TEXT

Top Abstract Text References

As was reported previously (7), oat spelt arabinoxylan-grown Cytophaga xylanolytica XM3 produced up to 15 electrophoretically separable endoxylanases but only a single -L-arabinofuranosidase (ARAF) activity, which was purified, characterized, and referred to as ArfI.

During the initial stages of purification of ArfI, there was a parallel effort to clone and sequence the gene encoding it. The first approach used was to shotgun clone (8) the ArfI-encoding gene into Escherichia coli, a strategy that was seemingly successful as it readily yielded an E. coli clone expressing ARAF activity. However, when ArfI was finally purified (7), it became apparent that the gene that had been cloned did not encode ArfI, because none of the amino acid sequences of the four internal peptide fragments of ArfI matched the amino acid sequence deduced from the clone. This indicated that C. xylanolytica contained more than one ARAF-encoding gene, including the gene that was initially cloned and expressed by E. coli but not expressed by C. xylanolytica under our growth conditions.

In this paper, we describe the cloning and sequencing of two ARAF-encoding genes from C. xylanolytica, the ARAF-encoding gene cloned initially (which we refer to as arfII) and the authentic ArfI-encoding gene (designated arfI) that was subsequently obtained by a PCR walking technique.

Bacterial strains and growth conditions. C. xylanolytica XM3 (= DSM 6779) was grown anaerobically at 30°C on oat spelt arabinoxylan as previously described (7). E. coli strains were routinely grown in Luria-Bertani (LB) broth at 37°C with shaking (8). E. coli DH5F' and TOP10F' were used as the recipient strains for recombinant plasmids pUC19 (4) and pCR2.1 (TA cloning kit; Invitrogen, Carlsbad, Calif.), respectively. To select for plasmid-containing transformants of E. coli, ampicillin was included in media at a concentration of 100 µg · ml¹. Solid media contained 15 g of agar per liter.

Cell extracts and enzyme assays. Soluble cell extracts of E. coli were prepared by sonication and centrifugation. These extracts were used either without further treatment (to determine the specific activities of the arabinofuranosidases with p-nitrophenylarabinoside as the substrate) or after concentration by ultrafiltration and dialysis (to determine arabinose-releasing ability with sugar beet arabinan or rye or wheat arabinoxylans as the substrates). The methods used for these procedures have been described previously (7).

Isolation and manipulation of DNA. C. xylanolytica XM3 genomic DNA and E. coli plasmid DNAs were isolated by using genomic and plasmid DNA isolation kits (Qiagen, Valencia, Calif.) according to the manufacturer's instructions. All PCR amplicons were purified with a QIAquick gel extraction kit (Qiagen) used according to the manufacturer's instructions. Standard procedures were used to digest DNA with restriction enzymes, to separate the fragments by gel electrophoresis, and to transfer the fragments to nylon membrane filters (8). Southern blots were prepared by using probe DNA labeled with a digoxigenin DNA labeling and detection kit (Boehringer Mannheim, Indianapolis, Ind.) used according to the manufacturer's instructions.

Amino acid sequences of ArfI peptides. To determine the partial amino acid sequence of ArfI, purified ArfI (ca. 15 µg per lane) was electrophoresed on a 16% (wt/vol) sodium dodecyl sulfate-polyacrylamide gel electrophoresis resolving gel with a 4% (wt/vol) polyacrylamide stacking gel and blotted onto an Immobilon P^SQ membrane as described previously (7). After blotting, the membrane was rinsed with H₂O, soaked in 100% methanol, stained with 0.2% amido black in 40% methanol for 30 s, and destained with multiple changes of H₂O. The band corresponding to the position of ArfI in each lane of the membrane was excised with a clean, sterile razor blade and placed in a sterile Eppendorf tube. The individual ArfI-containing membrane fragments were sent to the Worcester Foundation for Biomedical Research (Shrewsbury, Mass.) for sequence determination. Upon receipt, ArfI was digested with trypsin, and the oligopeptide fragments that were released were purified by reversed-phase high-performance liquid chromatography. The N-terminal amino acid sequences of three such fragments were then determined by the Edman degradation method.

Cloning of arfI and arfII. arfI was cloned by the PCR walking technique (Table 1 and Fig. 1). To do this, the amino acid sequences of three trypsin-generated fragments of ArfI were aligned with similar regions in the deduced amino acid sequence of an ARAF from Bacteroides ovatus V975 (asdII gene product; GenBank accession no. U15179 [10]), a protein to which the trypsin fragments were similar as determined by BLASTp analysis (1). Based on this alignment, the two trypsin fragments (peptides 1 and 3) (Fig. 2) presumed to be farthest apart in ArfI were used to design degenerate primers F1 and R1 for PCR 1, in which genomic DNA from C. xylanolytica was used as the template (all primer sequences are shown in Table 1). The resulting 599-bp amplified product (amplicon) was inserted into cloning vector pCR2.1 and sequenced. Based on the nucleotide sequence of the first amplicon, a new primer (primer R2) was designed and used in PCR 2, in which a pUC19 library of EcoRI-restricted C. xylanolytica genomic DNA was used as the template. The forward primer for this reaction (primer F2) corresponded to a region of pUC19 about 70 bp away from its own EcoRI restriction site, and the sequence of the resulting amplicon was aligned with the sequence of the homologous region from the previous PCR. Analogous procedures were used for PCR 3 and 4; in these PCR pUC19 libraries of C. xylanolytica DNA digested with KpnI and HindIII, respectively, were used as the templates.

View larger version (8K): [in this window] [in a new window]   FIG. 1.   PCR walking procedure used to clone and sequence arfI. Amplicons from PCR 1 through 5 were generated by using primer sets F1-R1 through F5-R5 (arrowheads), whose sequences (Table 1) were complementary to the ends of the amplicons indicated. The thick lines represent nucleotide sequences of C. xylanolytica DNA; the thin lines represent pUC19 vector DNA sequences. Breaks in the amplicons from PCR 3 and 4 represent portions of C. xylanolytica DNA that were sequenced but lie outside the PCR 5 amplicon. The putative ORF corresponding to arfI is represented by the box in the amplicon from PCR 5. The numbers at the bottom (drawn to scale) indicate nucleotide positions in amplicon PCR 5.

View larger version (67K): [in this window] [in a new window]   View larger version (72K): [in this window] [in a new window]   View larger version (71K): [in this window] [in a new window]   FIG. 2.   CLUSTAL W comparison of the deduced amino acid sequences of the arfI gene product (ArfI) and the arfII gene product (ArfII) with the amino acid sequences of homologous enzymes from B. ovatus V975 (asdII gene product), C. stercorarium (C. sterc.) (arfB gene product), and B. subtilis (putative arabinosidase). The boxes indicate regions where the level of sequence identity was 80%. Dashes in a sequence indicate gaps. The numbers to the left and right of each row indicate the numbers of the first and last amino acids, respectively, in the row. The shaded areas in the ArfI sequence correspond to internal peptides 1 (positions 155 to 175), 2 (positions 188 to 212), and 3 (positions 339 to 356) generated by digestion of ArfI with trypsin and internal peptide 4 (positions 269 to 288) generated by digestion with Endoproteinase Lys-C.

Sequences of arfI and arfII and their translation products. The putative ORF of arfI begins with the initiation codon ATG at position 231 and ends at position 1,763 with the stop codon TAA. The 509-amino-acid sequence deduced from this 1,533-bp ORF is shown in Fig. 2 (as ArfI) and corresponds to a 57.7-kDa polypeptide. This size agrees well with the subunit size of ArfI determined by sodium dodecyl sulfate-polyacrylamide gel electrophoresis (56 kDa) (7) and suggests that the native ArfI enzyme is a tri- or tetramer consisting of similar, if not identical, subunits. Consistent with this interpretation were the amino sequences of three trypsin-generated fragments of ArfI (two of which [peptides 1 and 3] were used to design primers for PCR 1) and a fourth peptide fragment (peptide 4) generated by Endoproteinase Lys-C digestion of ArfI, all of which were present in the arfI gene product (Fig. 2). The amino acid sequence of each peptide fragment determined by Edman degradation was identical to the amino acid sequence deduced from the corresponding nucleotide sequence, except for the following amino acids in peptide 4: the deduced T at position 282, which was reported as I; and W and T at positions 279 and 286, respectively, which were not unambiguously resolved after Edman degradation.

1

1

Relationship of ArfI and ArfII to other ARAFs. BLASTp analysis of the deduced amino acid sequences of ArfI and ArfII revealed significant similarities (smallest sum probability, 10¹³) to ARAFs from (in order of similarity) B. ovatus V975 (asdII gene product; GenBank accession no. U15179), Clostridium stercorarium (arfB gene product; GenBank accession no. AF002664), Bacillus subtilis (putative arabinosidase and putative ARAF; GenBank accession no. Z75208 and X89810, respectively), Streptomyces lividans (AbfA; GenBank accession no. U04630), and B. ovatus (asdI gene product; GenBank accession no. U15178). A CLUSTAL W alignment of ArfI and ArfII with the first three proteins mentioned above revealed numerous regions of conserved amino acids (Fig. 2). Based on this alignment, a consensus tree of these ARAFs was generated by parsimony analysis (Fig. 3). The bootstrap values indicate that the evolutionary relationships shown in the tree are strongly supported by the data. Not surprisingly, ArfI and ArfII are most closely related evolutionarily to the asdII gene product of B. ovatus, a member of the same 16S rRNA phylogenetic group as C. xylanolytica (i.e., the "Bacteroides group" of the Flexibacter-Cytophaga-Bacteroides phylum [3]). However, compared to ArfI, ArfII has diverged further from the hypothetical common ancestor. This may reflect its more specialized purpose in C. xylanolytica, a notion supported by its probable export out of the cellular compartment (i.e., cytoplasm) occupied by ArfI.

View larger version (9K): [in this window] [in a new window]   FIG. 3.   Consensus tree showing the evolutionary relationships among the sequences shown in Fig. 2. One hundred randomly sampled replications of the data set were created with SEQBOOT and were subjected to parsimony analysis with PROTPARS. A majority rule consensus tree was generated from the 100 output trees with CONSENSE. Bootstrap values that indicate the number of times that a given cluster was formed are to the right of each node. The lengths of the horizontal lines represent the relative rates of divergence. All of the programs used for this analysis were part of the PHYLIP package (2).

Nucleotide sequence accession numbers. The sequence of the 1,839-bp clone containing arfI has been deposited in the GenBank database under accession no. AF028018. The nucleotide sequence of the 1,940-bp EcoRI restriction fragment containing the putative arfII gene has been deposited in the GenBank database under accession no. AF028019.