A surface protease and the invasive character of plague

Substitution of two histidine residues in YadA protein of Yersinia enterocolitica abrogates collagen binding, cell adherence and mouse virulence

The plasmid-encoded surface protein YadA of Yersinia enterocolitica mediates binding to diverse extracellular matrix (ECM) proteins, adherence to epithelial cell lines, resistance to complement lysis, autoagglutination, and is required for mouse virulence. Using site-directed mutagenesis we attempted to analyse the relationship between structural domains and functions of YadA. In a first approach we could abrogate collagen binding by chemical modification of histidyl residues of YadA protein. This result prompted us to substitute histidyl residues (His) of conserved regions of YadA protein of Y. enterocolitica O8 by tyrosine residues using site-directed mutagenesis. Substitution of His-156 and His-159 (YadA-2 mutant) resulted in abrogation of binding to ECM proteins, of cell adherence, and in reduction of mouse virulence, whereas autoagglutination, serum complement resistance and oligomer formation remained unaffected. A striking result was obtained from the orogastric mouse-infection model: the YadA-2 mutant retained the ability to colonize the small intestine and to invade and multiply within the Peyer's patches but was impaired in colonizing mesenteric lymph nodes and spleen in comparison to the wild-type strain.

Differential effects of deletions in lcrV on secretion of V antigen, regulation of the low-Ca2+ response, and virulence of Yersinia pestis

The Yersinia pestis V antigen is necessary for full induction of low-calcium response (LCR) stimulon virulence gene transcription, and it also is a secreted protein believed to have a direct antihost function. We made four nonpolar deletions in lcrV of Y. pestis to determine if secretion, regulation, and virulence functions could be localized within the V antigen (LcrV). Deletion of amino acids 25 to 40 caused secretion of LcrV to be decreased in efficiency; however, removal of residues 108 to 125 essentially abolished LcrV secretion. Neither mutation had a significant effect on LCR regulation. This showed that LcrV does not have to be secreted to have its regulatory effect and that the internal structure of V antigen is necessary for its secretion. Both mutants were avirulent in mice, showing that the regulatory effect of LcrV could be separated genetically from its virulence role and raising the possibility that residues 25 to 40 are essential for the virulence function. This study provides the best genetic evidence available that LcrV per se is necessary for the virulence of Y. pestis. The repressed LCR phenotype of a mutant lacking amino acids 188 to 207 of LcrV raised the possibility that the deleted region is necessary for regulation of LCR induction; however, this mutant LcrV was weakly expressed and may not have been present in sufficient amounts to have its regulatory effect. In double mutants containing this mutant lcrV and also lacking expression of known LCR negative regulators (LcrG, LcrE, and LcrH), full induction of the LCR occurred in the absence of functional LcrV, indicating that LcrV promotes induction not as an activator per se but rather by inhibiting negative regulators.

Evidence for two evolutionary lineages of highly pathogenic Yersinia species

Sensitivity to Yersinia pestis bacteriocin pesticin correlates with the existence of two groups of human pathogenic yersiniae, mouse lethal and mouse nonlethal. The presence of the outer membrane pesticin receptor (FyuA) in mouse-lethal yersiniae is a prerequisite for pesticin sensitivity. Genes that code for FyuA (fyuA) were identified and sequenced from pesticin-sensitive bacteria, including Y. enterocolitica biotype 1B (serotypes O8; O13, O20, and O21), Y. pseudotuberculosis serotype O1, Y. pestis, two known pesticin-sensitive Escherichia coli isolates (E. coli Phi and E. coli CA42), and two newly discovered pesticin-sensitive isolates, E. coli K49 and K235. A 2,318-bp fyuA sequence was shown to be highly conserved in all pesticin-sensitive bacteria, including E. coli strains (DNA sequence homology was 98.5 to 99.9%). The same degree of DNA homology (97.8 to 100%) was established for the sequenced 276-bp fragment of the irp2 gene that encodes high-molecular-weight protein 2, which is also thought to be involved in the expression of virulence by Yersinia species. Highly conserved irp2 was also found in all pesticin-sensitive E. coli strains. On the basis of the fyuA and irp2 sequence homologies, two evolutionary groups of highly pathogenic Yersinia species can be established. One group includes Y. enterocolitica biotype 1B strains, while the second includes Y. pestis, Y. pseudotuberculosis serotype O1, and irp2-positive Y. pseudotuberculosis serotype O3 strains. E. coli Phi, CA42, K49, and K235 belong to the second group. The possible proximity of these two iron-regulated genes (fyuA and irp2), as well as their high levels of sequence conservation and similar G+C contents (56.2 and 59.8 mol%), leads to the assumption that these two genes may represent part of an unstable pathogenicity island that has been acquired by pesticin-sensitive bacteria as a result of a horizontal transfer.

The outer surface protein A of the spirochete Borrelia burgdorferi is a plasmin(ogen) receptor

The spirochete Borrelia burgdorferi is the causative agent of Lyme borreliosis (Lyme disease) and is transmitted to mammalian hosts by tick vectors. In humans, the bacteria induce a complex disease, which involves the skin, joints, heart, and nervous system. However, the pathogenic principles of this multisystem illness are far from being understood. To disseminate from the site of the tick bite and invade multiple organ sites, spirochetes have to penetrate normal tissue barriers, such as vascular basement membranes and other organized extracellular matrices. Substantial evidence from other invasive bacterial infections suggest that spirochetes may use endogenous or host-derived enzymes--in particular, proteinases--for this purpose. Here we show that B. burgdorferi binds human plasmin(ogen)--mainly via its outer cell surface lipoprotein A. Binding of plasminogen to spirochetal receptor leads to an accelerated formation of active plasmin in the presence of host-derived plasminogen activator. The cell-surface-associated plasmin cannot be regulated by the serum inhibitor alpha 2-antiplasmin and degrades high-molecular-weight glycoproteins, such as fibronectin. It is suggested that the acquisition of host-derived proteinase plasmin(ogen) contributes to the pathogenicity of B. burgdorferi.

Plasminogen, absorbed by Escherichia coli expressing curli or by Salmonella enteritidis expressing thin aggregative fimbriae, can be activated by simultaneously captured tissue-type plasminogen activator (t-PA)

Curli are fimbrial structures expressed by Escherichia coli that specifically interact with matrix proteins such as fibronectin and laminin. Similar structures are also expressed by Salmonella enteritidis and have been denoted thin aggregative fimbriae. Bacteria expressing curli and thin aggregative fimbriae were found to bind radiolabelled plasminogen as well as the tissue-type plasminogen activator (t-PA). By contrast, E. coli carrying a gene locus with an insertionally inactivated chromosomal curlin subunit were unable to bind the two human proteins. The purified subunit polypeptides of curli and thin aggregative fimbriae bound plasminogen and t-PA with high affinity (1 x 10(8) to 2 x 10(8) M-1). The binding of plasminogen and t-PA to curli-expressing E. coli was only partially inhibited by fibronectin and laminin. Plasminogen absorbed from human plasma by curli-expressing E. coli was readily converted to plasmin by t-PA; both plasmin and t-PA were functionally active when bound to the bacteria. A simultaneous binding of fibrinolytic proteins and matrix proteins to fimbriae of E. coli and S. enteritidis could provide these pathogens with both adhesive and invasive properties.

Passive immunity to yersiniae mediated by anti-recombinant V antigen and protein A-V antigen fusion peptide

LcrV (V antigen), a known unstable 37.3-kDa monomeric peptide encoded on the ca. 70-kb Lcr plasmid of Yersinia pestis, Yersinia pseudotuberculosis, and Yersinia enterocolitica, has been implicated as a regulator of the low-calcium response, virulence factor, and protective antigen. In this study, lcrV of Y. pestis was cloned into protease-deficient Escherichia coli BL21. The resulting recombinant V antigen underwent marked degradation from the C-terminal end during purification, yielding major peptides of 36, 35, 34, and 32 to 29 kDa. Rabbit gamma globulin raised against this mixture of cleavage products provided significant protection against 10 minimum lethal doses of Y. pestis (P < 0.01) and Y. pseudotuberculosis (P < 0.02). To both stabilize V antigen and facilitate its purification, plasmid pPAV13 was constructed so as to encode a fusion of lcrV and the structural gene for protein A (i.e., all but the first 67 N-terminal amino acids of V antigen plus the signal sequence and immunoglobulin G-binding domains but not the cell wall-associated region of protein A). The resulting fusion peptide, termed PAV, could be purified to homogeneity in one step by immunoglobulin G affinity chromatography and was stable thereafter. Rabbit polyclonal gamma globulin directed against PAV provided excellent passive immunity against 10 minimum lethal doses of Y. pestis (P < 0.005) and Y. pseudotuberculosis (P < 0.005) but was ineffective against Y. enterocolitica. Protection failed after absorption with excess PAV, cloned whole V antigen, or a large (31.5-kDa) truncated derivative of the latter but was retained (P < 0.005) upon similar absorption with a smaller (19.3-kDa) truncated variant, indicating that at least one protective epitope resides internally between amino acids 168 and 275.

How to become a pathogen

For most bacterial species, virulence is viewed as a derived state, whereby pathogens acquire certain loci and are rendered virulent. The majority of virulence genes in Salmonella are present in closely related nonpathogenic species, and most genes known to be confined to the salmonellae are not essential for virulence. Alternative evolutionary scenarios may explain the origins of pathogenicity in Salmonella.

Binding of tissue-type plasminogen activator (t-PA) to Neisseria meningitidis and Haemophilus influenzae

Forty-nine bacterial strains representing five species known to interact with human plasminogen were tested for the ability to bind the two major human plasminogen activators, t-PA and urokinase. The bacterial species tested included Haemophilus influenzae, Neisseria meningitidis, Streptococcus pyogenes, Streptococcus equisimilis and human group G streptococci. All N. meningitidis and 11 of 14 H. influenzae strains displayed substantial binding of t-PA with values in the range of 20-46%. On the contrary, none of the streptococcal strains bound significant amounts of tPA. With urokinase no binding could be found for any of the bacterial species tested. Scatchard analysis with a selected H. influenzae strain (HI23354) demonstrated 10,000 receptors per bacterium for t-PA with a Kd value of about 20 nmol l-1. The corresponding values with a selected N. meningitidis strain (Mo 52) was 8500 receptors per bacterium and 70 nmol l-1. t-PA binding could be reduced about 40% by the addition of 10 mmol l-1 of the lysine analogue epsilon-aminocaproic acd (EACA) whereas no inhibitory effect could be demonstrated with arginine. Addition of 2 mumol l-1 of plasminogen which is enough to occupy all bacterial sites for plasminogen did not interfere with the t-PA binding, suggesting that the receptors for t-PA and plasminogen are distinct. Using very high plasminogen concentrations however, t-PA binding could be reduced by about 50% possibly due to an interaction between t-PA and plasminogen in the fluid phase. Our results demonstrate the occurrence of a previously unknown type of bacterial receptor that is capable of specifically binding t-PA.

The pigmentation locus of Yersinia pestis KIM6+ is flanked by an insertion sequence and includes the structural genes for pesticin sensitivity and HMWP2

The pigmentation (Pgm+) phenotype of Yersinia pestis includes a number of different characteristics which appear to be associated with a 102 kb segment of chromosomal DNA known as the pgm locus. In Y. pestis KIM6+, the pgm locus is flanked by direct copies of a repeated element that probably plays a role in the spontaneous deletion of this region. We have sequenced the ends of these elements and shown that they have features in common with bacterial insertion sequences. In addition we show that a clone, pSDR498, from the pgm locus of KIM6+ restores pesticin sensitivity and the iron-regulated expression of three polypeptides, 240 kDa, 190 kDa, and 68 kDa in size, to Pgm- cells. In vitro transcription/translation assays and Escherichia coli minicells were used to analyse the products encoded by various subclones of pSDR498. Pesticin sensitivity mapped to a 5.9 kb fragment that encodes a 68 kDa protein derived from a 72 kDa precursor. Synthesis of the 190 kDa protein was restored by a 19.2 kb clone, indicating that the structural gene for this protein also resides within the pgm locus of Y. pestis KIM6+. Finally, a survey of our Pgm- strains indicates that 97% have also deleted the sequences encoding the 190 kDa protein and pesticin sensitivity.

Streptokinase activates plasminogen bound to human group C and G streptococci through M-like proteins

An ability to interact with plasminogen or plasmin could provide micro-organisms with a mechanism for invasion. Thus, group A, C and G streptococci secrete streptokinase which binds and activates plasminogen. Some streptococci also express surface structures which bind plasminogen without causing its activation. Plasminogen-binding surface proteins were extracted from one group C and one group G streptococcal isolate. Both proteins were found to bind plasmin, fibrinogen and serum albumin in addition to plasminogen. Gene fragments encoding the streptococcal proteins were amplified by PCR and were subsequently cloned and expressed in Escherichia coli. DNA sequence determination revealed for both genes open reading frames encoding proteins which contained repetitive domains and a carboxyl-terminal unrepeated region that were typical of M and M-like proteins. Though the amino-terminal regions of the group C and G streptococcal proteins demonstrated a rather high overall similarity between themselves, they were not similar to the variable regions of other M-like proteins with one exception: there was a 46% identity between the first 22 amino acids of the group G streptococcal protein and the corresponding sequence of PAM, the plasminogen-binding M-like protein of type M53 group A streptococci. Like the proteins extracted from the streptococci, the recombinant proteins bound plasminogen, fibrinogen and albumin. The three plasma proteins bound to separate sites on the streptococcal M-like proteins. Plasminogen bound by the group C and G streptococcal proteins was readily activated by streptokinase, providing evidence for a functional link between the secreted plasminogen-activator and proteins exposed on the bacterial surface.

Plasmid regulation and temperature-sensitive behavior of the Yersinia pestis penicillin-binding proteins

Six major bands corresponding to penicillin-binding proteins (PBPs) with molecular weights ranging from 43,000 to 97,000 were detected in cell envelopes of Yersinia pestis EV76 grown at 28 degrees C. When cells were transferred to 37 degrees C and incubated for extended periods of time, the amounts of all PBPs, except for PBP2, were gradually reduced in cell envelopes of a strain carrying a 75-kb virulence-associated plasmid (as measured by penicillin-binding capacity), whereas in a strain cured of the plasmid, all PBPs were stable. The results indicated that the stability and/or the expression of Y. pestis PBPs is affected by a temperature-inducible pathway associated with the virulence-associated plasmid.

Capturing host plasmin(ogen): a common mechanism for invasive pathogens?

Plasmin is a potent enzyme that can dissolve blood clots and degrade extracellular matrix proteins. A number of pathogenic bacteria produce plasminogen activators. Many of these organisms can also bind plasmin(ogen) to surface receptors and protect the active enzyme from physiological inhibition. Cell-surface localization of plasmin may be a common mechanism used by bacteria to facilitate movement through normal tissue barriers.

Families of serine peptidases

This chapter examines families of serine peptidases. Serine peptidases are found in viruses, bacteria, and eukaryotes. They include exopeptidases, endopeptidases, oligopeptidases, and omega peptidases. On the basis of three-dimensional structures, most of the serine peptidase families can be grouped together into about six clans that may have common ancestors. The structures are known for members of four of the clans, chymotrypsin, subtilisin, carboxypeptidase C, and Escherichia D-Ala-D-Ala peptidase A. The peptidases of chymotrypsin, subtilisin, and carboxypeptidase C clans have a common “catalytic triad” of three amino acids—namely, serine (nucleophile), aspartate (electrophile), and histidine (base). The geometric orientations of these are closely similar between families; however the protein folds are quite different. The arrangements of the catalytic residues in the linear sequences of members of the various families commonly reflect their relationships at the clan level. The members of the chymotrypsin family are almost entirely confined to animals. 10 families are included in chymotrypsin clan (SA), and all the active members of these families are endopeptidases. The order of catalytic residues in the polypeptide chain in clan SA is His/Asp/Ser.

New method for plague surveillance using polymerase chain reaction to detect Yersinia pestis in fleas

Yersinia pestis, the plague bacillus, infects a variety of mammals throughout the world and is transmitted by fleas. We developed a polymerase chain reaction (PCR) test using primers designed from the Y. pestis plasminogen activator gene to directly detect plague-infected fleas. As few as 10 Y. pestis cells were detected, even in the presence of flea tissue, by PCR and then agarose gel electrophoresis and ethidium bromide staining. The feasibility of the assay was demonstrated by using naturally infected Xenopsylla cheopis fleas. The detection of Y. pestis in fleas by PCR provides a rapid and sensitive way to monitor plaque in wild animal populations, allowing public health officials to better assess the potential risk of transmission to humans.