Lack of Functional Significance of C Y S ~ ~ ’ and Cys234 in Terminal Deoxynucleotidyltransferase*

Identification of the three functional regions (cata- lytic, nucleotide substrate-binding, DNA substrate-binding) of the monofunctional template independent DNA polymerase terminal deoxynucleotidyltransferase has not been completely established. The potential participation of 2 amino acid residues, Cys227 and CysZs4, has been controversial, and conflicting data have been published. To investigate the role of Cys227, the human terminal transferase cDNA was modified by site-directed mutagenesis to introduce a glycine codon at this position. Mutant and control wild-type human terminal transferase cDNAs had to be inserted into baculovirus genomes by homologous recombina- tion and overexpressed in Trichoplusia ni insect larvae because terminal transferase cDNAs have not been successfully expressed in bacterial systems. The C Y S ~ ~ ’ 4 Gly mutant and wild-type enzymes displayed similar k, values for both the nucleotide (dGTP) and DNA initiator (dAao) substrates. The kcat for the mutant enzyme (0.56 s-’) was comparable to that of the native enzyme (0.58 s-’). Additionally, catalysis by both mutant and wild-type enzymes was stimulated by Zn2+. These results together with the observation that the amino acid residue at position 234 is not conserved across species indicated that neither Cys234 nor Cys227 is an essential residue in the active site of terminal transferase. Terminal

Terminal deoxynucleotidyltransferase is a DNA polymerase that is specifically expressed in immature lymphocytes and is thought to be part of a complex of proteins that mediates rearrangements of immunoglobulin and T-cell receptor genes. This enzyme is not template directed and may catalyze random addition of nucleotides at the junctions of the rearranged genes (Alt and Baltimore, 1982;Alt et al., 1986). Terminal transferase consists of a single polypeptide chain of 58 kDa which displays a sole catalytic activity and bears three associated functional regions, the nucleotide-and DNA substratebinding sites and the catalytic site. As a result of its streamlined functional design, terminal transferase serves as an excellent model for mechanistic studies of DNA polymerization.
The nucleotide-binding domain of calf terminal transferase has been probed by direct cross-linking with dTTP (Pandy and Modak, 1988) and with a photoaffinity analog, 8-azido-dATP (Evans and Coleman, 1989;Evans et al., 1989). In the * This work was supported by Grant CA26391 from the National Institutes of Health. The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked "aduertisement" in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
presence of UV light, [ c x -~* P ]~T T P was reported to form covalent bonds with CysZz7 and Cys234 of calf terminal transferase (Pandy and Modak, 1988). The authors of this study postulated that these 2 residues form hydrogen bonds with the pyrimidine base in the nucleotide-binding domain. However, this model is inconsistent with the observation that calf terminal transferase does not require mercaptan as a protective agent and is not very sensitive to sulfhydryl inhibitors (Kato et al., 1967). By contrast, the photoaffinity probe 8azido-dATP covalently binds residues within the region G1y343 to Asp3" and does not interact with the region of the calf protein containing and Cys234 (Evans et al., 1989). The DNA-binding domain of calf terminal transferase has also been probed with photosensitive DNAs containing azido residues. Cross-linking was observed within the region of the protein which includes CysZz7 and Cys234 (Farrar et al., 1991). However, in the model for the DNA-binding domain proposed in this study, neither of these Cys residues is in close contact with the DNA. Moreover, Gly appears at position 234 in the human form of the enzyme and Ser appears at position 234 in murine terminal transferase (Koiwai et al., 1986). It is therefore unlikely that the amino acid in this nonconserved position plays an essential role in the active site of terminal transferase. By contrast, CysZz7 is conserved in the enzyme from calf, human, and murine species. To assess the putative involvement of CysZz7 in enzyme activity by an alternate strategy, we generated recombinant human terminal transferase in which CysZz7 was substituted with Gly by site-directed mutagenesis. Although bacterial expression systems offer certain advantages for rapid screening of site-directed mutants, expression of terminal transferase in these systems has not been achieved (Chang et al., 1988).' Therefore, we introduced human terminal transferase cDNA into a baculovirus and expressed the altered gene in the insect host. This procedure permitted purification of functional recombinant terminal transferase. Comparative kinetic analyses of the altered and wild-type recombinant enzymes established that CysZz7 was not critical for enzyme activity.

RESULTS AND DISCUSSION
The observation that the Cys residue at position 234 in calf terminal transferase was not conserved in either the human or mouse enzymes suggested that the residue was not crucial to the enzyme active site and prompted examination of the significance of the Cys at position 227 which is conserved in J. A. Medin and M. S. Coleman, unpublished analyses. Portions of this paper (including "Materials and Methods," part of "Results," Figs. 1 and 2, and Table 11) are presented in miniprint at the end of this paper. Miniprint is easily read with the aid of a standard magnifying glass. Full size photocopies are included in the microfilm edition of the Journal that is available from Waverly Press.

Significance of ana! Cys234 in Terminal Transferase
all three forms of the enzyme. Guided by the natural substitution of to Gly234 in human terminal transferase, CysZz7 of the human enzyme was replaced with Gly by site-directed mutagenesis of cloned cDNA in a baculovirus expression system. The specific activities observed in crude extracts of Sf9 cells or Trichoplusia ni larvae infected with recombinant baculovirus containing either wild-type or Gly227 mutant enzymes were comparable: in Sf9 cells, 1430-1830 (wild-type) versus 1470 (units/mg (mutant), and in larvae, 370 (wildtype) uersus 263 units/mg (mutant).
The retention of enzymatic activity, and the stability of the mutant form of the enzyme under the in vivo conditions employed permitted a more detailed functional comparison of the enzymes through kinetic analyses of partially purified preparations. The wild-type and mutant recombinant enzymes partitioned similarly during the purification procedure. In the case of both substrates, dAso and dGTP, the k, values for the wild-type and mutant enzymes were similar ( Table I) (Table I). These results indicated that CysZz7 was apparently not directly involved in binding either substrate and that the replacement of by Gly had only minor effects on the catalytic properties of the enzyme.
Low concentrations of Zn2+ have been demonstrated to increase the affinity of terminal transferase for the DNA initiator and to decrease the enzyme's affinity for dATP resulting in an overall increase in the apparent Vmax of the reaction (Chang and Bollum, 1990). The altered substrate affinities are presumably the result of a Zn2+-induced conformational change. This capacity to respond to Zn2+ has been retained in the Gly227 mutant enzyme (Table 11, Miniprint).
The region of terminal transferase which contains amino acid positions 227 and 234 has been correlated with the DNA initiator-binding domain (Farrar et al., 1991) and, on the basis of three-dimensional structure predictions, includes a putative a-helix (Matsukage et al., 1987). The nonconservative amino acid substitution of Cys to Gly can theoretically destabilize an a-helix (O'Neil and DeGrado, 1990). Therefore, the Cys to Gly substitution at position 227 could have altered the conformational stability of the enzyme, particularly within the DNA initiator-binding domain. Under standard reaction conditions no alteration in the stability of the mutant enzyme was detected, and the kinetic properties of the human wildtype and mutant enzymes were similar in these assays. However, when their thermal stabilities were compared over a range of temperatures, subtle differences in the functional stabilities of the enzymes were detected (Fig. 2, Miniprint).

TABLE I Kinetic properties of terminal transferases
The kinetic experiments were carried out as described under "Materials and Methods." All reactions were done in triplicate. Kinetic parameters were calculated as described under "Materials and Methods" (Miniprint Supplement). While the inflection point in the temperature inactivation curve occurred at about 45 "C for both enzymes ( Fig. 2A,  Miniprint), analyses of the time course of inactivation at 45 "C revealed differential thermostability which peaked at 10 min (Fig. 2B, Miniprint). This phenotype did not affect the specific activity of the mutant enzyme in crude extracts since these cells and insects were cultured at 27 "C. Indeed, the low optimal temperature employed in the baculovirus expression system facilitates the preparation of mutationally altered proteins with potential thermolabilities. The similarity between the kinetic properties of the wildtype and Gly227 mutant forms of human terminal transferase indicated that the presence of Cys in position 227 was not essential to enzyme activity. In order to assess the functional significance of the Cys which occurs at position 234 of the calf enzyme, the kat/k, ratios for the 58-kDa form of the calf enzyme (Robbins and Coleman, 1988) were calculated from unpublished analyses. The values (214 X lo3 s-l M-' for &, and 3.7 X lo3 s-' M-' for dGTP) were in agreement with corresponding values for the human wild-type and Gly2" mutant enzymes (Table I). Thus, the replacements of Cys with Gly at position 234 (human terminal transferase), and at both positions 234 and 227 (human G1927 mutant) were not accompanied by appreciable alterations in catalytic efficiency. The origin of the observed cross-linking of calf terminal transferase CysZz7 and Cys234 to dTTP in the presence of UV light reported in an earlier study (Pandey and Modak, 1988) is unknown. However, there is evidence that irradiation of sulfur-containing amino acids leads to the formation of radicals where an unpaired electron localizes on sulfur atoms (Vladmirov et al., 1970). It is conceivable that the 2 Cys residues in this region of the protein were activated and crosslinked to dTTP since it is expected that the nucleotide would be in close proximity with the DNA-binding domain. The data presented in this article, however, provided direct evidence that neither nor Cys234 is essential for enzyme activity of terminal transferase. -a i eggs were produced in our laboratory. Larvae were grown on a diet that human terminal transferase coding region was released from pBR (Riley d a . , t 9 8 8 ) . blunt-ended with Klenow iragment and redigested with Nco I. The DNA (1.5kb) was isolated from NuSieve GTG low-melting temperature agarose and cloned Into the transfer vector pAcC4 at the 5' NCO I site and a 3' blunt site. The transfer vector was selected and the construct was engineered lo produce a non-fusion protein product after homologous recombination wtth wild-type AcMNPV. Clones containing the terminal transferase cDNA (pAcC4TdT) in the rlght orlentation wlth respect to the polyhedrln promoter were identifled by restriction analysis.

recombinant circle polymerase chain reaction (RCPCR) method of Jones and Howard
Site-Specific Mulagenesis of cDNA Coding for Terminal Transferase-The (1990) was used to introduce the site-specific mutation that changed the Cys227 codon la GlyZ27. The terminal transferase coding sequence in pAcC4TdT was subcloned into PUC18 vector. The new vector was identlfled as pUCTdT. Four oligomers were synthesized for the mutagenescs step: A(ScacagaaggaancccGgcctggggtccaagg 3') and B(5'cccttgaccttggaccccaggcCgggaattcc 3') where lhe capital letters represent the specific mutation site; A(5'ctgatgattgtgaatggc 3') and E (5'ggagaltattgaagalgg 3'). These four primers were annealed to the pUCTdT plasmid in two separate reactions (A containing A primers and B containing B primers) and amplified by PCR as shown in Flg.
The resulting nicked-duplex circular product was used lo transform frozen competent I 1. Following the PCR step, the reaction products were combined, heated and annealed. fi larvae. When optimum levels of recombinant IErmlnal

EPLL
Purification of Recombinant Terminal Transferase-Two methods of enzyme purification were used. Frozen larvae (7-log) were added to 20 ml Cold BXtraCtiOn buffer (0.25M potasslum phosphate. pH 7.4 containlng tmM 2-mercaptoethanol) and homogenized with two 15-sec bursts of a Brinkman Tissuemlzer. Thls extract was centrlfuged at 20.000 xg for one hour. Protamine sulfate (26 mg) was used to preclpltate viral DNA. The clarlfied supernatant from thts step was precipitated with potassium phosphate, pH 7.4, 1mM 2-mercaptoethanol, diluted 9 fold with buffer. and ammonium sulfate (70% saturation). The protein pellets were dissolved in 50mM applied to a carboxymethyl cellulose column. The enzyme actlvity was eluted at 0.3M potassium phosphate. pH 7.4. The fract~ons containing the majority of enzyme actlvity were concentrated by ultrafiltration, dlluted wlth 50mM potassium phosphate, pH 7.4 and applied to an oligo(dT)-ceilulose Column equllibrated with the same buffer. The enzyme was eluted wlth a KC1 gradlent as previously described (Deibel 8 Coleman, 1980b). Enzyme Preparations were also partially purifled by substitutlng the carboxymethyl CeIIulOse chromatography with blnding of the crude extract to PhosPhoceIIuIose and elutlng with 0.25M Potassium phosphate, pH 7.4 The tractions glycerol at -20°C.
containing enzyme activlty were concentrated by ultrafiltration and stored ~n 50%   aThe experiments in which divalent Cation was added to the standard reactions were carried out as described in Matercals and Methods. The reactions were carried out in triplicate and the enzyme actlvlty values were the average values obtained. -""_""