ABSTRACT
Extracellular Yersinia disables the immune system of its host by injecting effector Yop proteins into host cells. We show that a Yersinia enterocolitica nonpolar lcrGmutant is severely impaired in the translocation of YopE, YopH, YopM, YpkA/YopO, and YopP into eukaryotic cells. LcrG is thus required for efficient internalization of all the known Yop effectors.
The capacity ofYersinia species (Yersinia pestis, Y. pseudotuberculosis, and Y. enterocolitica) to resist the immune systems of their hosts depends on the Yop virulon, which is encoded by the 70-kb pYV plasmid. This virulon allows extracellular bacteria adhering to the surfaces of eukaryotic cells to inject bacterial proteins into the cytosol in order to disable these cells (9). Translocation of the intracellular effectors (YopE, YopH, YpkA/YopO, YopM, and YopP) across the eukaryotic cell membrane requires at least two other secreted proteins, namely, YopB and YopD (5, 12-14, 19, 22, 28, 33-35). The Yop proteins are secreted outside the bacterial cell by a contact (type III) secretion apparatus called Ysc (1, 2, 4, 10, 16, 18, 23, 24, 37). The translocators YopB and YopD are encoded by a large operon that also encodes LcrV and LcrG (3, 20, 26). LcrG is a 96-amino-acid (11-kDa) protein that appears to be involved in the control of Yop release (31). In addition, LcrG has been shown to bind LcrV, a protein required for the secretion of YopB and YopD (21, 29). Yop secretion occurs only when bacteria are in contact with eukaryotic cells or deprived of Ca2+ ions. A nonpolar lcrG mutant of Y. pestis is Ca2+ blind, secreting large amounts of Yops in the absence as well as in the presence of eukaryotic cells or Ca2+ions, like the yopN mutants of Y. pseudotuberculosis and Y. enterocolitica and thetyeA mutant of Y. enterocolitica (5, 11, 14, 28, 31). In spite of their deregulated phenotype,yopN45 mutant bacteria (i.e., bacteria in which the yopN gene is interrupted after codon 45) can efficiently deliver Yop effectors into the cytosol of eukaryotic cells (5). YopN is thus thought to act at the level of Yop release as the stop valve of the secretion apparatus. TyeA is required for the translocation of a subset of Yop effectors (14). We have recently shown that LcrG can bind to HeLa cells via heparan sulfate proteoglycans and that addition of exogenous heparin can interfere with the translocation of Yops into HeLa cells (7). We inferred that LcrG could have an important role to play in translocation and that interaction with heparan sulfate could affect the activity of LcrG. In this work, we present evidence that LcrG is indeed essential for efficient translocation to occur.
Construction and characterization of an lcrGmutant.To investigate the role of LcrG in the secretion of Yops and their subsequent translocation into eukaryotic cells, we constructed an lcrG nonpolar mutant. First, we inactivated the chromosomal gene encoding β-lactamase A ofY. enterocolitica E40(pYV40) (34) with the mutator plasmid pKNG105 (15) to produce strain MRS40(pYV40). Next, 147 bp (bp 22 to 169) of lcrG were deleted from pMRS22 (Table 1) by site-directed mutagenesis (17) with oligonucleotide MIPA310 (5′-AGTCTTCCCATTTTGATAAGCTAGCGGAGCGCGAG-3′), which is identical to nucleotides 5 to 21 and nucleotides 170 to 187 oflcrG but which changes Pro58 to Leu. The mutated allele of lcrG, called lcrGΔ8–57, was verified by sequencing, cloned in a suicide vector, and introduced into MRS40(pYV40) to create strain MRS40(pMRS4043) (Table 1). The lcrG mutant strain was tested for Ca2+dependency and in vitro Yop secretion (2, 8). The mutant was unable to grow at 37°C in the presence or absence of Ca2+(data not shown) and as such was defined as growth thermosensitive. TheY. enterocolitica lcrG mutant secreted all the Yops in the presence and absence of Ca2+ (Fig.1) and was thus Ca2+ blind, as was previously described for Y. pestis(31). The translocators YopB and YopD, whose genes are situated downstream of LcrG, are efficiently secreted, demonstrating the nonpolarity of the lcrG mutation. Yop secretion was prevented by Ca2+ ions after the introduction of plasmid pMSK23, containing lcrG alone transcribed from theyopE promoter, into MRS40(pMRS4043) (Table 1; Fig. 1). This confirmed the nonpolarity of the lcrGmutation.
Plasmids used in this work
The lcrG mutant strain MRS40(pMRS4043) is Ca2+ blind. Yop secretion by theY. enterocolitica wild-type strain MRS40(pYV40) (lane 1), the lcrG mutant strain MRS40(pMRS4043) (lane 2), and the complemented strain MRS40(pMRS4043)(pMSK23) (lane 3) in the absence (−) and in the presence (+) of Ca2+ was analyzed. Bacteria were grown in brain heart infusion-oxalate or brain heart infusion-Ca2+, and Yop secretion was induced for 4 h at 37°C. Purified Yops were subjected to sodium dodecyl sulfate-polyacrylamide gel electrophoresis and stained with Coomassie blue.
LcrG is involved in the translocation of the YopE cytotoxin.Wild-type Y. enterocolitica induces a cytotoxic response on HeLa cells that is characterized by the rounding up and detachment of the target cells due to the disruption of actin microfilaments (25, 27). After 2 h of infection, thelcrG mutant bacteria were unable to induce this cytotoxicity (data not shown). This observation suggested that the lcrGmutant was impaired in its ability to internalize YopE, the major cytotoxin, inside HeLa cells. To investigate this further, we introduced plasmid pMS111, encoding YopE130-Cya (i.e., a hybrid protein made of 130 residues of YopE fused to Cya), into wild-type Y. enterocolitica, a yscNsecretion mutant, the lcrG mutant, a yopN mutant, and a yopB translocation mutant (Table 1). Cultured PU5-1.8 macrophages were infected with each of these strains in the presence of cytochalasin D. We monitored both the release of hybrid adenylate cyclase into the culture medium and the accumulation of cyclic AMP (cAMP) inside the eukaryotic cells. In good agreement with the Ca2+ blind phenotype, the lcrG andyopN mutant bacteria secreted much more YopE130-Cya into the culture medium than the wild-type strain (5) (Table 2). Hence, the lcrG mutant strain was able to efficiently secrete Yops in the presence of eukaryotic cells, but this Yop secretion was deregulated and probably independent of eukaryotic cell contact. Unlike the yopN mutant bacteria but like the yopB mutant bacteria, the lcrG mutant bacteria were unable to induce high levels of cAMP accumulation in the cytosol of PU5-1.8 macrophages (Table 2). LcrG was thus involved in the delivery of YopE into eukaryotic cells. Introduction of lcrG on plasmid pMSK23 into the lcrG mutant strain resulted in the recovery of the translocation ability of YopE, thus showing that the translocation phenotype was due solely to the defect in the lcrG gene (Table 2). To visualize directly the internalization of YopE inside eukaryotic cells, macrophages infected with wild-type and mutantlcrG isogenic Y. enterocoliticaoverproducing YopE from plasmid pMS3 (32) were subjected to immunostaining and examined by confocal microscopy. YopE appeared dispersed in the cytosol of macrophages infected with the wild-type bacteria but not in the cytosol of cells infected with the mutantlcrG bacteria (Fig.2). Taken together, these results led us to conclude that LcrG is essential for the efficient translocation of YopE across the eukaryotic cell membrane.
Role of LcrG in the translocation of YopE-Cya, YopH-Cya, and YopM-Cya into PU5-1.8 macrophages
Delivery of YopE into macrophages. PU5-1.8 macrophages grown on coverslips were infected for 2 h with Y. enterocolitica MRS40(pPW401)(pMS3), a yopB mutant bacterium overproducing YopE (A); MRS40(pYV40)(pMS3), a wild-type bacterium overproducing YopE (B); or MRS40(pMRS4043)(pMS3), anlcrG mutant bacterium also overproducing YopE (C). The asterisk indicates YopE inside the cytosol of the macrophages. The arrows indicate bacteria. After infection, the cells were fixed, incubated with purified anti-YopE antibodies, stained with fluorescein isothiocyanate-labelled anti-rabbit antiserum, and examined by confocal microscopy. The eukaryotic cell membranes were labelled with wheat germ agglutinin-Texas red. Each panel shows a single optical plane at the level of the nucleus. Note that in panel C, bacteria are heavily stained because of deregulated and depolarized Yop secretion.
We also tested the secretion and translocation phenotypes of aYersinia lcrG yopN double mutant strain (pMRS99) (Table 1). This strain was Ca2+ blind for Yop secretion like thelcrG and yopN individual mutant strains (data not shown). However, the lcrG yopN mutant strain did not significantly translocate YopE130-Cya into macrophages (Table 2). Thus, the function of LcrG is not solely to control the opening of the Yop secretion pore by YopN to allow Yop release and subsequent translocation. Rather, LcrG is itself independently required for optimal translocation of YopE.
LcrG is involved in the internalization of YopH, YopM, YopO, and YopP.We then investigated whether translocation of YopH99-Cya and YopM100-Cya was also dependent on LcrG (Table 1). Although the lcrG mutant bacteria secreted more YopH99-Cya and YopM100-Cya into the culture medium than the wild-type bacteria, they did not induce significant accumulation of cAMP in infected macrophages (Table 2). Thus, the efficient internalization of YopH and YopM was also dependent on the presence of LcrG.
We also wanted to look at the translocation of YopO143-Cya and YopP99-Cya into eukaryotic cells (Table 3). Because these Yops are not translocated as efficiently as YopE, YopH, and YopM, this must be studied in a Y. enterocolitica strain lacking the Yop effectors YopE, YopH, YopO, YopP, and YopM (12, 14, 35). Due to the lack of competition for the secretion and translocation apparatuses, the translocation of the Yop-Cya hybrid is optimized. We thus introduced the lcrGΔ8–57 allele into the Yop effector polymutant strain MRS40(pABL403) (Table 1). As can be seen in Table 3, the translocation of YopO143-Cya, YopP99-Cya, and the other Yop-Cya hybrid proteins was greatly reduced in the lcrG mutant strain compared to that in the parental strain. The level of translocation of each of the hybrid Cya proteins by the polymutantlcrG was almost similar to that of the polymutantyopB strain. Thus, LcrG is involved in the translocation of all the known effector Yops.
Role of LcrG in the translocation of YopO-Cya and YopP-Cya into PU5-1.8 macrophages
Conclusions.The phenotype of the newly constructedY. enterocolitica lcrG mutant is unique. Not only is it Ca2+ blind like the yopN and tyeAmutants (5, 11, 14, 28, 31), but it is also a weak Yop translocator like the yopB and yopD mutants (5, 22, 28). This phenotype clearly shows that LcrG is involved in translocation of all the Yop effectors. This is in contrast to the yopN mutant, which translocates all the Yops efficiently, and the tyeA mutant, which is required for the translocation of only a subset of Yop effectors, namely, YopE and YopH (5, 18).
There are several possibilities regarding the role of LcrG in translocation. LcrG could be an essential element of the translocation machinery along with YopB and YopD. It is also possible that LcrG is an element regulating the deployment of the translocation apparatus or the action of the translocation process itself. We have recently shown that LcrG can bind to HeLa cells via heparan sulfate proteoglycans and that heparin can interfere with the translocation of Yops inside HeLa cells (7). Thus, LcrG could be a Yop apparatus ligand whose interaction with heparan sulfate proteoglycans augments its function in the translocation of Yops into eukaryotic cells. We plan to investigate these possibilities in greater detail in our future work.
ACKNOWLEDGMENTS
We thank D. Desnoeck, I. Lambermont, and C. Kerbouch for excellent technical assistance.
M.R.S. was a recipient of a Sociéte Générale de Belgique fellowship and A.P.B. was a recipient of a Brenninkmeijer fellowship, both awarded by I.C.P. This work was supported by the Belgian FRSM (3.4595.97), the Direction générale de la Recherche Scientifique-Communauté Française de Belgique (ARC 94/99-172), and the Belgium Federal Office for Scientific, Technical and Cultural affairs (PAI 4/03). Confocal microscopy was funded by credit 9.4531.94F from FRSM (Loterie Nationale).
FOOTNOTES
- Received 11 November 1997.
- Returned for modification 31 December 1997.
- Accepted 2 February 1998.
- Copyright © 1998 American Society for Microbiology
