropecuaria Mandioca (MN) (site A), Santa Bárbara (MN) (site B), Los Corhs-QvJN) -___.- - (site C), Maturin (MN) (site D), Santa Rosa (near Upata, BV) (site E), El Tigre (AZ) (site F), Mucura (AZ) (site G) and Boca del Pozo (AZ) (site H) (Fig. 1). Three to five fields were visited at sites A, C, D and G, while only one field was sampled at sites By E, F and H. Distances between fields varied from 1 to 20 km. One to six leaves were collected from each cultivar, and at different locations in each field. Field size varied greatly, but the average was about 1 ha.

-.

Bacterial isolates

The leaves were stored at room temperature (20°C) for 10 days before isolation of the pathogen. Angular leaf lesions were removed and macerated in sterile distilled water. The resulting suspension was spread onto YPG medium (yeast extract 5gL-l, Bacto peptone 5 g L-I, glucose 5 g L-' and agar 15 g L-', pH 7.2) and plates were incubated at 30°C for 48 h. A single colony was selected per sample, and isolates were maintained on the same medium at 4°C (for routine use) or stored in 60% glycerol at -80°C. A total of 91 isolates was produced; collections CUT 1129 and Orst 2 from Maracaibo (Zulia) isolated in 1974 and 1971, respect- ively, were included in the study as reference strains (Table 1).

WLP analysis

Genomic DNA was extracted from 3mL of culture grown overnight in nutrient broth, as described by Berthier et al. (1993). The DNA pellet was dissolved in 100 pL of TE (10 mM Tris, 1 m EDTA, pH 8.0) and the

Figure 1 Geographical location of the different sites in Venezuela from which isolates of Xanthomonas avonopodis pv. manhotis were collected. Table 2 describes the sites.

O 1998 BSPP Plant Pathobgy (1998) 47, 601-608

Xanthomonaspopulation in Venezuela 603

Table 1 Source and origin of the isolates of Xanthomonas avonopodis pv. manihotk used in this study

Strain/isolate numbeP Locality (Site) State Haplotype Clustep

CIAT Il29 (reference strain) Maracaibo (Mc) Zulia * 8 ORST 2 (reference strain) . Maracaibo * 8' 8 CIO 201 A. ,Mandioca (A) Monagas v4 5 CIO 200,202,203 A v5 7 CIO 204.2~5,207,209,210.212,213,214 A V6 7 CIO 217,218,219 A v9 7

A VI 2 4 CIO 193,195,196,199 A VI 3 5 CI0191,192 '

CIO 198 A VI 4 4 CIO 215,216,220 A VI5 5 CIO 208 A VI7 8 CIO 206,221 A VI 8 5 CIO 197 A V24 4 CIO 194 , A V28 5 CIO 222 Santa Bárbara (B) Monagas V8 5 CIO 223 B v5 7 CIO 224,225 B V6 . 8 CIO 227 Los Corocitos (C) Monagas VI 6 n 6 %IO231 ' C VI 8 5 CIO 228 i c VI9 c 64

i CIO232 ' C 4 v21 5

CIO234 C V23 4 'i CIO 241,242,243,244,245 Maturin (D) Monagas V6 8 J CIO 239,240 D v7 5 2

F CIO 235 D v21 5 CIO 236,237 D v24 4 CIO 238 D V25 5

CIO 247 E v5 7 CIO 257 EI Tigre (F) Anzoátegui v4 5 CIO 248,249,250,251,253, 254,255,258 F v5 7 CIO 256 F V8 5 CIO 252 F V26 5 CIO 261,262,264,271 Mucura (G) Anzoátegui VI 8 CIO 259,260 G v5 7 CIO 263,265,267 G v9 4 CIO 266 G VIO 2 CIO 268 G VI 1 2 CIO 272 G V27 4 CIO 274,275,276,278,279,280,281,282,283,284 Boca del Anzoátegui VI 8

Pozo (H) CIO 277 H v2 1 CIO 285 H v3 3

aAbbreviations: CIAT, Xanthomonas collection Dr Lozano, CIAT, Cali, Colombia; CIO, ORSTOM-CIAT collection, CIAT, Cali, Colombia; ORST, Collection du Laboratoire de Phytopathologie, ORSTOM, Montpellier, France. b A s defined by the multiple correspondence analysis and as described in the text. 'As previously described (Verdier et a/., 1993).

CIO 226,229,230 t C 8 v20 6

' CI0233 C V22 4

p i CIO 246 Santa Rosa (E) Bolívar V8 5

concentration was estimated with a spectrophotometer. Bacterial DNA (7pg) was digested overnightwith EcoFU at 37"C, as recommended by the manufacturer (Gibco BRL, Life Technologies, MD, USA). DNA fragments were separated by electrophoresis in 0.7% agarose in 1 xTAE buffer ( 4 0 m Tris, l m ~ EDTA, adjusted to pH7-6 with glacial acetic acid) at 45V for 1 5 h and blotted onto Hybond Nf (Amersham International, UK)

by alkali transfer, as described by the manufacturer. A size marker was included in each gel (RaoulI, Eurogentec, Liege, Belgium). The pthB probe, a blue- script plasmid containing a 5.6-kb EcoFU fragment isolated from Xam CFBP1851, was used to detect polymorphism because it resolved more RFLP banding patterns than other probes (Verdier et al., 1993; Restrepo & Verdiel; 1997). Probe pBS6 was used

O 1998 BSPP Plant Pathology(l998) 47, 601-608

604 l? Verdier et al. b

because it contains a high copy-number repetitive element present in DNA from over 400 Xanthomonas azonopodis pv. manihotis isolates (Verdier et al., 1993; Restrepo & Verdier, 1997). Probe pBS8 is useful for distinguishing X a m strains that contain either a high or low copy-number repetitive element (Verdier e t al., 1993). Probes were labelled by random priming with [32P]dATP, according to the manufacturer’s instructions (Multiprime, Amersham). Membranes were exposed to X-ray films (Kodak) for 14h at -80°C with intensifying screens.

Data ,analysis

Each distinct RFLP -banding pattern revealed with each probe was regarded as a haplotype. Haplotypes generated for each isolate were first compared visually and grouped according to unique banding patterns. Representative isolates of a haplotype were then analysed in the same gel to confirm band positions. This allowed conversion to binary data, coded as 1 for presence or O for absence of a band. Band weights were determined at each position, using the size marker as reference. Each pattern was compared with data available in a haplotype database previously established with Xanz collections from Colombia (Restrepo & Verdier, 1997). A multiple correspondence analysis was generated with the SAS option CORRESP (SAS, Institute, Inc., Cary, NC), which determined the posi- tions of the isolates on a three-dimensional graph drawn by the program JMP (version 3.1, SAS). The number of clusters was determined by consensus among three clustering statistics (SAS). Average distances between and within clusters were obtained by the SAS IML software. For each pair of isolates the Jaccard index of similarity (s) was calculated and distances were obtained as (1 - s). The genetic diversity for all X a m isolates (HT) and from each site (Hs) was estimated by Nei’s diversity index fyei & Tajima, 1981), based on the following equatiori:

<

where X , is the proportion of the i& haplotype at each site or the frequency of the ith haplotype in the entire population and n is the number of isolates at each site .or in the entire population. The standard error (o) was estimated from the sampling variance for each popula- tion and for the total population as described by Vera Cruz et al. (1996). Significance was declared if Hs - 2 u was greater than HT.

Pathogenicity tests

All 91 X a m isolates were tested for virulence on the susceptible cultivar MCOL1522, in a glasshouse at 28/ 19°C (daylnight temperatures), under a 12-h daylight photoperiod and 80% relative humidity. The cassava plants were grown from mature stem cuttings in sterile

soil. Steins were inoculated as described by Verdier et al. (1994). Three plants, randomly distributed in the glasshouse, were inoculated with each isolate. Disease progress was monitored 7, 14 and 30days after inoculation, and disease severity was rated as follows: score O = no disease symptoms; 1 = necrosis around the inoculation point; 2 = gum exudation on stem; 3 = wilting of one or two leaves and exudation;-%= wilting of more than two leaves; 5 = complete wilting and dieback. Isolates with an average score ranging from O to 2.49 were considered as nonaggressive, from 2-5 to 3-49 as weakly aggressive and 3-5 to 5 as very aggressive.

In another experiment, a set of 28 isolates, represent- ing 28 haplotypes, was used to inoculate five cassava cultivars (Mcol 22, CM 2177-2, Mnga 2, SG 107-35 and CM 523-7) as described above. Three plants of each cultivar, randomly distributed in the glasshouse, were inoculated with each haplotype strain. Non-inoculated plants were included in each experiment as control. The five cultivars had previously shown differential reactions with &znz reference strains (Restrepo et al. unpublished data). For simplicity, plants with a disease reaction of 5 3 were grouped as resistant, while those with a difease reaction of >3 were classified as susceptible. Data on disease reaction of each cultivar-strain combination, i.e. differences among means, were examined by the Kruskal-Wallis test .(SAS).

Results

RFLP analysis

Twenty-eight distinct banding patterns, e x h regarded as a haplotype, were found with the pthB probe among the 91 Venezuelan isolates collected in 1995 (Table 1, Fig. 2). The 28 haplotypes were clearly distinct from the haplotypes previously described for Colombian Xanz collecfions. Between four and eight hybridizing bands were pbserved per isolate, and a total of 26 band positions was scored (Fig. 2). Fourteen haplotypes were represented by a single isolate each, while others were represented by as many as 15 isolates (haplotype VS). Nine haplotypes (Vl, V4, VS, V6, V8, V9, V18, V21 and V24) were detected in more than one site. By contrast, the 19 other haplotypes, including the 14 represented by only one ísolate each, as well as haplotypes V7 (two isolates, site D), V12 (two isolates, site A), V13 (four isolates, site A), V15 (three isolates, site A) and V20 (three isolates, site C) were site-specific (Table 1). The total haplotypic diversity (HT) of the X a m collection analysed by pthB RFLP was 0.92 (Table 2).

Diversity was highest at sites A (0.90) and C (0093)~ with 12 and 7 haplotypes among 30 and 9 isolates analysed, respectively. S i x haplotypes present in site A were also detected in one or more of the sites By Cy Dy E, F and G (Table l), which are located in the areas from which planting material used in site A originated. Diversity at sites E and B was not considered repre- sentative in this study because only a few isolates were

Q 1998 BSPP Plant Pafhology(l998) 47, 601-608

Xanthomonas population in Venezuela 605

a KbP

18 *

5.61)

3.9*

h

Figure 2 Southerh''B!ot 'analysis of €cod1 digested genomic DNA from X a n t b o m o n a s - a o d j s pv. 'manihoEs isolates probed with ' the 32P-labe¡led ;OthB probe;Of the.28 haplotypes (Ht) detected in Venezuela and.thetwo.reference strains (CIAT 1129 and ORST 2), 21 haplotypes are4liustrated in the following order: (a) lanel, HtV26 lane.2, HtV12;lane 3; HtVl;.lane 4, HtV4; lane 5, HtV5; lane 6, HtV9; lane 7. HtV23; iane.8, HtV24; lane.9, HtV22; lane 10, HtV10; lane 11, HtVl1; lane 12;st[ain ClAT,1129 Ht; (b), lane 1,.HtV18; lane 2 HtV21; lane.3, HtV15 lane 4, HtV20; lane.5; HtV25, iane 6, HtV4; "

lane 7, HtV19; lane,8, GtV7; lane.9, HtV28; R,'Raoull size' standard. . . , . . . . . . . . . . . . . . , i , I .

collected at these sites. Genetic diversity at sites H (0-33) and F (0.51) was significantly lower than haplotype diversity in the whole popidation (Table 2). '

A total of 28 isolates representing the 28 Venezuelan haplotypes defined with the ptbB probe, were selected and characterized with two repetitive probes, pBS6 and pBSS. The pBS6 probe hybridized with a high number of different fragments, whereas the pBS8 probe hybridized with only two to three fragments. Three haplotypes were found with pBS6 and four with pBS8 (data not shown). Haplotype pBS6-1 grouped four isolates belonging to ptbB-haplotypes V17, V18, V21 and V22; haplotype pBS6-2 contained two isolates representing pthB- haplotypes V8 and V26, and haplotype pBS6-3 grouped 22 isolates belonging to 22 different pthB-haplotypes (Vl, V2, V3, V4, V5, V6, V7, V9, V10, V11, V12, V13, V14, V15, V16, V19, V20, V23, V24, V25, V27 and V28). Of the four haplotypes found with pBS8, haplotype pBS8-1 grouped two isolates belonging to pthB-haplotypes V8 and V26; haplotype "pBS8-2 grouped three isolates (ptbB-haplotypes V10, V>1 and V28), haplotype pBS8-3 included fol; isolated8pthB- haplojypes V17, V18, V21 and V22); and haplotype 4 grouped 19 isolates representing 19 pthB-haplotypes. As hybridization with the pthB probe resulted in the revelation of the highest diversity, probes pBS6 and pBS8 were not used further in this study.

.

Cluster analysis

Based on the multiple correspondence analysis, the 93 isolates were grouped into eight clusters (Table 1, Fig. 3 ) . In general, clusters contained isolates collected in different sites (e.g. clusters 4, 5, 7, and 8). Clustër 7 contained 15 isolates from five different sites, all showing the same haplotype (V5). The two reference strains were included in cluster S. Clusters 1, 2 and 3 contained one or two isolates collected in the western sites of Múcura and Boca del Pozo. Mean genetic distances between and within clusters are shown in Table 3.

Table 2 Number of isolates, haplotypes and haplotypic diversity (Hs) in each site and for the Venezuelan Xantbomonas auonopod/s pv. sites Location manhotis collection (b)

Number of Number of pfbB haplotypes Hsb

A A. Mandioca MN 30 12 o.9oc B Santa Barbara MN 4 3 0.83

Statea isolates

C Los Corocitos MN 9 7 0.93 D Maturin MN 11 5 0.78 E Santa Rosa BV 2 2 1 .o0 F EI Tigre AZ 11 4 0.51 G Mucura AZ 12 6 0.85' H Boca del Pozo AZ 12 3 0.33'

Total 91 28 6 = 0.92

. =MN = Monagas: BV = Bolívar; AZ= Anzoátegui. H, = Haplotypic diversity.

CSignificantly.different from the total diversity as described in the text.

O 1998 BSPP Plant Pafbology(1998) 47,601-608

i

606 V.Verdier et al.

Cluster 2

Cluster 8 N=32

I Cluster 4

I I N = 1 6

, " * 'a

o o Cluster5 Cluster I N=15 N = 2 5

% e

Figure 3 fhre&dimensìonal.graph from a mdtiple correspondence analysis (version 6,:SAS/IML Software), showing the relationships of' Xantbomonasaxonopodis pv. manibofis clusters after RFLP analysis with pthB probe (N = number of isolates in each cluster). The graph was drawn by thq program JMP (version 31; SAS). Clusters were assigned by the average linkage method. The consensus. among three clustering statistics, pseudo F2, pseudo t2, and cubic clustering criterion indicated eight clusters. The first three dimensions (x, y, z) accounted for 17, 13 and 12% (total = 42%) of the variation for RFLP analysis. The haplotype and the number (N) of isolates contained in each cluster are indicated. Symbols'denote 'cluster members. . ' . ;

.

Virulence analysis

All the 91 Venezuelan X a m isolates caused characteristic CBB symptoms on the susceptible cultivar Mco11.522. Eighty-nine (98%) were very aggressive (disease hdex >3-5) &d only two (2%) were weakly aggressive (2.5 5 disease'hdex 5 2-5).

No significant differences were observed for disease index on MCOL1522 among the isolates belonging to the same haplotype (data not Shawn). Therefore, one

Table 3 Mean genetic distance between and within clusters, generated by the multiple correspondence analysis of 93 isolates of Xanthomonas axonopodis pv. manihotis obtained with RFLP-ptbB analysis

'

Cluster 1 2 3 4 5 6 7 8

1 1 0.08 0.27 0.18 0-15 2 0-29 0.05 0.19 0.08 3 1 0.14 0.17 4 - 0.52 0.25 5 0-51 6 7 8

0.25 0.14 0.13 0.03 027 0.13 0.4 027 0.23 0.27 0-21 0.2 0.21 0.28 0.19 1 0.35 .O27

1 0.31 0-7

isolate from each of the 28 haplotypes identified was

fuhess of RFLP analysis with pthB as a probe, already outlined in a~imilar study on populations from Colombia (Restrepo &Verdier, 1997), for investigating population structures of X a m in South

probes (such as aurXalO from Xaizthomonas oryzae pv. oryzae) are usefui in population studies because of their role in plant-pathogen interactions (Nelson et al., 1994; Adhikari et al., 1995). It is ndnetheless some- what surprising that pthB allowed detection of more polymorphism t6an either, pBS6 or pBSS, which are

1996; Vera Cruz et al., 1996). However, plasmidic probes have also proved useful i s population analyses of othei bacterial pathogens (Gagnevin et al., 1997).

These results indicate that considerable genetic diversity exists within the X a m population in Venezuela, which is consistent with the high level of diversity preseht within South American populations of this pathogen (Elango & Lozano, 1981; Maraite et al., 1981; Verdier; 1988; Grousson et al., 1990; Verdier et al., 1993; Restrepo & Verdier, 1997). This high level of diversity can be related to the centres of genetic diversity in the host, since the Manihot genus originates from Latin America, with Brazil as one of the main diversity centres (Rogers & Appan, 1973).

In Venezuela, the X a m population has shown a high degree of haplotypic diversity in each given site, e.g. in Agropecuaria Mandioca and other nearbv sites. This

i

- - - diversity may be related to the large number of host

génotypes (traditional and improved cultivars) grown at these sites. An additional factor may be the introduction of new strains of the pathogen with experimental

-material. This is particularly the case for site A I where, since 1992, vegetative planting material 'has

been introduced from different locations in Venezuela. The comparison of genetic patterns among isolates from various locations in Venezuela, suggests that the

O 1998 BSPP PlantPathology(1998) 47, 601-608

I Xanthomonas popdation in Venezuela 607

Table 4 Reaction of five cassava genotypes to the 10 pathotypes defined among the 28 representative haplotype strains of Xanthomonas axonopodis pv, manihofis from Venezuela

Disease reactiona

Representative haplotype strains Cluste? Pathotype MCOL 22 CM 2177-2 MNGA 2 S G 107-35 . CM 523-7

C10268, CIO285 2 , 3 1 s S S S S CI0238 5 2 S S S R S C10208, CI0257 8.5 3 S R R S S CI0256 5 4 S R S R S C10198, C10258, CI0252 4 , 7 , 5 5 s, , R R R S C10277, C10199, C10282, C10228, 1 , 5 , 8 , 4 6 s * S R R S c10230, CI0234 C10266, C10272, CIO1 94 2 , 4 , 5 7 S S R S R C10192, (30215, C10218, C10227, 4 ,5 ,6 8 S S R R R C10231, ‘210232, (30233, CI0236 CI0240 5 9 S S R R S CI0225 8 10 S R R R ‘R

=Disease reaction rated on a 0-5 scale, 30days after inoculation {53 = resistance (R) and >3 = susceptible (S)). b A s defined by the multiple correspondence analysis and as described in the text.

, . i

I

pathogen was introduced into site A from othedsites (such as By Cy Dy E, F or G).

The RFLP analysis conducted with the same probes diiferentiated Colombian isolates of X. axonopodis pv. manihotis in relation to the three ecozones from which they were collected. The highest levels of diversity were found in ECZ 1 and ECZ 2, probably as a result of the longer history of cassava cultivation in these zones (Restrepo & Verdier, 1997). The Venezuelan isolates were collected in one ecozone (ECZ 1-2), which had characteristics of ECZ 1 (subhumid tropics) and of ECZ 2 (acid-soil savannas) (Anonymous, 1987). To establish a possible relationship between DNA polymorphism and geographical origin, collecting and characterizing iso- lates from other ecozones in Venezuela, e.g. ECZ 7 (semiarid) should be undertaken.

Additionally, as the conditions of two different ecozones are found in ECZ 1-2, the range of cassava genotypes that can be introduced is broader. This amplifies the risk of pathogen dissemination through contaminated material, and the selection exerted by a larger number of genotypes will be enhanced.

Until now, existence of pathogenic races among Xurn isolates has not been demonstrated (Elango & Lozano, 1981). However; recent studies conducted on the interaction between various cassava varieties and a large set of Colombian isolates, representative of the genetic diversity encountered in Colombia, suggested that a quantitative relationship exists (Restrepo et al., unpublished data). The present study also suggested the existence of a specifìc interaction between cultivars and X a m isolates from Venezuela, but the interaction should be clarified by inoculating more cultivars. As shown by different authors, knowledge of pathogenic variability and its distribution, linked with genetic analyses, allows the identification of strains for resistance screening (Nelson et uZ., 1994; Adhikari e t al., 1995; Ardales e t al., 1996). From our studies in Venezuela, Xurn isolates

i.

* , ’ .

a E representing different pathotypes could be chodpn for screening cassava germplasm indigenous to the different regions studied. In addition, isolates could be further selected for an accurate study of the host-pathogen interaction.

Acknowledgements

We thank W.M.Roca (CIAT) and E.Alvarez for their support. We also thank the staff of the F O N N and Agropecuaria Mandioca experiment stations for their help during sampling. We thank E. de Paez (CIAT) for editing, D. Andrivon and the two anonymous reviewers for their helpful comments and suggestions. This research was supported by grants from ORSTOM and CIAT and by a doctoral fellowship awarded to S. Restrepo by ORSTOM.

References

Adhikari TB, Vera Cruz CM, Zhang Q, Nelson RJ, Skinner DZ, Mew TW, Leach JE, 1995. Genetic diversity of Xanthomonas oryzae pv. oryzae in Asia. Applied and Enviroirinerital Microbiology 61, 966-71.

Anonymous, 1987. Annual Report. Cali, Colombia: Centro htemacional de Agricultura Tropical (CIAT).

Ardales EY, Lemg H, Vera Cruz CM, Mew TW, Leach JE, Nelson RJ, 1996. Hierarchical analysis of spatial variation of the rice bacterial blight pathogen across diverse agroecosystems in the Philippines. Phytopathology 86,

Berthier Y, Verdier V, Guesdon JL, Chevrier D, Denis JB, Decoux G, Lemattre M, 1993. Characterization of Xanthomonas canzpestris pathovars by rRNA gene restriction patterns. Applied arid Environmental Microbiology 59, 851-9.

Boher B, Verdier V, 1994. Cassava bacterial blight in Africa. The state of knowledge and implications for designing control strategies. African Crop Science Journal 2, 1-5.

241-52.

O 1998 BSPP Plani Pafhology (1 998) 47, 601 -608

r

608 9

V: Verdier et al.

Nelson RJ, Baraoidan MR, Vera Cruz CM,Yap IV, Leach JE, Mew TW, Leung H, 1994. Relationship between phylogeny and pathotype for the bacterial blight pathogen of rice.

Elango F, Lozano JC, 1981. Pathogenic variability of Xanthomonas mattihotis, the causal agent of cassava bacterial blight. Fitopatologia Brasileira 6, 57-63.

Gagnevin L, Leach@, Pruvost O, 1997. Genomic variability of the Xanthoniottas pathovar mangiferaeindicae; agent of mango bacterial black spot. Applied and Environmental Microbiology 63,246-53.

Gonzales JA, 1973. Las enfermedades de la yuca. III Sentinario naciorral de Fitopatologia, Sociedad Venezolana de Fitopatologia. Boletin especial 3.

d'un agent pathogène, Xanthomonas campestris pv. mani~iotis, par l'analyse factorielle multiple. Agronomie 4, 627-40.

Hahn SK, 1979. Breeding of cassava for resistance to cassava mosaic disease and bacterial blight in Africa. In: Maraite H, Meyer YA, eds. Diseases of Tropi&l Food Crops. Louvain-la Neuve, Belgium: Université Catholique de Louvain, 211-19.

disease. Plant Disease 70, 1089-93.

Colombia: epidemiology and control. Phytopathology 64,

Grousson F, Pages J, Boher B, 1990. Etude de la variabilité

Lozano JC, 1986. Cassava bacterial blight: a manageable

Lozano JC, Sequeira L, 1974. Bacterial blight of cassava in

83-8.

cassava, cause of bacterial blight and bacterial necrosis. In: Swings JG, Civerolo EL, eds. Xanthomonas. London, UK: Chapman and Hall Ltd, 18-24.

Maraite H, Weyns J, Yimkwan O, Lipemba P, Perreaux D, 1981. Physiological and pathogenic variations in Xattthomonas campestris-pv. mattihotis. In: Lozano JC,

. Gwin P, eds. Proceedings of the Fifth International Conference on Plant Pathogenic Bacteria. Cali, Colombia:

Maraite H, 1993. Xanthomonas campestris pathovars on

CUT, 358-68. Nei M, Tajima F, 1981. DNA polymorphism detectable by

restriction endonucleasck Genetics 97,145-63.

Applied and Environmental Microbiology 60,3275-83. Restrepo S, Verdier V, 1997. Geographical differentiation of

the population of Xanthomonas axonopodis pv. manihotis in Colombia. Applied and Environmental Microbiology 63, 4427-34.

Rogers DJ, Appan SG, 1973. Maitihot manihotoides (Euphorbiaceae). (Flora Neotropica monograph no. 13). New York, NY: Hafner Press.

Trujill:, GE,'Subero LJ, Luciani J, 1980. Aiiublo bacteriano (Xanthomo?2as manihotis) en la zona central del pais. II Jornadas Agronomicas de la Facultad de Agronomia, Universidad Central Venezolana, Maracay (Summary).

Xanthonzorzas campestris pathovars by SDS-PAGE of protein. Journal of General Microbiology 137,1677-87.

Vera Cruz CM, Ardales E, Skinner RJ, Talag J, Nelson RJ, Louws FJ, Leung H, Mew TW, Leach JE, 1996. Measurement of haplotypic variation in Xanthomonas oryzae pv. oryzae withìì a single field by rep-PCR and RFLP analyses. Phytopathology 86, 1352-59.

Verdier V, 1988. Evaluation de la Variabilité de Xanthomonas campestris pv. manihotis, Agent Causal de la Bactériose Vmcuhre du Manioc (Manihot esculenta, Cran&). Paris, France: Université de Paris Sud-Orsay, PhD thesis.

Verdier V, Boher B, Maraite H, Geiger JI', 1994. Pathological and molecular characterization of Xanthomonas campestris strains causing diseases of cassava (Manihot esculenta). Applied and Environmental Microbiology 60,4478-86.

Vauterin L, Swings J, Kersters K, 1991. Grouping of

Verdier V, Dongo P, Boher B, 1993. Assessment of genetic diversity among strains of Xanthotnonas campestris pv. manihotis. Journal of General Microbiology 139,

I i I

2591-60 1. ?

I Ji i k ; I '

1 *

O 1998 BSPP PlantPathology(1998) 47,601-608

..-

Volume 47, Number 5, October 1998

Plant Pathology An International Journal edited by the

(@J B ritish Society for Plant Pathology BSPP AL

Senior Editor R. Johnson

I

b

w - U c O

a- t ?T s U

Edinburgh, UK 9-1 6 August

1998

I