Factors Required for Adhesion of Salmonella enterica Serovar Typhimurium to Corn Salad (Valerianella locusta)

Transmission of gastrointestinal pathogens by contaminated fresh produce is of increasing relevance to human health. However, the mechanisms of contamination of, persistence on, and transmission by fresh produce are poorly understood. We investigated the contributions of the various adhesive structures of STM to the initial event in transmission, i.e., binding to the plant surface. A reductionist system was used that allowed experimentally controlled surface expression of individual adhesive structures and analyses of the contribution to binding to leave surfaces of corn salad under laboratory conditions. The model system allowed the determination of the relative contributions of fimbrial and nonfimbrial adhesins, the type 3 secretion systems, the O antigen of lipopolysaccharide, the flagella, and chemotaxis of STM to binding to corn salad leaves. Based on these data, future work could reveal the mechanism of binding and the relevance of interaction under agricultural conditions.

ABSTRACT

Salmonella enterica is a foodborne pathogen often leading to gastroenteritis and is commonly acquired by consumption of contaminated food of animal origin. However, frequency of outbreaks linked to the consumption of fresh or minimally processed food of nonanimal origin is increasing. New infection routes of S. enterica by vegetables, fruits, nuts, and herbs have to be considered. This leads to special interest in S. enterica interactions with leafy products, e.g., salads, that are mainly consumed in a minimally processed form. The attachment of S. enterica to salad is a crucial step in contamination, but little is known about the bacterial factors required and mechanisms of adhesion. S. enterica possesses a complex set of adhesive structures whose functions are only partly understood. Potentially, S. enterica may deploy multiple adhesive strategies for adhering to various salad species and other vegetables. In this study, we systematically analyzed the contributions of the complete adhesiome, of lipopolysaccharide (LPS), and of flagellum-mediated motility of S. enterica serovar Typhimurium (STM) in adhesion to Valerianella locusta (corn salad). We deployed a reductionist, synthetic approach to identify factors involved in the surface binding of STM to leaves of corn salad, with particular regard to the expression of all known adhesive structures, using the Tet-on system. This work reveals the contribution of Saf fimbriae, type 1 secretion system-secreted BapA, an intact LPS, and flagellum-mediated motility of STM in adhesion to corn salad leaves.

IMPORTANCE Transmission of gastrointestinal pathogens by contaminated fresh produce is of increasing relevance to human health. However, the mechanisms of contamination of, persistence on, and transmission by fresh produce are poorly understood. We investigated the contributions of the various adhesive structures of STM to the initial event in transmission, i.e., binding to the plant surface. A reductionist system was used that allowed experimentally controlled surface expression of individual adhesive structures and analyses of the contribution to binding to leave surfaces of corn salad under laboratory conditions. The model system allowed the determination of the relative contributions of fimbrial and nonfimbrial adhesins, the type 3 secretion systems, the O antigen of lipopolysaccharide, the flagella, and chemotaxis of STM to binding to corn salad leaves. Based on these data, future work could reveal the mechanism of binding and the relevance of interaction under agricultural conditions.

INTRODUCTION

Salmonella enterica is one of the main bacterial pathogens leading to foodborne illnesses and thousands of fatal cases worldwide (1). Depending on the serovar, S. enterica causes gastroenteritis (nontyphoidal serovar, e.g., Typhimurium) or typhoid fever (typhoidal serovars, e.g., Typhi and Paratyphi). Focus has historically been on infection routes of Salmonella by animal products, although in recent years, an increasing number of infections caused by fresh produce has been reported. In addition to pathogenic Escherichia coli (e.g., E. coli O157:H7) or Listeria monocytogenes, S. enterica is also involved in such plant-associated infections (2–4). Several outbreaks were associated with contaminated vegetables (e.g., tomatoes and salad), fruits (e.g., watermelons and berries), nuts, herbs (e.g., basil), and sprouts (5, 6). Fresh produce can be contaminated either through cultivation (contaminated irrigation water or fertilizer) or during handling and processing. S. enterica may adhere to leaves and roots, colonize the plant, and further internalize into the plant tissue (7, 8). Once inside the plant, S. enterica potentially can replicate and persist (9, 10). Endophytic S. enterica cannot be removed by surface washing, and bacteria will thus be ingested if food is consumed after minimal processing.

While the adhesion of S. enterica to mammalian cells has been investigated in great detail, far less is known about the mechanisms of interaction of S. enterica with plants. Investigation of adhesion to plant surfaces should allow better understanding of contamination and colonization of plant-based products by S. enterica. For the analyses of contamination of salads by S. enterica, the leafy part is of special interest, and the initial binding to salad leaves is a key event in the adhesion to and further colonization of salad. While surface contamination may occur by irrigation water or fecal shedding, a certain degree of adhesion is expected to maintain bacterium-plant association in the production process from “farm to fork.”

In this study, we employed S. enterica serovar Typhimurium (STM) as a model pathogen causing gastroenteritis. STM possesses a large set of adhesive structures, including 12 chaperone-usher (CU) fimbriae, curli fimbriae assembled by the nucleation-precipitation pathway, two type 1 secretion system (T1SS)-secreted adhesins (BapA and SiiE), and three type 5 secretion system (T5SS)-secreted adhesins (MisL, ShdA, and SadA). Further, PagN and Rck are known outer membrane proteins (OMP) with putative adhesive features (reviewed in reference 11).

For most of the 12 CU fimbriae, little is known about their functional surface expression and binding properties (12). All operons encoding CU fimbriae consist of at least a fimbrial main subunit, a specific periplasmic chaperone, and a specific usher located in the outer membrane (13). The most prominent and best-studied fimbriae are Fim fimbriae encoded by the fim operon (fimAICDHF). Fim fimbriae are functionally expressed under static culture conditions and mediate binding to mannosylated proteins (14).

The Salmonella pathogenicity island 4 (SPI4) locus (siiABCDEF, Salmonella intestinal infection) encodes the giant adhesin SiiE, which is secreted to the bacterial surface by the T1SS SiiCDF (15). SiiE is known as the largest protein in STM, with 53 repetitive bacterial Ig (BIg) domains and a molecular mass of 595 kDa. Moreover, SiiE exhibits binding specificity for glycostructures with terminal N-acetylglucosamine (GlcNAc) and 2,3-linked sialic acid (16). SiiE mediates the first contact of Salmonella with polarized epithelial cells of mammalian hosts (e.g., MDCK cells), enabling subsequent invasion mediated by the SPI1-encoded T3SS (here also referred to as SPI1-T3SS) and various effector proteins (17, 18). The bap operon (bapABCD, biofilm-associated protein) encodes a T1SS including BapB (outer membrane protein), BapC (ATPase), and BapD (membrane fusion protein) which is necessary for the secretion of the adhesin BapA to the bacterial surface. The T1SS-secreted adhesin BapA has a molecular mass of 386 kDa, contains 28 BIg domains, and is involved in biofilm formation (19).

In addition, motility and chemotaxis mediated by flagellar rotation, as well as the adhesive effect of the lipopolysaccharide (LPS) layer, must be taken into consideration (11, 20). The specific binding properties of only a few adhesive structures of S. enterica are known, and thus, no educated guess can be made in regard to possible interactions with salad leaves. Several studies have investigated the adhesion of S. enterica serovars to various species of salad (9, 21–28), with a focus on individual adhesion factors. These studies succeeded in clarifying the first steps of colonization using wild-type (WT) strains or mutant strains defective in single adhesion factors. Prior work revealed the involvement of flagella and motility, as well as further virulence-associated genes, in adhesion to salad. Further, the impact of different salad species was evaluated. Yet most studies on plant-pathogen interactions only tested differences in adhesion of one Salmonella isolate to various plant species or adhesion of various Salmonella serovars to one plant species (22, 23, 29).

A major obstacle for many analyses of S. enterica adhesion to vegetables was the lack of surface expression of functional adhesins in order to test their involvement. Indeed, only a minor proportion of adhesins is known to be expressed under laboratory conditions or defined environmental conditions. For example, global transcriptional analyses of STM under 22 defined culture conditions or stress exposure revealed significant transcriptional changes for only 3 of 20 adhesins (30; unpublished observation). It can thus be speculated that a subset of adhesins is expressed under environmental conditions outside of a warm-blooded host organism, although a systematic analysis of such expression is pending. To circumvent this limitation and to functionally express the entire adhesiome of STM, we recently devised a simple and robust approach based on the use of the P_tetA promoter and induction by the nonantibiotic tetracycline (Tet) derivative anhydrotetracycline (AHT) (31). In the present study, we deployed this technique to investigate the contributions of the various adhesive structures of STM to adhesion to the surface of corn salad leaves.

We have analyzed the impact of, to our knowledge, all adhesive structures of STM in adhesion to corn salad (Valerianella locusta). Moreover, we have found factors that are involved in the adhesion of STM to salad. With this knowledge, we are potentially able to devise defensive strategies in growing, harvesting, and processing fresh produce in order to decrease the incidence of Salmonella infections.

RESULTS

We deployed a reductionist, synthetic approach to identify factors that contribute to the surface binding of Salmonella enterica serovar Typhimurium (STM) to leaves of corn salad. As with all S. enterica serovars studied so far, STM possesses a complex adhesiome. We expressed the various operons or genes encoding adhesins ectopically under the control of a tetracycline-inducible promoter, as previously described (31). Strains harboring these Tet-on plasmids were subsequently tested for their contribution to adhesion to Valerianella locusta (corn salad). We selected corn salad as a salad species that can be easily cultured and infected under laboratory conditions, as well as being a representative fresh produce relevant to consumer health. Thus, corn salad served as a model organism in the national research consortium Plantinfect. The infection of corn salad grown under aseptic conditions by STM was performed as described schematically in Fig. S1 in the supplemental material.

Prior to analyzing the contribution of adhesive structures in adhesion to corn salad, we tested different deletion strains for their suitability as a negative control and as a host strain for heterologous expression. The laboratory conditions for native expression of only a few adhesins such as fim fimbriae are known. Moreover, the expression of a fimbrial adhesin can impact the expression of other systems, including other adhesins (32, 33). To avoid potential interference by these factors, we generated a strain lacking all 12 CU fimbriae (SR11 Δ12). Furthermore, a strain was generated lacking all known and putative adhesive structures in SR11 (ΔfimAICDHF ΔstbABCD ΔsthABCDE ΔstfACDEFG ΔstiABCH ΔbcfABCDEFGH ΔsafABCD ΔpefACD-orf5-orf6 ΔstcABCD ΔstjEDCBA ΔstdAB ΔlpfABCDE::KSAC ΔmisL ΔsadA ΔshdA ΔSPI4 ΔbapABCD Δrck ΔpagN ΔcsgBAC-DEFG), which we termed SR11 Δ20. Under the assay conditions, both SR11 Δ12 and SR11 Δ20 showed the same level of adhesion to corn salad as WT SR11 (Fig. 1A). Therefore, we decided to use SR11 Δ12 in all further experiments to avoid any background expression of CU fimbriae during our assays. Furthermore, SR11 Δ12 strains with additional deletions of single adhesive structures showed no altered levels of adhesion compared to SR11 Δ12, except for deletion of SPI4 and of bapABCD (Fig. 1B). The deletion strain defective in SPI4, lacking SiiE, the corresponding T1SS, and accessory proteins, showed increased adhesion (129% on average). The loss of adhesin BapA and its cognate T1SS BapBCD (ΔbapABCD) led to significantly decreased adhesion (65% on average). Of interest, BapA was not detected on the bacterial surface in 3.5-h subcultures of parental strain SR11 Δ12 (Fig. S2C and D).

FIG 1.

Comparison of Salmonella NCTC 12023 WT, SR11 WT, SR11 Δ12, and SR11 Δ20 and impact of deficits in genes encoding putative adhesive structures and effector proteins of SPI1-T3SS STM adhesion to corn salad. Corn salad grown under aseptic conditions was infected with STM NCTC 12023 WT, SR11 WT, SR11 Δ12, and SR11 Δ20 (A) and with SR11 Δ12 with various deletions in genes encoding putative adhesive structures and effector proteins of SPI1-T3SS (ΔsopA ΔsopB ΔsopD ΔsopE2 ΔsipA [Δeffector proteins]) (B). Overnight cultures were diluted 1:31 in fresh LB, and bacteria were subcultured for 3.5 h and diluted in PBS for infection of corn salad. After infection for 1 h, corn salad segments were washed three times to remove nonadherent bacteria. For the quantification of adherent bacteria, corn salad leaf discs were homogenized in PBS containing 1% deoxycholate, and serial dilutions of homogenates and inoculum were plated onto MH agar plates for the quantification of CFU. Adhesion rates were determined by the ratio of CFU in inoculum and homogenate, and adherent bacteria normalized to SR11 Δ12 were set as 100% adhesion. Shown are the distributions of three biological replicates represented as box plots with medians. Statistical significances were calculated with Student’s t test and are indicated as follows: n.s., not significant; *, P < 0.05; **, P < 0.01; and ***, P < 0.001.

Contribution of fimbrial adhesins to adhesion to corn salad.

We analyzed adhesion to corn salad after P_tetA-induced expression of various CU fimbriae (Fig. 2A). The assay revealed distinct phenotypes of binding to corn salad. Expression of certain CU fimbriae by STM (Lpf, Bcf, Sth, Std, and Stj) resulted in adhesion levels similar to that of background strain SR11 Δ12, indicating that these adhesins do not have cognate ligands on corn salad. Adhesion to corn salad was impaired after expression of Fim, Pef, Stc, and Stb fimbriae (53%, 72%, 58%, and 60% mean adhesion rates, respectively, compared to that of SR11 Δ12), while expression of Sti fimbriae resulted in slightly, but not significantly, decreased adhesion. In contrast, AHT-induced expression of Saf and Stf fimbriae led to increased adhesion (166% and 116% mean adhesion rates, respectively). A clear contribution of Saf fimbriae in adhesion to corn salad was confirmed by the noninduced control, exhibiting no altered adhesion level compared to that of background strain SR11 Δ12. Of note, a nonsignificant increase in adhesion was observed for Stf fimbriae in the absence of the inducer AHT, which was comparable to the case with AHT-induced samples. Consequently, a clear role for Stf fimbriae in adhesion of STM to corn salad cannot be ascribed.

FIG 2.

Impact of chaperone-usher fimbriae and curli fimbria expression on STM adhesion to corn salad. Sterile grown corn salad was infected with S. enterica serovar Typhimurium strain SR11 Δ12 with the expression of various chaperone-usher fimbriae (A) and the expression of curli fimbriae (B). Expression of fimbriae was induced with 10 ng/ml AHT for 3.5 h in subculture. The adhesion and statistical significances were determined as described in the legend to Fig. 1.

Curli fimbriae are known to be involved in biofilm formation (34) and are encoded by two divergent operons, csgBAC and csgDEFG, with assembly occurring via the nucleation-precipitation pathway. AHT-induced expression of curli fimbriae showed a decreased adhesion to corn salad, whereas without AHT induction, no altered adhesion was observed (Fig. 2B).

Contribution of T1SS-secreted nonfimbrial adhesins to adhesion to corn salad.

As generation of a vector for Tet-on expression of the sii operon turned out to be problematic, we deployed an alternative approach to control expression of the native sii operon. Enhanced surface expression of SiiE was achieved by AHT-induced overexpression of hilD, the central transcriptional activator of the SPI1/SPI4 regulon (35). We observed that increased amounts of SiiE on the bacterial surface led to decreased adhesion to corn salad (77% mean [Fig. 3A]) compared to that of SR11 Δ12 with native expression of SiiE in 3.5-h subcultures (Fig. S2A and B). Without induction by AHT and therefore with almost natural SiiE expression, no differences in adhesion from that of the background strain SR11 Δ12 were observed. Since the expression of the regulator hilD also influences the expression of the SPI1-encoded T3SS and its effector proteins, the plasmid carrying hilD was tested under the control of the Tet-on system in further SPI1 and SPI4 deletion mutants. Overexpression of hilD in an SPI4 deletion mutant led to a significantly decreased adhesion (53% on average), indicating that the SPI1-T3SS rather than SiiE itself interferes with the adhesion to corn salad. This was further confirmed by an increased adhesion rate of a strain lacking invC (ATPase subunit of SPI1-T3SS), and thereby the SPI1-T3SS, harboring a plasmid for hilD overexpression (153%). The deletion of invC alone, as well as deletion of the effector proteins SopA, SopB, SopD, SopE2, and SipA (Fig. 1B), did not alter adhesion, leading to the hypothesis that the SPI1-T3SS affects adhesion to corn salad.

FIG 3.

Impact of T1SS-secreted adhesins and hilD expression on STM adhesion to corn salad. Corn salad grown under aseptic conditions was infected with STM strain SR11 Δ12 with the overexpression of the regulator hilD for analysis of the SPI4-encoded, T1SS-secreted adhesin SiiE and the SPI1-encoded T3SS (A). In addition, SR11 Δ12 expressing AHT-induced, T1SS-secreted wild-type adhesin BapA or the indicated BapA truncation mutants were tested (B). The adhesion and the statistical significances were determined as described in the legend to Fig. 1. (C) Schematic overview of truncated BapA forms used in adhesion assays.

AHT-induced expression of the bap operon led to increased adhesion to corn salad (124% mean [Fig. 3B]), whereas no significant differences were observed without AHT induction. To gain further insight into which structural features of BapA are essential for adhesion, we generated plasmids for Tet-on expression of bapABCD that encode BapA with deletions of BIg domains to various extents. Synthesis and secretion of truncated forms of BapA were confirmed by flow cytometry (Fig. S2C and D) and indicated that deletion of BIg1-28 and BIg15-24 ablated the surface expression of BapA. This observation is in line with the adhesion assay results for strains expressing BapA harboring a deletion of BIg1-28 or BIg15-24, which showed no increased adhesion to corn salad. Thus, the loss of BapA surface expression resulted in adhesion levels comparable to that of SR11 Δ12. In contrast, truncated forms of BapA with deletion of only one BIg domain, either BIg1 or BIg28, were detected on the bacterial surface by flow cytometry. Moreover, in adhesion assays, no increased adhesion was observed compared to that with wild-type BapA. Hence, the BIg1 and BIg28 domains might be relevant for proper binding to corn salad by BapA.

Contribution of autotransported adhesins to adhesion to corn salad.

STM expresses three autotransported adhesins: MisL, ShdA, and SadA. MisL and ShdA are monomeric adhesins, whereas SadA belongs to the class of trimeric adhesins. Previous studies have shown that MisL and ShdA are involved in binding to fibronectin, which impacts intestinal infection of mice (36, 37). SadA is possibly involved in adhesion to CaCo2 cells, as well as in biofilm formation, but only in a strain background with altered LPS structure (38). AHT-induced expression of misL did not alter adhesion to corn salad (Fig. 4A). In contrast, the AHT-induced expression of shdA led to a decreased average adhesion of 67%, whereas the noninduced strain displayed no changes in adhesion. The AHT-induced expression of sadA and its chaperone sadB led to a slight, but nonsignificantly, decreased adhesion (79% mean). Although we observed significantly higher adhesion (158% mean) without AHT induction, SadA surface expression was not detected by flow cytometry in noninduced samples (Fig. S2E and F).

FIG 4.

Impact of T5SS-secreted adhesins and of outer membrane proteins on STM adhesion to corn salad. (A) Corn salad grown under aseptic conditions was infected with STM strain SR11 Δ12 expressing the different T5SS-secreted adhesins MisL, ShdA, and SadA induced by AHT. (B) For the analysis of outer membrane proteins, SR11 Δ12 expressing rck and pagN by induction of AHT was used. The adhesion and the statistical significances were determined as described in the legend to Fig. 1.

Contribution of OMP adhesins to adhesion to corn salad.

The OMPs Rck and PagN are adhesive structures, and an involvement in SPI1-T3SS-independent invasion of epithelial cells has been reported (39, 40). AHT-induced expression of rck led to a significantly decreased adhesion to corn salad (65% mean), although even the noninduced sample exhibited decreased adhesion (65% mean [Fig. 4B]). In a previous study, Western blot analyses confirmed the absence of expression of Rck in noninduced cultures (31). The AHT-induced expression of PagN exhibited significantly reduced adhesion (59% on average), whereas the noninduced samples showed no altered adhesion level.

Contribution of flagellar filaments and motility to adhesion to corn salad.

The effect of flagella and motility on infection of various plants has been previously investigated for Salmonella and other pathogenic bacteria (25, 41, 42). In this study, we demonstrated the binding properties and the contribution of motility in adhesion to corn salad using four distinct deletion strains. The deletion of fliC and fljB, resulting in the loss of the flagellar filament, yielded a decreased adhesion (50% mean) which could not be restored to background strain level by centrifugation (Fig. 5A). This effect may thus be due to an adhesive feature of the flagellar filament or due to flagellum-mediated motility promoting contact with corn salad surfaces. To dissect the contribution of flagella, a motAB mutant strain was employed; such strains still produce a flagellar filament, but they are unable to energize the flagellar motor and are thus nonmotile. The ΔmotAB strain showed decreased adhesion for static and centrifuged samples (67% and 73% means, respectively). Thus, the presence of flagella without motility does not enable Salmonella to bind to corn salad. To gain further insight into how motility contributes to adhesion to corn salad, we deployed mutant strains with defective cheY, resulting in a strong bias toward smooth swimming, or defective cheZ, resulting in a strong bias for tumbling (Fig. 5C). The ΔcheY strain showed a decreased adhesion (71% mean) after centrifugation, whereas the deletion of cheZ led to a decreased adhesion which did not represent a statistically significant difference in static and centrifuged samples. We conclude that proper flagellum-mediated motility contributes to adhesion to corn salad surfaces and that this effect is not caused solely by the interaction of the flagellar filament with the leaf surface.

FIG 5.

Impact of defect in motility and flagellar assembly and deletion of LPS structure on STM adhesion to corn salad. Corn salad grown under aseptic conditions was infected with STM strain SR11 Δ12 with deletion of various motility and flagellum-associated genes (A) and deletion of LPS structure-related genes (B). The infection took place either under static conditions or after centrifugation at 500 × g for 5 min to compensate effects of mutations in motility genes. For deletion of genes involved in O-antigen (OAg) biosynthesis, only static samples are shown. The adhesion and the statistical significances were determined as described in the legend to Fig. 1. Models of the resulting phenotype depending on the different deletions in motility flagellar assembly and LPS structure are depicted in panels C and D. Panel D is based on reference 46. OS, oligosaccharide.

Contribution of O antigen to adhesion to corn salad.

The major constituent of the Gram-negative cell surface is LPS. In addition to stabilization of the cell envelope and protection against various environmental factors, LPS increases the negative charge of the cell envelope, and a putative adhesive role has been reported (43). To analyze the impact of LPS in adhesion to corn salad, we used mutant strains lacking various genes involved in the biosynthesis of the O antigen of LPS. WT Salmonella displays a heterogeneous distribution of long-chain O antigen (L-OAg) and very-long-chain O antigen (VL-OAg). Deletion of wzz results in the homogenous distribution of VL-OAg, deletion of fepE results in the homogenous distribution of L-OAg, and a strain lacking both genes (wzz fepE) can only synthesize short O antigen (S-OAg) (Fig. 5D). The deletion of rfaL leads to the lack of O antigen, resulting in LPS being restricted to the core oligosaccharides.

In this study, the deletion of wzz and wzz fepE led to a decreased adhesion (49% and 52%, respectively) in static samples (Fig. 5B). The deletion of rfaL yielded a decreased adhesion (82% mean) which did not represent a statistically significant difference. The strain lacking fepE showed no altered adhesion. These data suggest that the presence of only VL-OAg or only S-OAg impairs binding to corn salad, and as a consequence, the L-OAg has to be present. The observation that the rfaL deletion, resulting in a lack of O antigen, led to no significant decrease in adhesion could be explained by binding of the core oligosaccharide to corn salad.

DISCUSSION

To address the question of which factors of S. enterica are involved in adhesion to plant surfaces, we deployed a reductionist, synthetic approach. This allowed controlled surface expression of specific adhesive structures of STM, one at a time. The various adhesive structures were tested for their impact on adhesion to corn salad leaves as a representative fresh produce, and the results of this study are summarized in Fig. 6.

FIG 6.

Overview of the impact of the analyzed factors of STM in adhesion to corn salad. The absence of underlining indicates static samples, and underlining indicates centrifuged samples. Arrows indicate increased or decreased adhesion, and circles indicate that adhesion was not altered. OM, outer membrane; PP, periplasm; IM, inner membrane; CP, cytoplasm.

Several prior studies showed that absence of the flagellar filament had an influence on adhesion to various plants. Whereas Berger et al. (21) reported a decreased adhesion to basil leaves for a ΔfliC ΔfljB strain of S. enterica serovar Senftenberg, Iniguez et al. (44) revealed an enhanced colonization of Arabidopsis thaliana roots for a ΔfliC ΔfljB mutant of STM. Thus, there has to be a clear difference in the role of flagella between colonization of the rhizosphere and of the phyllosphere. For the colonization of roots, the presence of flagella is apparently obstructive, due to pathogen-associated molecular pattern (PAMP)-triggered immunity of Flg22 by receptor kinase FLS2 recognition in A. thaliana (44, 45). For the first contact of S. enterica and other pathogenic bacteria with leaf surfaces, the presence of flagella is of crucial importance. To investigate the possible binding of flagellar filaments, Rossez et al. (42) purified the flagellar filament of pathogenic enterohemorrhagic E. coli (EHEC) O157:H7 Sakai, enteropathogenic E. coli (EPEC) O127:H6, and nonpathogenic E. coli K-12 with flagellar serotype H48. They showed that the binding of purified flagellar filaments to multiple plant lipid species (SQDG [sulfated glycolipid], phosphatidylcholine, phosphatidylglycerol, phosphatidylinositol, and phosphatidylethanolamine) results in the assumption of an ionic adhesion by binding to sulfated and phosphorylated plant plasma membrane lipids with negative charge. In addition, E. coli strain TUV93-0 ΔfliC showed a decreased adhesion to Arabidopsis leaves which could be reversed through complementation by all three flagellar serotypes (42). Possibly, the ionic adhesion of flagellar filaments represents a conserved mechanism for adhesion to plant leaves among Gram-negative bacteria.

Despite analyses of flagellar filament involvement in adhesion to various plant organs, less is known about the impact of motility. Kroupitski et al. (25) showed that deletion of cheY in STM had no consequences for attachment to iceberg lettuce leaves, whereas the internalization of STM was affected. The authors hypothesize that STM cannot reach stomata due to the lack of directed motility. Directed motility conceivably enables STM to sense sucrose near stomata, facilitating internalization. Thus, internalization was impaired during an experiment performed in the dark with fusicoccin-treated leaves, leading to constitutively opened stomata without producing sucrose by photosynthesis (25). In this study, we detected decreased adhesion levels for strains lacking flagellar filaments (ΔfliC ΔfljB) or the energization of flagellar rotation (ΔmotAB), under conditions of either natural contact or forced contact. We therefore conclude that flagellar filaments are needed for not only adhesion to corn salad leaves but also motility. We observed only moderate effects on adhesion to corn salad leaves in the absence of either clockwise (CW) or counterclockwise (CCW) rotation, leading to the assumption that the flagellar filament and energization of at least CW or CCW rotation are necessary for binding to corn salad leaves. However, bacteria might utilize directed motility for accumulation near stomata and/or colonization of plant leaves.

The LPS layer of STM and other pathogenic bacteria was often examined with a focus on adhesion to, and invasion of, mammalian cells and for the impact on inflammatory responses. The impact of LPS on adhesion to plant leaves, roots, and fruits remained unclear. Mutant strains of STM lacking very long OAg, or long and very long OAg, revealed higher levels of invasion of HeLa and MDCK cells, whereas deletion of the whole OAg even led to a highly increased adhesion to both cell lines. Despite this virulence advantage for STM, immune escape was reduced due to higher effector protein translocation (46). In contrast to an enhanced adhesion to mammalian cells due to an altered LPS structure of STM, we found that an altered LPS structure resulted in decreased adhesion to corn salad leaves. Our findings are in line with a study by Jang and Matthews (47) revealing that a truncated OAg in pathogenic E. coli O157:H7 decreases the ability to survive and persist on Arabidopsis plants as well as on romaine lettuce. In addition to pathogenic bacteria, an intact LPS structure is also important in nonpathogenic bacteria, like Herbaspirillum seropedicae, which acts as a symbiont for many agriculturally important plants. An altered LPS structure in H. seropedicae led to decreased attachment to maize root surfaces and to further endophytic colonization (48). These results were also observed for WT H. seropedicae when LPS, N-acetylglucosamine, or glucosamine was added to act as a competitor for binding sites. Here we show the importance of STM LPS in adhesion to leaf surfaces.

Regardless of LPS and motility of STM, adhesion was increased by expression of different adhesins. Saf fimbriae (Salmonella atypical fimbriae) were the only fimbriae of the CU pathway found in this study to enhance STM adhesion to corn salad leaves. Salih et al. (49) revealed by electron microscopy the highly flexible linear structure of Saf fimbriae belonging to FG-loop long (FGL) fimbriae. In contrast to rigid, rod-shaped FG-loop short (FGS) fimbriae, which exhibit various subunits with a distal adhesive tip, FG-loop long fimbriae often display only two subunits (50). Therefore, the adhesive unit is likely formed by the most numerous subunits. Thus, FGL fimbriae, like Saf fimbriae, might bind to a high number of receptors or ligands (49). Nevertheless, binding properties of Saf fimbriae are unknown. Until now, Saf fimbriae were reported to be involved in biofilm formation and in binding to porcine intestine IPEC-J2 cells (51). In addition, expression of Saf fimbriae was observed only during infection of murine spleen (52). Genes of the saf operon are often pseudogenes in host-restricted S. enterica serovars (Typhi, Paratyphi, and Gallinarum) (20), indicating their potential contribution in STM to dispersal by farm animals and newly investigated environmental routes, e.g., leafy plants and other vegetables. To gain further insight into the contribution of Saf fimbria adhesion of STM to plants, binding properties of Saf fimbriae have to be investigated, for example, by glycan arrays (42, 53) or by a detailed mutagenesis of potential binding domains.

In this study, we showed that T1SS-secreted adhesins SiiE and BapA both contribute to adhesion to corn salad leaves. While SiiE involvement in adhesion to mammalian polarized epithelial cells by binding GlcNAc and sialic acid is well understood (16, 17), a potential role for SiiE in adhesion to plant surfaces is less likely. The tight control of expression of the SPI1/SPI4 regulon by host cell factors would exclude surface expression of SiiE under environmental conditions. A contribution to adhesion was shown for T1SS-secreted adhesin BapA, and BapA contributes to biofilm formation (54), especially for formation of pellicles on the air-liquid interface (19). Furthermore, deletion of BapA led to a decreased mortality in mouse infection. Our data obtained after 1 h of infection excluded the possibility of biofilm formation by BapA-expressing STM on corn salad leaves. However, specific binding properties are unknown. To gain further insight into properties of binding of BapA to corn salad leaves, various truncated forms of BapA were tested. Truncated forms of BapA lacking one BIg domain were surface expressed and showed no autoaggregation. Deletion of one or more BIg domains reduced BapA-dependent adhesion. Thus, we propose a diminished adhesion to corn salad leaves by a shortened BapA. This hypothesis is further supported by the fact that deletion of BIg1, possibly never reaching out of the OAg layer in WT BapA, results in a phenotype similar to that with deletion of BIg28, possibly reaching out of LPS layer first in WT BapA. Further characterization of BapA binding to corn salad leaves is necessary, including investigating the importance of proper folding of BapA in the presence of Ca²⁺ (55) and specific binding properties.

This study showed that adhesion of STM to corn salad leaves depends on an intact LPS layer and on flagellum-mediated motility. Further, we revealed the involvement in adhesion to corn salad leaves by expression of CU pathway-assembled Saf fimbriae, T1SS-secreted SiiE, and T1SS-secreted BapA. To gain further insight into adhesion of STM to salad, additional salad species should be investigated to assess if the detected contributing structures are also involved in adhesion to other salad species, or even to leafy plants in general. Moreover, a transcriptomic and proteomic analysis of the involved adhesins could further elucidate environmental conditions or conditions during colonization of plants. We used a synthetic system with controlled expression of one adhesive factor at a time. Whether the adhesive factors determined in this study are also expressed and functional under conditions of natural contamination of plants has to be investigated in further studies.

In summary, this work contributed to identification of STM adhesive factors required for adhesion to plants. To take these studies to a global context and to study the pathogen-plant interaction under field-like conditions, a more complex experimental setting is needed.

MATERIALS AND METHODS

Bacterial strains and culture conditions.

Bacterial strains used in this study are listed in Table 1. Unless otherwise mentioned, bacteria were routinely grown aerobically in LB (lysogeny broth) medium or on LB agar containing antibiotics if required for selection of specific markers. Carbenicillin (Carb), nalidixic acid (Nal), or kanamycin (Km) was used to a final concentration of 50 μg/ml if required for the selection of phenotypes or maintenance of plasmids. Chloramphenicol (Cm) was used at 30 μg/ml. When needed for cloning purposes, 5-bromo-4-chloro-3-indolyl-β-d-galactopyranoside (X-Gal) was added to LB agar at 20 μg/ml. For the induction of the Tet-on system, anhydrotetracycline (AHT) was used at a final concentration of 10 ng/ml or 100 ng/ml.

TABLE 1.

Bacterial strains used in this study

Strain	Relevant characteristics	Reference or source
E. coli CC118 λpir	Cloning strain for λpir-dependent plasmids	60
E. coli DH5α MCR	Cloning strain	61
E. coli NEB5α	Cloning strain	New England Biolabs
E. coli S17-1 λpir	Mobilization strain for plasmids containing oriVR6K and mobRP4	62
S. Typhimurium IR715	Salmonella enterica serovar Typhimurium ATCC 14028s, spontaneous Nalr	63
S. Typhimurium LT2	Wild type	64
S. Typhimurium NCTC12023	Wild type	NCTC
S. Typhimurium SR11	Wild type	65
AJB754	IR715 ΔstiABCH::KSAC	66
AJB786	IR715 ΔstbABCD::KSAC	66
EHW1	IR715 ΔbcfABCDEFGH::KSAC	66
EHW2	IR715 ΔfimAICDHF::KSAC	66
EHW3	IR715 ΔstfACDEFG::KSAC	66
EHW11	IR715 ΔstdAB::KSAC	66
SF22	IR715 ΔsthABCDE::KSAC	66
SPN191	IR715 ΔbcfABCDEFGH	This study
SPN192	IR715 ΔfimAICDHF	This study
SPN193	IR715 ΔlpfABCDE::KSAC	This study
SPN195	IR715 ΔsafABCD::KSAC	This study
SPN196	IR715 ΔstbABCD	This study
SPN198	IR715 ΔstdAB	This study
SPN199	IR715 ΔstfACDEFG	This study
SPN200	IR715 ΔsthABCDE	This study
SPN201	IR715 ΔstiABCH	This study
SPN202	IR715 ΔstjEDCBA::KSAC	This study
SPN226	IR715 ΔbcfABCDEFGH::pSF1	This study
SPN227	IR715 ΔfimAICDHF::pSPN22	This study
SPN230	IR715 ΔsafABCD	This study
SPN231	IR715 ΔstbABCD::pSF38	This study
SPN233	IR715 ΔstdAB::pSPN3	This study
SPN234	IR715 ΔstfACDEFG::pSF5	This study
SPN235	IR715 ΔsthABCDE::pSF25	This study
SPN236	IR715 ΔstiABCH::pSPN2	This study
SPN237	IR715 ΔstjEDCBA	This study
SPN251	IR715 ΔsafABCD::pSPN13	This study
SPN252	IR715 ΔstjEDCBA::pSPN14	This study
SPN334	IR715 ΔpefACD-orf5-orf6::KSAC	This study
SPN335	IR715 ΔpefACD-orf5-orf6	This study
SPN336	IR715 ΔpefACD-orf5-orf6::pSPN16	This study
SPN337	IR715 ΔstcABCD::KSAC	This study
SPN338	IR715 ΔstcABCD	This study
SPN339	IR715 ΔstcABCD::pSPN15	This study
SPN365	SR11 ΔfimAICDHF	This study
SPN366	SR11 ΔfimAICDHF ΔstbABCD	This study
SPN367	SR11 ΔfimAICDHF ΔstbABCD ΔsthABCDE	This study
SPN368	SR11 ΔfimAICDHF ΔstbABCD ΔsthABCDE ΔstfACDEFG	This study
SPN369	SR11 ΔfimAICDHF ΔstbABCD ΔsthABCDE ΔstfACDEFG ΔstiABCH	This study
SPN370	SR11 ΔfimAICDHF ΔstbABCD ΔsthABCDE ΔstfACDEFG ΔstiABCH ΔbcfABCDEFGH	This study
SPN371	SR11 ΔfimAICDHF ΔstbABCD ΔsthABCDE ΔstfACDEFG ΔstiABCH ΔbcfABCDEFGH ΔsafABCD	This study
SPN372	SR11 ΔfimAICDHF ΔstbABCD ΔsthABCDE ΔstfACDEFG ΔstiABCH ΔbcfABCDEFGH ΔsafABCD ΔpefACD-orf5-orf6	This study
SPN373	SR11 ΔfimAICDHF ΔstbABCD ΔsthABCDE ΔstfACDEFG ΔstiABCH ΔbcfABCDEFGH ΔsafABCD ΔpefACD-orf5-orf6 ΔstcABCD	This study
SPN374	SR11 ΔfimAICDHF ΔstbABCD ΔsthABCDE ΔstfACDEFG ΔstiABCH ΔbcfABCDEFGH ΔsafABCD ΔpefACD-orf5-orf6 ΔstcABCD ΔstjEDCBA	This study
SPN375	SR11 ΔfimAICDHF ΔstbABCD ΔsthABCDE ΔstfACDEFG ΔstiABCH ΔbcfABCDEFGH ΔsafABCD ΔpefACD-orf5-orf6 ΔstcABCD ΔstjEDCBA ΔstdAB	This study
SPN376 (=SR11 Δ12)	SR11 ΔfimAICDHF ΔstbABCD ΔsthABCDE ΔstfACDEFG ΔstiABCH ΔbcfABCDEFGH ΔsafABCD ΔpefACD-orf5-orf6 ΔstcABCD ΔstjEDCBA ΔstdAB ΔlpfABCDE::KSAC	This study
MvP493	ΔSPI4::aph	15
MvP681	ΔsadA::aph	67
MvP702	Δwzz::aph	46
MvP703	ΔfepE::aph	46
MvP813	ΔinvC::aph	68
MvP886	ΔrfaL::aph	69
MvP1208	ΔsopB::aph	70
MvP1209	ΔcheY::aph	71
MvP1210	ΔfliI::aph	71
MvP1412	ΔsopE2::aph	18
MvP1472	ΔsopA::aph	This study, construction intermediate
MvP1527	ΔcheZ::aph	71
MvP1611	ΔbapABCD::aph	This study, construction intermediate
MvP1663	ΔsadA::aph	This study, construction intermediate
MvP1754	ΔfliC::aph	72
MvP1755	ΔfljB::aph	72
MvP1760	ΔfliC ΔfljB::aph	72
MvP1825	ΔshdA::aph	This study
MvP1827	ΔmisL::aph	This study
MvP1842	PtetA::shdA	This study, construction intermediate
MvP1884	ΔsipA::aph	18
MvP1885	ΔsopD::aph	This study, construction intermediate
MvP2050	ΔmotAB::aph	73
MvP2447	Δ12 ΔmisL	This study
MvP2448	Δ12 ΔmisL ΔshdA::aph	This study, construction intermediate
MvP2449	Δ12 ΔmisL ΔshdA	This study, construction intermediate
MvP2456	Δ12 ΔmisL ΔshdA ΔSPI4::aph	This study, construction intermediate
MvP2457	Δ12 ΔmisL ΔshdA ΔSPI4	This study, construction intermediate
MvP2458	Δ12 ΔmisL ΔshdA ΔSPI4 ΔbapABCD::aph	This study, construction intermediate
MvP2486	Δ12 ΔmisL ΔshdA ΔSPI4 ΔbapABCD	This study, construction intermediate
MvP2487	Δ12 ΔmisL ΔshdA ΔSPI4 ΔbapABCD ΔsadA::aph	This study, construction intermediate
MvP2488	Δ12 ΔmisL ΔshdA ΔSPI4 ΔbapABCD ΔsadA	This study, construction intermediate
MvP2506	Δ12 rck::aph-I-SceI	This study
MvP2507	Δ12 pagN::aph-I-SceI	This study
MvP2508	Δrck::aph-I-SceI	This study
MvP2509	ΔpagN::aph-I-SceI	This study
MvP2518	Δ12 ΔmisL ΔshdA ΔSPI4 ΔbapABCD ΔsadA Δrck::aph-I-SceI	This study, construction intermediate
MvP2535	Δ12 ΔmisL ΔshdA ΔSPI4 ΔbapABCD ΔsadA Δrck	This study, construction intermediate
MvP2533	Δ12 ΔmisL ΔshdA ΔSPI4 ΔbapABCD ΔsadA Δrck ΔpagN::aph-I-SceI	This study, construction intermediate
MvP2537	Δ12 ΔmisL ΔshdA ΔSPI4 ΔbapABCD ΔsadA Δrck ΔpagN	This study
MvP2622	Δ12 shdA::aph	This study
MvP2623	Δ12 sadA::aph	This study
MvP2624	Δ12 SPI4::aph	This study
MvP2625	Δ12 bapABCD::aph	This study
MvP2702	ΔcsgBAC-DEFG::aph	This study, construction intermediate
MvP2703	Δ12 csgBAC-DEFG::aph	This study
MvP2706	Δ12 ΔmisL ΔsadA ΔshdA ΔSPI4 ΔbapABCD Δrck ΔpagN ΔcsgBAC-DEFG::aph	This study, construction intermediate
MvP2707 (=SR11 Δ20)	Δ12 ΔmisL ΔsadA ΔshdA ΔSPI4 ΔbapABCD Δrck ΔpagN ΔcsgBAC-DEFG	This study
MvP2710	Δ12 ΔmisL ΔsadA ΔshdA ΔSPI4 ΔbapABCD Δrck ΔpagN ΔcsgBAC-DEFG ΔfliI::aph	This study
MvP2711	Δ12 ΔmisL ΔsadA ΔshdA ΔSPI4 ΔbapABCD Δrck ΔpagN ΔcsgBAC-DEFG ΔmotAB::aph	This study
MvP2718	Δ12 ΔinvC::aph	This study
MvP2788	Δ12 ΔfepE::aph	This study, construction intermediate
MvP2789	Δ12 Δwzz::aph	This study, construction intermediate
MvP2790	Δ12 ΔrfaL::aph	This study, construction intermediate
MvP2798	Δ12 ΔfepE	This study
MvP2799	Δ12 Δwzz	This study
MvP2800	Δ12 ΔrfaL	This study
MvP2812	Δ12 ΔfepE Δwzz::aph	This study
MvP2819	Δ12 ΔsopB::aph	This study, construction intermediate
MvP2828	Δ12 ΔsopB	This study, construction intermediate
MvP2829	Δ12 ΔsopB ΔsopA::aph	This study, construction intermediate
MvP2831	Δ12 ΔsopB ΔsopA	This study, construction intermediate
MvP2832	Δ12 ΔsopB ΔsopA ΔsopE2::aph	This study, construction intermediate
MvP2835	Δ12 ΔsopB ΔsopA ΔsopE2	This study, construction intermediate
MvP2841	Δ12 ΔsopB ΔsopA ΔsopE2 ΔsopD::aph	This study, construction intermediate
MvP2843	Δ12 ΔsopB ΔsopA ΔsopE2 ΔsopD	This study, construction intermediate
MvP2844 (=SR11 Δeffector proteins)	Δ12 ΔsopB ΔsopA ΔsopE2 ΔsopD ΔsipA::aph	This study
MvP2864	Δ12 ΔmisL::aph	This study

Construction of Δ12 strain with a deletion of the chaperone-usher fimbrial gene cluster.

Strains are listed in Table 1, plasmids (and the extent of each fimbrial gene cluster deletion) in Table 2, and oligonucleotides in Table 3. For cloning, E. coli DH5α was used as a host for pCR2.1 and pBluescriptII-derived plasmids, whereas E. coli CC118 λpir was used as a host for pRDH10-derived plasmids. To generate the unmarked Δlpf, Δpef, Δsaf, Δstc, and Δstj allelic-exchange-mediated deletion constructs, upstream and downstream regions flanking the respective gene cluster to be deleted were amplified from the genome of S. enterica serovar Typhimurium LT2 by PCR with primers containing (i) restriction sites that enable ligation of the flanking regions together at their proximal ends, as well as that enable future introduction of an antibiotic resistance cassette, and (ii) restriction sites to enable subcloning of the deletion construct into the sucrose-counterselectable pRDH10 suicide vector. With the exception of the Δlpf construct, flanking region PCR products were gel purified (QIAEX II kit; Qiagen), digested with XbaI (New England BioLabs [NEB]), ligated with T4 DNA ligase (NEB), and then PCR amplified by utilizing the distal primer of each respective flanking region’s primer pair. Products were then cloned into pCR2.1 via the TOPO TA kit (Invitrogen), and correct inserts were confirmed by Sanger sequencing (SeqWright). For the Δlpf construct, each flanking region was PCR amplified, gel purified, cloned separately into pCR2.1, and then confirmed by sequencing. The flanking regions were then joined together by sequential subcloning into pBluescriptII KS+. The unmarked Δlpf, Δpef, Δsaf, Δstc, and Δstj constructs were then subcloned into pRDH10. To generate the unmarked Δstd and Δsti constructs in pRDH10, the Km resistance cassette was removed from pEW5 and pEW13, respectively, by restriction digestion, and then the vectors were gel purified and religated. As pSF2 (pRDH10 Δfim) did not confer appreciable sucrose sensitivity to strains harboring it, the Δfim construct was subcloned into another site in pRDH10: following EcoRI digestion of pSF2, the Δfim construct was gel purified, blunted (QuickBlunt; NEB), and subcloned into the blunted BamHI site of pRDH10, yielding pSPN22. To generate Km-marked deletion constructs, the KSAC cassette of pBS34 was excised with XbaI or PstI as relevant, gel purified, and then subcloned between the flanking regions of the Δlpf, Δpef, Δsaf, Δstc, and Δstj constructs in their respective pRDH10-based vectors. To enable their conjugation, all unmarked and KSAC-marked pRDH10-based fimbrial gene cluster deletion vectors were electroporated into E. coli S17-1 λpir.

TABLE 2.

Plasmids used in this study

Plasmid	Relevant genotype	Reference or source
pE-FLP	FLP recombinase expression	74
pKD4	aph resistance cassette flanked by FRT sites; Kmr Carbr	58
pKD13	aph resistance cassette flanked by FRT sites, temp-sensitive replication (30°C); Kmr Carbr	58
pWRG730	Red recombinase expression	59
p3313	pWSK29 rfaDFCL	69
p3773	tetR PtetA	31
p4253	tetR PtetA::bapABCD in pWSK29	31
p4318	tetR PtetA::bapA[ΔBIg1]BCD in pWSK29	This study
p4321	tetR PtetA::bapA[ΔBIg28]BCD in pWSK29	This study
p4320	tetR PtetA::bapA[ΔBIg15-24]BCD in pWSK29	This study
p4331	tetR PtetA::bapA[ΔBIg1-28]BCD in pWSK29	This study
p4380	tetR PtetA::csgBACEFG in pWSK29	31
p4389	tetR PtetA::stiABCD in pWSK29	31
p4390	tetR PtetA::stfABCDEFG in pWSK29	31
p4391	tetR PtetA::stbABCDEFG in pWSK29	31
p4392	tetR PtetA::fimAICDHF in pWSK29	31
p4393	tetR PtetA::safABCD in pWSK29	31
p4394	tetR PtetA::stdABCD in pWSK29	31
p4395	tetR PtetA::stjABCDE in pWSK29	31
p4396	tetR PtetA::pefACDEF in pWSK29	31
p4397	tetR PtetA::bcfABCDEFG in pWSK29	31
p4399	tetR PtetA::stcABC in pWSK29	31
p4400	tetR PtetA::sthABCDE in pWSK29	31
p4401	tetR PtetA::pagN in pWSK29	31
p4402	tetR PtetA::rck in pWSK29	31
p4403	tetR PtetA::misL in pWSK29	31
p4519	tetR PtetA::lpfABCDE in pWSK29	31
p4520	tetR PtetA::shdA in pWSK29	31
p4904	tetR PtetA::hilD in pWSK29	This study
p5035	tetR PtetA::sadBA in pWSK29	This study
pBluescriptII KS+	Cloning vector; Carbr	75
pBS34	pBluescriptII KS+ [XbaI][PstI]KSAC[PstI][XbaI]); Carbr Kmr	76
pCR2.1	TOPO TA cloning vector; Carbr Kmr	Invitrogen
pEW5	pRDH10 ΔstdAB (−60 to +3219)::KSAC; Cmr Tetr Kmr	12
pEW13	pRDH10 ΔstiABCH (+40 to +4992)::KSAC); Cmr Kmr	12
pRDH10	oriVR6K sacRB mobRP4; Cmr Tetr	77
pSF1	pRDH10 ΔbcfABCDEFGH (+47 to +6830); Cmr Tetr	12
pSF2	pRDH10 ΔfimAICDHF (+40 to +5970); Cmr Tetr	66
pSF5	pRDH10 ΔstfACDEFG (−122 to +5493); Cmr Tetr	12
pSF25	pRDH10 ΔsthABCDE (−6 to +5420); Cmr Tetr	12
pSF38	pRDH10 ΔstbABCD (−59 to +5183); Cmr	12
pSPN2	pEW13 ΔstiABCH (+40 to +4992); Cmr	This study
pSPN3	pEW5 ΔstdAB (−60 to +3219); Cmr Tetr	This study
pSPN5	pCR2.1 (LPF-FR1); Carbr Kmr	This study
pSPN6	pCR2.1 ΔsafABCD (−45 to +4364); Carbr Kmr	This study
pSPN7	pCR2.1 ΔstjEDCBA (−49 to +5185); Carbr Kmr	This study
pSPN8	pCR2.1 ΔstcABCD (−65 to +4827); Carbr Kmr	This study
pSPN9	pCR2.1 ΔpefACD-orf5-orf6 (−110 to +5610); Carbr Kmr	This study
pSPN12	pCR2.1 (LPF-FR2); Carbr Kmr	This study
pSPN13	pRDH10 ΔsafABCD (−45 to +4364); Cmr	This study
pSPN14	pRDH10 ΔstjEDCBA (−49 to +5185); Cmr	This study
pSPN15	pRDH10 ΔstcABCD (−65 to +4827); Cmr	This study
pSPN16	pRDH10 ΔpefACD-orf5-orf6 (−110 to +5610); Cmr	This study
pSPN17	pBluescriptII KS+ ([BamHI]LPF-FR1[PstI]); Carbr	This study
pSPN18	pSPN13 ΔsafABCD (−45 to +4364)::KSAC); Cmr Kmr	This study
pSPN19	pSPN14 ΔstjEDCBA (−49 to +5185)::KSAC); Cmr Kmr	This study
pSPN20	pSPN15 ΔstcABCD (−65 to +4827)::KSAC); Cmr Kmr	This study
pSPN21	pSPN16 ΔpefACD-orf5-orf6 (−110 to +5610)::KSAC); Cmr Kmr	This study
pSPN22	pRDH10 ΔfimAICDHF (+40 to +5970); Cmr	This study
pSPN26	pSPN17 ([BamHI]LPF-FR1[PstI]LPF-FR2[Acc65I]); Carbr	This study
pSPN27	pRDH10 ΔlpfABCDE (−60 to +5325); Cmr	This study
pSPN37	pSPN27 ΔlpfABCDE (−60 to +5325)::KSAC); Cmr Kmr	This study

TABLE 3.

Oligonucleotides used in this study

Oligonucleotide	Sequence (5′–3′)	Purpose and/or target
Gibson assembly
1r-ST-Ptet-sadA-pWSK29	CCGGGCTGCAGGAATTCATGGCATTATGCCATTGC	sadBA
1f-ST_Ptet-fim-pWSK29	TCGACGGTATCGATAAGCTTAGGGAAAAAGGTTATGCTGCT	sadBA
Vf-pWSK29	GAATTCCTGCAGCCCGGGG	Vector p4392
Vr-pWSK29-Ptet-rev	TTCACTTTTCTCTATCACTGATAGGGAGTGGTAAAATAACTCT	Vector p4392
1f-PtetA-hilD	CCCTATCAGTGATAGAGAAAAGTGAAAACATCAACAAAGGGATAATATG	hilD
1r-hilD-vec	CCCGGGCTGCAGGAATTCTGCCTGGCAGAAAGCTA	hilD
Vr-PtetA	TTCACTTTTCTCTATCACTGATAGGGAGTGGTA	Vector p4252
hVPS26B-Rev-EcoRI	TCAGAATTCTCTGCCTGCAGTTGTTGTCAGACAGC	Vector pWSK29
Vr-pWSK29	AAGCTTATCGATACCGTCGACCTC	Vector pWSK29
1f-FRT	CGACGGTATCGATAAGCTTGAAAGTTCCTATACTTTCTAGAGAATA	PtetA-misL
1r-misL	CCCGGGCTGCAGGAATTCCATGAAACCTATCAGCCAAA	PtetA-misL
Site-directed mutagenesis
ST-delbapA-Big1-Q5-fw	CCTCTCCCCGATACACCG	SDM p4253
ST-delbapA-Big1-Q5-rv	AGTATCTACAGGATTACTGCTACC	SDM p4253
ST-delbapA-Big28-Q5-fw	ATAACCAGTCTTGATCTGAC	SDM p4253
ST-delbapA-Big28-Q5-rv	CGTATCGACAATCACCGTC	SDM p4253
ST-delbapA-Big15-Q5-rv	GGTGTCGACGGTGAAGGTAAAGCTG	SDM p4253
ST-delbapA-Big24-Q5-fw	CTTGCGCCAACGGTTCCG	SDM p4253
Check of P22 transductions and aph cassette removal
K1-Red-Del	CAGTCATAGCCGAATAGCCT	aph
K2-Red-Del	CGGTGCCCTGAATGAACTGC	aph
CsgC-Check-rev	TGTTGCCCTACCGCAGAATG	csgBAC
CsgG-Del-Check-for	AGTGGGCTATGGCTGGCATC	csgBAC-DEFG
SPI4-Ctrl-For	CGGTAGAGAATGGTCGGTAT	SPI4
SPI4-Ctrl-Rev	GTGCTGACCTGATACGCTAT	SPI4
InvC-DelCheck-For	TGTATCAGCGTCAAGGACGA	invC
InvC-DelCheck-Rev	CGGCGAACAATAGACTGCTT	invC
SopB-DelCheck-Rev	CAATGGCATAAAGGGACAGC	sopB
SopB-DelCheck-For	TACGTATGGACGTCAGGATG	sopB
FLiI-DelCheck-For	CGATCCAACGTTGCATCACG	fliI
FliI-DelCheck-Rev	ACGCATTTCGCCCAGTAAAC	fliI
sopE2-check-rev4	GCGTCGCCATAAAAACGAATA	sopE2
SopE2-Red-Check-For	TGTGACGCAGTAGTTGAATTGAAG	sopE2
sopA-DelCheck-Rev	TTCGTACATGCGATGGTGAG	sopA
SopA-Check-For	CCTGCCAGATAACATGGTGAATT	sopA
BapA-Check-For	GTCAGGCACAAAAAACAAAGGGT	bapABCD
bapD-Check-Rev	CCGAAATTCCTACATCCTCGG	bapABCD
STM3690-sadA-For	GAGCATGGACAAACGTCACGC	sadA
SadA-end-Rev	GGCATTATGCCATTGCCTTTG	sadA
ShdA-DelCheck-For	GCCACAGCAAAGTTAAAGCG	shdA
ShdA-DelCheck-Rev	TGAAGTCAAATCCGTCACGC	shdA
MisL-DelCheck-For	TTTATGTGCATAAGCTGCGG	misL
MisL-DelCheck-Rev	CAGGGCCATCGTGGCTTTAT	misL
SipA-Red-Check-For	CACATTACAGACGCTGACGC	sipA
sopD-DelCheck-For	ACCACAAAGGATTACCAACC	sopD
sopD-DelCheck-For	GGCTGCATGAAGGGTAATTG	sopD
MotB-Check-Rev	CCTGCAGAATAGTGAAGCCG	motAB
MotA-Check-For	ATGAACAGATCGAACAGG	motAB
pagN-check-for	CGTAGAAGTGAAACCGTACG	pagN
pagN-check-rev	CAGCTATTTTACCGATAGTG	pagN
rck-check-for	GAGGATGAAGCGGCGTTACG	Rck
rck-check-rev	GTACCACACCACAAACCAGC	Rck
FliC-For-XhoI	GCGCTCGAGGCAAACAGTAGTTAAGCGCG	fliC
FliC-Rev-EcoRI	AGCGAATTCAGCTTTCGCTGCCTTGATTG	fliC
FljB-For-XhoI	AGTCTCGAGCAATTTGCACTAGTAAGCGC	fljB
FljB-Rev	GCAAGCATAGAATAATCCCG	fljB
CheZ-Check-For	AAAACCATTCGCGCCGATAG	cheZ
CheZ-Check-Rev	GGTAAAAAAGGCGGGGTTTAT	cheZ
CheY-DelCheck-Rev	TACCGATGCGCGCAATGATG	cheY
CheY-DelCheck-For	ACGAAGCAAGTTGTGTGGTG	cheY
Wzz-FepE-DelCheck-For	AAACTATCGGGCCCATCATC	fepE
Wzz-FepE-DelCheck-Rev	TGTTAAGCGATCTCAACCGC	fepE
WzzB-DelCheck-For	AAAAGTGTATACCCGCGATC	wzz
WzzB-DelCheck-Rev	AGTGATGTAGTGGCATTGAG	wzz
RfaL-DelCheck-For	GCTGGCTGGCGCAAAATTTG	rfaL
RfaL-DelCheck-Rev	TATTGTGCCATCTCAGGTTG	rfaL
Red deletion
CsgC-Red-del-rev	CCGCCACCATCAAAAACTACTGTGCAGAAGGCGGCCATTGTGTAGGCTGGAGCTGCTTCG	csgBAC
CsgG-Red-Del-for	CACGCTTTGTCGTATTCATCAGGATTCTGGCGGTACTGACATTCCGGGGATCCGTCGACC	csgBAC-DEFG
misL-Del13-For	AGACGCTTTACGCCATAATGCAGGAGGCAGAATGCCAACTATTCCGGGGATCCGTCGACC	misL
misL-Del13-Rev	ATCAGCGGCTCTGTTGTTACCTGAATCAGAAACTGTATTTTGTAGGCTGGAGCTGCTTCG	misL
ShdA-Del13-For	AATAAAAGCAACGCGCGCGGCGCTGGCTTGCGCCGTGGCTATTCCGGGGATCCGTCGACC	shdA
ShdA-Del13-Rev	GGCAGGGAACACCCGCCCGGTTTTGTCTAACTTACCAGTTTGTAGGCTGGAGCTGCTTCG	shdA
pagN-del-red-for	GAAACTTGTCTTTTAGCCCAATATTAAGGCAGGTTCTGAAAGGGTTTTCCCAGTCACGAC	pagN
pagN-del-red-rev	CATGAAGTCATTGGAGGCAGCCTTTGTGTCTGCATCATAATGCTTCCGGCTCGTATGTTG	pagN
rck-del-red-for	CATAACACAATGAACTTAACTGTGTTCAGGGAGTTTTATCAGGGTTTTCCCAGTCACGAC	rck
rck-del-red-rev	CGGAAGCCTGCGGCTCCGCTCCCTTTCCTGCTCTCCGTTATGCTTCCGGCTCGTATGTTG	rck
misL-Red-Ptet-For	TTTTATAGATCCGTTTCCATTTTTATTATTTCCATATTATTGTAGGCTGGAGCTGCTTCG	tetR PtetA
misL-Red-Ptet-Rev	ATGAGTAATTTTGGGGAGTTGGCATTCTGCCTCCTGCATTTTCACTTTTCTCTATCACTG	tetR PtetA
BapB-Del-Red-For	GTTCGGGGCAACAAGCGGTGATATTTAAAAGGGATAAACTGTGTAGGCTGGAGCTGCTTC	bapABCD
BapD-Del-Red-Rev	CACGCGTGACCAGCCCCCGTATCTTCTTATCTTCAACGATCATATGAATATCCTCCTTAG	bapABCD
SopA-Red-DEl13-For	CCAGACCGTTTTTCCATAATGATGTTGATAAGGAATTCTAATTCCGGGGATCCGTCGAC	sopA
SopA-Red-Del13-Rev	CAACGCTGTGTCCCTTAATTCCATGCGGGTTGAGGCTGGAGTAGGCTGGAGCTGCTTCG	sopA
sopD-Del13-For	GATATTGAATAATATAAATTTGAAGGAAAATATTATGCCAATTCCGGGGATCCGTCGACC	sopD
sopD-Del13-rev	CAGCCGGATTTTAAATTGGTTATATTACTGACTATCTTTATGTTGTAGGCTGGAGCTGCT	sopD
Δ12 construction: PCR primers for cloning flanking regions
100-LPF6-Bam	TATCGGGGATCCGGGTTGAGTCGTATGACC	lpf flanking region 1
63-LPF5-Pst	TATGCGCTGCAGGTGTATAGAGGTGGGTATTGG	lpf flanking region 1
64-LPF3-Pst	TATCGCCTGCAGCATCTGGTGGGGAGCAACAATAC	lpf flanking region 2
101-LPF7-Bam	TATCGGGGATCCGCCAAACAGTGAAAGAAGACGAAG	lpf flanking region 2
66-SAF1-Bam	ATAGGCGGATCCCTGCACTGAAAAGCGATACC	saf flanking region 1
67-SAF2-Xba	ATAGGCTCTAGAACGCCATACCAAATCTTACC	saf flanking region 1
92-SAF5-Xba	TATCGCTCTAGACTGTTCCACTCATACTTCC	saf flanking region 2
69-SAF4-Bam	TATGCGGGATCCTGGTCACAAGAAAGAGATGC	saf flanking region 2
70-STJ1-Bam	TTACGCGGATCCCCTTTTTCGCCCATTACG	stj flanking region 1
71-STJ2-Xba	TATCGGTCTAGAGGTCGGGATTCTATGAAG	stj flanking region 1
72-STJ3-Xba	TATCGGTCTAGAGAAGTGCTGACGAAATAAACG	stj flanking region 2
73-STJ4-Bam	ATACGCGGATCCGGCATGTTAGGTTTCACC	stj flanking region 2
78-STC5-Bam	TTTGCGGGATCCAAGAGAATATGACATTCACTGC	stc flanking region 1
79-STC6-Xba	ATAGCCTCTAGACATAGACAGGAAGTTATCGC	stc flanking region 1
80-STC7-Xba	ATAGGCTCTAGACGATAGGTGAATGAACTTCC	stc flanking region 2
81-STC8-Sal	TATGCCGTCGACCAGCAGAAATGATACACACG	stc flanking region 2
88-PEF5-Bam	TTTGGCGGATCCTAATCTCACAGCCCGAAGC	pef flanking region 1
89-PEF6-Xba	ATTGCCTCTAGACAGCTATGACGTGACATCG	pef flanking region 1
90-PEF7-Xba	ATAGCGTCTAGAATGCGTGGTGTACTGAGG	pef flanking region 2
91-PEF8-Sal	TAAGGGGTCGACGGCAGAAATGGTTTTGACG	pef flanking region 2
Δ12 construction: PCR primers for confirmation of deletions
34-KSAC-5out	GGCATAAATTCCGTCAGC	Amplify out KSAC 5′ end
35-KSAC-3out	TGATGACGAGCGTAATGG	Amplify out KSAC 3′ end
38-BCF-Up1	CATGATGACAAACGACTCC	bcf deletions
39-BCF-Down1	CGCCATTTGCAACATATCC	bcf deletions
40-FIM-Up1	CGTCTACGTCTTTATCTGG	fim deletions
41-FIM-Down1	GCACTTATCCTGTTGACC	fim deletions
42-LPF-Up1	GGGAGAATATCTGGAAAGC	lpf deletions
43-LPF-Down1	CAGCCACAATACAAAGTGC	lpf deletions
44-PEF-Up1	CGACAGGATATTTGCTCC	pef deletions
45-PEF-Down1	GTCAGTTTCCTTCATCACC	pef deletions
46-STB-Up1	ATATGTTCTCCCGAGTCG	stb deletions
47-STB-Down1	GTATGGCGGTATATTGTCG	stb deletions
48-STC-Up1	GGGGATATTCAGCTAACG	stc deletions
49-STC-Down1	GAGATCCAGGCAAAATCG	stc deletions
50-STD-Up1	TTCAGCAAACCCGTAAGG	std deletions
51-STD-Down1	GTGTAGCGATTCATCTGC	std deletions
52-STF-Up1	GCGTTTTACTGGTCTTTGC	stf deletions
53-STF-Down1	GTATCAACGGGAACTTTCG	stf deletions
54-STH-Up1	CCTTGTAGATGCCTATGC	sth deletions
55-STH-Down1	GGATTGGGACAACTTACC	sth deletions
56-STI-Up1	CAGAGACTGGTGACATCC	sti deletions
57-STI-Down1	AAGCTGAAATCGGAGACG	sti deletions
74-SAF-Up1	TATGATACCGAAGGAATACC	saf deletions
75-SAF-Down1	TCGACACGAAGCAAATCC	saf deletions
76-STJ-Up1	ACCCATGAACAGGTCTGC	stj deletions
77-STJ-Down1	ACTGAAGATGGCAACTCC	stj deletions
Δ12 construction: PCR primers to check for presence of predicted major subunit
145-bcfA1	GATACTACAACCGTCACT	bcfA presence
146-bcfA2	CCAACAGACGAGAAAAAAATCCCG	bcfA presence
147-fimA1	GCTGATCCTACTCCGGTG	fimA presence
148-fimA2	AAAATGGAACGCTGACGGGAGC	fimA presence
149-stbA1	GTTTCTGATAACACCATC	stbA presence
150-stbA2	GCTACCCAAAATAGTAACGCTCGC	stbA presence
151-stfA1	GCGGGCAGTAATACTGGT	stfA presence
152-stfA3	AGCCAGAACAATACCCACCACG	stfA presence
153-sthA1	TCCACACCGGTATTTGC	sthA presence
154-sth-II	GGCATCAAGGCGAAAAAGAGG	sthA presence
155-stiA	CAACAGGCAACAAAGCAACCC	stiA presence
156-stiC	CCGCCAAAGACGGCACCG	stiA presence
157-safA1-Bam	TTAGCGGGATCCGGCTCATTTTTGCCGAACTC	safA presence
158-safA2-Sal	TTCACCGTCGACTTAAGGTTGATATCCCACTACG	safA presence
159-stjE1-Bam	TTAGCGGGATCCGTTGAATCCACTGCTGTATTAAAACTG	stjE presence
160-stjE2-Sal	TATGCCGTCGACCTGGTTGTAGCAAAGGAAGC	stjE presence
161-lpfA1	GCTGAATCTGGTGACGGC	lpfA presence
162-lpfA2	GATTCTCTTCCTGAGCCTCCG	lpfA presence
163-pefA1	GCCAATGAAGTAACTTTCCTGG	pefA presence
164-pefA2	GTTCTGCTTACGGGGGATTATTTG	pefA presence
165-stcA1	GTTGATGAGTATGATTCAGGC	stcA presence
166-stcA2	AACGACTTCTTTCTTCTCTGCCG	stcA presence
167-stdAF	GCCGATACTACACCCACAGC	stdA presence
168-stdA2	CGACTTCAGGACGGAAAATGTC	stdA presence

S. Typhimurium IR715-derived strains harboring a single, KSAC-marked deletion of lpf, pef, saf, stc, or stj (e.g., SPN195 = IR715 Δsaf::KSAC) were generated by conjugation through mating of the respective S17-1 λpir pRDH10(Δ::KSAC) strain with IR715. Transconjugants were selected for on LB-Km-Nal agar, and those resulting from a double-crossover event were screened for by sensitivity to Cm and then validated by PCR using primer pairs to confirm that KSAC was located in the correct genomic context, as well as by being negative for PCR amplification of the relevant fimbrial gene cluster’s predicted major subunit gene.

Eleven (bcf, fim, pef, saf, stb, stc, std, stf, sth, sti, and stj) of the 12 KSAC-marked fimbrial gene cluster deletion strains were then converted to unmarked deletion strains (e.g., SPN230 = IR715 Δsaf) by mating the respective S17-1 λpir pRDH10(Δ) and IR715 Δ::KSAC strains. Transconjugants with pRDH10(Δ) integrated into the genome were selected for on LB-Cm-Nal agar, and colonies were then transferred to 5% sucrose agar (56) and incubated at 30°C. Sucrose-resistant (Suc^r) colonies lacking the pRDH10(Δ) vector and the Δ::KSAC locus were identified by screening for a Km^s Cm^s phenotype, and the presence of the unmarked deletion was then validated by obtaining the expected PCR product size when amplifying over the deleted region. To enable transduction of the unmarked deletions (57), we next generated IR715 Δ::pRDH10::Δ strains (e.g., SPN251 = IR715 Δsaf::pSPN13), thus reversibly marking the unmarked deletion with the Cm-selectable, sucrose-counterselectable pRDH10 suicide vector. The respective pRDH10(Δ) construct was thus conjugated back into the relevant IR715 unmarked deletion strain, transconjugants with the plasmid integrated into the genome were selected for on LB-Cm-Nal agar, and plasmid integration was further inferred by the inability to PCR amplify across the respective unmarked deletion region due to the size increase.

The S. Typhimurium SR11 strain with a deletion of all 12 chaperone-usher fimbrial gene clusters (Δ12; SPN376) was then generated, with a focus on minimizing the number of passages necessary for introducing each deletion. To begin, Δfim::pSPN22 of SPN227 was transduced via phage P22 HT105/1 int-201 into wild-type SR11, and transductants were selected for on LB-Cm agar. As SR11 accepts DNA from P22 but is resistant to lysis by the phage, phage cleanup was unnecessary. Transductants were thus struck immediately to 5% sucrose agar and incubated at 30°C. Suc^r colonies were then screened for Cm^s by streaking for single-colony isolation on both LB agar and LB-Cm agar. Colony PCR was performed to confirm Δfim status (positive for amplification across the unmarked deletion and negative for fimA amplification) of Suc^r Cm^s colonies. A validated colony was then grown in LB medium, an aliquot of which was used for creating a freezer stock (SPN365 = SR11 Δfim) and another aliquot of which was used in the next round of transduction. This process was then repeated for the remaining deletions. The unmarked deletions were transduced first, generating strains SPN366 to SPN375. For the final deletion, Δlpf::KSAC of SPN193 was transduced, yielding the Δ12 strain (SPN376). With each successive deletion, every deletion thus far introduced into the strain was reconfirmed by PCR, as was the expected presence/absence of every major fimbrial subunit gene.

Construction of strains and plasmids.

For introduction of the genes sadBA under the Tet-on system, template vector p4392 harboring tetR P_tetA::fimAICDHF was used. Amplification of sadBA from the genome of S. Typhimurium NCTC 12023 and the vector including the Tet-on system aph tetR P_tetA present on p4392 was done using oligonucleotides as listed in Table 3, and the PCR products were purified by PCR purification (NEB; Monarch). The PCR product containing sadBA and the PCR product from vector p4392 were assembled by Gibson assembly according to the manufacturer’s protocol (NEB; Monarch). For overexpression of the sii operon, a plasmid was generated for Tet-on expression of transcriptional regulator hilD. Using primers listed in Table 3, hilD was amplified from S. Typhimurium NCTC 12023 genomic DNA, and the vector including aph tetR P_tetA present on p4392 was amplified as described before.

Strains with deletion of csgBAC-DEFG, rck, and pagN were created using λ Red recombination in S. Typhimurium 12023 harboring pWRG730. One-step gene inactivation was performed as described previously (58) using oligonucleotides as listed in Table 3. Deletion was checked by colony PCR using oligonucleotides as listed in Table 3. Further deletion of aph was performed using pE-FLP encoding FLP recombinase as described (58). For strains lacking rck and pagN, further deletion of aph was performed using I-SceI counterselection as described previously (59). Generation of strains lacking all fimbrial operons (SR11 Δ12) and one further adhesive structure were created by transferring the deletion by P22 phage transduction. The several deletions were always checked by colony PCR using oligonucleotides as listed in Table 3.

Cultivation of sterile grown corn salad.

Corn salad seeds (Valerianella locusta Verte à cour plein 2, N.L. Chrestensen Erfurter Samen- und Pflanzenzucht) were kindly provided by Adam Schikora and Sven Jechalke (Justus Liebig University Giessen). Seeds were sterilized with 70% ethanol (EtOH) for 1 min followed by 3% NaClO for 2 min. Seeds were washed thrice with sterile ultrapure water (MilliQ) and allowed to dry for 30 min. Seeds were planted on Murashige-Skoog (MS) agar (per liter: 2.2 g of Murashige-Skoog medium including vitamins [Duchefa Biochemie; number M0222], 10 g of agar, and 0.5 g of morpholineethanesulfonic acid [MES; pH 5.4]) in sterile plastic containers with air filters (round model, 140 mm [Duchefa Biochemie; number E1674]) at 20°C with a 12-h/12-h day/night cycle for 8 weeks.

Adhesion to corn salad.

For infection of corn salad by Salmonella, leaf discs (8 mm average) of 8-week-old plants were punched out by biopsy punches immediately before infection process. Forty-eight-well plates were used with one leaf disc per well mechanically fixed by sterile stainless steel inlays. For each condition, three leaf discs were infected. For infection, overnight cultures of Salmonella strains were diluted 1:31 in LB (containing antibiotics if required) and grown for 3.5 h in test tubes with aeration in a roller drum. The cultures were diluted in phosphate-buffered saline (PBS) to obtain approximately 5.6 × 10⁷ bacteria/ml, and 50 μl of this inoculum was spotted onto one leaf disc. The infection process was carried out either for 1 h at room temperature (RT) under static conditions or for 55 min at RT after a centrifugation step at 500 × g for 5 min. After infection, leaf discs were washed once with PBS to remove nonbound bacteria. Three leaf discs were transferred to tubes and washed two further times with PBS by short mixing on a Vortex mixer. Plant tissue was homogenized with a pellet pestle motor in 600 μl of 1% sodium deoxycholate in PBS, and CFU were determined by plating serial dilutions of the lysates on MH agar plates (Mueller-Hinton agar plates) incubated overnight at 37°C. A noninfected sample was used in every assay to ensure the sterility of the corn salad.

Flow cytometry.

For analysis of surface expression of SadA and BapA by flow cytometry, 6 × 10⁸ bacteria were washed in PBS and then fixed with 3% paraformaldehyde–PBS for 20 min. Bacteria were blocked with 2% goat serum in PBS for 30 min and afterwards stained with the specific antiserum goat anti-SadA or goat anti-BapA diluted 1:250 and 1:1,000 in 2% goat serum-PBS for 2 h and goat anti-rabbit IgG antibody coupled to Alexa-Fluor 488 diluted 1:2,000 in 2% goat serum-PBS for 1 h. For analysis of surface expression of SiiE by flow cytometry, ca. 3 × 10⁸ bacteria were fixed in 3% paraformaldehyde in PBS for 20 min. Bacteria were blocked with blocking solution (2% goat serum and 2% bovine serum albumin in PBS) for 30 min and afterwards stained with the specific antiserum anti-SiiE C-terminally coupled to Alexa-Fluor 488 (1:100) for 1 h. Bacteria were measured with an Attune NxT flow cytometer (Thermo Fisher) and analyzed using Attune NxT software version 2.7. A mutant strain lacking the respective adhesive structure was used as a negative control for gating.