<<

Proc. Natl. Acad. Sci. USA Vol. 91, pp. 489-493, January 1994 Microbiology A cytolysin encoded by Salmonella is required for survival within (/pathogenesis) STEPHEN J. LIBBY*tt, WERNER GOEBEL§, ALBRECHT LUDWIG§, NANCY BUCHMEIER1, FRANCES BOWEl', FERRIC C. FANGt, DONALD G. GUINEYt, J. GLENN SONGER**, AND FRED HEFFRON*II *Department of Molecular Biology, The Research Institute of Scripps Clinic, 10666 North Torrey Pines Road, La Jolla, CA 92037; §Theodor-Boveri-Institut fOir Biowissenschaften, (Biozentrum) der Universitat Warzburg, Leherstuhl fur Microbiologie, Am Hubland, W-8700 Wurzburg, Germany; tDepartment of Medicine, University of California, San Diego, 9500 Gilman Drive, La Jolla, CA 92093-0613; lDepartment of Pathology V151, University of California, San Diego, 3350 La Jolla Village Drive, La Jolla, CA 92161; IlDepartment of Microbiology and 3181, Oregon Health Science University, S.W. Sam Jackson Park Road, Portland, OR 97201; and "*Department of Veterinary Science, University of Arizona, Tucson, AZ 85721 Communicated by Stanley Falkow, August 16, 1993

ABSTRACT A SalmoneUla gene encoding a cytolysin has was used in all virulence studies (18). Clinical isolates of been identified by screening for on blood agar. DNA Salmonella were obtained from the State ofCalifornia Health sequence analyses together with genetic mapping in Salmonella Laboratory or the County of San Diego Health Laboratory. suggest that it is unrelated to other or . The were cultivated in Luria-Bertani medium. Blood gene (slyA) is present in every strain of Salmonella examined, agar plates were made in trypticase soy agar (Difco) contain- in Shigela, and in enteroinvasive but not in ing 4% defibrinated sheep red blood cells (Colorado Serum, other Enterobacteriaceae. SlyA (salmolysin) purified from a Denver). Clinical isolates obtained for homology studies derivative of the original clone has hemolytic and cytolytic included Yersinia sp., Legionella, Chlamydia, Pasteurella, activity and has a molecular weight predicted by the DNA Acinetobacter, Haemophilus, Proteus, Klebsiella, Neisseria sequence. The median lethal dose and infection kinetics in mice sp., Citrobacter, Campylobacter, Franciscella, Brucella, suggest that the toxin is required for virulence and facilitates , Serupla sp., Serratia, Enterobacter sp., E. coli K-12 SalmoneUa survival within mouse peritoneal macrophages. LE392 and DH5a, and Aeromonas. Many Gram-negative and Gram-positive pathogenic micro- Molecular Techniques. A Sau3A1 partial-digest cosmid organisms produce toxins or hemolysins that lyse eukaryotic library of S. typhimurium 14028s was constructed in the cells and contribute to their pathogenicity (1). The most vector pLAFR2 (19, 20) and packaged into A phage particles intensively studied exotoxins produced by Gram-negative (Stratagene). The host for the cosmid library, E. coli LE392 are members of the RTX family (reviewed in refs. or DH5a (21), was grown on trypticase soy agar plates 2 and 3). The HlyA , present in uropathogenic containing sheep red blood cells and tetracycline (20 ug/ml) Escherichia coli, is the best studied of these. HlyA increases and incubated for 36 hr at 37°C. A strongly hemolytic colony E. coli virulence in a rodent peritonitis model (4). Although was chosen for further characterization. Subcloning, se- the precise mechanism of action is unclear, it has been quencing, and Southern analysis were performed by standard suggested that these toxins attenuate host phagocytic techniques. Hemolytic activity was subcloned on a 1.4-kb function (4-8). Several facultative intracellular pathogens, EcoRV-Cla I fragment (pSL1117). This fragment was se- including rickettsiae, shigellae, Trypanosoma cruzi, and List- quenced and an open reading frame from the Cla I end was eria monocytogenes, escape from the phagocytic vesicle of identified. Oligonucleotide primers (Fig. 1, Oligo 1 and Oligo professional phagocytic cells, but Salmonella does not (9- 3) were used to amplify a 680-bp fragment. This fragment was 12). A defined role in virulence is difficult to assign to most cloned into the EcoRV site of pSK and transformed into toxins, in part because ofthe lack of animal models for many DH5a, and the bacteria were plated onto blood agar. Cells organisms. An exception is listeriolysin made by L. mono- harboring this clone (pSL2070) were as hemolytic as those cytogenes. The toxin dissolves the phagocytic membrane, harboring pSL1117. Sequence manipulation and data-base allowing the bacteria to escape into the cytoplasm, and the searches were done with IBI MACVECTOR and GenBank bacteria are protected from the immune defenses ofthe host. (update 71, August 1992). Listeriolysin-negative mutants are avirulent (13-17). To complement the sly mutation in SL2161 for We have occasionally observed hemolysis by certain Sal- survival studies, the 680-bp fragment from pSL2070 was monella strains, particularly low-passage clinical isolates. removed by HindIII/BamHI digestion, made blunt-ended The gene (slyA) encoding the Salmonella hemolysin (salmol- with Klenow DNA polymerase, and cloned into the blunt- ysin) has been determined by screening a Salmonella typhi- ended EcoRI site of pSC101 (22). Colonies containing the murium gene bank in E. coli for hemolytic activity. Several cloned salmolysin gene in pSC101 (pSL2185) were hemolytic. related clones were identified, one of which was studied in Plasmid constructs were electroporated into S. typhimurium detail.tt A mutation in slyA was constructed, and the mutant wild-type (14028s) and salmolysin mutant (SL2161) cells. was found to be attenuated for both virulence in mice and The slyA gene was mapped on the S. typhimurium chro- growth in murine macrophages. mosome by using the Mud-P22 phage of Benson and Gold- man (23). Phage were induced with mitomycin C and pelleted MATERIALS AND METHODS from the resulting lysates, and DNA was prepared for slot blots and hybridized with the PCR-amplified salmolysin Strains, Media, and Culture Conditions. Wild-type Amer- structural gene. Hybridizations were done at 65°C in 6x ican Type Culture Collection strain S. typhimurium 14028s SSPE (lx is 0.15 M NaCl/0.01 M phosphate/l mM EDTA,

The publication costs of this article were defrayed in part by page charge tTo whom reprint requests should be sent at the t address. payment. This article must therefore be hereby marked "advertisement" ttThe sequence reported in this paper has been deposited in the in accordance with 18 U.S.C. §1734 solely to indicate this fact. GenBank data base (accession no. U03824). 489 Downloaded by guest on September 29, 2021 490 Microbiology: Libby et al. Proc. Natl. Acad. Sci. USA 91 (1994)

- OLIGO 1 - AT'C 100 OLIGO 2 Po TGAGAATA CACTAGGITC 200 K K L E S P L G S D SspI CTGGCACGG9rGGTGCGCAT_GGCGTGCTCTGATTGACC T CA TACTrGCACAATATTCATC 300 L A R L V R I W R A L I D H R L K P L E L T Q TH W V T L H N I H AATTGCCGCCTGACCAGTC GTTGGATCAACTTAAGATAAGGGGCT 400 Q L P P D Q S Q I Q L A K A I G I E Q P S L V R T L D Q LE D K G L AATTTCGCGGCAAACCTGCGCCAGCGATCGTCGCGCTAAGCGTAAACGACCACGACC 500 I S R Q T C A S D R R A K R I K L T E K A E P L I A E M E E V I H SspI AAAAC GCGCAr CCAACTGAACACAATATTAT_GAATTGCACT 600 KT R G F I L A G I S S E E I F L L I K L I A K L E H N I X E L H CTCACGATGCAGGGGCATACGTGTGGCCATGTGACCACACGTAAAGCCTG TT¶rAGCGTQGAGAGACGGTAACCTGGCTGCCGTTGCTGGCCAG 700 S H D **' 4 OLIGO 3 CACGACACGCTGACCTGCCG 800

FIG. 1. Sequence of the slyA gene. The initiation codon is underlined and the termination codon is denoted by three stars. The Ssp I sites used to construct pSL2145 are shown. Oligonucleotide primers used for PCRs are shown (Oligos 1-3). The amino acid sequence (one-letter symbols) is shown below the nucleotide sequence. pH 7.4)/0.5% SDS/lOx Denhardt's solution containing de- mM Tris HCl (pH 7.5) and containing phenylmethanesulfonyl natured salmon testis DNA at 200 ,ug/ml. Blots were washed fluoride (2 ug/ml). Column fractions were assayed for he- in O.lx SSPE/0.1% SDS at 65°C for 1-2 hr and exposed to molytic activity by adding 50 Al of each fraction to 700 ,ul of x-ray fim. 10% defibrinated sheep erythrocytes and incubated at 37°C Disruption of the slyA Gene. A detailed description of the for 1 hr. The amount of released was determined suicide vector system used to mutate slyA will be reported by measuring A595 of the supernatant. Salmolysin was eluted separately. In brief, slyA was disrupted by homologous at 0.15 M NaCl and could be further purified by preparative recombination insertion of a suicide vector derived from the isoelectric focusing with a Rotofor (Bio-Rad). All hemolytic RK2 replicon, pRRlOAtrfA (24), containing an internal Ssp I activity focused at a pl of 5.5. A hemolytic unit of salmolysin fragment of slyA. The suicide vector-slyA fragment, was defined as the amount of partially purified salmolysin pSL2145, was maintained in E. coli S17-1 (25) and transferred required to lyse 50% of a suspension of 10% erythrocytes at to wild-type Salmonella by conjugation. Transconjugants 37°C in 1 hr. One hemolytic unit of salmolysin was heated to were selected on brilliant green agar containing sulfadiazine 65°C for 15 min and then assayed for activity. EDTA (50 mM) (80 ,g/ml) and penicillin (300 ug/ml). Chromosomal DNA was added to 1 hemolytic unit of salmolysin and treated as from a penicillin-resistant colony now called SL2161 was above. Phospholipase D and oxidase activities digested with Pst I, transferred to a membrane, and probed were assayed as described (30). with the 32P-labeled PCR-amplified slyA gene to confirm the interruption of slyA. Macrophage Survival and Virulence Determination of sly RESULTS Mutants. Overnight cultures of wild-type (14028s) and sly- Cloning the Salmolysin Gene. Although Salmonella is not (SL2161) cells were grown in Luria-Bertani medium with and usually hemolytic on laboratory culture media, we had oc- without penicillin (200 ,g/ml), respectively, and given orally casionally observed weak zones of hemolysis around some to mice in 200-,ul doses with a feeding cannula. Intraperito- colonies, especially in low-passage clinical isolates. This neal and intravenous inoculations were given with bacteria variable phenomenon might be regulated and the hemolysin washed and diluted in phosphate-buffered saline. Median expressed only in vivo. We therefore attempted to clone the lethal doses (LD50 values) were determined over a 4-week cognate gene away from its normal regulatory elements. A period. To study the course ofinfection, tissues from infected cosmid bank was made from S. typhimurium 14028 and mice were homogenized in sterile water with a Stomacher screened for hemolytic activity on blood agar plates in E. coli (Tekmar, Cincinnati), and the homogenates were diluted in LE392 (20). The average insert size was =25 kb, and about phosphate-buffered saline and plated on Luria-Bertani agar 1 in 100-200 colonies exhibited weak hemolysis. This fre- to determine the number of bacteria per organ. Proteous quency suggests that the gene is present in a single copy in the peptone-elicited peritoneal macrophages were prepared from S. typhimurium genome. BALB/c mice, cultured, and infected as described (26). A series of deletion constructs from a representative he- Partial Purification of Recombinant Salmolysin. Oligonu- molytic clone were used to identify the location of the gene cleotide primers 2 and 3 (Fig. 1) were used to amplify a 515-bp encoding the hemolysin. A 1.4-kb EcoRV-Cla I fragment fragment with a new Nde I site modified at the start codon by (pSL1117) ofthe original 25-kb cosmid clone was required for PCR. This fragment was cloned into pRK172 (27) at the Nde hemolytic activity in E. coli. Exonuclease III deletions ini- I site, placing salmolysin expression under the control of the tiated from the Cla I end abolished hemolytic activity, T7 RNA polymerase promoter. This construct, pSL2042, suggesting that the salmolysin structural gene was near this was transformed into E. coli BL21(DE3) (28) and expression end (data not shown). The nucleotide sequence of the 1.4-kb of salmolysin was induced by 1 mM isopropyl ,B-D- fragment was determined and analysis of the sequence ad- thiogalactopyranoside in the presence of rifampicin (200 jacent to the Cla I site revealed a 436-base open reading frame pg/ml) and [35S]methionine (50 ,uCi/ml; 1 ,Ci = 37 kBq). that could encode a of 16,747 Da. PCR was used to were separated by SDS/12% PAGE and gels were amplify a 680-bp region from chromosomal DNA encompass- stained with Coomassie brilliant blue, dried, and exposed to ing the open reading frame, the amplified DNA was cloned in x-ray fim. To purify salmolysin for biochemical studies, E. pKS, and transformants were plated onto blood agar. The coli cells containing pSL1117 were subjected to osmotic sequence of this region is shown in Fig. 1. This clone, shock (29). Proteins were separated by FPLC Mono Q anion pSL2070, was equally hemolytic as pSL1117, indicating that exchange (Pharmacia) with a gradient of 0-1 M NaCl in 50 the 680-bp fragment contained all the genetic infor- Downloaded by guest on September 29, 2021 Microbiology: Libby et al. Proc. Natl. Acad. Sci. USA 91 (1994) 491 mation necessary for the hemolytic phenotype in E. coli. The of T7 RNA polymerase) and [35S]methionine and then were deduced protein sequence was used to search a translated separated by SDS/PAGE. A single 16-kDa protein was version of GenBank (August 1992), but no significant short specifically labeled (Fig. 3), in close agreement with the size homologies were found. Thus, a single gene, which we call of salmolysin predicted from the DNA sequence. To isolate slyA, for salmolysin, is sufficient to confer hemolytic activity functional salmolysin from E. coli containing cloned slyA on laboratory strains of E. coli. (pSL1117), overnight cultures were osmotically shocked Mapping and Conservation of the Salmolysin Gene in Sal- (25). Released salmolysin was partially purified by FPLC monella. The location of the gene on the Salmonella chro- Mono Q anion-exchange chromatography (Pharmacia); all mosome was determined by a modification of a published hemolytic activity was eluted at 150 mM NaCl. A 16-kDa procedure (23, 31). This procedure is based on Mud-P22 protein, identical in size to the protein overexpressed under hybrid phages inserted at defined locations in the Salmonella the control of T7 RNA polymerase, was enriched in this chromosome. After induction, each phage will package an fraction (Fig. 3). Identically processed osmotic-shock super- -3-min contiguous portion of the Salmonella chromosome. natants of E. coli with pKS alone had no protein in this size DNA is then purified from each of these phages, bound to a range that was eluted at 150 mM NaCl, nor was hemolytic membrane, and probed (see Materials and Methods). Using activity detected in any fraction. Salmolysin has an isoelec- this procedure, we mapped the slyA gene to a region between tric point of pI 5.5 determined by preparative isoelectric 28.5 and 30 min. Few genes have been identified in this focusing. The hemolytic activity of salmolysin was found to be sensitive to heat (65°C for 15 min) and insensitive to location, either in Salmonella or in the related region in E. protease inhibitors (phenylmethanesulfonyl fluoride, leupep- coli. tin, and pepstatin). Salmolysin was fully active in the pres- DNA was extracted from the parent strain, 17 clinical ence of 50 mM EDTA, unlike the hemolysin from uropatho- Salmonella isolates, three Shigella serotypes, enteroinvasive genic E. coli (32). No phospholipase or cholesterol oxidase E. coli, E. coli K-12 laboratory strains, and other pathogenic activity was detected. Experiments that will be reported bacteria (see Materials and Methods for a list). Southern elsewhere suggest that salmolysin is a pore-forming toxin hybridization using the 680-bp fragment as a probe showed able to lyse a variety of cultured mammalian cells (S.J.L., R. that the gene was present in all serotypes of Salmonella Benz, and W.G., unpublished data). On the basis of these examined, Shigella, and enteroinvasive E. coli (Fig. 2) but results, we call salmolysin a cytolysin. not in any of the other bacterial species examined. A single Salmolysin Is Required for Survival in Macrophages and for 7-kb Pst I fragment hybridized to DNA from most serotypes Virulence. The role of salmolysin in virulence was examined ofSalmonella. Three serotypes contained two fragments that by constructing a defined slyA- mutant of S. typhimurium. hybridized, suggesting that the slyA gene may be duplicated The chromosomal slyA gene of S. typhimurium 14028s was in some serotypes of Salmonella (Fig. 2). Preliminary exper- disrupted with an RK2-based suicide vector. Compared with iments suggest that the slyA cognates in Shigella and entero- the isogenic wild-type parent, SL2161 (slyA-) demonstrated invasive E. coli are on extrachromosomal plasmids (data not no difference in in vitro growth rate, cell or colonial mor- shown). These data show that the slyA gene is conserved phology, ability to grow on minimal medium containing only within serogroups of Salmonella. salts and glucose (M9 with 0.2% glucose), sensitivity to acid, Biological Properties of Salmolysin. The slyA gene product ability to invade and transcytose epithelial cell monolayers, was determined by using the PCR primers shown in Fig. 1 to or ability to plaque the temperate phage P22 (data not shown). place the coding region under the control of T7 RNA poly- The LD50 values of SL2161 were determined by the oral, merase. A 519-bp fragment was amplified, cloned into (intragastric), intraperitoneal, and intravenous routes of in- pRK172, and transformed into E. coli BL21(DE3) (28). fection after 28 days of infection. The parent 14028s has an Proteins were metabolically labeled in the presence of iso- LD50 of less than 10 organisms when administered intraperi- propyl ,B-D-thiogalactopyranoside (which induces expression toneally or intravenously and of 6 x 105 organisms when a b c de f g h i j k I m n o p q r given intragastrically. The LD50 of a slyA- mutant (SL2161) 1 2 3 4 5 6

8k b__

5kb -*o 5 kb_v-p :::.. ..: :.. : kDa 29.0 -

20.4 - 14.0- 2.5 kb-_ 6.1 - FIG. 3. Size determination and partial purification of salmolysin. The expression of salmolysin was placed under the control of the FIG. 2. Conservation of the slyA gene. Samples of chromosomal inducible T7 RNA polymerase promoter (see Materials and Meth- DNA from clinical isolates of Salmonella were digested with Pst I ods). Lanes 1 and 2, respectively, Coomassie blue-stained total and probed with the labeled PCR-amplified slyA fragment. A single, protein from uninduced and induced E. coli BL21(DE3)/pSL2042 7.0-kb fragment hybridized to the probe with the exception of S. (arrow indicates a protein of 16 kDa); lanes 5 and 6, autoradiograph worthington, S. infantis, and S. enteritidis. The slyA gene may be of lanes 1 and 2; lanes 3 and 4, osmotic-shock supernatants from E. multicopy in these serotypes. A 5-kb fragment hybridized in the S. coli DH5a containing pKS (lane 3) and supernatants from E. coli hadar sample. Lanes a-r, respectively: S. typhimurium 14028s, S. DH5a containing pSL1117 (lane 4) prepared identically and sepa- typhimurium clinical, S. typhimurium clinical, S. heidelberg, S. rated by Mono Q anion exchange chromatography with a gradient of montevideo, S. infantis, S. choleraesuis, S. newport, S. enteritidis, 0-1 M NaCl. A 16-kDa protein, indicated by the arrow, was enriched S. havana, S. hadar, S. barta, S. typhi, S. typhi 21A, S. worthington, in the 150 mM NaCl fraction obtained from cells expressing cloned S. cerro, S. arizona, and S. typhisuis. slyA (lane 4). Downloaded by guest on September 29, 2021 492 Microbiology: Libby et al. Proc. Natl. Acad. Sci. USA 91 (1994) Table 1. Lethal-dose determination of salmolysin mutants Bacteria Dose No. of surviving mice Intraperitoneal infection 14028s 1 x 103 0/6 SL2161 1 x 103 6/6 1 X 105 6/6 Oral infection 14028s 1 x 105 0/6 SL2161 1 x 107 6/6 1 x 109 6/6 Intravenous infection 14028s 4 x 102 0/4 SL2161 4 x 104 4/4 Six-week-old female BALB/c mice were administered S. typhi- 0 2 4 6 8 10 12 14 16 18 murium wild type (14028s) or slyA- mutant (SL2161) as described in Time, hours Materials and Methods. Mouse survival numbers were determined after 21 days. FIG. 5. Survival characteristics of the slyA- mutant strain SL2161 in murine elicited peritoneal macrophages. Wild-type 14028s is >105 for intraperitoneal administration and >109 when (x), 14028s/pSC101 (a), MS7953 (phoP) (v), SL2161/pSC101 (o), administered orally to mice (Table 1). This represents a and SL2161 (pSL2185-slyA gene) (o) were used to infect macro- difference of >10,000-fold intraperitoneally and >1000-fold phages. Samples were taken at the times indicated and plated onto orally. Later experiments suggested that the oral LD50 value Luria-Bertani agar to determine the number of viable bacteria. was >2 x 1010. The salmolysin mutant is avirulent by be the result of polar effects on another gene required for or intravenous infection at a dose of >104 bacteria. involved in virulence. We have addressed this possibility by The infection kinetics of the slyA- mutant were very complementing the slyA mutation with a plasmid construct different from those of wild-type Salmonella (Fig. 4). Mice containing the 680-bp slyA structural gene cloned in pSC101 were infected orally with about 2 x 1010 organisms of wild (22) (pSL2185). Complementation with pSL2185 resulted in type (14028s) or the slyA- mutant (SL2161). Within 1 day, restoration of wild-type survival within proteose peptone- 103-104 bacteria of the parent strain were present in the elicited mouse peritoneal macrophages as compared with the Peyer's patches and mesenteric lymph nodes, and lower but isogenic control with the vector alone (Fig. 5). However, we significant numbers were present in the liver and spleen. were unable to complement the virulence defect of SL2161 in During the next 6 days, exponential growth ofthe parent was mice by using the same clone. Nonetheless, our results observed in the liver and spleen, resulting in death of the confirm that salmolysin plays a key role in the ability of mice. In contrast to this, low numbers of SL2161 reached the Salmonella to survive and replicate within murine macro- mesenteric lymph nodes and Peyer's patches (about 100 phages and that disruption of the slyA gene profoundly versus 103-104). Lower numbers of SL2161 reached the liver attenuates Salmonella virulence. and spleen as compared with the wild-type parent. The salmolysin mutant grew very poorly in the liver and some- what more slowly in the spleen. Similar results were seen DISCUSSION following intravenous infection (data not shown). Thus, the We have identified an essential virulence factor in Salmo- sly lesion is highly attenuating and the most striking effect is nella, salmolysin, a hemolytic toxin. This toxin is required for to render the salmolysin mutant vulnerable to some compo- survival within peritoneal macrophages and for virulence in nent of the liver and spleen. a mouse infection model. The slyA gene has no significant Survival ofthe slyA- mutant was compared with that ofthe homologies to other reported genes, nor does it contain any parent within peritoneal murine macrophages (Fig. 5) (26, 33, readily identifiable motifs such as the RTX motiffound in the 34). The parent strain is resistant to the bactericidal effects of E. coli hemolysin (2). slyA is present in all Salmonella, macrophages, whereas a phoP mutant is extremely sensitive Shigella, and enteroinvasive E. coli strains but is absent from to these cells (18, 35, 36). The slyA- and phoP mutants were many other bacteria. Thus, salmolysin may define a new equally sensitive in these assays, suggesting that salmolysin family of hemolysins. Although we cannot formally rule out plays a direct role in the survival of Salmonella within that slyA encodes a positive regulator ofa cryptic E. coli K-12 macrophages. The avirulence ofthe slyA mutation might also hemolysin, we feel that this is unlikely. Despite the plethora

10' 10i5

Im 106,

co0 FIG. 4. Course of infection of a salmolysin mutant. Mice were inoculated orally with 2 x 1010 0 104, wild-type (14028s) (Left) and slyA 0 102 mutant (SL2161) (Right) bacteria. C.)E v 103, Mice were sacrificed on the days indicated. Spleens (o), mesenteric 0 lymph nodes (m), Peyer's patches 102 I (o), and livers (v) were dissected II and homogenized, and the number I . l * of bacteria in each organ was de- 10 1. 1.6 1 2 3 4 5 0 4 128 16 termined by plating homogenates Tlme after Infection, days onto Luria-Bertani agar. Downloaded by guest on September 29, 2021 Microbiology: Libby et al. Proc. Natl. Acad. Sci. USA 91 (1994) 493 of virulence genes and toxins that have been cloned in E. coli We thank Virginia Miller for probing blots ofenterobacterial DNA from both related and unrelated pathogens, such a putative with the salmolysin gene. We are grateful to Jeff Lipps for assistance regulatory gene has not been found. with the animal studies and to Susie Muir for help with the mapping It is surprising that Salmonella strains are not obviously studies. We thank the State of California Department of Health Services Microbial Diseases Laboratory and Christopher Peter and hemolytic on blood agar, given that they encode this potent Larry Samons at the County of San Diego Department of Health hemolysin. One explanation for this comes from preliminary Services Laboratory for supplying clinical isolates of pathogenic studies aimed at understanding its regulation. A transcrip- bacteria and Salmonella. The work was supported by National tional fusion of slyA to lacZ resulted in constitutive expres- Institutes of Health Grant AI22933 (F.H.), Program Project Grant sion of lacZ in E. coli just as we observe hemolytic activity AM-35-108, and Training Grant 5 T32 A107036. in E. coli. However, lacZ was not expressed in Salmonella from the same transcriptional fusion (data not shown). This 1. Mims, S. A. (1982) The Pathogenesis of Infectious Disease (Aca- suggests that slyA is tightly regulated in Salmonella and is demic, New York). probably expressed only in vivo. Using this same fusion, we 2. Welch, R. A. (1991) Mol. Microbiol. 5, 521-528. 3. Braun, V. & Focareta, T. (1991) Crit. Rev. Microbiol. 18, 115-158. have also found that expression of slyA is independent of 4. Welch, R. A., Dellinger, E. P., Minshew, B. & Falkow, S. (1981) phoP/phoQ, a major intracellular regulator, and of katF, the Nature (London) 294, 665-667. starvation regulator (37, 38). That a slyA cognate is found in 5. Bhakdi, S., Mackman, N. & Holland, I. B. (1986) Infect. Immun. Shigella and enteroinvasive E. coli suggests that these bac- 52, 63-69. teria may share a similar virulence mechanism, perhaps 6. Bhakdi, S., Muhly, M., Korom, S. & Schmidt, G. (1990) J. Clin. Invest. 85, 1746-1753. related to survival within macrophages. 7. Cavalieri, S. J. & Snyder, I. S. (1982) Infect. Immun. 37, 966-974. We propose that salmolysin acts on the eukaryotic cell 8. Konig, B., Konig, W., Scheffer, J., Hacke, R. & Goebel, W. (1986) membranes, although to do so, it must be exported or become Infect. Immun. 54, 886-892. associated with the bacterial outer membrane when ex- 9. Moulder, J. W. (1991) Microbiol. Rev. 55, 143-190. pressed. The protein sequence does not reveal a classic signal 10. Turco, J. & Winkler, H. (1991) Infect. Immun. 59, 1647-1655. sequence. However, the protein product of hlyA present in 11. Finlay, B. B. & Falkow, S. (1989) Microbiol. Rev. 53, 210-230. 12. Andrews, N. W., Abrams, C. K., Slatin, S. L. & Giffiths, G. (1990) certain uropathogenic E. coli strains also lacks an amion- Cell 61, 1277-1287. terminal signal sequence but is transported by the products of 13. Gaillard, J. L., Berche, P., Mounier, J., Richard, S. & Sansonettie, hlyB, hlyD, and tolC. Salmolysin has a significant hydropho- P. (1987) Infect. Immun. 55, 1641-1646. bic region at the carboxyl-terminal end (residues 115-133) 14. Kathariou, S., Metz, P., Hof, H. & Goebel, W. (1987) J. Bacteriol. which may form a membrane-spanning a-helix. Whether this 169, 1291-1297. 15. Portnoy, D., Jacks, P. S. & Hinrichs, D. J. (1989) Infect. Immun. sequence could be involved in or in pore formation 57, 477-486. in the target cell membrane has not been addressed experi- 16. Hage-Chahine, C. M., Giudice, G. D., Lambert, P. H. & Pechere, mentally. J. C. (1992) Infect. Immun. 60, 1415-1421. We were unable to complement the slyA mutation in vivo 17. Bielecki, J., Youngman, P., Connelly, P. & Portnoy, D. A. (1990) in the mouse by using the slyA gene cloned in pSC101, Nature (London) 345, 175-176. although we were able to fully complement the mutation in 18. Fields, P. I., Swanson, R. V., Haidaris, C. G. & Heffron, F. (1986) Proc. Natl. Acad. Sci. USA 83, 5189-5193. bacteria in macrophages. There may be two reasons for this. 19. Friedman, A. M., Long, S. R., Brown, S. E., Bukema, W. J. & Perhaps insertional inactivation ofslyA is polar on expression Ausubel, F. M. (1982) Gene 18, 289-296. of other, unidentified downstream genes. This hypothesis 20. Libby, S. J., Goebel, W., Muir, S., Songer, S. & Heffron, F. (1990) predicts that a downstream gene(s) is required for virulence Res. Microbiol. 141, 775-783. but not for survival and growth in macrophages (39). It is also 21. Hanahan, D. (1985) DNA Cloning: A Practical Approach, ed. possible that slyA is not correctly regulated in our construc- Glover, D. M. (IRL, Oxford), Vol. 1, pp. 109-135. 22. Chang, A. C. & Cohen, S. N. (1978) J. Bacteriol. 134, 1141-1156. tion and hence does not express salmolysin correctly in vivo. 23. Benson, N. R. & Goldman, B. S. (1992) J. Bacteriol. 175, 1673- Salmonella slyA- mutants are blocked at more than one 1681. step during in vivo infection; lower numbers ofbacteria reach 24. Roberts, R. C., Burioni, R. & Helinski, D. R. (1990) J. Bacteriol. the Peyer's patches and mesenteric lymph nodes, and they 172, 6204-6216. grow less in the spleen and liver. The spleen and liver are both 25. Simon, R., Priefer, U. & Puhler, A. (1983) BiolTechnology 1, with but the 784-791. initially infected about equal numbers, slyA- 26. Buchmeier, N. A. & Heffron, F. (1989) Infect. Immun. 57, 1. mutant is cleared more rapidly from the liver than from the 27. Rosenberg, A. H., Lade, B. N., Chui, D. S., Lin, S. W., Dunn, spleen. In course-of-infection experiments with many other J. J. & Studier, F. W. (1987) Gene 56, 125-136. attenuated mutants, we have never observed this phenotype. 28. Studier, F. W., Rosenberg, A. H., Dunn, J. J. & Duberndorff, The slyA- mutant may be unusually susceptible to a specific J. W. (1990) Methods Enzymol. 185, 6-89. factor(s) found within the liver. The difference may be due to 29. Neu, H. C. & Heppel, L. A. (1965) J. Biol. Chem. 240, 3685-3692. 30. Songer, J. G., Libby, S. J., landolo, J. J. & Cuevas, W. (1990) survival within a specific macrophage population, in this case Infect. Immun. 58, 131-136. the Kupffer cell. The slyA- mutant can survive in certain 31. Youderian, P., Sugiono, P., Brewer, K. L., Higgins, P. & Elliot, T. macrophages (J774.1 cell line) but not in others (data not (1988) 118, 581-592. shown). Our hypothesis is that this difference in survival in 32. Rennie, R. P., Freer, J. J. & Arbuthnott, J. P. (1974) J. Med. diverse macrophage populations lies at the level of slyA Microbiol. 7, 189-195. expression. If this were true, it would suggest that Salmo- 33. Buchmeier, N. & Heffron, F. (1991) Infect. Immun. 59, 2232-2238. 34. Buchmeier, N. & Heffron, F. (1990) Science 248, 730-732. nella can distinguish certain differentiated macrophages and 35. Groisman, E. A., Chiao, E., Lipps, C. J. & Heffron, F. (1989) Proc. that expression of the slyA gene is critical for survival within Natl. Acad. Sci. USA 86, 7077-7081. these macrophages. That slyA is found in Salmonella, Shi- 36. Fields, P. I., Groisman, E. A. & Heffron, F. (1989) Science 243, gella, and enteroinvasive E. coli suggests that these bacteria 1059-1062. may share a similar virulence mechanism, perhaps related to 37. Fang, F. C., Libby, S. J., Buchmeier, N., Loewen, P., Switala, J. survival within a & Guiney, D. (1992) Proc. Natl. Acad. Sci. USA 89, 11978-11982. macrophages. Perhaps salmolysin plays 38. Mulvey, M. R., Switala, J., Andrew, B. & Loewen, P. C. (1990) J. crucial part in defending the bacteria against the early on- Bacteriol. 172, 6713-6720. slaught ofphagocytic cells found within the lamina propria of 39. Camilli, A., Tilney, L. G. & Portnoy, D. A. (1993) Mol. Microbiol. the intestinal mucosa. 8, 143-157. Downloaded by guest on September 29, 2021