Previous Article | Next Article ![]()
Journal of Bacteriology, July 2005, p. 5013-5018, Vol. 187, No. 14
0021-9193/05/$08.00+0 doi:10.1128/JB.187.14.5013-5018.2005
Copyright © 2005, American Society for Microbiology. All Rights Reserved.
Biological Sciences Research Group, Centro de Engenharia Biológica e Química, Instituto Superior Técnico, Av. Rovisco Pais, 1049-001 Lisbon, Portugal
Received 24 January 2005/ Accepted 15 April 2005
|
|
|---|
|
|
|---|
Biochemical characterization of BceB. A 1,374-bp PCR product bearing bceB was amplified by using primers based on the genome sequence of B. cenocepacia J2315 (primer sequences available upon request) and cloned into the SalI and HindIII sites of pWH844 vector (16). The resulting recombinant plasmid, pBceB432, carries the bceB gene coding sequence, preceded by a sequence coding for six histidine residues (His6). Escherichia coli C43(DE3) (Avidis) was transformed with pBceB432 or pWH844. Cultures were incubated at 25°C in Lennox broth with ampicillin (150 mg/liter) until the culture reached an A640 of 0.6 ± 0.1. At this point, bceB transcription was induced by the addition of 0.1 mM IPTG (isopropyl-ß-D-thiogalactoside), followed by an additional period of 6 h of cultivation. Expression from plasmid pBceB432 resulted in a 48-kDa protein found in membrane extracts. This protein is smaller than the expected 52-kDa His6-BceB fusion protein. It was thus hypothesized that the BceB protein might have a cleavable signal peptide at the N-terminal site. In agreement with this prediction, the His6 Western blot analysis presented no signal, due to the His6 tag cleavage. The N-terminal amino acid sequence of BceB was then examined by using different signal peptide prediction programs. Only the SignalP program (http://www.cbs.dtu.dk) allowed the prediction of a signal peptide cleavage site, between positions 21 and 22 of the amino acid sequence. The BceB protein exhibits all the features needed for having a cleavable signal peptide (23), namely, a short N region composed of basic residues (MLSVLAR) followed by a region composed of hydrophobic residues (VIDIAMVVTG) plus a neutral but polar region (ALIAAA), with a maximal cleavage site probability between residues at position 21 and 22.
To assess the predicted BceB activity of undecaprenyl-phosphate glucosyl-1-phosphate transferase (UndPGlcPT), membrane extracts from IPTG-induced E. coli C43(DE3) cells carrying pBceB432 were used as a source of both BceB and membrane acceptor substrate (isoprenoid lipid) as described before (22). Enzyme assays were carried out in a final volume of 100 µl containing the following: 50 µg of membrane fraction, 50 mM Tris-HCl (pH 8), 5 mM EDTA, 1 mM phenylmethylsulfonyl fluoride, 1 mM dithiothreitol, and 10 mM MgCl2. The reaction was started by the addition of 0.10 µCi of the substrate UDP-[14C]glucose (UDP-[14C]Glc) (NEN Life Science Products). The radiolabeled sugars, covalently linked to the membrane acceptor, were extracted in the lipid fraction and measured as previously described (22). The extracts expressing the BceB protein [C43(pBceB432)] incorporated approximately 10-fold more radiolabeled sugars than extracts prepared from C43(pWH844) control cells (Fig. 1). UndPGlcPT was assayed at different temperatures, which revealed that BceB-mediated sugar incorporation was maximal at the lowest temperature assayed (10°C), decreasing with the increase of temperature. Such a low temperature is well below the optimal temperature for maximal productivity of cepacian biosynthesis in B. cepacia (15). A similar observation was reported before for UndPGPTs from other bacterial systems (10, 17). Replacement of UDP-[14C]Glc by UDP-[14C]glucuronic acid or UDP-[14C]galactose resulted in a significant reduction of the level of sugar incorporation into the lipid fraction (results not shown).
![]() View larger version (15K): [in a new window] |
FIG. 1. Incorporation of [14C]glucose ([14C]Gluc) at 10°C, catalyzed by UndPGPT activity in membrane extracts enriched in BceB, prepared from E. coli C43(DE3) cells carrying the void cloning vector (pWH844) ( ) or this vector with the B. cenocepacia IST432 bceB gene (pBceB432) ( ).
|
-glucosidase (G5003; Sigma) or with ß-glucosidase (49290; Fluka) according to the manufacturer's instructions. Procedures with controls lacking the enzyme were performed. After incubation, lipids were extracted as described above, and the activities of the glucosidases were assessed based on the levels of radioactivity detected in the aqueous fractions due to the enzyme-mediated release of the [14C]Glc from the lipids to the aqueous fractions. Only the aqueous fraction obtained from
-glucosidase digestion exhibited a significant level of radioactivity, and the radiolabeled sugar released comigrated with glucose (results not shown). These results indicate that, in the product of the reaction catalyzed by BceB, glucose is linked through an
-linkage similar to the linkage exhibited in the UDP-[14C]Glc substrate. Taken together, these results reinforce the concept that BceB is an UndPGPT that does not catalyze the formation of a glycosidic linkage but does catalyze the transfer of glucose-1-phosphate and the formation of a pyrophosphate linkage. Amino acid sequence analysis of BceB compared with analysis of priming GTs from other microbial sources. BceB exhibits high similarity to a number of established priming GTs from different species (Fig. 2). Two conserved regions from the Protein Families (Pfam) database were found in BceB: the most conserved is present in some bacterial priming GTs (Pfam accession number PF02397; residues 283 through 457), and the other region is present in diverse bacterial polysaccharide biosynthetic proteins, including priming GTs and epimerases (Pfam accession number PF02719; residues 118 through 235).
![]() View larger version (77K): [in a new window] |
FIG. 2. Comparison of the conserved PF02397 C-terminal region of BceB from B. cenocepacia IST432 (BceB|Bcen) with those from proteins assigned as UndPGlcPT (UDP-glucose specific), namely, Xanthomonas campestris GumD (GumD|Xcam) (Pfam accession number AAA86372), Methylobacillus sp. 12S EpsB (EpsB|M12S) (accession number BAC41337), Rhizobium leguminosarum PssA (PssA|Rleg) (accession number T10535), Sphingomonas sp. S88 SpsB (SpsB|SS88) (accession number AAC44071), Gluconacetobacter xylinus AceA (AceA|Axyl) (accession number T44788), Lactococcus lactis B40 EpsD (EpsD|LlacB40) (accession number NP_053030), Streptococcus pneumoniae Cps14E (Cps14E|Spne) (accession number CAA59777), and Streptococcus salivarius CpsE (CpsE|Ssal) (accession number CAC18355), and as UndPGalPT (UDP-galactose specific), namely, Streptococcus thermophilus EpsE (EpsE|Sthe) (accession number AAC44012), L. lactis B35 EpsD (EpsD|LlacB35) (accession number AAD22526), Sinorhizobium meliloti ExoY (ExoY|Smel) (accession number Q02731), Salmonella enterica serovar Typhimurium RfbP (RFBP|Styp) (accession number P26406), Erwinia amylovora AmsG (AMSG|Eamy) (accession number Q46628), and Escherichia coli WbaP (WbaP|Ecol) (accession number AAD21565). Partial multiple alignment was performed using CLUSTAL W (18) and refined manually from the results obtained by using secondary structure prediction methods. Invariant and conserved amino acids are indicated in white on a black background. The invariant tyrosine residue exclusive of UndPGalPTs (17) is indicated in a grey background. The putative DXD motif typical of general GTs (13, 21) is indicated inside a box. Asterisks signify the BceB Asp residues replaced in the present work by site-directed mutagenesis. BceB secondary structure elements, predicted by molecular biology programs, are indicated above (arrows correspond to ß sheets and dot lines to helices). Identity (similarity) values [I(S)] in reference to the BceB amino acid sequence, obtained by using BLAST searches, are indicated on the right.
|
The alignment of the most conserved region, PF02397, of the selected UndPGPT sequences showed the presence of several invariant residues, in particular, three aspartic acid (Asp) residues (Fig. 2). The invariant residue Asp339 in BceB may correspond to the second Asp residue within a DXD motif (Fig. 2), which is the most widespread short-sequence motif identified among general GTs (4) and is involved in bridging the substrate donor through a divalent metal ion (20).
According to secondary structure prediction programs (data not shown), the amino acid sequence comprising the second conserved region, PF02719, of BceB has a putative parallel
/ß/
structure. This structure characterizes the Rossmann fold-like domains, typical of the nucleotide binding site (14). The PF02719 region is also found in diverse bacterial polysaccharide biosynthetic proteins, including several sugar nucleotide epimerases, which share with UndPGPTs the same activated sugar substrate. Together, these data suggest that the PF02719 region is involved in protein interaction with the substrate donor, in particular, with its nucleotide portion. Nevertheless, the importance of the PF02719 region is still ambiguous, since it is not essential for the full UndPGPT activity of the two-domain Salmonella enterica serovar Typhimurium RfbP protein (24) or for UndPGPT proteins exhibiting just one domain and therefore lacking this region.
Effect on BceB activity of the conserved aspartic acid substitutions. Functional and structural information obtained from several enzymes belonging to the phosphotransferase family suggests a mechanism of phosphoryl transfer that involves an invariant aspartic acid residue, located on the catalytic cleft of the enzymes, that promotes substrate binding through a divalent metal ion, such as Mg2+ or Mn2+ (2, 8). The in vitro activity of BceB was found to be dependent on the presence of Mg2+ in the reaction mixture, decreasing as the result of enzyme reaction mixture supplementation with EDTA (data not shown).
To examine the role in BceB activities of three conserved Asp residues (Asp339, Asp355, and Asp439) in the PF02397 region (Fig. 2), these residues were replaced by alanine (Ala), glutamate (Glu), and asparagine (Asn) by use of the Stratagene QuikChange site-directed mutagenesis kit as recommended by the supplier and with plasmid pBceB432 as the template. The possibility that these amino acid replacements could affect either protein stability or targeting to the membrane was ruled out by examining the levels of expression of the various constructs in membrane preparations. Indeed, all the BceB mutated proteins exhibited levels of expression similar to that of the parental BceB protein, which was encoded by pBceB432 (data not shown).
When Asp339 and Asp355 were mutated to neutral residues Ala and Asn, the resulting incorporation activities of membranes enriched with these mutated proteins were dramatically reduced down to the basal levels exhibited by the host cells (Fig. 3). The substitution of these residues with the acidic residue Glu also abrogated BceB activity (Fig. 3), suggesting that both Asp339 and Asp355 residues are crucial for BceB catalytic activity. As shown below, the conserved Asp439 residue is located at the beginning of one of the BceB transmembrane segments (Fig. 4), which may justify the fact that Asp439Ala and Asp439Glu mutated proteins displayed reduced although not null UndPGlcPT activities, with 90% and 85% reductions, respectively (Fig. 3). Replacement of the Asp439 residue by Asn, its corresponding amide residue, results in a less significant reduction of enzyme activity (60%), suggesting that although Asp439 is not an essential residue for BceB function, its mutation affects its activity eventually by altering the conformation of the protein.
![]() View larger version (27K): [in a new window] |
FIG. 3. Incorporation of [14C]Glc into lipid acceptor fractions, catalyzed by UndPGPT activity in membrane extracts enriched with BceB or BceB derivatives. BceB was mutated by site-directed mutagenesis at the Asp residues marked with asterisks in Fig. 2. The incorporation activities of E. coli membranes enriched with BceB or BceB derivatives (white columns) and the percentages of BceB activity retained (calculated after subtracting the basal activities exhibited by membranes of the host cell harboring the cloning vector) relative to the native enzyme (considered 100%) (grey columns) are shown. Values with standard deviations are the averages of results from three enzyme assays.
|
![]() View larger version (28K): [in a new window] |
FIG. 4. Proposed topology of the BceB protein and location and activity of the indicated reporter fusions. (A) Topological model drawing was performed manually based on transmembrane segment predictions and experimental data and refined according to the BceB hydrophobicity profile and conformational parameters for beta turns (ProtScale tools). The first and last residues of the predicted transmembrane segments of BceB (I through VII) are indicated. White ellipses indicate the locations of the reporter fusions; the last amino acid prior to the fusion junction with LacZ or PhoA reporter proteins and the respective position determined from the amino terminus of BceB are indicated. Black ellipses signify the BceB Asp residues replaced by site-directed mutagenesis. The locations of the conserved regions are underlined. (B) Table representation of the reporter fusion activities. The ß-galactosidase activity in strain DH5 containing plasmid pLKC480 or each of the five derived bceB-lacZ fusions was determined. The alkaline phosphatase activity in strain DH5 containing plasmid pBAF or each of the five derived bceB-phoA fusions was measured.
|
erikw/toppred2/], and TMpred [http://www.ch.embnet.org/]) were also used and all predicted the existence of membrane-spanning peptide segments I, II, III, IV, and VI in BceB (Fig. 4A). However, the nucleotides defining their limiting borders were slightly different, depending on the program used. Nevertheless, the existence of segment V was predicted only by the HMMTOP program, while the TopPred program considered the presence of this segment uncertain (score, 0.629). With the exception of TMpred, all the programs used predicted the existence of segment VII (Fig. 4A).
To elucidate the number of transmembrane domains presumably present in BceB and the localization of the two major nonmembrane loops, we have created a series of BceB-LacZ and BceB-PhoA fusions in specific selected positions along the entire length of the protein. The reporter fusion sites chosen were located approximately in the middle of the predicted nonmembrane segments of BceB (Fig. 4A). A series of 3' truncated sections of the bceB nucleotide coding sequence were PCR amplified as EcoRI-SalI fragments that were ligated into vector pLKC480 (19) to create in-frame LacZ fusions or into vector pBAF (9) to create in-frame PhoA fusions. E. coli DH5
cells transformed with the fusion constructs or the cloning vectors were grown in Lennox broth containing 1 mM IPTG and 150 mg/liter of ampicillin, harvested at the exponential phase, and permeabilized as described previously (9). ß-Galactosidase activities of the BceB-LacZ protein fusions and alkaline phosphatase activities of the BceB-PhoA protein fusions were determined based on the methods of Miller (11) and Brickman and Beckwith (5), respectively, and values are shown in Fig. 4B. Only the fusions prepared at Val179 and Val370 showed detectable ß-galactosidase activity, indicating that the fused proteins promote exposure of LacZ to the cytoplasm, whereas the Asp94 and Lys266 fusions gave no detectable ß-galactosidase activity, suggesting that the fused LacZ protein is located in the periplasm. These results contrast with those obtained with the PhoA fusions at the same residues: while Val179 and Val370 were devoid of phosphatase activity, Asp94 and Lys266 fusions gave rise to detectable phosphatase activity, confirming that the fused proteins promoted the exposure of the PhoA protein to the cytoplasm and to the periplasm, respectively. Western immunoblotting analyses using antibodies against LacZ and PhoA (MAb3468 and MAb1012, respectively; Chemicon, Bethesda, Md.) confirmed the presence of these fusion proteins with the expected size (data not shown). Control assays prepared with the void pLKC480 or pBAF vector exhibited no enzyme activity (Fig. 4B).
Consistent with the predicted presence of a cleavable signal peptide between BceB residues at positions 21 and 22, thus located before the Arg34 fusions (Fig. 4A), these fusion proteins showed no detectable ß-galactosidase or phosphatase activity (Fig. 4B), and no immunoreactive signals against LacZ or PhoA antibodies were detected (data not shown).
From the analysis of results obtained with the BceB-PhoA and BceB-LacZ reporter fusion proteins and the indications given by topology prediction programs, we propose that BceB spans the membrane at least six times (segments I through VI [Fig. 4A]) and that it has two major nonmembrane loops located in the cytoplasm, approximately from position 118 to 240 and from position 292 to 439. These cytoplasmic loops match the two highly conserved regions found in UndPGPTs PF02719 and PF02397, which are likely to play a critical role in BceB activity. Since the assembly of the exopolysaccharide repeating unit is performed in the cytoplasm, this topology is consistent with the implication of the conserved regions, in particular, the invariant Asp339 and Asp355 residues (Fig. 4A), in BceB-mediated catalysis.
The topologies of priming N-acetylhexosaminyltransferases, such as E. coli WecA and MraY, were already proposed and tested (1, 3), but the resulting topological models are different from the one proposed here for BceB. WecA and MraY share a common topological model possessing 10 transmembrane segments, 5 cytoplasmic domains, and 6 periplasmic domains, including the N- and C-terminal ends. Priming GTs recognizing UDP-N-acetylhexosamines exhibit a very poor homology to BceB in spite of the fact that they charge the same membrane lipid. Moreover, the conserved residues found to be essential to WecA enzyme activity (1) were not detected in any of the UndPGPTs examined here, and it is likely that N-acetylhexosaminyltransferases form a family of bacterial priming GTs distinct from the proteins studied in this work. Considering the strong similarity and the presence of conserved regions PF02397 and PF02719 among the UndPGlcPTs and UndPGalPT proteins present in databases, we believe that the topology model described here for the first time for an UndPGlcPT protein may be extended to other two-domain UndPGPTs.
Plasmids pBAF and pLKC480 were kindly provided by H. L. T. Mobley, Department of Microbiology and Immunology, University of Maryland.
|
|
|---|
This article has been cited by other articles:
| |||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
Copyright © 2009 by the American Society for Microbiology. For an alternate route to Journals.ASM.org, visit: http://intl-journals.asm.org | More Info»