Isolation and characterization of a porin-like protein of 45 kilodaltons from Bacteroides fragilis

MICROBIOLOGY LETTERS

ELSEVIER

FEMS Microbiology Letters 166 (1998) 3477354

Isolation and characterization of a porin-like protein of 45 kilodaltons from Bacteroides fragilis M.F. Odou, E. Singer, M.B. Romond, L. Dubreuil

*

Lahoratoire de BactCriolo@e, Facultt des Sciences Pharmaceutiques et Biologiques, 3. rue du Professeur Laguesse, P.O. Box 83, 59006 Lille Cedex, France

Received 22 June 1998; accepted 31 July 1998

Abstract Resistance to the combination of amoxicillin and clavulanic acid in some Bacteroides fragilis strains may be associated with a lack of porin proteins. Comparison of outer membrane protein profiles from one resistant strain (B. fragilis CFPL 358) and two susceptible strains of B. fiagilis (ATCC 25285 and CFPL 92125) showed that a few proteins were missing in the resistant strain, especially a 45-kDa protein. To determine whether this protein was a porin-like protein, we attempted to isolate it from the two susceptible strains by using gel filtration (Sephacryl S-200, Superose 6) and ion exchange chromatographies (DEAE Trisacryl, DEAE Sepharose Fast Flow). Elution from DEAE resins was poor compared to the 60-67-kDa region, which suggested that the 45-kDa protein exhibited stronger cationic forms. The use of sodium dodecyl sulfate during elution improved the recovery of the 45-kDa protein, showing that detergent modified its conformation and its ionic bounds with the chromatographic matrices but it was not sufficient for good purification. Superose 6 gel filtration also failed to separate this protein from the 60-67-kDa region. The only method resulting in the positive recovery of a purified 45-kDa band from both susceptible B. jiragilis strains was electroelution from SDS-PAGE. The swelling assay showed that the 45-kDa protein was a porin-like protein. From this study, we concluded that the 45-kDa protein from B. fragilis was a porin-like protein which might be involved in the antibiotic resistance of a strain in which this protein was missing. 0 1998 Published by Elsevier Science B.V. All rights reserved. Keyword.v: Bacteroides fragilis; Porin; Electroelution; Chromatography; Liposome swelling assay

1. Introduction The Bacteroides frugilis group - rod-shaped anaerobic Gram-negative bacteria - belongs to the human intestinal flora but is also frequently isolated from clinical specimens as opportunistic pathogens. Among the anaerobes, this group is more resistant to

* Corresponding author. Tel.: +33 3.20.96.40.08; Fax: +33 3.20.96.40.08; E-mail: [email protected] 0378-1097/98/$19.00 0 1998 Published PII: SO378-1097(98)00354-l

most antimicrobial agents. Most wild strains produce a chromosomal p-lactamase responsible for resistance to aminopenicillins and cephalosporins (except cephamycins) [l]. Moreover, B. frugilis CFPL 358, a strain from our collection, was resistant to the combination of amoxicillin and clavulanic acid. In this case, the antibiotic resistance mechanism might be induced by limited outer membrane permeability, due to the lack of porin-like proteins [2,3]. Porin channels are outer membrane proteins forming large

by Elsevier Science B.V. All rights reserved.

water-filled pores and represent the main diffusion routes through the outer membrane for hydrophilic molecules including antibiotics [4]. Wexler et al. demonstrated in B. fi-trgilis ATCC Type strain 25285 the presence of one porin [5]. Kanazawa et al. also demonstrated in this type strain the presence of three other porins [6]. In the present study, our attention focused on the proteins missing in the crude outer membrane protein (OMP) extract from B. ,fivrgi/is CFPL 358 rather than the OMP electrophoretic profiles of susceptible strains. For this purpose. we used two susceptible strains: B. ,fPugi/i.s ATCC Type strain 25285 and B. ,fiugilis CFPL 92125 from our collection. In a second step, we attempted to isolate from both susceptible strains one protein that was missing in B. ,fkrgilis CFPL 358 OMP profile by using various purification protocols. The characterization of porin activity was then displayed in reconstituted liposomes through the swelling assay.

2. Materials and methods

Experiments were carried out on B. ,/j.ugi/i.s ATCC type strain 25285 (American Type Culture Collection) and on two strains belonging to our laboratory’s collection: B. ,fkq$is CFPL 92125 and B. ,jwgpi/i.s CFPL 358 (Collection de la Faculte de Pharmacie de Lille). Strains were kept frozen at -20°C on Rosenow broth (Sanoti Diagnostic Pasteur, Marcy I’Etoile, France). They were grown on brain heart infusion (BioMerieux. Marnes la Coquette, France), in anaerobic conditions (Forma Scientific chamber).

fugation (8OOOxg: 20 min; 4”C), the supernatant was treated with 2% Triton X-100 (Sigma, Saint Quentin Fallavier, France) (I ml in IO ml of supernatant) for 30 min at room temperature, to allow for the selective solubilization of the inner membrane. After centrifugation (60000Xg; 60 min; 4”C), OMPs were collected from the pellet and suspended either in 0. I M phosphate buffer pH 7.0 or in 50 mM Tris-HCI 20 mM EDTA 2% (w/v) SDS pH 7.4 buffer for the type strain, or in 20 mM Tris-HCI pH 7.4 buffer for the clinical strains. Further centrifugation was carried out in order to exclude the insoluble components. Protein content was estimated using both the Lowry [7] and Peterson methods [8].

Crude extracts of OMPs and fractions from each purification step were analyzed by SDS-PAGE in a modified Laemmli gel [9] with an acrylamide concentration of 5% for the stacking gel and a gradient from IO to 25% for the running gel. 5530 ,ug of protein per well was loaded on the gel. Gels were silver stained or stained with Coomassie brilliant blue [IO]. 2.4. Chwtnatogruphic~

Different protocols were assayed to isolate the 4% kDa protein (Table I). The fractions containing the 45-kDa protein were detected by SDS-PAGE after each chromatography, pooled and dialyzed before the next step. The chromatographic conditions were as follows : I.

7_. After 48 h of culture. bacteria were harvested by centrifugation (60OO~gp: 20 min; 4°C) washed three times in 0.1 M phosphate buffer pH 7.0 and resuspended in 10 mM EDTA 0. I M phosphate buffer pH 7.0. Then, cells were broken down using an ultrasonic disintegrator (Vibracell 72434. Bioblock, Illkirch, France) (3X30 s; 20 kHz: 4°C). After centri-

methods

3.

Sephacryl S200 column (90x I cm) (Pharmacia); samples were eluted with buffer I (0.1 M Tris-HCI. 0.2 M NaCI, 0.02% w/v NaN,] pH X.0) at a flow rate of 6 ml min-‘. DEAE Trisacryl column (I5 x 0.9 cm) (Pharmacia); fractions from Sephacryl S200 were eluted with buffer 11 (50 mM Tris-HCI, IO mM EDTA, pH 8.0) for 2 h, then with a linear gradient of sodium chloride from 0 to 0.5 M in the same buffer. DEAE Sepharose Fast Flow column (I 5 x 0.9 cm) (Pharmacia). Fractions from the DEAE

M.F. Odou et al. IFEMS

4.

Microbiology Letters 166 (1998) 347-354

Trisacryl column were eluted in buffer II, with a linear gradient of sodium chloride from 0 to 0.5 M and from 0.5 to 2 M. Crude OMPs solubilized in buffer II with or without 2% SDS (wl v) were eluted with buffer II with or without 0.2% SDS (w/v) for 3 l/2 h, then 2 h with the same buffer supplemented with NaCl 0.3 M, then with NaCl 0.5 M. Superose 6 column (Pharmacia); FPLC method (fast protein liquid chromatography) on a Beckman system (pump 126, detector 166, 210 A sample injection valve, System Gold 3.1 version software): samples of 200 pl of crude extracts of OMPs with or without 2% SDS (w/v), or DEAE Sepharose Fast Flow fractions dialyzed against buffer III (50 mM Tris-HCl pH 8.0), filtered on a 0.22 urn Nylon filter (AIT, France) were eluted with buffer III with or without 0.2% SDS (w/v) at 0.3 ml min-’ for 3 h.

2.5. Electroelution After separating crude OMP extract by SDSPAGE in a modified Laemmli gel (18%), the gel was negatively stained as described by FernandezPatron et al. [ll]. The band corresponding to the 45kDa protein was excised and soaked in 25 mM Tris 0.2 M glycine pH 8.3 buffer until the band beTable 1 Results in SDS-PAGE

of the different chromatographic Protocol

Elution buffer without

SDS

Protocol

Protocol

Elution buffer with 0.2% SDS (OMP solubilized with SDS buffer)

I

II

Protocol Protocol

III IV

Protocol

V

“Number of bands detected by silver stained SDS-PAGE. “Presence of strongly stained 60&67-kDa bands.

349

came completely transparent for protein mobilization. Electroelution was achieved with a 422 Electroeluter module (Bio-Rad) at 10 mA per tube, for 4 h. About 600 ~1 of the eluted protein was recovered and dialyzed for one night against buffer IV (5 mM Tris-HCl pH 7.4 buffer) to remove all traces of SDS. 2.6. Liposome swelling assay To display the pore-forming activity of the protein, a liposome swelling assay [12] was performed. Briefly, 300 ~1 of DMPC (dimyristoyl phosphatidylcholine) (Sigma), dissolved in chloroform (Prolabo) in round-bottom glass tubes, was dried to a thin film with a gentle stream of nitrogen and lyophilized (Freeze Mobile 12 SL, Virtis) for 60 min. Then, 400 pl of buffer IV containing the protein suspension at the required concentration (0, 20, 40 pg per liposome) was added to each tube. Samples were sonicated for 1 min at 4°C dried under nitrogen flow at 40°C and lyophilized for 60 min. They were resuspended by adding 1 ml of 15% (w/v) dextran T 40 (Sigma) in buffer IV. Samples were incubated for 2 h at 30°C and then vortex mixed. The swelling assay was made as follows: 800 pl of an isotonic solution of sugar (40 mM) in buffer IV was added to 40 pl of liposome. The following sugar solutions were used: L-arabinose (MI 150), p-glucose (M, 180), o-sucrose (Mr 342), o-raffinose (MI 595) (Prolabo). Each time,

protocols Column

Detection protein

Sephacryl S 200 DEAE Trisacryl

peak II peak III (0.3 M NaCl)

10

peak IV (2 M NaCI)

6

peak III (0.3 M NaCI)

-40

u FPLC Superose 6 FPLC Superose 6 FPLC Superose 6

peak III (poorly stained) peak III (poorly stained) peak III (well stained)

8”

DEAE Sepharose Fast Flow

peak III (0.3 M NaCl) (well stained)

IO

U DEAE Sepharose Fast Flow DEAE Sepharose Fast Flow

of the 45kDa

Contaminating proteins” -50

12h 8

the absorbance at 400 nm was recorded (spectrophotometer Hitachi u 2000).

for 2 min

3. Results and discussion 3.1. SDS-PAGE

puttrrns of’ B. ,fiwgili.s crude OMPs

Fig. I illustrates the SDS-PAGE profiles of the dialyzed crude OMPs of the three B. ,fiugilis strains. The two susceptible strains showed similar profiles. On the other hand, in the OMP profile of B. jivgilis CFPL 358, several bands were missing, in particular proteins of 135, 114, 98, 62. 45, 40, 29, 27, 21 and 16 kDa. The proteins of 135, 114, 98 and 62 kDa might be the same as the porin proteins described by Wexler and Getty [5] (200, 94, 62 kDa) and Kanazawa et al. [6] (125, 92 kDa). Both authors used the ATCC type strain 25285 as we did. In the different studies, the discrepancies observed in the apparent molecular mass of proteins over 100 kDa might be related to the concentration of gel used, as demonstrated by

A Fig. 3. SDS-PAGE profile of the 40 -Da electroeluted protein from B fkqrlic ATCC 25285. Lane A: without treatment at WC before SDS-PAGE. Lane B: with treatment at 95°C before SDS-PAGE. Standard protein markers on the left lane (molecular mass indicated in kDa). Silver stained.

Wexler and Getty [5]. As for the additional missing proteins in B. ,frugilis CFPL 358, those ranging from 30 to 20 kDa are described by Edwards and Greenwood [13] as proteins whose molecular mass varies according to the strains of B. ,jiagilis and those under 20 kDa might be just subunits of larger proteins obtained after SDS-PAGE treatment. We focused our attention only on the 40- and 45-kDa bands, and more specifically on the 45-kDa protein, which we assumed to be a porin- like protein.

II

I

c

Fig. I. SDS-PAGE patterns of crude outer membrane protein extract& from the strains of B. ,j%gi/l.,. Lane A: B. frcrEi/i.s ATCC 25285. Lane 6: B ,furgi/i~ CFPL 92125. Lane C: B. /iqqi/i.\ CFPL 358. The arrows indicate the missing protems, compared with the two other profiles. The two white arrows indicate the 40. and the 45-kDa proteins missing in this profile. Standard protein markers (Pharmacia) on the left lane (molecular mass illdicated in kDa). Stained with Coomassic brilliant blue.

3.2. Purtiul purificution of’ the 45-kDu protein ,fi~m B. ,fiagilis A TCC 25285 and B. jixgilis CFPL 92125 hi chrornatographic methods To isolate the 45-kDa protein from B. jiugili.s, we attempted to adapt the chromatographic methods usually used to isolate OMPs from Gram-negative bacteria [l&16]. The results obtained with the different protocols tested are summarized in Table 1. Gel filtration on Sephacryl S-200 was used in a

M.F. Odou et al. IFEMS

Microbiology

first step to remove proteins of high and low molecular masses. The fraction containing the 45-kDa protein was further fractionated by ion-exchange chromatography. Anionic matrices were chosen because porin proteins are usually positively charged (pHi < pH 7) [17]. The protocol however failed to isolate a pure 45-kDa protein. Moreover, electrophoretie analysis showed that it was poorly stained compared to the other bands in the area of 60-67 kDa. The ionic bonds of the 45kDa protein on the DEAE columns must be too strong to ensure good recovery of this protein. Ion exchange chromatography used in a single step (protocol II) gave a high number of bands contaminating the 45-kDa protein. Gel filtration was very useful in excluding many proteins and protein-lipopolysaccharide complexes. The use of Superose 6 FPLC improved the isolation procedure since some proteins were eliminated but the 60-67kDa ones remained with the 45-kDa protein. Furthermore, the 60&67-kDa contaminating proteins were always more strongly stained than the 45-kDa protein. When the OMP crude extract was directly applied on Superose 6 (protocol III), the 45-kDa protein was still contaminated and was very poorly stained. Because the protein quantity collected was too small, it was unlikely to purify the 45-kDa protein containing fraction any further. Using the same column with OMPs solubilized with SDS 2% in the final extraction buffer and 0.2% SDS in the elution buffer (protocol IV), it clearly appeared that the 45-kDa protein was better recovered, but the contaminating proteins were still in high concentrations. In the same way, protocol V with the DEAE Sepharose Fast Flow column increased the recovery of the 45-kDa protein which was, however, still contaminated. Kanazawa et al. [6] obtained from B. frugilis ATCC 25285 the isola-

Fig. 3. Liposome swelling assay. Recording of the OD at 400 nm for 120 s of the different liposomes with isotonic solutions of sugars (arabinose, glucose, sucrose, raffinose). Results are expressed in OD variation in % according to time. A: Negative control liposome (without any protein included). B: Positive control proteoliposome: 200 ug of crude OMP extract from ATCC 25285. C: Proteoliposome with 200 pg of crude OMP extract from CFPL 358. D: Proteoliposome with the 45-kDa protein (40 ug of protein per liposome).

Letters 166 (1998)

351

347-354

tion of a porin of 51 kDa using ion matography. The 45-kDa protein tightly bound to the DEAE columns isolated by Kanazawa et al. From

exchange chromust be more than the porin these results, it

A

Y

omomomuma

-MbCJP-CQlO(rl c

Time

c

(s)

+

arabinose

cl

-+--glucose & sucrose +-

raffinose

Ov)c)b7Uv)Uv)Q

-mtrOr--Grow

cc

Time

(s)

C

--K-raffinose

Time

0

a Time

(s)

% 8 (s)

a

seems that adapted to they led to and perfect

chromatographic methods are not well purifying the 45-kDa protein because a considerable dilution of the samples purification was not obtainable.

3.3. Pur$!cution of’ the 45-kDu protein by. rlectroelution To solve the purification problem of the 45kDa protein, we considered a new approach, inspired by the studies of Wexler et al. on B. distusonis [IS] and B. ,fiugilis [5]. We collected the 45-kDa protein from an electrophoretic gel. by electroelution. either from B. jLagi1i.r ATCC 25285 or from B ,fiwgilis CFPL 92125. This technique was easier to carry out. less time-consuming, and ensured good recovery of the protein. With a view to test the pore-forming ability of the electroeluted protein, electroelution was first performed on a native gel, without SDS. As the yield of elution was too low, SDS buffer was required after SDS-PAGE to electroelute the protein. Minetti et al. [19] showed that Escherichia coli OmpF was resistant to SDS but not to heating beyond 95°C. On that account, samples were either heated or not before SDS-PAGE. At this point in the study. we noticed that the electrophoretic profiles were different with or without heating. The absence of heating led to the detection of a new band at 40 kDa in place of the band at 45 kDa. Elution of this 40-kDa OMP showed, when again subjected to SDS-PAGE (Fig. 2), either a band at 40 kDa without treatment at 95°C or at 45 kDa after treatment at 95°C. So we assumed that the 40- and the 45-kDa proteins, both missing in the B. fiagilis CFPL 358 strain, were two forms of the same protein. This difference in mobility has already been described for many OMPs in Gram-negative bacteria [20]: they are called heat modifiable proteins. Bolla et al. [21] showed in Cump?Mx~cter jejuni porins that the heating and concentration of SDS modified the conformation of the proteins, which altered the mobility. and consequently the apparent molecular mass of the proteins. The protein concentration obtained after elution, estimated by the Lowry method, was about 25 ug ml-’ from one electrophoresis gel. With chromatographic methods, 3 days of manipulation gave only a few nanograms of non-purified protein. Electroelu-

tion appears to be an appropriate method to isolate satisfactory quantities of purified proteins. 3.4. Highlighting porin-like uctivity: welling proteoliposomes

assu,v oj’

i. 4. I. Swelling ussuy with crude OA4Ps If a protein has pore-forming ability, test solutes (i.e. various sugars) can penetrate into it when it is included in the bilayer membrane of phospholipid liposomes; this is followed by a rapid influx of water; the swelling of the liposomes reduces their turbidity, as can be noted on a spectrophotometer at 400 nm. With structural proteins, only a very slight but negligible shrinkage of vesicles is observed : there is no swelling of the liposomes. A proteoliposome with crude OMP extract from the ATCC 25285 strain at a concentration of 200 ug per liposome was used as positive control (Fig. 3B). We chose a high concentration of protein because the proportion of porin proteins in the whole OMP extract might not be very high. Nevertheless, the swellings were more significant than those observed with the crude OMP extract from the CFPL 358 strain at the same concentration (Fig. 3C). Those results confirm that the resistant CFPL 358 strain has a lower proportion of porins in its OMPs, which must play a role in antibiotic resistance. 3.4.2. Swelling a.ssuy tvith electroeluted proteins The test was performed with electroeluted proteins A negative control - liposome with no protein included ~- was carried out. No swelling phenomenon was observed (Fig. 3A). Another liposome, with a structural protein of 20 kDa, showed no swelling phenomenon either (data not shown). Results with the 45-kDa electroeluted protein, at a concentration of 40 ug per liposome, are presented in Fig. 3D. The observed swelling demonstrates the porin activity of the 4%kDa protein. The number of porin proteins per liposome influences penetration rate [22]; data with concentrations of IO and 20 ug of protein per liposome are not shown here because the OD variation was too weak. Monosaccharides (arabinose, glucose) and disaccharides (sucrose) provoked a swelling phenomenon. With trisaccharide (raffinose), no swelling was observed. These results indicate that penetration through porin channels is related to the

M.F.

Odou et al. IFEMS

Microbiology

molecular mass of the solutes. This agrees with the findings of Nakae et al. [23] on the porins of E. coli and those of Cuchural et al. [24] with the OMPs of B. fragilis. Kobayashi and Nakae [25] reported that the exclusion limits of OMPs of B. fragilis were similar to the molecular mass of disaccharides: 340400 kDa. The size of porin channels can be determined thanks to the relative penetration rates of solutes of different molecular masses.

Letters 166 11998)

[2]

[3]

[4] [5]

[6]

4. Conclusion Many investigations have been carried out on porins of different species. However, little is known about porins of anaerobic Gram-negative bacteria [26]. Only Wexler et al. [5,18] isolated major OMPs from B. distasonis and from B. fragilis ATCC 25285 and showed pore-forming abilities. Kanazawa et al. [6] also identified three porins from B. fragilis. Piddock and Wise [27] reported the deficiency of OMPs of 49-50 kDa in some clinical isolates of B. fragilis resistant to cefoxitin; this suggested the presence of porin in the outer membrane of susceptible strains of B. fragilis but this was not demonstrated. In our study, we have shown on both susceptible strains B. fragilis ATCC Type strain 25285 and B. fragilis CFPL 92125- that the 45-kDa protein is a porin-like protein. This protein is absent from B. fragilis CFPL 358, as well as other porin-like proteins. It is probably involved in the resistance phenomenon of this strain. Further works are under way to determine the importance of this protein in resistance to antibiotics and to characterize it more thoroughly. To our knowledge, this is the first time that an OMP of 45 kDa has been isolated from B. fragilis and that its porin like activity has been highlighted.

[7]

[8]

[9]

[lo]

[ll]

[12]

[13]

[14]

[15]

Acknowledgments This work was supported in part by Quick Medical Service group ADECCO.

[16]

References

[17]

[l] Dubreuil,

L. (1996) Bacteroides fragilis:

&tat de la sensibilite

[18]

347-354

353

aux antibiotiques, evolution des resistances. Med. Mal. Inf. 26(HS), 196207. Gutmann, L., Williamson, R. and Collatz, E. (1984) The possible role of porins in bacterial antibiotic resistance. Ann. Intern. Med. 101, 554557. Nguyen Van, J.C. and Gutmann, L. (1994) Resistance aux antibiotiques par diminution de la permeabilitt chez les bacteries a gram ntgatif. Presse Med. 23, 5222531. Nikaido, H. (1994) Porins and specific diffusion channels in bacterial outer membranes. J. Biol. Chem. 269, 39053908. Wexler, H.M. and Getty, C. (1996) The isolation and characterization of a porin protein from Bocteroides frugilis. Anaerobe 2, 3055312. Kanazawa, K., Kobayashi, Y., Nakano, M., Sakurai, M., Gotoh, N. and Nishino, T. (1995) Identification of three porins in the outer membrane of Bacteroides frog&. FEMS Microbiol. Lett. 127, 181-186. Lowry, O.H., Rosebrough, N.J., Farr, A.L. and Randall, R.J.R. (1951) Protein measurement with the Folin phenol reaction J. Biol. Chem. 193, 265-275. Peterson, G.L. (1977) A simplification of the protein assay method of Lowry et al. which is more generally applicable. Anal. Biochem. 83, 346356. Laemmli, U.K. (1970) Cleavage of structural proteins during the assembly of the head of bacteriophage T4. Nature 227, 680-685. Stephano, J.L. (1986) Advantage of picrate fixation for staining polypeptides in polyacrylamide gels. Anal. Biochem. 152, 308-311. Femandez-Patron, C., Calero, M., Rodriguez Collazo. P., Garcia, J.R., Madrazo, J., Musacchio, A., Soriano, F., Estrada, R., Franck, R., Castellanos Serra, L.R. and Mendez, E. (1995) Protein reverse staining: high efficiency microanalysis of unmodified proteins detected on electrophoresis gels. Anal. Biochem. 224, 203-211. Nikaido, H., Rosenberg, E.Y. and Foulds, J. (1983) Porin channels in Escherichiu coli: studies with B-lactams in intact cells. J. Bacterial. 153, 232-240. Edwards, R. and Greenwood, D. (1997) Distinctive outer membrane protein and lipopolysaccharide composition of Bacteroides fragifis strains that produce metallo-S-lactamase. Anaerobe 3, 2333236. Bakken, V., Aaro, S. and Jensen, H.B. (1989) Purification and partial characterization of a major outer membrane protein of Fusobacterium nucleatum. J. Gen. Microbial. 135, 32533262. Kawamoto, Y., Hayakawa, M. and Abiko, Y. (1991) Purification and immunochemical characterization of a recombinant outer membrane protein from Bacreroides gingivalis. Int. J. Biochem. 23, 1053-1061. Gotoh, N., White, N.J., Chaowagul, W. and Woods, D.E. (1994) Isolation and characterization of the outer membrane proteins of Burkholderia (Pseudomonas) pseudomallei. Microbiol. 140, 7977805. Jap, B.J. and Walian, P.J. (1990) Biophysics of the structure and function of porins. Q. Rev. Biophys. 23, 367403. We&r, H.M., Getty, C. and Fisher, G. (1992) The isolation

and characterization of a major outer membrane protein from Bucteroi&\ t/irtcrsor~l\. .I. Med. Microbial. 37. 165- 175. [19] Minetti. C.A.S.A., Tai, J.Y.. Blake, MS.. Pullen. J.K.. Liang. S.M. and Remeta, D.P. (1997) Structural and functional characterization of a recombinant PorB Class 2 protein from ,Ye/a.reriu nwningikh Conformational stability and porin activity. J. Biol. Chem. 272. IO710 10720. [20] Kennel. W.L. and Holt. S.C. (1991) Extractmn purification and characterization of major outer membrane proteins from WdineNo ~‘wto ATCC 33238. Infect. Immun 59. 3740 3749. [21] Bolla. J.M.. Loret. E.. Zaleuski, M. and Pages. J.M. (1995) Conformational analysis of the (irr,~p~lohn~,tcr jejwi porin. J. Bacterial. 177. 4266 4271. [22] Page, W.J.. Huyer. G.. Huycr. M. and Worobec. E.A. (198’)) Characterization of the porins of Cun~p~lohacter jt,juni and C’fuw&ohrrcrrr w/i and implications for antibiotic susceptibility. Antimicrob. Agents Chemother. 33, 297 303.

[23] Nakae, T.. Ishii. J. and Tokunaga, M. (1979) Subumt structure of functional porin oligomers that form permeability channels in the outer membrane of Escherich di. J. Biol. Chem. 254. 1457 1461. [24] Cuchural. G.J.. Hurlbut. S.. Malamy. M.H. and Tally, F.P. ( 1988) Permeability to p-lactams in B~~cterokk~ /rugi/i,v. J. Antlmicrob. Chemother. 22. 7X5%790. [25] Kobayashi, Y. and Nakae. T. (1986) The permeability propcrty of the outer membrane of Buctrroidcs /rqi/O. a strictly anaerobic opportunistic pathogen. Biochem. Biophys. Res. Commun. 141. 292 298. [26] Wexler. H.M. ( 1997) Port-formmg molecules in gram-negative anaerobic bacteria. Clin. Infect. Dis. 25 (Suppl. 2). 284 286. 1271 Piddock. L.J.V. and Wise. R. (1987) Cefoxitin resistance in Buctuwdc~.v species: evidence indicating two mechanisms causing decreased susceptibility. J. Antimicrob. Chemothcr. 19, 161 170.