INTRODUCTION

Top Abstract Introduction Materials and Methods Results Discussion References

Myxococcus xanthus is the best-characterized member of the myxobacteria family. These gram-negative soil bacteria are able to undergo a multicellular developmental program in response to starvation. Hundreds of thousands of bacteria glide to aggregation centers to form complex structures known as fruiting bodies. These specialized structures contain differentiated cells, the myxospores (9). During the developmental cycle of M. xanthus, cellular communication involving at least five different extracellular signals, known as A, B, C, D, and E, is required. The D signal corresponds to translational initiation factor 3 (IF3) (6, 7, 8), which contains a particular C-terminal extension absent in IF3s from other species. A mutation impeding development, mapping within this extension, suggested that the extension is necessary for developmental functions of the cell rather than for viability (8).

In prokaryotes, IF3 (encoded by the infC gene) is required for the initiation of translation with at least two other factors, IF1 (encoded by infA) and IF2 (encoded by infB), ribosomes, mRNA, fMet-tRNA_f^Met, and GTP (13). Our knowledge of this process comes essentially from studies with Escherichia coli. The three factors play essential roles in each step of translation initiation to control the correct entry into the first round of the elongation cycle (22). In E. coli, the infB gene is part of the nusA-infB operon (19, 23, 24, 29). IF2 is an essential GTP binding protein (20) which exists in two major forms in E. coli, IF2 (97.3 kDa) and IF2 (79.7 kDa). IF2 exists in two subforms (1 and 2), which are barely distinguishable and which are due to two internal in-frame initiation codons separated by only 18 nucleotides (26). The expression of  and  forms of IF2 has also been observed in Bacillus subtilis (31). On the other hand, only one form seems to exist in other bacteria such as Bacillus stearothermophilus, Streptococcus faecium, and the myxobacterium Stigmatella aurantiaca (3, 4, 11). There is thus no obvious pattern to the occurrence of multiple forms of IF2 in gram-negative and gram-positive bacteria or even in closely related bacilli.

The fact that a C-terminal extension in M. xanthus IF3 appears necessary for developmental functions is intriguing given the very general role of IF3. We have previously identified a similar extension in the N-terminal portion of IF2 in the closely related bacterium S. aurantiaca, which hints that translation IFs, and hence translation itself, might play an important role in differentiation. In order to gain further insight into the existence and characteristics of peculiar protein domains in translation factors, so far only observed in myxobacteria, we decided to study the infB gene of M. xanthus. We took advantage of the sequence conservation between all known IF2 proteins, which covers the central GTP-binding domain and the C-terminal domain, to identify and clone the M. xanthus infB gene by cross-hybridization. The sequence analysis of the open reading frame revealed an N-terminal extension similar to that already observed in S. aurantiaca IF2 followed by a peculiar domain just upstream of the GTP binding site. The expression and the potential occurrence of multiple forms of IF2 in M. xanthus were investigated.

	MATERIALS AND METHODS

Top Abstract Introduction Materials and Methods Results Discussion References

Bacterial strains and plasmids. Bacterial strains and plasmids used in this study are listed in Table 1. M. xanthus DK101 was grown to late exponential phase in 1% Bacto Casitone (Difco) with 8 mM MgSO₄ at 30°C and harvested at ~5 × 10⁸ cells/ml. E. coli strains were propagated at 30, 37, or 42°C in Luria-Bertani (LB) broth or on LB agar plates (1.5% [wt/vol]) (27). If required, ampicillin (100 µg ml¹), chloramphenicol (10 µg ml¹), isopropyl-thiogalactopyranoside (IPTG; 50 µg ml¹), and X-Gal (5-bromo-4-chloro-3-indolyl--D-galactopyranoside; 50 µg ml¹) were added. Growth of liquid cultures were monitored by measuring the optical density at 600 nm.

DNA manipulations, cloning, and transformation. Total genomic DNA was isolated from M. xanthus DK101 by the method of Starich and Zissler (34). Plasmids from E. coli were extracted and purified as previously described by Sambrook et al. (27). Plasmid and genomic DNA was digested with restriction enzymes (Gibco-BRL, New England Biolabs Inc.) according to the supplier's recommendations. DNA restriction fragments were purified from agarose gels using the Qiaquick gel extraction kit (Qiagen). Ligation was obtained using the T4 DNA ligase (Gibco-BRL) in accordance with the manufacturer's recommendations. E. coli competent cells were prepared and transformed as described by Huff et al. (16) or by electroporation as described previously (1).

Plasmid constructs. The 4.67-kb BamHI fragment containing the potential M. xanthus infB homolog (see Results and Fig. 1) was cloned in the pBluescript II SK+ vector (pTLC10).

Screening a M. xanthus  Gem12 library. The  Gem12 library was kindly provided by J. Guespin-Michel (Rouen, France). This library was constructed by partially digesting M. xanthus genomic DNA with Sau3A and selecting fragments ranging from 9 to 23 kb. These fragments were filled in using the Klenow fragment and cloned into the filled XhoI site of  Gem12. A 1.2-kb XhoI-PstI fragment containing the 3' end of the S. aurantiaca infB gene (3) was labeled with [-³²P]dCTP (Amersham) by random priming (10) and used to screen around 80,000 clones by colony hybridization (27).

Hybridization. Southern analysis of plasmid and chromosomal DNA fragments was performed as described previously (33).

DNA sequencing and computer analysis. The inserts of plasmids pTLC10, -11, and -12 were in part sequenced using specific oligonucleotides. In addition, random nested deletions were created using the Exo III mung bean nuclease (Stratagene) to generate plasmids for sequencing. Both strands of the 4.67-kb BamHI region were sequenced with the dideoxy chain termination method (30) using an automated sequencer (ABI 310 Prism; Perkin-Elmer, Applied Biosystems Division) with Taq FS polymerase. Sequence analysis was performed with the program of the Genetics Computer Group (Madison, Wis.) sequence analysis software package. Preliminary sequence data concerning microbial genomes were obtained from The Institute for Genomic Research website at http: //WWW.tigr.org.

Maxicell analysis. Expression of plasmid-encoded proteins in the strain CSR603 was analyzed as described previously (28).

Complementation of an E. coli infB null mutation. The curing of the  infB transducing phage (GJ9-2) from strain SL598R carrying the various test plasmids and analysis of the survivors were performed as previously described (20).

Blotting and immunodetection of protein IF2. The presence of IF2 was examined in cell extracts from M. xanthus and E. coli strains by immunoblotting, using an antibody to E. coli IF2 as described by Howe and Hershey (see Fig. 3B, lane 1, and Fig. 4) (15) or with antirabbit antibody horseradish peroxidase conjugate (Caltag Laboratories) revealed with peroxidase substrate (3-3'-diaminobenzidine tetrahydrochloride tablets; Sigma) (see Fig. 3B, lane 2).

Nucleotide sequence accession number. The overall nucleotide sequence for the 4.67-kb BamHI fragment from pTLC10 appears in the GenBank database under accession no. AF261103.

	RESULTS

Top Abstract Introduction Materials and Methods Results Discussion References

Identification and cloning of the infB locus. We used the previously cloned S. aurantiaca infB gene to screen a  Gem12 genomic library of M. xanthus DK101. A 1.12-kb XhoI-PstI fragment from the 3' end of S. aurantiaca infB served as a hybridization probe and allowed us to detect several positive clones, one of them carrying a 13-kb insert. Southern blot analysis of a BamHI digest of this recombinant lambda phage identified a 4.67-kb BamHI fragment that hybridizes to the 1.12-kb C-terminal probe as well as to a 1.62-kb BamHI-SphI fragment corresponding to the N terminus of S. aurantiaca IF2 protein (3).

View larger version (13K): [in this window] [in a new window]   FIG. 1.   Overview of the genetic organization around the M. xanthus infB gene. Shown are the locations and the orientations of ORF1, the infB gene, the 3' and the 5' regions of two putative ORF (which are similar to the nusA and ORF3 genes) and an ORF present on the complementary strand encoding 885 aa. A partial restriction map of the 4.67-kb BamHI fragment from pTLC10 is presented at the top. Bars, restriction fragments cloned within the pBluescript II SK+ plasmid (for pTLC10, pTLC11, and pTLC12), plasmid pTrc99A (for pTLC22), and plasmid pCL1921 (for pTLC30 and pTLC32).

View larger version (64K): [in this window] [in a new window]   FIG. 2.   Comparison of amino acid sequences in the N-terminal and G domain regions of five bacterial species. XEAP motifs present in the APE region are boxed. The A, P, V, and E residues in the APE regions are shaded. The region rich in G, R, and P is underlined. Boldface characters identify identical amino acids at the same position in at least three IF2 species. Consensus sequences G1 to G5, found in the G domains of GTP-binding proteins, are also indicated. Swiss-Prot accession numbers for the E. coli, B. subtilis, Synechocystis sp., and S. aurantiaca IF2 proteins are P02995, P17889, P72689, and P55875, respectively.

The protein sequence of IF2. The M. xanthus infB gene encodes a protein with a deduced M_r of 111,800, making it the largest of IF2 proteins characterized to date. Compared to its counterparts from S. aurantiaca, E. coli, B. subtilis, B. stearothermophilus, and S. faecium, it contains 16, 180, 354, 328, and 285 additional amino acids, respectively (3, 4, 11, 23, 31). M. xanthus IF2 displays extensive homology to the C-terminal two-thirds of other IF2s, and this holds especially true for the central region (residues 571 to 717 of M. xanthus IF2; Fig. 2) corresponding to the GTP-binding domain (88% identical residues and 95% similarity with the equivalent region of S. aurantiaca IF2 and about 70% with that of the other bacteria cited above). On the other hand, the N-terminal sequences are surprisingly divergent. The increased size of M. xanthus IF2 with respect to other IF2s is exclusively due to an extension of its N-terminal region. However, a hydrophobic-cluster algorithm (12), designed to identify common conformational features among distantly related proteins, revealed that the first 90 residues of the N-terminal domain of M. xanthus IF2 display a highly hydrophobic structure (data not shown) similar to those described for other IF2s (3, 31). The adjacent region (residues 61 to 249) has an abnormal amino acid composition, consisting essentially (67%) of four different amino acids: 48 alanine, 44 proline, 19 valine, and 16 glutamic acid residues (Table 2). Due to the conserved high content in alanine, proline, and glutamic acid we have named this region the APE sequence (3). This domain is found in at least four myxobacterial proteins: the C-terminal portion of IF3 from M. xanthus, the ORF3 protein of S. aurantiaca, and the N-terminal domain of IF2 from M. xanthus and S. aurantiaca. Interestingly, in contrast to those of IF3 and the ORF3 protein, the APE sequences of M. xanthus and S. aurantiaca IF2 contain a high percentage of valine residues and the pattern XEAP is repeated five times, with X being alanine or valine (Fig. 2). Another protein domain immediately adjacent to the APE domain (residues 254 to 309) is found exclusively in IF2 of M. xanthus. This region of 56 aa comprises only glycine, proline, and arginine residues (33, 16, and 7, respectively). Named the GPR domain, it features a particular motif, GGRPGGPGGP, repeated five times (GGPGG six times and GRP six times). Both the APE and the GPR domains contain high levels of proline, 23 and 28% respectively, suggesting a high degree of flexibility for this portion of the protein. The remainder of the N-terminal portion of M. xanthus IF2 (residues 310 to 570) does not display any particular structural features.

Characterization of the putative M. xanthus infB gene product. In order to assess whether the cloned M. xanthus chromosomal fragment actually codes for a gene product of the expected size, we performed a Maxicell analysis of an E. coli strain (CSR603) harboring plasmid pTLC22, which contains the putative infB gene without extensive upstream flanking sequences under the control of the IPTG-inducible Trc promoter (see Materials and Methods). A protein of the expected size (M_r, 111,000) was expressed from the plasmid-encoded gene (Fig. 3A, lane 3). As a control, we used strain CSR603 containing plasmid pLBYC8, which expresses the two forms of E. coli IF2 (IF2 and IF2) from the same promoter (Fig. 3A, lane 2) (3).

View larger version (21K): [in this window] [in a new window]   FIG. 3.   (A) Maxicell analysis. Cells were lysed by boiling them in sodium dodecyl sulfate (SDS) sample buffer. Proteins were separated by SDS-10% polyacrylamide gel electrophoresis. After being dried, the resulting gel was autoradiographed. Lanes: 1, TLC7; 2, TLC8; 3, TLC9. M.x., M. xanthus; E.c., E. coli. (B) Western blot analysis. Lane 1, E. coli crude cell extract; lane 2, M. xanthus crude cell extract.

M. xanthus infB complements an E. coli infB null mutation. We used the previously described "IF2 complementation strain" SL598R (20) to test whether expression of a plasmid-borne M. xanthus infB gene could assure survival of E. coli lacking its chromosomal infB gene. In this strain, a functional wild-type copy of infB is supplied in trans by a thermosensitive lysogenic  phage integrated at att  After heat induction at 42°C, the  infB phage can be cured without killing the bacteria, provided a viable infB allele is supplied in trans.

View larger version (42K): [in this window] [in a new window]   FIG. 4.   Western blot analysis of cured infB null mutant strains. Lanes: 1, E. coli SL598R at 30°C; 2, E. coli SL598R carrying pTRC99A at 30°C (before curing); 3, SL598R carrying pB18-1 at 30°C (before curing); 4, SLT18-1 at 37°C; 5, SL598R carrying pTLC22 at 30°C (before curing); 6, SL598R carrying pTLC22 at 30°C (before curing) with IPTG induction; 7, SLT22 at 37°C. M.x., M. xanthus; E.c., E. coli.

	DISCUSSION

Top Abstract Introduction Materials and Methods Results Discussion References

Even though M. xanthus is a gram-negative organism, the genetic organization around the infB locus more closely resembles that of the gram-positive bacterium B. subtilis than that of E. coli. The intergenic region between nusA and infB in B. subtilis contains two ORF (31), ORF1 and ORF2, encoding two putative proteins named p10A and p11, which are absent in E. coli. Genes encoding proteins similar to the M. xanthus ORF1 product are present in the same chromosomal position in B. subtilis (48% homology), T. thermophilus, Helicobacter pylori, Mycobacterium leprae, and M. tuberculosis. Homologous regions in the products of various ORF1 reading frames are essentially located in the N-terminal halves of the proteins.

In B. subtilis, but not in E. coli, infB is followed by another gene encoding a protein of unknown function, ORF3, whose start codon overlaps the infB stop codon. A homologous gene is also located downstream of M. xanthus infB but is separated from it by a putative transcription terminator. Among the two potential translation start sites for ORF3, the downstream one has a significantly stronger Shine-Dalgarno sequence (AAGGGGG) and the derived N terminus more closely resembles that of other ORF3-encoded proteins from various species. In addition to B. subtilis, a gene homologous to ORF3 can also be found in the same chromosomal position in T. thermophilus and in the closely related myxobacterium S. aurantiaca, while it is present in different locations in several other gram-negative and gram-positive organisms such as Clostridium, Streptomyces, D. radiodurans, Thermotoga maritima, and M. leprae. All these quite-different organisms share nevertheless one common property: they all can differentiate to form spores and/or specific multicellular structures.

Rather unexpectedly, we found an extensive potential ORF (ORF885) on the complementary strand with respect to the infB coding sequence. This ORF codes for a hypothetical protein of 885 aa (M_r of 94,500) and extends from the start codon of ORF3 to codon 223 of infB. A similarity search identified several short potential ORF in the same location (complementary to a large part of infB) in a variety of eubacteria. The best score was obtained with Bordetella pertussis, where a simple frameshift would create an ORF encoding a protein of almost 600 aa with >40% identical residues. Similar frameshifts in other bacteria would also lead to production of this protein, suggesting that this might be the remnant of a gene that has lost its function during evolution.

Western blot analysis showed that M. xanthus expressed a single full-length form of IF2 whereas two or even three forms of the factor, differing in the lengths of their N-terminal regions, have been described for B. subtilis and E. coli, respectively (15, 26, 31).

The role of the N-terminal domain of IF2 is still unclear since, in E. coli, it does not appear to be essential for the initiation of protein synthesis in vivo and in vitro (5, 20, 26). The M. xanthus IF2 protein is significantly larger than its counterpart from E. coli. Nevertheless, it was capable of replacing the endogenous factor in this bacterium. The somewhat slower growth observed with a strain expressing M. xanthus IF2 instead of E. coli IF2 may not be very significant and might simply reflect the higher expression (approximately twofold; Fig. 4, compare lanes 6 and 7) suggested by the Western analysis. On the other hand, removal of the N-terminal extension of S. aurantiaca IF2 improved growth of E. coli somewhat, especially at 30°C (3).

With an M_r of 111,800, IF2 of M. xanthus is larger than all presently known IF2 proteins. This is essentially due to the insertion of a characteristic sequence between residues 61 and 249, which is also present in the S. aurantiaca protein but completely absent from the IF2 proteins from other bacteria, except the cyanobacterium Synechocystis (Fig. 2) (18). Like the myxobacteria, Synechocystis belongs to the group of gliding bacteria capable of forming multicellular structures, with the difference that individual cells fulfill different functions.

The main characteristic of the APE sequence is its abnormal amino acid composition. We now have knowledge of four myxobacterial proteins carrying such a domain. While alanine, proline, and glutamic acid are extremely enriched in all APE domains, valine is also very abundant in IF2 (11%).

The N-terminal APE sequence of M. xanthus IF2 is longer than those of M. xanthus IF3 and the product of S. aurantiaca ORF3 (188 aa versus 66 and 119 aa, respectively) and the repeated XEAP motif is specific to IF2. Despite the differences in size and the periodicity of motifs between the M. xanthus IF2 and IF3 APE sequences, continuous alignment of both regions reveals a 34% similarity. M. xanthus IF2 differs from all other proteins described here by the presence of an additional motif, GGRPGGPGGP, repeated five times. This motif is not only absent from all IF2 proteins known to date but also has no similarity with any protein in the data banks. For now, we can only suggest that this portion corresponds to a separation domain between the atypical N-terminal region and the central protein domain carrying the GTPase activity.

The finding that a developmental mutation in M. xanthus maps to the APE domain in IF3 (7) makes it thus very tempting to speculate that the atypical N-terminal IF2 extension in M. xanthus as well as Synechocystis could play a common but as yet unknown role in development. Mutational analysis of the peculiar M. xanthus IF2 N-terminal domains is in progress and will hopefully provide clues to a possible involvement of IF2 in the developmental process.