The Chemical synthesis of a Gene Coding for Bovine Pancreatic DNase I and Its Cloning and Expression in

A gene coding for bovine pancreatic DNase I has been been tight

The Chemical synthesis of a Gene Coding for Bovine Pancreatic DNase I and Its Cloning and Expression in ~~c~e~~c~~~ coZi* (Received for publication,April 30,199O) Andrew F. Worrall and Hernard A. Connolly+ From the Institc&e of Biomolec~lar Science, Department of Biochemistry, University of Southampton, Bassett Crescent East, Southampton SO9 3TU, United Kingdom A gene coding for bovine pancreatic DNase I has been constructed from synthetic oligonucleotides, This gene has been cloned into a plasmid vector pDOC55 designed to allow very tight control of expression of potentially lethal proteins. Induction of protein synthesis from the gene yielded a peptide of molecular weight of approximately 3 1,000, consistent with DNase I. The yield of this protein from the pDOC55 construct (pAW5) was approximately 150 @g/liter of cell culture. Attempts to clone the gene into a less tightly controfIed expression vector based on the tucpromoter (pKK223-3) were unsuccessful, presumably due to the expected lethality of the product. Mutagenesis of the gene to replace the active site histidine  in the protein with glutamine yielded a gene readily clonable into both expression systems. Yields of the mutagenized protein were approximately 6 mg/ liter from a pDOC55 system and 20 mg/liter from a pKK223-3 system. The activity of the proteins were assayed using the Kunitz procedure and their cleavage sel~tivities by digestion of the ~sc~ericfii~ coli tyr T promoter. The recombinant native enzyme had both the same specific activity and DNA cleavage selectivity as the protein isolated from bovine pancreas using these two assays. The H134Q mutant had a specific activity of about 0.001% of the native protein but had an unaltered DNA cleavage selectivity.
Bovine pancreatic DNase I (EC 3.1.21.1) is an endonuclease which cleaves double-stranded DNA to yield 5'-phosphorylated polynucleotides (Moore, 1981). DNase I does not cleave all phosphate bonds in DNA with equal ease (Lomonossoff et al., 1981;Travers, 1984, 1985) and a limited digest of ~sc~eric~~a coli or T promoter DNA yields a characteristic cleavage pattern. Two features are apparent and have been described in terms of a global function and a local function of the DNA structure. With the global function the rates at which runs of phosphodiester bonds are cleaved vary slowly, Here the cutting probability of a particular phosphodiester on one strand correlated well with the probability of cleavage of the phosphodiester displaced by three bonds in the S'direction on the opposite strand. Ad~tionally the global cutting probability reached a minimum at runs of d(A-T) or d(G-C) base pairs. These results were interpreted in terms of DNase I binding in the minor groove of DNA. Phospha~s facing each other across the minor groove are displaced by three bonds in the 3'-direction. *Random sequence B-DNA has a minor groove width of 12 A, whereas runs of d(A-T) have a narrower minor groove width and d(G-C) a broader: it was proposed that DNase I binds well to normal minor groove widths but poorly to a smaller or larger groove. Because minor groove width is a slowly changing parameter along the DNA helix, the global function, i.e. a slowly changing cleavage rate, results. In contrast to the global function, neighboring phosphodiester bonds are often cut at very different rates. This effect was explained by changes in phosphate group orientation, DNase I cuts the bond between the 3'-oxygen atom and phosphorus by an S,2 general base-catalyzed nucleophilic attack of water on phosphors (see below for details). This mechanism is most favored when water attacks the phosphorus in line with and opposite to the bond to be cleaved. Phosphate orientations allowing water access opposite to the scissile bond are easily cut, whereas other orientations are more poorly cut. As phosphate bond orientation can vary dramatically between neighboring phosphates, the local cutting function is obtained. The structure of DNase I has been solved by x-ray crystallography and this has enabled an explanation of several of the above features (Suck et al., 1984(Suck et al., , 1988Suck and Oefner, Oefner and Suck, 1986). In particular, the 2-A resolution structure of DNase I complexed to a short oligonucleotide showed that an exposed loop of the protein centered on Y76 protruded into the minor groove of the DNA and 2 residues (Y76 and R41) completely filled the groove. Additionally a variety of contacts were made between phosphate groups on both strands of the DNA and amino acids in the area of the exposed loop. No contacts were formed between DNase I and the major groove. The DNA was significantly distorted and was found to be bent toward? the major groove with a widening of the minor groove by 3.0 A and a consequent narrowing of the major groove. As DNase I binds in the minor groove one would clearly expect catalysis to be dependent on minor groove dimensions as suggested above. However, although poor catalysis at narrow minor groove widths (d(A-T)-rich regions) is easily explained, the low rates observed at broad minor grooves (d(G-C)-rich regions) are less easy to rationalize since the enzyme widens a "normal" minor groove of DNA when binding this molecule. It was also proposed that as DNA distortion takes place on DNase I binding, the flexibility of the minor groove might be important for catalysis. Two amino acids, E78 and H134, were implicated in catalysis. It was proposed that E78 abstracted a proton from molecule. The resulting OH-acted as a nucleophile performing an S,2 attack on the scissile phosphate group. This mechanism is required to explain the local cleavage effect and is also in agreement with the inversion of configuration at phosphorus observed for the DNase I-catalyzed reaction (Mehdi and Gerlt, 1984).
The possession of an x-ray structure for DNase I makes it an attractive target for the study of protein-nucleic acid interactions by site-directed mutagenesis. The gene for this protein has not been cloned but the amino acid sequence has been determined by conventional protein sequencing techniques (Liao et al., 1973) and found to contain 257 amino acids. The closely related pig and sheep pancreatic DNase I sequences have been investigated Liao, 1986a, 1986b) and were both similarly suggested to contain 257 residues. However, in light of the x-ray structure, the sequence of the bovine enzyme was modified in three respects. In particular a tripeptide IVR was inserted after the IVR sequence at positions 25, 26, and 27 giving an IVR repeat. This means that DNase I has 260 amino acids. Throughout this paper we have numbered the amino acids l-260 to take into account the IVR insertion. In addition, T14 was changed to S14 and the positions of G and P at 227 and 228, respectively, were reversed to P and G. Rather than attempting the cloning of the gene for DNase I we have decided to synthesize it chemically based on this corrected amino acid sequence as a prelude to mutagenesis experiments.
Several genes have now been assembled by chemical synthesis (Groger et al., 1988) and this method is an attractive alternative to conventional gene cloning for the following reasons. Codon usage can be optimized for the proposed organism of expression favoring high yields of the protein (Grosjean and Fiers, 1982). Restriction sites can be built into the gene (using the redundancy of the genetic code) greatly facilitating subsequent gene manipulation and allowing mutagenesis by the cassette method. Specifically for pancreatic DNase I it is anticipated that it would be difficult to obtain the mRNA coding for the protein (for cDNA preparation) as the pancreas is known to be particularly rich in RNases. This paper describes the synthesis of a gene which codes for bovine pancreatic DNase I based on the amino acid sequence described above and also for several variants. In particular we show that the corrected amino acid sequence is still wrong. The most important mistake being that residues E38 and Q39 should be reversed to Q38 and E39. Three variants have been prepared; the first based on the published amino acid sequence (E38, Q39), the second having these residues reversed (Q38, E39), and the third having the active site histidine (H134) replaced by glutamine (Q38, E39, H134Q). We show that the first and third variants have very low DNase I activity and the genes can be overexpressed in E. coli in plasmid pKK223-3 ( Fig. la) under the control of the tat promoter (Brosius and Holy, 1984).

The second construct
gives an active DNase I but the gene is too lethal to be expressed in E. coli using this plasmid.
However, small quantities of active DNase I can be produced in E. coli using the plasmid pDOC55 (  mis, 1987). Based on studies with partially purified active DNase I we show that the recombinant protein has the same specific activity and the same cleavage selectivity for the E. coli tyr T promoter as the enzyme isolated from bovine pancreas.

Gene Design
The gene was designed ( Fig. 2) with the help of commercially available software (Unger, 1986).
Oligonucleotides were synthesized on a 0%fiM scale on an Applied Biosystems 381A DNA synthesizer using standard phosphoramidite chemistry. They were purified by electrophoresis on 20% polyacrylamide gels containing 8 M urea at 50 "C and visualized by UV shadowing. The products were eluted from the gel by incubation at 37 "C in 100 mM Tris, 0.5 M NaCI, 5 mM EDTA, pH 8.0.2-3 nmol of oligonucleotide were 5'-phosphorylated using T4 polynucleotide kinase (18 units). Oligonucleotides were diluted to a final concentration of 10 pmol/al (concentrations were determined by measuring the absorbance of solutions at 260 nm). Oligonucleotides 1A and 16B carry the self-complimentary 5'-overlaps of the complete gene and were therefore not 5'-phosphorylated to prevent concatemer formation on treatment with T4-DNA ligase.

Ligations
Oligonucleotide ligations to give the four gene quarters were carried out on a 200-pmol scale in a final volume of 200 ~1. After annealing oligonucleotide pairs by a heating/slow cooling cycle, incubation was at 37 "C! for 5 h with T4-DNA ligase (40 units). The resultant products were purified by electrophoresis on a 20% polyacrylamide gel run under nondenaturing conditions. Bands corresponding to products of the correct molecular weight (equivalent to approximately 200 base nairs) were visualized bv UV shadowing and eluted from the gel as described above. Final ligation of the four gene quarters to give the require 816-base pair fragment was carried out by mixing the products obtained above with T4-DNA ligase (2 units) in a total volume of 70 ~1 and incubating at 37 "C for 1 h. This gene was used for cloning into M13mp19 without further purification.

Cloning of the Gene into Ml3 and Sequencing
The DNase I gene prepared above was ligated into M13mp19 (Yanisch-Perron et al., 1985) previously cut with EcoRI and NindIII. Thus 0.1 rg of the linear M13mp19 was ligated with varying amounts (0.5-5 11) of the crude gene using T4-DNA ligase (1 unit) in a total volume of 10 ~1. Incubation was carried out at 15 "C for 16 h. The ligated gene/M13 was used to transfect E. coli TGl which were then grown on H-media (Maniatis et al., 1982) in the presence of IPTG and 5-bromo-4-chloro-3-indolyl-fl-u-galactopyranoside as an indicator for recombinant phage. Double-strandedphage DNA was prepared from the white plaques (Messing, 1983) and this was cleaved with EcoRI and Hind111 in an attempt to reisolate the 816-base pair fragment.
Where successful, the plaque was used to prepare singlestranded DNA (Messing, 1983) to be used in dideoxy sequencing (Sanger et al., 1977) (Maniatis et al., 1982), and then cleaved with Hind111 to yield an 816-base pair fragment which was isolated from the gel using "Geneclean" (Bio 101, CA  M72/JM83 (X) H134Q pDOC55 using a Soniprep 150 sonicator (M. S. E., United Kingdom). The sonicate was centrifuged to precipitate insoluble material and the soluble proteins were analyzed by SDS-gel electrophoresis as above. pDOC55 Derivatiues-E.
coli M72[pAW5]were grown to an Ass0 of approximately 0.4 at 30 "C in the presence of 2 mM IPTG. An equivalent volume of LB broth at 54 "C was added to immediately increase the culture temperature to 42 "C. Aliquots of cells were taken and pelleted by centrifugation. Soluble proteins prepared as above were monitored for DNase I activity using the Kunitz assay and this activity was expressed as units per ml of cell culture.

Protein Purification
Cells isolated from large scale inductions were frozen and stored at -20 "C. Four g of induced cells were thawed and resuspended thoroughly in 40 ml of 10 mM Tris, 2 mM CaCl?, 100 pM benzamidine, 100 PM phenylmethylsulfonyl fluoride, pH 7.6. The cells were disrupted by sonication at 4 "C. The insoluble debris was removed by centrifugation at 40,000 X g, 4 "C, for 1 h. The soluble proteins were applied to a column of DEAE-Sephacel (10 x 3 cm') equilibrated to the above buffer and the unbound proteins were washed through with this buffer. Bound proteins were eluted with a gradient of O-O.3 M NaCl in the same buffer over 1 liter, run at 60 ml/h. Proteins were detected by absorbance at 280 nm or by activity in the Kunitz assay. The desired fractions were concentrated to 2 ml using Centriprep 10 spun concentrators (Amicon). The resultant solution was applied to a Sephadex G75-SF column (90 x 3 cm*) equilibrated to and eluted with 10 mM Tris, 2 mM Ca& 30 mM NaCI, pH 7.6, at 20 ml/h. The eluate was analyzed as before and by SDS-polyacrylamide gel electrophoresis. The fractions containing the purest protein were combined and concentrated. Purified protein was shock frozen in liquid nitrogen and stored at -20 "C.

Kunitz Assay
The Kunitz assay for DNase I activity (Kunitz, 1950) was performed at pH 8.0 in 10 m&f Tris buffer in the presence of 0.1 mM CaCll and 1 mM MgCl, using 0.05 mg/ml calf thymus DNA (Sigma) as substrate. One Kunitz unit causes an increase in absorbance of 0.001 absorbance units/min for 1.0 ml of the above solution in a lcm path length at 254 nm.

Tyr T Promoter DNA Cleavage
Plasmid pKMA-98 (Drew and Travers, 1984) was cleaved with EcoRI and AuaI to yield a 160-base pair fragment. The fragment was treated with [oI-~'P]ATP (Amersham Int. United Kingdom, 3000 Ci/ mmol) and reverse transcriptase to label one end of the 3'-5' strand (the EcoRI cleaved end). Digestions of the DNA were carried out essentially as described before (Drew and Travers, 1984) in the presence of 0.3 mM MgC12, 0.3 mM MnS04, 10 mM NaCl, 5 mM Tris, pH 8.0. Aliquots of the digests were removed at 1, 5, and 30 min and the reaction stopped by adding the samples to an equivalent volume of 10 mM EDTA, 0.1% bromphenol blue in formamide. Samples were boiled immediately for 10 min and separated by electrophoresis on 8% polyacrylamide gels in the presence of 7 M urea. Autoradiography was performed at -70 "C with an intensifying screen for 16 h.

RESULTS
Cloning of the 816-base pair synthetic DNA fragment into M13mp19 yielded a single recombinant phage from which the fragment could be reisolated.
Sequencing of the gene revealed a single deletion of a T-A base pair at position 609. This was corrected by cassette mutagenesis using the S&I and SpeI restriction sites. Subcloning of this gene from M13mplSDNaseIrec2 (Table I) into the expression plasmid pKK223-3 yielded a plasmid (pAW2) which directed the production of a protein with a molecular weight of approximately 30,000. The synthesis of this protein increased as a function of time (Fig. 3) after induction with IPTG; however, it was found that after 24 h the majority of the protein was present in inclusion bodies in the insoluble fraction of the cells (data not shown). We were unable to reconstitute a soluble protein from these inclusion bodies. The optimum time for the production of soluble protein was 4 h (data not shown) at which time it formed 10% of the total cellular protein as judged by gel scanning (Fig. 3). The purified protein (Fig. 5) was inactive in the Kunitz assay. The yield of the protein was approximately 20 mg/liter of cell culture. Protein sequencing data (Table II) revealed some differences between the sequence of this protein and native bovine DNase I. The recombinant protein carried an NHp-terminal methionine not present in the bovine enzyme. The bovine enzyme sequenced for a threonine at position -14 in accordance with the original published sequence (Liao et al., 1973) but contrary to the x-ray derived corrections . Based on the xray data, the recombinant protein carries a serine at this position. No glycosylation was apparent at position -18 in the recombinant protein, this amino acid sequenced as an asparagine whereas no amino acid was observed here in the native enzyme. Lastly the bovine enzyme has the sequence Q38, E39. The published and therefore the recombinant sequence is E38, Q39. Thus the protein derived from pAW2 is designated DNase I (Q38E, E39Q).

Cassette mutagenesis
of MlSmplSDNaseIrecP using the EcoRV and X&I sites yielded MlSmplSDNaseIrec5 which carries a gene which codes for a protein in which 38EQ are switched to 38QE. This gene could not be subcloned into pKK223-3 but was subcloned into the alternative expression vector pDOC55. Cutting pDOC55 with EcoRI and Hind111 liberated a 500-base pair fragment in disagreement with the published restriction map for this vector (O'Connor and Timmis, 1987). Subsequent restriction mapping (data not shown) indicated that the single EcoRI site lay outside of the multiple cloning site. Therefore subcloning was achieved using the Coomassie Blue-stained polyacrylamide gel (Fig. 4). However, DNase I activity was easily detectable in the cleared cell sonicates prepared as described above using the Kunitz assay. This activity was absent from similarly induced M72[pDOC55] or M72[pAW6] cells (pAW6 carries a gene subcloned from M13mplSDNaseIrecG which directs the synthesis of DNase I (H134Q), H-134 being the active site histidine required for enzyme activity). This clearly ascribes the observed nuclease activity to a recombinant DNase I. The optimum time of induction for the production of recombinant DNase I is 2-5 h (Fig. 6). This protein was produced at levels from 0.1 to 1 mg/liter of cell culture and was partially purified (Fig. 5) using an identical two-column procedure to that used for DNase I (Q38E, E39Q). The specific activity of this protein corrected for purity was found to be identical to that of the native protein, measured at 5 x lo* Kunitz units/g of protein.
Similarly, DNase I (H134Q) was purified from induced JM103[pAW4] and M72[pAW6]. The M13mp19 clone which carries the DNase I (H134Q) gene was prepared from MlSmplSDNaseIrec5 using the XhoI and BssHII sites. The codon used for Q-134 was CAG. The yields of this protein were not optimized, 4-h induction times being used for both expression systems. The protein was easily purified from JM103[pAW4] (approximately 20 mg/liter of cell culture) and partially purified (Fig. 5) from M72[pAW6] (approximately 6 mg/liter of cell culture, estimated by gel scanning to determine protein purity). As expected, DNase I (H134Q) is inactive in the Kunitz assay. The cleavage selectivities of the three recombinant proteins towards the tyr T promoter DNA have been examined. Recombinant DNase I and native enzyme have identical hydrolysis patterns (Fig. 7). Despite the apparent inactivity of DNase I (Q38E, E39Q) and DNase I (H134Q) in the Kunitz assay, they show activity (when present in large amounts) in this more sensitive assay. The relative activities of the two enzymes can be estimated at approximately 0.01 and 0.001% of the native activity. While DNase I (H134Q) shows an identical hydrolysis pattern to native enzyme, that for DNase I (Q38E, E39Q) is somewhat altered.

DISCUSSION
A gene coding for bovine pancreatic DNase I, based on the currently published (Liao et al., 1973;Oefner and Suck, 1986) amino acid sequence, has been successfully constructed by the annealing and ligation of 32 synthetic oligonucleotides (Fig.  2). The gene was designed with the following features in mind. possible the codon usage is that found in highly expressed genes in E. coli (Grosjean and Fiers, 1982). (c) A Shine-Dalgarno sequence of 7 bases (AGGAGGT) was placed at nucleotides -8 to -14 with respect to the methionine start codon. This sequence was taken from the mRNA coding for the A protein of R17 phage which is highly expressed protein (Shine and Dalgarno, 1974;Steitz, 1969). (d) Two stop codons (TAA, TAG) were placed after the final amino acid codon to ensure efficient termination of translation with minimal read through. (e) The total 816-base pair fragment carries EcoRI and Hind111 cohesive ends. Each of the synthetic oligonucleotides is about 50 bases in length. We have investigated overlaps of 4,5, and 7 base pairs for the DNA ligase catalyzed joining of synthetic duplexes. Many investigators have used short 4-base overlaps for gene assembly presumably to minimize the chances of incorrect ligations. We observed much higher ligation yields with the 'I-base pair overlaps and would strongly recommend this number as the minimal overlap. Attempts to ligate all 32 oligonucleotides in one pot and therefore assemble the gene in a single step were unsuccessful. Therefore we adopted a strategy of an initial ligation of four quarters of the gene.
The differences in amino acid sequence between our initial protein product and the native bovine enzyme are shown in Table II. We do not believe that the NH2 terminal methionine, serine/threonine-14 differences, or lack of glycosylation at position -18 are responsible for the lack of enzymic activity of this protein, these changes are all at the surface of the protein and would not be expected to be involved in substrate recognition or catalysis. We believe that the key difference lies at positions 38 and 39. The original sequence was E38, Q39, however, our data shows that the true sequence in the bovine enzyme is Q38, E39. Interestingly a complete sequence of the pig enzyme also showed QE at this position (Paudel and Liao, 1986a). With the sheep enzyme this sequence was assigned as EQ (Paudel and Liao, 1986b). However, the methods used were not based on complete sequencing but rather peptide isolation, amino acid composition determination, and sequence assignment by homology comparison with the bovine enzyme. An incorrect EQ assignment with the bovine protein would automatically lead to this mistake for the sheep protein. Examination of the crystal structure of DNase I shows that the amino acid at position -39 has an important catalytic role, coordinating the essential metal ion (Ca" in the x-ray structure but probably Mg+/Mn'+ in uiuo). Obviously glutamine would not be able to bind a metal ion as efficiently as glutamate. Due to the inherent errors in protein sequencing and also crystallographic difficulties in distinguishing acid and amide amino acids, we cannot be sure if the sequence we have prepared represents the true DNase I sequence. Nevertheless the identical specific activities and tyr T promoter cleavage selectivities seen between recombinant DNase I and the bovine enzyme (see below) mean that any amino acid differences are silent mutations with no effect on DNA binding and catalysis. We are currently about 70% of the way through a complete resequencing of the bovine enzyme and will report our results more fully elsewhere.
We ascribe the lack of success in subcloning the DNase I gene from MlSmplSDNaseIrec5 into pKK223-3 to the toxicity of the gene product and the lack of total control over protein synthesis provided by the &m-promoter.
This was overcome by the use of pDOC55. This vector has been used to clone EcoRI endonuclease in the absence of the EcoRI methylase (O'Connor and Timmis, 1987). Expression of cloned genes is under the control of the XpL promoter which is highly repressed in E. coli hosts that produce the X-repressor either as the wild type (JM83(X)) or as a temperature-sensitive are grown in the presence of a kc-inducer, the kc-promoter directs the synthesis of mRNA, antisense to the normal sense message derived from the gene. It is assumed that if any sense message is produced from the XpL-promoter, this will form an RNA-duplex with the antisense message which is incapable of being translated by the ribosome. We have been unable to clone the DNase I gene under the control of the XpL-promoter alone,3 showing that in M72 cells this promoter is not sufficiently tightly controlled and the presence of the antisense promoter is an absolute requirement. Yields of the inactive mutants from the pDOC55 expression system were always a little lower than those obtained from the pKK223-3 system, perhaps indicating a slight repression of synthesis caused by the antisense k-promoter. However, the comparatively small and variable amount of active DNase I produced by the pDOC55 system is clearly a function of the toxicity of this gene product in E. coli. We have tried to utilize the secretion system offered by the plasmid PIN-III-ompA (Ghrayeb et al., 1984) as a possible method for overcoming the toxicity of DNase I, but with no apparent success. Although we have yet to purify the recombinant active DNase I to homogeneity, it is quite clear that the nuclease activity that we have partially purified is authentic DNase I. First, nuclease activity is easily detected in lysed cell suspensions of M72[pAW5] after induction by heat, but not of M72[pAW6] nor of M72[pDOC55]. Second, the partially purified nuclease has the same specific activity as bovine pancreatic DNase I when corrected for purity. However, the best evidence comes from the pattern of hydrolysis produced by DNase I on the tyr T promoter DNA from E. coli. Authentic bovine pancreatic DNase I gives a characteristic partial cleavage pattern (Fig. 7) and this pattern has been shown to be different for other nucleases such as DNase II (Drew and Travers, 1984), micrococcal nuclease (Fox and Waring, 1987), and Sl nuclease . Protein derived from M72[pAW5] gives an identical cleavage pattern to bovine DNase I and therefore shows the same characteristic selectivity in its cleavage reaction. It would he extremely unlikely that a nuclease contamination from E. coli would show an identical cleavage selectivity to the bovine protein. Fig. 7 also shows that at high concentrations (approximately 1.5 X lo5 times that for active recombinant DNase I) DNase I (H134Q) shows DNA cleavage and furthermore, the cutting selectivity is identical to the bovine protein. Given the role of H134 in the mechanism of DNase I as an essential base that serves to activate water during the hydrolysis reaction  it is perhaps surprising that DNase I (H134Q) shows any activity at all. A possible rationale is that although this enzyme lacks its catalytic base it behaves in a similar manner to a catalytic antibody. One would expect the DNase I (H134Q) to bind preferentially to the transition state of the substrate DNA and therefore to stabilize this state, possibly leading to a low catalytic rate. The identical cleavage selectivity shown by DNase I (H134Q) as compared to the bovine enzyme is not surprising as the x-ray structure shows that H134 plays no role in DNA binding. However, at this early stage it is not possible to deduce the quantitative importance of H134 in the catalytic efficiency of the enzyme. In another similar case (Hibler et al., 1987) the removal of the general base catalyst Glu-43 from staphylococcal nuclease and replacement with either charged or uncharged amino acids led to global changes in protein structure, making interpretation of kinetic effects difficult. We have also shown that DNase I (Q38E, E39Q) is also active when assayed by tyr T cleavage.
Interestingly, this variant shows some changes in the selectivity of DNA cleavage when compared with the other proteins studied. The global cleavage pattern is not affected although we see several significant changes in some of the local cleavage effects. As amino acid -39 binds the essential metal ion, these local changes could be due to a repositioning of this cation changing the susceptibility of individual bonds to hydrolysis. These effects will be reported in more detail in a later publication.
Our research is currently taking two directions. First, we are trying to engineer the production of active DNase I recombinants in larger quantities. Second, we are producing and characterizing interesting mutants of DNase I based on the crystal structure. These efforts will be reported on as and when they succeed.