Phytotoxic dioxolanones from Guignardia bidwellii can be described as potential virulence factors which cause the formation of lesions upon an infection by G. bidwellii. The toxin guignardic acid was found in planta of G. bidwellii-infected Vitis vinifera leaves, whereas no phytotoxic dioxolanones were detected in uninfected leaf material. Secondary metabolism analyses of further phytopathogenic fungi from the genus Guignardia led to the observation that all species investigated can produce the phytotoxins known from G. bidwellii. In addition to these studies, it was demonstrated that phenguignardic acid is biosynthetically derived from two molecules of phenylalanine and that phenylalanine is a key precursor in the biosynthesis of the two other phytotoxins – alaguignardic acid and guignardic acid.
Guignardia bidwellii (Ellis) Viala & Ravaz (currently valid name Phyllosticta ampelicida (Engelm.) Aa; Botryosphaeriales, Ascomycota) is the causal agent of grape black rot1. This disease can cause substantial losses of grapes in areas with humid summers2. Previous studies of submerged fermentations of G. bidwellii led to the identification of nine secondary metabolites, five of which are phytotoxic3, 4. The metabolites contain a dioxolanone moiety and are presumably biosynthesised from transaminated amino acids, such as phenylalanine, valine, tyrosine and alanine. This is supported by the fact that supplementation experiments with phenylalanine led to an enhanced biosynthesis rate of one of the phytotoxic dioxolanones. In addition, a cryptic NRPS-like gene called pgnA was identified recently in Aspergillus terreus and its heterologous expression in Aspergillus nidulans confirmed that this gene is solely responsible for the biosynthesis of phenguignardic acid5. Sun et al.5 proposed its formation by linking two phenylpyruvic acids. The toxins are non-host specific and a structure activity relationship indicated that a free carboxyl group appears to be required for the phytotoxic activity4.
The secondary metabolites are believed to be involved in the pathogenicity of G. bidwellii and can be described as potential virulence factors, due to the fact that the application of the pure compounds onto the host tissue causes specific disease symptoms6. Additional studies should be conducted to corroborate this hypothesis. It is necessary to analyse whether the toxins can be found upon infection by the fungus in planta. Furthermore, the questions whether other phytopathogenic species of the Guignardia relationship can produce the phytotoxic dioxolanones and whether apathogenic species are toxin-free need to be addressed.
These questions were considered in the present study. As for the question whether the toxin can be identified in planta after infection by G. bidwellii, infected and non-infected leaf material of Vitis vinifera was analysed for the presence of phytotoxic dioxolanones. Three phytopathogenic species and one biotrophic endophyte of the genus Guignardia were cultivated and the production of secondary metabolites was analysed to see whether these fungi can produce the phytotoxic principles. In addition, the precursors for the biosynthesis of the phytotoxic dioxolanones were elucidated by isotopic labelling.
Results and Discussion
Toxin detection in planta
The production of phytotoxins in planta is an important requirement to assume the involvement of the toxin in pathogenicity7. In this study, both G. bidwellii-infected and non-infected leaf material was analysed by analytical HPLC after extraction. Comparison of both chromatograms (Fig. 1A) led to the detection of a signal at a retention time of 8.7 min in the infected sample (Fig. 1A, red), which was not present in the non-infected sample (Fig. 1A, blue).
The UV spectra recorded for this compound (Fig. 1C) resemble those of the phytotoxic dioxolanones known from G. bidwellii (local maxima around 230 nm and 295 nm3, 4). By isolation of the metabolite and HPLC-MS analysis the metabolite was identified as guignardic acid. This compound is the first member of the new class of secondary metabolites described8. Thus, the production of one of the toxins of the fungus during active growth in planta was proven. The detection of phytotoxins in planta needs highly sensitive methods of measurement, because most toxins are highly potent, therefore, only small quantities are required to cause symptoms of infection9. A concentration of 45 µg/g leaf material (wet) of guignardic acid was determined experimentally in the infected sample.
Toxin production in other species of the genus Guignardia
Phytotoxins, described as virulence factors, are frequently characteristic for a whole genus10. Several phytopathogenic Fusarium species, for example, produce trichothecenes, which inhibit eukaryotic protein biosynthesis. If the biosynthesis of trichothecenes is impaired, virulence is reduced (refs 11, 12 and references therein). Isolates of the genus Cercospora produce cercosporin and the targeted disruption of its biosynthesis in C. nicotianae led to fewer necrotic lesions on inoculated tobacco leaves13, 14. Both groups of toxins are therefore regarded as virulence factors. Phytotoxic dioxolanones from G. bidwellii can also be considered as virulence factors. We hypothesise that closely related species produce similar compounds, so three further phytopathogenic species and one species reported as endophyte15,16,17,18 of the genus Guignardia were explored (Table 1).
All species investigated produce one or more phytotoxic dioxolanones known from G. bidwellii in submerged cultures. Table 1 gives an overview of the secondary metabolites produced and their amount determined experimentally.
The fact that all Guignardia species examined produce dioxolanone-type secondary metabolites known from G. bidwellii supports the assumption that the toxins are virulence factors.
Interestingly, even G. mangiferae produces guignardic acid, although the fungus was isolated as a biotrophic endophyte19. A possible explanation could be that usually silent gene clusters responsible for phytotoxin production were activated within the artificial environment during fermentation. Another possibility could be that G. mangiferae produces the toxin even within the host, but the lesion formation is inhibited by the plant defence and, therefore, the fungus was classified as an endophyte20. Kuldau and Yates21 demonstrated that experimental or suboptimal conditions could stress the host and, therefore, the plant defence is weakened. This can result in disease caused by a usually asymptomatic fungus. Thus, the possibility that a fungus becomes pathogenic, although isolated as endophyte, cannot be ruled out21.
The study demonstrated that all Guignardia species investigated can produce phytotoxic dioxolanones. Therefore, these compounds can be described as potential virulence factors in the infection process.
Elucidation of the precursor for the biosynthesis of phytotoxic dioxolanones
Phytotoxic dioxolanones from G. bidwellii have not been studied extensively to date4. Even though the genetic origin of phenguignardic acid was recently uncovered by the heterologous expression of a cryptic NRPS-like gene found in Aspergillus terreus in Aspergillus nidulans 5, the biosynthesis in G. bidwellii is not yet understood. Previous studies have led to the assumption that the compounds were built up from α-keto acids related to proteinogenic amino acids. An involvement of the keto-precursors to phenylalanine and valine in guignardic acid, two phenylalanines in phenguignardic acid, phenylalanine and alanine in alaguignardic acid and phenylalanine and tyrosine in guignardianone E and F is assumed3, 4.
Supplementation experiments with the precursor phenylalanine or with pyridoxal phosphate, the coenzyme of the amino acid metabolic pathway, were carried out during submerged fermentation of G. bidwellii and led to increased yields of phenguignardic acid4. These results support the theory that amino acids are biosynthetic precursors for the phytotoxic dioxolanones.
Supplementation experiments were performed with 13C-labelled phenylalanine in this study to verify this hypothesis.
If phenylalanine was the precursor for the biosynthesis of the toxins, an enrichment of 13C at the C-5 position (and, in addition, the C-6 position in phenguignardic acid) would be detectable. Figure 2 shows the hypothetical biosynthesis of phenguignardic acid.
After fermentation of G. bidwellii with supplemented 13C-labelled phenylalanine, the pure compounds alaguignardic acid, guignardic acid and phenguignardic acid were isolated by preparative HPLC and analysed by HPLC-MS and 13C NMR spectroscopy.
Strong phenguignardic acid signals at 163 and 169 ppm could be detected in the 13C NMR spectra of the labelled sample (Fig. 3A, black). By contrast, these signals had a lower intensity in the unlabelled sample of phenguignardic acid (Fig. 3A, red), while the remaining signals in the range of 125 and 135 ppm have equal intensity in both spectra. Compared with NMR data from Molitor et al.3 and the structural formula of phenguignardic acid (Fig. 3B,C), the labelled carbon atoms are located at the C-5 and C-6 position in phenguignardic acid. This agrees with the proposed biosynthetic pathway shown in Fig. 2, as well as that proposed by Sun et al.5.
The investigation of the 13C-labelled samples of guignardic acid and alaguignardic acid by NMR spectroscopy led to a strong increase of the signal at 162 ppm in both samples (Fig. 4A,C, black).
Only signals with lower intensity at this position could be found in the unlabelled samples of the compounds (Fig. 4A,C, red). Similar to phenguignardic acid, the signals in the range of 125 and 135 ppm have equal intensity in both spectra. The labelled carbon atom could be identified as C-5 by comparison with NMR data from Rodrigues-Heerklotz et al.8 and Buckel et al.4 and the structural formula of guignardic acid and alaguignardic acid (Fig. 4B,D).
The results of these studies verify the hypothesis that two molecules of phenylalanine are required for the biosynthesis of phenguignardic acid. Furthermore, an involvement of phenylalanine as an essential module in all compounds from G. bidwellii investigated was proven. These results support the proposed biosynthesis via an NRPS-like gene by Sun et al.5.
This is an important step towards the elucidation of the biosynthesis of phytotoxic dioxolanones in G. bidwellii. The next step is to characterize their genetic origin. Even if Sun et al.5 found the NRPS-like gene in Aspergillus terreus responsible for the biosynthesis of phenguignardic acid, nothing is known to date regarding the genus Guignardia and the other dioxolanones, such as guignardic acid and alaguignardic acid.
The phytotoxic secondary metabolites from G. bidwellii can be described as potential virulence factors due to the facts that the toxin guignardic acid is detectable in G. bidwellii-infected leaves of V. vinifera and all species of the genus Guignardia investigated can produce these phytotoxic dioxolanones. Therefore, these compounds seem to play a crucial role in the infection process of G. bidwellii. The secondary metabolites belong to a new class of natural products and their biosynthesis in G. bidwellii has still not been understood. It could be proven by feeding experiments with 13C-phenylalanine that all phytotoxic dioxolanones incorporate at least one phenylalanine. These results form a reliable base for further investigation into the pathogen-host-interaction between G. bidwellii and V. vinifera and can help to develop new defence strategies, even for applications in ecological vineyards.
Materials and Methods
General experimental procedures
Compounds were analysed by HPLC (Agilent 1100 Series) equipped with an Eclipse XDB Phenyl-column (3 × 150 mm; 3.5 µm, Agilent, Waldbronn, Germany) at 40 °C and a flow rate of 0.7 ml min−1. The elution gradient was composed of H2O + 0.1% v/v phosphoric acid and acetonitrile. Compounds were detected via UV at 300 nm. The quantification of guignardic acid and phenguignardic acid was performed using an external standard. A calibration curve was obtained by measuring standards at different concentrations with an injection volume of 10 µl. By the means of this calibration, the amount [mg l−1] of guignardic acid, respectively, phenguignardic acid was determined. The regression coefficient amounts to 1.00 for a quadratic curve type.
A Zorbax Eclipse XDB Phenyl-column (250 × 9.4 mm; Agilent, Waldbronn, Germany) with a flow rate of 7 ml min−1 was used for preparative HPLC. The compounds were fractionated in an isocratic manner with H2O + 0.1% v/v formic acid and acetonitrile.
The mass spectra were recorded by HPLC (Series 1200, Agilent, Waldbronn, Germany) with an UV-DAD and a coupled LC/MSD Trap XCT-ESI-mass spectrometer. An Ascentis® Express C18 column (3 × 2.1 mm; 2.7 µm, Supelco, Sigma-Aldrich Chemie GmbH, Steinheim, Germany) was used at 40 °C and a flow rate of 1.0 ml min−1. The elution was performed with a gradient of H2O + 0.1% v/v formic acid and acetonitrile.
NMR spectra were recorded on an Avance-III 600 MHz spectrometer (Bruker, Rheinstetten, Germany) equipped with a 5 mm cryo-TCI probe head using standard pulse sequences at 23 °C. The signals were referenced to the respective solvent signals of CDCl3 (δH 7.26, δC 77.16).
Fermentations were carried out in duplicate. Values given are means of the quantities determined in the individual fermentation experiments. The standard deviation did not exceed 5% ± of the total amount measured. All fermentations were carried out in 1 l malt extract medium composed of malt extract (20 g, Lindenmeyer GmbH & Co. KG, Heilbronn, Germany), peptone (1 g, Becton, Dickinson GmbH, Heidelberg, Germany) and glucose (20 g) per 1 l tap H2O in 2 l Erlenmeyer flasks. Agar plugs of well-grown cultures were transferred aseptically as inoculum. Fermentation was carried out at 22 °C on an orbital shaker at 120 rpm. A sample was taken every second day to monitor the production of secondary metabolites and to measure glucose content (Diabur-Test 5000, Roche Diagnostics, Mannheim, Germany) and pH value (pH 209, HANNA® Instruments Deutschland GmbH).
The samples were separated into mycelium and culture filtrate by filtration via a Büchner funnel for HPLC analysis. After extraction of the culture filtrate with an equal volume of EtOAc, the organic solvent was dried over Na2SO4 and evaporated in vacuo to dryness. The mycelium was discarded.
Strains examined: CBS 756.70, obtained from the Centraalbureau voor Schimmelcultures (CBS, Fungal Biodiversity Centre, The Netherlands) as Guignardia aesculi (Peck) V.B. Stewart, currently valid name Phyllosticta paviae Desm. (syn. Phyllosticta sphaeropsoidea Ellis & Everh., for phylogeny of this strain see ref. 22). CBS 111645, obtained from CBS as G. bidwellii (Ellis) Viala & Ravaz, currently valid name Phyllosticta ampelicida (Engelm.) Aa. This strain (cited as ex-type strain23) was recently classified as species new to science and called Phyllosticta parthenocissi K. Zhang, N. Zhang & L. Cai based on a phylogenetic analysis of combined sequence data of internal transcribed spacer (ITS), translation elongation factor 1-α, glyceraldehyde-3-phospatase dehydrogenase, and actin as well as culture characteristics of this strain. Unfortunately no ecological data are available for this strain23. CBS 123405, obtained from CBS as G. mangiferae A.J. Roy, currently valid name Phyllosticta capitalensis Henn., for a phylogenetic analysis including this strain see ref. 24. CBS 126.22, obtained from CBS as G. vaccinii Shear (ex-type), for a phylogenetic analysis including this strain see ref. 23. IMI 147360 (ATCC 28563), obtained from the International Mycological Institute (England) as G. musae Racib25. These fungi were maintained as described by the corresponding culture collection. The principle of ‘one fungus: one name’ was enforced in 2011 (Melbourne Code; ICN) and most Guignardia species (sexual forms) relevant for the present study are currently validly cited as species of Phyllosticta (former asexual forms; Wulandari et al.18 and references therein). Nevertheless, the names of the sexual forms are applied in the present document, because they are established as references in this field of study.
Toxin detection in planta
Fresh G. bidwellii-infected and non-infected leaf material of V. vinifera cv. Riesling was analysed for the detection of phytotoxic dioxolanones within infected vine leaf material. Infected (43 g) and non-infected material (50 g), respectively, were transferred to 500 ml EtOAc and shredded by a hand-held blender (ESGE Zauberstab®, Unold AG, Hockenheim, Germany). Another 500 ml EtOAc were added and the mixtures were incubated for 30 min at RT on an orbital shaker. After incubation, the extracts were separated from the solids by filtration, the extracts were dried with Na2SO4 and evaporated in vacuo to dryness. Phytotoxic dioxolanones were identified by analytical HPLC and HPLC/MS.
Toxin production in other species of the genus Guignardia
Four other species (phytopathogenic or endophytic) were cultivated, as described above, to examine whether other species of the genus Guignardia could produce phytotoxic dioxolanones. The production of secondary metabolites was monitored every second day. The fermentation was stopped when the free glucose was depleted. The culture fluids were extracted, as mentioned above, and secondary metabolites were identified and quantified by analytical HPLC and HPLC/MS.
G. bidwellii was fermented as described above. After the production of known secondary metabolites was detected by analytical HPLC, 1-[13C] phenylalanine was supplemented aseptically (final concentration 2.5 mM). Fermentation was stopped as soon as the production rate for the compounds had reached the maximum value. The culture fluid was separated from the mycelium by filtration and was extracted with an equal volume of EtOAc. The dried crude extract was applied onto Chromabond C18ec columns (Macherey-Nagel) and elution was performed in six steps with a gradient of H2O + 0.1% v/v formic acid and acetonitrile. The fractions were analysed by HPLC and relevant fractions were purified by preparative HPLC, as described above. Pure compounds of guignardic acid, phenguignardic acid and alaguignardic acid were analysed by NMR spectroscopy for 13C enrichment.
Ferrin, D. M. & Ramsdell, D. C. Ascospore Dispersal and Infection of Grapes by Guignardia bidwellii, the Causal Agent of Grape Black Rot Disease. Phytopathology 77, 1501 (1977).
Hoffman, L. E. & Wilcox, W. F. Factors Influencing the Efficacy of Myclobutanil and Azoxystrobin for Control of Grape Black Rot. Plant Disease 87, 273–281 (2003).
Molitor, D. et al. Phenguignardic acid and guignardic acid, phytotoxic secondary metabolites from Guignardia bidwellii. Journal of Natural Products 75, 1265–9 (2012).
Buckel, I. et al. Phytotoxic dioxolanone-type secondary metabolites from Guignardia bidwellii. Phytochemistry 89, 96–103 (2013).
Sun, W.-W., Guo, C.-J. & Wang, C. C. C. Characterization of the product of a nonribosomal peptide synthetase-like (NRPS-like) gene using the doxycycline dependent Tet-on system in Aspergillus terreus. Fungal Genetics and Biology 89, 84–88 (2016).
Molitor, D. & Beyer, M. Epidemiology, identification and disease management of grape black rot and potentially useful metabolites of black rot pathogens for industrial applications - a review. Annals of Applied Biology 165, 305–317 (2014).
Kohmoto, K. In Modern Methods of Plant Analysis (eds Linskens, H. F. & Jackson, J. F.) 51–73 (Springer Verlag, 1992).
Rodrigues-Heerklotz, K. F., Drandarov, K., Heerklotz, J., Hesse, M. & Werner, C. Guignardic Acid, a Novel Type of Secondary Metabolite Produced by the Endophytic Fungus Guignardia sp.: Isolation, Structure Elucidation, and Asymmetric Synthesis. Helvetica Chimica Acta 84, 3766–3772 (2001).
Kheder, A. A., Akagi, Y., Takao, K., Akamatsu, H. & Kodama, M. Fungal growth and in planta distribution of host-specific AAL-toxin in tomato plants infected with the tomato pathotype of Alternaria alternata. Mycotoxins 62, 7–13 (2012).
Berestetskiy, A. O. A review of fungal phytotoxins: from basic studies to practical use. Applied Biochemistry and Microbiology 44, 453–465 (2008).
Roncero, M. Fusarium as a model for studying virulence in soilborne plant pathogens. Physiological and Molecular Plant Pathology 62, 87–98 (2003).
Desjardins, A. E., Hohn, T. M. & McCormick, S. P. Trichothecene biosynthesis in Fusarium species: chemistry, genetics, and significance. Microbiological Reviews 57, 595–604 (1993).
Fajola, A. O. Cercosporin, a phytotoxin from Cercospora spp. Physiological Plant Pathology 13, 157–164 (1978).
Choquer, M. et al. The CTB1 Gene Encoding a Fungal Polyketide Synthase Is Required for Cercosporin Biosynthesis and Fungal Virulence of Cercospora nicotianae. Molecular Plant-Microbe Interactions 18, 468–476 (2005).
Baayen, R. P. et al. Nonpathogenic Isolates of the Citrus Black Spot Fungus, Guignardia citricarpa, Identified as a Cosmopolitan Endophyte of Woody Plants, G. mangiferae (Phyllosticta capitalensis). Phytopathology 92, 464–477 (2002).
Everett, K. R. & Rees-George, J. Reclassification of an isolate of Guignardia citricarpa from New Zealand as Guignardia mangiferae by sequence analysis. Plant Pathology 55, 194–199 (2006).
Wikee, S. et al. Phyllosticta capitalensis, a widespread endophyte of plants. Fungal Diversity 60, 91–105 (2013).
Wulandari, N. F., Bhat, D. J. & To-anun, C. A modern account of the genus. Phyllosticta. Plant Pathology & Quarantine 3, 145–159 (2013).
Wickert, E., Lemos, E. G. de M., Kishi, L. T., de Souza, A. & de Goes, A. Genetic Diversity and Population Differentiation of Guignardia mangiferae from ‘Tahiti’ Acid Lime. The Scientific World Journal 2012, 1–11 (2012).
Schulz, B. & Boyle, C. The endophytic continuum. Mycological Research 109, 661–686 (2005).
Kuldau, G. A. & Yates, I. E. In Microbial Endophytes (eds Bacon, C. W. & White, J. F.) 85–117 (CRC Press, 2000).
Wikee, S. et al. A phylogenetic re-evaluation of Phyllosticta (Botryosphaeriales). Studies in Mycology 76, 1–29 (2013).
Zhang, K., Zhang, N. & Cai, L. Typification and phylogenetic study of Phyllosticta ampelicida and P. vaccinii. Mycologia 105, 1030–1042 (2013).
Glienke, C. et al. Endophytic and pathogenic Phyllosticta species, with reference to those associated with Citrus Black Spot. Persoonia - Molecular Phylogeny and Evolution of Fungi 26, 47–56 (2011).
Punithalingham, E. & Holliday, P. Guignardia musae. IMI Descriptions of Fungi and Bacteria 47, 467 (1975).
This work was financially supported by the state of Rheinland-Pfalz, the Naturstoffzentrum Rheinland-Pfalz and the BASF SE. We are grateful to Prof. Dr. Beate Berkelmann-Löhnertz and Nicole Siebert for the kind provision of G. bidwellii-infected and non-infected leaf material of V. vinifera.
The authors declare that they have no competing interests.
Publisher's note: Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.
Electronic supplementary material
About this article
Cite this article
Buckel, I., Andernach, L., Schüffler, A. et al. Phytotoxic dioxolanones are potential virulence factors in the infection process of Guignardia bidwellii . Sci Rep 7, 8926 (2017). https://doi.org/10.1038/s41598-017-09157-6
Hunting modulators of plant defence: the grapevine trunk disease fungus Eutypa lata secretes an amplifier for plant basal immunity
Journal of Experimental Botany (2020)
The Journal of Antibiotics (2019)