Logo of pnasPNASInfo for AuthorsSubscriptionsAboutThis Article
Proc Natl Acad Sci U S A. Oct 1, 2002; 99(20): 13142–13147.
Published online Sep 23, 2002. doi:  10.1073/pnas.202358699
PMCID: PMC130600

HecA, a member of a class of adhesins produced by diverse pathogenic bacteria, contributes to the attachment, aggregation, epidermal cell killing, and virulence phenotypes of Erwinia chrysanthemi EC16 on Nicotiana clevelandii seedlings


Erwinia chrysanthemi is representative of a broad class of bacterial pathogens that are capable of inducing necrosis in plants. The E. chrysanthemi EC16 hecA gene predicts a 3,850-aa member of the Bordetella pertussis filamentous hemagglutinin family of adhesins. A hecA::Tn7 mutant was reduced in virulence on Nicotiana clevelandii seedlings after inoculation without wounding. Epifluorescence and confocal laser-scanning microscopy observations of hecA and wild-type cells expressing the green fluorescent protein revealed that the mutant is reduced in its ability to attach and then form aggregates on leaves and to cause an aggregate-associated killing of epidermal cells. Cell killing also depended on production of the major pectate lyase isozymes and the type II, but not the type III, secretion pathway in E. chrysanthemi. HecA homologs were found in bacterial pathogens of plants and animals and appear to be unique to pathogens and universal in necrogenic plant pathogens. Phylogenetic comparison of the conserved two-partner secretion domains in the proteins and the 16S rRNA sequences in respective bacteria revealed the two datasets to be fundamentally incongruent, suggesting horizontal acquisition of these genes. Furthermore, hecA and its two homologs in Yersinia pestis had a G+C content that was 10% higher than that of their genomes and similar to that of plant pathogenic Ralstonia, Xylella, and Pseudomonas spp. Our data suggest that filamentous hemagglutinin-like adhesins are broadly important virulence factors in both plant and animal pathogens.

Bacterial attachment to host tissues by various adhesins is a first step in the pathogenesis of many animal pathogens, but a role for attachment in plant pathogenesis is less clear. The exception is the tumorigenic pathogen Agrobacterium tumefaciens, whose attachment to plant tissues before transfer of T-DNA (the portion of the tumor-inducing Ti plasmid that is transferred to plant cells) involves a Ca2+-dependent adhesin (1), a repertoire of proteins encoded by att (attachment) genes (24), exo- and capsular polysaccharides (5, 6), and bacterial cellulose fibrils (5). In contrast, the role of attachment and adhesins in virulence is unclear for the more prevalent necrogenic (rather than tumorigenic) plant pathogens. These bacteria colonize the surface and intercellular spaces of plants and attack with various combinations of virulence effector proteins injected by type III secretion systems, extracellular pectic enzymes, and low molecular-weight toxins.

The necrogenic pathogens have been reported to produce a variety of potential adhesins, including fimbriae by Erwinia rhapontici, Erwinia carotovora, Pseudomonas syringae (7), Xanthomonas campestris (8), and Ralstonia solanacearum (9), type IV pili by P. syringae (10), and adhesive factors such as lipopolysaccharide by R. solanacearum (11). Attachment to leaf surfaces by type IV pili slightly promotes the epiphytic fitness of P. syringae (12), and the related process of self aggregation is also promoted in P. syringae by type IV pili and in X. campestris by FimA fimbriae (8). The recent report that the Xylella fastidiosa (13) and R. solanacearum (14) genomes encode multiple candidate adhesins, most notably the 14 homologs in R. solanacearum of the Bordetella pertussis filamentous hemagglutinin (FHA), indicates that the potential role of adhesins in plant pathogenesis warrants closer scrutiny.

Much more is known about the adhesins produced by animal pathogens. These are usually proteins that are assembled in a structure, such as a pilus or fimbriae, or are surface exposed (afimbrial adhesins). The afimbrial adhesins are anchored directly in the outer membrane with the adhesin domain exposed to the extracellular surface (15). One of the afimbrial adhesins from B. pertussis, FHA, has been extensively characterized because of its large size (220 kDa) and the possibility of its use for vaccination against whooping cough (16). FHA is secreted in a Sec-dependent manner by the two-partner secretion system, which involves an outer membrane-associated accessory protein (FhaC “partner” protein) interacting with a conserved FHA N-proximal secretion domain. FHA is proteolytically processed after secretion, has multiple substrate-binding domains and proapoptotic activity, and promotes bacterial aggregation (17).

The plant pathogens Erwinia chrysanthemi and E. carotovora also encode FHA homologs (18, 19). E. chrysanthemi and E. carotovora cause soft-rot diseases on a wide variety of plants and are representative of a major subgroup of necrogenic pathogens that rapidly kill and macerate seedlings or mature fleshy tissues. The adhesin-encoding hecA gene was partially sequenced because of its genomic proximity to genes encoding the type III secretion (hrp/hrc) genes in E. chrysanthemi EC16 (18). The E. chrysanthemi Hrp system contributes to the infectivity of strain EC16 on witloof chicory leaves and strain 3937 on African violet (20, 21), although the main virulence factors in soft-rot disease are the multiple pectate lyase isozymes encoded by pel genes (22) and secreted via the out gene-encoded type II secretion system (23).

The pectolytic attack of susceptible tissues by E. carotovora is thought to be activated by N-acyl-homoserine lactone-mediated induction of pectic enzyme genes that occurs when bacterial populations in infection sites achieve “quorum” densities (24, 25). Although the quorum-sensing system in E. chrysanthemi appears more complex, the dynamics of pectic enzyme production and pathogenesis appear similar (26). E. chrysanthemi strain 3937 produces five major Pel isozymes and several secondary isozymes, and this multiplicity may contribute to the wide host range of the bacterium, as suggested by the varying contributions of the major Pels to cumulative virulence in different hosts (27). Other factors that may contribute to virulence in different hosts include siderophores and exopolysaccharide (systemic invasion of Saintpaulia), the Sap system (resistance to antimicrobial peptides in potato), and methionine sulfoxide reductase for repair of oxidative stress (macerated lesions in chicory leaves) (2831).

Established virulence assays for soft-rot Erwinia spp. are based on the size of macerated lesions and/or numbers of plants locally or systemically diseased after artificial wound inoculations or on the numbers of axenic tobacco seedlings macerated after spray inoculation (27, 32). With the advent of E. chrysanthemi (www.ahabs.wisc.edu/~pernalab/erwinia/index.html) and E. carotovora (www.sanger.ac.uk/Projects/Microbes/) genomics projects, new assays are needed that provide additional phenotypes for virulence factors and a more detailed view of pathogenesis, particularly the early events. Green fluorescent protein (GFP) from Aequorea victoria has been used for this purpose with other plant pathogens, such as R. solanacearum (33). However, a previous attempt to use GFP in E. carotovora subsp. atroseptica had limited success because the level of fluorescence was not suitable for observing bacterial movement in planta, and GFP subsequently has not been used with soft-rot Erwinia spp.

In this study, we report that (i) pGFP-TIR (34), unlike many other available GFP variants, produces sufficient fluorescence for use with E. chrysanthemi in planta; (ii) virulence on unwounded Nicotiana clevelandii seedlings is strongly reduced in a hecA mutant and in mutants deficient in the major Pel isozymes; (iii) the hecA mutant is reduced in its ability to attach to and aggregate on leaves; (iv) the hecA mutant is similarly unable to cause an early, localized, pectic enzyme-dependent killing of N. clevelandii epidermal cells that is associated with bacterial aggregates; and (v) analyses of the complete sequence of HecA, the N-terminal (secretion) domain of several HecA- related proteins, and the associated bacterial rRNA sequences suggest that these adhesin genes have been horizontally acquired.

Materials and Methods

Bacterial Strains and Plasmids.

E. chrysanthemi wild-type strains (35, 36) and the mutants CUCPB5006 (PelABCE) (37), CUCPB5026 (Out) (37), and CUCPB5040 (Hrp) (20) have been described previously. E. chrysanthemi strains were grown at 30°C in King's medium B (KB) supplemented with 20 μg/ml of nalidixic acid, 50 μg/ml of kanamycin, 10 μg/ml of gentamycin, and 100 μg/ml of ampicillin. Overnight cultures were washed twice with KB broth and adjusted to an OD600 of 0.3. Plasmids were isolated from E. coli DH5α following standard procedures (38). To obtain plasmid pCPP2394 containing the hecA and hecB genes, pCPP2157 (39) was digested with BglII, and the 14-kb fragment was purified by using the Prep-a-Gene kit (Bio-Rad) and ligated to pBluescript SK(−) (Stratagene). To obtain plasmid pCPP3095 containing hecA, pCPP2157 was digested with NotI, and the 10-kb fragment was purified and ligated to pBluescript SK(−). Plasmids were introduced into E. chrysanthemi strains by electroporation by using a Gene Pulser (Bio-Rad).

DNA Sequencing.

The sequence of the hecA region was obtained by primer walking. DNA sequencing was performed with an ABI 3700 DNA sequencer (Applied Biosystems) at the Cornell University BioResource Center. Sequences were assembled by using seqman 5.03 (DNAstar, Madison, WI). The GenBank accession no. for hecA is AF501263.

hecA Mutagenesis.

pCPP3095 was mutagenized with Tn7 by using the Genome Priming System (New England Biolabs). The location of the transposon insertions was determined by sequencing outward of the transposon by using Tn7 primers gpsS and gpsN, provided by the manufacturer. Six plasmid clones containing transposon insertions in hecA gene were introduced into the chromosome of wild-type E. chrysanthemi strain AC4150 (nalidixic acid resistant derivative of EC16) by marker exchange, following standard procedures (40).

Plant Bioassays.

N. clevelandii seeds were grown on agar medium following previously described procedures (32) with some modifications. N. clevelandii seeds were surface sterilized and planted in 24-well tissue culture plates (Costar) on Murashige and Skoog media (GIBCO/BRL). Planted seeds were grown in a growth chamber at 25°C for 4 weeks with a photoperiod of 12 h per day. Four-week-old N. clevelandii seedlings were inoculated with bacterial suspensions prepared as described above. Inoculation was done by pipetting 50 μl of suspension to the apical meristem followed by incubation at 25°C. Maceration was scored at 24 and 48 h after inoculation.

Identification of an Available GFP Variant That Functions in E. chrysanthemi EC16.

Several plasmids expressing fluorescent proteins were tested in E. chrysanthemi: pGFP, pGFPuv, pEGFP, and pEYFP (CLONTECH) produced no detectable fluorescence in AC4150; pTB93G and pTB93F (41), produced fluorescence in only 1% of the transformants. pGFPmut3.1 (CLONTECH) (42) made AC4150 fluorescent, but it reduced bacterial viability and virulence, and single-copy integration into the genome via insertion into a mini-Tn5 transposon did not correct this problem. pDsRed1 (CLONTECH) (43) also failed to fluoresce in AC4150, although it did in E. chrysanthemi strain 3937. In contrast, pGFP-TIR (34) was fluorescent in both AC4150 and 3937 and did not abolish virulence.

Epifluorescence and Confocal Laser-Scanning Microscopy.

Four-week-old N. clevelandii leaves were detached and dipped in suspensions containing GFP-labeled bacteria and incubated for 1, 2, 4, and 8 h in a moist chamber. Bacterial behavior on N. clevelandii leaves was initially observed and described by using an Olympus (New Hyde Park, NY) BL-FLA microscope equipped with a FITC filter. To further compare the location of bacterial cells relative to the leaf tissue, we used a confocal laser-scanning microscope (Bio-Rad MRC600) equipped with two filter cubes (excitation 488, 568/dichroic 560/emission 522 DF35) at the Microscopy and Imaging Facility at Cornell University. Inoculated N. clevelandii leaves were observed at different time points, and the leaf epidermis was used to fix the starting point for the scanning. Scanning was done to 20 μm below the epidermis into the mesophyll cells and 40 μm above the epidermis. Images taken in the red and green channels were merged by using confocal assistant 4.02 (Bio-Rad).

Attachment, Aggregation, and Cell Death Assays.

N. clevelandii leaves were detached from 4-week-old plants and placed in a moist chamber. Leaf surfaces were inoculated with 10 μl of bacterial suspension containing 1 × 108 cfu per ml. Inoculated leaves were incubated for 1, 2, 4, and 8 h. After each incubation time, leaves were rinsed in 1 ml of H2O and vortexed at maximum speed for 20 sec. Bacterial cells that remained attached to the leaf surface were counted by using epifluorescence microscopy. Individual bacterial cells on the leaf epidermis were counted in 10 randomly selected microscopic fields from three leaves per bacterial strain. In assays of potential inhibitors, AC4150 was incubated for 30 min with chloramphenicol (20 μg/μl) at room temperature or proteinase K (100 μg/μl) on ice before inoculation on detached N. clevelandii leaves. Adhesion and aggregation were scored at 4 and 8 h postinoculation, respectively. To assess aggregation in colonies grown in culture, bacterial strains were grown overnight on King's medium B agar, then a loopful of cells was spread in 50 μl of 1× PBS on a microscope slide and observed immediately by bright-field microscopy with a ×100 objective lens as described previously (44). For plant cell death assay, detached N. clevelandii leaves were inoculated with bacterial suspensions as described above and incubated for 8 h. Then leaves were stained on a microscope slide with 50 μl of propidium iodide (100 μg/ml). After incubation for 10 min, the leaves were rinsed with water and observed under confocal laser-scanning microscopy.

HecA Protein Sequence and Phylogenetic Analyses.

The entire hecA sequence was used to search the GenBank nonredundant database by using the blastx program (45). Significant similarity (E values ≤1 × 10−18) was found to regions at the N and C termini of putative adhesins of the hemagglutinin/hemolysin type represented by the FHA from B. pertussis. The large size of these proteins, the presence of multiple copies in some genomes, and the potential redundancy for submissions of partial protein sequences led us to restrict the data set for the analysis to the well characterized B. pertussis FHA and proteins found in fully sequenced genomes and larger than 1,200 aa. The previously identified secretion domain located at the N terminus (46) was shown to be conserved and was chosen to align the proteins found in plant pathogenic bacteria. Alignment was done by using the clustalw method (47) with Gonnet protein weight matrix (48) from megalign (DNAstar). To determine whether the relative similarities of these proteins correlate with the phylogeny of the bacteria, a test of congruence between the secretion domain from all plant and animal pathogens and a 16S rRNA data set was conducted. The 16S rRNA sequences from the same set of fully sequenced genomes were recovered from GenBank and aligned by the clustalw method by using the PAM250 matrix (49) from megalign (DNAstar). The 16S rRNA alignment was combined with the secretion domain alignment, and the homogeneity partition test of Farris was performed (50).


E. chrysanthemi Mutant CUCPB5124 Carries a Tn7 Insertion in the 11.5-kb hecA ORF.

The hecA gene is downstream of hecB and is linked to a functional cluster of hrp/hrc genes in E. chrysanthemi AC4150 (18), as shown in Fig. Fig.11A. The hecA gene was sequenced and mutagenized with Tn7 in E. coli. Several hecA::Tn7 insertions were marker exchanged into AC4150. In mutant CUCPB5124, the hecA ORF is disrupted by Tn7 insertion at nucleotide 7686.

Figure 1
Physical map of the hecA mutation and reduced virulence phenotype of the hecA mutant in N. clevelandii seedlings. (A) The hecA and hecB genes are adjacent to the hrp/hrc cluster in E. chrysanthemi AC4150. (B) The virulence of E. chrysanthemi is ...

hecA Mutation Reduces E. chrysanthemi Virulence in N. clevelandii Seedlings.

To detect any virulence phenotype for hecA mutants, we inoculated various test plants with 1 × 108 cfu per ml of AC4150 and CUCPB5124. We wound inoculated potato tuber slices and chicory leaves and measured the resulting weight of macerated tissue (potato) and size of necrotic lesion (chicory leaf) but could detect no phenotype for the mutant. We then surface inoculated axenic seedlings of Nicotiana tabacum cv Samsun, N. tabacum cv Xanthi, Nicotiana benthamiana, and N. clevelandii. Only N. clevelandii showed a clearly differential response 24 h after inoculation between wild-type AC4150 and its hecA mutant derivative (Fig. (Fig.11B): wild-type strains AC4150 and 3937 caused maceration and total collapse of N. clevelandii seedlings within 24 h, whereas all hecA mutants showed reduced virulence and produced water-soaked lesions in only a few individual leaves at this time. However, seedlings completely succumbed to hecA mutants by 48 h after inoculation, and the wild-type phenotype was restored in hecA mutants with a plasmid construct harboring the hecA and hecB genes (pCPP2394) in trans. The hecA mutant CUCPB5124 was chosen for further study. We also assayed on N. clevelandii seedlings CUCPB5040 (Hrp), CUCPB5006 (PelABCE), and CUCPB5026 (Out) (Fig. (Fig.11B). The Hrp mutant was indistinguishable from the wild type, the Out mutant caused no disease during the 48-h assay period, and the PelABCE mutant was like the HecA mutant in showing partially reduced virulence.

Labeling E. chrysanthemi with GFP Reveals That HecA Enables the Bacterium to Attach to the Surface of N. clevelandii Leaves.

The reduction in virulence associated with the hecA mutation suggested that hecA may mediate an attachment event in pathogenesis. We examined the behavior of E. chrysanthemi AC4150 and CUCPB5124 on the surface of N. clevelandii seedling leaves by epifluorescence microscopy by using bacteria carrying pGFP-TIR (34). As early as 1 h after inoculating seedling leaves, GFP-labeled wild-type strain AC4150 and hecA-complemented strain CUCPB5124(pCPP2394) began to attach to the leaf surface, whereas the hecA mutant strain CUCPB5124 remained planktonic. Inducing bulk flow in water, by pressing on the coverslip, accentuated the planktonic and anchored and immobilized phenotypes, which are presented in the timeline in Fig. Fig.22A. To document with static images the inability of the hecA mutant to attach to the leaf surface, we used confocal laser-scanning microscopy to selectively view those bacteria that were at the plane of the leaf surface rather than in the volume of water further above the leaf. As shown in Fig. Fig.22B, many more AC4150 and CUCPB5124(pCPP2394) than CUCPB5124 cells were localized (immobilized) to the plane of the leaf surface 4 h after inoculation.

Figure 2
HecA-dependent adhesion of GFP-labeled E. chrysanthemi to N. clevelandii seedlings. (A) Epifluorescence microscopy was used to determine the time course of bacterial anchoring (bacteria movement in flowing water was limited and “anchored” ...

To quantify the role of HecA in attachment, we used epifluorescence microscopy to count the number of GFP-labeled bacteria that remained attached after vigorous washing of the inoculated leaves at different time points (Fig. (Fig.22C). This treatment demonstrated that wild-type AC4150 and complementing mutant CUCPB5124(pCPP2394) attach tightly to the plant surface, and that attachment is HecA dependent and increases over time. These results also confirmed the utility of using the 4-h postinoculation time point for additional studies of factors affecting attachment. We also tested for the ability of E. chrysanthemi to attach to the surface of polystyrene and glass by using a standard assay for biofilm formation (51). In contrast to leaves, neither polystyrene nor glass retained bacteria after washing (data not shown).

HecA Is Also Involved in Aggregation.

The attachment studies also revealed that 4 h after inoculation, bacteria began to form aggregates that eventually became large clumps of cells immobilized on the leaf surface. These appeared to develop primarily through recruitment of planktonic cells by freshly attached cells. That is, attachment preceded aggregation, and large aggregates were not observed to be planktonic. However, small groups of planktonic cells (typically two to four) could be seen to move in flowing water as if they were tethered together end to end. Aggregates were observed occasionally in stomates, but they formed predominantly on the surface of the leaf, and 8 h after inoculation there was no evidence that bacteria had invaded the mesophyll.

The formation of aggregates on the surface of the leaves depended on HecA, and the maximum difference between AC4150 and CUCPB5124 in this phenotype was observed 8 h after inoculation. Fig. Fig.33A shows aggregates of AC4150 and CUCPB5124(pCPP2394) on the surface of N. clevelandii seedling leaves that are visible by confocal laser-scanning microscopy. The occurrence of small groups of planktonic bacteria, tethered end to end, also depended on HecA. HecA-dependent aggregation also was observed in bacterial colonies on agar medium after transfer of bacteria to glass slides. Fig. Fig.33B shows bright-field microscopy images of AC4150 and CUCPB5124(pCPP2394) cells that are aggregated, typically as arrays of parallel cells. The hecA mutant CUCPB5124 cells appeared strikingly different in bright-field microscopy, because almost all cells moved independently in the water.

Figure 3
HecA-dependent aggregation of GFP-labeled E. chrysanthemi on the surface of N. clevelandii seedling leaves and during growth on agar medium. (A) Detached leaves from N. clevelandii seedlings were inoculated with AC4150 (wild type), CUCPB5124 (hecA::Tn ...

Inhibition of Adhesion and Aggregation.

To determine whether the HecA-dependent adhesion of E. chrysanthemi to N. clevelandii seedling leaves and aggregation depended on de novo protein synthesis and extracellular protein, we treated the bacteria with either chloramphenicol or proteinase K just before inoculation (Table (Table1).1). We then observed the GFP-labeled bacteria at 4 and 8 h. Incubation of the bacteria with chloramphenicol did not inhibit attachment but did inhibit aggregation, whereas proteinase K prevented both attachment and aggregation. We also tested the attachment and aggregation phenotypes of CUCPB5040 (a Hrp mutant blocked in the type III protein secretion pathway) and CUCPB5026 (an Out mutant blocked in the type II protein secretion pathway). Neither mutant was altered in adhesion or aggregation.

Table 1
Effects of different mutations and treatments on E. chrysanthemi adhesion, aggregation, and epidermal cell killing on N. clevelandii seedling leaves

Cell Death.

When observing E. chrysanthemi AC4150 aggregates on the surface of N. clevelandii seedling leaves by bright-field microscopy, we noted a localized translucence in the underlying tissue. To determine the viability of epidermal cells under these aggregates, we used propidium iodide, which accumulates in the nuclei of cells with compromised cytoplasmic membranes. Fig. Fig.44A summarizes the relative development of macroscopically visible necrosis in N. clevelandii seedlings inoculated with AC4150, CUCPB5124 (HecA), and CUCPB5026 (Out). Using confocal laser-scanning microscopy, we observed the appearance of clusters of dead epidermal cells ≈8 h after inoculation with AC4150 (Fig. (Fig.44B). These clusters were associated with aggregates of wild-type bacteria and were not observed 8 h after inoculation with either CUCPB5124 or CUCPB5026. The failure of Out mutant CUCPB5026 to cause macroscopic necrosis and maceration is likely because of the inability of the bacterium to secrete pectic enzymes. Consistent with this interpretation, CUCPB5006 (PelABCE) did not elicit clusters of propidium iodide-stained epidermal cells at 8 h postinoculation (Table (Table1).1). In contrast, CUCPB5040 (Hrp) retained the wild-type cell killing phenotype. These observations indicate that this early killing of epidermal cells depends on HecA and pectic enzymes but is independent of the Hrp system.

Figure 4
Reduced killing of N. clevelandii seedlings and leaf epidermal cells by E. chrysanthemi HecA and Out mutants. (A) The time course of macroscopically observable necrosis in seedlings was determined after inoculation as described in Fig. ...

hecA Encodes a Protein with Homology to FHA-Related Adhesins in Animal and Plant Bacterial Pathogens.

The hecA ORF is 11.5 kb in size and predicts a 3,850-aa protein. Fig. Fig.55 shows the signal peptide sequence as predicted by signalp (www.cbs.dtu.dk/services/SignalP/), repetitive sequences found by using Statistical Analysis of Protein Sequences (www.ch.embnet.org/software/SAPS_form.html), and other relevant structural features such as Arg-Gly-Asp motifs and Gly-rich regions. The overall structure of all of the FHA-like adhesins found by blastx search (Table 2, which is published as supporting information on the PNAS web site) is similar, so we assumed that the entire proteins are homologous in these taxa. Comparison of the N-proximal 110 aa revealed that the two-partner secretion domain noted in HecA (46) is conserved in all of the HecA homologs recently reported in other plant pathogens (13, 14). The congruency test between the secretion domains and 16S rRNA sequences found the two datasets to be fundamentally incongruent (P = 0.001). Thus, the phylogenetic history of HecA and HecA-related proteins does not correlate with the phylogeny of the bacterial species.

Figure 5
Structural features of E. chrysanthemi HecA. Putative domains of the protein are depicted with gray indicating the N-terminal signal peptide (with cleavage site), black indicating the conserved two-partner secretion domain, asterisks indicating Arg-Gly-Asp ...


The soft-rot diseases caused by E. chrysanthemi and E. carotovora represent the “common cold” of plants in their ubiquity and greater dependence on environmental factors rather than host genotype. Although dynamic aspects of the interaction appear to determine the development of these diseases, little is known of the underlying events, especially those occurring early in pathogenesis. Here, we have found that HecA, a member of a class of adhesins important in several animal pathogens, is important in the virulence of E. chrysanthemi on N. clevelandii seedlings. Closer inspection of the pathogenic defect in hecA mutant CUCPB5124 revealed three events, attachment, aggregation, and epidermal cell killing, which occur in sequence before the appearance of disease symptoms and that appear to represent the earliest steps in E. chrysanthemi pathogenesis on this host. These events and the structure of HecA invite further consideration of the initiation of the pectolytic attack on plant tissues, the multiple activities of HecA, similarities between plant and animal pathogens, and the evolution of FHA-like adhesins among pathogens.

The importance of pectic enzymes in the active maceration phase of E. chrysanthemi pathogenesis seems certain and is supported here by the loss in pathogenicity of mutant CUCPB5026 (which cannot secrete pectic enzymes) and by the delayed virulence of mutant CUCPB5006 (which can produce only the “secondary” Pel isozymes). How pectic enzymes are deployed at the initiation of the pectolytic attack is less clear. Induction of pectic enzyme production by E. carotovora in potato appears to be a social event depending on a quorum of bacteria in the infection site. In contrast, the parasitic benefits of pectic enzyme production for a mixed population of E. chrysanthemi wild-type and nonpectolytic mutants in chicory appear to go only to those individuals that produce the enzymes (52). Using propidium iodide, we determined here that the first N. clevelandii seedling cells to die were epidermal cells associated with aggregates of immobilized bacteria. The formation of these aggregates depended on HecA, and the epidermal killing in turn depended on both aggregation and production of the major Pel isozymes. These observations raise two new questions. First, is aggregation needed for epidermal cell killing because it promotes quorum-dependent pel induction or because it promotes more focused Pel activity? Second, is HecA uniquely important when inoculum is applied to the surface of a susceptible host rather than to a wound?

HecA and related proteins found in other plant pathogenic bacteria are highly diverse. However, they share structural features of FHA-like adhesins, including large size, the presence of an N-terminal signal peptide, a 110-aa N-proximal two-partner secretion domain, Gly-rich regions, repetitive sequences, and multiple Arg-Gly-Asp motifs. Furthermore, the function of HecA is independent of Hrp and Out systems, suggesting that the hecB gene, located upstream of hecA, should encode a functional TpsB transporter family (46) to aid in HecA secretion. HecA contains neither the heparin-nor the carbohydrate-binding domain found in FHA. We do not know yet whether HecA is like FHA in being proteolytically cleaved on secretion, or whether it has any independent biological activity on host cells (although it is not sufficient to kill epidermal cells). The effects of chloramphenicol and proteinase K pretreatment on E. chrysanthemi interactions with N. clevelandii seedlings are consistent with the hecA-dependent attachment ability being conferred by a constitutively expressed extracellular protein whose continued production is required for bacterial aggregation. We observed KdgR-like binding sequences upstream of hecA, but expression of a uidA insertion under the control of these hecA upstream sequences was not induced by polygalacturonic acid, and hecA appears to be expressed constitutively (data not shown).

The factors known so far to contribute to the virulence of both plant and animal pathogens include regulators, protein secretion systems, and stress tolerance mechanisms that operate within the pathogen (53). Extracellular virulence factors that target hosts of both kingdoms are less common. Examples have been found in the multikingdom opportunistic pathogen P. aeruginosa and in a small subset of the effector proteins secreted by the type III systems of different plant and animal pathogens (53, 54). Our data suggest that FHA-like adhesins represent a class of external virulence factors that are broadly important in both plant and animal pathogens. Several things are noteworthy here: (i) The diverse plant pathogens E. chrysanthemi, R. solanacearum (root-infecting wilt pathogen with a type III secretion system), X. fastidiosa (insect vectored xylem-limited pathogen lacking a type III secretion system), and the most recently sequenced Xanthomonas axonopodis pv citri and X. campestris pv campestris (55) all encode at least one FHA homolog, suggesting such proteins may be universal among necrogenic (but not tumorigenic) plant pathogens. (ii) The 19 proteins most closely related to HecA are all from pathogens, further supporting a role for these proteins in host interactions. (iii) The distribution of similarities among these proteins suggests that the encoding genes have been acquired horizontally.

Further support for the horizontal acquisition of FHA-like adhesins is found with HecA and the two adhesins in Y. pestis. The Y. pestis adhesins are the proteins that are most similar to HecA despite differences in host kingdom, and the G+C content of the encoding genes is more than 10% higher than that of the respective genomes for E. chrysanthemi and Y. pestis. This finding suggests that these adhesins were horizontally acquired relatively recently from another pathogen, such as P. aeruginosa, with a higher % G+C genome. In contrast, the G+C content of the genes encoding the other adhesins in Table 2 was approximately the same as the respective genomes. In summary, the ubiquity of FHA-like adhesins among plant pathogens suggests that these proteins provide good starting points for future investigations of early events in the development of many bacterial plant diseases.

Supplementary Material

Supporting Information:


We thank Pablo Rodriguez-Palenzuela for help in testing the efficacy of GFP variants in E. chrysanthemi, Adela Ramos and Ryan Rapp for useful suggestions, and Terrence P. Delaney for assistance with epifluorescence microscopy. This work was supported by U.S. Department of Agriculture National Research Initiative Competitive Grants Program Grant 97-35303-4488 (A.C.) and National Science Foundation Division of Environmental Biology Grant 0089483 (J.J.D.).


hemolysin-like protein in E. chrysanthemi
filamentous hemagglutinin


This paper was submitted directly (Track II) to the PNAS office.

Data deposition: The sequence reported in this paper has been deposited in the GenBank database (accession no. AF501263).


1. Smit G, Swart S, Lugtenberg B J J, Kijne J W. Mol Microbiol. 1992;6:2897–2903. [PubMed]
2. Matthysse A G, Yarnall H A, Young N. J Bacteriol. 1996;178:5302–5308. [PMC free article] [PubMed]
3. Matthysse A G, McMahan S. Appl Environ Microbiol. 1998;64:2341–2345. [PMC free article] [PubMed]
4. Matthysse A G, Yarnall H, Boles S B, McMahan S. Biochim Biophys Acta Gene Struct Expr. 2000;1490:208–212. [PubMed]
5. Chumakov M I. Microbiology. 1996;65:631–643.
6. Reuhs B L, Kim J S, Matthysse A G. J Bacteriol. 1997;179:5372–5379. [PMC free article] [PubMed]
7. Romantschuk M. Annu Rev Phytopathol. 1992;30:225–243. [PubMed]
8. Ojanen-Reuhs T, Kalkkinen N, Westerlund-Wikstrom B, vanDoorn J, Haahtela K, Nurmiaho-Lassila E L, Wengelnik K, Bonas U, Korhonen T K. J Bacteriol. 1997;179:1280–1290. [PMC free article] [PubMed]
9. Van Gijsegem F, Vasse J, Camus J C, Marenda M, Boucher C. Mol Microbiol. 2000;36:249–260. [PubMed]
10. Roine E, Raineri D M, Romantschuk M, Wilson M, Nunn D N. Mol Plant–Microbe Interact. 1998;11:1048–1056. [PubMed]
11. Whatley M H, Hunter N, Cantrell M A, Hendrick C, Keegstra K, Sequeira L. Plant Physiol. 1980;65:557–559. [PMC free article] [PubMed]
12. Hirano S S, Upper C D. Microbiol Mol Biol Rev. 2000;64:624–653. [PMC free article] [PubMed]
13. Simpson A J, Reinach F C, Arruda P, Abreu F A, Acencio M, Alvarenga R, Alves L M, Araya J E, Baia G S, Baptista C S, et al. Nature (London) 2000;406:151–157. [PubMed]
14. Salanoubat M, Genin S, Artiguenave F, Gouzy J, Mangenot S, Arlat M, Billault A, Brottier P, Camus J C, Cattolico L, et al. Nature (London) 2002;415:497–502. [PubMed]
15. Finlay B B, Caparon M. In: Cellular Microbiology. Cossart P, Boquet P, Normark S, Rappuoli R, editors. Washington, DC: Am. Soc. Microbiol.; 2000.
16. Locht C, Bertin P, Menozzi F D, Renauld G. Mol Microbiol. 1993;9:653–660. [PubMed]
17. Locht C, Antoine R, Jacob-Dubuisson F. Curr Opin Microbiol. 2001;4:82–89. [PubMed]
18. Kim J F, Ham J H, Bauer D W, Collmer A, Beer S V. Mol Plant–Microbe Interact. 1998;11:563–567. [PubMed]
19. Bell K S, Avrona A O, Holeva M C, Cardle L, Morris W, De Jong W, Toth I K, Waugh R, Bryan G J, Birch P R J. Microbiology. 2002;148:1367–1378. [PubMed]
20. Bauer D W, Bogdanove A J, Beer S V, Collmer A. Mol Plant–Microbe Interact. 1994;7:573–381. [PubMed]
21. Yang C-H, Gavilanes-Ruiz M, Okinaka Y, Vedel R, Berthuy I, Boccara M, Chen J W-T, Perna N T, Keen N T. Mol Plant–Microbe Interact. 2002;15:472–480. [PubMed]
22. Barras F, Van Gijsegem F, Chaterjee A K. Annu Rev Phytopathol. 1994;32:201–234.
23. Jafra S, Figura I, Hugouvieux-Cotte-Pattat N, Lojkowska E. Mol Plant–Microbe Interact. 1999;12:845–851.
24. Jones S, Yu B, Bainton N J, Birdsall M, Bycroft B W, Chhabra S R, Cox A J R, Golby P, Reeves P J, Stephens S, et al. EMBO J. 1993;12:2477–2482. [PMC free article] [PubMed]
25. Pirhonen M, Flego D, Heikinheimo R, Palva E T. EMBO J. 1993;12:2467–2476. [PMC free article] [PubMed]
26. Nasser W, Bouillant M L, Salmond G, Reverchon S. Mol Microbiol. 1998;29:1391–1405. [PubMed]
27. Beaulieu C, Boccara M, Van Gijsegem F. Mol Plant–Microbe Interact. 1993;6:197–202.
28. Enard C, Diolez A, Expert D. J Bacteriol. 1988;170:2419–2426. [PMC free article] [PubMed]
29. Condemine G, Castillo A, Passeri F, Enard C. Mol Plant–Microbe Interact. 1999;12:45–52. [PubMed]
30. Lopez-Solanilla E, Garcia-Olmedo F, Rodriguez-Palenzuela P. Plant Cell. 1998;10:917–924. [PMC free article] [PubMed]
31. El Hassouni M, Chambost J P, Expert D, Van Gijsegem F, Barras F. Proc Natl Acad Sci USA. 1999;96:887–892. [PMC free article] [PubMed]
32. Pirhonen M, Saarilahti H, Karlsson M-J, Palva E T. Mol Plant–Microbe Interact. 1991;4:276–283.
33. Aldon D, Brito B, Boucher C, Genin S. EMBO J. 2000;19:2304–2314. [PMC free article] [PubMed]
34. Miller W G, Lindow S E. Gene. 1997;191:149–153. [PubMed]
35. Chatterjee A K, Thurn K K, Feese D A. Appl Environ Microbiol. 1983;45:644–650. [PMC free article] [PubMed]
36. Hugouviewux-Cotte-Pattat N, Robert-Baudouy J. J Bacteriol. 1985;162:248–255. [PMC free article] [PubMed]
37. He S Y, Collmer A. J Bacteriol. 1990;172:4988–4995. [PMC free article] [PubMed]
38. Sambrook J, Fritsch E F, Maniatis T. Molecular Cloning: A Laboratory Manual. 2nd Ed. Plainview, NY: Cold Spring Harbor Lab. Press; 1989.
39. Ham J H, Bauer D W, Fouts D E, Collmer A. Proc Natl Acad Sci USA. 1998;95:10206–10211. [PMC free article] [PubMed]
40. Roeder D L, Collmer A. J Bacteriol. 1985;164:51–56. [PMC free article] [PubMed]
41. Gage D J, Bobo T, Long S R. J Bacteriol. 1996;178:7159–7166. [PMC free article] [PubMed]
42. Cormack B P, Valdivia R H, Falkow S. Gene. 1996;173:33–38. [PubMed]
43. Matz M V, Fradkov A F, Labas Y A, Savitsky A P, Zaraisky A G, Markelov M L, Lukyanov S A. Nat Biotechnol. 1999;17:969–973. [PubMed]
44. Taira S, Tuimala j, Roine E, Nurmiaho-Lassila E-L, Savilahti H, Romanstchuk M. Mol Microbiol. 1999;34:736–744.
45. Altschul S F, Gish W, Miller W, Myers E W, Lipman D J. J Mol Biol. 1990;215:403–410. [PubMed]
46. Jacob-Dubuisson F, Locht C, Antoine R. Mol Microbiol. 2001;40:306–313. [PubMed]
47. Thompson J D, Higgins D G, Gibson T J. Nucleic Acids Res. 1994;22:4673–4680. [PMC free article] [PubMed]
48. Gonnet G H, Cohen M A, Benner S A. Science. 1992;256:1443–1445. [PubMed]
49. Dayhoff M O. In: Atlas of Protein Sequence and Structure. Foundation N B R, editor. Vol. 5. Washington, DC: Georgetown Univ.; 1978.
50. Farris S J, Kallersjo M, Kluge A J, Bult C. Cladistics. 1995;10:315–319.
51. O'Toole G A, Pratt L A, Watnick P I, Newman D K, Weaver V B, Kolter R. Methods Enzymol. 1999;310:91–109. [PubMed]
52. Lopez-Solanilla E, Llama-Palacios A, Collmer A, Garcia-Olmedo F, Rodriguez-Palenzuela P. Mol Plant–Microbe Interact. 2001;14:386–393. [PubMed]
53. Cao H, Baldini R L, Rahme L G. Annu Rev Phytopathol. 2001;39:259–284. [PubMed]
54. Cornelis G R, Van Gijsegem F. Annu Rev Microbiol. 2000;54:735–774. [PubMed]
55. da Silva A C, Ferro J A, Reinach F C, Farah C S, Furlan L R, Quaggio R B, Monteiro-Vitorello C B, Van Sluys M A, Almeida N F, Alves L M, et al. Nature (London) 2002;417:459–463. [PubMed]

Articles from Proceedings of the National Academy of Sciences of the United States of America are provided here courtesy of National Academy of Sciences
PubReader format: click here to try


Related citations in PubMed

See reviews...See all...

Cited by other articles in PMC

See all...


Recent Activity

  • HecA, a member of a class of adhesins produced by diverse pathogenic bacteria, c...
    HecA, a member of a class of adhesins produced by diverse pathogenic bacteria, contributes to the attachment, aggregation, epidermal cell killing, and virulence phenotypes of Erwinia chrysanthemi EC16 on Nicotiana clevelandii seedlings
    Proceedings of the National Academy of Sciences of the United States of America. Oct 1, 2002; 99(20)13142

Your browsing activity is empty.

Activity recording is turned off.

Turn recording back on

See more...