Previous Article | Next Article ![]()
Applied and Environmental Microbiology, December 2003, p. 6994-7001, Vol. 69, No. 12
0099-2240/03/$08.00+0 DOI: 10.1128/AEM.69.12.6994-7001.2003
Copyright © 2003, American
Society for
Microbiology. All Rights Reserved.
Alimentary Pharmabiotic Centre,1 Department of Microbiology,2 Department of Food and Nutritional Sciences, National University of Ireland, Cork, Cork, Ireland3
Received 21 May 2003/ Accepted 8 September 2003
| ABSTRACT |
|---|
|
|
|---|
| INTRODUCTION |
|---|
|
|
|---|
The strategy of
constructing random translational fusions between potential
translocation signals and an export-specific reporter protein, designed
to isolate genes encoding exported proteins, was first described for
Escherichia coli, in which alkaline phosphatase (PhoA) was
used as a reporter (18,
29). More recently, this
strategy has been adopted for use in gram-positive bacteria
(27,
37,
38) by using such
reporters as staphylococcal nuclease
(24,
38) and
-amylase
from Bacillus
(19). The reporter in
such cases is translocation competent but is unable to direct its own
export (due to removal of the signal peptide [SP]), while its
activity depends on an extracytoplasmic location. Among a library of
sequences N terminally fused to such a reporter, only those fusions
having an appropriate export signal are directed by the Sec-dependent
secretion machinery to be translocated. In most cases, a prerequisite
for the release of the translocated protein from the membrane (and
subsequent secretion into the medium) is removal of the SP by a signal
peptidase (SPase) (48,
52). Notably, several
integral membrane proteins retain their SPs and diffuse laterally from
the translocase. Other proteins contain several membrane-spanning
domains that are required for insertion into the cytoplasmic
membrane.
At present, four major classes of amino-terminal SPs can be distinguished on the basis of the SPase recognition sequence. The first class is composed of classical SPs, which are present in preproteins that are cleaved by a type I SPase. A separate group of these SPs contains a so-called twin-arginine motif (RR motif), which may direct proteins into a distinct translocation pathway known as the twin-arginine translocation (Tat) pathway (for reviews see references 4, 53, and 55). The classical Sec-type SPs consist of an amino-terminal N domain containing at least one positively charged residue (7, 14), a central hydrophobic core (H region), and a C region with a consensus SPase recognition sequence, A-X-A at positions -3 to -1 relative to the SPase I cleavage site (39, 46, 54). The second major class of SPs is present in prelipoproteins, which are cleaved by the lipoprotein-specific (type II) SPase. Cleavage in this case occurs in front of a cysteine residue (39, 46, 54). The third major class is formed by SPs of prepilin-like proteins, in which the recognition sequence of the prepilin-specific SPase (unlike that of secretory proteins and lipoproteins) is localized between the N and H domains (28, 39). The fourth class of SPs is found in ribosomally synthesized bacteriocins and pheromones that are exported by dedicated ABC transporters (3, 36, 56). These SPs lack an H domain and are removed from the mature protein by a subunit of the ABC transporter.
Despite the assumed biotechnological importance of
surface-located and extracellular proteins in Bifidobacterium,
very few exported proteins have been identified in this genus to date.
There is only one other report of an SP in Bifidobacterium,
and this SP was identified in a probable extracellular
ß-galactosidase from Bifidobacterium bifidum
(30). Approximately 200
proteins with probable Sec-type SPs were recently proposed based on a
genomic sequence analysis of Bifidobacterium longum
(41). In this study, the
broad-host-range plasmid pFUN was utilized to identify exported
proteins in Bifidobacterium breve by a strategy based on
translational fusions with an export-specific reporter protein. The
Staphylococcus aureus secreted nuclease (Nuc) devoid of its
export signal (
SPNuc) was used as a reporter.
Nuclease activity was shown to require an extracellular location in
B. breve.
SPNuc translational fusions were
constructed in which the export signal was provided by inserted B.
breve chromosomal DNA. By using this strategy, seven previously
unknown exported proteins were identified for B. breve
UCC2003. From these results, combined with bioinformatics-based
comparative analyses, it appears that protein translocation in several
Bifidobacterium spp. occurs through a mechanism which is
comparable to the mechanisms previously identified for a large number
of gram-positive bacteria.
| MATERIALS AND METHODS |
|---|
|
|
|---|
(16) was grown in
Luria-Bertani medium at 37°C with agitation
(40). L. lactis
subsp. cremoris MG1363
(13) and its derivatives
were grown on M17 medium supplemented with 1% glucose at
30°C. Plasmids pFUN, pVE8009, and pVE8010
(38) and derivatives of
these plasmids were maintained by antibiotic selection by using
erythromycin (5 µg ml-1 for L.
lactis) or ampicillin (100 µg ml-1 for
E. coli). The E. coli vectors pBluescript (Stratagene
Ltd., Cambridge, United Kingdom) and pCR2.1-TOPO (Invitrogen BV,
Groningen, The Netherlands) were used for cloning
purposes.
DNA manipulations.
Plasmid DNA was obtained from E.
coli by using either an alkaline lysis method
(5) or a QIAprep Spin
plasmid miniprep kit (Qiagen GmbH, Hilden, Germany). L. lactis
plasmid DNA was prepared by using either a modification of the alkaline
lysis method as described previously
(40) or a similarly
modified QIAprep Spin plasmid miniprep kit procedure. The modification
included addition of lysozyme (10 µg ml-1)
after resuspension of the pelleted cells, followed by incubation at
37°C for 30 min. The normal manufacturer's protocol was
used after this step. Large-scale production of total DNA from B.
breve was performed as previously described
(34). Restriction
endonucleases, T4 DNA ligase, and calf intestinal alkaline phosphatase
were purchased from Roche Diagnostics Ltd. (Lewes, East Sussex, United
Kingdom) or New England Biolabs Ltd. (Hitchin, United Kingdom) and were
used as recommended by the manufacturers. Electroporation of plasmid
DNA into E. coli
(40) and L.
lactis (10) was
performed essentially as previously described. Electrotransformation of
B. breve with plasmid DNA was performed as follows.
Mid-logarithmic-phase cells (optical density at 600 nm, 0.5 to 0.6)
were chilled on ice for 20 min, and this was followed by
centrifugation. The cell pellet was washed twice and resuspended in 0.5
M sucrose-1 mM citrate buffer (pH 5.8). The cells were
incubated on ice for 10 min, and this was followed by
electrotransformation with a Bio-Rad Gene Pulser II apparatus under the
following conditions: 25 µF, 200
, and 2.0 kV
cm-1. Modified Rogosa medium
(20) was added to the
cells, and the mixture was incubated anaerobically at 37°C for
2.5 h prior to plating. PCRs were performed by using either
the Taq PCR Master Mix (Qiagen) or the Expand Long Template
PCR system (Roche Diagnostics GmbH, Mannheim, Germany) in accordance
with the manufacturers' instructions. PCRs were performed with an
Omnigene thermal cycler (Hybaid Ltd., Middlesex, United
Kingdom).
Screening of genomic libraries
and Nuc activity assays.
B. breve UCC2003 genomic
libraries were constructed by cloning Sau3A fragments ranging
in size from 0.5 to 1.8 kb into the unique dephosphorylated
BamHI, BclI, or BglII sites of pFUN.
Ligation mixtures were introduced into E. coli DH5
as
the intermediate host by electrotransformation. Libraries were then
established in L. lactis MG1363 by electroporation of plasmid
DNA prepared from approximately 10,000 E. coli transformants.
An insertion rate of approximately 70% was observed. Lactococcal
transformants which were Nuc+ were selected for
further study. A Nuc+ phenotype was detected by
using a chromogenic toluidine blue-DNA-agar overlay
(22,
23). Nuclease activity in
the supernatants of cultures (see below) was determined
spectrophotometrically, essentially as described previously
(17). For all nuclease
assay procedures, strains containing plasmids pVE8009 and pVE8010
(38) were used as
positive and negative controls,
respectively.
Sequence analysis and
bioinformatics.
The pFUN
derivatives of nuclease-positive clones that were obtained were
characterized by PCR amplification of the pFUN insert from plasmid DNA
preparations. Primers were designed for either side of the multiple
cloning site; one primer corresponded to a sequence in the
Rho-independent terminator (5' GTTAGCTCACTCATTAGG
3'), and the other was complementary to the
5' end of the truncated nuclease gene (5'
TGCACTTGCTTCAGGACC 3'). PCR products were purified
by using the CONCERT rapid PCR purification system (GibcoBRL, Paisley,
Scotland) and sequenced. Sequencing was performed by MWG-BIOTECH AG
(Ebersberg, Germany). Sequence data assembly and analyses were
performed by using the DNASTAR software (version 5.05; DNASTAR,
Madison, Wis.). Database searches were performed by using nonredundant
sequences at the National Center for Biotechnology Information web site
(http://www.ncbi.nlm.nih.gov)
and the tBlastN, tBlastX, and BlastP programs
(1,
2). Sequence alignments
were constructed by using the Clustal method of the MEGALIGN program of
the DNASTAR software package. Functional domains in deduced
proteins were identified by using the SMART database
(42,
43;
http://smart.embl-heidelberg.de).
Classical SPs and their cleavage sites were predicted by using the
Signal Peptide Prediction program with gram-positive data
(32,
33;
www.cbs.dtu.dk/services/SignalP).
Likely transmembrane domains were determined by using the HMMTOP server
(49;
www.enzim.hu/hmmtop)
and the DAS (Dense Alignment Surface) transmembrane prediction server
(8;
www.sbc.su.se/~miklos/DAS/maindas.html).
Location
studies, SDS-PAGE, Western blotting, zymograms, and quantitative
nuclease assays.
The
localization of all
SPNuc fusion proteins was
examined in whole cells, culture supernatants, and membrane vesicles of
mid-exponential-phase B. breve cells (optical density at 600
nm, 0.5 to 0.6). Whole cells and supernatant fractions were isolated
essentially as described previously
(38), with modifications
for Bifidobacterium. Briefly, following harvesting and washing
of the pellet, cells were resuspended in Tris-EDTA containing lysozyme
(final concentration, 30 mg ml-1) and incubated at
37°C for 30 min. The cells were then lysed with sodium
dodecyl sulfate (SDS) (final concentration, 4%) for 15 min on
ice. Membrane vesicles were isolated essentially as described
previously (35). The
procedure was slightly modified for use with Bifidobacterium
by using a variation of the lysozyme step (30 mg
ml-1 for 30 min at 37°C). Equal volumes of
loading buffer (60 mM Tris-HCl, 2% SDS, 10% glycerol,
0.01% bromophenol blue, 200 mM dithiothreitol) were added to
cell pellet, supernatant, or membrane fractions, and 30 µl was
loaded onto an SDS-polyacrylamide gel electrophoresis (PAGE) gel prior
to electrophoresis.
For immunoblot experiments, proteins separated by SDS-PAGE were blotted electrophoretically onto nitrocellulose filters (Schleicher and Schuell, Dassel, Germany) as described previously (25). Custom-made polyclonal anti-Nuc rabbit antibodies (raised against the peptide EFDKGQRTDKYGRG) were obtained from ProSci Inc. (Poway, Calif.) and used according to the instructions of the manufacturer. Immunodetection was performed with peroxidase-conjugated goat anti-rabbit immunoglobulins (Dako A/S, Glostrup,Denmark) and an enhanced chemiluminescence kit (Amersham Biosciences UK Limited, Little Chalfont, Buckinghamshire, United Kingdom) as recommended by the suppliers. Nuclease activity was evaluated on zymograms of SDS-PAGE gels after removal of SDS, as described previously (26). Samples of concentrated supernatant fractions of both B. breve and L. lactis were compared by using a spectrophotometric assay based on the release of acid-soluble oligonucleotides following nuclease digestion of DNA (17).
Nucleotide
sequence accession numbers.
The sequences of the pFUN inserts
encoding the polypeptides fused to
SPNuc have been
deposited in the GenBank database under the following accession
numbers:
AY297716
(Sec1),
AY297717
(Sec2),
AY297718
(Sec3),
AY297719
(Tmp1),
AY297720
(Tmp2),
AY297721
(Tmp3), and
AY297722
(Tmp4).
| RESULTS |
|---|
|
|
|---|
|
nuc (nuclease gene lacking the export signal). For
both of these plasmids, fusion expression is driven by the
usp45 promoter and its original translational start site,
followed by the intact Usp45 SP sequence in the case of pVE8009 or a
deleted and therefore inactive version of this sequence in the case of
pVE8010.
All vectors harboring
SPNuc fusions
and displaying a Nuc+ phenotype were introduced into
B. breve UCC2003 for location studies and activity assays.
Successful transformation of the appropriate plasmid DNA into B.
breve was confirmed by PCR and sequence analysis. All clones
displaying a Nuc+ phenotype in L. lactis
exhibited a similar Nuc+ phenotype in B.
breve. Ten nonredundant fusions were identified among the clones
sequenced. All seven fusions which displayed a strong
Nuc+ phenotype contained a putative export signal
(represented by a predicted SP sequence or putative transmembrane
domains) (see below). Three fusions which had a delayed
Nuc+ phenotype did not contain any recognizable
export signal and probably corresponded to cytoplasmic proteins (data
not shown). The phenotypes of these fusions may have been due to cell
lysis or leakage, as described previously
(38), and these proteins
are not discussed further here.
Sec
proteins.
Three B.
breve polypeptides isolated as active
SPNuc
fusions, designated Sec (for putative secreted protein), contained a
predicted N-terminal classical SP (i.e., an SP composed of an
N-terminal positively charged region, a central hydrophobic core, and a
C-terminal cleavage region) (Table
2). Sequence analysis with the sequenced genome of B. breve
UCC2003 (Leahy et al., unpublished) revealed the full protein-encoding
open reading frame in each case (Table
3). Sec1 is predicted to contain a classical SP with a single transmembrane
region. The portion of this protein fused to the nuclease comprises the
first 77 amino acids of a putative 602-amino-acid protein. This protein
is significantly similar to the permease component of an ABC-type
transport system, and clear homologues of the gene are found in B.
longum NCC2705 (accession no.
NP_695398)
and DJO10A
(ZP_00121339)
(Table 3). Homologues of
the gene are also found in Enterobacter (NP_815308)
and Lactobacillus (NP_786841) (data not shown), both
of which are present in the gastrointestinal tract, indicating that the
gene may be involved in a function specific to bacterial inhabitants of
the gastrointestinal tract.
|
|
Sec3 is predicted to contain a classical Sec-type SP. The fused portion of this protein is homologous to the N-terminal regions of approximately 530-amino-acid hypothetical proteins from B. longum NCC2705 (accession no. NP_695526) and DJO10A (ZP_00121034). Sec3 also displays significant similarity to a hypothetical protein from Thermobifida fusca (ZP_00057200) and a putative M protein from Streptomyces (NP_824034).
Tmp
proteins.
Four B.
breve UCC2003 polypeptides (Table
2) containing at least one
predicted transmembrane domain were identified as active
SPNuc fusions and were designated Tmp (for putative
transmembrane protein). Sequence analysis of B. breve revealed
the complete protein-encoding open reading frame in each case (Table
3). The fused portion of
Tmp1 corresponds to the N-terminal region of a hypothetical protein and
is predicted to have a C-out topology, which is expected for
transmembrane proteins (Nuc activity of a fusion indicates that the
SPNuc domain is exported). Proteins similar to Tmp1
have been found only in B. longum NCC2705 (accession no.
NP_696268)
and DJO10A
(ZP_00120937)
(Table 3), although no
function has been assigned to these proteins.
The portion of Tmp2
fused to
SPNuc contains a single transmembrane
domain and has a C-out topology. Tmp2 is similar to hypothetical
proteins from B. longum NCC2705 (accession no.
NP_695850)
and DJO10A
(ZP_00121150)
and to a lesser extent to a protein from Clostridium
(NP_782940)
and a hypothetical protein from Lactobacillus
(ZP_00047154).
No conserved domains were detected, and consequently no putative
function could be assigned to this protein.
Tmp3 contains a single transmembrane domain. B. longum NCC2705 contains a homologue of this protein (accession no. NP_695675); however, Tmp3 appears to be unique to these two bacteria and does not exhibit similarity to any other protein identified to date (Table 2). It is notable that no homologue of Tmp3 was identified in B. longum DJO10A; however, it is possible that this putative protein has not been identified yet, as the available genome sequence is incomplete. No conserved domains were detected; thus, it is not possible to attribute a (putative) function to this protein.
Tmp4 contains two transmembrane domains. This protein exhibits highly significant similarity to DNA gyrase B subunits from B. longum NCC2705 (accession no. NP_695821) and DJO10A (ZP_00121174). Homologues of Tmp4 are found in a wide variety of bacterial species and are known to be involved in DNA replication and repair.
Locations and activities of fusion
proteins.
The locations of
all exported
SPNuc fusions were examined (see
Materials and Methods), and the results are shown in Fig.
1. The mature NucA form was found to be mainly associated with the pellet
fraction in the case of pVE8010 and was almost exclusively located in
the supernatant fraction in the case of pVE8009 (Fig.
1).
Sec1-
SPNuc, Sec2-
SPNuc, and
Sec3-
SPNuc were found almost exclusively in the
medium, thus confirming that these fusion proteins are secreted (Fig.
1). Full-size
Tmp4-
SPNuc was located mainly in the membrane
vesicle fraction (Fig. 1).
Similar results were obtained for Tmp1-
SPNuc,
Tmp2-
SPNuc, and Tmp3-
SPNuc
(data not shown). Zymograms performed for SDS-PAGE gels by using
concentrated supernatant samples of B. breve UCC2003
derivatives harboring Sec1-
SPNuc,
Sec2-
SPNuc, and Sec3-
SPNuc
demonstrated the enzymatic activity of the full-length fusion forms
(data not shown).
|
|
| DISCUSSION |
|---|
|
|
|---|
This is one of the first reports of identification of exported proteins in the genus Bifidobacterium. In fact, there has been only one other report of an SP in Bifidobacterium, which was identified in a probable extracellular ß-galactosidase from B. bifidum (30). The nuclease gene, which was first isolated from S. aureus, has previously been used to identify exported proteins in a number of gram-positive hosts, including Lactococcus (38) and Corynebacterium (26). It was demonstrated in this study that nuc (in the vector pFUN [38]) is a reliable reporter for export studies in Bifidobacterium.
Three Sec-dependent SPs were identified in this study, and they have SP cleavage regions comparable to those previously identified for other gram-positive bacteria (39, 46, 54). The SPs appear to be longer in Bifidobacterium than in other gram-positive bacteria, comprising between 35 and 84 amino acids (this study), compared with an average length of 15 to 30 amino acids (51). This may be due to differences in the cell wall structure of Bifidobacterium; however, a preliminary investigation of the average SP length in B. longum NCC2705 yielded a length of 44 amino acids (unpublished data). The fusion proteins are also consistently longer than the proteins described in similar studies (in which other reporter systems were used). This result is consistent with results previously obtained with the pFUN vector (38). The -3 to -1 cleavage regions for the three SPs (Table 1) also appear to be similar to the consensus A-X-A motif. Also noteworthy are the two helix-breaking proline residues located at positions -5 and -7 in Sec2. It has previously been reported that 50% of SPs in Bacillus subtilis contain a helix-breaking residue (proline or glycine) at a position between positions -7 and -4 relative to the predicted processing site for SPase I (47).
Sec1 is a
probable permease component of an ABC transport system, a function that
is consistent with its membrane location; however, the fact that this
protein is subject to secretion is inconsistent with its assumed
permease function. It is possible that fortuitous cloning of this
fragment of the permease resulted in secretion of the nuclease as this
moiety functions as an artificial SP. Alternatively, the complete
protein may harbor a C-terminal anchor motif, which would target the
permease for retention in the cell wall or membrane. As only the
N-terminal region of this (putative) protein was cloned in this study,
it is possible that the permease component was erroneously released
from the membrane and secreted into the medium. Sec2 harbors an acid
phosphatase domain and also displays homology to a serine protease.
Sec3 has similarity with a cell surface-associated M protein and may be
involved in adhesion of the bifidobacterial strain to host epithelial
cells. Four of the SPs identified are putative transmembrane proteins
harboring one or more transmembrane domains. Tmp3 harbors two predicted
stretches of hydrophobic acids at either end of the fusion, which may
facilitate insertion of the protein into the membrane in a hairpin-like
manner. The position of a putative DNA gyrase (Tmp4) at the cell
surface is both unexpected and inconsistent with the function of this
protein. It is more likely that the transmembrane domains (fused to
SPNuc) caused improper insertion of the nuclease
into the cell membrane, thus translocating the nuclease to the cell
surface.
Tmp1 and Tmp3 appear to be unique to the three Bifidobacterium strains examined, as no homologues in other bacterial species were identified. Tmp3 was notably absent from B. longum DJO10A, although this may have been because the entire genome sequence was not available. Homologues of Tmp2 have been found in Clostridium and Lactobacillus, in addition to the two B. longum strains investigated. No functions could be assigned to these three putative exported proteins due to a lack of conserved domains or homology with proteins having known functions.
Western blotting with anti-Nuc polyclonal antibodies confirmed that the three Sec fusion proteins were exported to the culture supernatant, and the activities of the full-length mature forms were demonstrated by using zymograms. The nuclease activities of the secreted proteins were demonstrated to be substantially higher (in the supernatant fractions of both L. lactis and B. breve) than the nuclease activities of the other clones and were actually higher than the activity of the positive control (see above). Each of the four transmembrane proteins was shown by Western blotting to be mainly associated with the membrane fraction. Taken together, the results described above indicate that the polypeptides that are signals for protein export operate at comparable levels in L. lactis and B. breve.
There are a number of limitations when the screening strategy is employed in L. lactis. Screening for exported proteins of bifidobacterial origin in Lactococcus may result in a bias in the SPs identified, due in part to differences in both the G+C content and promoter motifs. Thus, there are likely to be (as-yet-unidentified) SPs in B. breve which do not target the nuclease for export in a heterologous host (namely, L. lactis). Although it would be preferable to screen for exported proteins directly in B. breve, this is not possible at present due to low electrotransformation frequencies (102 to 103 transformants per µg of DNA for this strain [unpublished results]). There is also a paucity of tools available for genetic manipulation of bifidobacterial species. It is noteworthy, for the purposes of future molecular cloning, that the pAMß1 origin (45) of pFUN, which replicates in various gram-positive bacteria, is also active in Bifidobacterium. We found that the low copy number of this replicon in B. breve is roughly comparable to the copy number in L. lactis (data not shown).
It can be concluded from this study that proteins exported by Bifidobacterium contain either an SP similar to that of other gram-positive bacteria or a number of transmembrane region(s). Also, a number of these signals are heterologous, in that they function in both Lactococcus and Bifidobacterium. It is likely that a number of these exported proteins are important in the interactions of B. breve with its host environment. Future work should focus on functional analysis of the identified exported proteins and should result in a greater understanding of Bifidobacterium protein export. In addition, optimization of such mechanisms may greatly enhance the probiotic properties of members of this genus.
| ACKNOWLEDGMENTS |
|---|
This work was financially supported by Enterprise Ireland (grant BR/1998/202), by the Higher Education Authority Programme for Research in Third Level Institutions, and by the Science Foundation Ireland Centre for Science Engineering and Technology.
| FOOTNOTES |
|---|
| REFERENCES |
|---|
|
|
|---|
-amylase probe vectors.Gene
118:21-30.[CrossRef][Medline]
This article has been cited by other articles:
| |||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| J. Bacteriol. | Microbiol. Mol. Biol. Rev. | Eukaryot. Cell | All ASM Journals |
|---|