INTRODUCTION

Top Abstract Introduction Materials and Methods Results Discussion References

The hrp ("harp") genes encode type III secretory pathways and are required by many phytopathogenic bacteria to elicit a hypersensitive response (HR) on nonhost or resistant host plants and for pathogenesis on susceptible hosts. The HR is a rapid localized death of the host cells that occurs upon pathogen infection and, together with the expression of a complex array of defense-related genes, is a component of plant resistance. The hrp genes were first identified in Pseudomonas syringae pv. phaseolicola, a bean pathogen (38). Since then, hrp genes from a variety of plant pathogenic bacteria, including Erwinia, Pseudomonas, Ralstonia, and Xanthomonas, have been characterized (for reviews, see references 2, 9, and 11). The specific functions of the hrp pathway in pathogenesis are not known. However, type III secretion pathways of animal and plant pathogens have been demonstrated to mediate the secretion of virulence factors into the extracellular melieu. Some of the proteins ultimately end up in the host cell cytoplasm (reviewed in references 11, 25, and 37). In mediating the interaction of the bacterium and the host plant, the hrp pathway presumably acts to prevent or inhibit a general resistance response or otherwise enhance the colonization of the plant by the bacteria.

Given the importance of the type III systems to pathogenicity, it can be expected that analysis of the systems in different species will provide insight into the adaptation of the species to their respective host plants. The hrp gene clusters appear to group into two types on the basis of sequence relatedness and operon organization (reviewed in reference 9). The hrp genes of Pseudomonas and Erwinia comprise one group, and the hrp genes of Xanthomonas campestris pv. vesicatoria and Ralstonia solanacearum comprise the second group. Our present understanding of the group 2 hrp genes is based almost entirely on the characterization of two strains representing R. solanacearum and X. campestris pv. vesicatoria. The genus Xanthomonas itself is comprised of a large number of different species and pathovars that colonize over 392 species of plants (36). Xanthomonas oryzae pv. oryzae is the causal pathogen of bacterial leaf blight on rice (54). We report here the cloning of a hrp cluster from X. oryzae pv. oryzae and sequence analysis of the left end of the region. In the process of characterization, we identified two novel loci that are associated with the hrp cluster and an insertion sequence (IS) element not previously characterized in X. oryzae pv. oryzae.

	MATERIALS AND METHODS

Top Abstract Introduction Materials and Methods Results Discussion References

Bacterial strains and plasmids. The bacterial strains and plasmids used in the experiments are listed in Table 1. Escherichia coli strains were grown in Terrific broth (TB) or on Luria agar plates at 37°C with the appropriate antibiotics. X. oryzae pv. oryzae strains were cultured on tryptone sucrose agar or in TB at 28°C. The genomic library of PXO86 in pHM1 was described previously (24). Carbenicillin, spectinomycin, and kanamycin were used at 100 µg/ml.

Recombinant DNA techniques. DNA manipulations were performed by standard procedures (6). Restriction enzymes, T4 DNA ligase, Klenow fragment of DNA polymerase I, and Taq polymerase were purchased from Life Technologies, Inc. (Gaithersburg, Md.) and Fisher Scientific (St. Louis, Mo.). Chemicals were purchased from Sigma Chemical (St. Louis, Mo.) and Fisher Scientific. BioTrace HP membrane (Gelman Sciences, Ann Arbor, Mich.) was used for Southern blot hybridization. EcoRI DNA fragments from p23-44, a cosmid clone containing hrp genes of X. oryzae pv. oryzae, were subcloned into pBluescript KS(+) (Stratagene, Inc., La Jolla, Calif.). The DNA sequences of the two strands of these subclones were determined by the DNA Sequencing Facility of Iowa State University. Amino acid alignments were constructed using Clustal W 1.7 (55). Similarity searches were performed using the BLAST program (4). Identity and similarity comparisons among proteins were performed using the Fasta program (45). Potential signal peptides at the N-terminal and transmembrane domains of the proteins were predicted by PSORT (41, 42) and TopPred2 (60), respectively. Specific motif searches were performed using the MOTIF program (43).

Transposon and deletion mutagenesis. Mutagenesis of p23-44 with Tn5-gusA1 was performed in E. coli strain DH5MCR. Cells growing exponentially in TB were infected with bacteriophage pA1 carrying Tn5-gusA1 (51). After 1 h of incubation at 37°C, which allowed the phage to be absorbed and phenotypic expression of kanamycin resistance, bacteria were plated on Luria agar medium supplemented with kanamycin and incubated overnight at 28°C. Plasmid DNA was extracted from single colonies, digested with EcoRI, and analyzed by electrophoresis on 1% agarose gels. Each plasmid with a transposon in the bacterial genomic DNA of the clone was transformed into PXO99^A by electroporation or by conjugation using the conjugal helper strain S17-1 (53). Marker exchange mutagenesis was performed using spontaneous homologous recombination and screening for kanamycin-resistant, spectinomycin-sensitive clones. Mutations were confirmed by Southern blot analysis of bacterial genomic DNA using ³²P-labeled probes for subclones of p23-44. Four isolates from each marker exchange were tested on tomato and rice plants. The p23-44 cosmid was reintroduced by electroporation into the hrp mutants for complementation tests.

Plant assays. Pathogenicity and hypersensitivity assays were performed as described previously (24). Tomato cv. VFN8 was used for the nonhost hypersensitivity test. Ten-day-old rice seedlings IRBB10 and IRBB7, containing corresponding resistance genes Xa10 and Xa7, respectively, were used for race-specific resistance assays. IR24 was used for pathogenicity testing. All plants were grown in growth chambers at 28°C (daytime) and 25°C (nighttime) with a 14-h photoperiod and 85% humidity. For hrp phenotype assays, inoculum concentrations were adjusted to an optical density at 600 nm of 1.0 (approximately 2 × 10⁹ CFU/ml) using a DU-64 spectrophotometer (Beckman Instruments). Inoculum concentrations for hpa1 and hpa2 mutation phenotype assays were adjusted to an optical density at 600 nm of 0.3. The differences between wild-type and mutant pathogenicity and hypersensitivity reactions were enhanced at the lower dilution. Growth of bacteria on rice after infiltration was monitored as previously described (24).

Sequence analysis of PXO99^A. A 651-bp region was amplified by PCR from PXO99^A using the primers 5'-GATTGTCTGCGGAAAATAG-3' (IS99FOR) and 5'-GGTACGCAGCAGATCTGGG-3' (IS99REV), cloned into pCR2.1-TOPO (52), and sequenced. The parameters used for PCR were as follows: step 1, 95°C for 2 min; step 2, 50°C for 30 s; step 3, 72°C for 90 s; step 4, 95°C for 30 s; step 5, 35 cycles from step 2 to step 4; step 6, 72°C for 2 min.

Southern hybridization analysis. Total DNA isolation and Southern hybridization analysis were as previously described (34). Probes were prepared by either PCR (49) or gel purification and random priming (6). The 408-bp EcoRI/XhoI fragment from hpa1 was used as a probe for hpa1-related sequences. A 321-bp probe for hpa1 was generated by PCR using internal primers 5'-AACAGGATCCAGATTGCTTCGAAGAGGCTGCC-3' (HPA2F1) and 5'-AACAAGGATCCGCATATTTATCACGCTCC-3' (HPA2R1). A 1522 base pair probe for IS1114 was amplified using primers 5'-AGTCGCCCCTGAAAAACCCCCAG-3' (ISHRPF1) and 5'-AAGTCGCCCCTGAAAAACCCTC-3' (ISHRPR1). PCR conditions for both probes were as described above. Probes were labeled using a Rediprime random-priming labeling kit (Amersham, Arlington Heights, Ill.).

Nucleotide sequence accession numbers. The DNA sequence for the hpa2-to-hrpB region has GenBank accession no. AF232057. The sequence for the region including IS1114 from PXO86 is under GenBank accession no. AF232058. The sequence of the corresponding region of IS1114 insertion in PXO99^A is under GenBank accession no. AF232714.

	RESULTS

Top Abstract Introduction Materials and Methods Results Discussion References

Isolation of the hrp region of X. oryzae pv. oryzae. Cosmids p23-14, p23-44, and p2-2 were recovered from a genomic library of strain PXO86 by using p83-15, which contained a portion of the X. campestris pv. vesicatoria hrp region (10), as a probe. The maps of the clones were characterized by endonuclease restriction digests, Southern blotting, and DNA sequence analysis and shown to cover the regions corresponding to hrpA, hrpB, hrpC, and hrpD of X. campestris pv. vesicatoria (Fig. 1). Thirty-five strains with Tn5-gusA1 insertions in the region covered by p23-44 were generated from strain PXO99^A of X. oryzae pv. oryzae. Eleven strains, with insertions in the 4.8- and 7.5-kb EcoRI fragments, had unaltered hrp gene function. Twenty-four strains, with insertions in the 8.3-, 1.4-, and 5.1-kb EcoRI fragments, lost the ability to elicit disease on rice and an HR on tomato (Fig. 1). The mapping data indicated that the insertions in hrp loci were located in the hrpB, hrpC, and hrpD operons. None of the Tn5-gusA1 insertions appeared to have inserted into the hrpA operon. Therefore, two insertions in hrpA were generated in pK4.0 by using pGPS-1 and recombined into the hrp region (Fig. 1, insertions 3230 and 2785). Both mutations resulted in hrp mutant phenotypes. The phenotypes of a representative hrp mutant after inoculation on rice and tomato are shown in Fig. 2A. Introduction of p23-44 into selected hrp mutants restored the ability to elicit disease on rice and an HR on tomato (Fig. 2A). In planta growth of a representative hrp mutant was reduced compared to that of the wild-type strain, while the levels of in planta growth of the mutant strains carrying p23-44 were similar to that of the parent strain PXO99^A (Fig. 2B).

View larger version (16K): [in this window] [in a new window]   FIG. 1.   Restriction fragment map of the hrp region in X. oryzae pv. oryzae. The name of each fragment also indicates the size in kilobases. Vertical lines above the map indicate positions of Tn5-gusA1 insertion. Lines with filled circles indicate that the insertion abolished hrp function; lines without circles indicate insertions that did not cause loss of hrp activity. Arrows indicate positions and orientations of hrp operons. E, EcoRI; K, KpnI.

View larger version (42K): [in this window] [in a new window]   FIG. 2.   Effects of mutations in the hrp genes of X. oryzae pv. oryzae. (A) Phenotype of rice (cultivar IR24) and tomato leaf (VFN8) reactions to the following strains: 1, PXO99A; 2, PXO99A9mx (hrpC); 3, PXO99A9mx (p23-44). Water soaking and disease symptoms were evidenced by discoloration at the inoculation site on rice leaves (left). Inoculation in the hrp mutants resulted in no change in leaf coloration (leaf 2). HR on tomato leaf (right) is indicated by light gray patches (leaves 1 and 3). (B) Effect of hrp mutation on growth of bacteria in rice leaves. Numbering as for panel A.

Two novel hrp-associated loci are located in the left end of the hrp cluster. The DNA sequence was determined for 12,165 bp from the BglI site in the 7.5-kb EcoRI fragment (E7.5) and extending 713 bp into the 1.4-kb EcoRI fragment of p23-44 (Fig. 3). The sequence data were organized into two portions. The first portion, starting at the second SalI site in E7.5, contained 10,096 bp and included transcription units A and B and two additional open reading frames, tentatively termed hpa1 and hpa2 (Fig. 3). Table 3 gives the positions and properties of the noted features in the first portion. The second portion, containing 2,075 bp and the IS element IS1114, extended upstream from the SalI site to a BglII site in the E7.5 fragment (Fig. 3).

View larger version (17K): [in this window] [in a new window]   FIG. 3.   Region of sequence analysis of the left hrp region of PXO86. Transcriptional units and open reading frames of the hrp genes are indicated by lines and open arrows, respectively. The direction of transcription and translation is indicated by direction of the arrow. Filled bars indicate sequenced regions. E, EcoRI; K, KpnI.

View larger version (57K): [in this window] [in a new window]   FIG. 4.   DNA sequence of hpa1. Sequences for a PIP box, restriction sites, and a ribosome binding site (SD) are underlined. The sequence of the deduced translation product is given in the single-letter code below the DNA sequence. The glycine-rich regions are double underlined. Asterisks indicate a potential transmembrane motif.

View larger version (62K): [in this window] [in a new window]   FIG. 5.   Sequence analysis of hpa2. (A) Sequence alignment of Hpa2 and selected proteins. Bkpm, Burkholderia pseudomallei (GenBank accession no. AAD05172); IagB, Salmonella enterica serovar Typhimurium (39); IpgF, Shigella flexneri (3); VirB1, Agrobacterium tumefaciens (40); LyzC, chicken (28). The conserved A and E -helices (B and E) and -sheet () in the structure of lysozyme are overlined. Asterisks indicate conserved asparagine and aspartate residues in the catalytic region of lysozyme. Gaps in the alignment are represented by dashes. Hpa2 is shown as starting from the isoleucine that is 16 residues upstream of the putative cleavage site. The triangle indicates the putative cleavage site. (B) Promoter region of hpa2. A HindIII site and imperfect PIP box (PIP-1) are underlined. Only amino acids starting from the last methionine before the conserved region are indicated. The predicted signal peptide in the protein product is underlined.

An IS element is located adjacent to hpa2 in PXO86. An IS element was identified 254 bp from the end of hpa2. The element is 1,519 bp long and bounded by an 18-bp perfect inverted repeat. IS1114 was similar in sequence to two recently described insertion elements from Rhizobium and X. campestris pv. campestris (Fig. 6A) (13, 50). The repeat was located within a 6-bp direct repeat (TAAAA) that may have been generated by the insertion process (Fig. 6B). Strain PXO99^A did not contain IS1114 adjacent to hpa2, and the sequence analysis of the same region revealed only the TAAAA sequence with no duplication (Fig. 6B). The facts that the TAAAA sequence, which is duplicated at the ends of IS1114 in PXO86, was not duplicated in PXO99^A and that no remnant of IS1114 was found at the corresponding region of PXO99^A suggest that the insertion is unique to the PXO86 lineage. The sequence from the BglII site to the end of hpa2 without the IS element did not match any entries in the GenBank database. Therefore, the element also does not appear to have inserted into an identifiable genetic locus. Genetic loci that are involved in pathogenicity have been associated with transmissible genetic elements (30). These so-called pathogenicity islands may also have G+C compositions that differ from the G+C content of the bulk of the chromosomal DNA, reflecting the transfer among different species of bacteria (reviewed in reference 22). However, no evidence for a dramatic shift in the G+C content at the element's boundary with the hrp region was observed. We therefore could not find that the element had any relevance to hrp function. We believe that IS1114 is present near the hrp region in some lineages and probably inserted into the region by chance.

View larger version (38K): [in this window] [in a new window]   FIG. 6.   Sequence analysis of IS1114. (A) Sequence similarity of IS1114 to insertion elements IS1478a (13) and ISRm220-13-5 (50). The alignment shows only 43 to 51 bases of sequence from one end, with the overall sequence identity of the entire element to IS1114 given in parentheses. The arrow indicates an 18-bp repeat. Identities in two of the three elements are shaded. (B) Sequence alignment of genomic DNA from strain PXO99A with the region of insertion in strain PXO86. Only the 18-bp direct repeats are shown (boldface). Dashes indicate spaces introduced for optimal alignment. A direct 6-bp repeat at the site of insertion in PXO86 is underlined.

Deletion of the left end affects virulence. A deletion mutant (PXO99^A2) covering the hpa1-hpa2 region was constructed by replacing the region between the SalI and XhoI sites of pK6.0B with the gene for resistance to kanamycin and subsequent recombination into the chromosome of PXO99^A to give strain PXO99^A2 (58). PXO99^A2 was therefore missing hpa2 entirely, and hpa1 was truncated. Upon inoculation to rice, PXO99^A2 was found to cause reduced disease symptoms and, when harboring avrXa10, elicited a weak HR on rice plants with resistance gene Xa10 in terms of the intensity of browning (Fig. 7A). The reduced disease symptoms were accompanied by reduced bacterial populations in the leaf tissue (Fig. 7B). Similarly, the HR of the mutant on tomato was delayed and weaker in terms of the area that collapsed after inoculation (not shown). The reduced pathogenicity of the mutant could be complemented by reintroduction of p23-44 (Fig. 7B).

View larger version (34K): [in this window] [in a new window]   FIG. 7.   Effect of an hpa1-hpa2 deletion on virulence and avrXa10 activity. (A) Phenotypes of PXO99A2 and PXO99A2(pFWX10-F2), which contains the avirulence gene avrXa10, on susceptible (IR24) and resistant (BB10) rice cultivars. Susceptible leaves (upper two leaves) and resistant leaves (lower two leaves) were photographed 3 days after inoculation with the strain indicated at the right. (B) Growth of PXO99A2 in susceptible rice cultivar IR24. 1, PXO99A; 2, PXO99A2; 3, PXO99A2(p23-44).

View larger version (12K): [in this window] [in a new window]   FIG. 8.   Map of single-gene mutations in hpa1 and hpa2. Insertion sites are indicated by lines with circles at the top. Filled circles indicate no loss of virulence; open circles indicate insertions that reduced virulence. The filled bar indicates the region missing in deletion 2. IS1114 is indicated by the hatched bar. B, BglII; E, EcoRI; H, HindIII; K, KpnI; S, SalI; X, XhoI.

Distribution of hpa1, hpa2, and IS1114 in species of Xanthomonas. To determine the prevalence of the newly identified elements in a variety of other strains of X. oryzae pv. oryzae and X. campestris, gene-specific and element-specific probes were generated from hpa1, hpa2, and IS1114 by PCR or gel purification of an internal restriction fragment. The probes were then used in Southern analyses of genomic DNA. Both hpa1 and IS1114 hybridized to DNAs from all of the strains of X. oryzae pv. oryzae and the one strain of X. oryzae pv. oryzicola that were tested (Fig. 9A and B, lanes 1 to 5). Signal for hpa1 was detected in DNA from X. campestris pv. alfalfae KX-1, X. campestris pv. malvacearum H, and X. campestris pv. phaesoli SC-4A and was not detected in X. campestris pv. vesicatoria 81-23, X. campestris pv. campestris KXXC-1, or X. campestris pv. holcicola (Fig. 9A, lanes 6 to 11). Outside X. oryzae, IS1114 was detected only in X. campestris pv. campestris KXXC-1 and X. campestris pv. malvacearum H (Fig. 9B, lanes 7 and 10). In contrast to hpa1 and IS1114, hpa2 was detected in DNAs from all of the tested strains (Fig. 9C).

View larger version (55K): [in this window] [in a new window]   FIG. 9.   Southern analysis of hpa1, hpa2, and IS1114. (A) DNA probed with the EcoRI/XhoI fragment of hpa1. (B) DNA probed with the PCR fragment from IS1114. (C) DNA probed with the PCR fragment from hpa2. Genomic DNA was extracted from 11 strains of Xanthomonas spp. and digested with EcoRI (A and B) or SalI (C). Lanes 1, X. oryzae pv. oryzae PXO86; lanes 2, X. oryzae pv. oryzae PXO99A; lanes 3; X. oryzae pv. oryzae PXO177; lanes 4, X. oryzae pv. oryzae C191; lanes 5, X. oryzae pv. oryzicola BLS 303; lanes 6, X. campestris pv. vesicatoria 81-23; lanes 7, X. campestris pv. campestris Kxxc1; lanes 8, X. campestris pv. alfalfae KX-1; lanes 9, X. campestris pv. holcicola; lanes 10, X. campestris pv. malvacearum H; lanes 11, X. campestris pv. phaesoli SC-4A. See Materials and Methods for PCR primers and conditions. Numbers at the left indicate kilobase pairs.

	DISCUSSION

Top Abstract Introduction Materials and Methods Results Discussion References

The results presented here demonstrate that a type III secretory system, also known as the hrp system, has a critical role for pathogenicity of X. oryzae pv. oryzae on rice. Type III secretory systems play central roles in the ability of many gram-negative bacteria to colonize plant and animal hosts. In general, the systems are envisaged to direct the assembly of a supramolecular secretory apparatus similar in structure to the flagellar apparatus, which is also the product of a type III secretory system. Some components of the type III systems are conserved due to the structural requirements, while other components can be expected to reflect the adaption of the system to the particular niche of the bacterium (26). The hrpA and hrpB operons represent part of the conserved core of genes necessary for the assembly of the hrp system. The predicted proteins of the hrpA and hrpB operons had 94% or higher amino acid residue identity with the proteins from X. campestris pv. vesicatoria (10, 57). HrpA is a member of the HrcC class of hrp proteins and is localized to the outer membrane, where it is thought to function the transporter past the outer membrane layer (63). As a class, HrpA proteins have similarity to a variety of proteins from other type III systems, and in fact, HrpA ancestry can be traced to other secretory systems (63).

HrpB3, HrpB5, HrpB6, and HrpB8 of the hrpB operon are related to components of other type III systems, including the flagellar assembly pathway (8). HrpB8 and HrpB3, for example, are similar in amino acid sequence to FliR and FliF, which have been determined to be components of the inner membrane basal body and M-ring portion of the flagellar secretory apparatus, respectively (16, 27). HrpB1, HrpB2, HrpB4, and HrpB7 are proteins that, on the basis of sequence similarity, are unique to Xanthomonas and R. solanacearum. These proteins either have diverged from the ancestral secretion system or represent unique adaptations of the type III system in Xanthomonas and R. solanacearum. Divergence between the latter four proteins in X. oryzae pv. oryzae and X. campestris pv. vesicatoria was no greater than that for proteins with recognizable counterparts in other type III systems, and they therefore appear not to have diverged along species lines as a possible consequence of adaptation to the particular host plants.

The hrp mutations in X. oryzae pv. oryzae resulted in the loss of pathogenicity and prevented the elicitation of the nonhost HR on tomato and the race-specific HR on incompatible rice plants due to the avirulence genes avrXa10 and avrXa7, which are homologs of avrBss3 from X. campestris pv. vesicatoria (24). Activity due to avrBs3 was also lost when X. campestris pv. vesicatoria was hrp deficient (32), and dependence on hrp function for avirulence activity could be bypassed when avrBs3 and other members of the family from X. campestris pv. malvacearum were expressed in the plant cells (15, 56). In a separate study, we have observed a 75% reduction in the number of transformation foci after particle bombardment of resistant rice leaves with a plant-expressed copy of avrXa10 (66). Therefore, the protein products of avrXa10 and avrXa7 along with possible virulence factors are likely to be secreted by a hrp-encoded type III secretory apparatus into the cells of the rice plant.

Despite the high degree of similarity, the analysis of the left end of the hrp region of X. oryzae pv. oryzae revealed two genes, named hpa1 and hpa2, that were not found in X. campestris pv. vesicatoria. The putative hpa1 product has an amino acid composition similar to the compositions of the harpin proteins of P. syringae pathovars and Erwinia species and the harpin-like PopA protein of R. solanacearum (5, 20, 23, 61). These proteins share regions of high glycine content and are secreted via the type III pathway. Whether Hpa1 is secreted is unknown at present. Conditions for hrp-dependent secretion by X. campestris pv. vesicatoria have recently been determined (48). Thus, it may be possible to adapt the conditions to X. oryzae pv. oryzae and determine if the hpa1 product is secreted in a hrp-dependent manner. Like popA, hpa1 has a PIP box immediately upstream of the coding sequence and is likely to be regulated by the hrpXo gene product, which is a member of the AraC class of transcription factors (29, 44, 62).

Harpins gained attention by the fact that some harpins and PopA can elicit hypersensitive reactions on certain plants simply by injection of the proteins into the leaf tissue (5, 23, 46, 61). The protein product of hrpW has similarity to both harpin and pectate lyase (12, 20, 31). However, the elicitor properties are not shared by all harpins, and in the case of harpin from P. syringae pv. tomato, the protein elicited an HR on the host plant (46). Mutations in hrpZ, hrpW, and popA have no observable effect on the pathogenicity (1, 5, 12). Therefore, the biological relevance of the latter genes or the elicitor activities of their products is unknown. On the other hand, mutations in hrpN, from which harpins were originally named, abolished pathogenicity of Erwinia amylovora (61). Mutations in hpa1 reduced pathogenicity due to X. oryzae pv. oryzae. hpa1 is the only gene from Xanthomonas with a harpin-like product. However, sequence similarity with the hrpW gene from P. syringae pv. tomato in the DNAs from several Xanthomonas species was recently reported (12). Thus, Xanthomonas, like P. syringae and Erwinia species, may produce a variety of harpin-like proteins depending on the species or strain.

The role of the hpa2 locus, which encodes a lysozyme-like protein, is unclear. However, an argument can be made that the locus is another core component of the type III system. The locus appears to be present in all of the species of Xanthomonas that were examined, and related genes can be found in association with a variety of secretory systems, including type III systems of animal pathogens. On the one hand, the mutation of hpa2 had no apparent effect on pathogenicity under the conditions of the assay. Indeed, whether this protein is produced in vivo remains in question. The lack of a phenotype with the hpa2 mutants may reflect the lack of a requirement for this gene under the conditions of testing or environment, yet the locus may be required under natural infection conditions. None of the lysozyme-related proteins appear to be essential for the related systems under the conditions of testing. Mutations in virB1 severely reduced but did not eliminate tumor formation (40). Similarly, mutations in gene 19, which encodes a lysozyme-like protein in the R1-16 plasmid conjugation pathway, reduced but did not eliminate conjugative transfer of R1-16 (7). A mutation in ipgF, which is a locus in a type III pathway of Shigella flexneri, was reported to have no effect on pathogenicity (3). One obvious possibility for the function of the lysozyme-like proteins would be in the degradation of the peptidoglycan layer of the bacterial cell wall to accommodate the respective secretory apparatuses or associated pili (19, 33, 47). This possibility remains to be tested.