Aggressiveness of eight Venturia inaequalis isolates virulent or avirulent to the major resistance gene Rvi6 on a non-Rvi6 apple cultivar

Plant Pathology (2010) 59, 1072–1080 Doi: 10.1111/j.1365-3059.2010.02345.x

Aggressiveness of eight Venturia inaequalis isolatesvirulent or avirulent to the major resistance gene Rvi6 ona non-Rvi6 apple cultivar

V. Caffiera*, F. Didelota, B. Pumob, D. Causeurc, C. E. Dureld and L. Parisia†aINRA, UMR77, IFR149, 42 rue Georges Morel, F-49071 Beaucouze; bAgrocampus-Ouest, 2 rue Andre Le Notre, F-49045 Angers;cAgrocampus-Ouest, 65 rue de Saint Brieuc, F-35042 Rennes; and dINRA, UMR1259, IFR149, 42 rue Georges Morel, F-49071

Beaucouze; France

For sustainable management of scab-resistant apple cultivars, it is necessary to understand the role of aggressiveness in the

adaptation of Venturia inaequalis populations and particularly the costs to the organism of acquiring additional virulence.

The aims of the present study were (i) to identify the quantitative variables that are most important in determining the differ-

ences in aggressiveness among groups of V. inaequalis isolates, and (ii) to ascertain whether virulent and avirulent isolates of

V. inaequalis differ significantly in aggressiveness. The aggressiveness of eight isolates that differed in their virulence to the

major resistance gene Rvi6 was compared on the non-Rvi6 apple cv. Gala. Three components of aggressiveness, namely

lesion density, the number of spores per square centimetre of leaf area, and the number of spores per lesion, were evaluated

21 days after inoculation, and the kinetics of lesion density over time were analysed in terms of maximum lesion density,

length of latent period and rate of lesion appearance. On the second youngest but fully developed leaf at the time of inocula-

tion, maximum lesion density in the virulent group was 20% lower and the latent period 7% longer, than in the avirulent

group. However, the alternative hypothesis, namely that isolates had adapted to quantitative resistance present in cv. Gala

depending on their cultivar of origin, could not be rejected. The analysis of the kinetics of lesion density by a non-linear

mixed-effect model proved useful in the assessment of aggressiveness.

Keywords: fitness, linear mixed-effect model, Malus · domestica, non-linear mixed-effect model, scab, Vf major

resistance gene

Introduction

Selection for quantitative traits related to pathogenicity(the components of aggressiveness) largely determinesthe composition of populations of fungal plant pathogens(Pariaud et al., 2009a). The dominance of particular races(Brodny et al., 1988; Pariaud et al., 2009b) and theemergence of new populations of a pathogen (Brasier,1991; Miller et al., 1998) can be explained in terms of thegreater aggressiveness of those races or populations. Theselection of aggressive isolates can be mediated by theclimate (Milus et al., 2009) and by the host (Lehman &Shaner, 1997; Cowger & Mundt, 2002; Andrivon et al.,2007). On the other hand, low aggressiveness can slowdown the development of isolates that are virulent againsta major resistance gene (Vera Cruz et al., 2000; Huanget al., 2006).

*E-mail: [email protected]

†Present address: INRA, UERI, Domaine de Gotheron,

F-26320 St Marcel-les-Valence, France.

Published online 26 August 2010

1072

Because disease control in the Venturia inaequalis –apple tree pathosystem is based mainly on spraying fungi-cides, resistant apple cultivars have been developed toreduce the chemical input. Most of the resistant cultivarscarry the major resistance gene Rvi6 (Vf in the old nomen-clature), which has resulted in the selection of virulentisolates (Parisi et al., 1993). Approaches are now beingdeveloped that aim at a choice of more durable genes forresistance and at sustainable management of these genes,which are considered a fragile genetic resource. Boththese objectives require a better understanding of thepathogen’s capacity to adapt through changes in bothqualitative and quantitative characters. The deploymentof quantitative resistance in a breeding programme forapples (Calenge et al., 2004) requires improved ways ofcharacterizing the aggressiveness in V. inaequalis popula-tions to evaluate the durability of the resistance. Identify-ing fitness penalties – the cost to the organism ofacquiring the required virulence (virulence cost) – canalso be particularly useful in deploying the available resis-tance gene more strategically (Sapoukhina et al., 2009).The hypothesis that isolates that are virulent to the majorresistance gene Rvi6 may be less fit than avirulent isolates

No claim to original US government works

Plant Pathology ª 2010 BSPP

Aggressiveness of Venturia inaequalis 1073

on non-Rvi6 cultivars (Guerin et al., 2007) is yet to betested experimentally.

Aggressiveness can be evaluated in terms of diseaseseverity or can be dissected into the components of theinfection cycle, namely the length of the latent period,infection efficiency, lesion growth, and the rate of spor-ulation. However, measuring those components is bothtechnically difficult and system-specific, i.e. the tech-niques or the results of such measurements cannot betransposed directly from one pathosystem to another,because of differences in biological features of the sys-tems. Several components of aggressiveness of differentisolates of V. inaequalis to a susceptible apple cultivarare known to vary: infection efficiency, latent period,lesion size, and the number of spores per squarecentimetre of lesion (Lalancette et al., 1987; Parisiet al., 2004). These components were shown to besignificantly different between benomyl-sensitiveand benomyl-resistant subpopulations of V. inaequalis(Lalancette et al., 1987). Population studies require test-ing a large number of isolates (Pariaud et al., 2009b),which limits the number of aggressiveness componentsthat can be measured at one time. Better characteriza-tion of the components and their relationships is there-fore necessary.

The aims of the present study were (i) to identify thequantitative variables that are most important indetermining the differences in aggressiveness of differentisolates of V. inaequalis, and (ii) to test, on a limited num-ber of Venturia inaequalis isolates, the hypothesis thatisolates that are virulent to the major resistance gene Rvi6are less fit than avirulent isolates on a non-Rvi6 applecultivar. Because no apple cultivar can be considered asuniversally susceptible (Barbara et al., 2008), the alterna-tive hypothesis, namely that the isolates may be adaptedto other potential resistance genes present in the non-Rvi6 cultivar depending on the host from which theyoriginated, was also analysed.

Materials and methods

Plant material

Gala (X4712 from the INRA depository, France) waschosen as a non-Rvi6 cultivar. It carries at least twoQTL (quantitative trait loci) of resistance to scab(Soufflet-Freslon et al., 2008), but is often used as thesusceptible reference standard in pathogenicity studies(Bus et al., 2009). Gala was grafted on rootstockMM106 (Pirard, France) and grown in 1Æ5 L pots filledwith a potting mix comprising 60% leaf mould, 30%peat and 10% disinfected garden soil. The pots wereplaced in a glasshouse (15–25�C). Only actively grow-ing plants with uniform growth were chosen and trans-ferred to a climate chamber. The youngest fullyexpanded leaf was labelled F0, the second youngestleaf below F0 was labelled F1, and so on (the higherthe number, the older the leaf). At the time of inocula-tion, each plant had 8–13 leaves.

Plant Pathology (2010) 59, 1072–1080

Experiments

Two experiments were performed: Experiment 1 in April2005, 6 weeks after grafting, and Experiment 2 in Sep-tember 2005, 4 weeks after pruning the same set ofplants. Each plant was pruned back to retain 2–3 buds.For each isolate, nine plants were inoculated in Experi-ment 1 and eight in Experiment 2.

Isolates of Venturia inaequalis

Eight monoconidial isolates collected in 1998 or 1999from six apple cultivars and from four European coun-tries (European project Dare ‘Durable Apple Resistancein Europe’ 1998–2002, Table 1) were chosen from 39 iso-lates that had been tested on a set of eight differentialhosts (Parisi et al., 2004), that, among others, carriedthree major resistance genes, Rvi6, Rvi1 and Rvi13 (Vf,Vg and Vd in the old nomenclature, respectively). Thischoice was based only on whether the isolates were aviru-lent or virulent to the major resistance gene Rvi6,although they may have differed in their virulence toother major genes. Four isolates were virulent to Rvi6(Group V) and four were avirulent to Rvi6 (Group A).Three of the eight isolates (EU-B04, EU-D42 andEU-NL24) have been used for testing Rvi6 cultivars every2 years since the first test of pathogenicity; the same viru-lence profile has been observed each time (F. Laurens,INRA, Angers, France, personal communication). Noisolate originating from cv. Gala was chosen in order toavoid any bias due to direct adaptation of such isolates tocv. Gala compared to that of isolates originating fromother cultivars. Because the isolates originated from culti-vars that differed in the degree to which they were relatedto cv. Gala, their genetic proximity to cv. Gala was calcu-lated (Table 1) based on their known genealogies usingWright’s coefficient of relationship (Wright, 1922). Theisolates were stored on malt agar and maintained at 4�C.Conidial suspensions were obtained from cultures grownon cellophane overlaid on malt agar in Petri dishes after10–12 days of incubation at 17�C with 16 h of light andinoculated onto seedlings of cv. Golden Delicious · cv.Granny Smith to obtain sufficient inoculum for the testsof aggressiveness. This multiplication was performedunder standard conditions (Parisi et al., 1993). Scabbedleaves were harvested 10–11 days after inoculation, driedand stored at –20�C.

Inoculations

Spore suspensions were prepared from the scabbed leavesand adjusted to a concentration of 110 000–130 000 spores mL)1 (Coulter counter, Beckman,6Æ2 lm < mean diameter < 12Æ8 lm). Three droplets ofthe suspension were deposited on malt agar to measurethe germination rate, which varied from 61% to 92%depending on the isolate and the experiment. The concen-tration of viable inoculum (Ct) was calculated asthe spore concentration of the inoculum suspension

Table 1 Description of the isolates of Venturia inaequalis used in this study

Isolatea Sampling date Country of origin Cultivar of origin

Relationship

coefficientb

Avirulence ⁄ virulence to:c

Rvi6 (=Vf) Rvi1 (=Vg) Rvi13 (=Vd)

EU-B04 1998 Belgium Golden Delicious 1 ⁄ 2 A V A

EU-B27 1999 Belgium President Roulin 0 A V A

EU-GR03 1998 Greece Starking Delicious 1 ⁄ 4 A V A

EU-NL19 1998 The Netherlands Golden Delicious 1 ⁄ 2 A V A

EU-D22 1999 Germany Florina 1 ⁄ 8 V V A

EU-D42 1999 Germany Prima 1 ⁄ 8 V V A

EU-NL05 1998 The Netherlands Malus floribunda 821 0 V A V

EU-NL24 1998 The Netherlands Prima 1 ⁄ 8 V V A

A = avirulent and V = virulent.aIsolates from the European Dare project (1998–2002).bGenetic proximity between the host of origin and the test cv. Gala computed using Wright’s coefficient of relationship.cData from Calenge et al., 2004; Parisi et al., 2004; Bus et al., 2005; L. Parisi, INRA, Gotheron, France, personal communication.

1074 V. Caffier et al.

multiplied by its germination percentage. For each iso-late, 20 mL of suspension was applied with a chromatog-raphy sprayer (mainly on the upper side of the leaves) toeight or nine plants of cv. Gala enclosed in a plastic cham-ber to avoid cross-contamination of isolates during inoc-ulation. For 44 h after inoculation, the leaves were keptwet with an atomizer and the plants kept in darkness at18�C. When the leaves were dry, the plants were placed atrandomly assigned positions in the climate chamber, andincubated under 16 h of light at 18�C and relative humid-ity fluctuating between 70% and 90%. During incuba-tion the leaves remained dry, which prevented secondaryinfections.

Measurements of aggressiveness

Aggressiveness was evaluated on leaves F0 and F1, whichwere the most susceptible to scab, and developed full andwell-defined lesions within 3 weeks.

Number of lesionsSeven days after inoculation, each plant was inspected toconfirm the absence of symptoms. From the 8th to the21st day, the number of new lesions on each leaf wasrecorded daily until the 11th or 12th day and every 2 or3 days thereafter. Each lesion was marked with a perma-nent marker to avoid recounting.

Sporulation rateTwenty-one days after inoculation, each leaf was shakenin distilled water with 0Æ01% Tween 20 in order to collectthe spores in a suspension. This was done for three plantsper isolate in Experiment 1 and for eight plants per isolatein Experiment 2. After diluting the suspension a 100-fold,the number of spores per millilitre of the suspensionwas calculated after counting the spores with a Coultercounter.

Leaf areaAt the end of the experiment, the length (l) and width (w)of leaves were measured for estimating leaf area (La)

based on the following equation established for cv. Galaunder controlled conditions (F. Didelot and L. Parisi,INRA, Angers, France, personal communication):

La ¼ 0�6997� l �w� 0�3886 ð1Þ

where 0Æ6997 and 0Æ3886 are the coefficients of linearregression (r2 = 0Æ9837, P < 0Æ001).

Plant growthAs the stage of plant growth is known to affect thedevelopment of scab through changes in ontogenicresistance (Schwabe, 1979), the number of inoculatedleaves from the youngest (F0) to the oldest (Li) andthe number of new leaves that had appeared after theinoculation (Ln) were counted for each plant 21 daysafter inoculation.

Data analysis

Twenty-one days after inoculation, three components ofaggressiveness were evaluated for leaves F0 and F1,namely lesion density (ld21), the number of spores persquare centimetre of leaf area (sps21), and the number ofspores per lesion (spl21). Lesion density (ld21) wasexpressed as the total number of lesions per square centi-metre of leaf area. As the number of conidia originallydeposited on the leaves at the time of inoculation was notknown precisely, the efficiency of infection could not beascertained, and lesion density was considered an accept-able substitute.

For each of these variables, denoted as Y hereafter, thefollowing linear mixed-effect model (Lme) was used withthe random factor ‘isolate’ nested in ‘group’ V or A (Pin-heiro & Bates, 2000):

Yegij ¼ b0 þ ae þ cg þ b1 � Ct þ b2 � La

þ b3 � Liþ b4 � Lnþ bi þ eegij

ð2Þ

where b0 is the general mean; ae is the effect of the ‘experi-ment’ factor, for e = 1, 2; cg is the effect of the ‘group’

Plant Pathology (2010) 59, 1072–1080

Aggressiveness of Venturia inaequalis 1075

factor, for g = 1, 2 (avirulent and virulent, respectively);b1, b2, b3, b4 are parameters associated to co-variables Ct,La, Li, Ln; bi is the random effect associated with the‘isolate’ factor, supposed to be normally distributedbi � N (0, r2

b); and eegij is the error associated with the jth

observation.Two models were tested for the error term eegij. The first

one assumed that eegij � N (0, r2ei), with the variance r2

ei

depending only on the ‘experiment’ and ‘isolate’ factors.The second model assumed that eegij�N (0, r2

eg), with thevariance r2

eg depending only on the ‘experiment’ and‘group’ factors. The better of the two models was selectedusing AIC (Akaike’s Information Criterion), and each co-variable was included in this model based on an ascen-dant selection using AIC (Pinheiro & Bates, 2000).Parameters were estimated by the maximum likelihoodalgorithm with constraints

Pae = 0 and c1 = 0, for the

parameters concerning the ‘experiment’ and the level‘avirulent’ of the ‘group’, respectively.

For each date of observations, the lesion density (ldt)was calculated as the cumulative number of lesions attime t per unit of the final leaf area (La). The kinetics oflesion density are the repeated measurements for eachindividual plant and can be represented by a logistic func-tion, commonly used to describe the evolution of plantdisease over time (Zadoks & Schein, 1979; Holb et al.,2005). The logistic function f(b, t) is given by the follow-ing equation:

ldt ! fðk;lat;r; tÞ ¼ k

1þ expðlat�tÞ

r

h i ð3Þ

where b is the vector (k, lat, r) of three parameters: k isthe asymptote, representing the maximum lesion density;lat is the latent period, the time t for 50% of the lesions toappear; and r is the relative rate of lesion appearance.

A non-linear mixed-effect model (Nlme) with linkfunction given by equation (3) was used (Davidian &Giltinan, 2003; Holb et al., 2005). The three parameters(k, lat, r) were linearly modelled depending on co-vari-ables Ct, La, Li, Ln, fixed factors ‘experiment’ and‘group’, and the random effect of ‘individual’. Further-more for the parameters k and lat, the random effect of‘isolates’ nested in ‘group’ factor was included in themodel. The variance r2

ei of the error term depended on the‘experiment’ and ‘isolate’ factors. The covariance matrixof the residual error was defined by a first-order autore-gressive model, which took into account the dependenceof residuals (Pinheiro & Bates, 2000). The best modelwas selected using the same criteria as those for the model(2) described above.

Correlations among aggressiveness components weretested by Spearman tests (Sprent & Smeeton, 2007) basedon the data for each plant. For each plant, maximumlesion density and length of the latent period were fittedon the logistic curve with the non-linear least squaresmethod (Bates & Watts, 1988). The Spearman tests werebased on the assumption that correlations werenot affected by the isolate and data were pooled over all

Plant Pathology (2010) 59, 1072–1080

isolates. Analyses were done separately for leaves F0 andF1 and for Experiments 1 and 2.

All statistical analyses were performed using R version2.8.0 (R Development Core team, 2008) with the ‘nlme’package (Pinheiro et al., 2008).

Results

Co-variables

Most variables were significantly affected by the co-variables linked to the growth of the plant (Table 2).The leaf area (La) had a significant effect on ld21,sps21, k and lat; the number of inoculated leaves (Li)had a significant effect on ld21 and lat; and thenumber of new leaves (Ln) had a significant effect onld21. There was no significant effect of the concen-tration of the viable inoculum (Ct) on any of thevariables.

Comparison of the two experiments

The number of leaves at the time of inoculation washigher in Experiment 1 than in Experiment 2 (12Æ5 and8Æ9, respectively), although the plants had been growingmore actively during Experiment 2 (4Æ5 new leaves and2Æ2 new leaves, respectively, on day 21). The mean leafarea (La) was similar in the two experiments. Twenty-one days after inoculation, sporulation (sps21 and spl21)was significantly lower in Experiment 2 than in Experi-ment 1, but the effect of ‘experiment’ on ld21, k, lat and rwas not significant (Table 2).

Comparison of the two groups of isolates

The eight isolates were divided into two groups, one viru-lent to Rvi6 and the other avirulent to it. For leaf F0, therewas no significant effect of ‘group’ for any of the variables(data not shown). For leaf F1, there was a significanteffect of ‘group’ on k and lat (Table 2). In group V, themaximum lesion density k was 20% less, and the latentperiod 7% longer, than group A; the differences aredescribed by two different estimated logistic curves(Fig. 1).

As there was a significant effect of ‘experiment’ onsps21 and spl21 (Table 2), variables evaluated 21 daysafter inoculation were analysed separately for eachexperiment (Table 3). There was no significant effectof ‘group’ on ld21 in either of the experiments. How-ever, there was a significant effect of ‘group’ on sps21

and spl21 in Experiment 1; in Group V, the numberof spores per unit leaf area was 30% less, and thatper lesion was 40% less, compared to Group A.There was a significant effect of interaction betweenthe isolates and the experiments (data not shown),largely due to isolate EU-NL19, which was relativelymore aggressive in Experiment 2 than in Experiment 1.In both experiments, isolate EU-NL05 was the leastaggressive isolate.

Table 2 Fixed effects of Lme modelling of three components of aggressiveness 21 days after inoculation of apple cv. Gala with eight isolates of Venturia

inaequalis pooled in two groups (avirulent or virulent to the major resistance gene Rvi6), and fixed effects of Nlme modelling of the kinetics of lesion density

(leaf F1)

Variable Model

Factor or

co-variablea Nob Value Standard Error d.f. P-valuec

ld21 (no of lesions per cm2 of leaf area) Lmed (Intercept)e 132 0Æ85 0Æ17 121 0Æ0000

La )0Æ02 0Æ00 121 0Æ0000

Li 0Æ03 0Æ01 121 0Æ0002

Ln )0Æ03 0Æ01 121 0Æ0174

sps21 (no of spores per cm2 of leaf area) Lmef (Intercept)e 75 304774 53139 65 0Æ0000

Exp1 33916 15368 65 0Æ0309

La )6082 1626 65 0Æ0004

spl21 (no of spores per lesion) Lmef (Intercept)e 75 268674 21698 65 0Æ0000

GroupV )132899 22936 65 0Æ0012

Exp1 91481 21698 65 0Æ0001

GroupV:Exp1 )96145 22936 65 0Æ0001

Kinetics of lesion density Nlmed k.(Intercept)e 661g 1Æ53 0Æ18 530 0Æ0000

k.GroupV )0Æ30 0Æ11 530 0Æ0048

k.La )0Æ02 0Æ01 530 0Æ0002

lat.(Intercept)e 13Æ82 0Æ60 530 0Æ0000

lat.GroupV 0Æ93 0Æ30 530 0Æ0022

lat.La 0Æ05 0Æ02 530 0Æ0033

lat.Li )0Æ17 0Æ04 530 0Æ0000

r.(Intercept)e 1Æ13 0Æ05 530 0Æ0000

aCo-variables: La = leaf area, Li = number of inoculated leaves, Ln = number of new leaves 21 days after inoculation.bNumber of individuals.cIn bold, P-value < 0Æ05 for Group V (virulent to Rvi6).dModelling of variance of the error term according to ‘isolate’ and ‘experiment’ factors.eIntercept represents Group A (avirulent to Rvi6).fModelling of variance of the error term according to ‘group’ and ‘experiment’ factors.gNull data were excluded from the data set.

Lesi

on d

ensi

ty (n

o of

lesi

ons

per c

m2

of le

af) EU-NL19

EU-GR03

EU-D22

EU-B04EU-B27

EU-NL24EU-D42

EU-NL05

10 15 20 25

0·0

0·5

1·0

Time (in days after inoculation)

Figure 1 Kinetics of measured lesion density on leaf F1 after

inoculation of apple cv. Gala with eight isolates of Venturia

inaequalis, pooled in two groups according to their avirulence (x) or

virulence (o) to the major resistance gene Rvi6. Solid line: logistic

curve estimated by Nlme modelling for the avirulent group; dashed

line: logistic curve estimated by Nlme modelling for the virulent

group. Experiments 1 and 2 were pooled.

1076 V. Caffier et al.

Correlations between variables

The results for F0 and F1 were similar, and only thosefor F1 are presented here (Table 4). As expected, lesiondensity 21 days after inoculation (ld21) was highly cor-related with maximum lesion density (k). These twovariables were correlated with sps21 but not with spl21.Latent period was moderately correlated with ld21

and k.

Discussion

Aggressiveness of V. inaequalis isolates was tested undercontrolled conditions on the non-Rvi6 apple Gala andanalysed in relation to their virulence or avirulence to themajor resistance gene Rvi6 or in relation to the cultivarfrom which the isolates had originated. On leaf F0, no sig-nificant differences were observed, probably because thisleaf was very susceptible to scab, as shown by rapid scabdevelopment and a latent period 1Æ5 days shorter thanthat on F1 (data not shown). On leaf F1, the group com-prising virulent isolates (group V) and that comprisingavirulent isolates (group A) differed significantly in twoparameters of the kinetics of lesion density: maximumlesion density and length of the latent period; in the viru-lent group maximum lesion density was 20% less and thelatent period 7% longer. However, the hypothesis that

Plant Pathology (2010) 59, 1072–1080

Table 3 Means and standard deviations of three components of aggressiveness 21 days after inoculating apple cv. Gala with eight isolates of Venturia

inaequalis

Experiment Isolate Group ld21 sda

Tukey

testb sps21 sd

Tukey

test spl21 sd

Tukey

test

1 EU-GR03 A 1Æ13 0Æ26 a 344525 112886 ab 298730 117721 a

EU-B04 A 0Æ98 0Æ62 a 295921 132220 ab 370837 127353 a

EU-B27 A 0Æ97 0Æ37 a 498710 80321 a 452971 203008 a

EU-D22 V 0Æ89 0Æ37 a 221334 217506 ab 223302 71715 a

EU-NL19 A 0Æ84 0Æ63 a 169787 260007 ab 439083 302873 a

EU-D42 V 0Æ65 0Æ37 ab 128213 11835 b 128237 18623 a

EU-NL24 V 0Æ57 0Æ29 ab 77326 65095 b 139792 66645 a

EU-NL05 V 0Æ21 0Æ16 b 25004 14313 b 255081 288322 a

Mean of the group A 0Æ98 0Æ47 a 327 236 146 359 a 390 405 187 739 a

V 0Æ58 0Æ30 a 112 969 77 187 b 186 603 111 326 b

2 EU-NL19 A 1Æ73 0Æ52 a 304110 72341 a 185849 49915 a

EU-D22 V 0Æ95 0Æ53 b 148811 96936 b 132425 35245 a

EU-GR03 A 0Æ90 0Æ59 b 112732 74757 b 170731 126458 a

EU-B04 A 0Æ58 0Æ41 cd 80996 35723 b 149321 96243 a

EU-NL24 V 0Æ57 0Æ39 cd 75960 48997 b 166711 94654 a

EU-B27 A 0Æ44 0Æ27 cd 65693 43226 b 173859 142369 a

EU-D42 V 0Æ35 0Æ34 cd 50991 39274 b 210369 237152 a

EU-NL05 V 0Æ03 0Æ05 d 23572 9314 b 247667 13671 a

Mean of the group A 0Æ91 0Æ45 a 140 883 56 512 a 169 940 103 746 a

V 0Æ48 0Æ33 a 74 833 48 630 a 189 293 95 180 a

A: avirulent to Rvi6; V: virulent to Rvi6.aStandard deviation.bDifferent letters indicate significant differences between isolates or between groups of isolates (P < 0Æ05).

Table 4 Spearman coefficients of correlation (rho) among components of aggressiveness for eight isolates of Venturia inaequalis inoculated on apple cv.

Gala (leaf F1)

Variable Experiment

k sps21 spl21 lat

rho P-value rho P-value rho P-value rho P-valuea

ld21 (lesion density) 1 0Æ82 0Æ0000 0Æ66 0Æ0020 0Æ28 0Æ2251 )0Æ43 0Æ0007

2 0Æ98 0Æ0000 0Æ85 0Æ0000 0Æ05 0Æ7755 )0Æ48 0Æ0022

k (maximum lesion density) 1 0Æ43 0Æ0593 0Æ24 0Æ2978 )0Æ12 0Æ3389

2 0Æ81 0Æ0000 0Æ02 0Æ9097 )0Æ43 0Æ0067

sps21 (no of spores per cm2 of leaf area) 1 0Æ86 0Æ0000 )0Æ48 0Æ0351

2 0Æ52 0Æ0009 )0Æ64 0Æ0000

spl21 (no of spores per lesion) 1 )0Æ33 0Æ1543

2 )0Æ42 0Æ0083

aBold characters indicate a significant correlation (P < 0Æ05).

Aggressiveness of Venturia inaequalis 1077

isolates may be differentially adapted to quantitativeresistance genes present in cv. Gala depending on the hostfrom which isolates originated, could not be rejected.

There was a good fit between the three-parameterlogistic function and the kinetics of lesion density. Thenon-linear mixed-effects model of kinetics gave informa-tion on two components of aggressiveness: lesion densityand length of the latent period, which are known to varyin populations of V. inaequalis (Lalancette et al., 1987).These two parameters are usually analysed separately.The non-linear mixed-effects model is noteworthybecause it compares the kinetics by taking into account

Plant Pathology (2010) 59, 1072–1080

both the parameters simultaneously. The model provedmore sensitive than Lme for the detection of significantdifferences between the two groups of isolates. Forinstance, differences between the two groups in lesiondensity were not significant when analysed by Lme, butsignificant when analysed by Nlme.

Maximum lesion density and length of the latent per-iod, as analysed by Nlme, were not dependent on theexperimental effect, whereas disease severity and the rateof sporulation in apple scab are reported to be stronglyinfluenced by the experimental factor (Lalancette et al.,1987). Furthermore, with regard to maximum lesion

1078 V. Caffier et al.

density and length of the latent period, differencesbetween the two groups were similar for Experiment 1and Experiment 2, although conditions at the time of thesecond experiment were less favourable to the disease.Greater differences in aggressiveness among isolates maybe expected to express more clearly under less favourableconditions (Pariaud et al., 2009a); however, the presentstudy did not bear out that expectation.

A good correlation was found between lesion densityand the number of spores produced per unit of leaf area21 days after inoculation, a result that is contrary to thefindings from other pathosystems, such as those involvingrust fungi (Kardin & Groth, 1989; Rime et al., 2005) orpowdery mildews (Rouse et al., 1984), in which sporula-tion reaches a plateau at higher lesion densities. The goodcorrelation observed in the present study may be due tothe fact that the number of spores produced per lesionwas not affected by lesion density, at least for the range oflesion density recorded in the present study. Because newlesions continued to appear over a long time, both oldand new lesions were present on the plant at the sametime, and, although the rate of sporulation of a lesionmay decline with age (Sache, 1997), the decline may havebeen compensated for by the younger lesions. Analysingthe kinetics of sporulation in relation to lesion densitywith a non-destructive method of measuring total conid-ial production across the entire disease span (Sache,1997; Suassuna et al., 2004; Rime et al., 2005) may makeit possible to test this explanation. A wide range of lesiondensity should be analysed for single isolates, to allow forthe possibility that the effect may vary with the genotypeof the isolate, as has been shown for bean rust (Kardin &Groth, 1989). If it is confirmed that there is a good corre-lation between lesion density and sporulation, analysis oflesion density could replace analysis of sporulation level,which is notoriously difficult to measure, but is importantfor understanding the epidemiology of apple scab.

The co-variables related to plant growth improved theestimation of aggressiveness. First, leaf area had a signifi-cant effect on all the components of aggressiveness, prob-ably because younger – and therefore smaller – leaves aremore susceptible to scab or because scab reduces leafgrowth. Whatever the reason, taking leaf area intoaccount helps in estimating the aggressiveness of thepathogen in the apple-scab pathosystem more reliably.Secondly, the number of inoculated leaves and the num-ber of new leaves produced after inoculation wereobserved to affect aggressiveness. This result fits in withthe fact that apple scab is affected by stage of growth ofthe plant, which affects its ontogenic resistance (Schwa-be, 1979). The concentration of viable spores, which ran-ged from 60 000 to 110 000 viable spores per millilitre ofinoculum suspensions in the present study, did not affectaggressiveness – an observation in line with the results ofearlier studies, which showed no effect despite a widerrange, namely 80 000–300 000 spores mL)1 (Hartmanet al., 1999; Lateur et al., 2000).

Fungal populations that are virulent against a majorresistance gene may pay a fitness penalty, as for

instance in Leptosphaeria maculans on oilseed rape forthe resistance gene Lm4 (Huang et al., 2006), but thisis not a general rule (Bahri et al., 2009). The presentstudy provides the first results for the scab-apple patho-system concerning the relative aggressiveness of isolatesvirulent or avirulent to Rvi6; the non-linear mixed-effects model of the kinetics of lesion density indicateda significant difference between the two groups of iso-lates pooled according to their virulence ⁄ avirulence toRvi6, which suggests that the breakdown of Rvi6 resis-tance resulted in reduced fitness of the virulent popula-tion. However, the least aggressive isolate also carrieda virulence factor to Rvi13. More isolates virulent toRvi13 need to be tested to determine whether thisreduced aggressiveness is due to the breakdown ofRvi6, or that of Rvi13, or of both.

As no apple cultivar can be considered universally sus-ceptible, an alternative hypothesis would be differentialadaptation of the pathogen to the QTL present in the cul-tivar chosen for testing. Adaptation to quantitative resis-tance has been shown, for instance, in Phythophthorainfestans on potato (Andrivon et al., 2007) and Mycosp-haerella graminicola on wheat (Cowger & Mundt,2002), and differential adaptation to cultivars consideredas susceptible has also been shown for apple scab (Sier-otzki et al., 1994; Barbara et al., 2008). Isolates of thetwo groups (A and V) were sampled from different culti-vars, without sampling any isolate from cv. Gala itself,but generally Rvi6 cultivars were more distant from cv.Gala than were non-Rvi6 cultivars (average coefficient ofrelationship of 0Æ09 and 0Æ31, respectively). As a conse-quence, the hypothesis that the isolates had adapted toquantitative resistance present in cv. Gala depending ontheir cultivar of origin could not be rejected. Further-more, the least aggressive isolate came from M.floribundaand not from M. · domestica, therefore the hypothesis ofa differential adaptation related to the host species(M. · domestica or M. floribunda) should also be consid-ered. Testing a larger number of non-Rvi6 cultivars andtesting more isolates from each host cultivar or specieswould strengthen the interpretation of these data.

In conclusion, the present study, based on a restrictedset of isolates from different cultivars and different coun-tries tested on cv. Gala, shows significant differences inaggressiveness between two groups of isolates, a differ-ence which may be explained either by reduced aggres-siveness of virulent populations (directly in the case of acost of acquiring virulence or indirectly in the case of achance mutation for virulence in an isolate of low aggres-siveness) or by adaptation to QTL present in the non-Rvi6 cultivar. Data on a larger number of isolates areneeded to compensate for the high diversity observed inV. inaequalis populations (Gladieux et al., 2008) anddata on a larger set of cultivars without major genes ofresistance are needed to take into account the factthat no cultivar can be considered universally susceptible(Barbara et al., 2008). The present study showedthe usefulness of a non-linear mixed-effects model ofthe kinetics of lesion density for comparison of

Plant Pathology (2010) 59, 1072–1080

Aggressiveness of Venturia inaequalis 1079

aggressiveness in different groups of V. inaequalisisolates. This model will be used in future studies ofthe apple-scab pathosystem extending to more isolatesrepresenting different populations.

Acknowledgements

We thank partners of the European project Dare(1998–2002) for sampling the isolates of Venturiainaequalis, Florian Blanchet for technical help in theexperiments, and Christian Lannou and NataliaSapoukhina for helpful comments on earlier versions ofthis paper.

References

Andrivon D, Pilet F, Montarry J et al., 2007. Adaptation of

Phytophthora infestans to partial resistance in potato: evidence

from French and Moroccan populations. Phytopathology 97,

338–43.

Bahri B, Kaltz O, Leconte M, de Vallavieille-Pope C, Enjalbert J,

2009. Tracking costs of virulence in natural populations of the

wheat pathogen, Puccinia striiformis f.sp. tritici. BMC

Evolutionary Biology 9. Http://www.biomedcentral.com/

1471-2148/9/26.

Barbara DJ, Roberts AL, Xu XM, 2008. Virulence characteristics

of apple scab (Venturia inaequalis) isolates from monoculture

and mixed orchards. Plant Pathology 57, 552–61.

Bates DM, Watts DG, 1988. Nonlinear Regression Analysis and its

Applications. New York, USA: John Wiley & Sons.

Brasier CM, 1991. Ophiostoma novo-ulmi sp. nov, causative agent

of current Dutch elm disease pandemics. Mycopathologia 115,

151–61.

Brodny U, Wahl I, Rotem J, 1988. Factors conditioning dominance

of race 276 of Puccinia coronata avenae on Avena sterilis

populations in Israel. Phytopathology 78, 135–9.

Bus VGM, Laurens FND, van de Weg WE et al., 2005. The Vh8

locus of a new gene-for-gene interaction between Venturia

inaequalis and the wild apple Malus sieversii is closely linked to

the Vh2 locus in Malus pumila R12740-7A. New Phytologist

166, 1035–49.

Bus V, Rikkerink E, Aldwinckle HS et al., 2009. A proposal for the

nomenclature of Venturia inaequalis races. Acta Horticulturae

814, 739–46.

Calenge F, Faure A, Goerre M et al., 2004. Quantitative trait loci

(QTL) analysis reveals both broad-spectrum and isolate-specific

QTL for scab resistance in an apple progeny challenged

with eight isolates of Venturia inaequalis. Phytopathology 94,

370–9.

Cowger C, Mundt CC, 2002. Aggressiveness of Mycosphaerella

graminicola isolates from susceptible and partially resistant

wheat cultivars. Phytopathology 92, 624–30.

Davidian M, Giltinan DM, 2003. Nonlinear models for repeated

measurement data: an overview and update. Journal of

Agricultural, Biological and Environmental Statistics 8,

387–419.

Gladieux P, Zhang XG, Afoufa-Bastien D, Valdebenito Sanhueza

RM, Sbaghi M, Le Cam B, 2008. On the origin and spread of

the scab disease of apple: out of central Asia. PLoS ONE 3,

e1455. doi:10.1371/journal.pone.0001455.

Plant Pathology (2010) 59, 1072–1080

Guerin F, Gladieux P, Le Cam B, 2007. Origin and colonization

history of newly virulent strains of the phytopathogenic fungus

Venturia inaequalis. Fungal Genetics and Biology 44, 284–92.

Hartman JR, Parisi L, Bautrais P, 1999. Effect of leaf wetness

duration, temperature, and conidial inoculum dose on apple scab

infections. Plant Disease 83, 531–4.

Holb IJ, Heijne B, Withagen JCM, Gall JM, Jeger MJ, 2005.

Analysis of summer epidemic progress of apple scab at different

apple production systems in the Netherlands and Hungary.

Phytopathology 95, 1001–20.

Huang YJ, Li ZQ, Evans N, Rouxel T, Fitt BDL, Balesdent MH,

2006. Fitness cost associated with loss of the AvrLm4 avirulence

function in Leptosphaeria maculans (phoma stem canker of

oilseed rape). European Journal of Plant Pathology 114,

77–89.

Kardin MK, Groth JV, 1989. Density-dependent fitness interactions

in the bean rust fungus. Phytopathology 79, 409–12.

Lalancette N, Hickey KD, Cole H, 1987. Parasitic fitness and

intrastrain diversity of benomyl-sensitive and benomyl-resistant

subpopulations of Venturia inaequalis. Phytopathology 77,

1600–6.

Lateur M, Lefrancq B, Wagemans C, 2000. Influence of scab

inoculum concentration in an apple breeding program

focused on quantitative resistance. Acta Horticulturae 538,

249–55.

Lehman JS, Shaner G, 1997. Selection of populations of

Puccinia recondita f.sp. tritici for shortened latent period on

a partially resistant wheat cultivar. Phytopathology 87, 170–6.

Miller JS, Johnson DA, Hamm PB, 1998. Aggressiveness of isolates

of Phytophthora infestans from the Columbia Basin of

Washington and Oregon. Phytopathology 88, 190–7.

Milus EA, Kristensen K, Hovmøller MS, 2009. Evidence for

increased aggressiveness in a recent widespread strain of Puccinia

striiformis f.sp. tritici causing stripe rust of wheat.

Phytopathology 99, 89–94.

Pariaud B, Ravigne V, Halkett F, Goyeau H, Carlier J, Lannou C,

2009a. Aggressiveness and its role in the adaptation of plant

pathogens. Plant Pathology 58, 409–24.

Pariaud B, Robert C, Goyeau H, Lannou C, 2009b. Aggressiveness

components and adaptation to a host cultivar in wheat leaf rust.

Phytopathology 99, 869–78.

Parisi L, Lespinasse Y, Guillaumes J, Kruger J, 1993. A new race of

Venturia inaequalis virulent to apples with resistance due to the

Vf gene. Phytopathology 83, 533–7.

Parisi L, Fouillet V, Schouten HJ et al., 2004. Variability of the

pathogenicity of Venturia inaequalis in Europe. Acta

Horticulturae 663, 107–13.

Pinheiro JC, Bates DM, 2000. Mixed-Effects Models in S and

S-PLUS. New York, USA: Springer-Verlag.

Pinheiro J, Bates D, DebRoy S, Sarkar D, R Core team, 2008.

Nlme: linear and nonlinear mixed effects models. R Package

Version 3.1-90. Http://www.R-project.org.

R Development Core team, 2008. R: A Language and

Environment for Statistical Computing. Vienna, Austria:

R Foundation for Statistical Computing. Http://www.

R-project.org.

Rime D, Robert C, Goyeau H, Lannou C, 2005. Effect of host

genotype on leaf rust (Puccinia triticina) lesion development and

urediniospore production in wheat seedlings. Plant Pathology

54, 287–98.

1080 V. Caffier et al.

Rouse DI, MacKenzie DR, Nelson RR, 1984. Density dependent

sporulation of Erysiphe graminis f.sp. tritici. Phytopathology

74, 1176–80.

Sache I, 1997. Effect of density and age of lesions on sporulation

capacity and infection efficiency in wheat leaf rust (Puccinia

recondita f.sp. tritici). Plant Pathology 46, 581–9.

Sapoukhina N, Durel CE, Le Cam B, 2009. Spatial deployment of

gene-for-gene resistance governs evolution and spread of

pathogen populations. Theoretical Ecology 2, 229–38.

Schwabe WFS, 1979. Changes in scab susceptibility of apple leaves

as influenced by age. Phytophylactica 11, 53–6.

Sierotzki H, Eggenschwiler M, Boillat O, McDermott JM, Gessler

C, 1994. Detection of variation in virulence toward susceptible

apple cultivars in natural populations of Venturia inaequalis.

Phytopathology 84, 1005–9.

Soufflet-Freslon V, Gianfranceschi L, Patocchi A, Durel CE, 2008.

Inheritance studies of apple scab resistance and identification of

Rvi14, a new major gene that acts together with other broad-

spectrum QTL. Genome 51, 657–67.

Sprent P, Smeeton NC, 2007. Applied Non-parametric Statistical

Methods. Berlin, Germany: Chapman & Hall.

Suassuna ND, Maffia LA, Mizubuti ESG, 2004. Aggressiveness

and host specificity of Brazilian isolates of Phytophthora

infestans. Plant Pathology 53, 405–13.

Vera Cruz CM, Bai J, Ona I et al., 2000. Predicting durability of a

disease resistance gene based on an assessment of the fitness loss

and epidemiological consequences of avirulence gene mutation.

Proceedings of the National Academy of Sciences, USA 97,

13500–5.

Wright S, 1922. Coefficients of inbreeding and relationship. The

American Naturalist 56, 330–8.

Zadoks JC, Schein RD, 1979. Epidemiology and Plant

Disease Management. New York, USA: Oxford University

Press.

Plant Pathology (2010) 59, 1072–1080