![[Feedback]](/icons/banner/feedbackACT.gif){kind=icon}
![[For Subscribers]](/icons/banner/subscriberACT.gif){kind=icon}
![[Archive]](/icons/banner/archiveACT.gif){kind=icon}
![[Search]](/icons/banner/searchACT.gif){kind=icon}
![[Contents]](/icons/banner/tocACT.gif){kind=icon}
Young Jin Lee,3,
Young Moo Lee,3,¶
Anup Madan,4,|| and
Michael J. Calcutt2
Department of Molecular Microbiology and Immunology,1 Department of Veterinary Pathobiology, University of Missouri-Columbia, Columbia, Missouri 65212,2 Molecular Structure Facility, University of California Davis, Davis, California 95616,3 Institute for Systems Biology, Seattle, Washington 98103-89044
Received 17 February 2006/ Accepted 21 April 2006
 |
ABSTRACT |
|---|
 Top Abstract IntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
0.6- to 1.4-Mb) mycoplasmal genomes suggest their key role in the adaptation and survival of these wall-less monoderms. Diversity through variable genes is less clearly established among phylogenetically similar mycoplasmas, such as the Mycoplasma mycoides cluster of ruminant pathogens, which vary widely in host range and pathobiology. Using (i) genome sequences from two members of this clade, Mycoplasma capricolum subsp. capricolum and M. mycoides subsp. mycoides small colony biotype (SC), (ii) antibodies to specific peptide determinants of predicted M. capricolum subsp. capricolum gene products, and (iii) analysis of the membrane-associated proteome of M. capricolum subsp. capricolum, a novel set of six genes (vmcA to vmcF) expressing distinct Vmc (variable M. capricolum subsp. capricolum) lipoproteins is demonstrated. These occur at two separate loci in the M. capricolum subsp. capricolum genome, which shares striking overall similarity and gene synteny with the M. mycoides subsp. mycoides SC genome. Collectively, Vmc expression is noncoordinate and combinatorial, subject to a single-unit insertion/deletion in a 5' flanking dinucleotide repeat that governs expression of each vmc gene. All vmc genes share modular regions affecting expression and membrane translocation. In contrast, vmcA to vmcD genes at one locus express surface proteins with highly structured size-variable repeating domains, whereas vmcE to vmcF genes express products with short repeats devoid of predicted structure. These genes confer a distinctive, dynamic surface architecture that may represent adaptive differences within this important group of pathogens as well as exploitable diagnostic targets. |
INTRODUCTION |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
Mycoplasma genomes have been reported to contain from 482 to 1,037 protein-coding genes (18, 43). Despite this limited capacity, the genomes of several mycoplasma species contain distinctive sets of genes that encode diverse surface lipoproteins. Products of these genes are anchored in the single plasma membrane of the mycoplasma and often are among the most abundant membrane proteins. These lipoproteins are commonly subject to high-frequency phase-variable expression, mediated by a variety of mechanisms causing localized reversible mutations that affect transcription or the translational reading frame of individual genes (35, 61). The diversity of the coding sequences comprising these gene "families" is remarkable, whether compared within a single species or across species from widely divergent clades (5, 35, 61). Nevertheless, unifying features shared among members of a particular family are commonly found, for example, (i) highly conserved coding sequences that present a shared segment of otherwise variable surface proteins, (ii) tandemly repeating coding sequences that confer variable length (and, in some instances, function) to the surface protein, and (iii) a common sequence motif appended to structural genes that mediates mutation-driven phase variation in expression. Overall, these systems provide a means of generating extensive surface diversity in rapidly mutating mycoplasma populations (35) and are prime examples of phase variation among bacteria (48). Importantly, these features are readily apparent in mycoplasmal genome sequences.
Variable gene families are widespread among phylogenetically distant mycoplasma species, including some for which complete genome sequences are now known (4, 32, 43). In contrast, comparison of variable gene families among closely related mycoplasmas within narrowly defined clades has only been partially explored (19). These comparisons may offer key insights into the evolution of variable gene systems within highly constrained phylogenetic frameworks and under influences driving genome reduction. This approach may also identify important features in mycoplasmal surface architecture and adaptability that determine factors such as host range and pathogenicity. In addition, these products are known to be targets of the adaptive immune response and contribute to the distinctive antigenic profile of the organism. Nevertheless, the detection and serological discrimination of closely related mycoplasmal species or strains can be confounded due to overlapping sets of antigens and their variable expression (40). The genomic repertoire, pattern of expression, and mechanism of variation of genes encoding surface proteins are therefore critically important in understanding the consequences of mycoplasmal surface variation.
The "Mycoplasma mycoides cluster" (16, 21, 34, 52), comprising a phylogenetic clade of very closely related mycoplasmas, offers particularly attractive opportunities for understanding the dynamics of surface diversification at the genomic level. Subspecies of this group are distinguishable by detailed analysis of 16S rRNA sequence divergence, although their resolution is near the limit of this approach (28, 34). Members of the M. mycoides cluster include two severe pathogens that cause reportable diseases listed by the Office International des Epizooties: contagious bovine pleuropneumonia is caused by Mycoplasma mycoides subsp. mycoides small colony biotype (SC), and contagious caprine pleuropneumonia is caused by Mycoplasma capricolum subsp. capripneumoniae. Four other members of this clade represent significant pathogens of ruminants, each with a distinctive pathobiology and host range. These are M. mycoides subsp. mycoides large colony type, M. mycoides subsp. capri, Mycoplasma sp. bovine serogroup 7, and M. capricolum subsp. capricolum, a caprine pathogen investigated in this report. M. capricolum subsp. capricolum has also been widely studied as a model to understand the basic molecular biology of mycoplasmas and as a representative of this phylogenetic cluster (9, 23, 31, 37, 44). Notably, subspecies within this cluster show considerable and inconsistent cross-reactivity in assorted serological assays designed to differentiate these organisms (8, 10, 14, 16, 42).
This study was prompted by the completion of the genome sequence of M. capricolum subsp. capricolum (GenBank accession no. NC_007633; http://www.ncbi.nlm.nih.gov), which allowed its comparison to that of M. mycoides subsp. mycoides SC type strain PG1 (M. mycoides subsp. mycoides SC PG1), the only other member of the M. mycoides cluster for which the whole-genome sequence is currently available (53). With information from the complete M. capricolum subsp. capricolum sequence determined from the type strain (Kid) and from the partial genomic sequence (described herein) determined for a clonal isolate derived from the same source of the organism, we examined a repertoire of genes encoding a diverse family of variable surface proteins, termed Vmc (variable Mycoplasma capricolum subsp. capricolum) lipoproteins, that differed from currently reported variable genes in M. mycoides subsp. mycoides SC PG1 (33). We characterized two sets of vmc-encoded size-variable proteins with highly structured or unstructured features. We further verified that the predicted phase-variable expression of vmc genes was governed by a reversible promoter mutation mechanism analogous to that occurring in the vmm gene of M. mycoides subsp. mycoides SC PG1 but with greater subtlety than previously known. These results document the expanded utilization of a multigenic, phase-variable system of contingency genes in the M. mycoides cluster of mycoplasmas and underscore the high degree of genomic variation and plasticity that these systems bring to otherwise conserved genomes of this important group. Our results also suggest applications for diagnostics and detection that may overcome current limitations in these areas.
|
MATERIALS AND METHODS |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
Clonal lineages. Clonal lineages were derived from the mck3 population as described in detail previously (39, 46, 65). Plated populations were screened by immunoblot analyses to score the Vmc expression phenotype, using specific polyclonal antibodies (PAbs) or monoclonal antibodies (MAbs) to synthetic peptides. Assurance of phenotypic switching in successive populations included the use of multiple rounds of colony isolation for a phenotype and the inclusion of lineages where sequential switching was selected to proceed through three alternative phenotypes, in addition to reciprocal switches between two phenotypes. For analysis of DNA sequence or protein expression in particular variants, cells were grown from the primary colony to approximately 10 ml, harvested and rinsed by centrifugation, and suspended in buffer for protein analysis as previously described (39) or for extraction of genomic DNA using a DNeasy tissue kit (QIAGEN, Valencia, CA).
Antibodies, immunological assays, and protein analysis. Synthetic peptides derived from the sequences of predicted Vmc lipoproteins were prepared as described previously, with a Cys residue appended at the N terminus for coupling reactions (6, 7). These included Vmc peptide A (pep A) (CEKQAKQQAEKNAKEELDKAEAELKTAREN), Vmc pep B (CAAKAKLNELQKPSDQAKQEELKKAQEAVT), Vmc pep C (CNEIDAAQNEVERAEAELKDAKENLQKLKK), Vmc pep D (CAVKEAEKKLEDLKKAQPENPKEKQLEEAQ), Vmc pep E (CGSGTVKPAEGGSGTVKPAEGGSGTVKPAR), and Vmc pep F (CSKPGTGSETAPSKPGTGSETAPSKPGK). For immunization, peptides were coupled to maleimide-activated keyhole limpet hemocyanin and inoculated into BALB/c mice as described previously in detail (6). Preimmune and postinoculation sera were collected and used at a 1:200 dilution, except where indicated. All procedures using animals were performed under an animal use protocol. Resulting PAbs used in all immunological assays were compared to control preimmune sera, which were uniformly negative. For immunological inhibition assays, peptides were added to primary Abs at a final concentration of 10 µM.
The enzyme-linked immunosorbent assay (ELISA) has been described in detail elsewhere (6). Briefly, peptides were conjugated with N-[6-(biotinamido)hexyl]-3'-(2'-pyridyldithio)propionamide (biotin-HPDP), immobilized on microtiter plates that were precoated with neutral avidin, and then incubated successively with PAb, secondary horseradish peroxidase (HP)-conjugated Ab, and the chromogenic substrate 2,2'-azino-bis(3-ethylbenz-thiazoline-6-sulfonic acid) to quantify binding that was monitored at an optical density at 410 nm (OD410). To measure the relative binding of each PAb to pep A through pep F, saturating amounts (1:1,000 dilutions) of each PAb were incubated with a panel representing all six immobilized peptides (each PAb yielded an OD of 1.0 with its cognate peptide). Binding to pep A through pep F was determined for each PAb as the percent relative normalized binding, calculated as the ratio of background-corrected OD values obtained with immobilized cognate peptide (normalized to 100%) versus that obtained with the other peptides. For one ELISA, a peptide with the same sequence as pep F, but with Cys appended to the C terminus, was used (SKPGTGSETAPSKPGTGSETAPSKPGKC).
Murine MAbs were constructed at the University of Missouri-Columbia Cell and Immunology Core Facility as described in detail previously (6, 7), using the ELISA as an initial screening assay, followed by other immunological assays described below. Isotypes of MAb 2H8 to VmcE {immunoglobulin G1(
) [IgG1()]}, 29E9 to VmcF [IgG2b()], and 14H2 to VmcF [IgG1()] were determined by ELISA, using HP-conjugated, isotype-specific secondary Abs (Southern Biotech, Birmingham, AL).
Triton X-114 (TX-114) detergent phase fractionation of broth-grown M. capricolum subsp. capricolum was performed using standard procedures described in detail elsewhere (57). Proteins were visualized after separation by Tris-glycine sodium dodecyl sulfate-polyacrylamide gel electrophoresis (SDS-PAGE) in 10% gels by staining them with Coomassie brilliant blue R250 (Serva, Heidelberg, Germany) or silver (Bio-Rad, Hercules, CA). Western blot analyses were performed as previously described (11). Briefly, separated proteins were transferred to nitrocellulose blots that were blocked with 10% fetal calf serum and incubated successively with primary mouse Ab, HP-conjugated secondary Ab, and 4-chloro-1-naphthol substrate in the presence of hydrogen peroxide to visualize the bound Ab. Colony immunoblot analyses were performed as described previously in detail (6, 39) by transfer of 3- to 4-day-old M. capricolum subsp. capricolum colonies to nitrocellulose filters that were then blocked and immunostained as described for Western blot analyses.
Proteomics. TX-114 phase proteins were prepared and precipitated with methanol as previously described (57). Precipitated proteins were digested either en masse in solution or in gel after their separation by SDS-PAGE. For in-solution digestion, the complex mixture of proteins was dissolved in 8 M urea-200 mM Tris buffer (pH 7.8), reduced with dithiothreitol, alkylated by iodoacetamide, and treated with trypsin (modified trypsin, sequencing grade; Promega, Madison, WI) at an enzyme-to-substrate ratio of 1:50 (wt/wt) overnight at 37°C. For in-gel digestion, the SDS-PAGE gel was cut into 40 bands that covered the entire length of the gel, and each band was digested according to a standard in-gel digestion protocol. Digested peptides were analyzed using a capillary liquid chromatography-tandem mass spectrometry (LC-MS/MS) system. They were loaded sequentially on an on-line trap column (0.25 mm by 30 mm, Magic C18AQ, 5 µm, 100 Å; Michrom BioResources, Auburn, CA) at a flow rate of 10 µl/min with buffer A (see below). After application and removal of salt and urea, the flow rate was decreased to 300 nl/min, and the trap column effluent was switched to a home-built fritless reverse-phase microcapillary column (0.1 mm by 180 mm, packed with Magic C18AQ, 5 µm, 100 Å) (17). The reverse-phase separation of peptides was performed using a Paradigm MG4 LC system (Michrom BioResources) and buffers of 5% acetonitrile-0.1% formic acid (buffer A) and 80% acetonitrile-0.1% formic acid (buffer B), using a 150-min gradient (0 to 10% buffer A for 20 min, 10 to 45% for 110 min, 45 to 100% buffer B for 20 min). Peptides separated by reverse-phase chromatography were sprayed directly into an ion-trap mass spectrometer (LCQ Deca XP Plus; Finnigan, San Jose, CA). An MS survey scan was obtained for the m/z range of 400 to 1,400, and MS/MS spectra were acquired for the three most intense ions from the survey scan. An isolation mass window of 3.0 Da was used for the precursor ion selection, and normalized collision energy of 35% was used for the fragmentation. Dynamic exclusion for a 2-min duration was used to acquire MS/MS spectra from low-intensity ions. SEQUEST (13) analysis software (Bioworks v3.1) was used to find the best matching peptides in a translated open reading frame (ORF) database generated in Artemis from the complete M. capricolum subsp. capricolum genome. DTA files (Bioworks v3.1) in ASCII format for each MS/MS spectrum were generated from the raw data for a peptide mass range of 500 to 3,500, a minimum ion count of 10, and a minimum signal of 105. Peptide (parent ion) tolerance of 2.5 Da and fragment ion tolerance of 1 Da were allowed, and fixed modification of carbamidomethylation on Cys (+57 Da) and differential modification of oxidation on Met (+16 Da) were used. ProteinProphet (30) was used to evaluate peptide and protein probability. A total of 185 expressed proteins were identified by the combined approach of a shotgun proteomics (en masse) and SDS-PAGE separation with a 99% confidence level (data not shown). Among these proteins were all six predicted Vmc products. Representative peptide sequences for each (all above 90% confidence level) are shown in Results. Full data on peptides mapped to Vmc or other ORFs described in the text are included in the supplemental material.
Genomic shotgun sequencing. Genomic DNA isolated by phenol extraction from M. capricolum subsp. capricolum clone mck3 was sequenced to 8x redundancy, using shotgun sequencing as described elsewhere (36, 41). Six different plasmid libraries with 3- to 5-kb DNA inserts were prepared using purified chromosomal DNA. For each library, genomic DNA was randomly fragmented by sonication and end repaired with T4 DNA polymerase. DNA fragments were resolved on a 1% agarose gel, and fragments of 3 to 5 kb in length were isolated from the gel, ligated with the pUC18 vector, and introduced into competent Escherichia coli host XL10-Gold cells (Stratagene, LaJolla, CA). Plasmid DNA templates from 9,000 transformants (1,500 from each library) were prepared in 96-well format using an Eppendorf 5-Prime Perfect station. DNA templates were sequenced from both ends using M13 forward and reverse primers and BigDye Terminator chemistry. Sequences were resolved on PE Biosystems 3700 capillary sequencers, assembled using Phred-Phrap (15), and viewed using consed (20). Sequences spanning vmcA to vmcD and vmcE to vmcF were submitted to the National Center for Biotechnology Information (GenBank; http://www.ncbi.nlm.nih.gov).
PCR, DNA sequencing, and molecular cloning. All PCRs were performed with a PerkinElmer GeneAmp 2400 or model 480 thermocycler, using a primer concentration of 0.2 µM. DNA sequencing of amplicons and engineered recombinant constructs was performed at the DNA Core Facility at the University of Missouri-Columbia, using BigDye Terminator chemistry on a PE Biosystems 3730 capillary DNA sequencer.
For analysis of the dinucleotide tract 5' of vmcF in clonal expression variants, 10 to 20 ng of M. capricolum subsp. capricolum genomic DNA template was used in a reaction mixture containing 200 µM deoxynucleoside triphosphate, 5 mM Mg2+, and 3 U of Taq polymerase (Promega), with primers (5' to 3') CAAGAATTCCATTTATAACAACTACTA and GATCCTGTTCCTGGTTTTTCTTC. PCR incubation conditions were as follows: one cycle of 1 min at 94°C, 35 cycles of 40 s at 94°C, 40 s at 44°C, and 50 s at 72°C, and a final extension for 10 min at 72°C. The PCR amplicon was purified using the QIAquick PCR purification kit (QIAGEN) and eluted in 35 µl of 10 mM Tris, pH 8.5 (QIAGEN buffer EB). Two microliters of this preparation was subjected to agarose gel electrophoresis to determine the size and quantity of amplicon by ethidium bromide staining and comparison to standard markers. Approximately 40 ng of the 0.4-kb amplicons was sequenced using the respective amplifying primers. Amplicons from representative expression variants were sequenced on both strands.
To confirm the chromosomal location of vmcF in clonal VmcF expression variants, PCR was performed using primers GAAGAAAAACCAGGAACAGGATC and GTTCGTGTGGATTTTAATGTACC and conditions similar to those used for analysis of the dinucleotide tract 5' to vmcF, except for the following incubation conditions: one cycle of 1 min at 94°C, 35 cycles of 40 s at 94°C, 30 s at 44°C, and 2 min at 72°C, and a final extension for 10 min at 72°C. To confirm the chromosomal location of vmcE in these same variants, PCR was performed using primers GACACAGAAAGGGAAAACC and ACCTTCAGCTGGTTTTTCTTC in a reaction mixture containing approximately 40 ng of template, 3 mM Mg2+, 600 µM deoxynucleoside triphosphate, and 3 U of Pfu Turbo polymerase (Stratagene). The following incubation conditions were used: initial denaturation for 1 min at 94°C, 32 cycles of 30 s at 94°C, 1 min at 46°C, and 18 min at 68°C, and a final extension of 10 min at 68°C. The respective amplicons linking vmcF (1.68 kb) or vmcE (9.88 kb) to their adjacent conserved chromosomal flanking regions were purified as described above and sequenced using the respective amplifying primers. To confirm the length of the repeating region in vmcA, a PCR was performed using genomic template from clone mck3 and primers GGCGATGATTCTGGAACTGG and TAGTGAATCTGATACTGTTAATAACC. The following incubation conditions were used: one cycle of 1 min at 94°C, 35 cycles of 1 min at 94°C, 1 min at 51°C, and 1 min 45 s at 72°C, and a final extension for 10 min at 72°C.
For the cloning and expression of VmcE and VmcF fusion proteins, amplicons were generated from M. capricolum subsp. capricolum mck3 genomic DNA, using a common forward primer, TTCCCCCGGGGATAGATCAAATACTGAA, with an engineered SmaI site and reverse primers CCAGGATCCGTGAGTTTTTTATTCTTCC (for vmcE) and GGAGGATCCGTGTTAGATAAGTGAGTT (for vmcF), each with engineered BamHI sites. A PCR was performed using Pfu Turbo, with the following incubation conditions: one cycle of 94°C for 1 min, 35 cycles of 1 min at 94°C, 1 min at 53°C, and 72 s at 72°C, and a final extension for 10 min at 72°C. Amplicons were purified by electrophoresis in a low-melting-point agarose gel, recovered with the QIAquick gel purification kit (QIAGEN), and ligated with pCR4Blunt-TOPO vector using the Zero Blunt TOPO cloning kit (Invitrogen, Carlsbad, CA). Inserts from randomly selected transformants were confirmed by sequencing, using M13 forward and reverse primers. The 0.9-kb insert containing vmcE was also sequenced using the initial amplifying primers, in order to determine the number of repeat sequences in the coding region. A complete sequence across the insert was not achieved due to a decrease in the quality of sequence over this length. Nevertheless, 22 tandem repeats were confirmed, consistent with the 26 predicted from shotgun sequencing. Recombinant vmcE and vmcF gene sequences (lacking signal peptide sequences) were recovered from their respective constructs by SmaI-BamHI digestion, purified in 1% low-melting-point agarose gels, and ligated with XmnI-BamHI-digested, and then purified, pMALc-2 (TEV) vector (24) to create in-frame malE::vmc gene fusions. Transformant colonies, grown in E. coli DH10B and selected with ampicillin, were screened for inserts by membrane hybridization as previously described (3), using digoxigenin-labeled oligonucleotides corresponding to those used for initial amplification of the respective genes. Plasmid DNA from positive transformants was sequenced using flanking, vector-specific primers GAAATCATGCCGAACATCCCGCAG and ATCCGAATTCTGAAATCCTTCCCT to confirm sequences. This also provided verification of the number of repeats in the recombinant vmcF sequence. Fusion protein was overexpressed from each recombinant construct and purified by amylose affinity chromatography as described previously (25).
Informatics. Nucleotide sequence comparisons were performed with BLASTn (without filter) through the National Center for Biotechnology Information (NCBI) resource (http://www.ncbi.nlm.nih.gov). Comparison of genome and regional genomic sequences utilized BLASTn output files visualized in ACT (http://www.sanger.ac.uk/Software/ACT/), with a default setting for the figure of 100 bp as the minimum size of the matches shown. ORF and DNA sequence analyses and ORF library construction were performed using Artemis (http://www.sanger.ac.uk/Software/Artemis/). Protein comparisons were performed with BLASTp through NCBI. The last search for sequence similarities was performed on 15 May 2006. Protein sequences were analyzed for secondary structure using the PELE suite of programs, and membrane topologies were analyzed using TMHMM and TMAP; other properties are available through Biology WorkBench (http://seqtool.sdsc.edu/CGI/BW.cgi#!). Disordered structure predictions were performed with DISOPRED (http://bioinf.cs.ucl.ac.uk/disopred/) and DisEMBL (http://dis.embl.de/). Coiled-coil structure was analyzed with COILS (http://www.ch.embnet.org/software/COILS_form.html) and MARCOIL (http://www.isrec.isb-sib.ch/webmarcoil/webmarcoilC1.html), including weighted programs for charged polypeptides and a window of 28 residues.
Nucleotide sequence accession numbers. The sequences of regions spanning the vmcA to vmcD genes and the vmcE to vmcF genes in M. capricolum subsp. capricolum clonal isolate mck3 have been assigned GenBank accession numbers DQ480439 and DQ480440, respectively.
|
RESULTS |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
90% of the genome covered by both sequencing efforts, with minor exceptions in variable regions that are discussed below in further detail. We compared the two available complete genomes that currently represent the M. mycoides cluster: M. mycoides subsp. mycoides SC PG1 (53), GenBank accession no. NC_005364 (1,211,703 bp), and M. capricolum subsp. capricolum, GenBank accession no. NC_007633 (1,010,023 bp). As expected from their phylogenetic classification, the overall chromosomal content reflected extensive regions of high sequence similarity as well as striking similarity in gene synteny, involving conserved housekeeping genes, hypothetical proteins, and intergenic regions. Superimposed on this overall similarity were several chromosomal inversions involving segments ranging in length from <1 to >300 kb. Global differences between the chromosomes were also apparent, such as the large number of putative genes (72 ORFs) related to insertion sequence (IS)-like elements in M. mycoides subsp. mycoides SC PG1 (53) compared to those in M. capricolum subsp. capricolum (2 ORFs). Some of these differences have been presented elsewhere (K. Röske, M. J. Calcutt, J. I. Glass, A. Persson, J. Westberg, and K. S. Wise, Abstr. 104th Gen. Meet. Amer. Soc. Microbiol., abstr. G-027, 2004). Further examination at finer resolution revealed several chromosomal regions containing putative genes that were selectively present in one genome or the other. Among these genes were several that encoded predicted surface lipoproteins, having characteristic signal peptide and lipobox signatures. These differences provided the impetus to examine the M. capricolum subsp. capricolum genome in order to identify variable surface lipoproteins that were novel within the M. mycoides cluster.
We searched for an initial set of lipoprotein genes selectively present in M. capricolum subsp. capricolum with features similar to those of other known mycoplasma variable lipoprotein gene families and, in particular, the prototypic vmm gene in M. mycoides subsp. mycoides SC PG1 (33). The vmm gene is subject to phase-variable expression that is associated with insertion/deletion (indel) mutations that occur in a region of dinucleotide (TA) repeats lying between a putative –10 and –35 box located 5' of the vmm structural gene. Transcripts of vmm and the Vmm lipoprotein translation product are documented to be expressed in spontaneous mutants having a TA repeat number of 10. In contrast, mutants containing 12 TA repeats do not express Vmm (33). This "poly(TA)" motif is one of the distinctive variable-number tandem repeat (VNTR) sequences found associated with phase-variable expression of surface proteins in assorted mycoplasma systems (35, 61) and is the only one characterized in the M. mycoides cluster. One of our initial search criteria was the presence of an appropriately placed poly(TA) tract in the noncoding 5' flanking region of candidate lipoprotein genes, emphasizing those that most closely resembled the vmm gene. An additional criterion was the presence of in-frame tandem repeat sequences in the coding region of candidate genes. Such repeats are a hallmark of some other variable gene families reported for mycoplasmas (35, 61) and can generate translation products of variable length on the surface of the organism, through indels that occur in repeated sequences (62).
Through this process, we identified two loci in the M. capricolum subsp. capricolum chromosome that contained six novel ORFs representing a potential family of genes encoding variable lipoproteins of M. capricolum subsp. capricolum. These loci are compared in Fig. 1 to the analogous regions in the M. mycoides subsp. mycoides SC PG1 chromosome, using a BLASTn comparison at intermediate stringency to depict regions of high similarity. These putative genes of M. capricolum subsp. capricolum encode predicted products termed Vmc lipoproteins and comprise the vmc gene family investigated in this report. Four of these (vmcA to vmcD) occur at one locus, and two (vmcE to vmcF) occur at another. These loci are separated by 40 kb of intervening sequence comprised predominantly of genes, including multiple housekeeping genes, with sequences and synteny highly similar to those of genes in the corresponding region of the M. mycoides subsp. mycoides SC PG1 genome. No characteristics suggested a genome "island" encompassing both loci. Within each locus, the vmc genes were adjacent, similarly oriented, and separated by relatively large intergenic regions ranging from 397 to 1,452 bp. No Rho-independent terminator motifs were associated with vmc genes. The G+C content of intergenic regions (19 to 20%) was lower than the whole-genome average (23.8%), whereas those of the vmc coding regions (vmcA to vmcD, 31 to 34%; vmcE to vmcF, 43 to 45%) were markedly elevated by comparison. Each of the two loci in M. capricolum subsp. capricolum is flanked by genes with sequences and synteny that are highly similar to those of their counterparts in M. mycoides subsp. mycoides SC PG1 (portions of the flanking regions are shown in Fig. 1).
|
90% sequence identity to an immunogenic surface lipoprotein encoded by the ortholog MSC_0397 of M. mycoides subsp. mycoides SC PG1 (27). In contrast to the conserved flanking regions, the sequence within this locus, including vmc genes (MCAP_0596 through MCAP_0593), shows distinctive features. With the exception of limited regions of sequence showing similarity between the two genomes, this genomic segment generally represents sequence that is selectively present in the M. capricolum subsp. capricolum genome (based on further comparison of the whole genomes) (data not shown). The corresponding locus of M. mycoides subsp. mycoides SC PG1 contains sequence most of which is absent from M. capricolum subsp. capricolum. Interestingly, one segment in this locus (Fig. 1) is precisely reiterated elsewhere in the M. mycoides subsp. mycoides SC PG1 genome and is related to the mobility element ISMmy1 (54), prevalent in this genome but absent from M. capricolum subsp. capricolum. Despite the distinctive features at this locus, in either genome, closer comparison showed that regions of some vmc genes have significant DNA sequence similarity to counterpart regions in the M. mycoides subsp. mycoides SC PG1 genome and that some of these same regions are redundant in the M. capricolum subsp. capricolum genome. A key focus in this regard is the vmm gene, located at the boundary of the locus in the same position as vmcA in the M. capricolum subsp. capricolum genome. Particularly striking is a region of sequence within and adjoining the vmm gene (MSC_0390) of M. mycoides subsp. mycoides SC PG1 that corresponds to similar regions in multiple vmc genes. This relationship is also apparent in a whole-genome comparison that additionally reveals similar conserved regions in vmcE and vmcF genes (data not shown). Another example of local similarity is the presence of limited DNA sequences within vmcC and vmcD coding regions that correspond to counterpart regions in MSC_0392 and MSC_ 0393, respectively. The significance of these local similarities is considered in the following section. The vmcE to vmcF locus is flanked by sequences that are highly conserved between the M. capricolum subsp. capricolum and M. mycoides subsp. mycoides SC PG1 genome (Fig. 1). The immediate 3' flanking region contains hypothetical proteins MCAP_0623 and MCAP_0624 that have corresponding orthologs in M. mycoides subsp. mycoides SC PG1. The 5' flanking region contains conserved housekeeping genes pgk (MCAP_0631) and gap (MCAP_0632) that also occur in both genomes. Within the locus are vmcE and vmcF, again separated by a relatively long intergenic region. The remaining sequence in this locus is absent from the M. mycoides subsp. mycoides SC PG1 genome. Within the locus is encoded a set of four adjacent ORFs (MCAP_0625 through MCAP_0628), each encoding a protein with membrane-spanning domains but with no known function. In a reciprocal manner, the analogous locus in the M. mycoides subsp. mycoides SC PG1 genome contains regions of sequence that are absent from M. capricolum subsp. capricolum, including a segment reflecting a complex insertional event involving the prevalent IS1296 and IS1694 elements and a segment encoding three hypothetical proteins. Overall, the vmc gene family of M. capricolum subsp. capricolum is distributed between two small regions of the chromosome that contain distinctive gene sets, compared in this binary manner with that of M. mycoides subsp. mycoides SC PG1. In addition, local segments of vmc-associated sequences are similar to those of the vmm gene of M. mycoides subsp. mycoides SC PG1.
Features of vmc genes and predicted translation products.
The features of vmc genes of M. capricolum subsp. capricolum and the predicted properties of their annotated translation products are shown in Fig. 2, which also shows the assembled sequences from the mck3 clonal isolate that we used for further experimentation. All vmc genes encode prolipoproteins with very similar signal peptide sequences, terminating with a standard mycoplasmal lipobox (AVA
C for VmcA to VmcD; VVAC for VmcE to VmcF). After prolipoprotein acylation, signal peptide processing, and translocation, each mature Vmc surface lipoprotein is predicted to contain a diacylglyceryl-Cys residue at the N terminus, anchored in the single limiting plasma membrane of the organism. Each mature Vmc represents a unique sequence among the set of deduced Vmc proteins. None of the predicted Vmc ORFs contains Trp codons. Several other ORFs overlap vmc gene coding regions (Fig. 2) and contain both UGA and UGG Trp codons, characteristic of mycoplasma genes (60). These were not annotated. Tandemly repeating in-frame DNA sequences occur in vmc coding regions and encode corresponding tandem repeat sequences in Vmc surface polypeptides. An exception is vmcC, which lacks such repeats but was included in our anaIysis due to similarities in other features. Both the sequence and the length of the tandemly repeated unit in each Vmc product are unique. Longer repeat units occur in VmcA (80 amino acids [aa]), VmcB (69 aa), and VmcD (68 aa), whereas products of the other vmc locus, VmcE (10 aa) and VmcF (11 aa), contain shorter repeat units analogous to those reported for some mycoplasma variable gene families (35, 61). Assembled shotgun sequencing data from the uncloned parent culture (GenBank accession no. NC_007633) revealed two differences from the mck3 clone: vmcA encoded 5 tandem repeats instead of 6, and vmcE encoded 16 repeats instead of 26. The numbers of repeats within vmcA and vmcE for the mck3 clonal isolate shown in Fig. 2 were verified by generating respective amplicons of the predicted size and sequence, using opposing primers to unique sequences flanking the repeat regions (data not shown). Together, these results suggest the possible dynamic variation in Vmc size based on indels in the repeat regions.
|
VmcE and VmcF products encoded in the other locus are organized quite differently. These mature lipoproteins share a nonrepeating N-terminal portion, and distinctive small tandem repeats extend to the C terminus of each protein. The entire sequence of each mature product is predicted to have random coil conformation without secondary structure. The common N-terminal sequence reflects sequence similarity at the DNA level and defines a clear boundary at which the C-terminal sequences of these products diverge. Interestingly, this common region encoded by vmcE and vmcF is shared in part with the vmm gene of M. mycoides subsp. mycoides SC PG1 (MSC_0390; refer to Fig. 2 for region of sequence similarity), even though the corresponding genes lie in a different locus from that of vmm (Fig. 1). More specifically, the first 11 N-terminal residues of the mature VmcE and VmcF polypeptides are identical to those of Vmm. This similarity further explains the previously reported (33) evidence of a "vmm-like gene" in M. capricolum subsp. capricolum, based on identification in Southern blot analyses of a HindIII genomic fragment hybridizing with a probe specific for this region (Fig. 2). The reported fragment is compatible with the 5,552-bp fragment bearing both vmcE and vmcF indicated in Fig. 1. Due to divergence in the region of the probe, vmcA to vmcD gene sequences would not be predicted to hybridize.
All six vmc genes of M. capricolum subsp. capricolum shared sequence features with the vmm gene of M. mycoides subsp. mycoides SC PG1 relevant to their possible phase-variable expression (33). Each vmc coding region adjoins a highly conserved 5' flanking sequence with similarity to that of vmm, containing –10 and –35 sequence motifs separated by a signature poly(TA) tract (see Fig. 5B for examples). The genome sequence assembled from the mck3 clonal isolate identified poly(TA) tracts ranging from 10 to 16 TA repeats among the vmc genes. Due to the possible presence of variant genomes having indel polymorphisms at this VNTR, expression patterns were not predicted but are addressed experimentally in a later section. Overall, Vmc sequences deduced from the M. capricolum subsp. capricolum genome predicted that (i) Vmc proteins may present distinct polypeptide products on the cell surface, (ii) complete Vmc products are predicted to occur selectively in M. capricolum subsp. capricolum, (iii) two distinct structural classes of Vmc product are encoded in corresponding loci of the M. capricolum subsp. capricolum genome, and (iv) vmc genes in both loci are potentially subject to mutations conferring phase-variable patterns of expression for each Vmc product.
|
|
As an entirely independent method to determine the expression of Vmc protein products as well as other predicted membrane proteins, we explored the use of a proteomics approach, LC-MS/MS, to identify products comprising the "membrane-associated proteome" of M. capricolum subsp. capricolum. For this analysis, TX-114 phase proteins were subjected to trypsin digestion, either en masse after precipitation with methanol and in-solution digestion or after SDS-PAGE separation and excision of adjacent regions in the gel spanning the entire spectrum of sizes, as described in Materials and Methods. Tryptic peptides were then analyzed by nanoflow capillary LC-MS/MS (17), and 30,000 tandem mass spectra were scanned against a library of the annotated ORFs encoded in the M. capricolum subsp. capricolum genome. Among several putative products predicted to be present in the TX-114 phase population, peptides unambiguously assignable to each Vmc product, including VmcC, were identified (Fig. 2). These results confirmed and extended our immunological data, demonstrating that the predicted translation products were expressed from the annotated ORFs of all six vmc genes. This sensitive technique, although not quantitative, was a critical adjunct to our immunological methods and provides a direct means of identifying mycoplasmal membrane surface protein expression in cases where identification of antigenic entities may be hampered by their physical state, low abundance, or variable expression.
Antigenic structure and surface architecture of Vmc lipoproteins. Unusual characteristics of both structured and unstructured Vmc proteins were apparent during their analysis by standard SDS-PAGE analysis. Although VmcB migrated with a measured relative molecular mass (Mr) similar to that predicted from its deduced sequence, VmcA showed a discrepancy between its Mr (74 kDa) and calculated mass (60.7 kDa). It is possible that the extensive structure imposed by the large coiled-coil region of this product, compared to a much smaller structured region in VmcB, could account for this anomaly. Also striking were the characteristics of VmcE and VmcF products expressed in M. capricolum subsp. capricolum and their similarity to properties observed for variable lipoproteins comprising short tandem repeats in other mycoplasma species (56, 61). Both lipoproteins showed characteristic ladder patterns in more-heavily loaded Western blots of whole organisms (Fig. 3C, insets); all members of these laddered sets partitioned into the TX-114 phase fraction (data not shown). For some mycoplasmal variable surface proteins (62), such patterns have been rigorously shown to result from indel mutations in tandem repetitive sequences within coding regions that generate heterogeneity in the number of repeats expressed by different progeny in a propagating population. Typically, a prevalent size of protein dominates in the cell population, with fewer members expressing products of different size. We did not segregate VmcE or VmcF size variants to formally distinguish this model from an alternative possibility that size variants represent precise degradation products of proteolysis. Nevertheless, our observation that size variant vmcE genes occurred in clonal versus parental populations of M. capricolum subsp. capricolum is compatible with the former model. These two mature lipoproteins also show a marked discrepancy between their Mrs determined from SDS-PAGE (VmcE, 46 kDa; VmcF, 26 kDa) and the masses predicted from DNA sequence derived from the mck3 clone (VmcE, 26.6 kDa; VmcF, 14.7 kDa). For each product, this feature is also reflected in the size interval between members of its laddered set, showing an analogous discrepancy between the measured (larger) and calculated (smaller) differences in mass between size variants proposed to differ by integral numbers of repeats (data not shown).
To further characterize the VmcE and VmcF polypeptides and the interactions of their repetitive sequences with cognate peptide-directed MAbs, recombinant plasmids encoding fusion proteins (rVmcE and rVmcF) were constructed and expressed in E. coli. These represented malE::vmc translational fusions expressing maltose binding protein (41 kDa) joined at its C terminus to the full-length mature VmcE or VmcF lipoprotein sequence, devoid of a signal peptide or a lipid modification site. Both fusion proteins were readily soluble in aqueous buffers and were purified by affinity chromatography. The number of repeats in the native genes was preserved in each of these recombinant products, as determined by sequencing amplicons generated from recombinant plasmid templates as described in Materials and Methods. In Western blot analysis, MAbs to VmcE (2H8) or VmcF (29E9 and 14H2) specifically bound to the corresponding rVmc product (Fig. 3D). The relative molecular masses measured in SDS-PAGE for rVmcE (90 kDa) and rVmcF(68 kDa) were greater than the masses predicted from their corresponding deduced sequences (rVmcE, 72.5 kDa; rVmcF, 60.6 kDa), consistent with the discrepancy observed for each of the native products. This result argued against postranslational adducts in the mycoplasma as a basis for the abnormal migration of VmcE or VmcF in SDS-PAGE and suggested that the unusual behavior was a feature of the polypeptide sequence per se. The predicted absence of structure in these products further discounted a structural basis for the effect. A final possibility is that the acidic nature of these products (pI 5.3 or 5.4, respectively) may prevent SDS binding during PAGE. This property may also explain their poor protein staining properties, a feature also noted with similar variable proteins, such as Vlp (39).
With MAbs to VmcE and VmcF, we could examine the binding of each specific reagent to its cognate epitope in multiple contexts. Binding to each corresponding rVmc polypeptide was completely and specifically inhibited in the presence of the soluble peptide initially used for immunization and Ab screening by ELISA. Additional studies with MAbs to VmcF showed that a synthetic peptide with identical sequence to Vmc pep F, but coupled through a Cys residue appended to the C terminus rather than the N terminus, bound as well as the original peptide to either MAb when used as an immobilized target in ELISA. These results confirmed that the repeat sequence was sufficient to mediate MAb binding as a soluble peptide or in a recombinant or authentic translation product containing many tandem repeats. By extension, we determined that binding of PAbs and MAbs to all Vmc proteins detected by Western or colony immunoblotting was similarly inhibited, selectively, by their cognate synthetic peptides (data not shown). Overall, these studies show that the detection system we developed for Vmc proteins can be defined by epitopes accessible on soluble polypeptides.
Phase-variable and combinatorial expression of Vmc products. An important consequence predicted by the conserved 5' flanking sequence of each vmc gene is that each gene (and corresponding Vmc lipoprotein) may be independently subject to phase-variable expression through indel mutations that alter the poly(TA) tract. The colony immunostaining patterns that we initially observed in the clonal mck3 population by using specific Ab to Vmc proteins (Fig. 3B) also supported this possibility. To more rigorously assess the pattern of phase-variable expression of Vmc proteins and to formally define the underlying mutational mechanism, we characterized several clonal populations varying in expression of VmcF, using MAb to this product as a precise tool to identify and score populations for subsequent analysis. In initial studies, using mck3 and its derivative clonal populations, three distinct and consistent phenotypes in colony immunostaining were observed. Individual colonies either were stained uniformly, showed no staining, or displayed a generally negative phenotype but with small sectors of staining in peripheral populations of the colony, termed "microsectors" (µS), as illustrated by the examples shown in Fig. 4A. We then used the MAb to identify and isolate well-separated colonies representing each of the three phenotypes. Isolates were subsequently plated following minimal replication, and the selected populations were scored and further subcloned to isolate each phenotype. Whereas the selected phenotype predominated in subcloned populations, all plated populations contained a small fraction (0.001 to 0.05) of colonies displaying each of the two nonselected phenotypes (see Fig. 5A for examples). These results indicated that the phenotypes were heritable and could interconvert, as previously demostrated for variable phenotypes in other mycoplasma systems (38, 39).
|
We then extended this analysis to determine whether variable expression of VmcF was independent of expression of other members of the Vmc family by monitoring VmcE as an alternative partner for comparison. VmcE was chosen due to the proximity of the corresponding vmcE and vmcF genes and the availability of a specific MAb to monitor expression. We first assessed the same clones showing variable VmcF expression (Fig. 4B) for the expression of VmcE. The three clones expressing variable amounts of VmcF by Western blot analyses all expressed VmcE. In one case, for the clone selected for the VmcF– phenotype (Fig. 4B, right lanes), the result formally showed that VmcE can be expressed without expression of VmcF (a result also consistent with the absence of transcriptional read-through from the vmcE gene). Similarly, another clonal population selected for the VmcF+ phenotype (Fig. 4B, left lanes) appeared to express both products in combination. To verify this possibility and to confirm generally that variable expression of VmcE and VmcF is noncoordinate, we used a second strategy wherein replicate lifts of the same plated clonal population were stained with MAb to VmcE or MAb to VmcF. The high resolution and specificity in these immunoblots revealed dramatic differences in the expression patterns of VmcF and VmcE, both in the plated clonal populations and within single colonies (Fig. 4C). Examples of coexpression, selective expression of either product, and distinct microsectoring patterns within single colonies clearly indicated that expression of these two products was noncoordinate. It is noteworthy that the expression patterns observed for VmcE also showed the three expression phenotypes characterized for VmcF (Fig. 4C). Finally, to determine the predicted independence of switching of other Vmc proteins, we established lineages selected for phase-variable expression of VmcB+ and VmcB– by using our specific PAb (data not shown). Variants of either VmcB phenotype showed noncoordinate switching of VmcE and VmcF products. Overall, these results demonstrated that the phenotypic switches in expression of Vmc gene products were independent, thereby providing a strong argument that a stochastic mutational event drives the variable expression of each product.
A hallmark of mutation-based phase-variable systems is the reversibility of the phenotypic change. Using rigorously cloned populations derived from the mck3 clone, we identified several independent examples showing conversion of any one of the three VmcF phenotypes (Fig. 4A) to both of the other phenotypes (data not shown). Proof of phase-variable switching requires the demonstration of switches in lineages representing successive progeny. To strictly establish the reversible phase variation of all three phenotypes and to gain access to the underlying mutational basis for this variation, we derived and examined multiple clonal lineages selected for variable expression of VmcF. These were selected to represent subsequent switches in populations from one phenotype to another and ultimately represented all possible switches among the three phenotypes. The three clonal lineages (I, II, and III) used in this analysis are shown in Fig. 5A. Through these lineages, we could formally establish that each phenotype defined for VmcF expression was reversible and that all three patterns were interchangeable in a phase-variable manner, as shown in Fig. 4A.
A single dinucleotide indel governs phase-variable vmc gene expression. Our cloned lineages comprised critical switched populations that were highly enriched in a particular VmcF phenotype. These provided a source of genomic templates that could be interrogated to determine the status of poly(TA) tracts 5' of the vmcF gene corresponding to each phenotype and the change of the poly(TA) regions associated with phase transitions in the lineage. We used specific primers that amplified part of the coding region and the 5' flanking region of vmcF (Fig. 2) to determine sequences from several phase variants, including all of those depicted in Fig. 5A. Precise and consistent differences in the poly(TA) tracts associated with each phenotype were observed (examples of sequence chromatograms of these tracts are shown in Fig. 5B). These differences were confirmed by sequencing both strands of selected variants representing each phenotype. Using the convention previously described for vmm variants of M. mycoides subsp. mycoides SC PG1 (33), all VmcF+ variants in the lineages contained 10 TA repeats, all VmcF– variants contained 12 TA repeats, and all microsectored variants contained 11 TA repeats. Because these populations represented formal lineages of progeny undergoing phase variation, these correlations were compelling in linking the length of the TA region with expression of the VmcF product and, by inference, the vmcF gene. The correlation of a specific poly(TA) repeat number with each phenotype was further supported by determining the sequences from several additional, individually cloned populations that did not formally represent switching lineages. For each phenotype, the numbers of TA repeats determined per clone assayed were as follows: VmcF+, n = 10 for all of the 6 clones assayed; VmcFµS, n = 11 for all of the 10 clones assayed; and VmcF–, n = 12 for 7 (and n = 13 for 1) of 8 clones assayed. To rule out other possible mechanisms involving rearrangement of vmcE and vmcF genes during these switches, amplicons linking each gene to the respective conserved flanking region of the vmcE to vmcF locus (Fig. 1) were generated and partially sequenced. In each of the multiple populations queried (Fig. 5A), representing the three variant phenotypes, the genomic configuration of the genes was unchanged.
This strict correlation allowed us to refine the parameters that were initially proposed to explain expression variants of Vmm in M. mycoides subsp. mycoides SC PG1 (33). poly(TA) tracts in the vmm gene of 10 and 12 TA repeats, respectively, correlated with Vmm+ and Vmm– variants, precisely corresponding to the poly(TA) tracts observed in VmcF+ and VmcF– variants. However, our discovery that populations generating microsectored colonies invariably contain poly(TA) tracts of 11 TA repeats suggests a more subtle level of control on vmc (and perhaps vmm) gene expression, dictated by an indel of only 1 TA repeat. The detailed immunostaining pattern within microsectored colonies indicated that the founder organism initially plated, and its early progeny, has a Vmc– expression phenotype and that only later in colony formation do organisms of the Vmc+ phenotype appear at the periphery. We interpret this to indicate that 11 TA repeats correlate with a Vmc– configuration for gene expression and that the phenomenon of microsectoring results from a high rate of switching to the Vmc+ configuration (presumably 10 TA repeats) for a portion of the subsequent progeny within colonies. We further suggest that the frequency of switching from 11 to 10 TA repeats may exceed that of switching from 12 to 10 TA repeats; this is consistent with the pattern seen for colonies switching directly between Vmc– and Vmc+, where a lower frequency for indels involving 2 TA repeats would be less likely to generate mixed populations within single colonies. While this model would also predict the occurrence in Vmc+ colonies of variants that switch from 10 to 11 TA repeats at very high frequency, the resultant Vmc– populations at the periphery of Vmc+ colonies may simply not be resolved by our staining method. Similarly, it may be very difficult to observe predicted high-frequency switches from 11 to 12 TA repeats in microsectored colonies, since the latter would have a Vmc– phenotype, undetectable in this context. It is possible that these more subtle populations were represented in sequence chromatograms of amplicons generated from clonal populations (Fig. 5B), wherein low-amplitude signals consistent with a mixture of indel mutations occurred following poly(TA) tracts (irrespective of the strand sequenced).
The pattern of poly(TA) mutations we observed and the limitation in defining the resultant subtle phenotypes precluded any meaningful assignment of switch frequencies to vmc gene expression. Populations scored for VmcE and VmcF expression displayed the microsectored phenotype, indicating switch frequencies higher than that quantifiable by these methods. This phenotype was less prevalent in lineages expressing variants of VmcB, which we monitored in colony immunoblot analyses with our specific PAb (data not shown). These showed positive, negative, and traditionally sectored phenotypes (Fig. 3B), all consistent with classic high-frequency phase variation systems in mycoplasmas (39, 57) but with a frequency below that of the hypervariable VmcE or VmcF. From our detailed analysis of the vmcF gene, we conclude that all vmc genes are likely to operate in a manner analogous to that of the vmm gene, through a spontaneous indel mutation in the poly(TA) tract that independently governs the expression of each gene. It is quite possible that more-subtle differences occur between these two systems or even between the sets of vmc genes in each locus of M. capricolum subsp. capricolum. For example, although the sequences 5' of vmcF and vmm coding regions are nearly identical, discrepancies occur at critical positions. The transcriptional start identified (33) for vmm (beginning 8 nt downstream of the –10 box with ATCT...) differs from the analogous region for vmcF (AGCT...), shown in Fig. 5B. In addition, the reported Shine-Dalgarno sequence of vmm (AGGAG) differs from that in the corresponding position in vmcE and vmcF (AGGAT), whereas this motif in each of the vmcA to vmcD genes is identical to that in vmm. These differences could confer properties to specific genes that affect aspects of variable expression, such as the frequencies of switching and hierarchical efficiencies of transcription or translation.
|
DISCUSSION |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
