Cloning, Expression, and Biochemical Characterization of Hexahistidine-tagged Terminase Proteins*

The terminase enzyme from bacteriophage λ is composed of two viral proteins (gpA, 73.2 kDa; gpNu1, 20.4 kDa) and is responsible for packaging viral DNA into the confines of an empty procapsid. We are interested in the genetic, biochemical, and biophysical properties of DNA packaging in phage λ and, in particular, the nucleoprotein complexes involved in these processes. These studies require the routine purification of large quantities of wild-type and mutant proteins in order to probe the molecular mechanism of DNA packaging. Toward this end, we have constructed a hexahistidine (hexa-His)-tagged terminase holoenzyme as well as hexa-His-tagged gpNu1 and gpA subunits. We present a simple, one-step purification scheme for the purification of large quantities of the holoenzyme and the individual subunits directly from the crude cell lysate. Importantly, we have developed a method to purify the highly insoluble gpNu1 subunit from inclusion bodies in a single step. Hexa-His terminase holoenzyme is functional in vivo and possesses steady-state and single-turnover ATPase activity that is indistinguishable from wild-type enzyme. The nuclease activity of the modified holoenzyme is near wild type, but the reaction exhibits a greater dependence onEscherichia coli integration host factor, a result that is mirrored in vivo. These results suggest that the hexa-His-tagged holoenzyme possesses a mild DNA-binding defect that is masked, at least in part, by integration host factor. The mild defect in hexa-His terminase holoenzyme is more significant in the isolated gpA-hexa-His subunit that does not appear to bind DNA. Moreover, whereas the hexa-His-tagged gpNu1 subunit may be reconstituted into a holoenzyme complex with wild-type catalytic activities, gpA-hexa-His is impaired in its interactions with the gpNu1 subunit of the enzyme. The results reported here underscore that a complete biochemical characterization of the effects of purification tags on enzyme function must be performed prior to their use in mechanistic studies.

Terminase enzymes are found in all of the large, tailed double-stranded DNA bacteriophages and are responsible, at least in part, for the insertion of a viral genome into an empty, pre-formed shell or procapsid (1)(2)(3). In bacteriophage , the enzyme is composed of large (gpA, 1 73.2 kDa) and small (gpNu1,20.4 kDa) subunits that are isolated as a gpA 1 ⅐gpNu1 2 holoenzyme complex (4 -6). These proteins are an integral part of a series of nucleoprotein intermediates involved in DNA packaging, however, and the subunit stoichiometry in each of these intermediates likely differs (7)(8)(9).
The preferred packaging substrate in phage consists of a linear concatemer of viral genomes, linked head-to-tail and up to 10 genomes in length (10). The cohesive end site (cos) of the viral genome represents the junction between successive genomes in the concatemer and is the site where the terminase subunits assemble to initiate DNA packaging (11)(12)(13)(14). A model for genome packaging has been proposed as follows (7-9) (see Fig. 1): 1) the terminase gpA subunit assembles as a symmetric dimer onto the cosN subsite of cos; 2) cooperative binding of gpNu1 to three repeating R-elements found within the cosB subsite is required for efficient gpA assembly at cosN and the stability of the resulting pre-nicking complex; 3) the endonuclease activity of the gpA subunit nicks the duplex at cosN and, after terminase-mediated strand separation, yields the mature 12-base single-stranded left end of DNA bound and protected by the terminase subunits; 4) this stable nucleoprotein intermediate, known as complex I, next binds to an empty procapsid that triggers an ATP-dependent translocation across the duplex and initiates active DNA packaging; 5) upon encountering the next downstream cos in the concatemer (the end of the viral genome), terminase again nicks the duplex at cosN and strand separation simultaneously releases the DNA-filled capsid and regenerates complex I, which again captures an empty procapsid; and 6) attachment of the tail to the DNA-filled capsid completes the assembly process and yields a fully infectious virus.
Genetic experiments have identified functional domains within the terminase subunits. An N-terminal domain of gpNu1 contains a putative helix-turn-helix DNA binding motif identified by sequence homology ((15) A. Becker, cited in Ref. 7), and site-specific DNA binding has been localized to this region of the protein (16,17). Terminase-procapsid interactions have been localized to the extreme C-terminal amino acids of the gpA subunit (18,19), and it has been postulated that gpA⅐gpA dimer formation occurs through the putative leucine zipper motif identified within the primary sequence of the large subunit (20). Protein-protein interactions in the holoenzyme are mediated through the association of an N-terminal domain of gpA with the C terminus of gpNu1 (16,21). Finally, biochemical experiments have identified two ATPase catalytic sites in the holoenzyme, a high affinity site in gpA and a low affinity, DNA-stimulated site in gpNu1 (22,23).
Our laboratories are interested in the genetic, biochemical, and biophysical properties of DNA packaging in phage and, in particular, the nucleoprotein complexes involved in these processes. These studies require a quick, simple, and efficient method to purify large quantities of protein, especially for the biophysical studies of the holoenzyme and its subunits. Toward this end, we have constructed hexahistidine (hexa-His)-tagged protein subunits and have developed simple, one-step protocols for the purification of phage terminase holoenzyme and the isolated enzyme subunits. Whereas these types of constructs are extremely useful for the rapid purification of protein, the purification tags unavoidably alter the primary sequence of the protein, even with "cleavable" purification tags where one or two amino acids remain after removal of the tag. It is thus imperative that the functional significance of these "minor" changes in protein structure be clearly defined prior to utilization of the constructs for detailed structural and mechanistic studies. We have therefore examined the effect of these hexahistidine tags on the catalytic competence of terminase holoenzyme and the isolated enzyme subunits.

EXPERIMENTAL PROCEDURES
Materials and Methods-Tryptone, yeast extract, and agar were purchased from Difco. Restriction enzymes were purchased from Promega. Mung bean nuclease was purchased from New England Biolabs. Guanidinium hydrochloride was purchased from Mallinckrodt. [␣-32 P]ATP was purchased from ICN. Unlabeled nucleoside triphosphates and ampicillin were purchased from Sigma. Ni-NTA-agarose was purchased from Qiagen. All other materials were of the highest quality commercially available.
Bacterial cultures were grown in shaker flasks utilizing a New Brunswick Scientific series 25 incubator-shaker. Bacterial growth media and agar were prepared as described by Sambrook and co-workers (24). When required, kanamycin and ampicillin were added at 50 and 100 g/ml, respectively. UV-VIS absorbance spectra were recorded on a Hewlett-Packard HP8452A spectrophotometer. Automated DNA sequence analysis was performed by the University of Colorado Cancer Center Macromolecular Resources Core facility or by the DNA sequencing facility at the University of Iowa. Both strands of the duplex were examined to verify the expected DNA sequence.
Bacterial Strains, DNA Preparation, and Protein Purification-The viral and bacterial strains and the plasmids used in these studies are shown in Table I. Plasmids pSF1 and pAFP1 were purified from the Escherichia coli strains C600[pSF1] (25) and JM107[pAFP1] (26), respectively, using Qiagen DNA Prep ® columns. Synthetic oligonucleotides used in this study were purchased from either Integrated DNA Technologies, Inc., or Life Technologies, Inc., and were used without further purification. Purification of wild-type terminase holoenzyme and the isolated wild-type gpA and gpNu1 subunits was performed as described previously (6,27). E. coli integration host factor was purified from HN880 (generously provided by H. Nash, National Institutes of Health, Bethesda) by the method of Nash et al. (28). All of our purified proteins were homogenous as determined by SDS-PAGE and densitometric analysis using a Molecular Dynamics laser densitometer and the ImageQuant ® data analysis package. Unless otherwise indicated, protein concentrations were determined spectrally using millimolar extinction coefficients (6,27).
Construction of pQH101, a Hexa-His Terminase Holoenzyme Overexpression Plasmid-A vector that overexpresses terminase holoenzyme with six histidines fused to the natural C-terminal glutamic acid of the gpA subunit (see Fig. 2) was constructed by PCR methods using pASY20 as a PCR template (Table I). This plasmid contains DNA extending from bp 2216 2 (SphI site) to bp 3522 (BssHI site) cloned into a pIBI30 (International Biotechnologies) background (29). pASY20 thus contains the wild-type sequence for the A gene (3Ј end), except for an XbaI site introduced at bp 2628 ( Fig. 2) (29). The forward PCR primer was complementary to bp 2206 -2229 in the A gene sequence and contained the SphI site. The reverse PCR primer was complementary to bp 2615-2638 in the A gene sequence, except that the TCC Ser codon at bp 2619 -2621 (Ser-637, Fig. 2) was changed to an AGC. This introduced an AflII site at bp 2615 in the A sequence but maintained a wild-type serine codon. PCR amplification yielded the expected 412-bp fragment that was isolated, digested with SphI and XbaI, and ligated into similarly digested pASY20. This afforded the plasmid pASY30 which was identical to pASY20 except for the AflII site introduced at bp 2615. pASY30 was cut with AflII, and the cohesive ends were removed by digestion with mung bean nuclease. The linearized plasmid was next digested with BspEI (bp 3329) which deleted a 714-bp AflII-BspEI fragment (Fig. 2). The following synthetic duplex 1 was ligated into doubly digested pASY30.  The cos region of the genome is shown at top. The three subsites, cosQ, cosN, and cosB, are indicated, as are the three gpNu1-binding elements (R1-R3) and the IHF binding element (I1) found within cosB. The terminase subunits and IHF assemble at cos forming a pre-nicking complex which, in the presence of Mg 2ϩ , nicks the duplex ultimately yielding the stable packaging intermediate complex I. Complex I binds an empty procapsid which initiates an ATP-dependent insertion of viral DNA into the capsid (Active DNA Packaging). We note that the stoichiometry of the terminase subunits in each of the nucleoprotein intermediates remains speculative. Details are presented in the text.
where the BspEI restriction site is indicated in italics. The resulting plasmid, pQH70, contains the 3Ј end of the A gene (bp 2633) with six histidine codons appended but is deleted for the BspEI segment between the sites at bp 2637 and 3329 of pASY20 (Fig. 2). To replace this missing plasmid DNA, pASY20 was digested with BspEI, and the 692-bp fragment was cloned into BspEI-linearized pQH70. The resulting plasmid was digested with Sse83871 (Panvera), and the 263-bp fragment containing the modified segment of the A gene (see Fig. 2) was ligated into the corresponding Sse83871 site of the terminase expression vector pCM101 (30) yielding the hexa-His terminase holoenzyme expression plasmid pQH101. The sequence from the SphI site at 2216 to the end of the A gene was found to be as predicted from the manipulations. We note that protein expression in this cell line is heat-inducible.
Construction of Phage -P1 A hexa-His -A lysogen that expresses hexa-His terminase (-P1 A hexa-His ) in place of the wild-type holoenzyme terminase was constructed as follows. The Sse83871 restriction segment extending from bp 2561 to 2824 (see Fig. 2) and containing the modified segment of A gene (see above) was cloned into the corresponding Sse83871 sites of pJM1 yielding pJM1-hexa-His. pJM1 is a derivative of pBR322 containing DNA extending from bp 44141 through cos to 5505 (Table I). This 9-kilobase sequence includes the late promotor pRЈ, the lysis genes S, R, and Rz, cos, the terminase genes Nu1 and A, and genes W and B. To cross the hexa-His modification of the A gene into phage , pJM1-hexa-His was crossed with -P1 Aam42. Aam42 is an amber mutation in the fifth-to-last codon of the A gene which renders the terminase enzyme inactive (19). Plasmids carrying wild-type (pJM1) and hexa-His-modified (pJM1-hexa-His) A genes were transformed into MF1427 (-P1 Aam42) yielding the lysogens MF1427 (-P1 A wild-type ) and MF1427 (-P1 A hexa-His ), respectively. Selection for transformed lysogens was accomplished by plating at 31°C on L agar plates containing kanamycin and ampicillin. The transformed lysogens were grown overnight in L broth plus antibiotics at 31°C. Prophages were induced to undergo lytic growth by incubation at 42°C for 20 min followed by incubation at 37°C for 60 min. The lysates were treated with CHCl 3 and plated on MF1427 to titer the -P1 A wild-type and -P1 A hexa-His recombinant phages.
Construction of gpA-Hexa-His and Hexa-His-gpNu1 Overexpression Plasmids-A vector that overexpresses the terminase gpA subunit with six histidines fused to the natural C-terminal glutamic acid (see Fig. 2) was constructed by PCR methods previously described (27). The plasmid pSF1 contains wild-type Nu1 and A genes cloned into a pBR322 background (Table I) and was used as a PCR template. The forward primer contained an EcoRI site and was complementary to the first 21 nucleotides of the A gene, except that the GUG initiation codon in A was changed to an AUG codon with the use of the PCR primer. The reverse primer contained a BamHI site, six histidine codons, and the penultimate 18 nucleotides of the A gene. A vector that overexpresses the terminase gpNu1 subunit containing a methionine initiation codon and six histidines fused to the natural N-terminal (initiation) methionine (see Fig. 2) was similarly constructed. The primer sequences used were as follows: gpA forward primer, 5Ј-GAA TTC ATG AAT ATA TCG AAC AGT CAG-3Ј; gpA-hexa-His reverse primer, 5Ј-GGA TCC TCA ATG GTG ATG GTG ATG GTG TTC ATC CTC TCC GAA TAA-3Ј; hexa-His-gpNu1 forward primer, 5Ј-GAA TTC ATG CAC CAT CAC CAT CAC CAT ATG GAA GTC AAC AAA AAG-3Ј; gpNu1 reverse primer, 5Ј-GGA TCC TTA ACC TGA CTG TTC GAT ATA-3Ј. The EcoRI and BamHI restriction sequences are indicated in italics, and the f-MET (forward primers) and stop (reverse primers) codons are shown in bold type. Sequences complementary to the terminase genes are underlined, and the hexa-His codons are double underlined. PCR amplification of the A and Nu1 genes and construction of BL21(DE3)[pH6-A] and BL21(DE3)[pH6-Nu1], the gpA-hexa-His and hexa-His-gpNu1 expressing cell lines, respectively, were performed as described by Meyer et al. (27). We note that protein expression in these cell lines is inducible with IPTG.
Expression and Purification of Hexa-His Terminase Holoenzyme-One liter of 2ϫ YT media containing 25 mM NaH 2 PO 4 , pH 7.2, 1% glucose, and 50 g/ml ampicillin was inoculated with a 10-ml overnight culture of E. coli OR1265[pQH101] and maintained at 30°C until an optical density of 0.6 (600 nm) was obtained. The cells were then heat-induced and harvested as described previously for wild-type holoenzyme (6). The cell pellet was resuspended in 100 ml of 20 mM Tris-HCl buffer, pH 8.0, containing 500 mM NaCl, 1 mM PMSF, 0.4 mg/ml lysozyme, and 10 g/ml aprotinin, and placed on ice for 20 min. The cells were then lysed by sonication, and the insoluble cellular debris was removed by centrifugation (12,000 ϫ g for 15 min, followed by 12,000 ϫ g for 30 min). The clarified supernatant was mixed with 5 ml of Ni-NTA agarose followed by gentle shaking on ice for 1 h. The mixture was loaded into an empty column and unbound protein was eluted with 2 ϫ 4-ml aliquots of 20 mM Tris-HCl buffer, pH 8, containing 500 mM NaCl. Bound protein was eluted in a stepwise fashion with 20 mM Tris-HCl buffer, pH 8.0, containing 500 mM NaCl and increasing concentrations of imidazole (4 ϫ 0.5 ml each, 100, 150, and 250 mM imidazole). Hexa-His terminase holoenzyme eluted in the 250 mM imidazole fractions that were pooled and dialyzed against 20 mM Tris-HCl buffer, pH 8.0, containing 2 mM EDTA, 7 mM ␤-ME, and 50% glycerol. The purified protein sample was stored at Ϫ70°C.
Expression and Purification of gpA-Hexa-His-One liter of 2ϫ YT Derivative of pKKT7(-H) expressing hexaHIS-gpNu1 This work media containing 25 mM NaH 2 PO 4 , pH 7.2, 1% glucose, and 50 g/ml ampicillin was inoculated with a 10-ml overnight culture of E. coli BL21(DE3)[pH6-A] cells, and the culture was maintained at 37°C until an optical density of 0.6 (600 nm) was obtained. The cells were then induced with the addition of IPTG to 1.2 mM, and the culture was maintained at 37°C for an additional 2 h. The cells were harvested by centrifugation, and the cell pellet was resuspended in 100 ml of 25 mM Tris-HCl, pH 8.0, buffer containing 100 mM NaCl, 2 mM EDTA, and 7 mM ␤-ME. Cell lysis was affected as described above for hexa-His terminase holoenzyme, and the clarified supernatant was dialyzed against 50 mM NaH 2 PO 4 buffer, pH 8.0, containing 500 mM NaCl and 25 mM imidazole. Five milliliters of Ni-NTA agarose was added to the protein solution, and the mixture was shaken gently on ice for 1 h. The mixture was then loaded into an empty column, and the unbound protein was eluted with 2 ϫ 4-ml aliquots of 50 mM NaH 2 PO 4 buffer, pH 8.0, containing 500 mM NaCl and 50 mM imidazole. gpA-hexa-His was finally eluted with 7 ϫ 0.5-ml aliquots of 50 mM NaH 2 PO 4 buffer, pH 8.0, containing 500 mM NaCl and 250 mM imidazole. The elution fractions were examined by SDS-PAGE, and the appropriate fractions were pooled and dialyzed against 20 mM Tris-HCl buffer, pH 8.0, containing 100 mM NaCl, 2 mM EDTA, 7 mM ␤-ME, and 50% glycerol. The purified protein sample was stored at Ϫ70°C. Expression and Purification of Hexa-His-gpNu1-Growth and induction of E. coli BL21(DE3)[pH6-Nu1] cells was performed as described above for gpA-hexa-His; however, similar to the wild-type gpNu1 (27,31,32), all of the expressed hexa-His-gpNu1 protein was found in the crude cell lysis pellet (data not shown). The protein pellet was resuspended in 50 ml of 25 mM Tris-HCl, pH 8.0, buffer containing 100 mM NaCl, 2 mM EDTA, and 7 mM ␤-ME and re-pelleted by centrifugation (11,500 ϫ g for 30 min). The washed pellet was next solubilized with 20 ml of 6 M guanidinium HCl (GdmHCl), pH 8.0, gently shaken on ice for 1 h, and insoluble material was removed by centrifugation (10,000 ϫ g for 30 min). Five milliliters of Ni-NTA-agarose was added to the clarified supernatant which was gently shaken on ice for 1 h. The mixture was then loaded into an empty column, and unbound protein was eluted with 2 ϫ 4-ml aliquots of wash buffer (10 mM Tris-HCl, 100 mM NaH 2 PO 4 buffer containing 6 M GdmHCl) at pH 5.9. Nonspecifically bound protein was eluted with 2 ϫ 4-ml aliquots of wash buffer at pH 5.5, followed by 1 ml of wash buffer at pH 5.0. Hexa-His-gpNu1 protein was finally eluted with 6 ϫ 0.5-ml aliquots of the same buffer at pH 4.5. The fractions were examined by SDS-PAGE, and the appropriate fractions were pooled, dialyzed against 25 mM Tris-HCl buffer, pH 8.0, containing 2 mM EDTA, 7 mM ␤-ME, 2.5 M GdmHCl, and 20% glycerol. The concentrated protein samples showed no signs of aggregation upon prolonged storage at Ϫ70°C in the presence of 2.5 M GdmHCl.
Reconstitution of Terminase Holoenzyme-Reconstitution of terminase holoenzyme from the individually purified subunits was accomplished by mixing gpA and gpNu1 (wild-type or hexa-His-tagged) in a 1:2 molar ratio and incubating on ice for 5 min. We note that gpNu1 is stored as a concentrated stock in 2.5 M GdmHCl. Dilution of the protein sample during reconstitution simultaneously diluted the GdmHCl thus allowing re-folding of gpNu1 into an active enzyme complex. Control experiments confirmed that GdmHCl concentrations as high as 50 mM had no effect on the catalytic activities of the enzyme (not shown).
In Vitro DNA Packaging Assay-The in vitro packaging assay was performed as described by Chow et al. (30). The reaction mixture (20 l) contained 10 l of a sonic extract of an induced culture of MF2517 (Table I) in 30 mM Tris-HCl buffer, pH 9.0, containing 10 mM MgCl 2 , 3 mM spermidine, 6 mM putrescine, 7 mM ␤-mercaptoethanol, 1.5 mM EDTA, 1.5 mM ATP, 1.5 nM of mature cI857 Sam7 DNA, and the indicated concentration of either wild-type or hexa-His-tagged terminase holoenzyme. The sonic extract was prepared as described previously (33) and provides proheads, tails, and assembly proteins required for virus assembly. The reaction samples were incubated at room temperature for 30 min to allow the assembly of infectious virus in vitro, and appropriate dilutions were plated on the supF strain MF1968 to determine virus yield.
In Vivo Virus Development (Virus Yield Assay)-MF1427 (IHF ϩ ) or MF1972 (IHF Ϫ ), lysogenized with -P1 carrying either a wild-type A gene (-P1 A wild-type ) or with -P1 carrying the hexa-His-A gene (-P1 A hexa-His ), were grown overnight with aeration in L broth plus kanamycin at 31°C. To determine the number of viable lysogens in each culture, the cultures were diluted into L broth (1:100 dilution) and grown to approximately 2 ϫ 10 7 cells/ml. Then portions of each culture were removed, diluted, and spread on L plates plus kanamycin; the plates were incubated overnight at 31°C. Lysogens were induced by thermal induction at 42°C for 20 min and then incubated at 37°C for 60 min. The lysates were treated with chloroform, clarified, and plated for phage yield on the IHF ϩ strain MF1427.
Terminase Activity Assays-The cos cleavage assay was performed as described previously using pAFP1 as a nuclease substrate (6,34). ATPase catalytic activity was examined as described previously (22). Where indicated, DNA (ScaI-linearized pAFP1) was added to the ATPase assay mixtures at a concentration of 25-50 nM. The concentration of protein used in these assays is indicated in each individual experiment.
Kinetic Analysis-Steady-state ATPase activity was analyzed using linear regression techniques as described previously (6,34). Only data within the linear portion of the reaction time course were used in the analysis. Data from single turnover experiments were analyzed according to both Equations 1 and 2, which describe monophasic and biphasic reaction time courses, respectively.
where Products refer to the ADP formed at time , and A is the extent of the reaction at ϭ ϱ. B and C describe the fraction of the observed rate associated with the slow and fast phases, respectively, and k 1 and k 2 represent the observed rate constants for the slow phase (k slow ) and fast phase (k fast ) of the reaction, respectively. The indicated constants were determined by nonlinear regression analysis of the experimental data using the Igor® data analysis program (Wave Metrics, Lake Oswego, OR) as described previously (6). A mono-exponential curve function was deemed appropriate to describe the data if 1) the values of the rate constants, k 1 and k 2 , obtained by nonlinear regression analysis of the data to Equation 2 differed by less than 10-fold and 2) the 2 value obtained from fitting to Equation 1 was within an order of magnitude of that obtained from fitting to Equation 2.

Construction, Expression, and Purification of Hexa-His-gpNu1, gpA-Hexa-His, and Hexa-His Terminase Holoenzyme-
Previous studies have demonstrated that the C-terminal 38 amino acids of gpA define a functional domain that is required for interaction of complex I with the procapsid (18,19). Interestingly, whereas the penultimate 5 amino acids are strictly required for procapsid binding (19), addition of up to 4 random amino acids to the C terminus of the gpA subunit did not appear to significantly affect the phage yield in E. coli (29). Based upon these data, we reasoned that the addition of six histidines to the C terminus of the gpA subunit might provide a convenient and efficient purification tag with little effect on enzyme function. Vectors that overexpress the hexa-His-tagged gpA subunit, alone and co-expressed with the wild-type gpNu1 subunit, were thus constructed as described under "Experimental Procedures." We further constructed a vector for the expression of an isolated gpNu1 subunit that contains a hexa-His tag at the N terminus. 3 The C-terminal amino acid sequence of gpA-hexa-His and the N-terminal amino acid sequence of hexa-His-gpNu1 are shown in Fig. 2. For the purpose of clarity, we use the term terminase holoenzyme to describe enzyme that was directly purified as a gpA 1 ⅐gpNu1 2 holoenzyme complex from cells simultaneously expressing both enzyme subunits and the term reconstituted terminase holoenzyme to describe enzyme that was prepared by mixing the individually purified subunits.
Both hexa-His terminase holoenzyme and the isolated gpAhexa-His subunit were efficiently expressed in E. coli (Fig. 3) and were found in the soluble fraction of the crude cell lysate (not shown). The proteins were purified from the clarified cell lysate in a single step as described under "Experimental Procedures" and yielded hexa-His terminase holoenzyme (10 mg/ liter cells) and gpA-hexa-His (12 mg/liter cells) preparations that were Ͼ95% homogenous as determined by SDS-PAGE. The isolated hexa-His-gpNu1 subunit was similarly efficiently expressed in E. coli (Fig. 3). However, as is observed with the wild-type protein (31,32,35), hexa-His-gpNu1 was found exclusively in the insoluble cell lysis pellet (data not shown). The protein was solubilized from these inclusion bodies using 6 M guanidinium hydrochloride and purified to homogeneity in a single step using the nickel-chelate column as described under "Experimental Procedures" (29 mg/liter cells, Ͼ95% pure). Hexa-His-gpNu1 has been stored as a concentrated protein solution in 2.5 M GdmHCl for up to 12 months with no evidence of aggregation and with no loss of catalytic activity.
In Vitro Packaging Activity of Wild-type and Hexa-His Terminase Holoenzyme-An in vitro packaging assay was used to examine the biological activity of the purified hexa-His-tagged terminase holoenzyme. This assay utilizes extracts of induced cultures of E. coli MF2517 (Table I) as a source of viral procap-sids, tails, and all the necessary assembly proteins, except for a functional terminase enzyme (30). Addition of viral DNA and terminase holoenzyme allows virus assembly in vitro and yields fully infectious phage that are quantitated by their ability to form plaques (pfu). Fig. 4 shows that both wild-type and hexa-His terminase holoenzymes are biologically active and may be used to assemble infectious virus in vitro. We note, however, that the concentration dependence of pfu formation is slightly greater for hexa-His terminase than for the wild-type enzyme, with Ϸ10-fold more enzyme required for similar phage yields. This difference disappears at enzyme concentrations greater than 10 nM, however, presumably because terminase is no longer limiting in the assay mixture.
ATPase Activity of Wild-type and Hexa-His Terminase Holoenzyme-We have previously identified two ATPase catalytic sites in terminase holoenzyme, a high affinity site in gpA (K m Ϸ5 M) and a low affinity, DNA-stimulated site in gpNu1 (K m Ϸ1,300 and 500 M, minus and plus DNA, respectively) (22,23). Fig. 5A shows the steady-state rate of ATP hydrolysis using an ATP concentration of 1 mM and thus examines ATPase activity of both catalytic sites in the holoenzyme. The figure shows that the ATPase activity of both wild-type and hexa-His-tagged terminase holoenzymes are essentially identical. Moreover, the ATPase activity of both the wild-type and mutant holoenzymes is significantly stimulated by DNA, and the degree of stimulation is virtually identical for each (Fig. 5A, Table II). Furthermore, the steady-state rate of ATP hydrolysis by both enzymes using an ATP concentration of 20 M, a concentration where catalytic activity is predominantly localized within the gpA subunit, is similarly identical (Fig. 5B and Table II).
In order to characterize more fully the effect of the hexa-His purification tag on ATP hydrolysis activity, we next examined the rate of ATP hydrolysis in a single-turnover experiment. In this experiment, the concentration of enzyme and ATP were both 25 nM, and a single catalytic turnover by the enzyme is observed. Furthermore, the low concentration of ATP used ensures that hydrolysis is observed at the high affinity gpA subunit of the enzyme only. Fig. 5C shows that under these experimental conditions, both wild-type and hexa-His terminase holoenzymes hydrolyze ATP with a similar time course and that both proteins are similarly stimulated by DNA. These data were analyzed as described under "Experimental Procedures" yielding the observed rate constants presented in Table  III.
Endonuclease Activity of Wild-type and Hexa-His Terminase Holoenzyme-Terminase holoenzyme possesses a site-specific

FIG. 4. In vitro DNA packaging activity of wild-type and hexa-
His-tagged terminase holoenzyme. The packaging assay was performed as described under "Experimental Procedures" using MF1427 (IHF ϩ ) as the plating bacteria. Phage -P1 A wild-type (q) possesses a wild-type terminase holoenzyme, and phage -P1 A hexa-His (E) possesses a gpA-hexa-His-tagged terminase enzyme.

FIG. 5. ATPase activity of wild-type and hexa-His-tagged terminase holoenzymes.
Circles represent data obtained with wild-type holoenzyme in the absence (E) and presence (q) of DNA. Triangles represent data obtained with hexa-His holoenzyme in the absence (‚) and presence (OE) of DNA. A, steady-state ATP hydrolysis by wild-type (solid line) and hexa-His (dashed line) terminase using an ATP concentration of 1 mM. The concentration of enzyme used in these experiments was 100 nM. B, steady-state ATP hydrolysis by wild-type (solid line) and hexa-His (dashed line) terminase using an ATP concentration of 20 M. Steady-state ATPase assays were conducted as described under "Experimental Procedures" using an enzyme concentration of 100 nM. C, single-turnover hydrolysis of ATP by wild-type (solid line) and hexa-His (dashed line) terminase. ATPase assays were conducted as described under "Experimental Procedures" using enzyme and ATP concentrations of 25 nM. nuclease activity that is required for excision of a single genone from a concatemeric precursor. The site at which terminase assembles and cleaves the duplex is known as cos, an abbreviation for the cohesive end site of the viral genome (see Fig. 1). Fig. 6A shows the results of a cos-cleavage activity assay for wild-type and hexa-His terminase holoenzymes. Whereas both enzymes exhibit significant activity in this assay, the time course for the hexa-His-tagged enzyme lags slightly behind that of wild-type terminase. We have suggested that assembly of the terminase subunits onto DNA is the rate-limiting step in the cos-cleavage reaction (6,34), and these data suggested that the hexa-His-tagged holoenzyme might have impaired DNA binding interactions. In order to explore more fully this possibility, we examined the cos-cleavage activity of these enzymes in the absence of E. coli integration host factor (IHF). IHF, although not strictly required for cos-cleavage activity in vitro or virus assembly in vivo, stimulates these reactions (6,36,37); however, IHF becomes essential for plaque formation when terminase assembly at cos is impaired by mutations in the cos sequence of the genome (26,38). Fig. 6B demonstrates that the cos-cleavage activity of hexa-His terminase is significantly reduced in the absence of IHF, whereas that of wild-type enzyme is only modestly affected. These data support the posit that the hexa-His-tagged holoenzyme is slightly impaired in its interactions with cos-containing DNA.
Catalytic Activity of the Isolated Wild-Type and Hexa-Histagged Terminase Subunits-We (17) and others (39) have previously demonstrated that whereas gpNu1 is devoid of coscleavage activity, the isolated gpA subunit possesses a weak nuclease activity that is strongly stimulated in the holoenzyme complex. Table IV demonstrates that the hexa-His-tagged subunits behave similarly to the wild-type subunits. Moreover, the mutant proteins may be reconstituted into catalytically competent holoenzyme complexes that possess cos-cleavage activity. This is particularly true for the hexa-His-tagged gpNu1 subunit that may be reconstituted into an enzyme complex that is fully active compared with wild-type reconstituted enzyme (Table IV). Consistent with the cos-cleavage activity of wild-type and hexa-His terminase holoenzymes, terminase reconstituted with a hexa-His-tagged gpA subunit possesses nuclease activity that is slightly impaired when compared with wild-type reconstituted enzyme (see Table IV and Fig. 6A).
Studies on the ATPase activity of the isolated terminase subunits have similarly revealed that the isolated gpNu1 subunit possesses weak ATPase activity while gpA efficiently hydrolyzes ATP (17,34,40). Consistently, the isolated hexa-His gpNu1 subunit possesses modest ATP hydrolysis activity, whereas hexa-His gpA hydrolyzes ATP at a rate similar to that of the wild-type subunit ( Table V). Reconstitution of terminase with wild-type gpA and hexa-His gpNu1 subunits affords a holoenzyme with ATPase activity that is virtually identical to that obtained with a fully wild-type enzyme (Table V). Interestingly, however, reconstitution of the isolated gpA-hexa-His subunit with either wild-type or hexa-His-gpNu1 yields an enzyme complex that does not possess ATP hydrolysis activity beyond that of the isolated gpA-hexa-His subunit alone (Table V).
In order to characterize more fully the ATPase activity of the isolated gpA-hexa-His subunit, a single-turnover kinetic analysis of ATP hydrolysis was performed, and the results are presented in Fig. 7 and Table VI. The wild-type gpA subunit hydrolyzes ATP with a time course that is well described by a single exponential curve function (Fig. 7A) and with an observed rate that is similar to that observed for terminase holoenzyme in the absence of DNA (compare Tables III and VI). This confirms that at this concentration of ATP, hydrolysis in the holoenzyme is limited to the gpA subunit of the enzyme.  6. Endonuclease activity of wild-type and hexa-Histagged terminase holoenzyme. cos-Cleavage activity of wild-type (q) and hexa-His terminase (OE) in the presence (A) and absence (B) of IHF. The assay was conducted as described under "Experimental Procedures" using an enzyme and DNA concentration of 400 and 100 nM, respectively. IHF was included at a concentration of 100 nM as indicated.
Addition of DNA to the reaction mixture strongly stimulates the ATPase activity of the wild-type gpA subunit and the rate of ATP hydrolysis is more appropriately described by a doubleexponential curve function. Analysis of these data yield the fast and slow rate constants presented in Table VI. Interestingly, the rate constant for the fast phase of the reaction is similar to that observed for terminase holoenzyme in the presence of DNA (compare Tables III and VI), suggesting that at this concentration of ATP, the observed stimulation of ATP hydrolysis in holoenzyme is mediated by DNA interactions with the gpA subunit.
Unlike wild-type protein, single-turnover ATP hydrolysis by the isolated gpA-hexa-His subunit exhibits biphasic behavior, even in the absence of DNA (Fig. 7B). Moreover, the rate constants obtained from analysis of these data are quite similar to the k slow and k fast rate constants obtained for wild-type gpA in the presence of DNA (Table VI). Only a small proportion of the time course (14 -22%) is attributable to k fast for the hexa-His-tagged protein, however, and unlike the wild-type protein, gpA-hexa-His is unresponsive to DNA (Fig. 7B and Table VI).
The Effect of Hexa-His Terminase on in Vivo Virus Development--Pl A hexa-His was constructed by crossing phage -P1 Aam42 (Table I) and plasmid pJM1-hexa-His as described under "Experimental Procedures." The Aam42 mutation is a lethal amber mutation located in the 5th-to-last codon of the A gene. Since there are only 4 codons between the Aam42 muta-tion in the phage and the six His codons of the modified A gene in the plasmid, it was expected that virtually all plaque-forming A ؉ recombinant phage would contain the hexa-His modification. 4 This was confirmed by sequencing studies on the -Pl A hexa-His recombinants, which directly demonstrated the presence of the hexa-His modification. The yield of plaque-forming (A ϩ ) recombinants was 1.43 ϫ 10 7 and 1.62 ϫ 10 7 pfu/ml in crosses with pJMI-A wild-type and pJM1-A hexa-His plasmids, respectively. A control cross with no plasmid yielded less than 1 ϫ 10 4 pfu/ml, indicating that revertants of the -P1 Aam42 phage were not contributing significantly to the titers of the cross-lysates. Since the frequencies of plaque-forming recombinants were essentially the same for each plasmid, we conclude that terminase with the hexa-His modification is functional in virus development.
Effect of IHF on -Pl A wild-type and -Pl A hexa-His in Vivo Virus Development-The distinct requirement for IHF in cos-cleavage by hexa-His terminase holoenzyme (Fig. 6) suggested that phage development in vivo might exhibit a similar requirement for IHF. This is especially important as the hexa-His purification tag might be expected to weaken interactions with the procapsid and thus require increased stability of complex I to ensure that progression toward active DNA packaging would occur. We thus examined the requirement for IHF on in vivo phage development by -P1 A wild-type and -P1 A hexa-His , viruses that express wild-type and hexa-His-tagged terminase holoenzymes, respectively (Table I). Initial studies demonstrated that although -P1 A hexa-His formed normal plaques an IHF ؉ host, the mutant phage formed minute pinpoint plaques on an IHF-deficient host (data not shown). To quantify these effects, virus burst studies were performed using -P1 A wild-type and -P1 A hexa-His lysogens in IHF ϩ and IHF Ϫ hosts. Table VII shows that the yield of -P1 A hexa-His was, within experimental error, identical to that of -P1 A wild-type when an IHF ϩ strain was used. This result is similar to that observed in the in vitro DNA packaging (Fig. 4) and cos-cleavage (Fig. 6) assays, where only modest effects are observed between the wild-type and hexa-His-tagged holoenzymes in the presence of IHF. Whereas both phages showed significantly reduced yields in the IHF Ϫ host, the deficit was much more pronounced for -P1 A hexa-His (Table VII). In fact, the observed decrease in burst size -P1 A hexa-His was sufficient to lower virus yield to a level just above that required for plaque formation on an IHF Ϫ host, thus yielding minute pinpoint plaques. DISCUSSION The terminase enzyme from bacteriophage lambda is responsible, at least in part, for the insertion of viral DNA into a procapsid (7)(8)(9). Our laboratories are interested in the genetic, biochemical, and biophysical aspects of phage assembly and, specifically, the nucleoprotein complexes required for genome packaging. These studies require a simple, rapid, and efficient purification scheme for the isolation of large quantities of wildtype and mutant terminase holoenzymes, as well as the isolated enzyme subunits. The use of terminase in the biotechnology industry further underscores the need for simple and efficient purification protocols. Several vectors have been developed for the expression of terminase holoenzyme and the individual subunits in E. coli (27,30), and several purification schemes have been published over the years (4,6,23,31,32). Even the most efficient protocols are relatively laborious and time consuming, however, for the routine purification of large quantities of enzyme. Purification of hexa-His-tagged proteins  using nickel-chelate chromatography has become routinely used for rapid and efficient protein purification, and we thus constructed vectors that express hexa-His-tagged holoenzyme and the individual terminase subunits. All of the vectors described here express significant quantities of protein, and our purification protocols allow their purification in one step from the crude cell lysate. Importantly, we have developed a method to purify the highly insoluble terminase gpNu1 subunit in a single step from the crude cell lysis pellet in the GdmHCldenatured state. This protein may be stored as a concentrated stock solution that, upon dilution of the GdmHCL, refolds into an active conformation. Expression from the vectors described above yield proteins that possess a hexa-His purification tag directly attached to the C-terminal glutamic acid of gpA or the N-terminal methionine of gpNu1 (see Fig. 2). Most vectors that express hexa-Histagged proteins provide a cleavage site that allows the removal of most, but not all, of the purification tag. Previous studies have suggested that addition of 4 amino acids to the C terminus of the gpA subunit did not significantly affect in vivo viral development (29), and we reasoned that the 6 histidines at the C terminus of gpA-hexa-His would similarly be tolerated. Our expression vectors were thus constructed to add a minimal purification tag that would remain part of the purified protein.
This concept provides a simple method to purify the proteins and avoids a proteolysis step that would add time and significant expense to the purification procedure.
Although the addition of 4 random amino acids to the C terminus of the gpA subunit did not affect terminase function in vivo, it was necessary to confirm that addition of the hexa-His purification tag similarly did not affect the catalytic activities of the enzyme. Initial experiments demonstrated that the tag only modestly affected phage development in vivo and phage assembly in vitro. A more detailed investigation is required, however, if these proteins are to be used for mechanistic studies on the enzyme, and we next examined the ATPase activity of the modified enzyme. Both steady-state and singleturnover kinetic experiments demonstrated that ATP hydrolysis by terminase holoenzyme is little affected by introduction of the hexa-His purification tag into the gpA subunit. Importantly, both enzymes hydrolyze ATP with identical rates, and both enzymes are similarly responsive to DNA. Similarly, only modest differences were observed between the enzymes in our standard cos-cleavage activity assay; however, further investigation demonstrated that hexa-His terminase holoenzyme has a mild DNA-binding defect that is masked by IHF, an E. coli protein known to stabilize terminase-DNA interactions. This result was also apparent in experiments that similarly showed a significant requirement for IHF in the in vivo development of a hexa-His terminase-containing phage.
The mild defect in hexa-His terminase holoenzyme was significantly magnified in the isolated gpA-hexa-His subunit. Unlike wild-type protein, gpA-hexa-His exhibited biphasic ATPase kinetics in the absence of DNA and was completely unresponsive to the addition of polynucleotide. These data suggest that while gpA-hexa-His in a holoenzyme complex is relatively "normal," the isolated subunit is more severely impaired, particularly in its interactions with DNA. Moreover, the isolated gpA-hexa-His subunit does not appear to interact appropriately with gpNu1 to form a "natural" holoenzyme complex. Although the nuclease activity of terminase reconstituted from the gpA-hexa-His subunit is near wild type, this reconsti-  a The data were equally well described by a single-exponential curve function, and the single rate constant is presented.

TABLE VII
Effect of IHF on phage development in vivo MF1427 and MF1972 are wild-type (IHF ϩ ) and IHF deficient (IHF Ϫ ) cell lines, respectively (see Table I). Phage -P1 A wild-type possesses a wild-type terminase, whereas phage -P1 A hexaHIS possesses a gpAhexa-His-tagged terminase. tuted holoenzyme does not possess ATPase activity beyond that observed with the isolated enzyme subunits. Conversely, however, terminase reconstituted from wild-type gpA and a hexa-His-tagged gpNu1 subunit yields a catalytically competent holoenzyme complex with wild-type nuclease and ATPase activities.
During the course of this investigation, we have uncovered an interesting aspect of ATP hydrolysis by terminase holoenzyme. Previous studies have suggested that the steady-state rate of ATP hydrolysis by the gpA subunit in terminase holoenzyme was unaffected by DNA and that DNA-mediated stimulation of ATPase activity occurred primarily at the low affinity, gpNu1 ATP-binding site of the enzyme (22). Contrary to these earlier results, however, the data presented here demonstrate that DNA directly stimulates ATP hydrolysis at the high affinity gpA subunit of the enzyme. Steady-state kinetic experiments performed at an ATP concentration of 20 M, a concentration well below the K m for ATP binding by the gpNu1 subunit (500 M), show a 4 -5-fold increase in the rate of ATP hydrolysis with the addition of DNA to the reaction mixture. Virtually identical results were obtained in the single-turnover experiments under conditions where ATP hydrolysis at the low affinity gpNu1 catalytic site is expected to be minimal. The above experiments were performed with terminase holoenzyme, however, and it was feasible that DNA interactions with the gpNu1 subunit were responsible for the observed increases in catalytic activity. Single turnover experiments with the isolated gpA subunit were thus performed to provide additional mechanistic insight. The single-turnover rate constant for ATP hydrolysis by the wild-type gpA subunit was virtually identical to that of the holoenzyme, confirming that ATP hydrolysis at the gpNu1 subunit under these conditions is essentially zero. Addition of DNA to gpA resulted in the appearance of biphasic kinetics and the introduction of a fast rate of ATP hydrolysis. Interestingly, the rate constant for the fast phase of the reaction was quite similar to the rate constant obtained with the holoenzyme, again suggesting that DNA stimulates ATP hydrolysis directly at the gpA subunit. However, although a single, fast rate was observed with the holoenzyme, only Ϸ50% of the time course may be ascribed to the fast phase with the isolated gpA subunit. These data suggest that protein-protein interactions with gpNu1 in the holoenzyme complex are required for full expression of DNA stimulation at the gpA subunit. The mechanistic implications of the biphasic kinetics observed in these single-turnover experiments, and the protein-protein interactions affecting these kinetic profiles is currently under investigation in our laboratory.
In conclusion, we have constructed vectors that efficiently express hexa-His-tagged terminase holoenzyme and the individual enzyme subunits, and we have developed simple purification protocols that allow their purification in one step from the crude cell lysates. The catalytic activities of the tagged holoenzymes is, at first glance, indistinguishable from the wildtype enzyme. Moreover, phages that express these mutant proteins efficiently replicate in vivo, perhaps the most stringent test of biological activity. Upon closer examination, however, the catalytic properties of these proteins reveal subtle defects. The primary defect appears to be in the interaction of gpA-hexa-His with DNA. Although most apparent with the isolated subunit, effects are also observed with the hexa-Histagged holoenzyme. The deficiency can be overcome with the addition of IHF to the reaction mixture, and these proteins will find utility in a number of experimental systems; however, the results reported here underscore that a complete biochemical characterization of the effects of purification tags on enzyme function must be performed prior to their use in mechanistic studies.