The structural and functional organization of H-NS ... - Semantic Scholar

Download PDF
72 downloads 1568 Views 1MB Size Report
Comment
BpH3 protein of Bordetella pertussis exhibits a low amino acid conservation with H-NS, particularly in the. N-terminal domain. To obtain information on the struc-.
Molecular Microbiology (1999) 31(1), 319–329

The structural and functional organization of H-NS-like proteins is evolutionarily conserved in Gram-negative bacteria Philippe Bertin,1* Nora Benhabiles,2 Evelyne Krin,1 Christine Laurent-Winter,3 Christian Tendeng,1 Evelyne Turlin,1 Annick Thomas,2 Antoine Danchin1 and Robert Brasseur 4 1 Unite´ de Re´gulation de l’Expression Ge´ne´tique, Institut Pasteur, 28 rue du Dr Roux, 75724 Paris Cedex 15, France. 2Unite´ INSERM U10, Hoˆpital Bichat-Claude Bernard, 75877 Paris Cedex 18, France. 3Laboratoire de Chimie Structurale des Macromole´cules, Institut Pasteur, 28 rue du Dr Roux, 75724 Paris Cedex 15, France. 4 Centre de Biophysique Mole´culaire Nume´rique, Faculte´ Universitaire des Sciences Agronomiques, 5030 Gembloux, Belgium. Summary The structural gene of the H-NS protein, a global regulator of bacterial metabolism, has been identified in the group of enterobacteria as well as in closely related bacteria, such as Erwinia chrysanthemi and Haemophilus influenzae. Isolated outside these groups, the BpH3 protein of Bordetella pertussis exhibits a low amino acid conservation with H-NS, particularly in the N-terminal domain. To obtain information on the structure, function and/or evolution of H-NS, we searched for other H-NS-related proteins in the latest databases. We found that HvrA, a trans -activator protein in Rhodobacter capsulatus, has a low but significant similarity with H-NS and H-NS-like proteins. This Gram-negative bacterium is phylogenetically distant from Escherichia coli. Using theoretical analysis (e.g. secondary structure prediction and DNA binding domain modelling) of the amino acid sequence of H-NS, StpA (an H-NS-like protein in E. coli ), BpH3 and HvrA and by in vivo and in vitro experiments (e.g. complementation of various H-NS-related phenotypes and competitive gel shift assay), we present evidence that these proteins belong to the same class of DNA binding proteins. In silico analysis suggests that this family also includes SPB in R. sphaeroides, XrvA in Xanthomonas oryzae and VicH in Vibrio cholerae. These results demonstrate Received 7 June, 1998; revised 16 September, 1998; accepted 18 September, 1998. *For correspondence. E-mail [email protected]; Tel. (1) 40 61 35 56; Fax (1) 45 68 89 48. Q 1999 Blackwell Science Ltd

that proteins structurally and functionally related to H-NS are widespread in Gram-negative bacteria. Introduction Prokaryotic cells contain many proteins involved in the organization and/or function of their chromosomal DNA. These proteins are generally small, abundant and basic (Hayat and Mancarella, 1995). Among them, HU, IHF, FIS and H-NS have been studied extensively during the past decade. HU is present in E. coli as a heterodimer capable of wrapping DNA into nucleosome-like particles. HU homologous proteins have been found in numerous prokaryotes as well as in organelles of eukaryotes (Oberto et al., 1994). IHF shows considerable amino acid sequence similarity with HU but, unlike it, IHF binds to a specific DNA sequence (Goosen and van de Putte, 1995). FIS is involved in recombination, replication and transcription processes (Finkel and Johnson, 1992). H-NS is one of the most abundant DNA binding proteins in enterobacteria. Numerous phenotypes have been associated with hns mutations, especially a modification in the expression of several plasmid and chromosomal genes. Most of these target genes are also regulated by environmental parameters, such as pH, osmolarity and temperature (Atlung and Ingmer, 1997). Moreover, a possible role of H-NS in many other biological processes such as replication, recombination and transposition has also been suggested (Ussery et al., 1994). H-NS is a neutral protein isolated more than 20 years ago as a transcription factor (Jacquet et al., 1971). It was later shown to be involved in the organization of the bacterial chromosome by affecting the level of DNA condensation (Spassky et al., 1984). H-NS exists essentially as a homodimer (Falconi et al., 1988) and binds preferentially to curved DNA in vitro (Bracco et al., 1989; Yamada et al., 1991; Owen-Hughes et al., 1992; Jordi et al., 1997). The three-dimensional structures of HU, IHF and FIS have been determined and their DNA binding domain identified (Tanaka et al., 1984; Yuan et al., 1991; Rice et al., 1996). In contrast, no information is available concerning H-NS structure, except for the organization of its C-terminal domain that has been resolved by nuclear magnetic resonance (NMR; Shindo et al., 1995). However, by characterization of dominant negative mutants and

320 P. Bertin et al. mutational analysis, it has been proposed that H-NS contains at least two distinct domains. The N-terminal region is implicated in protein–protein interactions and the C-terminal region in DNA binding (Ueguchi et al., 1996; Williams et al., 1996). H-NS exhibits strong amino acid conservation in enterobacteria (Atlung and Ingmer, 1997) as well as in closely related bacteria such as Erwinia chrysanthemi (GenBank accession number X89444) and Haemophilus influenzae (Fleischmann et al., 1995). Moreover, a protein (StpA) showing 57% identity with H-NS has been characterized in E. coli (Zhang et al., 1996). Outside these groups, only one H-NS-like protein has been characterized so far: BpH3 from Bordetella pertussis, a phylogenetically more distant bacterium (Goyard and Bertin, 1997). Using theoretical and experimental studies, we provide evidence that H-NSlike proteins are ever more widespread and that they represent a large family of functionally and structurally related DNA binding proteins. Results To search for H-NS-related DNA binding proteins, we screened the latest SWISSPROT database (release 35.0) using the BLITZ program (Smith and Waterman, 1981). We found a protein, HvrA, that shows a low (30%) but significant similarity to the whole sequence of E. coli H-NS (Fig. 1). HvrA is a 103-amino-acid protein (137 in H-NS) known as a trans -acting positive regulator involved in the control of photosystem gene expression in R. capsulatus (Buggy et al., 1994).

Effect of H-NS and H-NS-related proteins on various hns-related phenotypes Because of the low similarity between H-NS and HvrA, it was of interest to determine whether HvrA was able to reverse the effects of H-NS deficiency in a mutant strain. Mutations in the hns gene abolish motility in Salmonella typhimurium (Hinton et al., 1992) and in E. coli (Bertin et al., 1994). Therefore, we tested the effect of H-NS, StpA, BpH3 and HvrA synthesis on the swarming behaviour of

Fig. 1. Alignment of R. capsulatus HvrA sequence with the H-NS sequence of E. coli (Pon et al., 1988). Residues conserved in both sequences are in grey boxes. The MULTALIGN method (Corpet, 1988) was used for initial alignment and refined manually (see below).

Fig. 2. Motility assay on semi-solid medium. A. FB8 wild-type strain. B. BE1411 hns:: SmR mutant. C. BE1411/pLGH-NS. D. BE1411/pDStpA. E. BE1411/ pSUH3.19. F. BE1411/pDIA523. Plasmids pLGH-NS, pDStpA, pSUH3.19 and pDIA523 produce H-NS, StpA, BpH3 and HvrA respectively. The results are representative of three independent experiments. White bars represent 10 mm.

an hns mutant. As expected, the hns ::SmR mutant was non-motile on semi-solid medium (Fig. 2), while hns strains producing H-NS or BpH3 showed a motility close to that of the wild type (Bertin et al., 1994; Goyard and Bertin, 1997). Introduction of a plasmid encoding StpA or HvrA into an hns mutant also restored swarming but to a somewhat lesser extent (Fig. 2). Motility examination by light microscopy of the hns mutant expressing hns and hns related genes gave similar results (data not shown). Moreover, synthesis of H-NS, StpA, BpH3 and HvrA in a hns strain carrying a transcriptional fusion between the flagellin structural gene and lacZ resulted in increased b-galactosidase activity compared with the hns mutant (Table 1). Besides the loss of motility, numerous phenotypes have been associated with hns mutations, e.g. b-glucoside utilization (Defez and De Felice, 1981), serine susceptibility (Lejeune et al., 1989), mucoidy (Sledjeski and Gottesman, 1995) and reduced growth yield (Barth et al., 1995). Like H-NS, synthesis of StpA, BpH3 and HvrA in the hns mutant strain reversed, at least in part, the hns phenotype with regard to b-glucoside utilization, mucoidy, growth yield and/or serine susceptibility (Table 2). Q 1999 Blackwell Science Ltd, Molecular Microbiology, 31, 319–329

H-NS-like proteins in Gram-negative bacteria 321 Table 1. Expression of fliC–lacZ transcriptional fusion.

Strain

Genotype

Plasmid a

b-Galactosidase specific activity b

YK3421 BE1414 BE1414 BE1414 BE1414 BE1414

Wild type hns-1001 hns-1001 hns-1001 hns-1001 hns-1001

¹ ¹ pDIA547 pYCStpA pSUH3.19 pDIA523

5950 6 255 173 6 3 4140 6 143 2658 6 106 3293 6 71 1163 6 44

a. Plasmids pDIA547, pYCStpA, pSUH3.19 and pDIA523 produce H-NS, StpA, BpH 3 and HvrA respectively. b. Data are the mean values 6 standard deviations of three independent assays. b-Galactosidase activities are expressed as described previously (Bertin et al., 1994).

Proteome analysis of hns mutant producing H-NS and H-NS-related proteins The protein pattern of an hns strain has been drawn up recently using two-dimensional electrophoresis (LaurentWinter et al., 1997). To study the ability of the different H-NS-like proteins to substitute for H-NS in E. coli further, we evaluated their effects on the hns mutant proteome. In agreement with our previous results, the hns mutation decreased the accumulation of GlyA and OmpF proteins compared with the wild-type strain (Fig. 3) and increased the accumulation of DnaK, ProX, Gnd and GroEL. Synthesis of H-NS, StpA, BpH3 and HvrA in the hns strain restored, at least in part, the wild-type protein profile on a two-dimensional gel.

Structural and functional organization of the N-terminal domain in H-NS and H-NS-like proteins To identify structural and/or functional domains in H-NS, StpA, BpH3 and HvrA, sequences were aligned by MULTALIGN and hydrophobic cluster analysis (HCA) methods, and secondary structures were predicted (Fig. 4). These Table 2. Effect of H-NS and H-NS-like proteins on various H-NS-related phenotypes.

results suggest that the hydrophobic core of the N-domain is made up of two subdomains separated by a loop (loop 1) in H-NS and H-NS-like proteins (Fig. 4). The N-domain is linked to the C-terminal part by a second loop (loop 2). In H-NS, the first N-subdomain extends from the N-terminus to Val at position 37 and the second N-subdomain from Lys at position 57 to Val at position 82. The first N-subdomain is a mixed a-b structure, whereas the second is more helical. The level of sequence conservation in H-NS and StpA supports the idea that these structures are similar in both proteins. The homology of BpH3 and HvrA with H-NS and StpA is less prominent. In fact, the first N-subdomain and loop 1 are shorter in BpH3 and HvrA than in H-NS and StpA; loop 2 is longer in BpH3 than in any other protein and shorter in HvrA than in H-NS and StpA. The N-terminal domain seems to play a crucial role in the oligomerization process of H-NS and StpA. This has been demonstrated in vitro as well as in vivo (Williams et al., 1996; Ueguchi et al., 1997). These observations prompted us to see whether, despite a low amino acid conservation in their N-terminal domain, H-NS-like proteins such as BpH3 and HvrA were able to form dimers. To achieve high-level conditional expression of StpA and HvrA in E. coli, Nde I– Xho I DNA fragments including the stpA and hvrA coding regions were cloned into the expression vector pET-22b. The resulting plasmids pDIA536 and pDIA540 carry a recombinant gene that encodes a hybrid protein with six histidines at the C-terminus of StpA or HvrA respectively. Recombinant proteins were purified from E. coli extracts as described in Experimental procedures. The ability of these proteins to interact in vitro was analysed in EDC-NHS-catalysed cross-linking experiments (Fig. 5). Like H-NS (kindly provided by S. Rimsky) and StpA, BpH3 (Goyard and Bertin, 1997) and HvrA could be cross-linked into species with a twofold increase in their apparent molecular weight. This indicates that each

E. coli strain (relevant genotype)a

b-Glucoside utilization b

Serine susceptibility c

Mucoidy d

Growth yield e

FB8 (wild-type) BE1411 (hns:: SmR ) BE1411/ pLGH-NS BE1411/ pDStpA BE1411/ pSUH3.19 BE1411/ pDIA523

¹ þ ¹ ¹ 6 6

¹ þ ¹ ¹ ¹ 6

¹ þ ¹ ¹ ¹ þ

þ ¹ þ þ 6 þ

a. Plasmids pLGH-NS, pDStpA, pSUH3.19 and pDIA523 were introduced into BE1411 hns strain to synthezise H-NS, StpA, BpH3 and HvrA respectively. b. þ, b-Glucoside metabolism was revealed by the appearance of red colonies on MacConkey– salicin agar plates; ¹, no metabolism by white colonies; 6, intermediate metabolism by pink colonies. c. þ, No growth on 0.4% glucose M63 medium containing 40 mg ml¹1 L-serine; ¹, growth on the same medium; 6, intermediate growth level. d. þ, Mucoid phenotype was observed on 0.4% glucose M63 medium; ¹, no mucoid phenotype. e. þ, Growth yield was evaluated by measurement of the OD600 in stationary phase; ¹, 20 6 2% reduction in the growth yield in comparison with the wild type; 6, intermediate growth yield. Q 1999 Blackwell Science Ltd, Molecular Microbiology, 31, 319–329

322 P. Bertin et al.

Fig. 3. Effect of H-NS and H-NS-like proteins on the accumulation of various H-NS-dependent proteins. Quantitative data were obtained from Coomassie blue-stained two-dimensional gels. The level of protein accumulation is plotted relative to the level of synthesis in the wild-type strain, which was assigned a value of 100%. Plasmids pLGH-NS, pDStpA, pSUH3.19 and pDIA523 produce H-NS, StpA, BpH3 and HvrA respectively.

Fig. 4. Structurally based alignment and prediction of secondary structures of H-NS, StpA, BpH3 and HvrA. The alignment was achieved using both MULTALIGN and HCA plots. Identical residues in comparison with H-NS are in grey boxes. The secondary structures were predicted using the Chou–Fasman algorithm (Chou and Fasman, 1977) modified by Ralph et al. (1987). In bold is plotted the STRIDE assignment for H-NS C-terminal domain from NMR structure and for StpA, BpH3 and HvrA C-domains from MODELLER. H ¼ a-helix; G ¼ 310 a-helix; E ¼ b-strand; B ¼ b-bridge; and T ¼ turn. Q 1999 Blackwell Science Ltd, Molecular Microbiology, 31, 319–329

H-NS-like proteins in Gram-negative bacteria 323

Fig. 5. Analysis of the in vitro protein–protein interactions in H-NS and H-NS-like proteins by chemical cross-linking experiments. After 1 h incubation at room temperature with (þ) or without (¹) EDC-NHS cross-linking reagents, proteins were loaded onto a SDS–14% acrylamide gel and silver stained.

protein is able to interact with itself and to form dimers (at least) in vitro.

Structural and functional organization of the C-terminal domain in H-NS and H-NS-like proteins The C-terminal domain is the most conserved region in H-NS-like proteins. The amino acid composition of the

C-domain of StpA (residues 91–134) is 70% identical to that of H-NS (residues 91–137), while the C-domains of BpH3 (residues 82–123) and HvrA (residues 61–102) are both 40% identical to that of H-NS. Sequence conservation and distribution of hydrophobic clusters suggest that the secondary and tridimensional structures of all these C-domains are similar. On the basis of the alignment of H-NS, StpA, BpH3 and HvrA (Fig. 4), we calculated the three-dimensional structure of their C-domains by homology with the NMR structure of H-NS using MODELLER (Sali and Blundell, 1993). Three-dimensional structures were obtained by satisfying spatial restraints derived from the structure of the C-terminal domain of H-NS obtained by NMR experiments (Shindo et al., 1995). The all-atom models calculated for the C-domain of H-NS and H-NSlike proteins are shown in Fig. 6. The secondary structures of the C-domains were deduced using STRIDE, which assigns structure from the atomic co-ordinates of a protein (Fig. 4). After energy minimization, all secondary structures were restricted to a common core of b-turns and helices (a or 310 ). Structural topology of the C-terminal domain is conserved: the hydrophobic core of H-NS that holds this structure packed is built around the interactions of W109 and Y97, Y99, F134 and I135. Distances between Y97, Y99 and W109 are less than 5 A˚ . Y97 in H-NS is conserved in all proteins and corresponds to Y97, Y87 and Y67 in StpA, BpH3 and HvrA respectively. W109 is also strictly conserved and corresponds to W109, W96 and

Fig. 6. Three-dimensional structures of the C-domains of H-NS, StpA, BpH3 and HvrA. H-NS structure was determined by NMR (PDB code 1HNR), and the three other structures were calculated with MODELLER (Sali and Blundell, 1993) using the alignment presented in Fig. 4. Q 1999 Blackwell Science Ltd, Molecular Microbiology, 31, 319–329

324 P. Bertin et al.

Fig. 7. Competitive gel retardation assay with H-NS, StpA, BpH3, HvrA and restriction fragments derived from plasmid pBR322. This plasmid was cleaved by Taq I and Ssp I, and DNA fragments were incubated with the indicated concentrations of each purified protein. After protein–DNA complex formation, the fragments were resolved on a 3% MetaPhor agarose gel. The arrow indicates the position of the 217 bp fragment containing the bla promoter.

W77 in StpA, BpH3 and HvrA respectively. W109 is surrounded by two polar residues, threonine or serine, in all proteins. The strong amino acid conservation in the C-terminal domain of the various H-NS-like proteins prompted us to determine whether the DNA binding properties of H-NSrelated proteins were similar to those of H-NS, i.e. a preferential binding to the bla promoter region (Lucht et al., 1994; Zuber et al., 1994). The binding properties of H-NS, StpA, BpH3 and HvrA were evaluated in gel shift experiments. DNA of plasmid pBR322 was restricted by TaqI and Ssp I restriction enzymes to generate a 217 bp fragment encompassing the bla promoter region and additional small DNA fragments used as competitors for binding. As shown in Fig. 7, a low concentration of all proteins (from 0.25 to 2 mM, depending on the protein used) resulted in significant retardation of the 217 bp fragment carrying the bla promoter. Such an alteration in electrophoretic mobility was not observed with other DNA fragments. A twofold increase in concentration resulted in a full retardation of the bla promoter. This result is in agreement with previous data on H-NS, suggesting a co-operative binding of this protein to DNA (Rimsky and Spassky, 1990; Lucht et al., 1994). No such retardation was observed with other DNA fragments of plasmid pBR322. These data suggest that StpA, BpH3 and HvrA, like H-NS, have functional similarities with respect to DNA binding.

H-NS and H-NS-like proteins constitute a large class of DNA binding proteins As the C-terminal domain is highly conserved in the four proteins studied, we looked for similar domains in proteins of the databases. A 9-amino-acid sequence corresponding to the pattern centred around Gln at position 112 in H-NS (Fig. 4) was used to screen the latest databases using the

BLITZ program (Smith and Waterman, 1981). Two additional H-NS-like proteins were identified in Gram-negative bacteria. SPB is a 104-amino-acid transcription factor isolated from R. sphaeroides (Shimada et al ., 1996). XrvA is a 132-amino-acid protein that regulates virulence to rice in Xanthomonas oryzae (GenBank accession number X97866). Moreover, we recently identified VicH ( Vibrio cholerae H-NS-like protein), a 135-amino-acid protein (GenBank accession number AJ010791) showing a strong amino acid conservation between its C-terminal domain and that of H-NS and H-NS-like proteins (unpublished results). As in H-NS, StpA, BpH3 and HvrA, theoretical analysis suggests that the N-terminal domains of SPB, XrvA and VicH are mainly a-helical and able to adopt a coiled-coil conformation, frequently identified as the dimerization element in several transcription factors (Lupas, 1996). Moreover, they appear to be organized in two subdomains separated by a loop (data not shown). Finally, analysis of conserved residues in the C-terminal part supports the existence of the following consensus motif in the DNA binding domain of all these proteins: [AP]-K-Y-x(5,6)[GS]-[ED]-x(0,2)-T-W-[TS]-G-[QR]-G-[RK]-[TAK]-[PL]. A structurally based alignment of all H-NS and H-NS-like proteins was achieved using both MULTALIGN and HCA methods (data not shown). This multiple alignment was used to generate a phylogenetic tree using protein parsimony analysis (Fig. 8). H-NS-like proteins form three obvious clusters in a parsimony tree, which correspond to H-NS of enterobacteria, to H-NS of H. influenzae, StpA of E. coli and S. typhimurium and VicH of V. cholerae, and to H-NS-like proteins isolated from Rhodobacter and Bordetella genera.

Discussion During the past decade, genetic and biochemical analysis Q 1999 Blackwell Science Ltd, Molecular Microbiology, 31, 319–329

H-NS-like proteins in Gram-negative bacteria 325

Fig. 8. Phylogenetic tree constructed by a parsimony method (Felsenstein, 1989) from a structurally based alignment of H-NS-like protein sequences (Buggy et al., 1994; Shimada et al., 1996; Atlung and Ingmer, 1997; Goyard and Bertin, 1997) (GenBank accession numbers U82566, X89444, X97866, AJ006983 and AJ010791). Species abbreviations are as follows: ecoli, Escherichia coli ; shifl, Shigella flexneri ; salty, Salmonella typhimurium ; serma, Serratia marcescens ; erwch, Erwinia chrysanthemi ; provu, Proteus vulgaris ; haein, Haemophilus influenzae ; xanor, Xanthomonas oryzae ; borpe, Bordetella pertussis ; borbr, Bordetella bronchiseptica ; rhoca, Rhodobacter capsulatus ; rhosp, Rhodobacter sphaeroides ; vibch, Vibrio cholerae.

of the DNA binding protein H-NS has advanced our understanding of its structure and function in E. coli as well as in other enterobacteria (Ussery et al., 1994; Atlung and Ingmer, 1997). Recently, hns -like genes have been identified in Erwinia chrysanthemi (GenBank accession number X89444), in H. influenzae (Fleischmann et al., 1995) and in B. pertussis (Goyard and Bertin, 1997), but no information is available concerning the role of these proteins. Moreover, in a very recent report, it has been suggested on the basis of a multiple alignment that H-NSlike proteins could constitute a large family of proteins, at least in Gram-negative bacteria (Cusick and Belfort, 1998). In the present paper, using an approach involving theoretical analysis of functional and/or structural protein domains as well as in vivo and in vitro experiments, we provide evidence that, despite a low amino acid conservation, proteins such as H-NS and StpA of E. coli, BpH3 of B. pertussis and HvrA of R. capsulatus belong to the same family of structurally and functionally related DNA binding proteins. Synthesis of H-NS, StpA, BpH3 and HvrA in E. coli clearly showed that all these proteins were able to complement, at least in part, various phenotypes of an hns mutant such as loss of motility (Fig. 2 and Table 1), b-glucoside utilization, serine susceptibility, mucoidy and reduction in growth yield (Table 2). Moreover, expression of their Q 1999 Blackwell Science Ltd, Molecular Microbiology, 31, 319–329

structural gene in an hns mutant resulted in a protein pattern on two-dimensional electrophoresis close to that of the wild-type strain (Fig. 3). The difference between H-NS, StpA, BpH3 and HvrA in their ability to restore the wildtype pattern could be accounted for by the proteins themselves and/or the level of expression of their structural gene in the hns strain. Studies of mutant H-NS proteins suggest that the N-terminal domain plays a crucial role in transcriptional repression of the proU operon and in the oligomerization process (Ueguchi et al., 1996; Williams et al., 1996). Only a few amino acids are conserved in both regions in H-NS, StpA, BpH3 and HvrA. For example, the Leu residue in position 26 in H-NS and in StpA corresponds to Leu at positions 20 and 14 in BpH3 and HvrA respectively (Fig. 4). However, despite a low amino acid conservation in their N-terminal domain, we observed that H-NS, StpA, BpH3 and HvrA decreased, at least in part, the accumulation of ProX in an hns mutant (Fig. 3). This provides evidence that all proteins are able to repress proU expression at low osmolarity. Moreover, by chemical cross-linking experiments, we showed that BpH3 and HvrA are able to form homodimers in vitro like H-NS and StpA (Ueguchi et al., 1997; Williams and Rimsky, 1997) (Fig. 5). Taken together, these data reinforce the idea that similarities exist in the functional and structural organization of the N-terminal region in H-NS and H-NS-like proteins. Theoretical analysis of amino acid sequences further supports such a structural conservation. Indeed, the N-terminal domain was predicted mainly as a-helical in all four proteins (Fig. 4). In addition, the N-domain of H-NS and H-NS-like proteins is organized in two parts separated by a loop and connected to the C-domain by a second loop, which corresponds to a protease-sensitive linker recently identified in H-NS and StpA (Cusick and Belfort, 1998). Both loops are rich in positively charged amino acids. Their length seems to be the major structural difference between the four proteins (Fig. 4). A similar organization of the C-terminal part in H-NS, StpA, BpH3 and HvrA is suggested by secondary structure prediction (Fig. 4) and C-domain modelling (Fig. 6). These observations are corroborated by gel shift experiments. Indeed, like H-NS (Lucht et al., 1994; Zuber et al., 1994), StpA, BpH3 and HvrA were shown to bind preferentially to the bla promoter region (Fig. 7). Unlike the N-terminal domain, a strong amino acid conservation (i.e. more than 40%) was observed in the DNA binding domain of the four proteins (Fig. 4), as in that of other H-NS-related proteins such as VicH of V. cholerae (unpublished results). This observation supports a key role for these highly conserved residues in the function of all H-NS-related proteins. Mutagenesis of most of them in H-NS and StpA has been shown to result in an altered pattern of gene expression (Ueguchi et al., 1996; Williams et al., 1996). In particular,

326 P. Bertin et al. mutations affecting Y97 abolish H-NS and StpA function. W109 also seems to be important in this respect. In fact, substitution of this residue impairs H-NS binding to curved DNA but has no effect on binding to non-curved DNA (Spurio et al., 1997). The results presented in the present paper provide evidence that proteins of the H-NS family are widespread at least in Gram-negative bacteria. However, the physiological role of all these proteins as well as the genetic organization of their structural gene and the regulation of their expression present several differences between microorganisms. H-NS protein is involved in a number of cellular functions such as the modulated expression of numerous environmentally regulated genes The hns gene is not part of an operon in E. coli (Ussery et al., 1994; Atlung and Ingmer, 1997). In the presence of H-NS, the homologous protein StpA of E. coli is expressed at a low level. Inactivation (Sonde´n and Uhlin, 1996; Zhang et al., 1996) or overexpression of stpA (Fig. 3) results in altered transcription of some H-NS-regulated genes. However, unlike H-NS, StpA acts as an RNA chaperone in vitro (Zhang et al., 1996). In B. pertussis, bph3 is present as a single gene. Even if the expression of bph3 in E. coli hns mutants reversed numerous H-NS-dependent phenotypes to the wild type (Fig. 2 and Table 2), the role of BpH3 in the metabolism of the Bordetella genus is not yet known (Goyard and Bertin, 1997). Similarly, the precise function of XrvA in the physiology of X. oryzae has not been reported, except its involvement in virulence to rice. In contrast, the role of SPB and HvrA in the regulation of the puf operon in R. sphaeroides and in R. capsulatus, respectively, has been demonstrated (Buggy et al., 1994; Shimada et al., 1996). Both proteins play a role in the light-mediated expression of these genes by a specific binding to their promoter region. However, spb is transcribed monocistronically, and SPB is a repressor of the puf operon. In contrast, hvrA is transcribed polycistronically with the regulatory protein RegA structural gene, and HvrA is a trans -activator of puf expression. It is, therefore, tempting to speculate that proteins of the H-NS class have evolved from an ancestral protein (Fig. 8) of broad specificity (like H-NS in enterobacteria) to more specialized proteins either by duplication (like the paralogous protein StpA in E. coli and in S. typhimurium ) or by divergence (like the orthologous proteins BpH3, BbH3, XrvA, SPB, HvrA and VicH in B. pertussis, B. bronchiseptica, X. oryzae, R. sphaeroides, R. capsulatus and V. cholerae respectively). Experimental procedures

Bacterial strains, plasmids and growth conditions The hns:: SmR mutation of strain BE1411 results from the insertion into the FB8 wild-type strain (Bruni et al., 1977) of a SmR /SpcR cassette at the Hpa I site located at the 38th

codon of the hns gene. This strain was constructed by allele replacement using the M13mp8 phage (Blum et al., 1989). The hns-1001 mutation was introduced into YK3421 containing a fliC–lacZ transcriptional fusion (Komeda, 1982) by P1 transduction as described previously (Bertin et al., 1994), giving rise to BE1414. Rhodobacter capsulatus strain CIP 104408 was used for polymerase chain reaction (PCR) amplification of the hvrA gene. Plasmids pLGH-NS and pDStpA (Williams et al., 1996), pSUH3.19 (Goyard and Bertin, 1997) and pDIA523 were introduced into the hns strain by transformation to produce H-NS, StpA, BpH3 and HvrA, respectively, without the addition of IPTG. For compatibility reasons with resident KanR and ApR markers, plasmids pYCStpA (Williams et al., 1996) and pDIA547 were introduced into YK3421 to evaluate the effect of H-NS and H-NS-like proteins on flagella synthesis. To overproduce StpA-His6 and HvrA-His6 recombinant proteins, plasmids pDIA536 and pDIA540, respectively, were introduce into the BL21(DE3) strain (Novagen) in the presence of plasmid pDIA17 synthezising the lacI repressor (Munier et al., 1991). Bacterial growth was carried out at 378C in Luria–Bertani medium or L-M63 medium supplemented when required with 40 mg ml¹1 serine (Miller, 1972). Swarm plates containing 25 ml of tryptone medium were used to assess the motility of bacteria as described previously (Bertin et al., 1994). Metabolism of b-glucosides was tested on MacConkey indicator agar plates with 1% salicin as a carbon source. b-Galactosidase activity was measured as described previously (Bertin et al., 1994) from cells grown in M63 medium supplemented with 0.01% uracyl, 0.1% casamino acids and 0.4% succinate as a carbon source. When required, antibiotics were added at the following concentrations: kanamycin or ampicillin, 25 mg ml¹1; streptomycin, 10 mg ml¹1; chloramphenicol, 20 mg ml¹1.

DNA manipulations The hvrA gene of R. capsulatus with a putative ribosome binding site was PCR amplified as follows: one clone of R. capsulatus was resuspended in 50 ml of 10 mM Tris–1 mM EDTA buffer containing 100 mg ml¹1 Proteinase K. The suspension was incubated for 1 h at 658C. Proteinase K was inactivated by heating for 10 min at 958C, and the suspension was then centrifuged at 10 000 × g for 5 min. DNA amplification was carried out with 20 ml of supernatant using Pfu polymerase (Stratagene) and PCR core reagents from Perkin-Elmer as well as synthetic oligonucleotides 58-AACTGCAGTCCATGAAACCACTCACTGGAG-38 containing a Pst I site and 58CGGGATCCCCGGGTCGGCTGAACCCGGCGGCC-38 containing a Bam HI site. The final concentration of MgCl2 was adjusted to 2 mM. A first step, at 988C for 45 s, was followed by 30 cycles consisting of 45 s at 958C, 45 s at 568C and 45 s at 728C. The resulting DNA fragment of about 0.4 kb was purified on a Millipore Ultrafree-MC column. The hvrA gene was cloned into Pst I and Bam HI restriction sites of pSU19 plasmid (Bartolome´ et al., 1991) under the control of the lac promoter, giving rise to plasmid pDIA523. Plasmid pDIA547 was constructed by inserting the hns structural gene and its promoter region into the pSU19 plasmid in the opposite orientation with respect to the resident Plac promoter. Q 1999 Blackwell Science Ltd, Molecular Microbiology, 31, 319–329

H-NS-like proteins in Gram-negative bacteria 327 To overproduce HvrA-His6 and StpA-His6 proteins, their structural gene was PCR amplified from genomic DNA using primers introducing a Nde I cloning site at the 58 end and a Xho I cloning site at the 38 end of each fragment. PCR products were inserted into the Nde I and Xho I sites of the pET-22b vector (Novagen), giving rise to plasmids pDIA536 and pDIA540 respectively.

Preparation of total protein extracts and two-dimensional gel electrophoresis Protein samples were prepared by ultrasonic treatment, and detection of proteins was performed by Coomassie blue staining of two-dimensional gels as described previously (LaurentWinter et al., 1997). Gels were digitalized using a JX-330 scanner (Sharp). Quantitative data were obtained using the PD Quest software package (PDI).

Protein purification Recombinant proteins HvrA-His6 and StpA-His6 were purified from E. coli using NiSO4 chelation columns (Qiagen). BL21(DE3) strain, carrying either plasmid pDIA536 or plasmid pDIA540, was grown to mid-log phase in LB medium. Cultures of cells synthezising StpA and HvrA were induced by 0.5 mMIPT6 for 1 h and 3 mM IPTG for 2 h respectively. Bacterial pellets of 1 l culture of cells containing overexpressed proteins were disrupted by sonication in buffer A (50 mM NaH2PO4 , pH 8, and 300 mM NaCl) containing 20 mM and 70 mM imidazole for HvrA and StpA respectively. Cell debris was removed by centrifugation, and crude extracts were loaded onto a 3 ml NiSO4 chelation column. After adsorption on the column and washing with buffer A containing 100 mM imidazole, proteins were eluted with a 200–500 mM imidazole gradient in buffer A. Buffer was changed to 500 mM NaCl and 40 mM NaH2PO4 , pH 8. using a PD 10 desalting column (Pharmacia). Protein concentration was measured colorimetrically with the Bio-Rad reagent using the Bradford procedure. The purity of both HvrA and StpA preparations was estimated to be more than 95% on SDS–PAGE stained with Coomassie brilliant blue.

Protein–protein cross-linking Chemical cross-linking experiments with wild-type H-NS and recombinant StpA-His6, His6-BpH3 and HvrA-His6 proteins were performed using the chemical cross-linker 1-ethyl-3(38-dimethylaminopropyl) carbodiimide (EDC) and the catalyst N -hydroxysuccinimide (NHS) as described previously (Williams et al., 1996). Reaction mixtures containing each protein at a final concentration of 25 mM were loaded onto an SDS–14% acrylamide gel (Prosieve) and silver stained.

Gel retardation assay DNA fragments were obtained by digestion of plasmid pBR322 with Taq I and Ssp I restriction enzymes and incubated for 15 min at room temperature with H-NS, StpA, BpH3 or HvrA in 10 ml of reaction mixture containing 40 mM HEPES, pH 8, Q 1999 Blackwell Science Ltd, Molecular Microbiology, 31, 319–329

100 mM potassium glutamate, 10 mM magnesium aspartate, 0.022% NP40 and 0.1 mg ml¹1 BSA. Protein–DNA complexes were resolved as described previously (Goyard and Bertin, 1997).

Sequence analysis Analysis of sequence homology was carried out using the BLITZ algorithm (Smith and Waterman, 1981). The MULTALIGN method (Corpet, 1988) was used for sequence alignments, refined by the hydrophobic cluster analysis (HCA) method (Lemesle-Varloot et al., 1990). Identity and similarity scores were calculated by reference to H-NS. Secondary structure prediction was performed using the Chou and Fasman algorithm (Chou and Fasman, 1977) as modified by Ralph et al., (1987). The STRIDE algorithm was used for protein secondary structure assignment from atomic co-ordinates (Frishman and Argos, 1995). Tridimensional structures of the C-terminal domain of BpH3, HvrA and StpA were calculated by homology with the NMR structure of the C-terminal domain (47 amino acids) of H-NS (Shindo et al., 1995) using MODELLER (Sali and Blundell, 1993). Phylogenetic relationships were investigated using the protein parsimony method (PROTPARS) from the PHYLIP package (Felsenstein, 1989).

Acknowledgements We are grateful to P. Glaser, I. Martin and T. Pugsley for critical reading of the manuscript. We thank S. Rimsky and C. Badaut for providing us with H-NS protein and various plasmids and for helpful advice. We also thank S. Ngo for help in some experiments. Financial support came from the Institut Pasteur and the Centre National de la Recherche Scientifique (URA D1129). N.B. was supported by a grant from FNRSTelevie.

References Atlung, T., and Ingmer, H. (1997) H-NS: a modulator of environmentally regulated gene expression. Mol Microbiol 24: 7– 17. Barth, M., Marschall, C., Muffler, A., Fischer, D., and HenggeAronis, R. (1995) Role for the histone-like protein H-NS in growth phase-dependent and osmotic regulation of s-dependent genes in Escherichia coli. J Bacteriol 177: 3455–3464. Bartolome´, B., Jubete, Y., de Martinez, E., and la Cruz, F. (1991) Construction and properties of a family of pACYC184-derived cloning vectors compatible with pBR322 and its derivatives. Gene 102: 75–78. Bertin, P., Terao, E., Lee, E.H., Lejeune, P., Colson, C., Danchin, A., et al. (1994) The H-NS protein is involved in the biogenesis of flagella in Escherichia coli. J Bacteriol 176: 5537–5540. Blum, P., Holzchu, D., Kwan, H.S., Riggs, D., and Artz, S. (1989) Gene replacement and retrieval with recombinant M13mp bacteriophages. J Bacteriol 171: 538–546. Bracco, L., Kotlarz, D., Kolb, A., Diekman, S., and Buc, H. (1989) Synthetic curved DNA sequences can act as transcriptional activators in Escherichia coli. EMBO J 13: 4289–4296.

328 P. Bertin et al. Bruni, C.B., Colantuoni, V., Sbordone, L., Cortese, R., and Blasi, F. (1977) Biochemical and regulatory properties of Escherichia coli K-12 his mutants. J Bacteriol 130: 4–10. Buggy, J.J., Sganga, M.W., and Bauer, C.E. (1994) Characterization of a light-responding trans -activator responsible for differentially controlling reaction center and light-harvesting-I gene expression in Rhodobacter capsulatus. J Bacteriol 176: 6936–6943. Chou, P.Y., and Fasman, G.D. (1977) Beta-turns in proteins. J Mol Biol 115: 135–175. Corpet, F. (1988) Multiple sequence alignment with hierarchical clustering. Nucleic Acids Res 16: 10881–10890. Cusick, M.E., and Belfort, M. (1998) Domain structure and RNA annealing activity of the Escherichia coli regulatory protein StpA. Mol Microbiol 28: 847–857. Defez, R., and De Felice, M. (1981) Cryptic operon for b-glucoside metabolism in Escherichia coli K12: genetics evidence for a regulatory protein. Genetics 97: 11–25. Falconi, M., Gualtieri, M.T., La Teana, A., Losso, M.A., and Pon, C.L. (1988) Proteins from the procaryotic nucleoid: primary and quaternary structure of the 15-kDa Escherichia coli DNA binding protein H-NS. Mol Microbiol 2: 323–329. Felsenstein, J. (1989) PHYLIP – phylogeny inference package (version 3 2). Cladistics 5: 164–166. Finkel, S.E., and Johnson, R.C. (1992) The Fis protein: it’s not just for DNA inversion anymore. Mol Microbiol 6: 3257–3265. Fleischmann, R.D., Adams, M.D., White, O., Clayton, R.A., Kirkness, E.F., Kerlavage, A.R., et al. (1995) Whole-genome random sequencing and assembly of Haemophilus influenzae Rd. Science 269: 496–512. Frishman, D., and Argos, P. (1995) Knowledge-based protein secondary structure assignment. Proteins 23: 566–579. Goosen, N., and van de Putte, P. (1995) The regulation of transcription initiation by integration host factor. Mol Microbiol 16: 1–7. Goyard, S., and Bertin, P. (1997) Characterization of BpH3, an H-NS like protein in Bordetella pertussis. Mol Microbiol 24: 815–823. Hayat, M.A., and Mancarella, D.A. (1995) Nucleoid proteins. Micron 26: 461–480. Hinton, J.C.D., Santos, D.S., Seirafi, A., Hulton, C.J., Pavitt, G.D., and Higgins, C.F. (1992) Expression and mutational analysis of the nucleoid-associated protein H-NS of Salmonella typhimurium. Mol Microbiol 6: 2327–2337. Jacquet, M., Cukier-Kahn, R., Pla, J., and Gros, F. (1971) A thermostable protein factor acting on in vitro DNA transcription. Biochem Biophys Res Commun 45: 1597–1607. Jordi, B.J.A.M., Fielder, A., Burns, C.M., Hinton, J.C.D., Dover, N., Ussery, D.W., et al. (1997) DNA binding is not sufficient for H-NS-mediated repression of proU expression. J Biol Chem 272: 12083–12090. Komeda, Y. (1982) Fusions of flagellar operons to lactose genes on a Mu lac bacteriophage. J Bacteriol 150: 16–26. Laurent-Winter, C., Ngo, S., Danchin, A., and Bertin, P. (1997) Role of Escherichia coli histone-like nucleoid-structuring protein in bacterial metabolism and stress response. Eur J Biochem 244: 767–773. Lejeune, P., Bertin, P., Walon, C., Willemot, K., Colson, C., and Danchin, A. (1989) A locus involved in kanamycin,

chloramphenicol and L-serine resistance is located in the bglY-galU region of the Escherichia coli K12 chromosome. Mol Gen Genet 218: 361–363. Lemesle-Varloot, L., Henrissat, B., Gaboriaud, C., Bissery, V., Morgat, A., and Mornon, J.P. (1990) Hydrophobic cluster analysis: procedures to derive structural and functional information from 2-D representation of protein sequences. Biochimie 72: 555–574. Lucht, J.M., Dersch, P., Kempf, B., and Bremer, E. (1994) Interactions of the nucleoid-associated DNA-binding protein H-NS with the regulatory region of the osmotically controlled proU operon of Escherichia coli. J Biol Chem 269: 6578–6586. Lupas, A. (1996) Coiled coils: new structures and new functions. Trends Biochem Sci 21: 375–382. Miller, J.H. (1972) Experiments in Molecular Genetics. Cold Spring Harbor, NY: Cold Spring Harbor Laboratory Press. Munier, H., Gilles, A.M., Glaser, P., Krin, E., Danchin, A., Sarfati, R.S., et al. (1991) Isolation and characterization of catalytic and calmodulin-binding domains of Bordetella pertussis adenylate cyclase. Eur J Biochem 196: 469–474. Oberto, J., Drlica, K., and Rouvie`re-Yaniv, J. (1994) Histones, HMG, HU, IHF: meˆme combat. Biochimie 76: 901–908. Owen-Hughes, T.A., Pavitt, G.D., Santos, D.S., Sidebotham, J.M., Hulton, C.S.J., Hinton, J.C.D., et al. (1992) The chromatin-associated protein H-NS interacts with curved DNA to influence DNA topology and gene expression. Cell 71: 255–265. Pon, C.L., Calogero, R.A., and Gualerzi, C.O. (1988) Identification, cloning, nucleotide sequence and chromosomal map location of hns, the structural gene for Escherichia coli DNA-binding protein H-NS. Mol Gen Genet 212: 199–202. Ralph, W.W., Webster, T., and Smith, T.F. (1987) A modified Chou and Fasman protein structure algorithm. Comput Appl Biosci 3: 211–216. Rice, P.A., Yang, S.-W., Mizuuchi, K., and Nash, H.A. (1996) Crystal structure of an IHF–DNA complex: a proteininduced DNA U-turn. Cell 87: 1295–1306. Rimsky, S., and Spassky, A. (1990) Sequence determinants for H1 binding on Escherichia coli lac and faI promoters. Biochemistry 29: 3765–3771. Sali, A., and Blundell, T.L. (1993) Comparative protein modelling by satisfaction of spatial restraints. J Mol Biol 234: 779–815. Shimada, H., Wada, T., Handa, H., Ohta, H., Mizoguchi, H., Nishimura, K., et al. (1996) A transcription factor with a leucine-zipper motif involved in light-dependent inhibition of expression of the puf operon in the photosynthetic bacterium Rhodobacter sphaeroides. Plant Cell Physiol 37: 515–522. Shindo, H., Iwaki, T., Ieda, R., Kurumizaka, H., Ueguchi, C., Mizuno, T., et al. (1995) Solution structure of the DNA binding domain of a nucleoid-associated protein, H-NS, from Escherichia coli. FEBS Lett 360: 125–131. Sledjeski, D., and Gottesman, S. (1995) A small RNA acts as an antisilencer of H-NS-silenced rcsA gene of Escherichia coli. Proc Natl Acad Sci USA 92: 2003–2007. Smith, T.F., and Waterman, M.S. (1981) Identification of common molecular subsequences. J Mol Biol 147: 195– 197. Q 1999 Blackwell Science Ltd, Molecular Microbiology, 31, 319–329

H-NS-like proteins in Gram-negative bacteria 329 Sonde´n, B., and Uhlin, B.E. (1996) Coordinated and differential expression of histone-like proteins in Escherichia coli : regulation and function of the H-NS analog StpA. EMBO J 15: 4970–4980. Spassky, A., Rimsky, S., Garreau, H., and Buc, H. (1984) H1a, an E. coli DNA-binding protein which accumulates in stationary phase, strongly compacts DNA in vitro. Nucleic Acids Res 12: 5321–5340. Spurio, R., Falconi, M., Brandi, A., Pon, C.L., and Gualerzi, C.O. (1997) The oligomeric structure of nucleoid protein H-NS is necessary for recognition of intrinsically curved DNA and for DNA bending. EMBO J 16: 1795–1805. Tanaka, I., Appelt, K., Dijk, J., White, S., and Wilson, K. (1984) 3-A resolution structure of a protein with histonelike properties in prokaryotes. Nature 310: 376–381. Ueguchi, C., Suzuki, T., Yoshida, T., Tanaka, K., and Mizuno, T. (1996) Systematic mutational analysis revealing the functional domain organization of Escherichia coli nucleoid protein H-NS. J Mol Biol 263: 149–162. Ueguchi, C., Seto, C., Suzuki, T., and Mizuno, T. (1997) Clarification of the dimerization domain and its functional significance for the Escherichia coli nucleoid protein H-NS. J Mol Biol 274: 145–151. Ussery, D.W., Hinton, J.C.D., Jordi, B.J.A.M., Granum, P.E., Seirafi, A., Stephen, R.J., et al. (1994) The chromatinassociated protein H-NS. Biochimie 76: 968–980.

Q 1999 Blackwell Science Ltd, Molecular Microbiology, 31, 319–329

Williams, R.M., and Rimsky, S. (1997) Molecular aspects of the E. coli nucleoid protein, H-NS: a central controller of gene regulatory networks. FEMS Microbiol Lett 156: 175–185. Williams, R.M., Rimsky, S., and Buc, H. (1996) Probing the structure, function and interactions of the Escherichia coli H-NS and StpA proteins using dominant negative derivatives. J Bacteriol 178: 4335–4343. Yamada, H., Yoshida, T., Tanaka, K.I., Sasakawa, C., and Mizuno, T. (1991) Molecular analysis of the Escherichia coli hns gene encoding a DNA-binding protein, which preferentially recognizes curved DNA sequences. Mol Gen Genet 230: 332–336. Yuan, H.S., Finkel, S.E., Feng, J.A., Kaczor-Grzeskowiak, M., Johnson, R.C., and Dickerson, R.E. (1991) The molecular structure of wild-type and a mutant Fis protein: relationship between mutational changes and recombinational enhancer function or DNA binding. Proc Natl Acad Sci USA 88: 9558–9562. Zhang, A., Rimsky, S., Reaban, M.E., Buc, H., and Belfort, M. (1996) Escherichia coli protein analogs StpA and H-NS: regulatory loops, similar and disparate effects on nucleic acid dynamics. EMBO J 15: 1340–1349. Zuber, F., Kotlars, D., Rimsky, S., and Buc, H. (1994) Modulated expression of promoters containing upstream curved DNA sequences by the Escherichia coli nucleoid protein H-NS. Mol Microbiol 12: 231–240.