Thermotoga neapolitana Homotetrameric Xylose Isomerase Is Expressed as a Catalytically Active and Thermostable Dimer in Escherichia coli

Abstract

The xylA gene from Thermotoga neapolitana 5068 was expressed in Escherichia coli. Gel filtration chromatography showed that the recombinant enzyme was both a homodimer and a homotetramer, with the dimer being the more abundant form. The purified native enzyme, however, has been shown to be exclusively tetrameric. The two enzyme forms had comparable stabilities when they were thermoinactivated at 95°C. Differential scanning calorimetry revealed thermal transitions at 99 and 109.5°C for both forms, with an additional shoulder at 91°C for the tetramer. These results suggest that the association of the subunits into the tetrameric form may have little impact on the stability and biocatalytic properties of the enzyme.

Because of difficulties in cultivating hyperthermophilic microorganisms (e.g., unusual fermentation conditions, low cell yields, toxic and/or corrosive metabolites) (1, 15), obtaining large amounts of a potentially useful thermostable biocatalyst from the natural host is often impractical. Today’s molecular biology tools allow the expression of a desired gene product in a foreign host. Expressing enzymes from hyperthermophiles in mesophilic hosts, though, raises questions about the properties of the recombinant enzyme versus the native enzyme. In addition to the problems normally encountered in expressing recombinant proteins, the correct folding of a hyperthermophilic protein at significantly lower temperatures is a key concern. Many genes from hyperthermophiles have been successfully expressed in mesophilic hosts (2, 15), and the properties of the recombinant and native enzymes have been found to be indistinguishable. As a wider variety of hyperthermophilic proteins are expressed in mesophilic hosts, it remains to be seen whether the large differences in growth temperatures between mesophiles and hyperthermophiles affect protein folding, resulting in differences between native and recombinant hyperthermophilic proteins.

Xylose isomerase (XI) (EC 5.3.1.5) is a well-studied enzyme, in part because of its industrial significance as an immobilized biocatalyst in the production of high-fructose corn syrup (14, 18). Amino acid sequences have been reported for at least 25 XIs (GenBank), and three-dimensional structures [characteristically (α/β)₈ barrels] have been resolved for several of them (6, 9, 11, 12, 15, 20, 21, 29). Most known XIs are homotetramers with molecular masses of approximately 45 to 50 kDa per subunit, although some XIs have been found to be dimeric (3, 15, 17, 25). Typically, two divalent cations (Mg²⁺, Co²⁺, or Mn²⁺) per monomer are required for catalytic activity and stability. The XIs whose amino acid sequences are available form two subclasses, type I and type II; the latter enzymes have an N-terminal 50-amino-acid insert. Since high-resolution three-dimensional structures have been reported only for type I enzymes, it is not known whether this insert has a structural or catalytic role.

XIs have been isolated from bacteria with very different growth temperatures and from the eukaryote Hordeum vulgare (barley) (16). Currently, the most thermostable XIs are those from members of the hyperthermophilic eubacterial genus Thermotoga. XIs from Thermotoga maritima (5) and two strains of Thermotoga neapolitana, strains 5068 and 4359, have been purified and characterized (28). These enzymes in their native forms are all type II homotetramers which are optimally active at a temperature of 95°C or above. The T. neapolitana 5068 xylA gene was cloned, sequenced, and expressed in Escherichia coli, which yielded a recombinant XI (TNXI) with catalytic characteristics identical to those of the native enzyme (28). The recombinant form appeared to be predominantly dimeric, however, in contrast to the tetrameric native form (26, 28). This result raises a question concerning whether the two enzyme forms differ in terms of biochemical properties and thermostability.

MATERIALS AND METHODS

Production of the recombinant TNXI.

The original plasmid construct for TNXI expression (28) was uninducible and unstable during fermentation. Because of these difficulties, the T. neapolitana xylA gene was subcloned from pBluescript (Stratagene, La Jolla, Calif.) into a more tightly regulated vector, pET22B⁺ (Novagen, Madison, Wis.). Oligonucleotides 5′-GGGCATATGGCTGAATTCTTT and 5′-CCAAGCTTCACACTCTGTTTC were purchased from Integrated DNA Technology (Coralville, Iowa); the former oligonucleotide created an NdeI restriction site overlapping the xylA initiation codon (underlined), which allowed T. neapolitana xylA to be inserted directly under the T7-lac fusion promoter of pET22B⁺, and the latter oligonucleotide created a HindIII restriction site downstream of the xylA stop codon, which allowed directional cloning. The gene was subcloned by PCR with Deep Vent polymerase (New England Biolabs, Cambridge, Mass.). The sequence of the amplified gene was verified by sequencing. Up to 50 mg of TNXI per liter of culture was produced in E. coli BL21(DE3) (Novagen). Cells were grown to the mid-exponential phase in Terrific Broth (23), induced with 1 mM IPTG (isopropyl-β-d-thiogalactopyranoside), and allowed to grow to the late exponential phase, which took approximately 3.5 to 4 h. The concentrated cells were pelleted by centrifugation (4,000 × g, 10 min, 4°C), resuspended in 50 mM MOPS (morpholinepropanesulfonic acid) buffer (pH 7.0; pH adjusted at room temperature) (buffer A) containing 5 mM MgSO₄ and 0.5 mM CoCl₂, washed twice, and disrupted by passage through a French pressure cell. Cellular debris was removed by centrifugation at 20,000 × g for 20 min and the resulting soluble fraction was used as the starting point for purification.

Production of native TNXI.

T. neapolitana 5068 was grown in RDM medium (22) supplemented with 0.1% yeast extract and 0.5% xylose at 80°C in sealed culture bottles. The cells were harvested in the late exponential phase, chilled on ice, and pelleted by centrifugation (10,000 × g, 40 min, 4°C). The cell pellets (approximately 0.8 g [wet weight] per liter of culture volume) were washed twice with buffer A and resuspended in approximately 1/1,000th the original volume. The cells were then disrupted by sonication (Heat Systems Inc., Farmingdale, N.Y.). Cellular debris was removed by centrifugation in a microcentrifuge (16,000 × g, 10 min, 4°C), and the supernatant was used for Western blot analysis.

Purification of the recombinant TNXI.

For enzyme purification we employed the following steps: (i) heat treatment of the cell extract at 75°C for 20 min, (ii) pelleting of denatured E. coli proteins by centrifugation at 20,000 × g for 20 min, and (iii) column purification by fast protein liquid chromatography (Pharmacia, Uppsala, Sweden). All chromatographic media and columns were purchased from Pharmacia (Uppsala, Sweden). The heat-treated cell extract was loaded onto a type XK 50 DEAE-Sepharose column equilibrated with buffer A and was eluted with 1 M NaCl in buffer A. TNXI eluted at a high salt concentration, 25%. The fractions containing XI activity were pooled and concentrated with a stirred cell concentrator (Amicon, Wooster, Mass.). The salt was removed by using a Sephadex G-25 desalting column equilibrated with 100 mM MOPS (pH 7.0)–10 mM MgSO₄–1 mM CoCl₂. Molecular masses were determined by using a Superdex S-200 (16/60) column calibrated with protein standards having molecular masses of 443, 200, 150, and 66 kDa (Sigma Chemical Co., St. Louis, Mo.); the flow rate used was 0.3 ml/min.

Western blot analysis.

Enzyme samples were electrophoresed on a 10% native acrylamide gel. The samples were transferred to nitrocellulose with a Hoefer semidry blotter (Pharmacia Biotech, Piscataway, N.J.). Polyclonal antibodies raised against the T. maritima XI (5) were used to detect the native and recombinant TNXIs. The procedures used have been described previously (5).

Enzyme inactivation.

Enzyme samples (0.025 mg/ml) were incubated at 95 ± 0.5°C in 100 mM MOPS (pH 7.0; pH measured at room temperature)–10 mM MgSO₄–1 mM CoCl₂. All samples were analyzed in duplicate and were immediately chilled in an ice water bath following heating. Residual activity was determined at 80°C, as described previously (28).

Differential scanning calorimetry.

Enzyme samples (1.3 ± 0.15 mg/ml) were dialyzed overnight against 500 volumes of 20 mM MOPS (pH 7.0)–2 mM MgSO₄–0.2 mM CoCl₂. The dialysate was used to generate a baseline scan. Samples were scanned at temperatures from 30 to 125°C with a Nano-Cal differential scanning calorimeter (Calorimetry Sciences Corp., Provo, Utah) by using scan rates of 0.5 and 1°C/min. There were no noticeable differences between the results of the 0.5 and 1°C/min scans; therefore, a scan rate of 1°C/min was used for comparative studies.

Isoelectric focusing of TNXI.

Isoelectric focusing of TNXI was done by using a Phast system (Pharmacia Biotech, Piscataway, N.J.). The pH range of the gel was 4.0 to 6.5. The markers used had pI values of 4.55 (trypsin inhibitor), 5.2 (lactoglobulin A), and 5.85 (bovine carbonic anhydrase). The gels and markers were purchased from Pharmacia (Piscataway, N.J.).

RESULTS

Identification of the dimeric and tetrameric forms of TNXI.

When the molecular mass of the recombinant TNXI was determined, two peaks were obtained that exhibited XI activity; a relatively small peak eluted at a molecular mass of 210 ± 20 kDa, and a much larger peak eluted at a volume corresponding to 100 ± 13 kDa (Fig. 1). Both samples yielded a 50-kDa band on a sodium dodecyl sulfate–12.5% polyacrylamide gel electrophoresis gel when they were first boiled for 15 min in 1% sodium dodecyl sulfate (data not shown). The recombinant TNXI (500 ng) and T. neapolitana cell extract (approximately 20 μg) were electrophoresed on a 10% polyacrylamide native gel and subjected to Western blot analysis by using polyclonal antibodies raised against the native T. maritima XI (5). Enzymes with M_r of 200 and 100 kDa were recognized; however, the only species found in the T. neapolitana cell extract was a tetramer (results not shown). This result indicated that the recombinant TNXI existed as a dimer as well as a tetramer. The ratio of dimer to tetramer was approximately 20:1, based on total protein assay data (4). Addition of sodium chloride to a final concentration of 4 M or ammonium sulfate to saturation did not noticeably change the ratio of tetramer to dimer of the recombinant TNXI (results not shown). Purification of the native TNXI and purification from the first construct in E. coli involved either hydrophobic interaction chromatography or ammonium sulfate fractionation (28); however, it did not appear that the differences in the purification procedures were responsible for the different forms.

FIG. 1.

Identification of the dimeric and tetrameric forms of TNXI by gel filtration chromatography. Molecular masses were determined by using a Superdex S-200 (16/60) column calibrated with protein standards having molecular masses of 443, 200, 150, and 66 kDa and a flow rate of 0.3 ml/min.

Comparison of biochemical and biophysical properties of the dimer and the tetramer.

The biochemical and biophysical properties of the two recombinant TNXI forms were investigated. The two forms had similar pH and temperature optima (7.0 and 95°C, respectively), which corresponded to the results obtained previously with the native tetrameric form (28). Inactivation at 95°C did not follow a first-order decay profile for either form (Fig. 2). After an initial rapid decrease in activity, the inactivation rate decreased considerably; the reason for this unusual inactivation is currently being investigated. Two sequential first-order decay profiles were fitted to the data. The calculated rate constants k_d1 and k_d2 were 0.06 and 0.0020 min⁻¹, respectively, for the dimer and ∼0.06 and 0.0031 min⁻¹, respectively, for the tetramer (where 1 and 2 refer to the initial and secondary inactivation phases, respectively). The tetramer lost more activity in the initial phase than the dimer lost. Studies of the tetrameric XI from Arthrobacter strains revealed three possible dimers, only one of which was active (19). Thus, it is possible that disassociation of the tetrameric TNXI may result in a mixture of dimers, a minority of which are inactive or unstable.

FIG. 2.

Thermoinactivation of the two forms of TNXI. Enzyme samples (0.025 mg/ml) were incubated at 95 ± 0.5°C in 100 mM MOPS (pH 7.0; pH measured at room temperature)–10 mM MgSO₄–1 mM CoCl₂. Residual activity was measured at 80°C to determine if there were differences between the thermoinactivation behaviors of the dimeric and tetrameric TNXIs.

When subjected to differential scanning calorimetry, both forms of the enzyme showed separate thermal transitions at 99 and 109.5°C (Fig. 3). The tetramer showed an additional shoulder at approximately 91°C. No activity was recovered from samples after calorimetry, indicating that the unfolding was irreversible.

FIG. 3.

Differential scanning calorimetry of the two forms of TNXI. Dimeric and tetrameric TNXI enzyme samples (1.3 ± 0.15 mg/ml) were dialyzed overnight against 500 volumes of 20 mM MOPS (pH 7.0)–2 mM MgSO₄–0.2 mM CoCl₂. The dialysate was used to generate a baseline scan. Samples were scanned at temperatures from 30 to 125°C by using a scan rate of 1°C/min.

DISCUSSION

The occurrence of the two functional forms of the recombinant TNXI was interesting since it has been shown previously that the native and recombinant TNXIs are homotetrameric (28). When an alternative expression system in E. coli was used, higher levels of expression of the xylA gene led to the production of an active homodimer which was found to be the predominant form. This result is not completely surprising. Several other bacterial XIs have been reported to be dimeric (3, 17, 25), has the barley XI (16) (Table 1). Another thermostable multimeric enzyme, T. maritima lactate dehydrogenase, has also been expressed in E. coli as a tetramer and an octamer (8). The two forms showed similar properties when they were subjected to guanidinium chloride-induced denaturation, near-UV circular dichroism, and fluorescence emission analyses. The interactions between the two tetramers in the octamer were shown to be hydrophobic in nature and did not significantly alter the conformation of the tetramers.

TABLE 1.

Levels of identity and similarity between previously reported XI sequences and the T. neapolitana 5068 XI sequence

XI sequencea	% Identity to T. neapolitana 5068 XI sequence	% Similarity to T. neapolitana 5068 XI sequence	Form(s)b
Type II XIs
Thermoanaerobacter ethanolicus 39E	71.4	84.0	T
Thermoanaerobacterium thermosulfurogenes 4B	70.4	81.8	T, D
Thermoanaerobacterium saccharolyticum B6A-RI	69.9	81.3	T
Clostridium thermosaccharolyticum	67.7	79.7	T
Bacillus subtilis	65.2	81.2	T
Bacillus megaterium	64.2	77.8	ND
Bacillus licheniformis	62.4	77.8	ND
Streptomyces xylosus	60.0	75.5	ND
Lactobacillus brevis	56.3	74.7	T
Lactobacillus pentosus	54.4	74.9	ND
Escherichia coli	48.3	65.3	D
Hordeum vulgarec	57	72	D
Type I XIs
Thermus thermophilus	28.5	46.1	T
Streptomyces rubiginosus	27.4	48.8	T
Streptomyces violaceoniger	26.2	48.1	T
Actinoplanes missouriensis	26.2	46.2	T

^aSequences were obtained from GenBank. 

^bT, tetramer; D, dimer; ND, not determined. 

^cEukaryotic XI. 

Why the dimeric form of the recombinant TNXI produced in E. coli is predominant is not clear. Since E. coli XI has been reported to be a dimer (3), it is possible that the cellular mechanisms of the host are responsible for this particular state of assembly. The thermostable XI from Thermoanaerobacterium thermosulfurogenes has been shown by gel filtration to be produced exclusively as a dimer in E. coli (26a). However, other foreign genes encoding XIs have been expressed in E. coli and have been shown to produce tetramers, including the XI from the thermophile Thermus aquaticus (7, 10, 29). The Thermus XI lacks the 50-amino-acid insert present in both the Thermoanaerobacterium and Thermotoga XIs, so it is possible that this insert may play a role in the level of association of the enzymes. The available crystal structures for type I XIs show that intradimer interactions are much stronger than interdimer interactions (11, 12, 29). Moreover, the active site architecture is complete within the dimer. This was confirmed by studies of the tetrameric type I XI from an Arthrobacter species, which is active as a dimer in the presence of denaturants (19). Therefore, the differences in cellular environments (temperature, salt concentration, pH, etc.) between E. coli and T. neapolitana might be responsible for preventing the relatively weak dimer-dimer interactions necessary for the tetramer to be formed in E. coli. Analytical isoelectric focusing revealed only one species with a pI of approximately 5.5, which is in good agreement with the calculated pI of 5.49 (results not shown). Thus, it is unlikely that a posttranslational modification, such as incomplete removal of the amino-terminal formyl, is responsible for preventing formation of the tetramer. There were no indications that the recombinant dimeric form reverts to the tetrameric form under the conditions studied.

Differential scanning calorimetry was employed to see if there were any fundamental differences in the folding of the two forms. Calorimetric measurements showed that the dimeric TNXI goes through thermal transitions at 99 and 109.5°C; the same transitions occur in the tetrameric version, except that there is a small shoulder at approximately 91°C. These transitions could correspond to the release of the dimer from the tetramer (91°C), the breakdown of the dimer into the monomer (99°C), and the irreversible unfolding of the monomer into the unstructured polypeptide (109.5°C). There are other reasons for multiple thermal transitions (for example, release of ligands or the presence of intermediate species). Biochemical and biophysical characterization of the TNXI will be reported elsewhere (27).

The sole difference between the calorimetric results obtained for the tetramer and the dimer is the transition at 91°C. Above this temperature the two enzymes should be in the same form, and their inactivation behaviors should be comparable. This is shown in Fig. 2, which shows that the inactivation rate constants for both phases of inactivation are similar for the dimer and the tetramer. All of these temperatures are above the normal growth temperature for T. neapolitana (optimum temperature, 80°C; maximum temperature, 90°C), which supports the hypothesis that the native structure is homotetrameric in vivo. Indeed, the native TNXI purified directly from T. neapolitana cell extracts (cells grown at 80°C) was tetrameric (28).

In any case, although the structural features of the recombinant and native forms of TNXI differ in the degree of subunit assembly, their functional properties do not differ. The dimer is a catalytically viable and stable form of the enzyme. Some hyperthermophilic enzymes (13, 24) show a higher level of assembly than their mesophilic homologs. It has been suggested that hyperthermophilic enzymes derive some of their remarkable thermostability from higher levels of assembly through subunit interaction (13). This does not seem to be the case here.