Eight of these genes were strongly upregulated in all four treatments, indicating that B. cinerea may produce these toxins regardless of the ripening stage of the fruit. F. acuminatum genes annotated with this GO term included enzymes involved in the biosynthesis of the toxin fumonisin. Several of these genes showed significant upregulation in infections of MG fruit at 1 dpi or RR fruit at both time points. Fungal proteolysis-related genes were found to be enriched during MG and RR inoculations with F. acuminatum as well as RR inoculations with R. stolonifer at 3 dpi. Though not enriched, several genes with this GO term were also found to be expressed during fruit inoculation by B. cinerea, mostly in RR fruit. Across all treatments, F. acuminatum was found to produce 28 genes with this GO term, while B. cinerea was found to produce 29, and R. stolonifer produced 44 in RR fruit alone . Seven members of the B. cinerea aspartic proteinase family were upregulated in at least one of the fruit inoculations, though none were upregulated at 1 dpi in RR fruit. Thus, fungal proteases are likely to be a strategy used by all three pathogens. Other GO terms served as a proxy for successful growth. Enrichments of genes involved in protein translation initiation , glycolytic process , and DNA replication initiation were found in compatible interactions with RR fruit. Notably, DEGs involved in glycolytic process were enriched in MG inoculations for F. acuminatum at 1 dpi, which is consistent with visual observations of mycelium growth on inoculated fruit. A similar pattern was observed for the chitin catabolic process term, vertical aeroponic tower garden which are involved in the continuous fungal cell wall remodeling during hyphal growth . Multiple GO terms relating to carbohydrate metabolism were found to be enriched across multiple fruit inoculation treatments.
The corresponding genes included those involved in breakdown of the cell wall polysaccharides, metabolism of host sugar sources, and production of fungal polysaccharides. As both the cell wall properties and sugar biochemistry differ between MG and RR fruit, we hypothesized that the fungi employ different classes of CAZymes depending on the ripening stage as already demonstrated for B. cinerea . To test this, we examined the expression profiles of CAZyme families among the DEGs for each pathogen . CAZyme families involved in catabolism of cellulose, hemicellulose, pectin, and monosaccharides were detected, along with families with non-carbohydrate substrates and several responsible for polysaccharide biosynthesis. In B. cinerea and F. acuminatum inoculations, families involved in the degradation of cellulose and hemicellulose were more prominent during infections of MG fruit than RR fruit. Moreover, the CE5 family, which contains cutinases and acetylxylan esterases, was also especially utilized at 1 dpi in MG fruit. In B. cinerea, this family included the cutA gene previously shown to be expressed in tomato fruit infection . MG infections also exhibited higher percentages of families involved in the degradation of cellobiose, a disaccharide of β-1,4-linked glucose molecules that results from the breakdown of cellulose and glucan-based hemicelluloses. A similar trend was found for pectin-degrading families, particularly polygalacturonases and pectate lyases , though the PL1-4 subfamily appeared to be prominent in RR infections as well. Enzymes involved in metabolism of simple sugars, most notably GH32 in F. acuminatum and AA3- 2 in B. cinerea, showed greater prominence in RR infections. Chitin and chitosan biosynthesis and processing families were also detected in B. cinerea and F. acuminatum. In B. cinerea, chitin synthases were generally equally expressed in all fruit inoculations, though chitin deacetylases , which produce chitosan, were only particularly prominent in RR infections at 3 dpi.
In contrast, F. acuminatum produces multiple CE4 enzymes at 3 dpi in MG infections in addition to RR infections. CE10 enzymes were especially prevalent in both B. cinerea and F. acuminatum infections. Members of the CE10 family include lipases, which catalyze the hydrolysis of fatty acids. The previously described B. cinerea gene lip1 was upregulated at both 1 and 3 dpi in MG fruit, but not RR fruit. Additionally, B. cinerea and F. acuminatum both produced multiple AA7 family enzymes in both MG and RR fruit. Many of these genes showed significant similarity to three genes of the PHI database: ZEB1 in F. graminearum, CTB5 from Cercospora nicotianae, and sol5 from Alternaria solani. Each of these PHI genes is involved in the biosynthesis of polyketide mycotoxins in those plant pathogens . Thus, these B. cinerea and F. acuminatum genes may be involved in similar roles. Detection of CAZymes during infection by R. stolonifer was only possible in RR fruit due to the low number of DEGs determined in MG fruit. However, sizable numbers of genes from families detected in B. cinerea and F. acuminatum infections were also discovered in R. stolonifer. These include xyloglucanases , cellobiose dehydrogenases , pectin methylesterases , and polygalacturonases . In addition, multiple enzymes involved in chitin/chitosan biosynthesis were prevalent in RR fruit inoculations, which is indicative of the particularly aggressive hyphal growth of R. stolonifer on these fruit. In RR fruit at 3 dpi, R. stolonifer also produced six enzymes of the AA1 family, which consist of laccases, ferroxidases, and multi-copper oxidases. Each of these enzymes showed significant similarity to FET3 enzymes from Colletotrichum graminicola in the PHI database and to genes of the TCDB class 2.A.108.1.4, the latter being iron transport multicopper oxidase FET5 precursors.
This finding is also consistent with the enrichment of transmembrane transport genes during RR infection at 3 dpi for R. stolonifer.We inoculated fruit of the non-ripening tomato mutant to verify the effect of the ripening stage on the infection success of B. cinerea, F. acuminatum and R. stolonifer. Fruit from the nor mutant do not show ripening-associated processes, such as carotenoid and sugar accumulation or cell wall disassembly, and therefore resemble wild-type MG fruit even at a comparative RR-like stage. None of the three pathogens were able to infect nor fruit at any ripening stage . No hyphal growth of B. cinerea and R. stolonifer was apparent, whereas F. acuminatum formed visible mycelia especially at 3 dpi in MG and RR-like fruit. Like on wild-type MG fruit, all three fungi induced necrotic rings in nor fruit. When inoculated in RR-like fruit from nor, the three fungi displayed similar growth and morphology as in MG fruit from wild-type and nor, indicating that for compatible interactions to occur, tomato fruit needs to undergo certain ripening processes that facilitate fungal colonization and spread.Successful infections of B. cinerea, F. acuminatum and R. stolonifer in tomato fruit are dependent on the host developmental stage. In MG fruit, all three fungi were able to grow on the surface, but none of the pathogens was able to cause rot. In contrast, ripe tomato fruit represented a compatible system for infections as all three fungi induced lesions that spread rapidly. This contrasting ability to cause disease in fruit has been previously reported for a variety of fungal pathogens, particularly those displaying necrotrophic behavior . The tomato gene SlWRKY33 has been shown in leaves to be expressed in response to B. cinerea inoculation, vertical gardening in greenhouse and deletion of this gene leads to increased susceptibility, indicating its role in defense response . We demonstrated that, as in leaves, B. cinerea is capable of inducing SlWRKY33 in MG and RR fruit. Furthermore, F. acuminatum also induced SlWRKY33 in MG and RR fruit, and R. stolonifer did so substantially in RR fruit. These findings indicate that all three pathogens triggered disease responses in the host and that the strength of the response was reflective on the success of the infection process. During interactions with tomato fruit, B. cinerea, F. acuminatum, and R. stolonifer employed a variety of pathogenicity and survival strategies that involved redox processes, carbohydrate catabolism, and proteolysis. Moreover, the degree to which particular strategies were used varied according to the ripening stage of the fruit, as certain processes were emphasized in either MG or RR fruit . These observations suggest that the fungi can sense the physiological environment of the fruit and react accordingly with suitable infection, growth, or quiescence strategies. Though these fungi are incapable of causing disease symptoms in MG tomato fruit, this study demonstrates that they do make attempts to either establish infections or create a suitable environment in fruit for fungal growth and do not merely die on the host tissues. However, when the conditions in fruit are highly unsuitable , the infection strategy of the fungal pathogen is often insufficient to cause successful infections. In many cases, when fungal pathogens encounter incompatible conditions, like in unripe fruit, they enter a quiescent phase with limited growth and activity . During ripening, the physicochemical properties of the fruit tissues change, resulting in compatible conditions for the fruit-pathogen interaction and the reactivation of quiescent pathogens . In this context, it would also be interesting to investigate the strategies employed by the three fungi during inoculations of other plant organs such as leaves.
Our initial tests, however, indicated that both F. acuminatum and R. stolonifer are incapable of infecting tomato leaves even when leaves were senescing. This observation may suggest that the isolates of these two fungi are exclusive fruit pathogens and lack the molecular toolset to grow on leaves. The redox environment of the plant-pathogen interface influences the outcome of the interaction. Upon pathogen detection, ROS are rapidly produced by the host, triggering a downstream signaling of various defense responses . The enzymatic agents of this oxidative burst are respiratory burst oxidative homologs , which generate superoxide O − 2 in the apoplast . This oxidative burst has been previously reported in incompatible tomato-Botrytis interactions , including MG fruit, in which the appearance of a necrotic ring is associated with resistance to B. cinerea . However, necrotrophic pathogens can exploit this ROS response by overwhelming the host with their own ROS production . In leaves of French bean , B. cinerea has been shown to produce ROS as virulence factors by activating the NADPH oxidases BcnoxA and BcnoxB , coupled with the regulatory protein BcnoxR . Although we did not detect strong upregulation of these genes during inoculation of fruit, other ROS producing systems, including laccases and glucose oxidases were upregulated during inoculations of tomato fruit. In F. acuminatum, a BcnoxA homolog FacuDN4838c0g1i1 and BcnoxB homolog FacuDN3221c0g1i1 were induced in specific treatments. A BLAST search did not reveal anyhomologs of BcnoxA or BcnoxB in R. stolonifer, nor were any homologs of Bclcc8 or BcGOD1 detected in either F. acuminatum or R. stolonifer. In addition to ROS generation machinery, fungal pathogens must protect themselves against the oxidative stress of the infection site. Methods of ROS scavenging in phytopathogenic fungi include enzymatic and non-enzymatic mechanisms . SODs catalyze the conversion of O− 2 produced by RBOHs into the less reactive hydrogen peroxide . B. cinerea mutants lacking the BcSOD1 gene have been shown to have reduced virulence on tomato leaves . In tomato fruit, BcSOD1 is upregulated for both MG and RR ripening stages, which suggests it is also a critical gene for fruit colonization. H2O2 can be converted to water by either catalases or peroxidases such as GPXs or PRXs. All three pathogens demonstrated upregulation of specific mechanisms of catabolizing H2O2, but only F. acuminatum showed enrichment of genes involved in the H2O2 catabolic process. The usage of these H2O2 catabolizing systems varied between the pathogens. While B. cinerea utilized catalases in MG fruit at 1 dpi, F. acuminatum and R. stolonifer produced more catalases and peroxidases in RR fruit at 1 dpi. In each pathogen, multiple genes involved in protein degradation were found to be upregulated during fruit inoculations. The strong enrichment of proteolysis-related genes may indicate that protein degradation is important for pathogenicity of F. acuminatum and R. stolonifer but not B. cinerea. Some pathogen-derived proteases, such as Sep1 and Mep1 in Fusarium oxysporum, are known to serve as suppressors of host-immune response in plant-pathogen interactions . Even though their specific roles in pathogenesis are not fully characterized, several aspartic proteinases in B. cinerea have been described . Three of the B. cinerea aspartic proteinases that we found to be induced in tomato fruit were also found to be upregulated during infection of grape berries . Aspartic proteinases were also found to be among the upregulated proteinases in F. acuminatum and R. stolonifer , though all three pathogens appeared to utilize a diverse suite of proteinases of different families.