EISEVIER
FEMS Microbiology
Letters 136 (1996) 9 l-97
Crucial domains are conserved in Enterobacteriaceae ValCrie Simonet, Monique MalEa, Didier Fourel, Jean-Michel Jean-Marie Pa& * Etweloppe
et permkation
chez les Entt!robacte’ries, Joseph-Aiguier,
Received
UPR 9027, IBSM,
Centre National
B.P. 71. 13402 Marseille
18 November
1995; accepted
porins Bolla,
de la Recherche Scienti$que,
31 Chemin
Cedex 20, France
1 December
1995
Abstract With the recent resolution of the crystal structures of several bacterial porins, it is worthwhile to define the generality of their organization throughout the Enterobacteriaceae. The distribution of specific epitopes was analysed among various Gram-negative bacterial porins using anti-peptide antibodies specific to exposed, transmembrane spanning, or pore-forming regions of Escherichiu coli porins. Anti-peptide antibodies which recognized the exposed epitopes indicated a great variability among the bacterial porins analysed. Interestingly, an antigenic site located in the internal loop L3 constricting the pore diameter was present in the majority of the bacterial porins tested. Two epitopes located in domains involved in subunit interaction were also highly conserved. The presence of these common peptides suggested a conservation of specific regions involved in the functional organization of the enterobacterial porins. Keywords:
Escherichia
coli; Enterobacteriaceae;
Bacterial
porins; Anti-peptide
1. Introduction The major outer membrane proteins OmpF and OmpC of Escherichia coli K-12 are bound to lipopolysaccharide and are organized as trimeric complexes [l]. The primary amino acid structures of OmpC, OmpF, Lc, NmpC, and PhoE porins are quite similar [2,3] and a high conservation of postulated transmembrane segments has been noted [4]. They form large, water-filled channels allowing the diffusion of small hydrophilic molecules into the periplasm [5,6]. klactam antibiotics use the porin as
* Corresponding author. Tel.: +33 91 16 45 32; Fax: +33 91 71 2 1 24; E-mail:
[email protected]. Federation of European Microbiological SSDI 0378-1097(95)00499-8
Societies
antibodies;
Epitopes;
Antibiotic
resistance
the major pathway to cross through the outer membrane which protects Gram-negative bacteria against harmful compounds [7]. The three-dimen$onal structure of OmpF has been solved at 2.4 A and each monomer is folded as a B-barrel built by 16 antiparallel e-strands containing the channel [8]. The gap between one loop (L3) and the inner wall of the barrel forms a narrow constriction in the pore [8,9]. It is therefore valuable to compare the reactivity of the non-specific porins of other Enterobacteriaceae with antibodies specific for defined segments of the OmpF. Several peptides were selected to generate anti-peptide antibodies and used to investigate the distribution of the epitopes among various enterobacterial porins.
2. Materials
rabbits. Sera were collected and specific antibodies were prepared by using affinity columns.
and methods
2.1. Strains, plasmids
and medium 2.3. Dot blot anal_vsis
The strains used were E. coli B and BZB 1107 [lO,l l]. The plasompF (E. coli BE, ompF::Tn5) mids used (pFDl19, pFD284, pFD285, pLc4, pLG361, and pMY 150) have been previously characterized [2,11,12]. The Enterobacteriaceae came from the “Collection de I’Institut Pasteur” (CIP) and the “Service des EndrobactCries” (Institut Pasteur), except E. coli strains and plasmids (laboratory collection), and A. baumanii and E. cloacae (L. Gutmann, UniversitC Paris VI, France). Bacteria were routinely grown in Luria-Bertani (LB) medium at 28°C or 37°C. 2.2. Antisera and anti-peptide
Purified porins (0.5 pug) were applied to nitrocellulose sheet after a pretreatment at 37°C or 96°C. An initial saturation with TBS (50 mM Tris . HCl, 150 mM NaCl, pH 8) containing 10% bovine serum was made overnight at 4°C. The nitrocellulose sheet was then incubated in TBS containing 0.1% octylpolyoxyethylene (Octyl-POE, Bachem AG-Switzerland) and 10% bovine serum for 2 h at room temperature with either polyclonal or anti-peptide antibodies directed against the porin. The detection was then performed with alkaline phosphatase-conjugated goat anti-rabbit 1gG antibodies.
antibodies
Polyclonal antibodies directed against denatured OmpC or OmpF monomers were as used previously [ 10,131. Various synthetic peptides (Fig. 1) corresponding to different regions of the OmpF or OmpC porins (commercially acquired from Neosystem Laboratoire) were used to produce the antibodies in
2.4. SDS-polyactylamide munodetection
L5
and im-
Bacterial pellets (corresponding to E&t”,,, = 0.02) were solubilized in loading buffer at 96°C [ 10,141 and loaded onto SDS-polyacrylamide gels (10% polyacrylamide, 0.1% SDS). The immunodetection
antibody
L6
gel electrophoresis
peptide sequence
L*3
hxatton residues
F1
YNKDGNKVDL
4.13
F2
YARLGFKCETQ
40-w 71.79
F3
ECADAQTGN
F4
DMLPEFGGDTAY
F6
YEYEGFGIVGA
ltt6”190
F7
TGLKYDANNI
216.225
F8
AQYQFDFGLRPS
261.272
F9
YTKSKAKDVEG
275-285
Cl CL Fd FL Cd MoF21
113-124
KNGNPSGEGFTSGVTNNG
antiserum
prepared
antiserum antiserum monwlonat
against the unfolded
prepared prepared
antibody
152-169 197-212
ISSSKRTDAQNTAAYI
against the native against the unfolded
OmpF OmpF OmpC
selected against the native OmpF
Fig. 1, Presentation of the anti-peptide antibodies and location of the respective antigenic sites. The regions involved in the antigenic sites are indicated on the OmpF topological model (LI, LZ, and L4 to L8, exposed loops: L * 3. internal loop located inside the pore lumen; Fl-F9, antigenic sites). The antigenic peptides (commeicially acquired from Neosystem Laboratory) are numbered starting from the N-terminus of the mature porin (OmpF or OmpC).
V. Simonet
et al. / FEM.5
Microhiolog~
with antibodies was carried out after electrotransfer onto nitrocellulose as described above.
Letters
Table I Immunorecativity Antibody
of the OmpC and OmpF porins OmpF
OmpC 96°C
37°C
96°C
FI F2 F3 F3 F6 F7 F8 F9 Cl C2
_ _ _ _ _ _ _ _ _ _
+ + + + + + + + _ _
_ _ _ _ _ _ _ _
+ + _
Fd Ft Cd MoF2 1
+ ++
++ + + _
+/_
+ -
+ _
++
37°C
3. Results 3. I. Immunoanalyws of OmpC and OmpF pot-ins The sequence of the various antigenic peptides and their respective location in OmpF are presented in Fig. 1. Among these, one (F4) corresponds to the internal constriction loop L3 of the channel, three (Fl, F2. F3) are involved in subunit association, three (F6, F7, F8) are located on the B-strands or turns to the periplasmic side, and finally one (F9) corresponds to a &strand close to the external surface [8]. Two regions are specific to the OmpF (F3, F9), the others being conserved in the OmpC and Lc porins [4]. In addition. two specific peptides (Cl and C2l are only present in the OmpC sequence [2,3]. The anti-peptide antibodies directed against the exposed part of OmpF did not react with the protein in immunodotting analyses (Table 1). A positive signal was expected with F3 and F9, since the respective antigenic sites are on the exterior of the molecule according to the three-dimensional structure of OmpF IS]. The loop L2, containing the F3 site, interacts with loops L2, L3, and L4 from the adjacent subunit [8], and consequently these numerous interactions probably mask the epitope. Interestingly, loops L4 to L7 are necessary for the detection of the epitope recognized by the monoclonal antibody MoF2l directed against cell surface exposed OmpF regions [ IO, 1 I]. The interaction between these loops may hide the area containing the F9 epitope on the trimer surface. In the case of antigenic sites located close to the periplasmic turns, the smooth end of the barrel [8], their availability may be hindered by the O-strand interactions in the native trimer. To increase the accessibility of the various epitopes which are masked in the native conformation, pretreatment at 96°C. which completely converted native trimers to unfolded monomers [ 151, was used. Under these conditions (Table 11, a signal was obtained with all of the antibodies tested except for MoF21, which only recognizes trimeric OmpF [lo]. The antibodies directed against the specific OmpC peptides generate a very similar profile, reflecting
93
136 (I YY6) Y 1 -Y7
+/+
The respective
antibodies
were obtained
and tested by immunodotting. (trimer)
or heat-denaturated
+ + + + _ + +
from immunized rabbits
0.5 g.ug of purified,
(unfolded
under native
after incubation at 96°C)
porin:, were applied onto nitrocelluloae. After the saturation step, the nitrocellulose antibodies
sheet was incubated
and then with alkaline
first with
the specific
phosphatase-conjugated
goat
anti-rabbit antibodies. -.
no detection: + / -
to + +. signal with increasing intensity.
the masking of the putative epitope by other regions of OmpC. However, unfolded OmpC reacted with Cl and C2 as well as with the antibodies specific for peptides shared by OmpC and OmpF. 3.2. l$ect site
of amino acid substitution
on antigenic
Four substitutions have been previously isolated in the OmpF [ 111. Among them three are located in the selected immunogenic peptides (F4 and F9, respectively). The substitution G119D is located inside the internal loop L3 and changes the channel properties [ 161. The two others, E284K and G285D are in loop L7 [I I]. No effect of G119D was observed on the F4 site (Fig. 2). The detection with the F9 antibody was conserved on the G285D mutated porin. In contrast, the E284K substitution induced the loss of the antigenic site (Fig. 2). These results indicate that the E284 is involved in the epitopic area corresponding to F9 antibody. E284K caused the loss of the E21 antigenic site while G285 eliminated the E18, El9 and E20 epitopes [I I].
V. Simonet et al. / FEMS Microbiology Letters 136 (1996) 91-97
94
3.3. Cross-reactivity bacteria
of porins from
mentary domain specifically present in the E. coli OmpC porin [3]. In the case of antibodies directed against the conserved regions (Fig. 11, several crossreactivities were observed with Fl, F2 and to a lesser extent with F6, F7, and F8. The F2 site was detected in the majority of enterobacterial porins used here. An interesting point concerns the antigenic site recognized by the Fl antibody (located in the amino terminal part) which is detected in most tested strains. In the crystal structure, a salt-bridge connects the amino and carboxy termini of the monomer creating the 16th &strand [8]. Since the carboxy-terminal amino acid (phenylalanine) in the porin sequences is highly conserved due to its strategic role [8,17], it seems reasonable to propose that the conservation of
Gram-negative
The anti-peptide antibodies were tested with a range of Gram-negative bacteria. Bacterial pellets were solubilized at 96°C before SDS-PAGE and the reactivity of the monomeric porins to the antibodies was assayed. A large number of strains reacted with polyclonal antibodies directed against the denatured E. coli porins (Table 2). The F3 antibody only detected the E. coli OmpF, while F9 also recognized a porin in E. cloacae (Table 2). A similar response was obtained with Cl and C2, which detect only the E. coli OmpC. It is interesting to note that the region containing the Cl epitope corresponds to the supple-
Table 2 Conservation
of specific antigenic
E. coli BZBl107 BZBllO7, BZB1107, BZBl107,
sites among Enterobacteriaceaee Antibodies
Strains
plG361 pMYl50 plc
Porin 0 OmpF OmpC Lc
Cedecea davisae Citrobacter freundii Citrobacter koseri Edwardsiella tarda Enterobacter aerogenes Enterobacter cloacae Hafnia alvei Kluyvera ascorbata Leminorella grimontii Proteus mirabilis Proteus vulgaris Prooidencia stuartii Salmonella typhimurium Serratia marcescens Shigellaflexneri Yersinia enterocolitica Yersinia pestis Yersinia pseudotuberculosis Acinetobacter baumanii Campylobacter jejuni Erwinia carotauora Erwinia chrysantemi
Cd
Fd
F3
F9
Cl
C2
F2
F6
F4
Fl
F7
_ + + +
+ + +
+ _
_ + -
+ -
_ + _
+ + +
+ + +
+ + +
+ + +
+ + +
+ + + + + + + + + + + + + + + + + + _ + + +
+ + + + + + + + + + + + + + + + + + + +
_ _ _ _ _ _ _ _ -
+ _ _ _ _ _ _ -
_ _ _ _ _ -
_ _ _ _ _ -
+ + + + + + + + + + + -/+
_ + +
+ + +
_ + _ -/+
+ + + + + +
+ + + + + + + + + + + + + + + + + + _ + +
+ + _ + _ + +
IfI + -/+ 1 -_
-/+ - -- -_ + - + -
+ +
+ + + _ +
F8
Similar amounts of bacterial cells grown in LB medium were solubilized at 96°C under standard conditions. Proteins were separated by SDS-PAGE and electrotransferred onto nitrocellulose. The detections were performed with the antibodies against denaturated OmpC (Cd) or OmpF (Fd) and with the various anti-peptide antibodies. - , no detection; - / + to + , signal with increasing intensity; nd, not determined.
V. Simonet et al. / FEMS Microbiology
B
C
Fig. 2. Immunoanalysis of the subsituted OmpF porins. After solubilization in usual loading buffer, the samples were separated by SDS/PAGE. The immunodetections were carried out with polyclonal antibodies directed against the denatured OmpF (A), with the F4 antibody (B), or with the F9 antibody (C). Arrows indicate the monomer. Only the relevant parts of the immunoblots are shown.
the first turn may reflect a similar behaviour in the outer membrane insertion of enterobacterial porins. In addition, Fl and F2 peptides are located in subunit contact regions [81. The unexpected conservation of these epitopes, noted here, may reflect a well-adapted organization, explaining the unusually strong stability of the trimeric conformation [8,14,15]. 3.4. Detection OmpF lumen
of the antigenic sites located in the
The presence of the F4 epitope, which was detected in the majority of the tested porins (Table 2), suggests that a region which overlaps the consensus sequence reported in the porin family [4] is present in most enteric bacterial porins. The conservation of this region probably reflects a high selective pressure on this functional porin domain. For the antigenic site recognized by F2, the corresponding peptide contained the residue Arg42 located in the lumen of the pore [8]. In this narrow region, the positive
Letters
136 (19%)
91-97
95
charges (Lysl6, and Arg42, 82, and 132) face towards the negative charges (Asp1 13, Glul17) located in loop L3, generating an electrostatic field in transverse orientation in the porin channel [9]. The analyses carried out with these two antibodies, F2 and F4, indicate the conservation of at least two domains which are strategic in the structure of OmpF pore [8,9]. Edwardsiella and Proteus-Prollidencia did not react with the F4 antibody, although a conservation of the F2 site was detected. In the case of strains producing several porins, e.g. the D and F porins of E. cloacae, separated by using SDS-urea-PAGE, the F4 antigenic site has been recently detected on both porins [ 181. In addition, positive responses were also obtained with some species classified in the subgroup y-2 of purple bacteria 1191 (E. chryantemi, E. carotoaora). In the case of C. jejuni, for which a cross-immunoreactivity with the OmpC porin was reported previously [ 13,201, the various anti-peptide antibodies failed to detect a common epitope. In addition, the conservation of specific regions between E. coli, S. ~phimurium and S. marcescens porins described here fits in with the recently reported gene sequences (EMBL, 23 1594; GenBank, L24960).
4. Discussion Among the various immunological probes used in this study, two were directed against OmpF-specific regions containing three punctual mutations previously described [l 11. Although the substitution G119D, located in the eyelet region of the pore, induced drastic alterations of the thermal stability and channel properties [ 11 ,161, no modification of the immunological profile was observed with the F4 antibody corresponding to this region. Interestingly, the two other mutations, E284K and G285D located in the L7 cell-surface-exposed loop, abolished the colicin N binding [I 11. E284K impaired the recognition with the anti-peptide antibodies directed against the F9 region while no effect was obtained with G285D. Taking into account a similar response previously reported [l l] with the monoclonal antibody MoF21 prepared against the native trimer, these
results suggest a strategic role of the E284 in the antigenic structure. The L2 loop, containing the F3 site, forms numerous interactions with the adjacent subunit to stabilize the trimer [8]. Since the compatibility between loops L2, L3, and L4 is necessary for the trimer stability [14], this suggests a specific role for the L2 loop in the trimerization process. A similar hypothesis could be proposed for the Cl and C2 sites, although the OmpC structure has not yet been reported. OmpC displays a homologous antigenic design with the various anti-peptide antibodies, indicating the conserved immunoreactivity of the selected regions. A puzzling question concerns the immunoreactivity of the trimeric conformation after SDS-PAGE and electrotransfer onto nitrocellulose sheets. Under these conditions, the F4 antibody reacts weakly with this epitope on OmpC and OmpF trimers (data not shown). Interestingly, a proteinase K cleavage site was previously identified at position Glyl 10 [21] in OmpF trimers. Taken together, these observations indicate that under certain conditions (electrophoresis and electrotransfer or long time incubation with protease) an internal domain of OmpF or OmpC becomes accessible to some agents without obvious alteration of their trimeric organization. Knowledge of the structural organization of nonspecific pore proteins in Enterobacteriaceae is critical for understanding of the molecular sieve present in the membrane and of antibiotic diffusion through the outer membrane. Analyses of the responses obtained with the different probes used in this study, in conjunction with the three-dimensional structure of the OmpF porin, indicate high conservation of the peptides corresponding to regions which are buried within the outer membrane. The most noteworthy conservation of the porin organization reported in this work concerns loop L3 which contributes to the pore lumen, and the regions involved in subunit association [8]. The relative diffusion rate of various cephalosporins through E. coliporins [22,23] and the identification of regions critical for the organization of the pore lumen could help explain the efficiency of 13-lactam antibiotics against diverse members of the enterobacterial family. Concerning the diffusion of antibiotics through the channel, several mutations located in the eyelet region of OmpC or OmpF modify the drug susceptibility by increasing
the pore size [24,25]. Conversely, the insertion of a small peptide in the restriction region of the PhoE pore confers measurable resistance to the hydrophilic antibiotic cephaloridine [26]. In addition, we recently described a point substitution, Gll9D, located in the OmpF eyelet which alters porin properties [ 161 and causes decreased susceptibility to various 13-lactam antibiotics (Fourel et al., manuscript in preparation). Since this domain is well conserved in Enterobacteriaceae, mutagenesis of the residues located in the constriction area may elucidate the mechanism of antibiotic diffusion through the pore.
Acknowledgements We thank Professors A. Engel, P. Grimont, L. Gutmann, S. Mizushima, J.P. Rosenbusch, P.J. Sansonetti, C.A. Schnaitman, M. Simonet, and the Immunotech, for the generous gifts of plasmids, purified porins, strains, and monoclonal antibody. We gratefully acknowledge A.P. Pugsley for helpful suggestions. This work is supported by the Centre National de la Recherche Scientifique and the Institut National de la Sante et de la Recherche Medicale (CRE9306 10).
References 111Nikaido,
H. and Vaara, M. (1987) Outer membrane. In: co/i and Salmonellu t\phimurium: cellular and molecular biology (Neidhardt, F.C., Ingraham, J.L., Low, K.B., Magasanik, B., Schaechter, M. and Umbarger, H.E., Eds.), Vol. I, pp. 7-22. American Society for Microbiology. Washington. D.C. t21 Blasband, A.J., Marcotte, W.R. and Schnaitman, C.A. (1985) Structure of lc and nmpC outer membrane porin protein gene of lamboid bacteriophage. J. Biol. Chem. 261, 1272312732. 131 Mizuno, T., Chou, M.Y. and Inouye, M. (1983) A comparative study on the genes for three porins of the E. co/i outer membrane. J. Biol. Chem. 258, 6932-6940. [41 Jeanteur, D., Lakey, J.H. and Pattus, F. (1991) The bacterial porin superfamily: sequence alignment and structure prediction. Mol. Microbial. 5, 2153-2164. 151 Benz, R. and Bauer, K. (1988) Permeation of hydrophilic molecules through the outer membrane of Gram-negative bacteria: review on bacterial porins. Eur. J. Biochem. 176, l-19. [61 Nikaido, H. (1994) Porins and specific diffusion channels in bacterial outer membranes. J. Biol. Chem. 269, 3905-3908. Escherichia
V, Simonet et ~1. / FEMS Microhio/og.v
H. (1989) Outer membrane barrier as a mechanism of antibacterial resistance. Antimicrob. Agents Chemother. 33, 1831-1836. [8] Cowan, S.W., Schirmer, T., Rummel, G.. Steiert, M., Ghosh, R.. Pauptit, R.A.. Jansonius, J.N. and Rosenbusch, J.P. (1992) Crystal structures explain functional properties of two E. co/i porins, Nature 358, 727-733. [9] Karshikoff, A., Cowan, S.W., Spassov. V., Ladenstein, R. and Schirmer, T. (I 9941 Electrostatic properties of two porin channels from Eschrrichicr co/i. J. Mol. Biol. 240, 372-384. [lo] Lupi. N., Bourgois. A., Bernadac, A., Laboucarie, S. and Pages. J.-M. (1989) Immunological analysis of porin polymorphism in E. co/i B and K-12. Mol. Immunol. 26, 10271036. [I I] Fourel, D.. Mizushima, S., Bernadac, A. and Pages, J.-M. (19931 Specific regions of Eschrrichia co/i OmpF protein involved in antigenic and colicin receptor sites and in stable trimerization. J. Bacterial. 175, 2754-2757. [ 121 Nogami, T.. Mizuno. T. and Mizushima, S. (1985) Construction of a series of ompF-ompC chimeric genes by in vivo homologous recombination in Escherichia co/i and characterization of the translational products. J. Bacterial. 164, 797-80 1. [13] Kervella, M., Fauchere. J.-L., Fourel, D. and Pages, J.-M. ( 19921 Immunocrossreactivity between outer membrane pore proteins of Cumpylohacter jejuni and Escherichicr co/i. FEMS Microbial. Lett. 99, 28 l-286. [14] Fourel, D., Bernadac, A. and Pages, J.-M. (1994) Involvement of exposed loops in trimeric stability and membrane insertion of Eschrrichin co/i OmpF porin. Eur. J. B&hem. 222, 625-630. [ 151 Rosenbusch. J.P. (19741 Characterization of the major envelope protein from E. co/i. J. Biol. Chem. 249, 8019-8029. [16] Jeanteur. D., Schirmer, T.. Fourel, D., Simonet, V., Rummel, G.. Widmer. C., Rosenbusch, J.P., Pattus, F. and Pages, J.-M. ( I9941 Structural and functional alterations of a col-
Letters
[7] Nikaido,
[ 171
[18]
[ 191 [20]
[21]
[22]
[23]
[24]
[25]
[26]
136
f 1996)91-97
‘)I
icin-resistant mutant of OmpF porin from Ercherichirr co/i. Proc. Natl. Acad. Sci. USA 91, 10675-10679. StruyvC, M.. Moons, M. and Tommassen. J. (I991 1 Carboxy-terminal phenylalanine is essential for the correct assembly of a bacterial outer membrane protein. J. Mol. Biol. 218. 141-148. MallCa, M., Simonet, V.. Eun-Hee, L., Collatz, E.. Gervier. R.. Gutmann, L. and Pages, J.-M. (19951 Biological and immunological comparisons of Enterohtrctrr c~/otrccw and Escherichir~ co/i porins. FEMS Microbial. Lett. 129. 27% 280. Woese, C.R. (19871 Bacterial evolution. Microbial. Rev. 5 I, 221-271. Bolla, J.-M.. Loret. E., Zdlewski, M. and Pages. J.-M. (19951 Conformational analysis of the Cnrnpvlohtrcter ,jrjurli porin. J. Bacterial. 177. 4266-427 I. Hoenger, A.. Pages. J.-M., Fourel. D. and Engel, A. (I9931 The orientation of porin OmpF in the outer membrane of E. co/i. J. Mol. Biol. 233. 400-413. Yoshimura. F. and Nikaido, H. (19851 Diffusion of B-lactam antibiotics through the porin channel of E. co/i K-12. Antimicrob. Agents Chemother. 27. 84-92. Nikaido, H. and Normark, S. (19871 Sensitivity of f
