Cloning, Expression, and Sequence Analysis of the Gene Encoding the Alkali-Stable, Thermostable Endoxylanase from Alkalophilic, Mesophilic Bacillus sp. Strain NG-27

Abstract

Alkalophilic Bacillus sp. strain NG-27 produces a 42-kDa endoxylanase active at 70°C and at a pH of 8.4. The gene for this endoxylanase was cloned and sequenced. The gene contained one open reading frame of 1,215 bases. An active site characteristic of the family 10 β-glycanases was recognized between amino acids 303 and 313, with the active glutamate at position 310. Though highly thermostable, the enzyme contains no cysteine residue.

Hemicellulose is the second most abundant renewable polysaccharide in nature after cellulose (3). β-1,4-xylan is a major component of hemicellulose and has a backbone of β-1,4-linked d-xylopyranoside residues substituted with acetyl, arabinosyl, and uronyl side chains (2). Complete breakdown of xylan requires the action of several hydrolytic enzymes (2, 4, 25), the most important of which is the endo-1,4-β-d xylanase (EC 3.2.1.8). Xylanases have been widely detected in bacteria and fungi and have been well characterized for their properties (24, 25). Microbial xylanases usually have acidic or neutral pH optima. Alkaline xylanases are of special interest, as xylan is more readily soluble in alkaline solutions than at a neutral pH. An exciting prospect for recombinant xylanases is their use in the paper and pulp industry (21, 24) to reduce the use of chlorine. However, for xylanase to be used in the mill, xylanase pretreatment has to take place at a high temperature and in alkaline conditions, for which thermostable xylanases capable of acting at a high pH are of great importance (21). Earlier, we reported the isolation of a mesophilic, obligate alkalophilic Bacillus sp. strain NG-27, which can produce xylanase(s) that is optimally active at 70°C and a pH of 8.4 and that can retain 70% of its activity at a pH of 11 (7). To understand the molecular basis of its thermostability, this xylanase has been cloned and sequenced.

A number of microorganisms produce more than one type of xylanolytic enzyme (25). To see if this were also the case with Bacillus sp. strain NG-27, the xylanase(s) was first purified as follows: 100 ml of Luria-Bertani (LB) medium supplemented with 0.5% oat spelts xylan (Sigma) and 1% (wt/vol) Na₂CO₃, was inoculated with 1 ml of a 14-h culture (10⁹ cells ml⁻¹) followed by incubation at 37°C for 24 h. The culture was then centrifuged at 16,266 × g for 10 min. Xylanase activity was assayed in the cell-free supernatant exactly as described before (7), except that oat spelts xylan was used. Zymogram analysis of this preparation, performed essentially as described by Nakamura et al. (20), with minor modifications (8), showed a number of bands with the major activity centered around a broad band of 42 kDa (Fig. 1A, lane 1). These bands could be due either to distinct independent endoxylanases or to a mixture of xylanolytic enzymes (25). Endoxylanase activity was differentiated from the general xylanolytic activity by using the carboxymethyl derivative of xylan (CMX), a substrate specific for the endoxylanases (17). On a zymogram developed with CMX, only one activity band at 42 kDa was seen (Fig. 1A, lane 2), suggesting that this protein is an endoxylanase.

FIG. 1.

(A) Zymographic analysis of preparation of xylanase from Bacillus sp. strain NG-27 on xylan (lane 1) or CMX (lane 2). (B) Zymographic analysis of xylanase on Xylan. Lane 1, enzyme preparation from Bacillus sp. strain NG-27; lane 2, xylanase purified from E. coli DH5Δlac carrying the plasmid pGNG19. The positions of standard protein molecular weight markers (Sigma Chemical Co.) run on the same gel are indicated at left. The sizes are as follows: 66 kDa (bovine albumin), 45 kDa (egg ovalbumin), 36 kDa (glyceraldehyde-3-phosphate dehydrogenase), 29 kDa (carbonic anhydrase), and 24 kDa (trypsinogen).

To clone the gene for the 42-kDa endoxylanase, a shotgun library was constructed in Escherichia coli, at the HindIII site of pBR322, using standard methods (22). On LB agar plates supplemented with 0.5% xylan and 100 μg of ampicillin ml⁻¹, one clone out of a total of 5 × 10³ recombinants showed a zone of clearing around it when the plates were stained with congo red. This recombinant clone, designated pGNG-17, was analyzed further and was found to carry an insert of 4.3 kb.

To localize the gene of the 42-kDa endoxylanase on the 4.3-kb insert, first a detailed restriction map of the 4.3-kb DNA fragment was constructed. The various restriction fragments were then subcloned in pBR322 and expressed in E. coli DH5Δlac. The plasmid carrying the smallest piece of DNA, a 2.6-kb HindIII-HpaI fragment able to express the endoxylanolytic activity, was designated pGNG19 and was chosen for further analysis. Comparison of the activity patterns produced by the cloned enzyme and the NG-27 enzyme on a zymogram with xylan as the substrate confirmed that the 42-kDa protein encoded by pGNG19 was indeed the 42-kDa thermostable, alkali-stable endoxylanase produced by Bacillus sp. strain NG-27 (Fig. 1B, lane 2). This result was further confirmed by zymography with CMX as the substrate (data not shown).

The amount of xylanase produced by a 24-h culture of E. coli DH5Δlac containing pGNG19, grown in LB medium plus 100 μg ampicillin ml⁻¹, was estimated both in the cellular as well as in the cell-free supernatant fractions. A total activity of about 7 U ml⁻¹ was obtained. It was found that only about 15% of the xylanase was secreted into the medium by the recombinant (with the average values, based on three independent experiments, being 15.1% secreted versus 84.9% retained). The properties of the cloned enzyme and those of the 42-kDa endoxylanase purified to homogenecity (as evidenced by the presence of a single band of 42 kDa on a sodium dodecyl sulfate [SDS]-polyacrylamide gel from NG-27) were found to be identical. Both preparations were active over a pH range of 6 to 11 and a temperature range of 25 to 85°C, with the optimum activity at 70°C and pH 8.4. The half-life under optimum conditions was 30 min (data not shown).

To obtain the sequence of the 2.6-kb HindIII-HpaI fragment, various subcloned fragments and appropriately designed synthetic primers were used. Sequencing was done with the Sequenase version 2.0 kit of U.S. Biochemical Co. (Cleveland, Ohio). A 2,580-nucleotide-long sequence was analyzed with MicroGenie (Beckman), Sequaid II version 3.81, and CLUSTAL V (14). It revealed one large open reading frame of 1,215 nucleotides encoding a protein of 405 amino acids, with its first 28 amino acids resembling a bacterial signal sequence. Upon comparison, this putative signal sequence was found to have 57.0 and 46.0% homologies with the signal sequences of the xylanases from Bacillus sp. strain C-125 (13) and Bacillus stearothermophilus T-6 (6) respectively. The calculated molecular mass (44.5 kDa) of the protein differed from the value obtained by SDS-polyacrylamide gel electrophoresis (42 kDa) for the secreted endoxylanase by 2.5 kDa, a difference that could be attributed to a possible deletion of the putative signal sequence (equivalent to around 2.8 kDa) from the proenzyme before its secretion into the medium. Sequence analysis revealed the presence of an AGGAG motif, similar to the consensus ribosome binding site for Bacillus species, 8 nucleotides upstream of the translation initiation codon ATG. A well-conserved −10 element (TATAAT) similar to the Bacillus subtilis consensus promoter recognized by ς⁴³, was detected 125 bases upstream of the ribosome binding site. To determine the transcription initiation site, primer extensions were performed using the Preamplification Superscript kit (Gibco-BRL). The Xyla1 primer (5′ CGTTTTTAGCATGTGATAATCTCC 3′), corresponding to nucleotides 1339 to 1362, was labeled with [γ-³²P]ATP and T4 polynucleotide kinase (Promega). Five-microgram quantities of RNA isolated, following standard procedures (22), from Bacillus sp. strain NG-27 and E. coli carrying the recombinant xylanase plasmid pGNG19, were used in each primer extension reaction. The sizes of the extension products were determined by comparison with the DNA sequences generated by using the same primers. Results showed that transcripts initiated from two nucleotides (C and G) located 131 and 132 nucleotides upstream of the start codon in Bacillus sp. strain NG-27 (Fig. 2, lane B), confirming that the conserved −10 motif is indeed recognized by the ς⁴³ polymerase. Similarly, transcripts were found to be initiated from the same nucleotides when the recombinant plasmid pGNG19, carrying the xylanase gene, was expressed in E. coli (Fig. 2, lane E). It may be mentioned here that initiation of multiple transcripts is a common feature in several prokaryotic promoters. A sequence motif similar to the consensus sequence of the catabolite responsive element in Bacillus (14), was detected at nucleotides 1213 to 1226, immediately downstream of the promoter element. The presence of the catabolite responsive element in a number of Bacillus strains has been reported, but its location with respect to the promoter element has been found to vary considerably (14).

FIG. 2.

Mapping of the 5′ ends of the xylanase gene transcripts by primer extension. The partial nucleotide sequence of pGNG19 generated using Xyla1 primer is represented (lanes A, T, G, and C). The extension products of Xyla1 primer obtained by using total RNA from E. coli carrying pGNG19 plasmid and Bacillus sp. strain NG-27 are shown (lanes E and B, respectively). The nucleotide sequence surrounding the transcription initiation sites is shown, and the −10 sequence motif is boxed. The numbers −131 and −132, in parentheses and highlighted by asterisks, indicate the transcription initiation sites for the xylanase gene when expressed in Bacillus and E. coli.

Xylanases and cellulases have been grouped into 45 families on the basis of their active sites (10, 11). The active site of family 10 is centered around a conserved glutamic acid residue which is directly involved in the glycosidic bond cleavage (24). In the NG-27 endoxylanase sequence, a region nearly identical to the family 10 active site was recognized between the amino acid positions 303 and 313, with the catalytically important glutamate at position 310. Lee et al. (18) demonstrated, by site-specific mutagenesis, that in Thermoanaerobacterium saccharolyticum xylanase, Asp-537, Asp-602, and Glu-600 are essential for activity. Amino acids corresponding to these could be identified on the NG-27 xylanase at positions 247, 312, and 310, respectively. Tryptophan has also been shown to be involved in the activity of different xylanases (15, 16). All of the six tryptophan residues in NG-27 endoxylanase appear in the conserved region, but their importance is not known. A surprising observation was that in spite of the high thermostability of the enzyme, it does not contain any cysteine residue, suggesting that disulphide bonds are not responsible for its thermostability.

A BLASTP (1) search for proteins homologous to the NG-27 xylanase, carried out in early June of 1999, produced 124 hits, of which 112 corresponded to xylanolytic enzymes. Barring duplicate entries, a majority of these hits (62%) were against large xylanases containing a family 10 catalytic domain in addition to other domains performing different functions. Among domains not involved in binding or degrading cellulose or xylan, there is one (roughly half the size of the family 10 catalytic domain and located immediately adjacent to its N terminus) which is found to be present in many xylanases from thermophilic sources and which has been proposed to have a thermostabilizing function (5). Since the NG-27 xylanase is produced as a single-domain protein, its stability must be due to features that are intrinsic to its own structural fold and sequence. Clues to such features could potentially emerge from a comparison of the enzyme's sequence with those of other family 10 xylanases also produced as single-domain proteins.

Of the 112 hits, 11 corresponded to single-domain xylanases of known (or likely) thermal stability, while 30 corresponded to mesostable xylanases. Figure 3 shows representative aligned sequences of thermostable xylanases from organisms from among these 11 that are either thermophiles (B. stearothermophilus, Clostridium stercorarium, and Thermotoga neapolitana) or mesophiles (Bacillus sp. strain C-125 or Bacillus sp. strain NG-27), along with those of mesostable xylanases from mesophiles (Aspergillus kawachii and Cellvibrio mixtus). All the xylanases are seen to possess an N-terminal secretory signal corresponding to the region between residues 1 and 27 of the NG-27 sequence, confirming their identities as xylanases lacking the proposed thermostabilizing domain. The xylanases from B. stearothermophilus and C. mixtus are seen to have (different) inserts in their C-terminal halves that do not occur in any other xylanase. With the interesting exception of the T. neapolitana xylanase, all xylanases in the figure that have optimal activity at high temperatures appear to have insert regions in the same locations, corresponding to residues 28 to 58 and 150 to 165, respectively, of the NG-27 xylanase. Interestingly, Thermotoga maritima and Thermoascus aurantiacus xylanases (not shown) also lacked these regions. The absence of these regions in certain single-domain, thermostable xylanases prevents speculation about any significant role of these specific regions in thermal stabilization. Indeed, surface salt bridges and improved packing of nonpolar residues have been implicated in the thermal stabilization of many proteins, including the single-domain xylanase from T. aurantiacus. Such features involve residues that tend to be distributed along the amino acid sequence, so it may be simplistic to expect a contiguous stretch of residues to define a sequence feature universally responsible for thermal stability in all single-domain xylanases. That being so, the occurrence of a region corresponding to residues 150 to 165 of the NG-27 sequence in four thermostable xylanases does, however, point to a common evolutionary origin of these enzymes. Similarly, homology in the region corresponding to residues 28 to 58 of the NG-27 sequence with the Bacillus sp. strain C-125 xylanase points to a common evolutionary origin of these two enzymes.

FIG. 3.

CLUSTAL V-based alignment of the sequence of the thermostable, alkalistable xylanase from the mesophilic Bacillus species strain NG-27 (BAC.SP.NG-27) with sequences of xylanases from related Bacillus species (XYNA_BACS5), thermophilic microbes (XYN1_BACST, CEXY_CLOSR, XYNB_THENE), and mesophilic microbes (XYNA_ASPAK, CELLVIB). Single, fully conserved residues are boxed, and strong or weak conservation of residue type is indicated by dark or light shading, respectively.

A fascinating functional correlate for the existence of a heat-loving active site in this mesophilic enzyme emerges from consideration of what the enzyme is and what it might do for the organism that secretes it. The bacillus that makes the xylanase is capable of using xylan both as a carbon source and as an inducer of xylanase expression and secretion. It was isolated (7, 8) from cellulose-rich waste and compost that lies out in the open under the hot sun in the plains of northern India (where temperatures reach 48 to 50°C in the shade, and perhaps even higher in the open). In the laboratory, Bacillus can use a variety of carbon sources and shows an optimum growth temperature of 27°C when either xylan or other carbon sources are used, suggesting that other enzymes in the bacterium work optimally at, or around, this temperature. Under the midday Indian sun, the organism would probably go into suspended animation and heat shock, but a secreted extracellular xylanase could continue to degrade xylan at high temperatures and produce food that the organism could use once temperatures cool down at night.

Nucleotide sequence accession number.

A 2,580-nucleotide-long sequence analyzed in this study was submitted to GenBank under accession no. XYN AF 015445.