Characterization of monilia disease caused by Monilinia linhartiana on quince in southern Spain

Authors


E-mail: trapero@uco.es

Abstract

This study elucidates the aetiology and epidemiology of monilia disease of quince caused by the fungus Monilinia linhartiana in Spain. Disease incidence and the dynamics of apothecial development and ascospore discharge were quantified and the pathogen was characterized using morphological and molecular methods. The pathogen did not produce conidia or apothecia on agar media but produced conidia on leaves showing symptoms and apothecia on mummified young quince fruit. Monilinia linhartiana was not pathogenic on ripe quince fruit but was readily isolated from developing, mummified fruit (pseudosclerotia). Phylogenetic analysis based on 5·8S-ITS region sequences placed M. linhartiana in the Disjuntoriae section of Monilinia species infecting rosaceous hosts. Studies during 2004–2008 in four commercial orchards in southern Spain determined two major infection periods for the disease. The first coincided with the unfolding of the first leaves and resulted in leaf blotch and shoot blight. The second coincided with flowering and led to mummification of developing young fruit. Foliar infection was apparently initiated by airborne ascospores produced on pseudosclerotia that overwintered on the soil surface, while flower infection was probably initiated by conidia produced on leaf lesions. Incidence of diseased shoots ranged from 1 to 91% and was correlated with calculated inoculum potential, based on the density and maturity of apothecia formed on pseudosclerotia. This epidemiological study has made it possible to characterize the life cycle of monilia disease on quince in southern Spain, which will help the development of new control strategies.

Introduction

Quince (Cydonia oblonga) is a traditional crop in Spain where the annual production is 20 000 tonnes; production is concentrated in the Subbética district of the Andalucía region in southern Spain. Quince has been cultivated in Andalucía since the 1st century ad (Columela, 1st Century) but its agronomic features have seldom been studied in this region. Quince is of enormous cultural significance in the Subbética district, where it is an important component of the diet and grows close to olive orchards (Cabello, 2008).

The most important disease of quince in southern Spain and one of the most destructive diseases of quince worldwide, is monilia disease caused by the fungus Monilinia linhartiana (anamorph: Monilia cydoniae) (Batra, 1991). This pathogen often destroys the entire quince crop (Moral et al., 2008). However, there are no estimates of the incidence of this disease in southern Spain because monilia disease is often confused with other diseases or disorders, such as frost damage, physiological fruit drop or black rot caused by Botryosphaeria obtusa (Moral et al., 2007; Cabello, 2008).

Species in the genus Monilinia have been divided into two sections: Disjunctoriae and Junctoriae. These sections were defined on the basis of conidial morphology, fungal life cycle and host specialization, and have been accepted by most researchers (Batra, 1991; Holst-Jensen et al., 1997). The difference between the Disjunctoriae and Junctoriae sections has been confirmed by studies of internal transcribed spacers (ITS1 and ITS2) and the 5·8 subunit of the nuclear ribosomal DNA (Holst-Jensen et al., 1997; Harada et al., 2005; Takahashi et al., 2005). Monilinia linhartiana is in the Disjunctoriae section but there have been few studies on its identification and biology (Batra, 1991). A sequence of the ITS1-5·8S-ITS2 region of M. linhartiana would increase understanding of its relationship to other Monilinia species and the phylogenetics of this genus (Holst-Jensen et al., 1997).

The life cycle of M. linhartiana on quince trees may be similar to that of M. vaccinii-corymbosi on blueberry, with two separate cycles of infection: foliar infection initiated by ascospores produced on mummified fruit (pseudosclerotia) that overwinter on the soil surface, and flower infection initiated by conidia produced on the diseased leaves (Batra, 1991). However, the life cycle of M. linhartiana on quince is not well understood and has never been studied in Spain.

Disease management requires information about inoculum dynamics and the timing of symptom development; such information helps orchard managers to time fungicide treatments and cultural practices (Campbell & Madden, 1990; Michailides et al., 2007). The dynamics of apothecial development and ascospore production by M. linhartiana in relation to infection are unknown. The management of M. linhartiana is based on control of primary infections caused by ascospores, as is the case with management of other pathogens that cause monocyclic diseases or epidemics with few secondary cycles (Campbell & Madden, 1990).

The overall objective of this study was to increase understanding of the aetiology and epidemiology of monilia disease of quince. The specific objectives were: (i) to characterize the pathogen using morphological and molecular methods, (ii) to quantify pathogen development (apothecial development and ascospore discharge), and (iii) to quantify disease incidence in crops. The critical time for control of M. linhartiana on quince (in terms of host phenology) was also determined. A preliminary report of this study has been published (Moral et al., 2008).

Materials and methods

Disease incidence

The occurrence of monilia disease in the Subbética district of southern Spain was evaluated in quince orchards from 2004 to 2008. In 2004, disease was assessed in 10 orchards distributed across all the main areas where quince is cultivated. No fungicide sprays were applied in any of these orchards. From 2005 to 2008, disease occurrence was assessed in four orchards that were selected from the original 10 as representative of the region: El Algar, La Mendaña, El Palancar and Zagrilla Baja. In May of each year, 10 trees (replications) of cv. Común in each field were selected and 250 to 500 shoots per tree were assessed for disease incidence, calculated as the percentage of shoots showing typical disease symptoms (leaf blotch, shoot blight, or mummification of developing fruit). In 2004, the data for percentages of shoots affected by M. linhartiana were analysed by a one-way analysis of variance (anova) to determine whether disease incidence differed among the 10 orchards. For the four selected orchards that were assessed from 2004 to 2008, a combined anova was used to determine whether disease incidence was affected by orchard, year or the interaction between orchard and year (Gomez & Gomez, 1984). Before analyses, percentage data were converted into proportions that were then arcsine transformed to homogenize the variances. When anovas indicated that differences between orchard or year were significant, transformed mean values were compared by the Fisher’s Protected Least Significant Difference (LSD) test at P = 0·05. Data from all experiments were analysed using statistix 9 (Analytical Software).

Morphology of M. linhartiana

In 2005 and 2006, 12 isolates of M. linhartiana were collected from naturally infected leaves on quince trees in two orchards (Zagrilla Baja and El Palancar) (Table 1). Conidia from infected leaves were placed in 9 cm diameter Petri dishes containing potato dextrose agar (PDA; Difco Laboratories) acidified with 2·5 mL L−1 of a 25% lactic acid solution. Petri dishes were incubated at 23 ± 2°C with a 12 h photoperiod of fluorescent light (350 μmol m−2 s−1) for 1 week. Pure cultures of M. linhartiana were obtained by transferring hyphal tips from the margins of colonies growing on acidified PDA to other Petri dishes. The cultures were maintained in the fungal collection of the Department of Agronomy at the University of Córdoba, Córdoba, Spain. Three of the isolates (MON-06, MON-07 and MON-08) were also cultured on PDA where they developed microconidia after 15 days. Length, width, shape and colour of 50 conidia per isolate were determined using a compound microscope (Nikon Eclipse 80i; Nikon Corp.) and nis-element d software (Nikon Instruments, Inc.). Two isolates of M. laxa and three of M. fructigena (Table 1) were obtained from A. de Cal of the National Institute for Agricultural Research (INIA, Madrid) and were used for comparison with M. linhartiana.

Table 1.   Isolates of Monilinia spp. used in this study
IsolateaSpeciesHostOriginYear
  1. aIsolates of M. laxa and M. fructigena were provided by Dr A. de Cal from diseased fruit, and isolates of M. linhartiana were collected by authors of this study from diseased leaves.

  2. bProvinces of Spain.

  3. cQuince orchards in Córdoba Province.

MON-01M. laxaPrunus armeniacaZaragozab2005
MON-02M. laxaP. armeniacaMurciab2005
MON-03M. fructigenaMalus domesticaLa Coruñab2005
MON-04M. fructigenaP. domesticaLa Coruñab2005
MON-05M. fructigenaM. domesticaZaragozab2005
MON-06M. linhartianaCydonia oblongaEl Palancarc2005
MON-07M. linhartianaC. oblongaEl Palancarc2005
MON-08M. linhartianaC. oblongaZagrilla Bajac2005
MON-09M. linhartianaC. oblongaEl Palancarc2005
MON-10M. linhartianaC. oblongaEl Palancarc2006
MON-11M. linhartianaC. oblongaEl Palancarc2006
MON-12M. linhartianaC. oblongaZagrilla Bajac2006
MON-13M. linhartianaC. oblongaZagrilla Bajac2006
MON-14M. linhartianaC. oblongaEl Palancarc2006
MON-15M. linhartianaC. oblongaEl Palancarc2006
MON-16M. linhartianaC. oblongaEl Palancarc2006
MON-17M. linhartianaC. oblongaEl Palancarc2006

Apothecia of M. linhartiana were obtained from mummified quince fruit (pseudosclerotia) that had overwintered on the soil surface. In total 60 pseudosclerotia with apothecia were collected from two orchards (Zagrilla Baja and El Palancar) and placed in plastic moisture chambers (22 × 16 × 10 cm) at 10°C with 100% relative humidity (RH). In the laboratory, the apothecia were classified using a 1 to 5 scale (Cabello, 2008), where 1 = apothecium cup not formed; 2 = initial formation of the cup; 3 = cup concave with the edge enlarged; 4 = cup completely developed with a smooth edge; and 5 = senescing apothecium with few or no ascospores. In this classification, stages 3 and 4 indicate apothecia with mature ascospores. Forty apothecia in stages 3 and 4 were dissected from the pseudosclerotia, and the diameter of the apothecial cup and the width and length of the apothecial stipe were measured. Length and width of 60 asci and 600 ascospores from 20 pseudosclerotia were measured.

Characteristics of M. linhartiana on agar culture

Production of conidia in culture media and other characteristics of growth in culture were evaluated by growing isolates of M. laxa (MON-01 and MON-02), M. fructigena (MON-04 and MON-05) and M. linhartiana (MON-06 and MON-07) on seven media. Five of the media were prepared according to Dhingra & Sinclair (1985): PDA, saline PDA, oat meal agar (OMA), V-8 juice agar (V8) and peach agar (PA). The other two media were PDA plus acetone (Pascual et al., 1990) and V-8 juice agar with a cellulose membrane (Brewster et al., 1995). A 5 mm diameter plug from the growing margin of a 5-day-old colony was placed in the centre of a 9 cm diameter Petri dish containing one of the seven media. The Petri dishes were incubated at 23 ± 2°C with a 12 h photoperiod. Colony diameter was measured after 7 and 14 days, and data were converted to radial growth (mm per day). After 14 days, the shape of the colony and the presence of zonation, microconidia and conidia were determined. Three replicate Petri dishes were used for each combination of isolate and medium. The test was done twice and the similar data from the two experiments (= 0·465) were averaged. anova was used to determine whether radial growth was affected by treatment variables isolate, medium, or the interaction between isolate and medium. Mean values were compared using the Fisher’s Protected LSD test at P = 0·05.

Molecular characterization of M. linhartiana

Fungal mycelium for DNA extraction was obtained from four isolates of M. linhartiana (MON-07 to MON-10) grown on PDA for 10 days. Mycelium was collected from the surface of the PDA. Genomic DNA was extracted using the FastDNA Kit (BIO 101, Inc.) following the manufacturer’s instructions. The nuclear ribosomal DNA repeat including ITS1, 5·8S rRNA and ITS2 and portions of the genes encoding both small and large subunit rRNAs were amplified using primers ITS1 and ITS4 according to White et al. (1990). The PCR products were purified using an Ultra Clean PCR Clean-Up kit (Mobio Laboratories, Inc.). The resulting amplicons were sequenced in both directions using an automated sequencer (ABI Prism 3130XL; Applied Biosystems) by the Central Sequencing Service of the University of Córdoba in Spain. The nucleotide sequences were read and edited with FinchTV 1·4·0 (Geospiza Inc.; http://www.geospiza.com/finchtv). All sequences were checked manually and nucleotide arrangements at ambiguous positions were clarified using direction sequences of both primers. Additionally, sequences of the strains isolated in this study were compared with those Monilinia sequences available in GenBank: M. amelanchieris (Z73769), M. aucupariae (Z73771), M. azaleae (AB182266), M. baccarum (Z73773), M. cassiopes (Z73776), M. fructicola (Z73777), M. fructigena (Z73781), M. gaylussaciae (Z73782), M. jezoensis (AB182265), M. johnsonii (Z73783), M. laxa (Z73786), M. mali (AB125615), M.  megalospora (Z73788), M. mumecola (AB125618), M. oxycocci (Z73789), M. padi (Z73791), M. polycodii (Z73782), M. polystroma (Y17876), M. seaverii (Z73793), M. urnula (Z73794), M. vaccinii-corymbosi (Z73796), Ovulinia azaleae (Z73797) and Sclerotinia sclerotiorum (Z73783) (Holst-Jensen et al., 1997; Harada et al., 2005; Takahashi et al., 2005). The 5·8S-ITS region sequences were aligned using the computer program mega v4·0 (Tamura et al., 2007). To obtain appropriate substitution models for Maximum Likelihood (ML) analysis (Felsenstein, 1981), the alignment was analysed with jModelTest 0·1·1 using the Akaike information criterion (AIC) (Posada, 2008). Searches for the best ML tree and 1000 bootstrap replications were done with the fast likelihood software phyml 3·0, using identical settings.

Pathogenicity tests on ripe fruit and infection of developing fruit

Completely gold-yellow ripe fruit of cv. Común collected during autumn were inoculated with two isolates of M. linhartiana (MON-08 and MON-09), two isolates of M. laxa (MON-01 and MON-02) or two isolates of M. fructigena (MON-04 and MON-05) (Table 1). Before inoculation, the fruit were washed and then surface sterilized by immersing them in a 10% solution of commercial NaClO (50 g Cl L−1). Surface sterilized fruit were inoculated by placing a 9 mm diameter PDA plug of mycelium of each isolate on them. Control fruit were treated with a sterile PDA plug. Control and inoculated fruit were placed in plastic moisture chambers (41 × 26 × 20 cm) at 100% RH and 22 ± 2°C under fluorescent lights (350 μmol m−2 s−1). There were five fruit per isolate and the experiment was repeated three times. Fruit were examined weekly for disease symptoms, and lesions that formed around the agar plugs were measured after 1 month. The effects of species and isolate on lesion diameter were determined by an anova and means were compared by the Fisher’s Protected LSD test at P = 0·05.

To study whether the fungus infects the developing fruit, a total of 10 aborted fruit (<10 mm diameter) with initial symptoms of mummification were collected from each of two quince orchards during the spring. Mummified fruit were washed, surface sterilized and air dried on a laboratory bench and small pieces (3 × 3 × 3 mm) were placed on acidified PDA in Petri dishes (five pieces per fruit). The Petri dishes were incubated at 23 ± 2°C with a 12 h photoperiod. The percentage of fruit pieces with M. linhartiana was assessed 7 days after the pieces had been placed on PDA. The experiment was repeated three times, i.e. 300 pieces of fruit from 60 mummified fruit were examined.

Effect of temperature on mycelial growth and apothecial development

Petri dishes containing potato sucrose agar (PSA; Harada et al., 2005) were inoculated with 5 mm diameter mycelial plugs (one plug per plate) taken from the edge of actively growing 10-day-old cultures of M. linhartiana. The dishes were independently incubated in the dark at 5, 10, 15, 20, 25, 30 and 35°C. Colony diameter was measured daily. There were five replicate dishes per isolate (MON-08 and MON-09) and temperature, and the experiment was done twice. The results of the two trials were similar, and thus the data were averaged across trials. For each isolate, regressions of radial growth (mm per day) against temperature were done. The Student’s t-test was used to compare optimal temperatures (ºC) for radial growth, maximum daily radial growth (mm) and area under the growth curve of the isolate.

The effect of temperature on apothecial development was also determined using 48 pseudosclerotia with apothecial initials (phenological stage 1 and 2 of the maturity scale) that were collected from the Zagrilla Baja quince orchard. The mummified fruit were divided into four replicate groups (12 fruit per group) so that the groups did not differ in number and maturity of apothecia. Each group was placed in a separate plastic moisture chamber (22 × 16 × 10 cm, 100% RH) at 6, 12, 18 or 24°C under fluorescent lights (350 μmol m−2 s−1). The phenological stage of each apothecium was assessed daily on a 1 to 5 scale. The experiment was done twice with similar results and the data were combined for analysis. The relationship between incubation temperature and apothecial development rate was evaluated by linear regression using an equation forced through the origin. In all linear regression analyses, the following were determined: the significance of the regression, the coefficient of determination (R2), the coefficient of determination adjusted for degrees of freedom (inline image) and the pattern of residuals. The slopes of the four regression lines generated by different incubation temperatures were compared after doing the Bartlett test for equality of variance by using the least significant difference test at = 0·05.

Apothecial development and ascospore dynamics in orchards

Apothecial development and ascospore dynamics were quantified in two quince orchards (El Palancar and Zagrilla Baja) from budbreak to fruit set (February to April) of 2006 and 2007 and in one quince orchard (El Palancar) in the same period of 2008. Five trees were selected in each orchard, and 4 m2 of soil surrounding each tree were examined every 7–10 days. The following variables were quantified: the number of mummified fruit with at least one apothecium, the number of apothecia per mummified fruit, the phenological stage of the apothecia (1–4; apothecia at stage 5 were not counted because they release few or no ascospores), the number of mummified fruit and the number of apothecia per square metre. Inoculum potential (IP) for each tree was calculated as:

image

where ni = number of apothecia, ei = phenological stage (<5) of each apothecium and = sampled area (4 m2).

Airborne ascospores of M. linhartiana were quantified with a spore sampler (Burkard Manufacturing Co. Ltd.) that was placed 1·5 m above ground level in one quince orchard (El Palancar) for 2 years (2006 and 2007). The trap was operated from February to April and the sticky tape that collected the spores was replaced weekly. Once removed from the field, the tape was cut into daily segments (48 mm), which were examined every other day. The entire area of each daily segment was scanned with a microscope (×200 magnification) and the number of M. linhartiana ascospores per hour (per 2 mm of tape) was counted.

Relationships between inoculum, disease and weather variables

Pearson correlation and linear regression analyses were used to evaluate the relationships between average and maximum values of inoculum variables (mummified fruit per square metre, apothecia per mummified fruit, apothecia per square metre, inoculum potential) and the percentage of diseased shoots in two orchards (El Palancar and Zagrilla Baja).

Multiple linear regression analysis was used to characterize relationships between daily concentration of airborne ascospores and the daily average of different weather variables (rainfall, temperature, relative humidity and wind speed) for one orchard in two years. Weather variables were measured at a weather station located 12 km from the experimental orchard. Influence of weather on days xn (= 1–7) on ascospore concentration on day x was also studied. The best model was selected using a stepwise linear regression method (Statistix 9). Thus, the hourly distribution of ascospore counts from the sampling period was studied. The cumulative number of ascospores collected over time was fitted using the logistic model [Ln (Yi)/(Ymax− Yi)], where Yi is the number of ascospores collected per day and Ymax is the maximum number of ascospores collected in 1 day. Logit-transformed ascospores data were regressed against days (Campbell & Madden, 1990). Finally, Pearson’s correlation test and linear regression were used to assess relationships between the mean disease incidence (%) in each year and the rainfall during the first 25 days of March.

Results

Disease incidence

All quince trees examined during the study (2004–2008) had symptoms of monilia disease. The pathogen caused two types of symptom in orchards, leaf blotch and shoot blight (Fig. 1a,b) and mummification of developing young fruit (Fig. 1c,d). The diseased leaves infected by M. linhartiana smelt aromatic, mimicking the smell of quince flowers. Incidence of diseased shoots ranged from 0·9% (in the Zagrilla Baja orchard in 2008) to 91·5% (in the El Palancar orchard in 2006) (Fig. 2). Sampling year, orchard, and the interaction between year and orchard significantly influenced (P < 0·001 in all cases) the percentage of shoots infected by the pathogen. Because the interaction between year and orchard was significant, orchards were compared within each year. Disease incidence across the orchards was greatest in 2006, when 75% of the shoots showed symptoms, whereas the average disease incidence was only about 6% in 2005 and 2008 (Fig. 2).

Figure 1.

 Symptoms of monilia disease of quince caused by Monilinia linhartiana: (a) shoot blight, (b) leaf blotch with abundant production of M. linhartiana conidia mainly along the leaf veins, (c) infected young fruit (arrow) together with a healthy fruit, (d) apothecia produced from mummified fruit.

Figure 2.

 Incidence (%) of quince shoots with symptoms of Monilinia linhartiana infection in different orchards and years. Bars represent the average of 10 trees. Vertical lines represent the standard errors of the means.

Morphology of M. linhartiana

The pathogen produced chains of up to 30 conidia with disjunctors on diseased leaves during the spring. Conidia were limoniform and hyaline and had smooth walls. There were small differences in the size of conidia between isolates (data not shown), which ranged from 6·4–17·2 × 5·1–13·6 μm. The mean ± SD conidium length and width were 10·5 ± 1·7 × 8·5 ± 1·1 μm. On PDA, the pathogen developed microconidia, which were subglobose to pyriform and hyaline and measured 3·5 ± 0·4 × 3·2 ± 1·1 μm. Monilinia linhartiana developed solitary or multiple apothecia (2–12) on mummified fruit that had overwintered on the soil surface or were slightly buried in the soil. Mature apothecia were brown to pale brown and cupulate or planoconvex. Apothecia at phenological stage 3 (cup concave with the edge enlarged) had a stipe that was 30·7 ± 15·7 mm long × 0·8 ± 1·1 mm wide and a cup diameter of 20·7 ± 5·8 mm. Apothecial stage 4 (cup completely developed with a smooth edge) had a stipe that was 56·1 ± 34·9 × 0·7 ± 0·3 mm and cup diameter of 30·6 ± 9·4 mm. Asci of M. linhartiana were cylindrical to claviform and measured 98·6 ± 24·7 × 7·2 ± 1·9 μm. Ascospores were ellipsoid, hyaline and smooth-walled. Ascospores measured 10·1 ± 1·0 × 5·6 ± 1·1 μm. Paraphyses were cylindrical to claviform and slightly longer (1–2 μm) than asci.

Characteristics of M. linhartiana in agar culture

Isolates of Monilinia species grew on all culture media. The species, medium, and interaction between species and medium significantly (P < 0·001) affected mycelial growth (data not shown). In general, M. linhartiana grew better in media with potato extracts than in other media. Colonies of M. linhartiana were white, almost grey in the centre, fluffy and had complete margins after 10 days on PDA at 23 ± 2°C. Monilinia fructigena colonies were brownish grey and felt-like, with dense mycelium and complete margins on PDA. In contrast, M. laxa colonies were yellowish grey and felt-like, with lobed margins. Production of conidia and microconidia differed greatly between Monilinia species. Monilinia fructigena formed conidia on all media except PDA but M. laxa and M. linhartiana did not form conidia on any media. Monilinia linhartiana developed microconidia on all media but M. fructigena and M. laxa did not form microconidia.

Molecular characterization of M. linhartiana

The 5·8S-ITS region of M. linhartiana contained 480 bp. The alignment of the ITS sequences showed that ITS1 was more variable than ITS2, i.e. ITS1 had 42 nucleotide sites while ITS2 had 38 sites that varied among the isolates studied. The AIC criterion as implemented in jModelTest suggested GTR+G as the most appropriate substitution model. The log likelihood of the best ML tree found was −1862. The 25 representative isolates of Monilinia spp. formed three major groups (I, II and III) supported by bootstrap values >65% (Fig. 3). Group I (bootstrap value 81%) included all of the Monilinia species that belong to the Junctoriae section and that infect rosaceous hosts; M. johnsonii, Ovulinia azaleae and S. sclerotiorum were also in group I. Group II (bootstrap value 67%) included those Monilinia species that belong to the Disjunctoriae section and that infect ericaceous and rosaceous hosts. Group III (bootstrap value 100%) included only those Monilinia species that belong to the Disjunctoriae section and that infect the Ericaceae. Monilinia linhartiana belonged to group II and exhibited similar sequences to, and was clustered close to, the reference sequence of M. amelanchieris. The 5·8S-ITS sequences of M. linhartiana and M. amelanchieris differed in only one nucleotide. The sequence of isolate MON-09 contained one nucleotide more (thymine) at position 122 than the rest of the M. linhartiana isolates studied. Sequences of the ITS region from two representative M. linhartiana isolates (MON-09 and MON-10) were deposited in GenBank (Accessions HM581953 and HM581954, respectively).

Figure 3.

 Maximum likelihood phylogram inferred with phyml from the 5·8S-ITS sequences of Monilinia spp., Monilia mumecola, Ovulinia azaleae and Sclerotinia sclerotiorum, under a GTR+G nucleotide substitution model considering Monilinia megalospora as root. Branch lengths are scaled in terms of the expected number of substitutions per site. Numbers above branches represent ML bootstrap values (only values greater than 50% are shown).

Pathogenicity tests on ripe fruit and infection of developing fruit

The species M. linhartiana was not pathogenic on ripe fruit. Isolates of M. laxa and M. fructigena were pathogenic on ripe fruit of quince cv. Común. The first symptoms (soft rot) were observed 12 days after fruit were inoculated. Uninoculated fruit did not develop any disease symptoms. Both species developed abundant conidia on the fruit surface. Monilinia laxa and M. fructigena did not differ in disease severity, i.e. lesion diameters on inoculated fruit were similar (P = 0·456) for both species.

In quince trees infected with M. linhartiana, first symptoms on developing fruit appeared soon after fruit set. Infected fruit stopped growing and then stromatization of mycelia and mummification occurred as hard and black areas under the fruit peel. Monilinia linhartiana was identified from all 60 fruit with symptoms collected from the quince orchards. The fungus, which was identified by its characteristic mycelium, grew from 96% of pieces of mummified fruit that were placed on acidified PDA. Most affected fruit fell from the tree to the orchard floor during the spring and early summer. However, some fruit remained attached to the shoots for at least 1 year.

Effect of temperature on mycelial growth and apothecial development

Both isolates of M. linhartiana grew on PSA at temperatures ranging from 5 to 35°C (Fig. 4). Growth rate (mm per day) was plotted against temperature and the best relationship was described by a third degree polynomial (Y = aT+ bT+ cT + d; Fig. 4). The four model parameters (a, b, c, d) were significant (P < 0·05) for both isolates. The optimal temperature for mycelial growth was 22·5°C for MON-08 and 22·3°C for MON-09. Optimal temperatures and area under the curve of mycelial growth did not differ between the two isolates (Student’s t-test; P > 0·05) but the maximum growth rate was greater for MON-08 than for MON-09 (Student’s t-test; P = 0·011).

Figure 4.

 Effect of temperature on mycelial growth of Monilinia linhartiana on potato sucrose agar. Each value is the mean of three replications.

Apothecia of M. linhartiana developed on mummified quince fruit (Fig. 1d) at temperatures ranging from 6 to 24°C (Table 2). Rate of development of apothecia increased linearly with increasing incubation temperature (Table 2). However, longevity was greater at low than at high temperatures (data not shown).

Table 2.   Effect of temperature on development of apothecia of Monilinia linhartiana on quince fruita
Temperature (ºC)R2 bRa2 bP valuebSlope (= a)b
  1. aMummified quince fruit with apothecial initials were incubated at four constant temperatures in moisture chambers and the phenological stage of apothecia [from 1 (apothecium cup not formed) to 5 (senescing apothecium with few or no ascospores) (Cabello, 2008)] was assessed periodically.

  2. bParameters of regressions for apothecial development against time [phenological stage of apothecia (0–5) = × t (= time in days)].

  3. cMeans (= 24) followed by the same letter are not significantly different according to least significant difference test at = 0·05.

60·900·89<0·0010·092cc
120·960·95<0·0010·131b
180·900·89<0·0010·167b
240·980·93<0·0010·263a

Apothecial development and ascospore dynamics in orchards

Monilinia linhartiana apothecia developed on aborted and mummified fruit that remained on the orchard floor from February to April. These fruit were buried, partly buried or on the soil surface. Fallen quince leaves and weeds usually protected the apothecia from direct sunlight and desiccation. In 2006 and 2007, the first apothecia were observed at the beginning of March, coinciding with quince budbreak. In 2008, the first apothecia were observed on 21 February, again coinciding with budbreak. The number of apothecia per square metre ranged from 5·3 to 28·7 and differed between years and orchards. Inoculum potential also differed between years and orchards (Table 3).

Table 3.   Inoculum variables and incidence of quince shoots with symptoms of Monilinia linhartiana infection in different years and different orchards
OrchardYearMummified fruit m−2Apothecia/mummified fruitApothecia m−2MIPabIncidenceb (%)
  1. aMaximum inoculum potential (MIP) was the maximum annual value of inoculum potential (IP) calculated as IP = (∑ni × ei)/S, where ni = number of apothecia, ei = phenological stage (<5) of each apothecium and S = sampled area (4 m2).

  2. bDisease incidence was correlated with MIP: incidence (%) = 1·58 × MIP [P = 0·002; R2 = 0·922; inline image = 0·902 (Campbell & Madden, 1990)].

Zagrilla Baja20064·73·114·635·360·2
El Palancar20066·72·718·047·091·5
Zagrilla Baja20079·63·028·727·721·9
El Palancar20076·33·119·221·520·5
El Palancar20082·22·45·310·810·5

Because of their characteristic morphology (shape, size and colour), M. linhartiana ascospores were easily identified on the spore sampler tape. Empty asci or asci with two to eight ascospores were also sometimes present on the tape. Spores were collected from 3 March to 6 April in 2006, and from 5 March to 23 April in 2007; spore concentration in the air was much greater in 2007 than in 2006 (Fig. 5). In 2006, ascospore concentration was maximal in mid- to late March, which coincided with budbreak, and then decreased gradually until late April, when no spores were detected.

Figure 5.

 Numbers of Monilinia linhartiana ascospores collected per cubic metre per day, mean relative humidity, mean temperature and rainfall in a quince orchard in southern Spain. Ascospore density in the air was measured with a Burkard volumetric spore trap located 1·5 m above soil surface in 2006 (a) and 2007 (b). Phenological stages of quince from B (bud swelling) to H (fruit setting) are according to Martínez-Valero et al. (2001).

Relationship between inoculum, disease and weather variables

The incidence of monilia disease (the percentage of shoots with symptoms of M. linhartiana infection) increased linearly with maximum inoculum potential (P = 0·002; R2 = 0·922; inline image = 0·902; Table 3). The incidence of monilia disease was not correlated (Pearson test; P > 0·05) with numbers of mummified fruit per square metre, numbers of apothecia per mummified fruit or the numbers of apothecia per square metre (Table 3).

In 2006, daily ascospore concentration in the air was only related to rainfall on the same day of capture (multiple linear regression: P = 0·007), but it was not related to rainfall on the days before ascospore capture or to the rest of the weather parameters studied (temperature, relative humidity and wind speed) on the same day or on days before ascospore capture. In 2007, ascospore concentration was not related (> 0·05) to any weather parameter (Fig. 5). The cumulative numbers of ascospores collected were described by a logistic model in both years (P < 0·001; R2 and inline image > 0·90). The fitted model was Ln (Yi/(1 – Ymax) = –25·092 + 0·315Xi in 2006 and Ln (Yi/(1 − Ymax) = −10·031 + 0·12Xi in 2007. The number of ascospores collected on a particular day was positively correlated with inoculum potential on that day (Pearson test; r = 0·707; P = 0·022) based on data from both years. This correlation improved when the number of ascospores collected on a particular day was replaced by the number of ascospores accumulated during a 72 h period that included the day on which inoculum potential was assessed and one day before and one day after assessment (Pearson test; r = 0·851; P = 0·002). The linear relationship between the number of spores collected on a particular day and inoculum potential calculated on that day was also indicated by regression (R2 = 0·918; inline image = 0·906; P < 0·001; Fig. 6). The number of ascospores collected did not have a diurnal pattern (data not shown).

Figure 6.

 Linear relationship (Y = 6·538X; R2 = 0·918; inline image = 0·906; P < 0·0001) between the density of Monilinia linhartiana ascospores in air (spores per cubic metre air per day) and Mlinhartiana inoculum potential in a quince orchard (El Palancar) in 2006 and 2007. Inoculum potential (IP) was calculated as IP = (∑ni × ei)/S, where ni = number of apothecia, ei = phenological stage (1–4) of each apothecium, and S = sampled area (20 m2).

Discussion

The importance of epidemics of monilia disease on quince in southern Spain was confirmed, with as many as 91% of the quince shoots in one orchard diseased. Incidences of monilia disease near 90% have also been reported in Italy and Turkey (Altinyay, 1972; Batra, 1991). However, in other years in the current study the pathogen caused a natural and beneficial fruit thinning and infected fewer than 7% of the shoots, suggesting that the development of monilia epidemics could depend greatly on the quantity of primary inoculum and weather conditions, which were very variable among studied years.

Apothecia were found on mummified fruit, which overwintered on the soil surface or were buried <2 cm. Morphological characteristics of the apothecia, ascospores, asci and paraphyses of isolates of M. linhartiana from Spain agreed with descriptions of previous studies in France, Italy, Portugal and Turkey (Altinyay, 1972; Tomaz & Costa, 1980; Batra, 1991). Like other Monilinia species in the Disjunctoriae section (Batra, 1991), conidia of M. cydoniae, the anamorph of M. linhartiana, were formed on affected leaves and shoots (Fig. 1b), but not on mummified young fruit or culture media. Poor sporulation in culture media of Monilinia species in the Disjunctoriae section could be related to the parasitic specialization of these pathogens (Batra, 1991).

Monilinia linhartiana was not pathogenic on ripe quince fruit, unlike M. fructigena and M. laxa, but was easily isolated from mummified developing fruit that remained hanging on trees or fell to the orchard floor. Younger fruit are less susceptible to infection by M. fructigena conidia even when fresh wounds are inoculated (Xu & Robinson, 2000). The affected young fruit of quince were probably infected by conidia via the gynoecial pathway of open flowers (Batra, 1991).

Genetic information has been published on Monilinia species that cause brown rot of pomes and stone fruits (Carbone & Kohn, 1993). In this study, the 5·8S-ITS region of M. linhartiana was sequenced for the first time. Phylogenetic analysis based on 5·8S-ITS region sequences showed that Monilinia is a polyphyletic taxon formed from three groups. However, groups I and II show relatively low bootstrap values (81% and 67%, respectively), with the species within each group having common morphological and biological characteristics (Batra, 1991; Holst-Jensen et al., 1997). Based on the current study, isolates of M. linhartiana fall into the Disjunctoriae section (group II) of Monilinia spp. and they showed greater affinity with M. amelanchieris and M. mali than with other species infecting rosaceous hosts. These results support those of Holst-Jensen et al. (1997) for the isolate (Z73783), which was previously described as M. johnsonii but seems to have been misidentified. Small differences in the 5·8S-ITS region and the temperature effect were found among the M. linhartiana isolates. These differences may be derived from recombination via sexual reproduction observed in this species.

Mycelia of M. linhartiana grew at temperatures ranging from 5 to 30°C with an optimum near 22·5°C. Tomaz & Costa (1980) reported that optimal temperature for mycelial growth of this species was 24°C. Monilinia linhartiana ascospores can germinate at temperatures ranging from 7 to 30°C (Altinyay, 1972). In general, apothecia of the pathogen developed faster at higher temperatures in the current study but the longevity of apothecia decreased as temperatures increased from 6 to 24°C. Similar results were reported for M. fructicola and M. vaccinii-corymbosi (Hong & Michailides, 1998; Wharton & Schilder, 2005). This suggests that during the cool weather of late winter and early spring, M. linhartiana inoculum might be available for long periods.

The number of apothecia of M. linhartiana on the soil surface ranged from 5·3 to 28·7 apothecia m−2, which is greater than the number (4·4 apothecia m−2) of M. fructicola apothecia that was found in peach orchards in New Zealand (Tate & Wood, 2000). This higher number of apothecia for M. linhartiana than for M. fructicola could result from the fact that M. linhartiana requires a sexual stage to complete its life cycle (Fig. 7; Cabello, 2008) whereas M. fructicola does not. Both M. fructicola ascospores and conidia can cause primary infection of blossoms, leading to blossom blight in spring (Michailides et al., 2007).

Figure 7.

 Life cycle of monilia disease of quince tree caused by Monilinia linhartiana in southern Spain.

In this study, the number of ascospores collected was related to inoculum potential (Fig. 6) but not to the percentage of shoots infected by M. linhartiana. The lack of relationship between ascospore concentration in the air and shoot infection during 2006 and 2007 was probably due to different weather conditions during leaf development (from C1 to D2 phenological stages) in the two years. In 2006, these host development stages coincided with the rainy season and shoot infection was 91%. This heavy infection left a large number of mummified fruit as inoculum source for the next year. However, in 2007 shoot infection was only 22%, probably because rain was very limited during this period. After quince leaves have developed, the risk of leaf infection is probably low, as has been shown for blueberry (Milholland, 1974; Ramsdell et al., 1975).

The life cycle of M. linhartiana in southern Spain was closely related to host phenology (Fig. 7) and was similar to the life cycles of other Monilinia spp. in the section Disjunctoriae (Lehman & Oudemans, 1997, 2000; Scherm & Savelle, 2003). Such synchrony between pathogen and host is reasonable because M. linhartiana ascospores are capable of infection only during the very short periods when young vegetative tissues are present on the host (Batra, 1991; Lehman & Oudemans, 1997; Scherm & Savelle, 2003). For M. vaccinii-corymbosi on blueberry, pathogen-host synchrony occurs because development of apothecia and budbreak require a similar cold period (chill hours) (Milholland, 1974; Lehman & Oudemans, 1997). It is possible that the same biological mechanism of synchronization occurs for M. linhartiana because apothecial development was induced when mummified fruit were incubated for 3 months at 4°C (J. Moral and A. Trapero, unpublished data).

In this study, M. linhartiana ascospores apparently infected quince leaves and produced conidia, which serve as secondary inoculum for flower infection. Foliage infected by M. linhartiana emitted an aromatic odour like the quince flowers. Blueberry foliage modified in a similar manner by M. vaccinii-corymbosi attracts flower-foraging insects, which act as vectors of conidia (Batra, 1991). This observation was already noted by Schellenberg for quince leaves infected by M. linhartiana, suggesting that insects are attracted by the odour and carried the conidia of the fungus to the stigmas of the flowers (Wormald, 1926).

The life cycle of monilia disease of quince in southern Spain established in this study (Fig. 7) should be used to define the strategy for disease control. Since primary infection is caused by ascospores from mummified fruit on the soil, coinciding with the presence of young and expanding leaves in the host, control strategies should be aimed at reducing the inoculum source (mummified fruit that overwinter on the soil surface) and protecting young leaves with fungicides. This strategy subsequently limits secondary infection (flower infection) due to conidia because it reduces the source of conidia produced on infected leaves, and the epidemic has only one secondary cycle without the ripe fruit rot phase characteristic of some other species of Monilinia.

Acknowledgements

This research was funded by the Spanish Ministry of Education and Science and Junta de Andalucía (projects AGL2004-7495 and P08-08365, both co-financed by the European Union FEDER Funds). We thank F. Luque and E. Garcia-Cuevas for their skilful technical assistance in the laboratory, A. de Cal for the isolates of M. fructigena and M. laxa and A. Rueda for his help in the quince orchards. We also thank Walter J. Kaiser, Richard C. Cobb and Bruce Jaffee for critical reviewing of this manuscript.

Ancillary