The now frequent, reoccurrence of foodborne illness cases associated with consumption of fresh produce requires an in-depth assessment of environmental factors that increase the risks of continued outbreaks. In our current study, we examined the population dynamics of a foodborne pathogen through an entomological perspective, analyzing the tri-trophic interactions between S. enterica, plants, and phytophagous insects. Specifically, we investigated how changes in the phyllosphere resulting from unique insect feeding styles, impacted the longevity and persistence of S. enterica populations on the leaf surface.
Previous literature demonstrated that insects can manipulate human enteric bacterial pathogen populations directly, and indirectly. Within poultry dominated environments, cockroaches may mechanically transmit S. enterica by traversing from contaminated egg surfaces to uncompromised substrates, consequently facilitating the movement of bacteria25. Seaweed flies, intimately associated with decaying and pathogenic seaweed beds, excrete viable bacterial populations within intertidal zones, enhancing the potential transmission of E. coli26. Within commercial agriculturally centered models, M. quadrilineatus enhanced transmission of S. enterica from contaminated leaves to clean leaves or adjacent plants13. Furthermore, excretion of viable S. enterica from M. quadrilineatus has also been documented27. Yet, how phytophagous insects influence this increase of S. enterica persistence on leaves remains mostly unexamined.
In earlier studies, we observed that only M. quadrilineatus infestation led to an increase in S. enterica persistence, but no observed benefit occurred following M. persicae infestation8. The findings of this investigation point towards differences between the inter- and intracellular penetrative styles of feeding between these two taxa, and the resulting effects these styles may hold for S. enterica population dynamics within the phyllosphere. While both insects possess similar mouthpart structures, collectively referred to as stylets, their modes of reaching vascular tissues are very distinct. Aphids, or intercellular feeders, begin probing at the junction of two epidermal cells and guide their stylet through intercellular spaces in the mesophyll and towards vascular bundles28. Leafhoppers, considered as intracellular feeders, similarly begin feeding at a cell junction, but distinctly pierce through leaf mesophyll to reach the phloem29. Comparisons of the electrical conductivity response of leaflets infested by inter- and intracellular penetration revealed that M. quadrilineatus infestation elicits a greater magnitude of electrolyte leakage, and consequently greater cellular damage than M. persicae on tomato plants (Fig. 1; Supplemental Fig. S1). Furthermore, our current study demonstrated that plants contaminated by S. enterica and infested with M. quadrilineatus had the highest overall populations of bacteria and resulted in the greatest magnitude of electrolyte leakage (measured as electrical conductance) (Fig. 3a-b). In addition to their distinct feeding behaviors, leafhoppers possess a stylet bundle 5-times wider than those found on aphids30,31. To compensate for the lesser stylet, we investigated the influence of higher aphid populations in a complementary experiment, yet found no measurable impact on enhanced electrolyte leakage, or cellular damage (Supplemental Fig. S2). Taken together, the wider stylet paired with intracellular lacerating types of feeding behavior by M. quadrilineatus may partially explain the enhanced magnitude of cellular damage on the phyllosphere of tomato plants (Fig. 1). These findings lead us to conclude that cellular damage induced by M. persicae probing behaviors does not manipulate the phyllosphere to the same extent as M. quadrilineatus.
As previously mentioned, S. enterica and M. quadrilineatus co-habitation on the same leaflet resulted in higher S. enterica populations and measured electrolyte leakage (aka cellular damage) compared to water inoculated leaflets with or without insects (Fig. 3a-b). These elevated levels of cellular damage are likely the result of greater probing frequencies and may indicate an unfavorable feeding environment for the insect, prompting them to more frequently probe and search for alternative food sources. Previous studies identified clusters of gustatory neurons, which when combined, functionally create taste receptors within insects32. When encountering food contaminated by lipopolysaccharides (LPS), a ubiquitous component found on gram-negative bacterial cells, Drosophila melanogaster not only avoids E. coli-contaminated foods but also commence a hygienic grooming regimen33. This prompted behavior suggests that some insects can discriminate between LPS contaminated and non-contaminated food sources via gustatory cues. Although many of these studies focus on insects with sponging-sucking mouthparts, such as flies, a genome analysis identified both odorant and gustatory receptor genes in aphid and mosquito genomes, both of which possess piercing-sucking mouthparts comparable to that of M. quadrilineatus34. In our experiments where we confined M. quadrilineatus and S. enterica together in more proximal environments, we propose that the adult leafhoppers could encounter higher traces of LPS and may modify their normal feeding behavior as a consequence. Due to the restricted movement in these instances, we surmise the heightened magnitude of electrolyte leakage is driven by a constant search for a non-contaminated substrate and thus, heightened occurrences of probing for a new food source on S. enterica inoculated plant (Fig. 3b). To further evaluate whether S. enterica presence alters M. quadrilineatus’ movement, we provided M. quadrilineatus with contaminated (S. enterica) and non-contaminated (sterile water) tomato leaflet surfaces and monitored their resting or feeding locations every 15 minutes thereafter for over a two-hour period. Throughout the time course of these observational experiments, a pattern of substrate discrimination occurred (Supplemental Fig. S8). Most insects initially landed on the plastic container housing the experiment, but over time began to immigrate more often to water-inoculated surfaces than those with S. enterica. In a complementary experiment, adult M. quadrilineatus exposed to tomato leaflets inoculated at either tip or basal regions of leaves similarly preferred water inoculated regions at 2 hours post exposure, but predominantly emigrated to the experimental container walls after 48 hours (Fig. 5). Altogether, leaflets entirely or partially inoculated with S. enterica were less frequently visited at the last time point (72 hours post infestation), whereas leaflets inoculated solely with water were occupied throughout the experiment. Contrasting this behavior, apterous M. persicae exhibited no preference between S. enterica or alternative surfaces (Supplemental Fig. S9). This lack of substrate preference may result from the largely sessile lifestyle of aphids, in contrast to more mobile and alate leafhoppers. These avoidance behaviors by M. quadrilineatus in response to S. enterica inoculated leaflets suggest a capability of recognizing contaminated substrates similar to the responses described for D. melanogaster.
To evaluate the extent by which M. quadrilineatus might influence the distribution of bacterial populations across leaflets, we first defined the distribution of S. enterica and the magnitude of electrolyte leakage across tomato and lettuce leaves in the absence of any insects. Morphological features between pre-reproductive lettuce and tomato plants are vastly distinct and were hypothesized to impact the distribution of bacterial populations and electrolyte leakage. In our study, the leaf tips were the lowest positioned part of tomato leaflets and exhibited half a log higher S. enterica populations in comparison to basal regions (Fig. 2b). Here again, the nominal architecture of tomato leaves results in a natural ‘drooping’ of fully expanded leaf tips. In a complementary experiment, tomato leaflets were modified to reverse this normal positioning of leaf tips to basal regions, and we did observe a corresponding re-distribution of S. enterica where accumulations were enhanced on basal portions of leaves (Supplemental Fig S4). These findings suggest that during the application of an aqueous solution – such as contaminated irrigation water or even foliar-applied crop inputs – factors including gravitational force may influence aggregations of aqueous solutions on leaves35. This suite of findings identified leaf positioning and morphology, in conjunction with gravitational forces, as dominant influences of S. enterica population distribution across tomato leaflets while unaffecting the degree of electrical conductivity estimates, or associated electrolyte leakage of leaf electrolytes (Supplemental Fig. S3b). Despite S. enterica populations being highest at the tips of unaffected leaflets, bacterial populations were comparable at the tip and middle portions of leaflets only after adult M. quadrilineatus infestation, suggesting an insect mediated influence (Fig. 4). To this finding, we hypothesized that leafhopper feeding is not uniform or homogeneous across whole leaflets and that the distribution of leaf vascular bundles may influence where adult leafhoppers find preferential feeding sites. The diameter of primary and secondary angiosperm vascular bundles typically narrows from the base to the tip of leaves, presumably to maximize the efficiency of hydraulic conductivity using adhesive and cohesive forces36,37. This natural tapering of vascular structures at the tips of leaves provides piercing-sucking insects with some limitations in the number of ideal feeding locations and we hypothesize that the variation in the dendritic nature of leaf venation may alter the distribution of M. quadrilineatus feeding sites, explaining the higher S. enterica populations in the middle of infested tomato leaflets38. In addition to frequently observing leafhoppers in middle portions of leaflet regions, salivary sheathes were also predominantly found in similar regions of water-inoculated leaflets indicating preferences for these vascular bundles across leaflets (Fig. 6). Despite being their preferred feeding site, S. enterica inoculation at the middle of leaflets appeared to influence adult M. quadrilineatus towards feeding at the non-contaminated basal and tip regions, away from the S. enterica middle regions (Fig. 5). Similarly, leaflets partially inoculated at the base and tip had the least amount of salivary sheathes at their base and tip, respectively. This consistent pattern of probing avoidance of contaminated regions suggests that M. quadrilineatus may exhibit discriminatory behaviors against leaflets where S. enterica was present, indicating that even limited exposure to S. enterica holds potential to alter natural feeding behaviors as seen on water inoculated leaflets.
Although M. quadrilineatus exhibited avoidance behaviors of partially inoculated leaflets, their mobile lifestyle illustrates their potential as a biological multiplier for S. enterica. During their exposure to partially inoculated leaflets, salivary sheathes were identified at the base, middle and tip, although nonuniformly, suggesting an exploratory behavior (Fig. 6). This movement across S. enterica contaminated leaflets and the subsequent aversion suggest a likelihood for emigrating to alternative food sources (Fig. 7). Logically, if M. quadrilineatus have previously encountered S. enterica contaminated leaves or plants, then mechanical transmission of bacteria could further exacerbate the likelihood of S. enterica dissemination within contaminated agricultural crops and promote the possibility of food borne outbreaks.
Within this study, we aimed to characterize insect feeding behaviors which could directly enhance S. enterica populations on tomato leaflets. Although we directly focused on cellular damage by stylet penetration, a suite of other phenomena (i.e. honeydew production and plant immunity regulation) occurring in tandem necessitate further investigation. While these biological factors likely co-occurred, we identified prominent insect-mediated interactions involving cellular damage, unique insect feeding behaviors, and S. enterica populations, thereby demonstrating intracellular stylet penetration by M. quadrilineatus as a beneficial insect behavior for S. enterica persistence. Furthermore, we demonstrated that plant morphology directs the distribution of bacterial populations when dispersed aqueously yet may be manipulated in the presence of M. quadrilineatus due to increased stylet probing at preferred feeding sites. Although our results were collected under laboratory conditions, our findings elucidate how insects interact within the phyllosphere, and in turn, influence S. enterica population dynamics.