Mol Gen Genet (1995) 249:474 486

© Springer-Verlag 1995

A l b r e c h t L u d w i g • C l a u d i a T e n g e l • S u s a n n e Bauer A n d r e a s Bubert • R o l a n d B e n z • H a n s - J o a c h i m Mollenkopf . Werner Goebel

SlyA, a regulatory protein from Salmonella typhimurium, induces a haemolytic and pore-forming protein in Escherichia coil Received: 21 July 1995 / Accepted: 27 July 1995

A chromosomal fragment from Salmonella typhimurium, when cloned in Escherichia coli, generates a haemolytic phenotype. This fragment carries two genes, termed slyA and slyB. The expression of slyA is sufficient for the haemolytic phenotype. The haemolytic activity of E. coli carrying multiple copies of slyA is found mainly in the cytoplasm, with some in the periplasm of cells grown to stationary phase, but overexpression of SlyB, a 15 kDa lipoprotein probably located in the outer membrane, may lead to enhanced, albeit unspecific, release of the haemolytic activity into the medium. Polyclonal antibodies raised against a purified SlyA-HlyA fusion protein identified the overexpressed monomeric 17 kDa SlyA protein mainly in the cytoplasm of E. coil grown to stationary phase, although smaller amounts were also found in the periplasm and even in the culture supernatant. However, the anti-SlyA antibodies reacted with the SlyA protein in a periplasmic fraction that did not contain the haemolytic activity. Conversely, the periplasmic fraction exhibiting haemolytic activity did not contain the 17 kDa SlyA protein. Furthermore, S. typhimurium transformed with multiple copies of the slyA gene did not show a haemolytic phenotype when grown in rich culture media, although the SlyA protein was expressed in amounts similar to those in the recombinant E. coli strain. These results indicate that SlyA is not itself a cytolysin but rather induces in E. coli (but not in S. typhimurium) the synthesis of an uncharacterised, haemolytically active protein which forms pores with Abstract

Communicated by E. K. F. Bautz A. Ludwig ( I ~ ) • C. Tengel . S. Bauer . A. Bubert H.-J. Mollenkopf • W. Goebel Biozentrum der Universit~it Wiirzburg, Theodor-Boveri-Institut, Mikrobiologie, Am Hubland, 97074 Wfirzburg, Germany R. Benz Biozentrum der UniversitS.t Wiirzburg, Theodor-Boveri-Institut, Biotechnologie, Am Hubland, 97074 Wiirzburg, Germany

a diameter of about 2.6 nm in an artificial lipid bilayer. The SlyA protein thus seems to represent a regulation factor in Salmonella, as is also suggested by the similarity of the SlyA protein to some other bacterial regulatory proteins, slyA- and slyB-related genes were also obtained by PCR from E. coli, Shigella sp. and Citrobacter diversus but not from several other gramnegative bacteria tested. Key words Salmonella typhimurium • SlyA • Virulence factor • Regulation of gene expression • Escheriehia coli" Haemolysin

Introduction

Salmonellae are the causative agents of a variety of diseases in a large number of different hosts. The clinical manifestation of these diseases may vary from local infections of the intestinal tract to systemic forms like typhoid fever, depending on the species and serotype of the infecting bacteria as well as on the host species. Salmonella typhimurium, for example; is the prevalent etiologic agent of localized gastroenteritis in humans, but it has also the capacity to cause a typhoid-like disease in immunocompromised humans and in mice. In the past decade, substantial progress has been made in unravelling the molecular basis of the virulence of this facultative intracellular organism (for recent reviews see Groisman et al. 1990; Guiney et al. 1994; Finlay 1994). Several genes have been identified which affect either the uptake of S. typhimurium by nonprofessional phagocytic mammalian cells (Galan and Curtiss 1989; Ginocchio et al. 1992; Lee et al. 1992; Altmeyer et al. 1993) or its survival within macrophages (Fields et al. 1989; Miller et al. 1989, 1990). In addition, it was shown that at least some of the virulence genes are part of regulons that are controlled by global regulatory proteins, such as the two-component system PhoP/PhoQ (Fields et al. 1989; Groisman et al. 1989;

475

Miller et al. 1989) or KatF (Fang et al. 1992; Norel et al. 1992). Little is known, however, of the functional properties of the virulence genes identified in S. typhimurium. Production of exotoxins by Salmonella is not obvious when these bacteria are grown extracellularly in rich culture media. Production of enterotoxin encoded by the chromosomal stn gene has been reported (Chopra et al. 1991; Prasad et al. 1992), but the molecular properties of this "choleratoxin-like" protein of S. typhimurium are still obscure. Salmonella sp. are normally non-haemolytic when grown on blood agar plates. Hence the isolation of a haemolytic clone obtained by cloning chromosomal DNA fragments of S. typhimurium into E. coli K-12 was rather unexpected (Libby et al. 1994). The ability to give rise to the haemolytic phenotype was restricted to a fragment of 1.3 kb, and its nucleotide sequence revealed two genes which are now called slyA and slyB (Libby et al. 1994 and Fig 1A). It was further shown by Libby et al. (1994) that slyA is apparently sufficient for the generation of the haemolytic phenotype in E. coll. The product encoded by the slyA gene was therefore called salmolysin (SlyA). Disruption of the chromosomal slyA gene in S. typhimurium by insertional mutation was shown to cause a significant reduction in the virulence of this strain when tested in a mouse model (Libby et al. 1994). However, the purification of the SlyA protein proved to be difficult and the haemolytic (cytolytic) activity of SlyA therefore could not be rigorously proven in this earlier work. We have carried this study further and show here that SlyA, although it strongly affects the virulence of S. typhimurium, is not itself a haemolysin. When present in high copy number, SlyA seems to induce the synthesis and partial release of a haemolytic and pore-forming protein in E. coli cultured in rich media. We further show that slyA- and slyB-related genes are also present in E. coli, Shioella sp. and Citrobacter diversus. Interestingly, multiple copies of the slyA-related gene from E. coli also cause the expression of the cryptic haemolysin in E. coli, but induction of haemolysin synthesis by the SlyA-related protein from E. coli is apparently not as efficient as that caused by SlyA from S. typhimurium.

Materials and methods Bacterial strains, plasmids and culture conditions The chromosomal DNA fragment carrying slyA and slyB was originally cloned from wild-type American Type Culture Collection strain Salmonella typhimurium 14028s (Libby et al. 1994). Recombinant multicopy plasmids carrying slyA, sIyB or slyA-slyB were propagated in E. coli CCl18 [araD139, A (ara, leu) 7697, A lacX74, phoAA20, galE, 9aIK, thi, rpsE, rpoB, argEam, recAll and S. typhimurium 14028s. The existence of sIyA- and slyB-homologous genes was tested by Southern hybridization and PCR in several Enterobacteriaceae, including pathogenic and non-pathogenic E. coli strains (EIEC W7062, EHEC W7063, EPEC W6146, ETEC

C9221a, E. coli HB101, E. coli CCl18), Shigella sp. and Citrobacter sp., as welt as in other gram-negative bacteria (species of Vibrio, Yersinia, Pseudomonas and Rhizobium). Plasmids pMOhly6 and pCT1 were propagated in E. coli 5K (Smr, lacY1, tonA21, Z-, thr-1, supE44, thi, r-, m+). The bacteria used in this study were usually cultivated in double-concentrated YT medium, containing 16 g tryptone (Difco, Detroit, Mich.), 10 g yeast extract (Difco) and 10 g NaC1 per litre. Blood agar plates were prepared from YT medium supplemented with 4% defibrinated sheep erythrocytes (Oxoid).

Enzymes and chemicals Restriction endonucleases and DNA modifying enzymes were obtained from Boehringer (Mannheim, Germany) or Pharmacia (Uppsala, Sweden) and used as recommended. Radioisotopes were purchased from Amersham-Buchler (Braunschweig, Germany).

Recombinant DNA methods Chromosomal DNA, plasmid DNA and RNA was isolated according to Sambrook et al. (1989). In addition, DNA cloning procedures, Southern hybridizations and primer extensions were performed largely as described by these authors. Site-directed mutagenesis was performed according to the method of Kramer et al. (1984). The nucleotide sequence of DNA fragments was analysed using the dideoxynucleotide chain-terminationmethod of Sanger et al. (1977). The amplification of DNA sequences by polymerase chain reaction (PCR) was performed following protocols described by Innis et al. (1990).

Isolation of periplasmic and cytoplasmic protein fractions of E. coli and S. typhimurium strains The periplasmic protein fraction of exponential or stationary phase cells was obtained by osmotic shock as described by Neu and Heppel (1965). Briefly, cells from a 10 ml culture were harvested by centrifugation, washed three times with 10mM TRIS-HC1 pH 8.0 and resuspended in 2.5 ml 20% sucrose/30 mM TRIS-HC1 pH 8.0/1 mM EDTA. Following incubation for 10 min at room temperature, the cells were pelleted and resuspended in 2.5 ml ice-cold water, which causes rupture of the outer membrane. After incubation for 10min on ice, the cells were removed by centrifugation. The supernatant contained the periplasmic proteins. The cytoplasmic protein fraction of E. coli and S. typhimurium strains was prepared by resuspending the osmotically shocked cells in 1 ml 20% sucrose/30 mM TRIS-HC1 pH 8.0 and adding 1 mg/ml lysozyme. The mixture was incubated on ice for 10 min. Subsequently, the cells were lysed by repeated freezing ( - 80° C) and thawing. The cell lysate was mixed with 10 mM MgC12 and 300 U DNase I and incubated for 30 min at room temperature. Finally, the cell debris was removed by centrifugation and the supernatant taken as the cytoplasmic protein fraction.

Determination of haemolytic activity (liquid assay) The haemolytic activity in the culture supernatants, in the periplasmic protein fractions and in the cytoplasmic fractions of E. coli and S. typhimurium strains was determined by mixing different amounts of these preparations with 600 gl of a sheep erythrocyte suspension in 0.9% NaC1 or 0.9% NaC1/20 mM CaC12. After incubation at 37°C for 30 min, the erythrocytes were pelleted by centrifugation and the amount of haemoglobin released from the red blood cells

476 into the supernatant was determined spectrophotometrically at 543 nm.

Separation of native proteins on non-denaturing polyacrylamide gels Native proteins were dissolved in sample buffer free of sodium dodecyl sulfate (SDS) and/?-mercaptoethanol and separated within 18-24h at 4°C (50V) on a non-denaturing polyacrylamide gel containing no SDS. Protein fractions of interest were electroeluted from the gel (50V, 15 h, 4°C) into a buffer containing 25 mM TRIS-HC1 pH 8.0 and 192 mM glycine.

Analysis of proteins by SDS-polyacrylamide gel electrophoresis (SDS-PAGE) Extracellular, periplasmic or cytoplasmic proteins were concentrated by precipitation with ice-cold trichloroacetic acid (TCA, final concentration 10%). The precipitated proteins were collected by centrifugation, washed with acetone, dried under vacuum and dissolved in sample buffer containing 50 mM TRIS-HC1 pH 6.8, 10% glycerol, 5% fi-mercaptoethanol, 2% SDS and 0.05% bromophenol blue. Samples were neutralized by addition of TRIS solution. Denaturation of the proteins was achieved by boiling for 4 min. SDSPAGE of the proteins was performed as described by Laemmli (1970). The protein bands were visualized by staining with Coomassie brilliant blue or by fluorography.

Immunoblot analysis of proteins The proteins were separated by SDS-PAGE and transferred to a nitrocellulose filter according to the method of Towbin et al. (1979). Subsequently, the proteins were probed with an antiserum of interest. Proteins reacting with the antiserum were detected by addition of peroxidase-conjugated immunoglobulins directed against these antibodies, followed by colorimetric development with chloronaphthol/H202.

Lipid bilayer assay The pore-forming characteristics of the haemolysin induced in E. coli by multiple copies of slyA were analysed in a lipid bilayer system which was described previously (Benz et al. 1978, 1989).

Labelling of proteins in vivo with [U-14C] palmitic acid Exponentially growing E. coli strains carrying different plasmids were incubated with 3 gCi/ml [U-14C] palmitic acid (824 mCi/ mmol; 200 gCi/ml) for 2 h at 37°C. Subsequently, the cells were pelleted by centrifugation and washed four times with fresh culture medium to remove non-incorporated radioactive palmitic acid. Finally, the cells were boiled for 5 min in SDS sample buffer and the extracted proteins analysed by SDS-PAGE and fluorography.

Lactate dehydrogenase activity assay The release of the cytoplasmic enzyme lactate dehydrogenase (LDH) from E. coli cells harbouring different plasmids was determined using the CytoTox96 Non-Radioactive Cytotoxicity Assay (Promega). To measure the total LDH activity of these strains, the

cells were lysed by treatment with lysozyme followed by sonication. Determination of the EDTA sensitivity of E. coli strains Cultures of exponentially growing E. eoli strains carrying different plasmids were divided in two aliquots. EDTA was added to a final concentration of 10 mM to one aliquot and then both aliquots were further incubated with shaking at 37°C. At different times (0,1, 2, 4 h), samples were taken from each aliquot and the colonyforming units (cfu) were counted.

Results

The haemolytic phenotype of E. coli strains carrying slyA and slyB of S. typhimurium Plasmid pAL101 was constructed by cloning the chromosomal 1.3 kb ClaI-EcoRV fragment, carrying slyA and slyB from S. typhimurium, into the multicopy vector pBluescriptSK ÷ (pSK). When this recombinant plasmid was transformed into the nonhaemolytic E. coli strain CC118, all transformant colonies obtained showed clear haemolytic zones on blood agar after growth for two days (Fig. 1B). Deletion of a SacII fragment from pAL101 resulted in plasmid pAL102, which carries only the intact slyA gene on a 0.84 kb ClaI-SaclI insert. The colonies of E. coli CCl18/pAL102 produced slightly smaller haemolytic zones than E. coli CCll8/pAL101. Even transformation of E. coli CCl18 with plasmid pAL103, containing only the slyB gene on a 0.83 kb Sau3AI-EcoRV fragment in vector pSK, gave rise to colonies with small haemolytic zones (Fig. 1B). In a liquid haemolysin assay using washed sheep erythrocytes, little haemolytic activity was detected in the culture supernatants of E. coli CCl18 harbouring pAL101 but no significant activity was found in the supernatants of E. coli CCll8/pAL102 and E. coli CC118/pAL103. However, high haemolytic activity was measured in the cytoplasmic fraction of the strain carrying only slyA (pAL102) and significant haemolytic activity (about 10% of the cytoplasmic activity) was also found in the periplasmic fraction (obtained by osmotic shock) of this strain. The haemolytic activity reached its highest level in stationary phase. The cytoplasmic and periplasmic haemolytic activity seemed to be substantially lower (< 10% as compared to E. eoli CCl18/ pAL102) in E. coli CCl18 carrying both slyA and slyB (i.e., pAL101). No specific haemolytic activity was observed in E. coli carrying only slyB on plasmid pAL103. These data confirm the previous results of Libby et al. (1994) who also showed that the haemolytic phenotype was associated with SlyA. They extend these earlier data by showing that SlyB may enhance the release of the haemolytic activity into the culture supernatant, probably by unspecific cell lysis (see below). This extracellular release of the haemolytic activity by SlyB seems, however, to be accompanied by a decrease of the total haemolytic activity, which may again be due to SlyBmediated cell lysis.

477 Fig. 1 A Arrangement of slyA and slyB in the chromosome of Salmonella typhimurium. The amino acid sequences of SlyA and SlyB were deduced from the nucleotide sequences of the genes. The putative N-terminal signal sequence of SlyB is indicated by a bar. B Haemolytic phenotype on sheep blood agar of E. coil CCl18 harbouring pAL101 (sIyA and slyB), pAL102 (slyA) and pAL103 (slyB), respectively

A

c~

styA

100 bp

EcoRV

s~B

._.a

ORF 465 bp

ORF438 bp

SlyA (16.7 kDa) M KL E SP LGSDLAR L KP LE LTQTHWVTL OIQLAKAIGI

ELL

SlyB (15.5 kDa) IWRAL

H N I HQL

I DH R

M I KRV

P P DQS

RAKR

E EV I H KTR

I KL IAKL

GEI

I KLTE LAG

LAV$

GDVYTAS

EQPSLMRTLDQLEDK

GL I S RQTGASDR L IAEM

LVR

LMGL

SLAGCVNNDSLS

EAKQVQNMTYGT

I VNVR

P

VQIOGGDDSNVIGAIGGAMLGGFLG KAEP

I S SEE

NT I GGGTGR I

GMQSAMN

EH N I M E L H S H D

VVQKQGNTR

S LATAAGAMAGGVAGQ

KTQGVE

L E I R KDDGNT

FSAGQRVM

LASN

I M

GSQV

TVSPR

B

pAL101

cm= slyA s~B ~=~i~i~i~i~i~i~i~i~i~!~!~i~i~!~!~i~!~!~i~i~i~i~i~i~i~i~i~i~i~i~i~i~i~

Clal

S/FA

Ec?RV

Sacll

S~uaAI

slyB

EcoRV

pAL103

Interestingly, reintroduction of multiple slyA gene copies (i.e., plasmid pAL102) into S. typhimurium 14028s did not cause haemolytic activity in this organism when the bacteria were grown on blood agar or in rich media. Transcription of slyA and slyB in S. typhimurium and

E. coli The transcription of slyA and of the adjacent slyB gene was assayed by primer extension analysis. As shown in Fig. 2, slyA is transcribed very weakly in the S. typhimurium wild-type strain and transcription apparently starts 40 nucleotides upstream of the putative ATG start codon of the slyA structural gene, although no typical -10 box can be directly attributed to this site. There is a significant increase in slyA transcription starting at this site when strains possess multiple slyA copies, i.e. in S. typhimurium transformed with pAL101 and E. coli CC118/pAL101. In addition, several other primer extension products can be identified when multiple slyA copies are present; these may represent degradation products of the primary transcript, or transcripts starting at other sites. In fact, the DNA

sequence upstream of slyA is very AT-rich and at least four overlapping putative promoters were identified within the 130 bp region proximal of slyA, which could account for the numerous transcriptional start sites (Fig. 2). Interestingly, the ribosome binding site in the sIyA upstream region is located immediately in front of the putative ATG start codon of slyA, suggesting that codon 3 (TTG) of the open reading frame may represent the actual start of the slyA structural gene. The adjacent gene, slyB, is transcribed in the opposite orientation to slyA. Again, the slyB gene is only weakly transcribed in the wild-type S. typhimurium strain, starting 46 nucleotides upstream of the open reading frame of sIyB (Fig. 2). The efficiency of slyB transcription starting at this site is apparently strongly enhanced when multiple slyB copies are present, i.e. in S. typhimurium/pALlO1 and E. coli CCll8/pAL101. Under these conditions, we also recognized a second transcriptional start site located 100 bp upstream of the start codon of slyB. In agreement with this, two promoters with typical -10 and -35 regions were identified in the slyB upstream region in front of the transcriptional start sites, the distance between the -10 box and the start site being 7 bp in both cases.

478 Fig. 2 Analysis of the transcription of slyA and slyB. Primer extension experiments were performed with RNA isolated from wild-type S. typhimurium (lanes 1) and E. coIi CCl18 containing multiple copies of the cloned sIyA and sIyB gene (pAL101) (lanes 2). Numbers next to the primer extension products indicate the distance (in bp) of the 5' end of the mRNA from the ATG start codon. In the slyA- and slyB-proximal sequence, the main transcriptional start sites are indicated in white on a dark background. Additional transcriptional start sites detected only in the presence of multiple gene copies are underlined. The -10 and -35 boxes of putative promoters are indicated by bars. The oligonucleotides used as primers are shown by arrows. 'S.D.' indicates putative ribosome binding sites. The intergenic DNA sequence between slyA and slyB, containing the putative rho-dependent termination signal, is shown below the promoter regions of these two genes

s@A

-3s Iv -1 o Iv -35 ill GTGa~CTATGTCTGAAAAAATGTC~GGTAAGCAAATTTAGC'AATA -3s, ,

-lo,~ ,

-~s~

~

~

-~o,

,

1

2 -90

-~o~ ,

CATTTGTTTTGAGAATAGAAATACTGCACACTATT~TCAGGATAA -49 - 40

~_I~CZT~GCM~__~TT__T_~G~~TCGCCACTAGQT S.D.

M

K

L

E

S

P

L

G

-30 -22

TCTGATCTGGCAGGGTTGGTGCGCATTTGGCGTGCTCTGATTGACCAT~ S

D

L

A

R

L

V

R

I

W

R

A

L

I

D

H

-14 -8 -6

R

CCTCAAGCCTCTGGAATTGACGCAGACACATTGGGTCACGTTGCACAATA L K P L E L T Q T H W V T L H N

-35 II

s@B

i

1

t

TATCTTTTCTTGCTTCCGACTTCGTTTAAGATTGGTTAATTAAGTTCTGA -10 II

2

iii iiiiii

100

-35 I

TTATGATTTTCACATTCTTTCGGCAGAAGATATGACCTTAACTGCATAAA -10 I IlllBfllmllm

TGCGATATAATTTAGCI~TAGGGATGCTGAAATAGCGTTTGGTGAAGAAC S.D.

q

AGGAGATGGGTA/~TTAAACGTGTACTGI~CCGTTTCACTGATGGGGT M I K R V L A V S L M G

46

TATGCCTAGCGGGGTGTGTTAACAATGATAGCCTTTCGGGTGATGTTTAT L S L A G C V N N D S L S G D V Y

8/yA

. . .

s~

cAce ' : i 'rc, H

D

*

There is a putative rho-dependent termination signal located between slyA and slyB (Fig. 2), suggesting that the two genes are divergently transcribed into monocistronic mRNAs. This was confirmed by Northern blot analysis (data not shown). In primer extension experiments performed with E. coli CC118 and S. typhimurium clones harbouring pAL102, pAL103 or other recombinant plasmids carrying either a functional slyA or slyB gene, the efficiency of transcription of slyA was found to be independent of the presence and copy number of slyB and vice versa. In addition, primer extension experiments with RNA isolated from strains harbouring recombinant plasmids with a truncated slyA and/or slyB gene indicated that the transcription of these genes is independent of their own gene products, i.e. SlyA and SlyB apparently do not influence the transcription of their own or the other gene. The mode of slyA and sIyB transcription makes it unlikely that insertions in slyA or slyB will exert polar effects on other adjacent genes. The data presented up to this point therefore allow the conclusion that the insertional inactivation of slyA is the reason for the reduced virulence of the slyA mutant of S. typhimurium observed by Libby et al. (1994).

*

R

P

SlyB is probably a lipoprotein in the outer membrane of

S. typhimurium The amino acid sequence of SlyB (Fig. 1A) contains an N-terminal transport signal sequence of 17 amino acids which contains at its C-terminal end a typical recognition sequence (LAGC) for lipid modification and processing by the signal peptidase II (Wu and Tokunaga 1986). In addition, the SlyB protein, which has a calculated molecular weight of 15.5 kDa and a pI of 9.84 (in the non-processed form), shows significant sequence homology (47% identity, 68% similarity) to PCP, an outer membrane lipoprotein of Haemophilus influenzae (Deich et al. 1988, 1990), suggesting that the mature SlyB protein is also a lipoprotein located in the outer membrane (Fig. 3A). In agreement with this, growth of E. coli CCll 8/pAL103 in the presence of [-U-14C] palmitic acid resulted in the specific radioactive labelling of a cellular protein of about 15 kDa, as detected by fluorography following separation of the cellular proteins by SDSpolyacrylamide gel electrophoresis (see arrowhead in Fig. 3B). The molecular weight of this protein corresponds well to that expected for processed and lipidmodified SlyB. In the isogenic E. coli control strain containing only the vector plasmid without slyB,

479 Fig. 3 A Alignment of SlyB from S. typhimuriurn and PCP from Haemophilus influenzae. B Lipophilic modification of SlyB. Exponentially growing cells of E. coli CCI18/pSK (lane 1) and E. coli CC118/pAL103 (lane 2) were labelled with [U-14C]palmitic acid as described in Materials and methods. The cellular proteins from 0.3 ml culture were analysed by SDS-PAGE and fluorography. The molecular weights of size markers are given in kDa

A

SlyB (S. ~/phimurium)

PCP (H. influenzae)

B 1 MIKRVLAVSLM,GLSLAGCVNNDSLSGDVYTASEAKQVQN

I I :l:.l:

::l:.ll.l.I

1 MKKTNMALALLVAFSVTGCANTDIFSGDVYSASOAKEARS

4 o-

VTYGTIVNVRPVQIQGGDDSNVIGAIGGAVLGGFLGNTIG

I:l.:ll:.lll:

2

97°4~ 66.2-

:lllll.ll:ll:...

:llllll.llll.ll:::::

1

I.Ill

iiii!i!iii!I

31.0 -

ITYGTIVSVRPVKIOADNQG.VVGTLGGGALGGIAGSTIG GGTGRSLATAAGAVAGGVAGQGVQSAMNKTQGVELEIRKD

II I..:l...ll::l::ll

::. I .... I.II

I:ll

21.5 -

I

GGRGQAIAAVVGAIGGAIAGSKIEEKMSQVNGAELVIKKD 14.4 -

II..l:llll

::..I

:.:.ll :.ll.

II.ll

DGQEIVVVQK,ADSSFVAGRRVRIVGGGSSLNVSVL

this protein could not be detected by fluorography (Fig. 3B). It has already been shown that overexpression of the major E. coli lipoprotein causes cell lysis (Nakamura et al. 1982). Cell lysis was also observed when lipoproteins from Bacillus subtilis and Streptococcus pneumoniae were expressed in E. coli (Martin et al. 1989; Gomez et al. 1994). Similarly, we observed that the overexpression of the SlyB protein, alone or in combination with SlyA, in E. coli CCl18 carrying either pAL103 (slyB) or pAL101 (slyA and slyB) leads to a significant reduction in the growth rate of these strains as compared to E. coli CCl18 carrying multiple slyA copies (pAL102) or the vector plasmid pSK alone. Furthermore, we observed that the SlyB-overproducing strains become leaky and release substantial amounts of cytoplasmic and periplasmic proteins into the culture supernatants, as shown by the release of up to 40% of the whole cellular lactate dehydrogenase activity (measured by an LDH activity assay) and almost 90% of the periplasmic/Mactamase (detected by immunoblot analysis) into the supernatant by an E. coli strain carrying multiple copies of the slyB gene (E. coli CC118/pAL103) (Fig. 4). No such release was observed for the E. coli strains carrying the multicopy vector plasmid alone or with the cloned slyA gene. E. coli cells overexpressing SlyB, or the combination of SlyA and SlyB, also exhibited an increased sensitivity to 10 mM EDTA as compared to E. coli strains harbouring only the vector pSK or multiple slyA copies (pAL102). Again, a similar effect was observed when genes encoding lipoproteins of other bacteria were overexpressed in E. coli (Martin et al. 1989; Gomez et al. 1994), further confirming that SlyB represents a lipoprotein that is localized to the outer membrane. The observed leakiness of the SlyB-overproducing E. coli strain probably accounts for the weak haemolytic phenotype of this strain on blood agar, in agreement with the previous observation (Wagner et al. 1988) that stationary grown E. coli K-12 cells exhibit a low cytoplasmic haemolytic activity. This leakiness may also be

66.2

-

45.0

-

31.0

-

~

I -.,,= i

DGNTIMVVQKQGNTRFSAGQRVVLASNGSQVTVSPR 155 6.5-

154

A

B

123

123

Fig. 4 Release of/3-1actamase from E. coli cells overexpresslng SlyB from S. typhirnurium. Extracellular proteins (A) and periplasmic proteins (B) from 0.5ml culture of exponentially grown E. coli CC118/pSK (lane 1), E. coli CCll8/pAL102 (lane 2) and E. coli CC118/pAL103 (lane 3) were separated by SDS-PAGE and transferred to a nitrocellulose filter. ]?-Lactamase (arrowhead) was detected with polyclonal anti-]?-lactamase antibodies. Molecular weights of marker proteins are given in kDa

responsible for the larger haemolytic zones around the colonies of E. coli CCl18 carrying slyA and slyB on plasmid pAL101 as compared to those of the isogenic E. coli strain carrying slyA alone (see Fig. 1). SlyA induces a haemolytic activity in E. coli Comparison of the predicted amino acid sequence of SlyA with the sequences of the known cytolysins from gram-negative and gram-positive bacteria did not show any significant similarities. However, there are substantial sequence similarities between SlyA and several bacterial regulatory proteins, such as MprA (Del Castillo et al. 1991), PecS (Reverchon et al., unpublished), MarR (Cohen et al. 1993; Ariza et al. 1994) and Hpr (Perego and Hoch 1988), as recently pointed out by Dehoux and Cossart (1995). It was therefore of

480 Fig. 5 A Construction of a SlyA-HlyA fnsion protein, which is secreted from E. eoli via the E. coli c~-haemolysin (HlyA) transport apparatus (details are given in the text). B Detection of the extracellular SlyA-HlyA fusion protein (arrowhead) by immunoblot analyses using a polyclonal anti-HlyA antiserum (I) and a polyclonal antiserum raised against the purified fusion protein (II). The proteins of 1 ml culture supernatant from E. coli 5K/pMOhly6 (lanes 1) and E. coil 5K/pCT1 (lanes 2) in the late exponential growth phase and the purified 23 kDa SlyA-HlyA fusion protein (lane 3) were separated by SDS-PAGE, transferred to a nitrocellulose filter and probed with the antisera

A

hlyR

htyC hlyA(3')

hlyB

hlyD pMOhly6

Nsi]

hlyR

NsiI NsiI

hlyC slyA-hlyA

146)

I

S

hlyD

pCT1

C-terminal61 amJr~aei6s of HlyA [

I 1

crucial importance for the understanding of the involvement of SlyA in virulence to study whether SlyA is itself a haemolysin (cytolysin) or a regulatory protein that induces a cytolytic activity in E. coli. In order to purify SlyA and to study its relationship with the haemolytic activity, we converted SlyA into a fusion protein that could be secreted by E. coli into the culture supernatant via the E. coli ~-haemolysin (HlyA) secretion apparatus. This was achieved by linking the genetic information for the C-terminal 61 amino acids of HlyA, containing the HlyA transport signal (HlyA~; Jarchau et al. 1994), in frame to the Y end of the full-length slyA gene, using plasmid pMOhly6. This plasmid is a derivative of plasmid pANN202-312*, which contains the entire c~-haemolysin determinant of E. coli in the vector pACYC184, including the transcriptional activator hlyR and the structural genes necessary for the synthesis and activation of ~-haemolysin (hlyA and hlyC, respectively) and for the secretion of ehaemolysin (hlyB and hlyD) (Hess et al. 1986). In pMOhly6 the e-haemolysin structural gene hlyA contains a large in-frame deletion and consists only of the extreme 5' end and the 61 3' terminal codons. These two sequences are separated by a NsiI site since the truncated hlyA gene was constructed by introduction of a NsiI site in codon 2/3 of hlyA, followed by an in-frame deletion of the NsiI fragment encoding the amino acids 4-963 of HlyA. In order to construct the SlyA-HlyA~ fusion protein, we amplified the sIyA gene by PCR from plasmid pAL102 and introduced a NsiI site at both ends of this gene. Subsequently, the slyA gene was inserted in-frame into the NsiI site of the truncated hlyA gene of pMOhly6, resulting in plasmid pCT1 (Fig. 5A). E. coli 5K carrying this construct was not haemolytic, but, as shown by immunoblot analysis, this strain efficiently secreted a protein with the size

hlyB

II 2

I

97.4 -

97.4

66.2

-

66.2-

45.0

-

45.0

2

3

-

-

31.0-

31.0-

21.5-

21.5-

expected for the SlyA-HlyAs fusion (23 kDa) into the supernatant. This protein reacted with a polyclonal antiserum directed against the C-terminal portion of HlyA (Fig. 5B). E. coli 5K harbouring pMOhly6 did not secrete this protein. The fusion protein was purified by preparative SDS-polyacrylamide gel electrophoresis and used as antigen for the generation of SlyA-specific antibodies in a rabbit. The polyclonal antiserum obtained reacted strongly with the SlyA-HlyAs fusion protein (Fig. 5B) and weakly with E. coli ~-haemolysin (HlyA). The weak reaction of the antiserum with HlyA is in accord with previous observations showing that the HlyA signal sequence (HlyAs) represents a weak antigen (Jarchau et al. 1994). In the cytoplasmic fractions of E. eoli and S. typhimurium strains carrying slyA on the multicopy plasmid pAL102, the anti-SlyA antiserum recognized specifically a protein of 16-17 kDa which corresponds well to the expected molecular mass of monomeric SlyA (Fig. 6). Significantly smaller amounts of this protein were also identified in the periplasmic fractions and

481 1 a b c a b c i

97.4

2 a

3 I bc

A

1

2

3

B

1

2

22 a b

3 3 3 a b 97.466.2 -

I

-

66.2 -

45.0 i

45.0 -

I

I

31.0 21.5-

~

: ....

14.4-

14.4-

Fig. 6 Detection of the SlyA protein in different fractions of E. coli and S. typhimurium strains carrying multiple copies of the slyA gene from S. typhimurium. Extracellular (a), periplasmic (b) and cytoplasmic (c) proteins from 0.5 ml overnight culture of E. coli CC118/pSK (1), E. coli CC118/pAL102 (2) and S. typhimurium 14028s/pAL102 (3) were separated by SDS-PAGE, transferred to a nitrocellulose filter and probed with the polyclonal antiserum raised against the SlyA-HlyA fusion protein. The molecular weights of size markers are given in kDa. The 17 kDa SlyA protein is indicated by an arrowhead

even in the culture supernatants of both strains when grown to the stationary phase. However, haemolytic activity in the periplasmic and cytoplasmic fraction was only identified in E. coli but not in S. typhimurium harbouring pAL102. Furthermore, the haemolytic activity in the periplasmic fraction of E. coli CCll8/pAL102 could not be inhibited with the antiSlyA antiserum. The antiserum reacted in the extracellular and (to a lesser degree) in the periplasmic fractions of both slyA-carrying strains also with one or more proteins of about 60-65 kDa (Fig. 6). But in the corresponding fractions of the isogenic control strains without the cloned slyA gene, the same protein bands reacted with this antiserum at similar intensities. Apparently, the 60-65 kDa proteins crossreact with the anti-SlyA antiserum but do not seem to be related to SlyA. This assumption is further supported by the fact that various conditions (low and high pH, dithiotreithol, SDS and other detergents in combination with elevated temperature), which should disrupt oligomeric forms of SlyA, failed to convert these proteins into the 17 kDa monomeric SlyA. The anti-SlyA antibody even detected in the cytoplasmic fraction of stationary phase cells of the control strain, E. coli CCll8/pSK, a very weak band which corresponded in size to the 17 kDa SlyA protein (see Fig. 6). As shown below, this reflects the fact that all E. coli strains tested harbour a chromosomal slyA-related gene, which is expressed with similar efficiency to slyA of S. typhimurium. The SlyA-related protein from E. coli also reacts with the anti-SlyA antibodies.

Fig. ? A Fractionation of the periplasmic proteins from E. coli CC118/pSK (lane 1), E. coli CCllS/pAL102 (lane 2) and S. typhimurium 14028s/pAL102 (lane 3). The periplasmic proteins were isolated by osmotic shock and separated on a non-denaturing polyacrylamide gel followed by an incubation of the gel for 30 minutes on blood agar. The proteins in bands 2a, 2b, 3a and 3b were eluted from the gel. B Immunoblot analysis of the unfractionated periplasmic proteins from E. coli CCllS/pSK (1), E. coli C C l l S / p A L I 0 2 (2) and S. typhimurium 14028s/pAL102 (3), as well as of the protein fractions 2a, 2b, 3a and 3b. The proteins were separated by SDS-PAGE, transferred to a nitrocellulose filter and probed with the polyclonal anti-SlyA antiserum. The SlyA protein is indicated by an arrowhead. Molecular weight standards are given in kDa

When the strongly haemolytic periplasmic fraction of E. coli CC118/pAL102, carrying multiple slyA gene copies, was fractionated on a non-denaturing polyacrylamide gel, haemolytic activity was detected in a defined band by placing the gel on blood agar (Fig. 7A). The haemolytic band (band 2a) and the band directly below (band 2b), which did not exhibit haemolytic activity, were cut out, the proteins were eluted and separated by SDS-PAGE. Subsequently, the proteins were transferred onto a nitrocellulose filter and probed with the anti-SlyA antiserum. As shown in Fig. 7B, in the haemolytic fraction 2a the antiserum reacted mainly unspecifically and detected only traces of the 17 kDa SlyA protein, whereas the nonhaemolytic fraction 2b contained most of the 17 kDa SlyA protein that specifically reacts with the anti-SlyA antiserum. The periplasmic proteins prepared from S. typhimurium/pAL102 were also fractionated as described for E. coll. In this case, the band 3a, corresponding to the haemolytic band 2a of the E. coli clone, was non-haemolytic, but both bands, 3a and 3b, yielded protein data similar to that mentioned above for E. coli (Fig. 7), i.e., fraction 3a contained only marginal amounts of SlyA, whereas fraction 3b contained again most of the SlyA protein specifically reacting with the anti-SlyA antiserum. Interestingly, prolonged incubation of the non-denaturing polyacrylamide gel on blood agar even led to the detection of a very weak haemolytic band in the periplasmic fraction of the stationary phase control strain E. coli CCll8/pSK.

482 Fig. 8 A, B Pore-forming activity of the haemolysin induced in E. coli by multiple copies of slyA. The periplasmic fraction of E. coli CC118/pAL102 was fractionated on a non-denaturing polyacrylamide gel. Following an incubation of the gel on blood agar, the haemolytically active protein fraction (band 2a in Fig. 7) was eluted from the gel and incubated in the presence of 1 M KC1 with an artificial lipid bilayer formed from erythrocyte lipids. The applied transmembrane voltage was 10mV (A) and 50mV (B), the temperature was 25 ° C

A

B

20 nS

i

2 0 0 pA

w

The position of this band corresponded exactly to that of the strongly haemolytic band 2a of E. eoli CCll8/pAL102. On the other hand, the periplasmic fraction of S. typhimurium/pALl02 remained inactive against erythrocytes under these conditions. In summary, these data show that the haemolytic activity detected in E. coli strains carrying multiple copies of slyA is not directly associated with the SlyA protein. Under in vitro cultivation conditions, SlyA rather appears to induce in E. coli, but not in S. typhimurium, the efficient synthesis of a haemolysin which is otherwise poorly expressed. The haemolytic activity of E. coli carrying multiple slyA copies is caused by a pore-forming protein To further characterize the haemolytic activity which is observed in the cytoplasmic and periplasmic fractions of E. coli CCl18 and other E. coli K-12 strains transformed with multiple copies of slyA, we used an artificial lipid bilayer system. This system was recently applied successfully to characterize the pores formed by E. coli c~-haemolysin (HlyA) (Benz et al. 1989). With the total periplasmic fraction of E. coli CCll8/pAL102 and the haemolytic fraction 2a (devoid of the 17 kDa SlyA protein, see Fig. 7), obtained by separating the periplasmic proteins on a non-denaturing polyacrylamide gel, efficient pore formation was observed (Fig. 8). Both protein preparations yielded the same defined pores. In the presence of 1M KC1 and at a membrane voltage of 10 to 20 mV these pores were stable and had a single channel conductance of about 10 nS, which corresponds to a pore diameter of approximately 2.6 nm. At a membrane voltage of 50 mV the pores had a shorter lifetime. In contrast, the periplasmic fraction 2b, which

,

30 s

,

1 nA

30 s

contains most of the 17 kDa SlyA protein, did not form pores in this system (data not shown). Likewise, no pore formation was obtained with identically prepared periplasmic fractions from an isogene E. coli strain lacking the cloned slyA gene. These data suggest that the pore-forming activity and the haemolytic activity are probably caused by the same, as yet unknown, protein present in the cytoplasm and in the periplasmic fraction 2a of E. coli carrying multiple copies of slyA. The data also indicate that neither the haemolytic activity nor the pore-forming activity is directly associated with the 17 kDa SlyA protein. However, the overexpression of SlyA is necessary for the efficient synthesis of the haemolytic (and pore-forming) protein in E. coli since these two activities were not found in significant amounts in E. coli strains that do not carry multiple slyA copies. slyA- and slyB-related genes in other Enterobacteriaceae Using a slyA-specific gene probe, Libby et al. (1994) observed hybridization with the chromosomal DNA of all Salmonella strains tested, which suggested that this gene is present in all Salmonella serotypes and species. In addition, these authors obtained positive hybridization data with chromosomal DNA of Shigella and enteroinvasive E. coli strains. We extended these hybridization studies and identified slyA- and slyB-homologous sequences in all pathogenic and nonpathogenic E. coli strains tested (Fig. 9). In addition, using slyAspecific primers we could amplify by PCR, and subsequently sequence, slyA-related genes from several E. coli strains (including the enteroinvasive E. coli (EIEC) strain W7062, E. coli HB101 and E. coli CCl18), as well as from Shigella flexneri and Citrobacter diversus. No PCR products were obtained from other

483

A

B

1 2 3 4 5 6

1 2 3 4 5 6

Fig. 9 A, B Identification of slyA- and slyB-homologoussequences in pathogenic and non-pathogenic E. coIi strains by Southern hybridization. HindIII-digestedchromosomal DNA isolated from S. typhimurium14028s(1), EIEC W7062 (2), EHEC W7063 (3), EPEC W6146 (4), ETEC C9221a (5) and E. coliHB101(6) was separated by agarose gel electrophoresis, transferred to a nitrocellulose filter and hybridized with a slyA-specificgene probe (291 bp SspI fragment of slyA) (A) and a slyB-specificgene probe (411 bp EaeI fragment of slyB) (B). DNA fragments carrying slyA- or slyB-relatedsequences were identified by autoradiography

Enterobacteriaceae (including C. freundii) and several other gram-negative

bacteria (different species of

Vibrio, Yersinia, Pseudomonas and Rhizobium). A comparison of these slyA-related genes indicates that the D N A sequences in all E. coli and Shigella strains tested are identical and share 81% homology with the slyA gene of S. typhimurium. The amino acid sequence of the SlyA-related protein of E. coli (Shigella) exhibits 90% identity and 95% similarity to that of SlyA of S. typhimurium. The differences are found in several amino acid positions, particularly in the C-terminal part of the protein. Even more amino acid exchanges are found in the SlyA-related protein of C. diversus relative to SlyA of S. typhimurium (Fig. 10). In all bacterial strains that carry a slyA-related gene we also identified a highly conserved slyB-homologue, which is located in opposite orientation immediately downstream of the slyA-related gene, suggesting that the chromosomal context of the slyA and slyB genes is conserved in all these bacteria. Again, the sequence of the slyB-related gene was found to be identical in S. typhim.

several E. coli strains studied (HB101, EIEC W7062) and Shigella flexneri, sharing 80% homology with slyB of S. typhimurium. The SlyB proteins from S. typhimurium and E. coli (Shigella) are even more highly conserved than the SlyA proteins (95% identity, 97 % similarity). Specifically, there are only eight amino acid exchanges in the SlyB protein of E. coli as compared to SlyB of S. typhimurium (A7V, L10M, M l l V , A16V, $23T, T41S, I78V, A137P) and the first four of these substitutions are located within the putative N-terminal signal sequence of the protein. While the slyA and slyB genes are highly homologous in S. typhimurium, E. coli/Shigella and C. diversus, the sequences upstream of these genes and in the intergenic region between slyA and slyB are less conserved but seem to be identical within strains of E. coli/Shigella. The chromosomal D N A fragment carrying the slyAand slyB- related genes from E. coli was cloned into the vector pSK, resulting in the recombinant plasmid pAL104. Primer extension experiments performed with R N A isolated from E. coli CCl18 and E. coli C C l 1 8 / pAL104 showed that the sly genes from E. coli and the slyA and slyB sequences from S. typhimurium are transcribed with similar efficiency in E. coli. The results of these primer extensions further indicated that the transcriptional start sites in the sequences upstream of slyA and slyB from E. coli largely correspond to those found in the promoter regions of slyA and slyB from S. typhimurium (data not shown). In agreement with this, putative promoters in the slyA- and slyB- proximal regions of E. coli are located at corresponding positions to those in the sequences preceding slyA and slyB of S. typhimurium (the putative promoter IV in the slyA upstream region of S. typhimurium is not conserved in E. colO. We also identified a putative rho-dependent termination signal in the region between slyA and slyB of E. coll. When E. coli CCl18 was transformed with plasmid pAL105, a recombinant derivative of pSK carrying only the slyA-related gene from E. coIi, the overexpressed SlyA-homologous protein could be detected in Fig. 10 Multiple alignment of SlyA from S. typhimuriumand SlyAhomologous proteins from E. coli, Shigellaflexneri and Citrobacter diversus.Only deviations from the sequenceof SIyAare shown. The C-terminal amino acid sequence of the SlyA-homologousprotein from C. diversushas not been determined upto now 75

MKLESPLGSDLARLVRIWRALIDHRLKPLELTQTHWVTLHNIHQLPPDQSQIQLAKAIGIEQPSLVRTLDQLEDK

E. coil

S. flexneri C. diversus

s. typhim. E. coli S. flexneri C. diversus

. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

V

......

146

GLISRQTCASDRRAKRIKLTEKAEPLIAEMEEVIHKTRGEILAGISSEEIELLIKLIAKLEHNIMELHSHD* ...........................

S...A..N...A...H...A..L.Q..T

.........

I..QAKG*

...........................

S...A..N...A...H...A..L.Q..T

.........

I..QAKG*

.........

T...A.TN.P

T .....

A ...........

.....

E..PA

?

484

97.4 66.2 45.0 31.0

21.5 14.4 Fig. 11 Overexpression of the SlyA-related protein from E. coli in E. coli CC118/pAL105. Extracellular (a), periplasmic (b) and cytoplasmic (c) proteins of 0.5 ml culture (stationary growth phase) from E. coli CCll8/pSK (1) and E. coli CCll8/pAL105 (2) were separated by SDS-PAGE. The 17 kDa SlyA protein (see arrowhead) was detected by immunoblot analysis of the proteins using the polyclonal anti-SlyA antiserum. Molecular weights of marker proteins are given in kDa

the cytoplasmic, periplasmic and extracellular fractions of the strain by the anti-SlyA antiserum in amounts similar to those found for SlyA of S. typhimurium (Fig. 11). In addition, when present in high copy numbers, the slyA-related gene from E. coli also caused a haemolytic phenotype in E. coli. However, the haemolytic activity induced in E. coli by the SlyA-related protein seems to be only 20-40% of that evoked by SlyA from S. typhimurium. This may be due to slightly different activities of the SlyA proteins from S. typhimurium and E. coli resulting from differences in the amino acid sequences of these homologous proteins.

Discussion

The data of Libby et al. (1994) suggested that the slyA gene of S. typhimurium might encode a novel bacterial haemolysin (cytolysin). However, the experiments did not rigorously exclude the possibility that SlyA might represent a regulatory factor that induces the synthesis of an unknown cytolysin in E. coli when present in multiple copies. The recent finding (Dehoux and Cossart 1995) that SlyA shares homologies with several known bacterial regulatory proteins indeed strongly favoured the latter explanation. Our data presented here clearly show that the haemolytic and pore-forming activity induced by the introduction of multiple slyA copies into E. coli is not encoded by slyA.

The most extensive sequence homology (28% identity, 51% similarity) was found between SlyA and MprA, a negative transcription factor of E. coli (Del Castillo et al. 1991) which regulates the synthesis of microcins B17 and C7 in the stationary growth phase. When present in E. coIi in high copy numbers, the mprA gene blocks the osmoinduction of proU and the growth phase-dependent induction of the mcb and mcc operons, which are required for the production of the microcins B17 and C7, respectively (Del Castillo et al. 1991). Interestingly, multiple copies of sIyA introduced into a slyA mutant of S. typhimurium failed to restore the virulence of this mutant in mice, although this complementation restored the ability to survive in cultured macrophages as shown by Libby et al. (1994), suggesting that a null mutation in slyA but also high gene dosage of slyA reduces the in vivo virulence ofS. typhimurium. Our data suggest that enhanced synthesis of SlyA and of the haemolytic (pore-forming) component in E. coli occurs in the stationary phase. Furthermore, preliminary data (not shown here) indicate that in the stationary growth phase the expression of many proteins is affected in the slyA mutant of S. typhimurium, but also in the isogenic S. typhimurium strain carrying multiple slyA copies, when compared to the parental S. typhimurium strain. This suggests that SlyA, like MprA, is indeed a regulatory element that modulates the expression of multiple genes in the stationary phase. The SlyA-related protein of E. coli, identified in this study, is obviously not identical to MprA ofE. coli. It is therefore likely that there exists in E. coli (and probably also in S. typhimurium) a family of functionally related regulatory proteins that may modulate the transcription of specific promoters in the stationary growth phase. In the case of MprA it has been suggested that this protein may interact, like some histone-like E. coli proteins, directly with DNA or with proteins associated with the nucleoid, thereby locally altering DNA topology (e.g., supercoiling and bending) and consequently affecting transcription (Del Castillo et al. 1994). Our data show that most of SlyA remains in the cytoplasm, which is in accord with the assumption that SlyA is a regulatory protein. The observed partial localization of SlyA to the periplasm is apparently due to the fact that the cytoplasmic membrane of E. coli cells carrying multiple copies of slyA or of its own slyArelated gene becomes leaky in the stationary growth phase. This possibility may also explain the partial release of the haemolytic (pore-forming) component in the slyA-carrying E. coli strain. The occurrence of a cytoplasmic haemolytic activity in stationary phase E. coli K-12 cells has been previously reported (Wagner et al. 1988) and our data suggest that SlyA (and the SlyA-related protein of E. coli) may enhance the synthesis of this haemolytic protein. The gene slyB, located immediately downstream of slyA, encodes a lipoprotein with substantial homology to the lipoprotein PCP of H. influenzae. When present

485

in E. coli in multiple copies together with slyA, it increases the haemolytic zones around the colonies growing on blood agar. However, the involvement of the gene product, SlyB, in haemolysis seems to be indirect, as it was shown that the overexpression of SlyB causes an unspecific leakiness of the inner and outer membrane of E. coli and may thus enhance release of the haemolytic activity from the periplasm into the extracellular environment. Our data thus show that the slyA gene of S. typhimurium does not encode a haemolysin (cytolysin) as previously assumed (Libby et al. 1994). Multiple copies of this gene rather induce in E. coli (but not in S. ryphimurium grown in rich culture media) the synthesis of a yet uncharacterized haemolytic and pore-forming protein. Based on the sequence homology with other regulatory proteins, especially MprA (Del Castillo et al. 1991), and our own preliminary experimental data, we suggest that SlyA modulates the transcription of specific genes, some of which are involved in determining the virulence of S. typhimurium. Studies have been initiated to clarify the precise function of SlyA as a regulatory factor in Salmonella and E. coli and to isolate the gene for the novel haemolytic and pore-forming protein which seems to be present in all E. coli strains in a silent state but which can be induced by high copy numbers of slyA.

AcknowledgementsThis work was supported by a grant from the Deutsche Forschungsgemeinschaft (SFB 176) and by the Fonds der Chemischen Industrie.

References Altmeyer RM, McNern JK, Bossio JC, Rosenshine I, Finlay BB, Gatan JE (1993) Cloning and molecular characterization of a gene involved in Salmonella adherence and invasion of cultured epithelial cells. Mol Microbiol 7 : 89-98 Ariza RR, Cohen SP, Bachhawat N, Levy SB, Demple B (1994) Repressor mutations in the marRAB operon that activate oxidative stress genes and multiple antibiotic resistance in Escherichia coll. J Bacteriol 176:143-148 Benz R, Janko K, Boos W, L~iuger P (1978) Formation of large, ion-permeable membrane channels by the matrix protein (porin) of Escherichia coll. Biochim Biophys Acta 511:305-319 Benz R, Schmid A, Wagner W, Goebel W (1989) Pore formation by the Escherichia coli hemolysin: evidence for an associationdissociation equilibrium of the pore-forming aggregates. Infect Immun 57 : 887-895 Chopra AK, Peterson JW, Houston CW, Pericas R, Prasad R (1991) Enterotoxin-associated DNA sequence homology between Salmonella species and Escherichia coll. FEMS Microbiol Lett 61 : 133-138 Cohen SP, HS,chler H, Levy SB (1993) Genetic and functional analysis of the multiple antibiotic resistance (mar) locus in Escherichia coli. J Bacteriol 175 : 1484-1492 Dehoux P, Cossart P (1995) Homologies between salmolysin and some bacterial regulatory proteins. Mol Microbiol 15 : 591-592 Deich RA, Metcalf BJ, Finn CW, Farley JE, Green BA (1988) Cloning of genes encoding a 15,000-dalton peptidoglycan-associatedouter membrane lipoprotein and an antigenicaUy related 15,000-dalton protein from Haemophilus influenzae. J Bacteriol 170:489-498

Deich RA, Anilionis A, Fulginiti J, MetcaIf BJ, Quataert S, QuinnDey T, Zlotnick GW, Green BA (1990) Antigenic conservation of the 15,000-dalton outer membrane lipoprotein PCP of Haemophilus influenzae and biologic activity of anti-PCP antisera. Infect Immun 58 : 3388-3393 Del Castillo I, Gonzalez-Pastor JE, San Millan JL, M oreno F (1991) Nucleotide sequence of the Escherichia coli regulatory gene mprA and construction and characterization of mprA-deficient mutants. J Bacteriol 173:3924-3929 Fang FC, Libby SJ, Buchmeier NA, Loewen PC, Switala J, Hatwood J, Guiney DG (1992) The alternative sigma factor KatF (RpoS) regulates Salmonella virulence. Proc Natl Acad Sci USA 89:11978-11982 Fields PI, Groisman EA, Heffron F (1989) A Salmonella locus that controls resistance to microbicidal proteins from phagocytic cells. Science 243 : 1059-1062 Finlay BB (1994) Molecular and cellular mechanisms of Salmonella pathogenesis. Curr Top Microbiol Immunol 192:163 185 Galan JE, Cur~tiss, R I I I (1989) Cloning and molecular characterization of genes whose products allow Salmonella typhimurium to penetrate tissue culture cells. Proc Natl Acad Sci USA 86: 6383-6387 Ginocchio C, Pace J, Galan JE (1992) Identification and molecular characterization of a Salmonella typhimurium gene involved in triggering the internalization of salmonellae into cultured epithelial cells. Proc Natl Acad Sci USA 89: 59765980 Gomez A, Ramon D, Sanz P (1994) The Bacillus subtilis lipoprotein LplA causes cell lysis when expressed in Escherichia coli. Microbiology 140:1839-1845 Groisman EA, Chiao E, Lipps CJ, Heffron F (1989) Salmonella typhimurium phoP virulence gene is a transcriptional regulator. Proc Natl Acad Sci USA 86:7077-7081 Groisman EA, Fields PI, Heffron F (1990) Molecular biology of Salmonella pathogenesis. In: Iglewski BH, Clark VL (eds) The bacteria, Vol. XI. Molecular basis of bacterial pathogenesis. Academic Press, San Diego-London, pp 251-272 Guiney DG, Fang FC, Krause M, Libby S (1994) Plasmid-mediated virulence genes in non-typhoid Salmonella serovars. FEMS Microbiol Lett 124:1-10 Hess J, Wels W, Vogel M, Goebel W (1986) Nucleotide sequence of a plasmid-encoded hemolysin determinant and its comparison with a corresponding chromosomal hemolysin sequence, FEMS Microbiol Lett 34: 1-11 Innis MA, Gelfand DH, Sninsky JJ (1990) PCR protocols. A guide to methods and applications. Academic Press, San Diego-London Jarchau T, Chakraborty T, Garcia F, Goebel W (1994) Selection for transport competence of C-terminal polypeptides derived from Escherichia coli hemolysin: the shortest peptide capable of autonomous HlyB/HlyD-dependent secretion comprises the C-terminal 62 amino acids of HlyA. Mol Gen Genet 245 : 53-60 Kramer W, Drutsa V, Jansen H-W, Kramer B, Pflugfelder M, Fritz H-J (1984) The gapped duplex DNA approach to oligonucleotide-directed mutation construction. Nucleic Acids Res 12 : 9441-9456 Laemmli UK (1970) Cleavage of structural proteins during the assembly of the head of bacteriophage T4. Nature 227 : 680-685 Lee CA, Jones BD, Falkow S (1992) Identification of a Salmonella typhimurium invasion locus by selection for hyperinvasive mutants. Proc Natl Acad Sci USA 89:1847-1851 Libby SJ, Goebel W, Ludwig A, Buchmeier N, Bowe F, Fang FC, Guiney DG, Songer JG, Heffron F (1994) A cytolysin encoded by Salmonella is required for survival within macrophages. Proc Natl Acad Sci USA 91:48%493 Martin B, Alloing G, Boucraut C, Claverys J-P (1989) The difficulty of cloning Streptococcus pneumoniae real and ami loci in Escherichia coli: toxicity of malX and amiA gene products. Gene 80:227-238 Miller SI, Kukral AM, Mekalanos JJ (1989) A two-component regulatory system (phoP phoQ) controls Salmonella typhimurium virulence. Proc Natl Acad Sci USA 86:5054-5058

486 Miller SI, Pulkkinen WS, Selsted ME, Mekalanos JJ (1990) Characterization of defensin resistance phenotypes associated with mutations in the phoP virulence regulon of Salmonella typhimurium. Infect Immun 58 : 3706-3710 Nakamura K, Masui Y, Inouye M (1982) Use of a lac promoteroperator fragment as a transcriptional control switch for expression of the constitutive Ipp gene in Escherichia coll. J Mol Appl Gen 1 : 289-299 Neu HC, Heppel LA (1965) The release of enzymes from Escherichia coli by osmotic shock and during the formation of spheroplasts. J Biol Chem 240 : 3685-3692 Norel F, Robbe-Saule V, Popoff MY, Coynault C (1992) The putative sigma factor KatF (RpoS) is required for the transcription of the Salmonella typhimurium virulence gene spvB in Escheriehia coli. FEMS Microbiol Lett 99:271-276 Perego M, Hoch JA (1988) Sequence analysis and regulation of the hpr locus, a regulatory gene for protease production and sporulation in Bacillus subtilis. J Bacteriol 170 : 2560-2567

Prasad R, Chopra AK, Chary P, Peterson JW, (1992) Expression and characterization of the cloned Salmonella typhimurium enterotoxin. Microb Pathog 13 : 109-121 Sambrook J, Fritsch EF, Maniatis T (1989) Molecular cloning. A laboratory manual. Cold Spring Harbor Laboratory Press, Cold Spring Harbor, New York Sanger F, Nicklen S, Coulson AR (1977) DNA sequencing with chain-terminating inhibitors. Proc Natl Acad Sci USA 74: 5463-5467 Towbin H, Staehelin T, Gordon J (1979) Electrophoretic transfer of proteins from polyacrylamide gels to nitrocellulose sheets: procedure and some applications. Proc Natl Acad Sci USA 76:4350-4354 Wagner W, Kuhn M, Goebel W (1988) Active and inactive forms of hemolysin (HlyA) from Escherichia coll. Biol Chem HoppeSeyler 369: 39-46 Wu HC, Tokunaga M (1986) Biogenesis of lipoproteins in bacteria. Curr Top Microbiol Immunol 125:127-157