Greeneria uvicola causes bitter rot on Vitis vinifera (bunch grapes) and Muscadinia rotundifolia (muscadine grapes) in warm moist temperate and subtropical regions. This study investigated the phylogenetic relationship of G. uvicola representatives from Australia (67 isolates), the USA (31 isolates), India (1 isolate) and Costa Rica (1 isolate) and compared their pathogenicity and fungicide sensitivity. Differences in cultural and conidial morphology were observed between the isolates from Australia and the USA. Phylogenetic relationships were determined based on three gene regions: the ribosomal DNA (rDNA) internal transcribed spacer 1 (ITS1–5∙8S–ITS2), 28S large subunit (LSU) nuclear rDNA and β-tubulin-2. Greeneria uvicola isolates were clearly differentiated into four groups: isolates from Australia and India; USA isolates from V. vinifera; USA isolates from M. rotundifolia; and the isolate from Costa Rica. All isolates were pathogenic on V. vinifera (cv. Chardonnay) berries although those originating from M. rotundifolia were not as aggressive as isolates from V. vinifera, irrespective of geographical origin. Sensitivity to pyraclostrobin and salicylhydroxamic acid (SHAM) was studied. Despite differences in fungicide applications, hyphal growth inhibition was not significantly different for geographical location, cultivar, tissue, year of collection or different spray regimes. For the Australian and USA isolates, fungal growth inhibition was significantly greater for pyraclostrobin than for SHAM, and was significantly greater for the combined treatment than for each of the fungicides applied singly. The aetiological and epidemiological knowledge of bitter rot collected through this study will aid better prediction and management strategies of this pathogen.
Greeneria uvicola (syn. Melanconium fuligineum) causes bitter rot on Vitis vinifera (European bunch grape; Fig. 1a), Muscadinia rotundifolia (muscadine grape; Fig. 1b) and Vitis × labruscana (fox grape) in warm moist temperate and subtropical grape-growing regions throughout the world. It has been identified in Greece, Italy, Japan, Taiwan, China, India, Australia, Costa Rica, Brazil, South Africa and the USA (Critopoulos, 1961; Ridings & Clayton, 1970; McGrew, 1977; Sutton & Gibson, 1977; Abrahão et al., 1993; Wu & Chang, 1993; Yan et al., 1998; Castillo-Pando et al., 2001; Pfenning et al., 2005; Navarrete et al., 2009). The pathogen causes yield losses to the grapevine industry and has been reported to affect as much as 30% of the crop (Ridings & Clayton, 1970). It is also responsible for girdling and dead-arm dieback of shoots (Critopoulos, 1961; Prakash et al., 1974; Navarrete et al., 2009). All vinifera and muscadine cultivars examined are susceptible to infection by G. uvicola although differences in the levels of infection have been observed (Ridings & Clayton, 1970; Longland & Sutton, 2008). Production of wine from as little as 10% infected fruit results in a rather unpleasant product with a bitter taste (McGrew, 1977), hence the name bitter rot.
The fungus reproduces asexually; a sexual stage has not been observed (Farr et al., 2001). Greeneria uvicola overwinters as a saprophyte on canes and mummified berries on the vine or on the ground (Kummuang et al., 1996b). Conidia are splash dispersed by moisture and can infect all parts of the plant (Luttrell, 1953). Humidity is vital for the development of the disease (Reddy & Reddy, 1983). The pathogen remains latent until the beginning of ripening (Kummuang et al., 1996a), as optimum temperatures for sporulation are between 8 and 24°C while optimum mycelial growth and infection occurs at 27–28°C (Ridings & Clayton, 1970; Steel et al., 2011). Penetration takes place through the berry epidermis via hyaline formation, appressorium and penetration peg (Kao et al., 1990) with hyphae colonizing the fruit tissue intracellularly (Kummuang et al., 1996a). Secondary infection may also occur through wounds caused by insects, birds and mechanical injuries. Characteristic black concentric rings of acervuli can be observed on infected berries that eventually shrivel to mummies.
Phylogenetic analysis based on molecular sequence data is used for reconstructing fungal evolutionary relationships at intraspecific and interspecific levels (James et al., 2006). Farr et al. (2001) placed G. uvicola into the order Diaporthales, class Pyrenomycetes, phylum Ascomycota based on morphological and molecular analyses of the 28S rDNA. Navarrete et al. (2009) compared G. uvicola isolates collected from wood tissues in Uruguay and their relationship to an isolate from Ohio described by Farr et al. (2001) on the basis of 28S rDNA. The authors also provided the first sequence data of the 5∙8S rDNA internal transcribed spacer 1 (ITS) of this species. The β-tubulin-2 gene sequence has also been used extensively during phylogenetic studies because it is slightly more informative than analysis of the ITS region as shown in a study of ripe rot Colletotrichum acutatum (Whitelaw-Weckert et al., 2007). However, β-tubulin-2 data for G. uvicola have not been reported.
Best management programmes include a combination of sanitation measures and chemical control; however, the few epidemiological studies that have been conducted focused on commercially unimportant wine grape cultivars and species and the fungicides recommended are now outdated. For example, the benzimidazole fungicide benomyl, that was found to be the most effective for the chemical control of G. uvicola (Ridings & Clayton, 1970), is no longer registered.
Understanding the development of a fungal pathogen is essential to plan and execute effective control measures. Greeneria uvicola has been identified on different grape species around the world but the relationships between different populations have not been investigated. The aim of this study was to enhance our knowledge of the aetiology and epidemiology of G. uvicola by investigating representative isolates from different parts of the world. In order to fulfil these aims, the phylogenetic relationships among G. uvicola isolates from V. vinifera and M. rotundifolia from the USA, Australia, India and Costa Rica were analysed by genotyping three conserved nucleotide regions: the ribosomal DNA (rDNA) internal transcribed spacer regions (ITS1–5∙8S–ITS2), 28S large subunit (LSU) nuclear rDNA, and β-tubulin-2. In addition, the morphological features, pathogenicity and in vitro sensitivity to pyraclostrobin of G. uvicola were also studied.
Materials and methods
Sixty-seven G. uvicola isolates were collected from V. vinifera cultivars in subtropical grape-growing regions in eastern Australia (Table 1). Seventeen G. uvicola isolates were collected from V. vinifera cultivars, 13 from M. rotundifolia cultivars and one from the interspecific French–American hybrid Chambourcin, in southeastern USA. Two isolates, one from India and one from Costa Rica, were obtained from the Centre for Agricultural Bioscience International (CABI, www.cabi.org). Single-spore pure cultures were collected from berries and wood following a protocol established by Steel et al. (2007). Representative isolates are maintained at the National Wine and Grape Industry Centre (Wagga Wagga, NSW, Australia), North Carolina State University (Raleigh, NC, USA) and the Plant Pathology Herbarium (DAR), Department of Primary Industry (Orange, NSW, Australia). Morphological characterization was performed according to Sutton & Gibson (1977) and Kornerup & Wanscher (1984).
Table 1. Greeneria uvicola isolates included in this study
Greeneria uvicola isolates were grown on potato dextrose agar (PDA; Oxoid) at 27°C in the dark for 2 or more weeks until conidia were formed. Conidia from each isolate were mounted on a microscope slide and fixed in 85% lactic acid. Images of conidia from each isolate were captured using Olympus colorview analySIS LS Research software (Soft Imaging System) with an Olympus AX70 True Research System microscope. The length and width of 50 conidia per isolate were measured.
Fungal DNA was extracted and molecular identification of pure cultures was achieved by PCR amplification of three highly conserved consensus regions using standard oligonucleotides as previously described (Table 2; Samuelian et al., 2011). Amplified PCR fragments were purified with the QIAquick PCR purification kit (QIAGEN) and ligated into the pGEM-T vector (Promega) following the manufacturers' instructions. Colonies with recombinant plasmids were grown in 3·5 mL liquid LB media containing ampicillin overnight at 37°C. Plasmids were purified with the QIAprep Spin Miniprep Kit (QIAGEN) following the manufacturer's protocol and sequenced in both directions with the M13 forward and reverse universal primers at the Australian Genome Research Facility (University of Queensland, Australia). DNA sequence chromatograms were analysed with chromas v. 1.45 (http://www.technelysium.com.au/chromas.html). Multiple sequence alignments, contig assembly and comparison of nucleotide sequences were performed with dnastar (DNASTAR Inc.). Nucleotide homology searches and comparison of identical sequences were performed with blast (http://www.ncbi.nlm.nih.gov). Genetic distances were calculated and maximum parsimony trees constructed for the three genes with mega v. 5.0 (Tamura et al., 2007). Nucleotide sequences obtained during this study were deposited at the National Center for Biotechnology Information (NCBI; www.ncbi.nlm.nih.gov). Where several isolates had the same DNA sequence, only one accession number was generated to represent identical isolates.
Table 2. Oligonucleotide primers used to amplify three conserved DNA regions from Greeneria uvicola
Three Australian isolates (AU01, AU67 and AU49) and 12 USA isolates (US01, US04, US06, US09, US12 and US16 from V. vinifera cultivars; US17, US19, US20, US24, US25 and US26 from M. rotundifolia cultivars) were analysed for their pathogenicity on V. vinifera cv. Chardonnay berries. These isolates were selected based on their differences within the three genomic regions analysed (Table 1; Fig. 2). AU67 is a representative of subgroup 1; AU49 is a representative of subgroup 2 (Fig. 2). US12 was chosen because this isolate, together with US13, was different from all other isolates based on the 28S LSU rDNA sequence (Fig. 2a). Disease-free mature berries (21·8°Bx) were collected from the Charles Sturt University vineyard, Orange Campus, NSW, in February 2011 and placed individually into microtitre plate wells (24-well, flat-bottom, Iwaki microplates) with pedicels facing downwards. Inoculation with G. uvicola spores was performed as previously described (Steel et al., 2007). Berries were incubated for 7 days at 27°C under a 12-h light/12-h dark photoperiod and examined for fungal infection with a dissecting microscope. Results were presented as the mean percentage of berries infected and were analysed by anova using GenStat (VSN International Ltd).
In vitro assessment of fungicide sensitivity
An in vitro pilot study performed with a small number of isolates and a range of fungicides and concentrations determined that the most effective fungicide to suppress G. uvicola development at 50% compared to the control (ED50 value) was pyraclostrobin (Cabrio; Sigma-Aldrich). (The fungicides tested in the pilot study belonged to the anilide (boscalid), anilinopyrimidine (pyrimethanil), chlorophenyl (chlorothalonil), dicarboximide (iprodione), dinitroaniline (fluazinam), phthalimide (captan) and strobilurin (azoxystrobin, pyraclostrobin, trifluxystrobin) chemical groups.) Based on these findings all G. uvicola isolates were screened for sensitivity to pyraclostrobin at 4 μg L−1 and salicylhydroxamic acid (SHAM; Sigma-Aldrich) at 100 μg mL−1 as these exhibited the most effective combination and doses to reduce mycelium growth. Amended PDA plates (Mondal et al., 2005) were prepared by diluting active ingredients in acetone and adding 1 mL L−1 of each to molten agar. The centre of each plate was inoculated with a 5-mm agar plug of mycelium obtained from the edge of an actively growing culture on PDA, with three replicates per treatment. Plates were maintained in the dark at 25°C. Measurements of fungal growth were performed simultaneously for all four treatments (i.e. no-fungicide control, SHAM, pyraclostrobin and pyraclostrobin × SHAM) and for each individual isolate, when mycelium of the no-fungicide control cultures of that individual isolate covered approximately 65–85% of the plates. Colony diameter was measured across two perpendicular axes for each plate and the two diameters averaged. The experiment was conducted twice as each isolate was evaluated in triplicate. Results were expressed with respect to the percentage inhibition of growth compared to the controls.
The statistical software R (R Development Core Team, 2010) was used to perform statistical analyses. anova techniques were used to quantify statistical differences in baseline sensitivities with regard to fungicide treatment application at and between geographical locations, cultivars, collection dates and host tissue. The linear mixed model approach ASREML-R (restricted maximum likelihood) (Butler et al., 2007) was employed to account for the random variation associated with sets (replicates) and thus minimize residual variability. Tukey's HSD all-pairwise comparisons test was used for comparison of means of fungal growth at P =0·05.
The mycelium of most G. uvicola isolates grown on PDA was creamy white at the outer edge of the colony, turning to greyish brown at the centre; however, isolates AU13, AU32, AU35, AU55 and AU65 were uniformly dark brown, and isolates US17, US26 and US29 were uniformly black (Fig. 3; Table S1). The mycelium of the USA isolates was much raised and ropey to slightly ropey while mycelium of the remaining isolates remained flat and appressed to the agar. Average lengths and widths of the conidia of G. uvicola populations are presented in Table 3 (conidia measurements of all isolates are presented in Table S2). The conidia of the Australian isolates were the smallest (Fig. 4). USA isolates identified on vinifera cultivars were significantly longer and broader than all other isolates. In addition, the conidia of the USA isolates identified on muscadine cultivars and the Indian and Costa Rican isolates were similar but significantly different from the other isolates.
Table 3. Conidia dimensions of the Greeneria uvicola isolates analysed in this studya
Mean, mean value for all conidia of the corresponding group; Max, mean of the maximum values for all isolates; Min, mean of the minimum values for all isolates. Standard errors are presented in brackets. Values with the same upper case letters indicate significantly similar groups, based on conidia length and width using Tukey's HSD all-pairwise comparisons test at a 5% similarity confidence level.
Lengths and widths of 50 conidia were measured for each isolate.
7·75 (± 0·02)A
3·00 (± 0·01)A
10·06 (± 0·14)
3·97 (± 0·06)
5·81 (± 0·09)
2·21 (± 0·04)
8·99 (± 0·04)C
4·25 (± 0·01)C
11·12 (± 0·21)
4·89 (± 0·08)
6·75 (± 0·19)
3·23 (± 0·07)
8·48 (± 0·03)B
3·36 (± 0·02)B
10·48 (± 0·15)
3·91 (± 0·06)
6·66 (± 0·16)
2·53 (± 0·07)
India and Costa Rica
8·47 (± 0·10)B
3·29 (± 0·04)B
10·97 (± 0·34)
4·34 (± 0·57)
6·81 (± 0·33)
2·55 (± 0·09)
Amplification of the three genes resulted in fragments of approximately 600 bp for the ITS sequences, 1410 bp for 28S rDNA and 960 bp for β-tubulin-2. blast analyses revealed homology to several G. uvicola isolates from Uruguay (Navarrete et al., 2009) and one from Ohio (Farr et al., 2001) in the 28S rDNA and ITS sequences. Homology to β-tubulin-2 sequences from other studies was not found in the public databases.
All of the Australian isolates and the isolates from India and Costa Rica had the same 28S rDNA sequence (‘AU, India, Costa Rica_28S’ in Fig. 2a). All muscadine isolates also had the same 28S rDNA sequence (‘US_28S, muscadine’, Fig. 2a). Most of the North American vinifera isolates possessed identical sequences in that region (‘US_28S’, Fig. 2a) with two exceptions – US12 and US13 (‘US12_13_28S’, Fig. 2a). Only one 28S rDNA sequence from USA (Farr et al., 2001) was available from NCBI and it had higher homology with the sequences from Uruguay compared with the isolates identified in this study.
The phylogenetic trees based on ITS and β-tubulin-2 regions were largely consistent with each other. However, the tree based on the β-tubulin-2 region (Fig. 2c) was more informative than the one based on the ITS regions (Fig. 2b). Isolates US05, US13, US16, US22, US27, US29 and US33 have unique β-tubulin-2 sequences; however isolate US09 has a unique ITS sequence but the same β-tubulin-2 region as two other isolates. Three well-defined clusters were formed based on all analysed genomic regions: (i) a population identified on muscadine grapes, (ii) a population identified on US vinifera grapes, (iii) a population from Australia, India and Costa Rica (Fig. 2). Interestingly, the two USA groups were more closely related to the third group and the isolates from Uruguay than to each other. The differences in the ITS and β-tubulin-2 regions among the Australian isolates were only one and two nucleotides, respectively, and the isolates were combined in two subgroups. The isolate from India was 100% identical to one of the Australian subgroups. According to the genes analysed, there were no differences between the Australian isolates with regards to different geographic regions and host tissues (berries/wood). The isolate from Costa Rica was clustered independently based on the ITS, but was close to the Australian and Indian isolates based on the β-tubulin-2 genes. The isolates from Uruguay were also grouped together for both 28S rDNA and ITS. Unfortunately, more conclusive comparisons with these sequences could not be performed as the sequences available at GenBank are incomplete at their 5′ and 3′ ends with a number of unknown (‘N’) nucleotides.
Berry infection studies
Symptoms of fungal infection, including skin cracks, oozing, splitting, browning, bleaching, mycelial growth and presence of black acervuli were noted. All G. uvicola isolates infected Chardonnay berries at 27°C and a relative humidity of c. 100% (Fig. 5). Infection rates of isolates identified on vinifera cultivars from Australia and the USA were not significantly different (P <0·05). The number of berries infected by the muscadine isolates was slightly higher at the third day after inoculation compared to the Australian and USA vinifera isolates but was significantly lower at the end of the observation period.
In vitro assessment of fungicide sensitivity
The sensitivity of all isolates was similar for each individual treatment: SHAM, pyraclostrobin and pyraclostrobin × SHAM (Table 4). Pyraclostrobin and SHAM significantly reduced colony growth of all isolates. Fungal inhibition by pyraclostrobin × SHAM (75·6 ± 9·2%) was significantly higher than pyraclostrobin alone (38·7 ± 10·3%), which was significantly higher than SHAM alone (20·4 ± 8·1%). Thus, the inhibitory effect of pyraclostrobin was increased significantly by SHAM.
Table 4. In vitro sensitivity of Greeneria uvicola to pyraclostrobin (4 μg L−1) and salicylhydroxamic (SHAM; 100 μg mL−1)
Greeneria uvicola isolates from subtropical regions of Australia, the USA, Costa Rica and India were studied for their morphology on artificial media, fungicide sensitivity, pathogenicity to V. vinifera and intraspecific variation based on three gene regions. Isolates from Australia, Costa Rica and India were all obtained from V. vinifera. The isolates from the USA were from either V. vinifera or the endemic M. rotundifolia. Muscadines are native to North America and currently only grown in southeastern USA under hot and humid conditions unsuitable for vinifera cultivars. Information regarding the V. vinifera cultivars from which G. uvicola were collected in India and Costa Rica was not available.
Phylogenetic analyses of β-tubulin-2 and ITS regions differentiated G. uvicola isolates in much more detail than the 28S rDNA sequences. The β-tubulin-2 tree had a slightly higher resolution than the ITS tree. This is in agreement with phylogenetic studies of Ophiostoma minus (Gorton et al., 2004) and C. acutatum (Whitelaw-Weckert et al., 2007). Characterization of G. uvicola populations by sequencing three genome regions clearly distinguished isolates into those from V. vinifera and those from M. rotundifolia cultivars. The Australian isolates were not differentiated by 28S rDNA sequences and divided only into two subgroups by β-tubulin-2 and ITS sequences. In addition, one of the Australian subgroups was identical to the isolate from India but not to the isolates from the USA and Costa Rica based on two genomic regions. A much higher intraspecific variation was observed amongst the USA isolates based on β-tubulin-2 and ITS sequences, suggesting that G. uvicola may have originated from North America and has arrived and spread within Australia from a single source. In addition, the finding that G. uvicola could be differentiated into vinifera and muscadine populations, even though they were collected from neighbouring vineyards raises additional questions as to how G. uvicola spreads.
Most of the isolates were morphologically identical, with five exceptions from the Australian isolates and three exceptions in the muscadine isolates. This variation in morphology could be misleading for identification purposes. The reasons for such differences in fungal morphology are currently unknown.
Pathogenicity of G. uvicola was studied on Chardonnay fruit as this cultivar has been reported to be highly susceptible to bitter rot (Miranda, 2004). Isolates of G. uvicola obtained from V. vinifera, regardless of country of origin and cultivar, were significantly more aggressive on detached Chardonnay berries than isolates obtained from M. rotundifolia. This is in agreement with the study conducted by Miranda (2004) and might explain the higher level of losses in V. vinifera vineyards caused by bitter rot in contrast to M. rotundifolia vineyards (Clayton, 1975; Miranda, 2004). In addition, the vinifera G. uvicola isolates used to study the pathogenicity of the fungus on Chardonnay berries were able to infect M. rotundifolia berries (data not shown).
Recommendations for bitter rot control on both V. vinifera (Sutton & Burrack, 2012) and M. rotundifolia (Cline & Burrack, 2012) cultivars in North Carolina involve application of strobilurin and other fungicides up to 14 days prior to harvest. These management options are not available to wine grape growers in Australia due to restrictions on pesticide use for wine destined for European export markets. For example, pyraclostrobin cannot be used after bunch closure or within 63 days of harvest. Pyraclostrobin was introduced in the late 1990s in the USA, and is moderately to highly effective for the control of a number of fungal diseases. In Australia it was approved for use in 2003 (Australian Pesticide and Veterinary Medicines Authority, www.apvma.gov.au). However, frequent application of strobilurin may lead to resistance development (Timmer, 2004; Mondal et al., 2005). Several fungicides (azoxystrobin, boscalid, captan, chlorothalonil, fluazinam, iprodione, pyraclostrobin, pyrimethanil and trifluroxystrobin) were tested in a pilot study for their effectiveness against bitter rot, with pyraclostrobin being found to be one of the most effective inhibitors (data not shown). This current study tested whether there was an indication of G. uvicola developing resistance to pyraclostrobin based on longer and more frequent use of the fungicide in the USA. The results indicated that, despite differences in pyraclostrobin usage, there was no difference between the G. uvicola isolates even though some were collected before the introduction of pyraclostrobin in viticulture. Many fungi possess an alternative respiration pathway that avoids electron transport through the cytochrome b target site of pyraclostrobin fungicides. SHAM blocks this alternative respiration pathway and has been used to investigate fungal sensitivity to pyraclostrobin fungicides (Olaya & Köller, 1999; Sierotzki et al., 2000; Avila-Adame et al., 2003; Mondal et al., 2005). Mondal et al. (2005) found that a number of citrus pathogens including C. acutatum, Alternaria alternata, Elsinoe fawcettii, Diaporthe citri and Mycosphaerella citri do not use the alternative pathway to avoid the toxic effect of the strobilurin fungicides. Greeneria uvicola growth was inhibited by SHAM independently and in combination with pyraclostrobin, leading to the conclusion that G. uvicola is using the alternative respiration pathway and therefore has the potential to develop resistance to pyraclostrobin and other strobilurins. Interestingly, fungicide sensitivity to strobilurins studied in vitro may differ from actual sensitivity to these chemicals in vivo (Olaya et al., 1998; Olaya & Köller, 1999; Bartlett et al., 2002; Avila-Adame & Köller, 2003; Avila-Adame et al., 2003; Mondal et al., 2005) and further research is necessary to determine whether that is true for G. uvicola.
In this study, independent populations of G. uvicola were differentiated on the basis of region of distribution and/or host plant. Understanding the genetic diversity and origin of the fungus should be useful to regulatory agencies in helping to prevent the introduction of the pathogen into countries where it does not exist, or determining the origin of an introduction. The data may also help plant breeders select isolates when screening populations of grapes for resistance. Finally, it provides some evidence that fungicide control programmes developed in one area may be effective in others as well. For example, despite the greater applications of pyraclostrobin in the USA compared to the Australian vineyards, there was no significant difference in pyraclostrobin sensitivity between isolates from the two regions.
This work was conducted within the Winegrowing Futures Program, a joint initiative of the Grape and Wine Research and Development Corporation (GWRDC) and the National Wine and Grape Industry Centre, Charles Sturt University. We thank Jim Hardie for critical review of the manuscript; Chris Haywood (New South Wales Department of Industry and Investment) for managing the experimental sites; Rujuan Huang, Ekaterina Simova, Lisa Greer and Stacey Greer for technical assistance; Matthew Prescott and Leigh Schmidtke for help with statistical analyses; and the grape growers and winemakers of New South Wales, Queensland and North Carolina who provided access to their vineyards and participated in this study.