Phytoplasmas cannot be cultivated in a pure culture in vitro and sustaining the phytoplasma infection sometimes appears difficult, as previously shown in in vitro grown grapevine (Eveillard et al. 2016). Therefore, studies dealing with investigation of interactions between grapevine and FDp remain almost exclusively based on sampling and analysis of field-grown plants (Abbà et al. 2014; Gambino et al. 2013; Margaria et al. 2014; Prezelj et al. 2016; Roggia et al. 2014). Initially, it was believed that changes to host physiology are an indirect response to the infection; however, more recent studies revealed that phytoplasma genomes possess a wide variety of effector molecules which enable them to directly modulate host physiology (Tomkins et al. 2018). Although whole genomes of some phytoplasmas are extensively studied and many of their effectors recognized, for example for ‘Ca. P. solani’ (Šeruga Musić et al. 2019), detailed whole genome analyses of FDp and functional analyses of putative effectors still await. Although there are some disadvantages of this approach, studying the physiological responses of field-infected grapevine could give an insight into potential targets of FDp effectors, and assist the identification of potential effectors of this phytoplasma. By combining the molecular identification with assessing the impact on grapevine physiology, this study will contribute to unravelling the mechanisms of FDp pathogenicity. Furthermore, to our knowledge, no previous study attempted to investigate the impact of different FDp strains, based on the phylogenetic relationship between relevant genes used in genotyping, on specific targets in grapevine physiology and correlate it with potential differences in pathogenicity.
3. 1. Indications for the epidemiology of FDp
FDp was detected in all symptomatic leaves at all three sampling time points. Phylogenetic analysis of map gene amplicons revealed that all the FDp-positive samples were infected by the FD genotypes belonging to the mapFD2 genetic cluster (Fig. 1a). This is unsurprising, considering previous studies revealed that isolates belonging to the mapFD2 cluster were predominant in Croatian vineyards (Plavec et al. 2019). However, two different genotypes within the mapFD2 cluster were detected, M38 and M54. Identified FDp isolates within the mapFD2 genetic cluster comprise two genotypes with map sequences 100% identical to sequences classified as genotypes M38 and M54. The same study (Plavec et al. 2019) discussed that M54 genotype within mapFD2 cluster could be the prevalent one in Croatian vineyards and its highly efficient transmission by the vector S. titanus was hypothesised. This genotype was identified as responsible for heavy damage to the viticultural production in Istria region of Croatia (Plavec et al. 2019). Prevalence of this genotype was also observed in Italian (Rossi et al. 2019) and French vineyards, where this genotype represented 85% of all FD cases in 2000s (Malembic-Maher et al. 2020). Present study revealed higher relative FDp DNA abundance in M54-infected grapevine, than in M38-infected grapevine, pointing to a higher phytoplasma titre (Fig. 1b). This could be linked to higher replicative potential of this genotype within grapevine plants and potentially, easier acquisition by the vector, leading to its higher geographic distribution. Indeed, Galetto et al. (2014) confirmed that acquisition capability of the FDp by the vector S. titanus correlates with phytoplasma titre in the source plant. Therefore, ability to attain higher titre in infected grapevine could be a potential reason for the higher geographic distribution of M54 genotype. Relative FDp DNA abundance was quantified using a real-time PCR assay (Pelletier et al. 2009) through the development of infection at three time points for each identified FDp genotype. The trend of FDp DNA between different time points was similar for samples of both genotypes. As expected, for both M38- and M54-infected samples, the lowest DNA abundance was observed at the first time point in June, earlier in the vegetative season. The peak in FDp DNA abundance was in late summer (August time point) for both M38- and M54-infected samples. A lower DNA abundance at the third time point compared to the August time point was more pronounced in M54-infected grapevine, with a somewhat lower reduction of FDp DNA abundance in M38-infected grapevine. This dynamics is in agreement with the previous results for the FDp titre in field grown grapevine through the vegetative season (Roggia et al. 2014), reflecting the similar and conserved seasonal dynamics of FDp replication in grapevine. Interestingly, more effective phytoplasma acquisition by mature S. titanus was observed, rather than S. titanus larvae. Considering that S. titanus matures in July/August (Chuche and Thiéry 2014), the highest phytoplasma titre in grapevine at this time point could be the phytoplasma mechanism to maximize the rate of transmission to mature potential vectors who can acquire the phytoplasma more effectively than larvae.
3. 2. Oxidative stress conditions affected by FDp genotypes
Both M54- and M38-infected grapevine had characteristic GY symptoms at the first time-point: yellowing, leaf rolling and fruit drying. Morphologically, no difference in disease severity could be observed. However, at the second, and third time point particularly, stronger disease symptoms were observed in M54-infected grapevine. Malondialdehyde (MDA) content, an indicator of the level of membrane lipid peroxidation, is an often-used biomarker of stress conditions in plants, due to the susceptibility of membranes to ROS-induced damage (Hodges et al. 1999). Enhanced ROS synthesis is a common defence mechanism against plant pathogens. This increase in ROS synthesis often leads to oxidative stress in infected plants (Suman et al. 2021). The level of lipid peroxidation was higher in infected plants at all three time points in comparison to the corresponding controls (Fig. 2a). Increased level of lipid peroxidation was previously reported in phytoplasma-infected plants, compared to the uninfected plants (Kiprovski et al. 2018; Mitrovic et al. 2021). At the first and second time point, MDA content was the highest in M38-infected leaves. However, at the third time point, the MDA content was the highest in M54-infected leaves. Considering that FDp DNA abundance was higher in M54-infected leaves, higher level of lipid peroxidation was expected in these leaves as a result of the oxidative burst. However, it is known that FDp genomes possess superoxide dismutase (sod) gene (Carle et al. 2011), encoding an antioxidative enzyme responsible for quenching superoxide radicals. This potential phytoplasma-mediated suppression of plant-derived ROS could have led to the lower level of lipid peroxidation in grapevine leaves infected with higher phytoplasma titre at the first and second time points, as a mechanism for the protection from plant-derived oxidative burst. Only at the third time point, in which relative FDp DNA abundance in M54-infected leaves dropped more severely (which could have led to a more drastic drop in FDp SOD activity) than in M38-infected leaves, was the level of lipid peroxidation higher in M54- than in M38-infected leaves, corroborating this hypothesis. Also, phytoplasma effectors could potentially manipulate the induction of synthesis of antioxidative compounds (for example, phenolics) in plants to quench protective ROS, as proposed previously (Prezelj et al. 2016). Interestingly, absolute MDA content positively correlated with the L-ascorbic acid (AA) content (r = 0.69), as did their relative content (r = 0.74) (Fig. 3, Supplementary material 1). Considering that the AA (vitamin C) is a strong antioxidant, its synthesis could be induced upon oxidative damage caused by the oxidative burst, thus establishing a link between the MDA and AA content. Indeed, AA levels were higher in infected leaves at all three time points (Fig. 2c). Based on the metabolite and the expression of genes involved in AA biosynthesis, Xue et al. (2020) also recognised the involvement of AA in phytoplasma-infected plants. Indirect role of AA in plants battling biotic stress and its induction as a response to the oxidative burst have been reviewed (Khan et al. 2012), suggesting to it being a general response mechanism in plant-pathogen interactions, not specific to the phytoplasma infection. For M38-infected leaves, the highest AA concentration in comparison to the controls was at the second time point, and the lowest at the first time point. Likewise, in M54-infected leaves the lowest concentration was at the first time point, however, the highest was at the third time point (Fig. 2d). Also, similarly to the MDA content, AA content was higher in M38-infected leaves at the first and second time points. However, at the third time point, AA content was also the highest in M54-infected leaves. This further supports the hypothesis that the mechanisms of induced ROS synthesis against FDp invasion are followed by responses to minimize oxidative damage by inducing the synthesis of antioxidants such as AA. Activation of responses to minimize oxidative damage could either be phytoplasma-induced, plant-induced (to minimize damage to its cells) or a combination of both and should be a subject of further studies.
3. 3. Osmolyte accumulation in FDp-infected grapevine
Both proline and sugars, which often accumulate in plants as a response to abiotic and biotic stressors, are reliable biomarkers of stress conditions in plants (Jeandet et al. 2022; Liang et al. 2013; Sharma et al. 2019). Soluble sugars (SS) content at all three time points followed the same trend, being the highest in M38-infected leaves, and lowest in uninfected leaves (Fig. 2e). Because phytoplasmas colonize phloem tissues in infected leaves, this causes the physical obstruction of transport of sugars due to callose deposition in phloem cell walls, leading to elevated sugar levels in leaves (Musetti et al. 2013). Furthermore, a substantial upregulation of genes involved in sugar biosynthesis was observed in FDp-infected leaves (Prezelj et al. 2016). However, it is still unknown whether this is induced by FDp effectors or is a more of an unspecific response of grapevine to the infection. Nevertheless, this sugar-rich environment creates a perfect setting for FDp replication, due to its metabolism, which is strictly oriented to glycolysis (Carle et al. 2011). Interestingly, a negative correlation was observed between relative SS content and FDp phytoplasma DNA abundance (r = -0.90) (Fig. 3b, Supplementary material 1). Particularly, lower SS content was recorded in M54-infected leaves, which had higher FDp DNA abundance. This could potentially be due to the M54 genotype having a less severe impact on sugar metabolism in infected grapevine than the M38 genotype. Also, the higher titre of M54 genotype could cause higher usage of leaf sugars by phytoplasmas, lowering the total SS in M54-infected leaves compared to M38-infected leaves. The same dynamics of relative SS content was observed for both genotypes, with the highest being at the first time point (Fig. 2f). Similarly, Negro et al. (2020) reported lower relative SS in September than July in BNp-infected grapevine. This could be due to the lower photosynthetic capacity of infected grapevine at later time points, resulting in lower relative SS content (Rasool et al. 2020).
Similarly to SS, upregulation of proline biosynthesis was also previously reported in phytoplasma-infected American cranberries (Pradit et al. 2019), as well as in phytoplasma-infected sweet orange (Rasool et al. 2020). At all three time points, proline content was the highest in M54-infected leaves (Fig. 2g). Levels in M38-infected leaves were higher than the control only at the second time point. Proline acts as a protective molecule in plants during various stress conditions acting as an osmolyte, but also by being involved in the stabilization of proteins, direct scavenging of ROS and promoting downstream cellular signalling pathways related to plant defence mechanisms (Liang et al. 2013). Its relative content positively correlated (r = 0.85) with relative FDp DNA abundance (Fig. 3b, Supplementary material 1). This points that proline accumulation is a protective response mechanism against FDp infection. Although it appears to be a direct response mechanism and it is being affected by phytoplasma titre, the hypothesis is that this response is not a part of the PAMP-triggered immunity. Rather, our hypothesis is that proline accumulation is a response to FDp-induced symptoms in infected grapevine leaves, due to its general and fundamental role in stress response (Liang et al. 2013). But, contrary to SS, at all three time points, proline content was higher in M54-infected leaves than M38-infected leaves. It was previously discussed that proline could be associated with avirulent pathogen recognition (Zeier 2013). Potentially, higher proline content in M54-infected leaves could be linked to previously observed lower effect of the M54 genotype on SS and MDA levels. Therefore, proline levels reflect the FDp abundance in infected leaves better than the previously assessed parameters. Even though proline and soluble sugars both primarily act as osmolytes and have a synergistic effect in battling various abiotic stresses (Suprasanna et al. 2015), their dynamics in FDp-infected grapevine was dramatically different. This could possibly be due to the specific impact of FDp infection on sugar metabolism, which had a specific impact on grapevine SS content. Therefore, in this case, SS probably do not have a substantial protective role in battling FDp infection in grapevine. However, proline accumulation could possibly alleviate some symptoms of FDp-induced damage to grapevine. Nevertheless, both of those metabolites could be used as biomarkers of FDp infection.
3. 4. Differential effect of FDp genotypes on HO production in grapevine
Role of hydrogen peroxide (H2O2) in phytoplasma-infected plants and its potential involvement in recovery has been heavily discussed (Gambino et al. 2013; Musetti et al. 2007). Musetti et al. (2007) hypothesised that H2O2 appears to be involved in counteracting phytoplasma virulence. Indeed, higher content of H2O2 in infected leaves at the first time point in comparison to the uninfected leaves points to the protective role of H2O2, especially active at the earlier stages of the infection (Fig. 2i). However, at the later time points, except in the M38-infected leaves at the third time point, H2O2 content was lower in infected leaves than uninfected leaves. Lower H2O2 content in FDp-infected leaves in comparison to the uninfected grapevine leaves was previously observed (Gambino et al. 2013). The same study recognised that overexpression of some H2O2 scavenging genes (APX2, POX4, POX5) reduced the quantity of H2O2. Similarly, a recent study also confirmed that symptomatic FDp-infected leaves of grapevine cultivar ‘Tocai friulano’ also accumulate less H2O2, accompanied by the upregulation of some H2O2 scavenging-related genes (Casarin et al. 2023). Our hypothesis is that this overexpression could potentially be mediated by certain FDp effectors. This is supported by a negative correlation (r = -0.86) between relative FDp DNA abundance and relative H2O2 content. Therefore, higher FDp DNA abundance (i.e., phytoplasma abundance in infected leaves) caused the higher reduction in H2O2 synthesis, potentially mediated by some FDp effector molecules inducing the overexpression of H2O2 scavenging genes at later stages of the infection. This mechanism reduces the oxidative damage to FDp and enables higher level of FDp replication, supported by the fact that the more abundant M54 genotype caused the higher reduction of H2O2 content. Interestingly, in M38-infected leaves at the third time point, H2O2 content was higher than the control and M54-infected leaves, pointing to the higher ability of ‘Pinot gris’ leaves to overcome the proposed impact of M38 genotype, compared to its ability to battle the effect of potential M54 effectors.
3. 5. Impact of FDp infection on photosynthetic pigments
Total chlorophyll content was the highest in uninfected leaves at all three time points during the development of infection, with M38-infected leaves having the lowest total chlorophyll content at all three time points (Table 1). Considering that the leaf chlorosis is a common symptom of FDp infection in many grapevine varieties (Teixeira et al. 2020), observed severe reduction in photosynthetic pigment content was expected. Even though initial studies argued that reduction of photosynthetic pigments could be linked to the increase in chlorophyllase activity in phytoplasma-infected grapevine (Bertamini and Nedunchezhian 2001), more recent studies concluded that no increase in pigment degradation is evident. Rather, general decrease in influx of chlorophyll precursors causes this severe reduction in photosynthetic pigments (Teixeira et al. 2020). Evidence point to leaf yellowing being a result of sugar accumulation in leaves infected by various phytoplasmas, leading to the metabolic feedback inhibition of photosynthesis, as previously reviewed (Dermastia et al. 2019). It is accompanied by a downregulation of genes involved in both light- and dark-dependent reactions and other genes related to photosynthetic processes in FDp-infected grapevine (Prezelj et al. 2016). In both M38 and M54-infected leaves, relative total chlorophyll content was the highest at the first time point (Supplementary material 2). In M38-infected leaves there was no statistically significant difference between relative chlorophyll content at the second and third time point. However, in M54-infected leaves, there was an increase in relative chlorophyll content at the third time point compared to the second time point. In fact, M54 had a lower impact on total chlorophyll content at all three time points. Potentially, M54 induced a higher accumulation of protective compounds by FDp effector proteins that protected photosynthetic pigments from UV-stress, which enabled an increase in their content. Similar dynamics was observed for chlorophyll a and b. For chlorophyll a, there was also an increase in its content at the third time point in M38, as well as in M54-infected leaves. On the other hand, in M38-infected leaves, relative chlorophyll b content dropped at the third time point, while in M54-leaves it increased. This opposite dynamics of chlorophyll a and b content at the third time point in M38-infected leaves probably lead to the total chlorophyll content being unchanged between second and third time points. Chlorophyll a/b ratio, however, had the same dynamics for M38- and M54-infected leaves, with it being reduced in infected leaves at all time points. This points to the higher impact of FDp infection on chlorophyll a, than chlorophyll b. Similarly, Teixeira et al. (2020) initially recorded a drop in chlorophyll a, but not chlorophyll b, pointing to its lower susceptibility to phytoplasma infection. Severe reduction in chlorophyll content of FDp-infected leaves causes inevitable reduction in photosynthetic performance. Importantly for agricultural production, this causes reduction in yield and drastic production losses in FDp-infected grapevine, as previously reported (Oliveira et al. 2020). Considering that chlorophylls are the major class of porphyrins, similar dynamics observed for chlorophylls was observed for the levels of total porphyrins. Interestingly, a negative correlation (r = -0.67) was observed between absolute MDA levels and porphyrin content, highlighting the impact of oxidative stress in FDp-infected grapevine on the reduction in levels of porphyrins. Carotenoids (both total and individual) content also exhibited the same trend, with M38 genotype impacting carotenoids more severely and the lowest content recorded at the second time point for both M38- and M54-infected leaves, when the highest FDp titre was detected. Furthermore, ratio chlorophylls/carotenoids was reduced in FDp-infected leaves.
However, no linear correlation between relative values of photosynthetic parameters and relative FDp DNA abundance was observed (Supplementary material 1), despite the obvious association between the dynamics patterns of those parameters. The reason behind this is that in M54-infected leaves, which had higher FDp DNA abundance, lower reduction of photosynthetic parameters was observed, and vice versa for M38-infected leaves, which prevented the establishment of a linear correlation. Therefore, separate correlation analyses were performed correlating the FDp DNA abundance with photosynthetic parameters for leaves infected with two separate genotypes independently (Supplementary material 3). Those results established a high positive correlation between the dynamics of all photosynthetic parameters and FDp DNA abundance. This indicates that dynamics of photosynthetic pigments in FDp-infected leaves is highly dependent of phytoplasma titre. However, there is an evident role of the FDp-genotype in the severity of the impact on photosynthetic parameters. Results point that M54 causes less severe changes in infected grapevine than M38 on the level of photosynthetic pigments. Considering that M54 caused a lower accumulation of SS than M38, and had a lower impact on photosynthetic pigments, because the reduction of pigments is caused by a metabolic inhibition loop, lower increase of SS potentially caused lower reduction of pigments in M54-infected grapevine. All photosynthetic pigment parameters mutually correlated, signifying that repression of their synthesis in FDp-infected leaves is probably trackable to the genes involved in the biosynthesis of mutual precursors of both chlorophylls and carotenoids, early in the isoprenoid biosynthetic pathway. Previous findings combining metabolomic and expression data (Teixeira et al. 2020) also point to a general repression of isoprenoid biosynthetic pathway, by the downregulation of key genes. However, further studies should focus on investigating the exact mechanism of FDp impact on photosynthetic pigments and take into consideration the impact of different FDp genotypes.
3. 6. Chemometric analyses
Principal component analysis (PCA) and hierarchical clustering were used to establish the interrelationship between groups based on tested physiological parameters (Fig. 4). Unsurprisingly, PCA revealed a clear separation of uninfected leaves based on the evaluated physiological parameters according to the principal component 1 (PC1). Photosynthetic pigments highly attributed to the separation of uninfected leaves from the infected (Fig. 4b), presumably because of the high impact of FDp on their content in infected leaves. Based on PC1, M54-infected leaves positioned more closely to the uninfected leaves, than M38-infected leaves at the June time point, pointing to the lower impact of M54 at the beginning of the vegetative season based on the level of tested parameters. AA levels attributed to the grouping of June-positive groups together with Sep-M54, presumably due to the highest AA levels observed at this time points. Results were also confirmed by the hierarchical clustering of the groups presented along with the heatmap based on the tested parameters (Fig. 4c). At the August time point, both M54- and M38-infected leaves grouped together based on both PC1 and PC2. However, at the September time point, M54- and M38-infected leaves clearly separated from each other, with M54-infected leaves remaining close to its position at the August time point, while M38-infected leaves separated more severely from its August time point. Both PCA and hierarchical clustering results significantly point to the lower impact of M54 infection on assessed grapevine physiological parameters. Particularly interesting is the high M38 impact at the end of the vegetative season considering the clear separation of this group, while the M54-induced changes to the parameters appear to be less-severely modified at the latest analysed time point. This could possibly be linked to the higher drop in FDp titre in M54-infected leaves, than in M38-infected leaves.