Even if honey bees displace wild insects at the flower scale , this is unlikely to scale up to the field, as indicated by our data, if mass flowering crops provide floral resources in excess of what can be exploited by local pollinator populations. Therefore, although insect pollinators appear not to be limited by crop floral resources, yield was commonly pollen limited, as crops set more fruit in fields with more visitation by pollinators . Contrary to the fourth prediction, fruit set increased with flower visitor richness independently of honey bee visitation . Correspondingly, the CVs of fruit set decreased with richness; in contrast, evenness did not affect the mean or CV of fruit set . Visitation by wild insects increased strongly with richness , and improved model fit , even when richness was included in the model . However, richness did not enhance model fit when added to a model with wild-insect visitation , suggesting that the effects of richness on fruit set reflect increased wild-insect visitation . Like wild-insect visitation , richness did not correlate with honey bee visitation . Previous studies have shown that agricultural intensification reduces both species richness of pollinator assemblages and wild-insect visitation . Our results for multiple crop systems further demonstrate that fields with fewer species experience less visitation by wild insects and reduced fruit set, independent of species evenness or honey bee visitation. Globally, plastic gutter wild-insect visitation signals both species richness and pollination services, and is a relatively inexpensive indicator that can be standardized easily among observers in field samples .
Large, active colonies of honey bees provide abundant pollinators that can be moved as needed, hence their appeal for pollination management in most animal-pollinated crops . By comparison, methods for maintaining diverse wild insects for crop pollination are less developed, and research on such pollination services is more recent . Although honey bees are generally viewed as a substitute for wild pollinators , our results demonstrate that they neither maximize pollination, nor fully replace the contributions of diverse, wild-insect assemblages to fruit set for a broad range of crops and agricultural practices on all continents with farmland. These conclusions hold even for crops stocked routinely with high densities of honey bees for pollination, such as almond, blueberry or watermelon . Dependence on a single species for crop pollination also carries the risks associated with predator, parasite and pathogen development . Our results support integrated management policies that include pollination by both wild insects as ecosystem service providers, and managed species, such as honey bees, bumble bees , leaf cutter bees , mason bees , and sting less bees as agricultural inputs . Such policies should include conservation or restoration of natural or semi-natural areas within croplands, promotion of land use heterogeneity , addition of diverse floral and nesting resources, and consideration of pollinator safety as it relates to pesticide application . Some of these recommendations entail financial and opportunity costs, but the benefits of implementing them transcend the supply of pollination services alone and extend to, for example, mitigation against soil erosion, and improved pest control, nutrient cycling and water use efficiency .
Without such changes, the on-going loss of wild insects is destined to compromise agricultural yields worldwide.Half of all fruit and vegetables produced globally are lost each year . While the causes of losses vary by region and commodity, fungal phytopathogens have a widespread role, as 20–25% of all harvested fruit and vegetables are lost to rotting caused by such fungi . In fleshy fruits, this issue is exacerbated because, in general, fruit become more susceptible to fungal pathogens as they ripen . Ripening-associated susceptibility has been demonstrated in multiple commodities including climacteric fruits such as tomato, stone fruit, banana, apple, and pear, as well as non-climacteric fruits such as strawberry, cantaloupe, citrus, and pineapple . The most devastating post harvest pathogens in fruit are those with necrotrophic lifestyles, which deliberately kill host tissue, resulting in rotting. Example pathogens include the model necrotrophic fungi Botrytis cinerea and Sclerotinia sclerotiorum as well as Monilinia spp., Alternaria spp., Rhizopus spp., Penicillium spp., and Fusarium spp. . Plant immune responses against necrotrophic fungi are multilayered, involving recognition of pathogen-associated molecular patterns, such as chitin or chitosan, by pattern recognition receptors, intracellular signaling through mitogen-activated protein kinase cascades, induction of downstream defenses by coordinated activity of phytohormones, particularly ethylene and jasmonic acid , cell wall fortifications, and production of various secondary metabolites and antifungal proteins . However, most defense strategies have been studied in leaves, and their utilization and effectiveness in fruit have been assessed only with single pathogens .
The outcome of any fruit–necrotroph interaction relies on the balance between the presence or induction of defenses and the contributions of susceptibility factors. Though induced defenses are heavily studied in plant immunity, the impact of preformed defenses and susceptibility factors are less researched . Preformed defenses include structural barriers, such as the cell wall and cuticle, and the accumulation of secondary metabolites , while susceptibility factors include the abundance of simple sugars and organic acids or activity of host cell wall modifying proteins . A sufficient understanding of ripening-associated susceptibility requires a characterization of the ripening program’s impact on the ability of the host to express necessary defense genes upon pathogen challenge, the integrity of preformed defenses, and the abundance of susceptibility factors. In this study, we first applied a transcriptomic approach to characterize core tomato fruit responses to three fungal pathogens and changes in gene expression that occur during ripening to promote susceptibility. To identify core responses that are not merely pathogen-specific, we used three pathogens with necrotrophic infection strategies: B. cinerea, Rhizopus stolonifer, and Fusarium acuminatum. Using well-established defense gene classifications, we developed profiles of host defense gene expression responses in unripe and ripe fruit. We then determined the susceptibility phenotypes of three non-ripening mutants: Colorless non-ripening , ripening inhibitor , and non-ripening , which have unique defects in ripening features . After demonstrating that each mutant has distinct susceptibility to disease, we identified ripening genes whose expression changes may impact the disease outcome. By integrating our transcriptomic data and mutant analyses, we found preformed defenses and susceptibility factor candidates associated with B. cinerea infections. Using CRISPR-based mutants, we established that one candidate, the pectin-degrading enzyme pectate lyase, is indeed a disease susceptibility factor in ripe tomato fruit.To characterize tomato fruit responses to fungal infection at unripe and ripe stages, we inoculated fruit with B. cinerea, F. acuminatum, or R. stolonifer spores. Each pathogen successfully infected RR fruit, producing visible water-soaked lesions and mycelial growth by 3 dpi, whereas MG fruit remained resistant and, except in samples inoculated with R. stolonifer, had a dark, necrotic ring around the inoculation sites , a feature of the pathogen response that did not appear in wounded fruit. Thus, MG fruit resistance and RR fruit susceptibility are a feature common to multiple necrotrophic infections. We hypothesized that these susceptibility phenotypes are the result of differences in immune responses at each ripening stage and developmental processes during ripening that alter the levels of preformed defenses and susceptibility factors. First, we assumed that, compared with arobust immune response in MG fruit, RR fruit have a weaker response, consisting of fewer genes induced, less diverse functionality, and absent expression of critical genes. Additionally, blueberry container we predicted that ripening may decrease the expression of preformed defenses and increase the expression of susceptibility factors, which create a more favorable environment for infection.
To test if immune responses to fungal pathogens are compromised in RR compared with MG fruit, we sequenced mRNA from B. cinerea-, F. acuminatum-, and R. stolonifer-inoculated fruit at 1 dpi, an early time point at which either a resistant or a susceptible phenotype becomes apparent. A principal component analysis of the mapped normalized reads for all tomato genes revealed that the major driver separating sample data was the ripening stage , while inoculation status accounted for less of the separation . The one exception to this pattern was the R. stolonifer-inoculated MG samples, which clustered with the healthy and wounded MG samples, suggesting that unripe fruit did not display strong responses to this pathogen and yet remained resistant. However, quantification of normalized pathogen reads confirmed that all three pathogens were detectable at 1 dpi even in MG samples. To identify the responses for each ripening stage common to all three pathogens, we performed a differential expression analysis between inoculated and healthy samples for MG and RR fruit. We chose the healthy samples as controls for these comparisons in order to capture responses to necrotrophic infection, which may share features with mechanical wounding. Of all 34 075 protein-coding genes found in the tomato transcriptome, 9366 were found to be differentially expressed in response to inoculation in fruit at 1 dpi in at least one comparison . Of these, 475 genes were significantly up-regulated in MG fruit in response to all three pathogens, corresponding to the MG core response , whereas 1538 genes formed the RR core response . The MG core response overlapped substantially with the wounding response in MG fruit , which suggests that unripe fruit activate similar functions when responding to pathogen attack and mechanical damage. However, this large overlap is also due to the similarity between the gene expression profiles of wounded and R. stolonifer-inoculated samples as seen in the PCA . In contrast, the lack of a strong wounding response in RR fruit indicates that nearly all RR core response genes were strictly pathogen-related . Downregulated genes in response to infection were largely unique to each pathogen, with only 57 and 225 down-regulated across all three pathogens in MG and RR fruit, respectively, and thus we decided to continue our analysis only on the up-regulated core response genes. Complete lists of gene set intersections of up-regulated and down-regulated genes are in given in Supplementary Table S5. We then assessed the MG and RR core responses for the presence of various well-established gene classifications relatedto pathogen defense, including selected GO terms, KEGG pathways, transcription factor families, hormone biosynthesis, signaling and response genes, and receptor-like kinase genes . For each category, we performed enrichment analyses to identify classifications of particular importance in both MG and RR core responses. A total of 70 defense genes were identified in the MG core response. Interestingly, these were enriched in only two categories: chitin catabolic process and RLK genes. The RR core response was enriched in 13 defense categories, including the plant–pathogen interaction and MAP kinase signaling pathways , secondary metabolite biosynthesis pathways , WRKY and ethylene responsive factor transcription factors, RLKs, and JA biosynthesis. Altogether, 302 defense genes were identified among the RR core response. Thus, in contrast to their respective susceptibility phenotypes, RR fruit appear to mount a more robust and diverse immune response than MG fruit early during inoculation, demonstrating that, contrary to our initial hypothesis, weakened immune responses in RR fruit are not a contributor to ripening-associated susceptibility. However, it is possible that tomato fruit resistance to necrotrophs could be determined by a small number of genes that were exclusive to the MG core response. Out of the 70 defense genes in the MG core response, 27 were not found in the RR core response . These 27 genes are heterogeneous, representing 12 different defense categories. Notable genes in this category include a three-gene cluster of PR-10 family proteins , a chitinase previously identified during infections of tomato with Cladiosporum fulvum , and an ERF active at the onset of ripening . Although these 27 genes were not in the RR core response, most of them were induced during RR infections by one or two of the pathogens studied. Only seven were not up-regulated by any of the three pathogens in RR fruit, including the ERF mentioned above , as well as three RLK genes, two glutaredoxin genes involved in the response to oxidative stress, and a cysteine protease. Given that each of these genes belongs to a large family of genes whose members are often functionally redundant, and their average expression levels in infected MG fruit were fairly low , we consider it unlikely that the lack of these genes in the RR core response contributes heavily to susceptibility.