The Caulobacter crescentus outer membrane protein Omp58 (RsaF)1 is not required for paracrystalline S-layer secretion

 
CONTINUE READING
FEMS Microbiology Letters 201 (2001) 277^283
                                                                                                                            www.fems-microbiology.org

The Caulobacter crescentus outer membrane protein Omp58 (RsaF)1
       is not required for paracrystalline S-layer secretion
                          Mike Reichelt, Bernd-Ulrich von Specht, Heinz P. Hahn *
                    Chirurgische Forschung, Chirurgische Universita«tsklinik, Hugstetter Strasse 55, D-79106 Freiburg i. Br., Germany

                                Received 19 January 2001; received in revised form 25 May 2001; accepted 6 June 2001

                                                          First published online 28 June 2001

Abstract

  To identify the outer membrane protein component of the Caulobacter crescentus CB2 surface-layer export machinery we used the
Serratia marcescens LipD protein to find homologs in the CB2 genome. From two homologous sequences found, one encodes a putative
OMP with a predicted molecular mass of 57.5 kDa, termed Omp58 (formerly RsaF). Comparison of membrane protein profiles revealed a
protein with an appropriate molecular mass present in wild-type, but not CB2 omp58: :kanamycin, a mutant strain with an inactivated
omp58 gene. Disruption of omp58 did not affect surface-layer production, suggesting that Omp58 is not involved in surface-layer protein
secretion and, thus, may not be the outer membrane protein component of the C. crescentus surface-layer export system. ß 2001
Federation of European Microbiological Societies. Published by Elsevier Science B.V. All rights reserved.

Keywords : ATP-binding cassette exporter; Allelic replacement; Outer membrane protein ; Surface layer; Caulobacter

1. Introduction                                                                 uncleaved C-terminal secretion signal and occurs via a type
                                                                                I or ATP-binding cassette (ABC) exporter system [5,6,7].
   Protein surface layers (S-layers) are widely spread                             In Gram-negative bacteria, the type I secretion appara-
among Caulobacter and many other eubacterial and ar-                            tus consists of three components forming a transmem-
chaebacterial genera (for review see [1]). Generally,                           brane complex, by which a C-terminal signal in the cog-
S-layers consist of regularly arranged protein subunits,                        nate protein is recognized, initiating ATP-dependent
forming paracrystalline surface arrays as the outermost                         secretion from the cytoplasm directly into the extracellular
layer on the bacterial cell envelope. Therefore, S-layers                       medium [8,9]. An ABC translocase, which contains an
are directly exposed to the environment and, in such posi-                      ABC, is embedded in the inner membrane with a cytoplas-
tion, they may have the role of a protective barrier for the                    mic domain. The membrane fusion protein (MFP) is also
cells. Since the S-layer proteins assemble on the outside of                    anchored in the inner membrane, but spans the periplas-
the cell they are considered secreted proteins. The Caulo-                      mic space. The third component resides in the outer mem-
bacter crescentus S-layer is composed of a 98-kDa regular                       brane and interacts with the MFP. One of the best char-
surface array protein (RsaA) which is non-covalently an-                        acterized type I exporter systems is that required for E.
chored to the cell surface via a speci¢c smooth lipopoly-                       coli K-hemolysin secretion, where the secreted hemolysin,
saccharide [2,3]. In contrast to most S-layers, where S-layer                   the ABC translocase, and the auxiliary MFP are geneti-
protein secretion depends on N-terminal signal peptides                         cally linked in a plasmid-borne operon, while the second
and is mediated by the general secretory pathway of the                         auxiliary component, the outer membrane protein (OMP),
cell [4], in C. crescentus, transport of RsaA relies on an                      is provided by the chromosomally encoded multifunction-
                                                                                al TolC protein [10,11].
                                                                                   So far, two putative type I secretion systems have been
  * Corresponding author. Tel. : +49 (761) 270 2743 ;                           identi¢ed, mediating the secretion of S-layer proteins in
Fax: +49 (761) 270 2579.
                                                                                Gram-negative bacteria. In Serratia marcescens, besides
  E-mail address : hahn@ch11.ukl.uni-freiburg.de (H.P. Hahn).
                                                                                the S-layer protein, three other non-related proteins are
  1
   The DNA sequence data of omp58 gene have been deposited in                   secreted by the same ABC exporter with the ABC trans-
GenBank with accession number AF331067.                                         locase and the two auxiliary proteins genetically linked in
0378-1097 / 01 / $20.00 ß 2001 Federation of European Microbiological Societies. Published by Elsevier Science B.V. All rights reserved.
PII: S 0 3 7 8 - 1 0 9 7 ( 0 1 ) 0 0 2 8 4 - 1

                                                                FEMSLE 10036 16-7-01
278                                 M. Reichelt et al. / FEMS Microbiology Letters 201 (2001) 277^283

the lipBCD operon [12]. In Campylobacter fetus, as for the              Biosystems 377 automated sequencer. Electrocompetent
S. marcescens lip system, the proposed OMP gene (sapF)                  cells of CB2 were prepared from mid-exponential phase
of the S-layer secretion system is part of the ABC exporter             cells as described by Gilchrist and Smit [15]. Prior to the
determinant [13].                                                       electrotransformation, cells were incubated with the DNA
   Both the genetic organization and regulation of the S-               on ice for 30^60 min. Electrotransformation conditions
layer secretion system in C. crescentus is still poorly under-          were essentially similar to that for E. coli [14], using 0.2-
stood. Besides the rsaA gene, the genes for the proposed                cm interelectrodal gaps with electroporation unit set to 2.5
ABC translocase (rsaD) and MFP (rsaE) have previously                   kV and 25 WF and resistance adjusted to 200 6. Subse-
been identi¢ed [6]. These genes are located immediately                 quent outgrowth was performed at 30³C in PYE medium
downstream of rsaA, but may not be combined with                        for a 2-h period.
rsaA to a transcriptional unit, because of a potential
rho-independent terminator sequence at the end of rsaA.                 2.3. PCR ampli¢cation and cloning of omp58 gene
On the other side, the OMP component has not been
identi¢ed yet. However, similar to tolC, the OMP gene                      The omp58 structural gene was ampli¢ed by PCR from
in C. crescentus is presumably not linked to the rasADE                 chromosomal CB2 DNA, using the primers 5P-CCAGA-
gene cluster [6].                                                       TCTCCATGCGAGTGCTGTCGAAAG-3P (OMP58F)
   To reconstruct the RsaA secretory apparatus in a het-                and           5P-CCAGATCTCTAGTTGCGGGGCGC-3P
erologous organism and generate de¢ned secretion-de¢-                   (OMP58R) (introduced £anking BglII restriction sites are
cient mutants, we have initiated attempts to identify the               in italics). Under optimized conditions, 100-Wl reaction
OMP component based on homologies to known OMPs                         volumes contained 100 ng of DNA template, 0.4 WM of
of other type I secretion systems. In the present work, we              each primer, 150 WM of dNTP mix, 2% dimethyl sulfoxide,
provide detailed evidence that omp58, a gene previously                 and 2.5 U polymerase. After initial incubation for 90 s at
introduced in a data base entry as the OMP component,                   94³C, 40 cycles were run, each consisting of 45 s at 94³C,
thus termed rsaF, is not required for S-layer secretion in              45 s at 56³C, and 210 s at 72³C, followed by a ¢nal in-
C. crescentus.                                                          cubation of 10 min at 72³C. The ampli¢ed 1.6-kb fragment
                                                                        was puri¢ed and cloned in frame with the GST gene se-
                                                                        quence into the single BamHI site of plasmid pGEX-3X
2. Materials and methods                                                (Amersham Pharmacia) to form plasmid pGEX-Omp58.
                                                                        DNA of one clone was double-strand sequenced using a
2.1. Bacterial strains and culture conditions                           set of strand-speci¢c primers. To avoid the potential prob-
                                                                        lem of PCR-induced errors, both strands of a second in-
   C. crescentus wild-type strain CB2 (ATCC 15252) and                  dependent clone were also sequenced. The integrity of the
the constructed mutant strain CB2 omp58: :kanamycin                     reading frame was further con¢rmed by SDS^PAGE anal-
(Kan) were grown in PYE medium (0.2% peptone, 0.1%                      ysis of IPTG-induced cell lysates. Homology searches were
yeast extract supplemented with 2 ml l31 each of 10%                    performed via the NCBI BLAST server, utilizing the im-
CaCl2 and 10% MgSO4 ). E. coli XL-1 Blue (Stratagene),                  proved gapped BLAST algorithm [16].
used for cloning procedures and propagation of plasmids,
was grown in LB broth or agar [14]. Ampicillin and Kan                  2.4. Allelic exchange of omp58 gene
were used for selection in C. crescentus and plasmid-bear-
ing E. coli strains at 100 Wg ml31 and 50 Wg ml31 , respec-                A 1.2-kb SalI fragment with the KanR gene from plas-
tively.                                                                 mid pUC4K (Amersham Pharmacia) was ¢lled in with
                                                                        Klenow enzyme. The blunted fragment was cloned into
2.2. Recombinant DNA methods                                            the 5P-PvuII site of the omp58 gene in plasmid pGEX-
                                                                        Omp58 in transcriptional orientation of the genetic
   C. crescentus chromosomal DNA was prepared with                      GST^Omp58 fusion, giving plasmid pGEX-Omp58: :Kan.
Qiagen DNeasy Tissue Kit according to the manufactur-                   E. coli cells bearing the recombinant plasmid were resis-
er's protocol. Recombinant DNA techniques were per-                     tant to Kan, indicating that the expression of the KanR
formed using standard procedures [14] with DNA-modify-                  gene was also driven by the tac promoter upstream of the
ing enzymes used in accordance with the supplier's                      GST gene. For allelic exchange, pGEX-Omp58: :Kan,
instructions (New England Biolabs). Relevant DNA frag-                  which cannot replicate in Caulobacter, was transformed
ments were excised from agarose gels and extracted with                 into CB2, and clones were selected for locus-speci¢c chro-
QIAquick Gel Extraction Kit (Qiagen). Native Pfu poly-                  mosomal insertion on Kan. To di¡erentiate between single
merase (Stratagene) was used for polymerase chain reac-                 and double cross-over events, transformants were initially
tion (PCR) according to the manufacturer's protocol. Se-                screened by chromosomal PCR ampli¢cation for the pres-
quencing reactions were prepared using the ABI PRISM                    ence or absence of the wild-type omp58 gene sequence,
Big-Dye Terminator chemistry and run on an Applied                      using primers OMP58F and OMP58R.

                                                        FEMSLE 10036 16-7-01
M. Reichelt et al. / FEMS Microbiology Letters 201 (2001) 277^283                                        279

2.5. Southern blot hybridization                                               2.7. Biotinylation of cell surface proteins

  Genomic DNA was digested with NcoI and KpnI, sep-                               After removal of S-layer by low-pH extraction step, S-
arated on 1% agarose gels (2 Wg per lane) and Southern                         layer de¢ciency of C. crescentus cells was checked by SDS^
blotted onto nylon membranes (Hybond-N, Amersham                               PAGE analysis of cell lysates. Biotinylation of S-layer-
Pharmacia) by the standard capillary and cross-linking                         cured cells was essentially performed as described by
method [14]. PCR-ampli¢ed omp58 gene fragment was                              Abath et al. [20] for Yersinia pestis. Brie£y, S-layer-cured
randomly labeled with 32 P, using the High Prime DNA                           C. crescentus cells (see Section 2.6) were washed twice with
Labeling Kit (Roche) and Nick Spin columns (Amersham                           100 mM carbonate bu¡er (pH 8.5) and resuspended to
Pharmacia) for removal of unincorporated nucleotides ac-                       approx. 5U108 cells in a ¢nal volume of 100 Wl. After
cording to the manufacturer's instructions. Hybridizations                     addition of an equal volume of freshly dissolved sulfobio-
and stringent washes were performed at 65³C by standard                        tin-N-hydroxysuccinimide ester (1.1 mg ml31 , dissolved in
protocol [14].                                                                 carbonate bu¡er) (Calbiochem^Novabiochem), cells were
                                                                               incubated for 30 min on a rotatory shaker. Reactions were
2.6. Immunodetection procedures for S-layer                                    stopped with 100 Wl of 1 M lysine^HCl and cells washed
                                                                               three times with carbonate bu¡er (pH 8.5). Preparation of
   SDS^PAGE and Western immunoblotting were done as                            membrane fractions was performed as described earlier
previously described [17] with the exception that protein                      [17]. Biotinylated proteins were detected after Western
samples were denatured for 1 h at room temperature prior                       blotting by streptavidin^alkaline phosphatase conjugate,
to loading. Blots were probed with a 1:3000 dilution of a                      with membranes from approx. 1U107 cells loaded per
polyclonal antibody against S-layer [18], and antibody                         well.
binding was detected with goat anti-rabbit IgG conjugated
to alkaline phosphatase (Sigma-Aldrich). To detect S-layer
on whole cells, washed cell suspensions were dotted onto                       3. Results
nitrocellulose and membranes processed, as for Western
blots, using horseradish peroxidase-labeled secondary anti-                    3.1. Identi¢cation and cloning of omp58
bodies. S-layer shedding assay was essentially performed
as described by Bingle et al. [19]. S-layer proteins                             To identify gene candidates for the S-layer OMP com-
were extracted from Caulobacter strains with HEPES bu¡-                        ponent, the C. crescentus genome data base released by
er (pH 2.0) following the procedure of Walker et al.                           The Institute of Genomic Research (TIGR) was searched
[18].                                                                          on the protein level for gene sequences homologous to

Fig. 1. Alignment of deduced amino acid sequences of Omp58, LipD [12], SapF [13], and TolC [21]. The comparison was performed using the ClustalW
algorithm implemented in the DNA Star program package (version 4.0) with the identity residue weight table utilized for graphical display. Structurally
similar residues (maximal distance of two) are shaded, while identical residues are highlighted in black. Asterisks above the Omp58 sequence indicate
amino acid residues which di¡er from the predicted RsaF sequence.

                                                               FEMSLE 10036 16-7-01
280                                       M. Reichelt et al. / FEMS Microbiology Letters 201 (2001) 277^283

                                                                              to GST was chosen for two reasons. First, the GST moiety
                                                                              should mask the putative signal peptide sequence, thus
                                                                              preventing direction of the native protein to the E. coli
                                                                              cell membrane. Second, the GST tag should allow easy
                                                                              puri¢cation of the Omp58 protein. All recombinant plas-
                                                                              mid clones analyzed further carried the expected 1.6-kb
                                                                              insertion. Furthermore, all those clones with the fragment
                                                                              inserted in the transcriptional orientation of the GST gene
                                                                              gave a dominant induction protein band (apparent molec-
                                                                              ular mass 80 kDa) when protein pro¢les of IPTG-induced
                                                                              and non-induced cell homogenates were compared by
                                                                              SDS^PAGE (data not shown). Thus, neither plasmid nor
                                                                              fusion protein stability appeared to be a problem.
                                                                                Sequence determination of the cloned fragment in
                                                                              pGEX-Omp58 revealed a continuous 1584-bp open read-
                                                                              ing frame encoding a putative protein with a predicted
                                                                              molecular mass of 57 466 Da. Sequencing of the omp58
                                                                              gene in a second independent clone gave an identical se-
                                                                              quence, indicating that no PCR-based errors had oc-
Fig. 2. Construction of omp58 mutant strain by gene replacement. A:
Schematic representation of the wild-type omp58 gene and the omp58
                                                                              curred. Compared to the rsaF sequence, 14 single-base
gene interrupted by a KanR cassette. The horizontal arrow indicates the       di¡erences were detected in the omp58 sequence, two of
transcriptional direction of the omp58 and the KanR gene, respectively.       which resulted in an amino acid exchange (Glu229 CLys229
The small hooked arrows symbolize the PCR primers OMP58F and                  and Gln422 CHis422 ) (Fig. 1).
OMP58R. The relative positions of restriction sites for ClaI (C), HindIII
(H), KpnI (K), NcoI (N), PstI (Ps), PvuII (P), and XmaI (X) are shown.
B: Chromosomal PCR analysis of mutant strains. Lanes: 1, plasmid
                                                                              3.2. Inactivation of omp58 gene
pGEX-Omp58; 2, plasmid pGEX-Omp58: :Kan; 3, CB2 omp58
omp58 : :Kan; 4, CB2 omp58 : :Kan-3; 5, CB2 omp58 : :Kan-17; S, DNA              An omp58 deletion mutant was constructed by insertion
fragment standard. C: Southern hybridization of genomic NcoI/KpnI-di-         of a promoter-less KanR cassette and subsequent replace-
gests of CB2 wild-type and mutant strains. Lanes: 1, CB2 wild-type ; 2,       ment of the chromosomal wild-type gene by the mutant
CB2 omp58: :Kan-3; 3, CB2 omp58: :Kan-17 ; 4, CB2 omp58 omp58: :
Kan. The arrows mark the ampli¢ed and hybridizing omp58 and
                                                                              allele by homologous recombination (Fig. 2A). Initial
omp58 : :Kan bands, respectively. Numbers along the sides indicate the        PCR screening results revealed that from 13 KanR trans-
sizes of the DNA standard.                                                    formants, two clones, termed CB2 omp58: :Kan-3 and
                                                                              CB2 omp58 : :Kan-17, underwent a double cross-over
                                                                              event, leading to a single copy of the mutant gene in the
known OMPs of S-layer-secreting ABC exporter. While                           chromosome (Fig. 2B). Due to a single cross-over event,
homology searches with the predicted C. fetus SapF se-                        the remaining clones carried both the wild-type and the
quence lead to BLAST hits with rather moderate scores,                        mutant omp58 allele. Data were con¢rmed by Southern
S. marcescens LipD revealed a low, but signi¢cant homol-                      blot hybridization (Fig. 2C). Both CB2 omp58 : :Kan
ogy to two ORFs, encoding putative protein products of
527 aa and 482 aa in length with predicted molecular
masses of 57.5 kDa and 50 kDa, respectively. Because of
their unknown function, the postulated genes were termed
omp58 and omp50. Sequence identity between Omp58 and
LipD was 18.1% and sequence similarity reached 34.9%.
As outlined in Fig. 1, Omp58 also displayed some homol-
ogy to C. fetus SapF (12.7% identical and 28.4% similar
residues), while the degree of homology to E. coli TolC
(18.6% identical and 32.3% similar residues) was compa-
rable to that of LipD. The identi¢ed omp58 sequence was
identical with rsaF, a sequence earlier deposited in the                      Fig. 3. Comparison of biotin-tagged proteins in membrane fractions of
GenBank (AAF03164). Without any experimental evi-                             CB2 wild-type and mutant strain. The detail on the left provides an en-
dence published, rsaF has been described as the S-layer                       largement of the molecular mass range between 50 and 75 kDa. The ar-
                                                                              row points at the putative Omp58 protein candidate. Lanes: 1, CB2
OMP component of C. crescentus. Because of this rather
                                                                              wild-type ; 2, CB2 omp58 : :Kan-3; 3, CB2 omp58: :Kan-17. Molecular
speculative assumption, we decided to functionally char-                      masses of standard proteins are shown on the right. Prominent biotinyl-
acterize the omp58 gene ¢rst.                                                 ated proteins appear at 81-, 63-, 49-, and 42-kDa apparent molecular
  For cloning of omp58, a genetic fusion protein approach                     mass.

                                                                FEMSLE 10036 16-7-01
M. Reichelt et al. / FEMS Microbiology Letters 201 (2001) 277^283                           281

                                                                              running underneath the OMP63 protein was found in
                                                                              the wild-type strain, but not the mutant strain, suggesting
                                                                              that this band most probably represents the Omp58 pro-
                                                                              tein. Omp58 is among those proteins which become rap-
                                                                              idly biotinylated over a wide range of di¡erent biotin to
                                                                              cell ratios, supporting that Omp58 must be easily accessi-
                                                                              ble on the C. crescentus cell surface once the S-layer has
                                                                              been removed (data not shown).

                                                                              3.3. Omp58 is not required for S-layer secretion

                                                                                 To investigate the e¡ect of Omp58 on S-layer secretion
                                                                              in C. crescentus, di¡erent approaches were applied.
                                                                              Whole-cell immunodot blot analysis of S-layer revealed
                                                                              no striking di¡erences in S-layer amounts on wild-type
                                                                              and mutant strains when identical cell numbers were
                                                                              loaded onto the membrane (Fig. 4B). These results were
Fig. 4. S-layer production and shedding by C. crescentus omp58 mutant         con¢rmed by semi-quantitative Western blot analysis of
strains. A: Western blot analysis of isolated S-layers. The arrow indi-       isolated S-layer (Fig. 4A). Results clearly indicated that,
cates the wild-type S-layer protein band. Numbers mark the molecular          despite functional inactivation of Omp58, S-layer secretion
masses of standard proteins loaded in lane M. B: Whole-cell dot blot
                                                                              and assembly was apparently not a¡ected in strain CB2
analysis. For each strain, 105 (upper panel) and 106 cells (lower panel)
were loaded, respectively. C: S-layer shedding assay. Each strain was         omp58: :Kan, suggesting that Omp58 might not be the
applied in triplicate. The weak shedding-positive phenotype seen for          OMP component of the S-layer secretion system.
CB2 omp58 : :Kan-3 (area III) was not reproducible. Strains: I, CB2
omp58 omp58: :Kan-1; II, CB2 omp58 omp58: :Kan-2; III, CB2
omp58 : :Kan-3; IV, CB2 omp58 : :Kan-17; V, CB2 wild-type; VI, CB2A
                                                                              4. Discussion
RsaA: :PAKPil(450); VII, CB2A. CB2A is an RsaA-negative derivative
of CB2 [5]. CB2A RsaA: :PAKPil(450) carries a plasmid directing the
synthesis of a recombinant shedding-type RsaA protein variant [19].             While the structural genes encoding the ABC transport-
                                                                             er protein and the MFP of the C. crescentus S-layer trans-
                                                                             port machinery have been characterized [6], the pore-form-
clones revealed a similar shift in the omp58 gene, re£ecting                 ing OMP component has not been identi¢ed yet. As for
the 1.2-kb insertion of the KanR cassette. Expectedly, mu-                   many other type I secretion systems, the gene for the OMP
tant clones with an integration of the complete pGEX-                        component is obviously not linked to the remaining genes
Omp58 plasmid (CB2 omp58 omp58: :Kan) showed both,                           of the export apparatus, since sequence analysis of the
the wild-type and mutant gene band.                                          immediate 3P- and 5P-£anking area of the rsa gene cluster
   To demonstrate the absence of the omp58 gene product                      did not reveal any open reading frames for a putative
in the mutant strain, protein pro¢les of crude membrane                      OMP. We therefore decided to identify gene candidates
fractions prepared from S-layer-cured CB2 omp58: :Kan                        for the OMP by their proposed structure-function homol-
and CB2 wild-type were compared by SDS^PAGE analy-                           ogy to known OMPs of type I S-layer secretion systems.
sis (data not shown). Particularly, in the molecular mass                    By homology searches, we found two predicted open read-
range between 50 and 60 kDa no obvious di¡erences in                         ing frames for putative OMPs sequences clearly standing-
protein band pattern could be observed (data not shown),                     out from the remaining BLAST hits by their score. One of
suggesting that Omp58 must be a minor OMP of C. cres-                        the open reading frames was identical to rsaF, which, in
centus. To increase the sensitivity and speci¢city, we there-                our opinion rather speculatively, has been postulated as
fore established a biotinylation approach to detect selec-                   the OMP component of the S-layer secretion system.
tively OMP candidates on the surface of S-layer-free C.                      Cloning of the omp58 sequence ¢nally revealed two amino
crescentus cells. Prior to membrane preparation, surface-                    acid residue di¡erences to rsaF, which could be strain-spe-
exposed proteins were tagged with an activated biotin de-                    ci¢c.
rivative followed by streptavidin-mediated detection of la-                     Since all type I exporter systems of Gram-negative bac-
beled proteins. Fig. 3 compares the biotinylated membrane                    teria characterized so far require an OMP component for
protein pro¢les of strain CB2 omp58 : :Kan with that of the                  secretion [8], we hypothesized that inactivation of the cor-
wild-type strain, indicating the presence of four prominent                  responding gene should result in an S-layer-negative phe-
biotin-tagged protein bands in both strains. According to                    notype. To test this prediction, we constructed an omp58
their apparent molecular mass, these proteins were desig-                    deletion mutant by allelic replacement of the wild-type
nated OMP81, OMP63, OMP49, and OMP42, respec-                                gene on the CB2 chromosome. One step further, we dem-
tively. Besides these bands, an additional band closely                      onstrated that inactivation of the gene results in loss of a

                                                               FEMSLE 10036 16-7-01
282                                M. Reichelt et al. / FEMS Microbiology Letters 201 (2001) 277^283

protein with appropriate molecular mass in the mutant                  candidate we have identi¢ed is possibly the unknown
strain. Both the accessibility of this protein for biotinyla-          OMP component of the S-layer secretion system.
tion in S-layer-cured cells and the presence of this protein
in crude membrane fractions strongly suggested that
Omp58 is indeed localized in the outer membrane. How-                  Acknowledgements
ever, neither by whole-cell immunodot nor Western blot
analysis of released S-layer could we de¢ne any di¡erence                 We gratefully acknowledge J. Smit for the generous gift
in the S-layer coating of the cells and the amount of re-              of the S-layer antibody and C. crescentus strains CB2A
leasable S-layer, respectively, between wild-type and the              and CB2A Rsa: :PAKPil(450). Preliminary C. crescentus
mutant strain. Those data demonstrated that the omp58                  sequence data were obtained from TIGR website at
gene function is not essential for S-layer secretion in C.             http://www.tigr.org. Sequencing of the C. crescentus ge-
crescentus, implementing that Omp58 is most probably not               nome was accomplished with support from DOE Biolog-
the OMP component of the S-layer secretion machinery.                  ical and Environment Research program. This work was
However, we may face a situation where a second func-                  supported by the Deutsche Bundeswehr.
tionally and structurally similar OMP component can
complement Omp58, thus restoring full S-layer secretion
competence in an Omp58-de¢cient C. crescentus strain.
                                                                       References
Although we did not investigate this possibility further,
the second putative omp gene, which we have identi¢ed                   [1] Sara, M. and Sleytr, U.B. (2000) S-Layer proteins. J. Bacteriol. 182,
by its sequence homology, may be a candidate for such                       859^868.
a complementing protein. Despite this possible scenario,                [2] Gilchrist, A., Fisher, J.A. and Smit, J. (1992) Nucleotide sequence
the proposed name omp58 may be more suitable than                           analysis of the gene encoding the Caulobacter crescentus paracrystal-
                                                                            line surface layer protein. Can. J. Microbiol. 38, 193^202.
rsaF, earlier suggested by Awram and Smit [6].
                                                                        [3] Walker, S.G., Karunaratne, D.N., Ravenscroft, N. and Smit, J.
   One step further, one could speculate that Omp58 is the                  (1994) Characterization of mutants of Caulobacter crescentus defec-
OMP component of a second, so far unknown, type I                           tive in surface attachment of the paracrystalline surface layer.
secretion system in C. crescentus. This assumption is based                 J. Bacteriol. 176, 6312^6323.
on the fact that the RsaD protein revealed signi¢cant ho-               [4] Boot, H.J. and Pouwels, P.H. (1996) Expression, secretion and anti-
                                                                            genic variation of bacterial S-layer proteins. Mol. Microbiol. 21,
mologies to several other open reading frames in the C.
                                                                            1117^1123.
crescentus genome, indicating the existence of additional               [5] Nomellini, J.F., Kubcu, S., Sleytr, U.B. and Smit, J. (1997) Factors
ABC transporter systems in this organism (unpublished                       controlling in vitro recrystallization of the Caulobacter crescentus par-
results). We therefore initiated preliminary attempts to                    acrystalline S-layer. J. Bacteriol. 179, 6349^6354.
identify such an exporter system by the secreted protein,               [6] Awram, P. and Smit, J. (1998) The Caulobacter crescentus paracrys-
                                                                            talline S-layer protein is secreted by an ABC transporter (type I)
based on the rational that inactivation of the OMP gene
                                                                            secretion apparatus. J. Bacteriol. 180, 3062^3069.
will lead to an impaired secretion of a soluble protein                 [7] Bingle, W.H., Nomellini, J.F. and Smit, J. (2000) Secretion of the
component. Besides certain enzyme activities, such as for                   Caulobacter crescentus S-layer protein : Further localization of the C-
proteases, we compared secreted protein patterns in cul-                    terminal secretion signal and its use for secretion of recombinant
ture supernatants of wild-type and Omp58-negative strain                    proteins. J. Bacteriol. 182, 3298^3301.
                                                                        [8] Young, J. and Holland, I.B. (1999) ABC transporters : bacterial ex-
(data not shown). However, we failed to identify a se-
                                                                            porters-revisited ¢ve years on. Biochim. Biophys. Acta 1461, 177^200.
creted protein candidate, which is a¡ected by the absence               [9] Binet, R., Lëto¡ë, S., Ghigo, J.M., Delepelaire, P. and Wandersman,
of functional Omp58. Interestingly, omp58 is located with-                  C. (1997) Protein secretion by Gram-negative bacterial ABC export-
in a cluster of lipopolysaccharide (LPS) synthesis genes,                   ers ^ a review. Gene 192, 7^11.
responsible for production of a speci¢c smooth LPS.                    [10] Koronakis, V., Shar¡, A., Koronakis, E., Luisi, B. and Hughes, C.
                                                                            (2000) Crystal structure of the bacterial membrane protein TolC cen-
Along with Ca2‡ , this LPS is involved in the anchoring
                                                                            tral to multidrug e¥ux and protein export. Science 405, 914^919.
of S-layer to the cell surface [3,5], giving rise to the as-           [11] Wandersman, C. (1996) Secretion across the bacterial outer mem-
sumption that Omp58 could have a function in LPS secre-                     brane. In: Escherichia coli and Salmonella typhimurium : Cellular
tion or S-layer anchoring. S-layer shedding analysis how-                   and Molecular Biology (Neidhardt, F.C., Ed.), Vol. 1., pp. 955^
ever, revealed no signi¢cant halo formation and thus, a                     966. American Society for Microbiology, Washington, DC.
                                                                       [12] Kawai, E., Akatsuka, H., Idei, A., Shibatani, T. and Omori, K.
shedding-negative phenotype for both, the wild-type and
                                                                            (1998) Serratia marcescens S-layer protein is secreted extracellularly
mutant strain (Fig. 4C).                                                    via an ATP-binding cassette exporter, the Lip system. Mol. Micro-
   Conclusively, this report provides evidence that omp58,                  biol. 27, 941^952.
formerly termed rsaF, represents an actively transcribed               [13] Thompson, S.A., Shedd, O.L., Ray, K.C., Beins, M.H., Jorgensen,
gene, encoding a non-essential OMP of, so far, unknown                      J.P. and Blaser, M.J. (1998) Campylobacter fetus surface layer pro-
                                                                            teins are transported by a type I secretion system. J. Bacteriol. 180,
function in C. crescentus. Currently, we are pursuing more
                                                                            6450^6458.
systematic approaches to shed some light on the role of                [14] Ausubel, F.M., Brent, R., Kingston, R.E., Moore, D.D., Seidman,
this protein. Besides that, we are applying a similar gene                  J.G., Smith, J.A., and Struhl, K. (1995) Short Protocols in Molecular
deletion approach to prove whether the second omp gene                      Biology. John Wiley and Sons, Inc., New York.

                                                       FEMSLE 10036 16-7-01
M. Reichelt et al. / FEMS Microbiology Letters 201 (2001) 277^283                                          283

[15] Gilchrist, A. and Smit, J. (1991) Transformation of freshwater and         [19] Bingle, W.H., Nomellini, J.F. and Smit, J. (1997) Cell-surface display
     marine caulobacters by electroporation. J. Bacteriol. 173, 921^925.             of a Pseudomonas aeruginosa strain K pilin peptide within the para-
[16] Altschul, S.F., Madden, T.L., Scha«¡er, A.A., Zhang, J., Zhang, Z.,             crystalline S-layer of Caulobacter crescentus. Mol. Microbiol. 26,
     Miller, W. and Lipman, D. (1997) Gapped BLAST and PSI-BLAST:                    277^288.
     a new generation of protein database search programs. Nucleic Acids        [20] Abath, F.G., Almeida, A.M. and Ferreira, L.C. (1992) Identi¢cation
     Res. 25, 3389^3402.                                                             of surface-exposed Yersinia pestis proteins by radio-iodination and
[17] Li, Y., Hess, C., von Specht, B.-U. and Hahn, H.P. (2000) Molecular             biotinylation. J. Med. Microbiol. 37, 420^424.
     analysis of hemolysin-mediated secretion of a human interleukin-6          [21] Niki, H., Imamura, R., Ogura, T. and Hiraga, S. (1990) Nucleotide
     fusion protein of Salmonella typhimurium. FEMS Immunol. Med.                    sequence of the tolC gene of Escherichia coli. Nucleic Acids Res. 18,
     Microbiol. 27, 333^340.                                                         5547.
[18] Walker, S.G., Smith, S.H. and Smit, J. (1992) Isolation and compar-
     ison of the paracrystalline surface layer proteins of freshwater caulo-
     bacters. J. Bacteriol. 174, 1783^1792.

                                                                   FEMSLE 10036 16-7-01
You can also read