HOME HELP FEEDBACK SUBSCRIPTIONS ARCHIVE SEARCH TABLE OF CONTENTS		QUICK SEARCH:   [advanced] Author: Keyword(s): Year:  Vol:  Page:

This Article Abstract Full Text (PDF) Alert me when this article is cited Alert me if a correction is posted Citation Map Services Email this article to a friend Similar articles in this journal Similar articles in PubMed Alert me to new issues of the journal Download to citation manager Citing Articles Citing Articles via HighWire Citing Articles via CrossRef Citing Articles via Google Scholar Google Scholar Articles by Puech, V. Articles by Daffé, M. Search for Related Content PubMed PubMed Citation Articles by Puech, V. Articles by Daffé, M. Agricola Articles by Puech, V. Articles by Daffé, M.

Structure of the cell envelope of corynebacteria: importance of the non-covalently bound lipids in the formation of the cell wall permeability barrier and fracture plane

Institut de Pharmacologie et Biologie Structurale, Centre National de la Recherche Scientifique/Université Paul Sabatier (UMR 5089), 205 route de Narbonne, 31077, Toulouse Cedex 04, France¹
Laboratoire des Biomembranes, UMR 8619 CNRS-Université Paris-Sud, 91405 Orsay Cedex, France²
Lehrstuhl für Biotechnologie, Biozentrum der Universität Würzburg, Am Hubland, D-97074 Würzburg, Germany³
Institut Pasteur, Service de Microscopie électronique, 25 rue du Docteur Roux, 75724 Paris Cedex 15, France⁴
Centre de Génétique Moléculaire, CNRS, 91190 Gif-sur-Yvette, France⁵

Author for correspondence: M. Daffé. Tel: +33 561 175 569. Fax: +33 561 175 994. e-mail: daffe{at}ipbs.fr

	ABSTRACT

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

 
With the recent success of the heterologous expression of mycobacterial antigens in corynebacteria, in addition to the importance of these bacteria in biotechnology and medicine, a better understanding of the structure of their cell envelopes was needed. A combination of molecular compositional analysis, ultrastructural appearance and freeze-etch electron microscopy study was used to arrive at a chemical model, unique to corynebacteria but consistent with their phylogenetic relatedness to mycobacteria and other members of the distinctive suprageneric actinomycete taxon. Transmission electron microscopy and chemical analyses showed that the cell envelopes of the representative strains of corynebacteria examined consisted of (i) an outer layer composed of polysaccharides (primarily a high-molecular-mass glucan and arabinomannans), proteins, which include the mycoloyltransferase PS1, and lipids; (ii) a cell wall glycan core of peptidoglycan-arabinogalactan which may contain other sugar residues and was usually esterified by corynomycolic acids; and (iii) a typical plasma membrane bilayer. Freeze-etch electron microscopy showed that most corynomycolate-containing strains exhibited a main fracture plane in their cell wall and contained low-molecular-mass porins, while the fracture occurred within the plasma membrane of strains devoid of both corynomycolate and pore-forming proteins. Importantly, in most strains, the amount of cell wall-linked corynomycolates was not sufficient to cover the bacterial surface; interestingly, the occurrence of a cell wall fracture plane correlated with the amount of non-covalently bound lipids of the strains. Furthermore, these lipids were shown to spontaneously form liposomes, indicating that they may participate in a bilayer structure. Altogether, the data suggested that the cell wall permeability barrier in corynebacteria involved both covalently linked corynomycolates and non-covalently bound lipids of their cell envelopes.

Keywords: cell wall, corynebacteria, mycolic acid, polysaccharide, porin

	INTRODUCTION

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

 
Corynebacteria are widely distributed in nature and are important in several ways; apart from the human pathogens, e.g. Corynebacterium diphtheriae, serious economic losses are caused by corynebacterial diseases of animals (Coyle & Lipsky, 1990 ; Funke et al., 1997 ). In addition, some corynebacteria are of considerable economic benefit, for instance in the industrial production of amino acids (Krämer, 1994 ). They belong to a distinctive suprageneric actinomycete taxon, which also includes mycobacteria, nocardiae, rhodococci and closely related genera; all these bacteria share with corynebacteria the property of having an unusual cell envelope composition and architecture, compared to those of other Gram-positive micro-organisms. Based on most recently published data, the envelope layers of mycobacteria consist of a typical plasma membrane of phospholipid and protein, a characteristic wall of unusual structure and a complex outer layer (Daffé & Draper, 1998 ). Although detailed studies have not been carried out on the cell envelopes of corynebacteria and closely related genera, it is evident from published data that these are very similar to that of mycobacteria, specially in terms of ultrastructure and cell wall chemical composition (Barksdale, 1970 ; Marienfeld et al., 1997 ; Sutcliffe, 1997 ).

The chemical structure of the cell wall skeleton of corynebacteria, mycobacteria and related genera has been extensively studied (for reviews see Daffé & Draper, 1998 ; Minnikin & Goodfellow, 1980 ; Minnikin & O’Donnell, 1984 ; Minnikin et al., 1978 ; Sutcliffe, 1997 ); this is formed by a thick meso-diaminopimelic acid-containing peptidoglycan covalently linked to arabinogalactan, which in turn is esterified by long-chain -alkyl, ß-hydroxy fatty acids. While these fatty acids in mycobacteria, called eumycolic acids, possess a very long chain (C_60–90) and may contain various oxygen functions in addition to the ß-hydroxyl group (Daffé & Draper, 1998 ), mycolic acids found in other actinomycetes consist of homologous mixtures of saturated and unsaturated acids and contain shorter chains, e.g. C_40–50 in nocardomycolic acids and C_22–36 in corynomycolic acids (Collins et al., 1982 ; Minnikin & Goodfellow, 1980 ; Minnikin & O’Donnell, 1984 ; Minnikin et al., 1978 ). Although they are Gram-positive bacteria, corynebacteria, mycobacteria and closely related micro-organisms share with Gram-negative bacteria the property of forming in their envelope an outer barrier that is distinct from the plasma membrane. While this additional barrier in Gram-negative micro-organisms is a typical bilayer of phospholipid and lipopolysaccharide, in mycobacte 9ria and corynebacteria the cell-wall-linked mycolates and corynomycolates certainly participate in this barrier since the disruption of genes that code for mycoloyltransferases, the antigen-85 complex and PS1, respectively, causes a decrease in the amount of cell wall-bound mycolates and affects the permeability of the envelope of the mutants (Jackson et al., 1999 ; Puech et al., 2000 ). Evidence has also been presented that the chemical structure of mycolic acids plays a role in determining the fluidity and permeability of the mycobacterial cell wall (George et al., 1995 ; Liu et al., 1996 ; Dubnau et al., 2000 ). The existence in mycobacteria, corynebacteria and related genera of an outer membrane diffusion barrier is reinforced by the characterization of cell envelope proteins with pore-forming ability (Kartmann et al., 1999 ; Lichtinger et al., 1998 , 1999 ; Rieß et al., 1998 ; Mukhopadhyay et al., 1997 ; Trias et al., 1992 ; Trias & Benz, 1994 ; Senaratne et al., 1998 ).

In all currently proposed models (Minnikin 1982 ; Rastogi, 1991 ; Liu et al., 1995 ) the outer permeability barrier of mycobacteria consists of a monolayer of mycoloyl residues covalently linked to the cell wall arabinogalactan and includes other lipids which are probably arranged to form a bilayer with the mycoloyl residues. Although no sign of a second lipid bilayer has ever been reported in thin sections of mycobacterial cells (Draper, 1998 ), freeze-fractured samples of mycobacteria, corynebacteria and related genera (Barksdale & Kim, 1977 ; Benedetti et al., 1984 ; Chami et al., 1995 ; Takeo et al., 1984 ) showed that these organisms had two such planes of weakness in their envelopes, in addition to the expected plasma membrane fracture. A second fracture plane, close to the cell surface of mycobacteria, was observed. The existence of distinct lipid domains in the mycobacterial cell envelope has been indicated by studies using progressive erosion of the envelope (Ortalo-Magné et al., 1996 ) or selective lipophilic probes (Christensen et al., 1999 ). Freeze-fracture electron microscopy also showed the presence of ordered arrays on the surface of Corynebacterium glutamicum and other Corynebacterium spp. (Chami et al., 1995 ; Peyret et al., 1993 ; Soual-Hoebeke et al., 1999 ); these surface layer proteins (S-layer) have been shown to be composed of the major corynebacteria-secreted protein, PS2 (Chami et al., 1995 ; Peyret et al., 1993 ; Soual-Hoebeke et al., 1999 ). Such crystalline structures are not visible in published freeze-fractured micrographs of mycobacteria despite a singular report of a crystalline protein layer on the outside of Mycobacterium bovis BCG (Lounatmaa & Brander, 1989 ).

From recent advances in molecular biology, it appears that corynebacteria represent a model suitable for functional studies of expression of mycobacterial genes, primarily because (i) they possess the simplest cell wall structure, but one that is functionally and structurally close to those of mycobacteria and related genera; (ii) they are rapid growers and do not form clumps; and (iii) heterologous expression of mycobacterial antigens in C. glutamicum has proved to be effective (Salim et al., 1997 ; Puech et al., 2000 ). However, it has to be remembered that, unlike the other related micro-organisms, corynebacteria are widely used in biotechnology for the production of amino acids (Krämer, 1994 ); this singular property of corynebacteria suggests that, despite the global similarity of their cell envelope composition with those of related genera, they differ from them in the arrangement of its constituents. To gain further insights into the cell envelope structure of corynebacteria, various representative corynebacterial strains were examined in the present study by combining biochemical analyses and electron microscopy studies. These included the Corynebacterium genus type species C. diphtheriae, the amino-acid-producing strain C. glutamicum, corynomycolate-less strains, i.e. C. amycolatum (Collins et al., 1988 ; Barreau et al., 1993 ) and defined corynebacterial cell wall mutants (Chami et al., 1995 ; Peyret et al., 1993 ; Joliff et al., 1992 ; Puech et al., 2000 ).

	METHODS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

 
Strains and growth conditions.
The strains used were Corynebacterium amycolatum ATCC 49368^T and other corynebacterial strains reclassified in this species (Barreau et al., 1993 ), i.e. CIP (Collection Institut Pasteur) 79.37, CIP 100836, CIP 103818, CIP 103819, CIP 103820 and CIP 103821; C. diphtheriae strain C8_r(-) Tox^- (Barksdale, 1959 ); C. glutamicum CGL2005, its isogenic csp1-inactivated mutant CGL2022 (Puech et al., 2000 ), cspB-disrupted mutant CGL2025 (Peyret et al., 1993 ), and CGL1009; C. pseudodiphtheriticum (formerly C. hofmanii) strain Breuillaud (Welby-Gieusse et al., 1970 ); and C. xerosis ATCC 373^T, ATCC 7094, ATCC 7711 and ATCC 9016. All these strains were routinely grown at 30 °C with shaking (200 r.p.m.) for 24 h in BHI (3·7% Difco brain–heart infusion). For electron microscopy analysis, strains of C. xerosis were also grown for 16, 24, or 48 h in BHI, and C. glutamicum CGL2005 and its isogenic cspB-disrupted mutant CGL2025 were grown on BHI-containing agar (15 g l^-1) plates. For the structural analysis of polysaccharides, strains were grown on a corynebacterial minimal medium (Huchenq et al., 1984 ) to avoid possible contamination with polysaccharides originating from the BHI medium.

Isolation of extracellular materials.
Bacteria were grown for 48 h on minimal medium and the culture filtrates were recovered by centrifugation, filtered through a 0·2 µm sterile filter (Nalgene) and concentrated under vacuum to 1/10 of the original volume; the corresponding cells were dried and weighed. Chloroform and methanol were added to a portion of the filtrates to isolate pids as described below and cold ethanol (6 vols) was added to the remaining portion; after centrifugation of the ethanol precipitates, the pellets were extensively dialysed against distilled water and extracellular materials were obtained and analysed for carbohydrate and protein content as described by Lemassu & Daffé (1994) .

Isolation of the crude surface-exposed materials.
Cells grown in either BHI or minimal medium were harvested by centrifugation, extensively washed and gently shaken for 1 min with 10 g glass beads (4 mm diameter) per 2 g of wet cells (Ortalo-Magné et al., 1995 ). The resulting material was then suspended in distilled water and immediately filtered through a 0·2 µm sterile filter (Nalgene) and concentrated under vacuum to 1/10 of the original volume. Cold ethanol (6 vols) was added to a portion of the crude surface-exposed material, and the ethanol-precipitates were extensively dialysed against distilled water, lyophilized, and analysed for their carbohydrate and protein compositions.

CHCl₃ and CH₃OH were added to the remaining filtrate to give a final one-phase mixture of H₂O/CHCl₃/CH₃OH (0·8:1:2, by vol.; Bligh & Dyer, 1959 ) and lipids were extracted for 1 h at room temperature; a two-phase partition mixture was obtained by adding 1 vol. CHCl₃/H₂O (1:1, v/v) to the one-phase solution and lipids were recovered in the lower organic phase, dried and analysed by thin-layer chromatography (TLC) as described below.

Isolation, fractionation and analysis of whole-cell lipids.
Lipids were obtained and analysed as previously described (Puech et al., 2000 ). Briefly, lipids were extracted from wet cells for 16 h with CHCl₃/CH₃OH (1:1, v/v) at room temperature with continuous stirring; the bacterial residues were re-extracted three times with CHCl₃/CH₃OH (2:1, v/v) and the organic phases were pooled and concentrated. The crude lipid extracts were partitioned between the aqueous and the organic phases arising from a mixture of CHCl₃/CH₃OH/H₂O (8:4:2, by vol.); the lower organic phases were collected, evaporated to dryness to yield the crude lipid extracts from each strain and comparatively examined by TLC on silica gel-coated plates (G-60, 0·25 mm thickness, Merck) developed with CHCl₃/CH₃OH (9:1, v/v) or CHCl₃/CH₃OH/H₂O (30:8:1 or 65:25:4, by vol.). Detection of all classes of lipids was performed by spraying the TLC plates with either rhodamine B or 20% H₂SO₄ in water, the latter followed by heating at 110 °C; glycolipids were revealed by spraying plates with 0·2% anthrone (w/v) in concentrated H₂SO₄, followed by heating at 110 °C. The Dittmer–Lester reagent (Dittmer & Lester, 1964 ) was used for visualizing phosphorus-containing lipids.

In a parallel experiment, surface-exposed lipids were isolated from C. glutamicum cells by extracting bacterial cells with octylglucoside. Briefly, C. glutamicum CGL2005 cells were grown overnight in BHI-rich medium. After extensive washing (twice the culture volume), cells were incubated for 1 h in 25 mM Tris/HCl, pH 6·8 buffer containing 1% octylglucoside. The crude detergent extract was recovered by centrifugation at 8000 r.p.m. for 15 min, extensively dialysed for 24 h against 25 mM Tris/HCl, pH 6·8 buffer, and then centrifuged at 200000 g for 30 min; the pellet was analysed by freeze-fracture electron microscopy.

Quantification of corynomycolic acids was performed as follows. Delipidated cells (1·5 g dry weight) and lipid extracts (100 mg) of the various strains were dried under vacuum prior to weighing and were saponified (Daffé et al., 1983 ). The saponified products were acidified with 20% H₂SO₄ and the resulting fatty acids were extracted with diethyl ether, converted to methyl esters with diazomethane and dried under vacuum, dissolved in petroleum ether and applied to a Florisil (60–100 mesh, Merck) column equilibrated in petroleum ether. The column was irrigated stepwise with increasing concentrations of diethyl ether in petroleum ether. Fractionations were monitored by TLC on silica gel-coated plates using dichloromethane and fractions containing the same lipid compounds (non-hydroxylated fatty acid methyl esters or corynomycolates) were pooled and weighed. Three sequential determinations from separate preparations of delipidated cells were performed (Puech et al., 2000 ).

Fatty acid methyl esters from bacteria, delipidated cells and extractable lipids (1–2 mg) were treated with trimethylsilyl reagents (Sweeley et al., 1963 ) to derivatize hydroxylated components of the mixtures, i.e. corynomycolates, and analysed by GC. The detector response for the various classes of fatty acid methyl esters was determined using authentic samples of C_16:0 and C_32:0 corynomycolate methyl esters. Identification of non-hydroxylated fatty acid methyl esters and corynomycolate derivatives was achieved by gas chromatograph–mass spectrometry (GC–MS).

Production of cell wall, plasma membrane and cytosol fractions.
The cell fractions of the various corynebacteria were produced as previously described for mycobacteria (Daffé et al., 1990 ; Raynaud et al., 1998 ). Wet cells (5 g) were suspended in 20 ml phosphate buffer (50 mM, pH 7·5) and the resulting bacterial suspension was passed through a cell disrupter and then centrifuged at 3000 r.p.m. for 15 min to eliminate unbroken cells; cell walls were recovered from the supernatant by recentrifugation at 10000 g for 1 h. The resulting pellet was extracted with aqueous 2% SDS at 95 °C for 1 h to remove soluble proteins and finally pelleted at 10000 g; the cell walls were washed three times with 80% acetone to remove SDS and lyophilized (Daffé et al., 1990 ). The 10000 g supernatant was recentrifuged at 100000 g for 2 h to yield the membrane fraction in the pellet; the supernatant was considered as the cytosol (Raynaud et al., 1998 ).

Analysis of the cell wall arabinogalactans.
O-Methylation of cell wall arabinogalactan was performed by suspending lyophilized walls (5 mg) in 3 ml of dimethylsulphoxide; 300 µl of dimethylsulphinyl carbanion was added and the mixture was stirred for 5 h. CH₃I was slowly added and the suspension was stirred for 2 h. A mixture of CH₃OH/H₂O (1:1, v/v) was added and the reaction mixture was dialysed against water. The addition of dimethylsulphoxide, dimethylsulphinyl carbanion and CH₃I, followed by dialysis, was repeated four times. The retentate was lyophilized and the per-O-methylated arabinogalactan was hydrolysed with 2 M CF₃COOH at 110 °C for 2 h. The hydrolysed products were reduced by NaBH₄, acetylated and the resulting partially O-methylated, partially O-acetylated alditols were analysed by GC-MS (Daffé et al., 1990 ).

Production and analysis of lipoarabinomannans and lipomannan.
Delipidated cells (7 g) were extracted with 100 ml ethanol/water (1/1, v/v) for 2 h at 70 °C; the bacterial residues were recovered by filtration and re-extracted with the same solvent mixture. The two extracts were pooled and dried; then, 100 ml hot phenol/water (1/1, v/v) was added and the mixture was heated for 5 min under continuous stirring followed by partitioning of the two phases. The phenol phase was discarded and the upper phase was extensively washed and dried. The extract was solubilized in water and Triton-X114 (2% w/v) was added to the cooled suspension; the mixture was stirred for 5 min and then heated at 50 °C until two phases had formed. The detergent phase was recovered, diluted by adding 2 ml water and washed three times with CHCl₃. The resulting aqueous phase was concentrated to a final volume of 0·5 ml and the macromolecules were precipitated with 6 vols cold ethanol. The precipitate was recovered, resuspended in water (0·5 ml) and the macromolecules were reprecipitated with 6 vols cold ethanol. The resulting pellet was dissolved in 100 mM sodium acetate buffer (pH 4·7), containing 15% (w/w) propan-1-ol and fractionated by hydrophobic interaction chromatography (Leopold & Fisher, 1993 ) by loading the material onto a column of octyl-Sepharose CL-4B (Pharmacia LKB) equilibrated with the buffer; the column was first irrigated with this buffer and then with the buffer containing 20, 30, 40, 50 and 60% (v/v) n-propanol. Fractions were collected in bulk, dialysed and monitored for carbohydrate; a portion of each dialysate was analysed for its carbohydrate and lipid compositions. The remaining portion was analysed by SDS-PAGE; authentic samples of mycobacterial lipoarabinomannan and lipomannan from Mycobacterium bovis BCG, kindly provided by T. Brando (IPBS, Toulouse, France) were used as standard.

Preparation of proteins for SDS-PAGE and Western blot analysis.
Cell wall proteins from the various strains were released by incubating the cell pellet with 2% (w/v) SDS in 50 mM Tris/HCl (pH 6·8) at 100 °C for 2 min (50 µl buffer per equivalent of 0·4 ml bacterial suspension at OD₆₅₀ 1·2); the suspension was centrifuged at 12000 g for 3 min and the supernatant, which contains only cell wall-associated proteins and not cytoplasmic proteins (Peyret et al., 1993 ), was collected. Cell wall proteins from roughly 0·4 ml bacterial suspension at OD₆₅₀ 1, were analysed by SDS-PAGE. The pore-forming proteins were isolated from the culture filtrates and different cell fractions as previously described (Lichtinger et al., 1998 ). Proteins were separated by SDS-PAGE (Laemmli, 1970 ) with a 4% stacking gel and a 7% or a 10% running gel. Samples were denatured in the presence of 2% SDS in 50 mM Tris/HCl (pH 6·8). After electrophoresis, gels were either stained with Coomassie brillant blue R-250 or blotted onto nitrocellulose membrane (Towbin et al., 1979 ). In this latter case proteins were probed with rabbit polyclonal antibodies anti-PS1 (Joliff et al., 1992 ), anti-PS2 (Peyret et al., 1993 ), or anti-C. glutamicum porin (Lichtinger et al., 1998 ). Bands were detected using alkaline-phosphatase-conjugated antibodies and nitroblue tetrazolium (NBT)/5-bromo-4-chloro-3-indoxyl phosphate (BCIP) p-toluidine (Promega) as substates.

Sugar compositional analysis.
The sugar constituents of the various materials were determined after acid hydrolysis either with 2 M CF₃COOH at 110 °C for 2 h or with 1 M methanolic HCl at 80 °C for 16 h; the mixture of hydrolysed products was dried, treated with trimethylsilyl reagents (Sweeley et al., 1963 ) to derivatize monosaccharides and analysed by GC for their sugar.

Gas chromatography and mass spectrometry.
GC was performed using a Hewlett Packard HP4890A equipped with a fused silica capillary column (25 m length by 0·22 mm internal diameter) containing WCOT OV-1 (0·3 mm film thickness, Spiral). A temperature gradient of 100–290 °C at 5 °C min^-1, followed by a 10 min isotherm plateau at 290 °C, was used.

GC-MS analysis was conducted on a Hewlett Packard 5890 gas chromatograph connected to a Hewlett Packard 5989A mass spectrometer. Samples were injected in the splitless mode. The column was a 12 m HP-1. A temperature gradient of 100–290 °C (8 °C min^-1) was used.

Experiments with lipid bilayer membranes.
Black lipid bilayer membranes were formed as described previously (Benz et al., 1978 ). The instrumentation consisted of a Teflon chamber with two aqueous compartments connected by a small circular hole. The hole had a surface area of about 0·5 mm². Membranes were formed across the hole by painting on a 1% solution of a mixture (molar ratio 4:1) of diphytanoyl phosphatidylcholine (PC) and phosphatidylserine (PS) (Avanti Polar Lipids, Alabaster AL) in n-decane.

Electron microscopy
Transmission electron microscopy.
Bacterial pellets were fixed for 1 h at 4 °C with 2·5% (w/v) glutaraldehyde and 0·05 M lysine in a 0·1 M cacodylate buffer (pH 7·4) containing, or not, 0·075% (w/v) ruthenium red. Cells were washed five times in the same buffer, i.e. with or without 0·075% (w/v) ruthenium red, and postfixed for 1 h at room temperature in 1% (w/v) osmium tetroxide containing, or not, 0·075% (w/v) ruthenium red in cacodylate buffer (pH 7·4), and then rinsed with distilled water. Bacteria were suspended in 1% (w/v) aqueous uranyl acetate for 1 h at room temperature and then washed five times in distilled water. Suspended cells were embedded in 2% molten agar type IX (Sigma) before dehydration through a graded ethanol series. Dehydrated cells were embedded in Spurr medium with intermediate 1,2-epoxypropane infiltration. Blocks were conventionally cut, stained and examined with a Philips CM12 microscope operating under standard conditions.

Freeze-etched preparations for electron microscopy.
Bacterial suspensions were centrifuged at 5000 g. A drop of the pellet was placed between thin copper holders and quenched in liquid propane. The frozen samples were fractured at -125 °C in a vacuum of about 10^-7 torr as described by Aggerbeck & Gulik-Krzywicki (1986) . The fractured samples were etched at -100 °C for 3 min at 1·3x10^-5 Pa and then replicated with 1–1·5 nm of deposits of platinum-carbon, and backed with about 20 nm of carbon. The replicas were cleaned overnight with chromic acid, washed with distilled water and observed with a Philips 410 electron microscope.

	RESULTS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

 
Ultrastructure of the cell envelopes of corynebacteria
Electron microscopic examination of ultrathin sections from conventionally fixed corynebacterial cells (Fig. 1a, d, e) revealed a cell envelope structure composed of (i) a plasma membrane (PM) of 6–7 nm, (ii) a thick electron-dense layer (EDL) of 15–20 nm, (iii) an electron-transparent layer (ETL) of 7–8 nm and (iv) a thin outer layer (OL) of 2–3 nm. This ultrastructural appearance of the cell envelope of corynebacteria is similar to that found in mycobacteria (Daffé & Draper, 1998 ; Draper, 1982 ; Paul & Beveridge, 1992 ; Rastogi et al., 1986 ). However, in contrast to mycobacteria, in which a space is observed between the PM and EDL corresponding to a hypothetical periplasmic space (Daffé & Draper, 1998 ), the PM is tightly associated to the EDL in thin sections of corynebacteria, an observation consistent with a previous report on C. glutamicum (Marienfeld et al., 1997 ). When ultrathin sections were stained with ruthenium red, a stain that has been previously shown to react strongly with the surface of mycobacteria (Rastogi et al., 1984 ), a much thicker OL (35–40 nm) was seen around the various corynebacteria (Fig. 1b, c, f). Thus, as previously suggested (Daffé & Draper, 1998 ), the thin margin observed for OL in conventionally fixed cells and freeze-substituted samples of both corynebacteria (Marienfeld et al., 1997 ) and mycobacteria (Paul & Beveridge, 1992 ) may not reflect the original thickness of the layer. In addition, no difference was seen between ultrathin sections of the parent strain of C. glutamicum (Fig. 1b) and its isogenic mutant devoid of S-layer (Fig. 1c), in terms of either staining properties or thickness; this observation indicated that, in contrast with a previous suggestion (Marienfeld et al., 1997 ), the OL is not synonymous with the S-layer.

View larger version (80K): [in this window] [in a new window]   Fig. 1. Thin sections of corynebacteria prepared by conventional embedding. Bacterial pellets were fixed for 1 h at 4 °C with 2·5% (w/v) glutaraldehyde and 0·05 M lysine in a 0·1 M cacodylate buffer containing (b, c, f) or not (a, d, e) 0·075% (w/v) ruthenium red. (a) and (b) C. glutamicum CGL2005 (PS2+); (c) C. glutamicum CGL2025 (PS2-); (d) C. xerosis (ATCC 373T); (e) and (f) C. amycolatum (ATCC 49368T). OL, outer layer; ETL, electron-transparent layer; EDL, electron-dense layer; PM, plasma membrane. Bars, 200 nm.

Chemical analysis of the corynebacterial outermost polysaccharides
Analysis of the acid hydrolysis products originated from the extracellular and surface-exposed macromolecules of the various strains by GC showed that the sugar composition of the outermost constituents of the vast majority of the corynebacterial species analysed was similiar to that in mycobacteria (Lemassu & Daffé, 1994 ; Lemassu et al., 1996 ; Ortalo-Magné et al., 1995 ); these were 10–20% arabinose (Ara), 20–35% mannose (Man) and 50–70% glucose (Glc). The outermost polysaccharides from C. amycolatum were almost exclusively composed of Glc. To characterize the polysaccharides that are composed of these sugar residues, the crude ethanol precipitates from the extracellular and surface-exposed macromolecules were further investigated; the precipitates were first treated with proteases and then chromatographed on an ion-exchange column (Lemassu & Daffé, 1994 ). In both cases, most of the carbohydrate material (>90%) was eluted from the column with the buffer containing no salts (data not shown), indicating that they were composed of neutral substances; gel-filtration chromatography over a Bio-Gel P-10 column led to the isolation of three fractions (Fig. 2a). The major fraction (peak A) was exclusively composed of Glc and was eluted in the void volume; gel-filtration chromatography over a Sephadex G-200 column of the glucan fraction led to the isolation of a single peak A₁ at a position corresponding to an apparent molecular mass of 110 kDa (Fig. 2b). The second peak of the Bio-Gel P-10 column (peak B, Fig. 2a) contained Ara and Man (in a molar ratio of 1:1) and exhibited an apparent molecular mass of 13 kDa. The last fraction eluted from the Bio-Gel P-10 column was a large, broad peak (peak C, Fig. 2a) and contained Ara, Man and Glc; gel-filtration chromatography over a Bio-Gel P-4 column of this material gave two peaks (Fig. 2c) at positions corresponding to apparent molecular masses of 1·7 kDa (peak C₁, glucan, Fig. 2c) and 1 kDa (peak C₂, arabinomannan, Fig. 2c). A similar pattern was observed when the macromolecules derived from the culture filtrate of C. xerosis were analysed. Thus, like mycobacteria (Lemassu & Daffé, 1994 ; Lemassu et al., 1996 ; Ortalo-Magné et al., 1995 ), C. xerosis elaborates extracellular and surface-exposed polysaccharides composed of a 110 kDa high-molecular-mass glucan and a 13 kDa arabinomannan. In contrast to what was found in mycobacteria, low-molecular-mass arabinomannan and glucan were present in the mixture of polysaccharides from the strains examined.

View larger version (12K): [in this window] [in a new window]   Fig. 2. Gel filtration profiles of the extracellular and surface-exposed polysaccharides of C. xerosis (ATCC 373T) grown on synthetic medium. (a) Neutral polysaccharides were fractionated on a column of Bio-Gel P-10 (90 cmx1·8 cm); (b) fraction A from the Bio-Gel P-10 column was rechromatographed on a column of Sephadex G-200 (90 cmx1·8 cm). (c) fraction C from the Bio-Gel P-10 column was rechromatographed on a column of Bio-Gel P-4 (90 cmx1·8 cm). The eluant of the gel filtration columns consisted of 1% (a, b) and 0·5% acetic acid (c). Fractions, 5 ml in (a), 9 ml in (b) and 6·5 ml in (c), were collected, and the carbohydrate content of each fraction was determined by colorimetric method. V0, void volume.

View larger version (51K): [in this window] [in a new window]   Fig. 3. SDS-PAGE analysis and hydrophobic interaction chromatography of the lipoglycans extracted from corynebacteria. (a) Gel electrophoresis of an authentic sample of mycobacterial lipoarabinomannan (mLAM) and lipomannan (mLM) (lane 1) and of hot phenol/water extracts from C. glutamicum CGL2005 (lane 2), C. diphtheriae (strain C8r(-) Tox-) (lane 3), C. xerosis (ATCC 373T) and C. amycolatum (ATCC 49368T). Positions of size markers are shown on the left. cLAM1, corynebacterial lipoarabinomannan with a long polysaccharide chain; cLAM2, corynebacterial lipoarabinomannan with a short polysaccharide chain; cLM, corynebacterial lipomannan. (b) Hydrophobic interaction chromatography profile of the crude aqueous phase from the phenol/water extract from C. glutamicum on a column of octyl-Sepharose (10 cmx0·5 cm). The column was equilibrated in 0·1 M sodium acetate buffer containing 15% propan-1-ol and first eluted with the buffer (3 ml) and then with a discontinuous gradient of propan-1-ol (15–100%, 3 ml each). Fractions of 1 ml were collected, and the carbohydrate content of each fraction was determined by a colorimetric method.

Distribution of the corynomycoloyltransferase PS1 and the S-layer constituent PS2 in corynebacteria
It was recently demonstrated that one of the two major secreted protein from C. glutamicum, PS1, possesses a mycoloyltransferase activity in this species (Puech et al., 2000 ). However, like the other major secreted S-layer constituent, PS2, the occurrence of PS1 in different corynebacteria has not been investigated. To determine the distribution of PS1, proteins extracted from the outermost cell envelope compartment were analysed by SDS-PAGE. Compared to those of mycobacteria (Ortalo-Magné et al., 1995 ), the protein profiles of corynebacteria were less complex (Fig. 4a, consisting only of a few major bands, notably three bands with electrophoretic migrations similar to those of PS1 and PS2 (in the 60–70 kDa region). To characterize these proteins further, Western blot analyses were performed on them using polyclonal antibodies directed against PS1 and PS2. The anti-PS1 antibodies reacted strongly with polypeptide bands exhibiting the corresponding molecular mass in C. glutamicum (positive control, Fig. 4b, lane 1), one strain of C. xerosis (ATCC 9016; Fig. 4b, lane 3) and C. amycolatum (Fig. 4b, lane 5), and gave a faint reaction with proteins from C. diphtheriae (Fig. 4b, lane 4) and the type strain of C. xerosis (Fig. 4b, lane 2). In these latter strains, stronger reactions were observed with other polypeptides which may share some epitopes with PS1. Although the corynomycoloyltransferase activity of PS1 (Puech et al., 2000 ) was not investigated, these data suggest the presence of PS1 in all the strains examined, including in those devoid of corynomycolates, i.e. C. amycolatum (Collins et al., 1988 ; Barreau et al., 1993 ). In contrast, the anti-PS2 antibodies reacted only with proteins originated from the PS2⁺ parent strain of C. glutamicum (positive control); proteins from the PS2^- strain CGL2025 of C. glutamicum (cspB-disrupted mutant, negative control), as well as those from the other corynebacterial strains examined did not react with the anti-PS2 antibodies (data not shown). Although the production of PS2 was shown to greatly depend on the carbon source (Soual-Hoebeke et al., 1999 ), the failure to detect this protein was not attributed to the growth medium used to analyse the other corynebacterial strains since all the strains examined were grown on the same BHI medium, which had been shown to allow C. glutamicum strain CGL2005 to produce a highly ordered surface layer composed of PS2 (Chami et al., 1995 ; Peyret et al., 1993 ). It was thus concluded that the S-layer component PS2 was absent from most of the strains examined herein.

View larger version (50K): [in this window] [in a new window]   Fig. 4. SDS-PAGE and Western blot analyses of cell wall proteins of corynebacteria. Protein associated with the cell wall of bacteria were extracted with 2% (w/v) SDS in 50 mM Tris/HCl at 100 °C for 2 min and analysed on acrylamide gels; protein bands were either stained with Coomassie blue (a) or blotted onto nitrocellulose membrane, probed with rabbit anti-PS1 polyclonal antibodies and bands detected using alkaline-phosphatase-conjugated antibodies and nitro blue tetrazolium (NBT)/5-bromo-4-chloro-3-indoxyl phosphate (BCIP) p-toluidine as substrates (b). Positions of size markers are shown on the left. (a) Lane 1, C. glutamicum CGL1009 (overexpressing PS2); lane 2, C. amycolatum (ATCC 49368T); lane 3, C. xerosis (ATCC 373T); lane 4, C. diphtheriae (strain C8r(-) Tox-); lane 5, C. pseudodiphtheriticum (strain Breuillaud). (b) Lane 1, C. glutamicum CGL2005; lane 2, C. xerosis (ATCC 373T); lane 3, C. xerosis (ATCC 9016); lane 4, C. diphtheriae (strain C8r(-) Tox-); lane 5, C. amycolatum (ATCC 49368T).

View larger version (150K): [in this window] [in a new window]   Fig. 5. Freeze-fractured and deep-etched preparations of corynebacterial strains grown on BHI-containing agar plates (a–f) and cell envelope outermost lipid material (g). (a) C. glutamicum CGL2025 (PS2-); (b) C. diphtheriae (strain C8r(-) Tox-); (c) C. pseudodiphtheriticum (strain Breuillaud); (d) C. xerosis (ATCC 7711); (e) C. xerosis (ATCC 373T); (f) C. amycolatum (ATCC 49368T). Note the absence of the ordered surface layer on all the strains examined. Depending on the strains, however, a fracture plane is seen either in the cell wall fracture plane (2) close to the bacterial surface (1) in (a)–(d) or in the plasma membrane (3) in (e) and (f). (g) Crude octylglucoside extract recovered by centrifugation, extensively dialysed and then pelled at 200000 g. Note the fracture plane that occurs within the homogeneous smooth vesicles and indicates that the material analysed spontaneously forms liposomes. Bars, 500 nm (a–f); 400 nm (g).

View larger version (20K): [in this window] [in a new window]   Fig. 6. Distribution and sublocalization of the low-molecular-mass porin in corynebacteria and functional analysis of the cell wall porin of C. diphtheriae (strain C8r(-) Tox-). (a) Culture filtrates (CF), surface-exposed (S), cell wall (CW) and plasma membrane (PM) material of various strains were analysed by Western blotting using polyclonal antibodies directed against the low-molecular-mass porin from C. glutamicum. Positions of size markers are shown on the right. (b) Histogram of the probability P (G) of the occurrence of a given conductivity unit observed with membranes formed of phosphatidylcholine/phosphatidylserine (molar ratio 4:1)/n-decane in the presence of the detergent-solubilized cell wall extract of C. diphtheriae. P (G) is the probability that a given conductance increment G is observed in the single-channel experiments. It was calculated by dividing the number of fluctuations with a given conductance increment by the total number of conductance fluctuations. The aqueous phase contained 1 M KCl, pH 6·0; the applied potential was 20 mV at 20 °C. The mean single-channel conductance was 2·3 nS for 148 single-channel events. (c) Single-channel recording of a PC:PS (molar ratio 4:1)/n-decane membrane in the presence of 10 µl the cell wall fraction of C. diphtheriae treated with 1% Genapol. The aqueous phase contained 1 M KCl, pH 6·0. The applied membrane potential was 20 mV and the temperature was 20 °C.

Non-covalently associated cell envelope lipids
To explain the occurrence of the cell wall fracture plane in strains exhibiting a very low cell-wall-associated corynomycolate content, we postulated that other lipids may replace corynomycoloyl residues in the inner monolayer of the outer membrane, based on the structural similarities between C_32–36 corynomycolate-containing molecules, e.g. TDCM and TMCM (with two parallel C_16–18 chains), and C_16–18-containing lipids such as phospholipids. Therefore, non-covalently associated lipids were extracted from bacterial cells grown with organic solvents and weighed (Table 3). Interestingly, strains of C. amycolatum and the type strain of C. xerosis contained 3–6% of extractable lipids whereas the other corynebacterial strains produced significantly more lipids. TLC analysis of these lipids showed that their composition was similar to that found in the extracellular and surface-exposed materials; TDCM, TMCM and phospholipids were the main substances detectable in all the strains, except C. amycolatum, where only phospholipids were observed. Compared to the outermost lipid materials, however, many more phospholipids were detected in bacterial lipids, as expected from their partiality for plasma membrane. It was thus concluded that the presence of significant amounts of extractable lipids, e.g. TDCM and TMCM, probably contributes to the occurrence of a cell wall fracture plane in corynebacteria.

To evaluate the possible contribution of non-covalently bound lipids in the cell wall permeability barrier, the outermost material of C. glutamicum was extracted with octylglucoside; this method did not affect the integrity of bacterial cells, as judged by freeze-fracture microscopy (data not shown). The crude detergent extract, which contained mostly polysaccharides and lipids, was extensively dialysed to eliminate the detergent and to reconstitute organized structures; it was then centrifuged and the pellet was analysed by freeze-fracture electron microscopy (Fig. 5g). Homogeneous smooth vesicles whose diameter varied from 50 to 500 nm were observed throughout the samples analysed, indicating that the material examined spontaneously forms liposomes upon dialysis of the detergent; analysis of the lipid material by TLC showed that it primarily consisted of trehalose dicorynomycolates, with small amounts of phospholipid and trehalose monocorynomycolates. It follows then that the non-covalently bound lipids from the cell envelope of corynebacteria are strong candidates for participating in a membrane-like structure.

	DISCUSSION

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

 
Corynebacteria, mycobacteria, nocardiae, rhodococci and closely related genera are members of a distinctive suprageneric actinomycete taxon which includes bacterial species of such biotechnological, medical, veterinary and environmental importance that it is crucial to understand fully their physiology and metabolism (Sutcliffe, 1997 ). From the enormous work that has been devoted to the characterization of members of this taxon, it appeared that many of their distinctive properties were related to their unusual cell envelope composition and architecture, compared to those of other Gram-positive micro-organisms. Consequently, models of the mycobacterial envelope, consistent with their chemical and physical properties, have been proposed (Daffé & Draper, 1998 ; Liu et al., 1995 ; Minnikin, 1982 ; Rastogi, 1991 ) and extrapolated to other members of the taxon (Minnikin & O’Donnell, 1984 ; Sutcliffe, 1997 ).

As far as the ultrastructural appearance is concerned, the cell envelope of corynebacteria is similar to that of mycobacteria, consisting of a typical plasma membrane bilayer, a thick electron-dense layer (EDL), an electron-transparent layer (ETL) and an outer layer (OL). The electron density of the EDL makes it likely that it contains the cell wall peptidoglycan, which possesses charged groups able to bind metallic stains used in electron microscopy. The EDL is surrounded by a thin ETL which is traditionally considered to consist of mycolic residues, based on the transparency of this layer to electrons; indeed, the occurrence of eumycoloyl, nocardomycoloyl, or corynomycololyl residues covalently linked to the cell wall arabinogalactan of mycobacteria and related genera, on the one hand, and the effect of the removal of lipids from isolated mycobacterial cell walls, which causes the disappearance of the layer, on the other hand, support this interpretation. Surprisingly, however, this layer was seen in all corynebacterial species examined, including C. amycolatum, which is devoid of corynomycolates (Fig. 1); indeed, this observation could be explained by the replacement on the arabinogalactan of C. amycolatum of corynomycoloyl residues by normal-chain fatty acyl residues, i.e. C_16–18. This hypothesis was ruled out by the analysis of the purified cell walls of the bacterium, where no fatty acyl substituent could be detected, indicating that the ETL was not synonymous with a lipid layer. Furthermore, the thickness of this layer in corynebacteria is comparable to that measured in mycobacteria (6–7 nm). Assuming a linearity between the thickness of the hydrocarbon portion and the number of carbon atoms, the theoretic