Designing a nine cysteine-less DNA packaging motor from bacteriophage T4 reveals new insights into ATPase structure and function

Abstract

The packaging motor of bacteriophage T4 translocates DNA into the capsid at a rate of up to 2000 bp/s. Such a high rate would require coordination of motor movements at millisecond timescale. Designing a cysteine-less gp17 is essential to generate fluorescently labeled motors and measure distance changes between motor domains by FRET analyses. Here, by using sequence alignments, structural modeling, combinatorial mutagenesis, and recombinational rescue, we replaced all nine cysteines of gp17 and introduced single cysteines at defined positions. These mutant motors retained in vitro DNA packaging activity. Single mutant motors translocated DNA molecules in real time as imaged by total internal reflection fluorescence microscopy. We discovered, unexpectedly, that a hydrophobic or nonpolar amino acid next to Walker B motif is essential for motor function, probably for efficient generation of OH⁻ nucleophile. The ATPase Walker B motif, thus, may be redefined as “β-strand (4–6 hydrophobic-rich amino acids)–DE-hydrophobic/nonpolar amino acid”.

Introduction

DNA packaging in most tailed bacteriophages and herpes viruses is a complex process involving recognition of a concatemeric viral genome, generation of an end by endonucleolytic cleavage, and translocation of the DNA into an empty prohead (Black and Rao, 2012; Casjens, 2011; Rao and Feiss, 2008). A packaging machine consisting of a dodecameric portal and a pentameric motor (in phages T4, T7, and phi29) assembled at the special five-fold vertex of the capsid carries out this process. After filling the head with one or slightly more than one viral genome (headful packaging), a second cut is made to terminate packaging. The motor-DNA complex dissociates from the full head and assembles on another empty prohead, whereas the portal provides a platform for assembly of neck and tail proteins to generate an infectious virion.

In phage T4 the motor protein is encoded by the large terminase protein gp17. It contains all the catalytic functions necessary for packaging including nuclease, ATPase, and translocase activities (Alam et al., 2008; Baumann and Black, 2003; Leffers and Rao, 2000; Rao and Black, 1988). A small terminase protein, gp16, is also required but it is dispensable in vitro. It aids in the recognition of the viral genome and regulates gp17 functions (Al-Zahrani et al., 2009). The atomic structures of full-length gp17, individual gp17 domains, and the central oligomerization domain of gp16 have been determined. gp17 consists of two domains, an N-terminal ATPase domain and a C-terminal nuclease/translocase domain (Sun et al., 2008) (Fig. 1a). The ATPase domain has two subdomains. The larger subdomain (NsubI) contains Walker A, Walker B, catalytic carboxylate, and adenine binding motifs that are involved in ATP binding and hydrolysis. The smaller subdomain (NsubII) is a regulatory (transmission) domain, connecting the ATPase catalytic center to the nuclease and translocation sites of the C-domain via a flexible hinge. This architecture is conserved in other phage and herpes virus DNA packaging motor proteins (Feiss; and Rao, 2012; Nadal et al., 2010; Roy and Cingolani, 2012; Smits et al., 2009; Zhao et al., 2013).

Fig. 1.

Domain organization and positions of cysteine residues on the phage T4 large terminase gp17 polypeptide chain (a). The positions of cysteine residues are marked with vertical red bars. Various domains and functional motifs are shown. (b) Alignment of large terminase sequences from five T4-like phages. The cysteines present in phage T4 large terminase and the corresponding residues in other large terminases are bolded and highlighted in yellow. The numbers shown above refers to the cysteines of phage T4 gp17. Various functional motifs and the critical residues corresponding to that motif are shown in the same color. The Walker B motif residues are boxed. (c) Schematic of SOE-PCR mutagenesis. Cysteine residue numbers shown in blue are mutated to serine in the first round of PCR to obtain the 5cys mutant. The PCR fragments (blue bars) amplified with mutant primers and the stitch PCRs to fuse the fragments to construct the 5cys (M1) template are shown (See Materials and Methods for details). Using this template, the residues shown in black were mutated to serine to obtain the 9cys mutant (M2) in the second round. (For interpretation of the references to color in this figure legend, the reader is referred to the web version of this article).

An electrostatic force driven mechanism was proposed for phage T4 DNA packaging (Sun et al., 2008), which was supported by mutational studies and force measurements (Migliori et al., 2014). Each gp17 subunit of the pentameric motor is proposed to alternate between two conformational states, “extended” (relaxed) and “compact” (tensed). In the extended state, ATP binds to the N-domain and DNA to the C-domain. ATP hydrolysis and charge repulsion between the products ADP (3⁻) and Pi (3⁻) causes NsubII to rotate by 6°, aligning the charge pairs present at the interface of N- and C-domains. This generates electrostatic force attracting the C-domain-DNA complex towards the N-domain, resulting in ~7 Å (2 bp) movement of DNA into the capsid (compact state). Release of products causes NsubII to rotate back to the original position, misaligning the charge pairs and returning the C-domain to the extended state. DNA is then captured by another gp17 subunit and this translocation cycle is repeated.

Single molecule optical tweezers studies show that the phage T4 packaging motor is the fastest and most powerful reported to date (Fuller et al., 2007). Given a packaging rate of up to ~2000 bp/s, the catalytic processes at the active sites and the accompanying movements of the motor parts must be orchestrated in the timescale of milliseconds. We have recently developed a single molecule fluorescence assay to visualize DNA translocation in real time using total internal reflection fluorescence microscopy (TIRF) (Vafabakhsh et al., in press). Individual packaging machines immobilized on a glass surface were shown to translocate DNA molecules into the capsid, one after another. This system could be adapted to measure distance changes between domains, on the order of ~10–40 Å, by installing acceptor and donor fluorescent probes at defined positions of the gp17 molecule and quantifying fluorescence resonance energy transfer (FRET) during DNA translocation. But it would require covalent modification of the −SH group of appropriately placed cysteine(s) with the maleimide-tagged fluorescent dye molecules such as cy3 and cy5 (Heyduk, 2002; Rasnik et al., 2004). This means that all nine cysteines of gp17 must be replaced and then new cysteine(s) must be introduced at desired positions. At minimum 10–11 different mutations must be introduced while at the same time retaining correct protein folding, solubility, and DNA packaging, a significant challenge considering that the cysteines are distributed across the 610-amino acid sequence of gp17 (amino acids 113, 125, 158, 179, 257, 402, 453, 481 and 500). Moreover, some of these cysteines are strictly conserved (Fig. 1b) or located at structurally vulnerable positions of the molecule (see below).

Here, guided by bioinformatics and structural modeling, and by employing a combination of splicing by overlap extension (SOE), combinatorial mutagenesis, and recombinational genetic rescue, we describe the design of cysteine-less and single cysteine gp17 mutants. Biochemical analyses showed that these mutants retained ATPase and DNA packaging activities but lost the nuclease activity. We further demonstrate, by using our newly developed single molecule fluorescence assay, that the mutants like the wild-type (WT) initiate and re-initiate DNA packaging, although the frequency of reinitiation was reduced in the mutants. We found, unexpectedly, that the cysteine-257 residue that is located directly adjacent to the conserved catalytic glutamate of Walker B motif (Mitchell and Rao, 2006) is essential for function. Substitution with a polar or charged amino acid, including a conservative serine, resulted in loss of function whereas substitution with a structurally unrelated nonpolar or hydrophobic amino acid retained function. We speculate on the implications of this finding which potentially expands the definition of the canonical Walker B motif.

Results

Replacement of all the nine cysteines of gp17 with serine or alanine results in loss of DNA packaging activity

To construct the cysteine-less gp17 mutant, we first replaced each of the nine cysteine codons with either serine or alanine codons by SOE-PCR mutagenesis (see Materials and Methods for details). In the first round, the cysteines at positions 113, 158, 257, 453 and 500 were replaced by amplifying five contiguous pieces of gene 17 DNA. Appropriate mutations were introduced into the primers. These were stitched together through a series of SOE-PCRs (Horton et al., 1989) to generate the 5cys (M1) mutant (Fig. 1c). The 5cys mutant was then used to construct the 9cys mutants (M2 and M3) by replacing the rest of the four cysteines at amino acids 125, 179, 402, and 481. This required 9 additional PCRs including 4 stitches. Both the 5cys and 9cys mutants showed poor solubility. However, we could purify significant amounts of 5cys gp17 in soluble form but it showed no detectable DNA packaging activity (Table 1). These results indicated that the 5cys and 9cys mutants did not fold correctly.

Table 1.

Cysteine-less mutants of gpl7, their solubility and DNA packaging activity. Substitutions made based on data from combinatorial mutagenesis (Fig. 2b and c) are underlined. Solubility was assessed using B-PER reagent (Thermo Scientific). Bulk DNA packaging assays were carried out as described in Materials and Methods.

gpl7 cysteine-less mutants	Solubility	DNA packaging
M1 (5cys mutant) [C113S] [C158S] [C257S] [C453S] [C500S]	<10% Soluble	No activity
M2 (9cys mutant) [C113S] [C125S] [C158S] [C179S] [C257S] [C402S] [C453S] [C481S] [C500S]	Insoluble	ND
M3 (9cys mutant-1) [C113A] [C125A] [C158S] [C179S] [C257S] [C402S] [C453S] [C481S] [C500S]	~10%	Not active
M4 (9cys mutant-2) [C113A] [C125A] [C158S] [C179S] [C257G] [C402S] [C453S] [C481S] [C500S]	~10 %	Not active
M5 (9cys mutant-3) [C113A] [C125A] [C158S] [C179S] [C257A] [C402S] [C453S] [C481S] [C500S]	~10%	Not active
M6 (9cys mutant-4) [C113A] [C125A] [C158S] [C179S] [C2571] [C402S] [C453S] [C481S] [C500S]	~10%	Not active
M7 (9cys mutant-5) [C113A] [C125A] [C158S] [C179S] [C257F] [C402S] [C453S] [C481S] [C500S]	~10%	Not active
M8 (9cysM1) [C113A] [C125A] [C158A] [C179A] [C257A] [C402S] [C453S] [C481S] [C500A]	~20%	~18%
M9 (9cysM2) [C113A] [C125S] [C158A] [C179S] [C257A] [C402S] [C453S] [C481S] [C500S]	~10%	~5%
M10 (9cysM3) [C113S] [C125A] [C158S] [C179A] [C257A] [C402S] [C453S] [C481S] [C500S]	Insoluble	~
M11 (9cysM4) [C113S] [C125S] [C158S] [C179S] [C257A] [C402S] [C453S] [C481S] [C500A]	Insoluble	~
M12 (9cys-C158V) [C113A] [C125A] [C158V] [C179A] [C257A] [C402S] [C453S] [C481S] [C500A]	~20–30%	~30%
M13 (9cys-C158H) [C113A] [C125A] [C158H] [C179A] [C257A] [C402S] [C453S] [C481S] [C500A]	~20–30%	~50–80%
M14 (9cys-C158F) [C113A] [C125A] [C158F] [C179A] [C257A] [C402S] [C453S] [C481S] [C500A]	~20–30%	~50–60%
M15 (9cys-C158Y) [C113A] [C125A] [C158Y] [C179A] [C257A] [C402S] [C453S] [C481S] [C500A]	~20–30%	~60–80%

Substitution of cysteine 257 with serine results in loss of DNA packaging activity

The lack of DNA packaging activity with the 5cys mutant suggested that at least one of the five cysteines at amino acids 113, 158, 257, 453 or 500 must be critical for function. By a process of elimination, we predicted that Cys257 might be the critical residue.

Of the nine cysteine residues, sequence alignments showed that the cysteines at amino acids 158, 179, 453, and 481 are not conserved (Fig. 1b). Our previous mutagenesis experiments showed that Cys402 is not essential because several substitutions including serine were tolerated at this position (Rentas and Rao, 2003). This leaves four cysteines at amino acids 113, 125, 257, and 500 that are well conserved and substitution of one or more of these with serine resulted in loss of function.

Of these, Cys257 is immediately adjacent to the critical Walker B residues, Asp255 and catalytic carboxylate Glu256 (Figs. 1b and 2a) (Goetzinger and Rao, 2003; Mitchell and Rao, 2006). Previous structural and biochemical analyses suggested that Asp255 coordinates with Mg^{2 +} −ATP complex through a bound water molecule, whereas Glu256 activates a water molecule for nucleophilic attack on the γ-phosphate of the bound ATP (Erzberger and Berger, 2006; Sun et al., 2007). Examination of the gp17 X-ray structure (Sun et al., 2008; Sun et al., 2007) showed that the Cys257 forms a hydrogen bond with another critical residue, Thr285, one of the ATP hydrolysis sensor residues (Draper and Rao, 2007), which in turn interacts with Glu256 (Fig. 2a). Changing this cysteine to a functionally conserved serine residue, often used in mutagenesis studies, was not expected to abolish function. However, the –SH group has lesser electronegativity and dipole moment than the –OH group, hence the S–H bond is less polar than the O–H bond. Therefore, it is conceivable that if the catalytic center is sensitive to electrostatic potential of its microenvironment, this change might affect catalysis. In addition, the orientation of Asp255 and Glu256 carboxylates could be altered by the introduction of −OH group, which could reconfigure the above hydrogen bond network.

Fig. 2.

X-ray structure of phage T4 gp17 ATPase domain complexed with ATP. The region containing the Cys158 and Cys257 residues and their contacts is enlarged (a). The phenotypes of amino acid substitutions at positions Cys257 (b) and Cys158 (c). Amino acid substitutions shown below the native residue (red) are null mutants and the ones above (green) are functional substitutions. “sp” refers to small plaque phenotype. (For interpretation of the references to color in this figure legend, the reader is referred to the web version of this article).

To analyze the functional significance of Cys257, we constructed a combinatorial library by replacing the cysteine with every possible amino acid and transferring each mutation (randomly) into T4 genome and testing for plaque forming ability by recombinational rescue using the T4 Tyr253am mutant phage (see Materials and Methods for details). About half of the Cys257 mutants were functional and the other half null. Some of these clones have been sequenced to determine the amino acid substitution corresponding to the respective phenotype, thereby establishing the phenotypes for fourteen of the twenty amino acid substitutions (Fig. 2b). Analysis of these results confirmed that serine substitution is indeed lethal. Importantly, the data showed that only hydrophobic or nonpolar amino acid substitutions were tolerated (Fig. 2b). These include aromatic (Phe, Tyr), hydrophobic (Val, Ile, Ala), or small (Gly) amino acids. Substitution with a polar or charged amino acid such as Ser, Asp, or Arg resulted in null phenotype and substitution with Asn resulted in a small (minute) plaque phenotype (Fig. 2b). Proline substitution resulted in null phenotype which was expected because its rigid pyrrolidine ring would have caused significant structural perturbation at this catalytically sensitive region.

Four different 9Cys recombinant mutant plasmid clones (M4-M7) were then constructed by substituting Cys257 with Gly, Ala, Ile, or Phe, all being functional substitutions by genetic assay. Though these mutant proteins exhibited better solubility (~10%), they aggregated and eluted near the void volume following size-exclusion chromatography (Superdex 200) and none retained the DNA packaging activity (Table 1).

Substituting some of the serines with alanine partially restored the DNA packaging activity

Using the 9Cys-C257A mutant (M5) as a template, we then introduced alanines at positions 113, 125, 158, and 500 in various combinations while retaining serine at positions 402, 453, and 481. A total of 12 mutant clones were constructed and some of these (e.g., M8-M11, Table 1) have been screened for solubility followed by purification of monomeric gp17 and retention of DNA packaging activity. Only two of these mutants, M8 and M9 (Table 1), purified as monomers but the yields were quite low. Both retained low levels of DNA packaging activity, M8 mutant ~18% of WT activity and M9, ~5% activity.

Substitution of cysteine 158 with tyrosine restored most of the DNA packaging activity

Structural analyses showed that Cys158 is in the middle of a β-strand and it forms a hydrogen bond with Thr286 from the adjacent β-strand, both of these strands being part of the six-stranded parallel β-sheet (Rossmann fold) (Fig. 2a). Thr286 is part of the previously identified ATPase coupling motif, TTT_285–287, which is critically required for DNA Translocation (Draper and Rao, 2007). Substitutions at Cys158 could affect folding and/or function of gp17, possible reasons for low solubility and poor packaging activity of M8 and M9 mutants. To determine which amino acid substitution(s) would be optimal at Cys158, we constructed a combinatorial library and determined the tolerant and null substitutions following recombinational rescue using the T4 Ser161am mutant phage (Rao and Mitchell, 2001) (Fig. 2c). Based on this profile, five different mutant clones were constructed by substituting Cys158 with Ala, Thr, Val, Phe, or Tyr, all being functional substitutions. All the five mutant proteins were purified and tested for DNA packaging activity (M11-M15; Table 1). Of these, the C158Y mutant was found to be the best variant as it retained 60–80% of the DNA packaging activity and was stable to repeated freezing and thawing (Table 1).

The cysteine-less mutants also retained the gp16-stimulated ATPase activity [Fig. 3a; note the hydrolysis of γ³²P-ATP (spot at the bottom) and appearance of ³²P (spot at the top) in WT as well as M8 and M15 mutants] but the nuclease activity was lost (Fig. 3b and c; note the appearance of DNA smear upon cleavage in vivo and in vitro by WT and K577 gp17s but not by M15 mutant). The loss of nuclease activity might be due to substitution of the strictly conserved Cys500 residue which makes intra-helix contacts with Lys496 that is part of the DNA binding groove of the nuclease catalytic site (Ghosh-Kumar et al., 2011; Sun et al., 2007; Sun et al., 2008). However, since the nuclease activity is not essential for DNA packaging, and since it might be advantageous not to have the nuclease activity which could otherwise cleave the DNA substrate, we did not attempt to restore the nuclease function by further mutagenesis.

Fig. 3.

The cysteine-less gp17 mutants retain gp16-stimulated ATPase activity but lost nuclease activity. (a) The gp16-stimuated ATPase activity of the cysteine-less gp17 mutants M8 and M15 (Table 1) was determined using purified proteins (~0.5 or 1 mM) at a ratio of one gp17 monomer to one gp16 oligomer. (b) In vivo nuclease assay showing a DNA smear as a result of cleavage of the plasmid DNA by WT and K577 protein but not by the cysteine-less M15 protein. 0, 60, and 120 at the bottom of the figure represent minutes after IPTG induction of gp17 expression in E. coli. (c) In vitro nuclease activity was determined by incubating the purified WT, K577, and M15 gp17 proteins (1 or 2 mM, as shown at the bottom of the figure) with phage λ DNA. "C" represents untreated control DNA. See Materials and methods for the details of the assays.

Introduction of single cysteines into the functional 9Cys mutant

The functional 9Cys mutant M15 (Fig. 4a) was used as a template to introduce single cysteines at specific positions of the gp17 molecule. Several solvent accessible non-conserved residues at positions that could show distance changes during translocation were selected as targets for substitution with cysteine. Twelve such mutant clones were constructed, the mutant proteins purified and screened for DNA packaging activity. Of these, two mutants (R517C and G358C; Fig. 4b–d) retained up to 60–80% of the DNA packaging activity (Fig. 4e). However, the yield of the proteins was quite low because most of the purified protein aggregated and eluted near the void volume of the gel filtration profile. Only a small peak was seen at the monomeric position, the active form of gp17 (the aggregate fraction had no in vitro DNA packaging activity). The aggregation was probably because the exposed –SH group formed non-specific inter-molecular disulfide bonds. Consistent with this prediction, addition of strong reducing agents such as 5 mM DTT (dithiothreitol) or 2 mM TCEP (Tris 2-carboxyethyl phosphine) to the purification buffers (Liu et al., 2010) had shifted most of the aggregated protein to the monomeric position.

Fig. 4.

Structural models of cysteine-less and single cysteine gp17 mutants (a). The models were created with Pymol using the gp17 X-ray structure (Protein Data Bank ID: 3CPE). The amino acid substitutions at various positions of the molecule are shown. Cysteines were replaced with alanine (red), serine (yellow) or tyrosine (blue). (b and c) Single cysteine mutants showing R517C (orange) and G358C (purple) substitutions. (d) SDS-polyacrylamide gel showing the purified gp17 proteins. Lanes: std (molecular weight standards); 1 (WT); 2 (R517C); and 3, (G358C). The purified gp17 shows 2-3 bands because of non-specific proteolysis at the C-terminal end (Goetzinger and Rao, 2003). (e) Bulk packaging assay using ladder DNA as substrate. The packaging activity of the gp17 mutants was compared with the WT gp17 used in each experiment. (f) Single molecule fluorescence assay using 45-bp Cy5 labeled dsDNA. (g) Average number of fluorescent spots per imaging area (70 μm × 35 μm) for each gp17 preparation. Error bar represents 7 standard error of mean of 15 different imaging areas. (h–j) Fluorescence intensity time traces of single packaging machines using WT (H), R517C (I), or G538C (J) gp17 packaging motors. The inset in H shows an example of increase in intensity as a discrete step due to packaging of one DNA molecule and a discrete drop in intensity due to photobleaching of a single fluorophore. (For interpretation of the references to color in this figure legend, the reader is referred to the web version of this article).

Single mutant motors assembled with single cysteine gp17 mutants packaged DNA

To evaluate if the single cysteine gp17 mutants are functional in our recently developed single molecule fluorescence assay (Vafabakhsh et al., in press), individual packaging machines were assembled using these mutant proteins and analyzed for DNA translocation in real time by total internal reflection fluorescence microscopy (TIRF). Emptied phage heads (Zhang et al., 2011) preassembled with purified WT gp17 or single cysteine R517C or G358C mutant proteins were flowed in (see Materials and Methods for details). After the unassembled gp17 was washed off, ATP and Cy5-labeled 45-bp dsDNA were introduced into the chamber. Fluorescent spots appeared due to packaging of the Cy5 fluorophore labeled DNA into the head (Fig. 4f and g) (buffer lacking ATP showed only a few background spots, Fig. 4f). The fluorescence intensity of each spot increased with time in discrete steps (e.g., inset to Fig. 4h, 1), each step due to packaging of a new Cy5 DNA molecule (Fig. 4h–j). This was confirmed by discrete drops in the intensity traces (e.g., inset to Fig. 4h, 2) upon photobleaching of the Cy5 fluorophores. These data demonstrated that the single cysteine mutant motors like their WT counterpart can repeatedly initiate DNA packaging and can translocate a series of oligonucleotide molecules into the capsid, one after another. The mutant motors, however, showed lower DNA packaging activity (Fig. 4g) and fewer packaging initiations (Fig. 4h–j) when compared to the WT motors.

Discussion

In this study, by incorporating multiple approaches; sequence alignments, structural modeling, SOE and combinatorial mutagenesis, and genetic rescue, we constructed functional cysteine-less and single cysteine mutant motors that package DNA. Such a multi-pronged approach was essential, first to replace all nine cysteines of gp17, then to find the right amino acid to replace with, in order to retain correct folding and DNA packaging activity. As demonstrated, single mutant motors could be imaged in real time, showing translocation of a series of DNA molecules. The mutant gp17 proteins also retained the ATPase activity, as would be expected because this activity provides energy for DNA translocation (Sun et al., 2008). But the nuclease activity was lost presumably due to substitution of the conserved Cys500 residue which might be important for the orientation of Lys496, one of the residues lining the nuclease DNA groove. These approaches serve as a model to design mutant motors for single molecule fluorescence studies, not only for the phage T4 motor but also for motors from the other well-studied phages such as phi29 and λ (Chemla; and Smith, 2012; Chistol et al., 2012; Tsay et al., 2009). These would also provide a framework to measure distance changes between motor domains during DNA translocation.

The ATPase motif is considered to be one of the most ancient motifs in biological systems (Saraste et al., 1990; Thomsen and Berger, 2008). The Walker B box, a critical component of this motif, contains a β-strand followed by an aspartate that coordinates with the Mg–ATP complex and a catalytic glutamate that activates a water molecule for nucleophilic attack on the γ-phosphate of ATP (Erzberger and Berger, 2006). The carboxylate moiety of glutamate is oriented into the catalytic pocket, at a distance of few Å from the γ-phosphate (Goetzinger and Rao, 2003; Sun et al., 2007; Story et al., 1993). An unexpected finding of our studies was that the function of the amino acid residue next to catalytic glutamate is also critical. This was not evident in numerous sequence alignments we and others have done over the years because the residue itself is not conserved. The conservation was at the level of hydrophobicity of the sidechain but not its structure. Substitution with polar or charged amino acids including a conservative serine (extensively used in mutational studies) resulted in null phenotypes. This means that the serine motor containing a perfectly good canonical Walker B motif is defective. To determine if this restriction was unique to the T4 packaging ATPase or was broadly seen in other ATPases, we have analyzed several hundreds of terminase and helicase sequences in the database and found that it is, indeed, a conserved feature of these ATPase motors (Draper and Rao, 2007; Fairman-Williams et al., 2012). As summarized in Table 2, a hydrophobic or nonpolar amino acid is found after the DE residues of Walker B motif in >93% of terminases and >99% of helicases.

Table 2.

The amino add distribution of the residue corresponding to the Cys257 position next to the catalytic glutamate of Walker B motif in gpl7 (Fig. 1b) was determined from sequence alignment of the viral terminases (Draper and Rao, 2007) or of the helicases (Fairman-Williams, Guenther, and Jankowsky 2012).

Amino acid	Terminases (out of 247)	Helicases (out of 208)
Ala	49 (20%)	130 (63%)
Val	36 (15%)	19 (9%)
Ile	30 (12%)	14 (7%)
Leu	30 (12%)	3 (1%)
Tyr	30 (12%)	3 (1%)
Cys	19 (8%)	15 (7%)
Phe	17 (7%)	6 (3%)
Pro	11 (4%)	2 (1%)
Glu	5 (2%)	2 (1%)
Gly	5 (2%)	12 (6%)
Met	4 (2%)	1 (0.5%)
Thr	4 (2%)	0
Ser	3 (1%)	0
His	1 (0.4%)	0
Trp	1 (0.4%)	0
Arg	1 (0.4%)	0
Asp	1 (0.4%)	0
Gln	0	1 (0.5%)
Asn	0	0
Lys	0	0

What is the functional significance of a hydrophobic amino acid next to the catalytic glutamate of the Walker B motif? First, a hydrophobic environment might be essential in the catalytic space surrounding the two charged residues (DE) of the Walker B motif and the ATP γ-phosphate. The β-strand residues of Walker B motifs, MIYI_251–254 in the case of gp17 (Fig. 1b, boxed), are hydrophobic. Our previous mutagenesis studies showed that introduction of a charged amino acid into this stretch, e.g., M251R or Y253K, resulted in a null phenotype (Mitchell and Rao, 2006). The hydrophobic environment would exclude solvent water molecules from this catalytic space which otherwise might interfere with the function of the bound water molecules. The aspartate bound water molecule coordinates with the Mg–ATP complex and the glutamate bound water molecule attacks the γ-phosphate (Story et al., 1993). Second, the presence of hydrophobic or non-polar amino acid next to glutamate might be essential for efficient activation of the water molecule. Recent molecular dynamics simulations showed that the carboxylate moiety of glutamate accepts a proton from the water molecule thus generating OH⁻ nucleophile that attacks the β–γ phosphoanhydride bond of ATP (Afanasyeva et al., 2014). A polar or charged amino acid might interfere with this process as it can form a donor H-bond with water oxygen rather than an acceptor H-bond. A hydrophobic amino acid, on the other hand, would decrease the negative electrostatic potential thus reducing the activation barrier for transfer of proton to glutamate.

Our previous studies showed that the phage T4 packaging motor packages DNA about 7-times faster than the phi29 motor (Fuller et al., 2007; Kottadiel, Rao, and Chemla, 2012). This is consistent with the fact that the T4 motor needs to package about 171 kb genome, ~8.5 times longer than that of phi29, in the same amount of time to generate an infectious virus (Fuller et al., 2007). How is the motor speed regulated to fit the “life-style” of a virus? We speculate, based on the above analysis, that introduction of appropriate polar, charged, or hydrophobic amino acids into the catalytic space of Walker B motif could modulate the rate of ATP hydrolysis and in turn, the speed of the motor. Mutant motors with 3–10 times reduced speeds than the WT have been reported with the large terminase gpA of phage λ and the dsDNA helicase SpoIIIE of Bacillus subtilis (Burton et al., 2007). These mutations were found to be localized in the hydrogen bonding network of the Walker B and ATPase coupling motifs. The T4 gp17 Walker B mutants I252S (Mitchell and Rao, 2006) and C257N that affected this network (Fig. 2a) and resulted in small/minute plaque phenotypes might belong to this category. We plan to do biochemical analyses on these mutants as well as attempt to rationally design motors with different speeds using this principle.

In conclusion, we designed a T4 DNA packaging motor with ten different mutations spread across the molecule yet retaining the DNA packaging function. Mutational data uncovered a new pattern in the Walker B motif, the presence of a hydrophobic or nonpolar amino acid next to the catalytic glutamic acid. That this feature is conserved in hundreds of terminases and helicases means that the definition of the canonical Walker B motif might need to be expanded to: “β-strand (4–6 hydrophobic-rich amino acids)–DE-Z” where “Z” represents a hydrophobic or nonpolar amino acid. Our analysis of the phenotypes provide a mechanistic basis, however speculative, to design motors with different speeds by manipulating the hydrophobic/electrostatic environment in the vicinity of Walker B glutamate and ATP γ-phosphate.

Materials and methods

Construction of cysteine-minus gp17 mutant clones

The gene 17 DNA fragments were amplified by PCR using purified phage T4 DNA as template and appropriate primers. Each of the nine native cysteine of gp17 was replaced with serine by PCR-directed SOE strategy (Horton et al., 1989; Rao and Mitchell, 2001) (Fig. 1c). In the first round, five cysteines at positions 113, 158, 257, 453 and 500 (highlighted in blue in Fig. 1c) were replaced using appropriate primers. PCR amplified DNA fragments obtained (amino acids 1–113, 113–158, 158–257, 257–453, 453–500, 500 – 610) were then stitched progressively; the 1–113 fragment was stitched with the 113–158 fragment to obtain the 1–158 fragment which was then stitched with the 158–257 fragment to obtain the 1–257 fragment and so on (see Fig. 1c). This resulted in the 5cys mutant M1 (Table 1). The M1 DNA was then used as a template for replacing the remaining four cysteines with serine in a similar manner to produce the 9cys mutant M2 (Table 1). Both M1 and M2 clones were tested for protein expression by first transforming into E.coli XL10 Gold cells and then into the expression hosts BL21 (DE3) pLysS and BL21 (DE3) RIPL (Stratagene, CA; see below).

The M2 mutant DNA was used as template to introduce alanines at various positions (e.g., mutant M3 which has alanine at positions 113 and 125). The DNA template of M3 was then used for the construction of a combinatorial library at Cys257 (Fig. 2b). Based on the phenotypic analyses using the Tyr253am mutant T4 phage, several mutants with a functional amino acid substitution at Cys257 were constructed and over-expressed (mutants M4-M7, Table 1). The mutant proteins were purified and DNA packaging activity was assessed. Using the mutant M7 as template, alanine was again introduced at various positions (mutants M8-M11, Table 1). Since M8 showed small but significant DNA packaging activity and better solubility but was unstable, a combinatorial library of mutants at Cys158 was constructed using the mutant M8 mutant plasmid DNA as template. Functional substitutions were determined by recombinational marker rescue using the Ser161am mutant T4 phage (Fig. 2c). Four of these mutations tested for expression, solubility, purification, and DNA packaging (Table 1). Mutant M15 (C113A-C125A-C158Y-C179A-C257A-C402S-C453S-C481S-C500A) that has a substitution of tyrosine at position Cys158 was selected as it showed better solubility and 60–80% of WT DNA packaging activity (Table 1).

For all the mutant constructions, the amplified DNAs were purified by agarose gel electrophoresis, digested with NheI and XhoI restriction enzymes (these restriction sites were included into the terminal primers), and ligated with the pET-28b plasmid DNA linearized with NheI and XhoI enzymes. In the ligated DNA, the N-terminal end of the mutants would have fused in frame with the hexa-histidine tag of the vector. As a result, the each mutant protein produced will contain a hexa-histidine tag for purification by Niagarose chromatography. The ligated DNAs were first transformed into E. coli XL10 Gold cells (Stratagene, CA) and mini-prep plasmid DNAs were prepared by the alkaline lysis procedure. The inserted DNA was sequenced in its entirety to ensure 100% accuracy of the mutant clones (Macrogen). The clones were then transferred into the expression strains, E. coli BL21(DE3) pLys-S or E. coli BL21 (DE3) RIPL (Studier et al., 1990).

Construction of combinatorial libraries

The SOE PCR strategy was used to construct combinatorial libraries at residues Cys257 and Cys158 using appropriate primers. The amplified gene 17 mutant DNA was digested with NheI and XhoI and ligated with pET28b DNA linearized with NheI and XhoI enzymes. The recombinant DNA was then transformed into E. coli XL10 Gold cells. About 100 random mutant colonies from each library were grown and the mutations were transferred to phage T4 by recombinational marker rescue using the previously described amber mutant phages in gene 17, Tyr253am (for the Cys257 library) and Ser161am (for the Cys158 library) (Mitchell and Rao, 2006; Rao and Mitchell, 2001). Each mutant was spot-tested and phenotypically scored as functional (lysis on the spot), null (no plaques or lysis), or small plaque (sp) (small to minute plaques) (Rao and Mitchell, 2001). Plasmid clones corresponding to a few of the functional and null phenotypes, and the only clone from the Cys257 library that gave rise to sp phenotype were sequenced to determine the amino substitution that gave rise to the respective phenotype (Fig. 2b and c). This established the phenotypes for fourteen and nine of the twenty amino acid substitutions at Cys257 and Cys158, respectively.

Purification of gp17 mutant proteins

The E. coli cells containing the recombinant gp17 mutant clones were induced with 1 mM IPTG at 30 °C for 2 h to over-express the His-tagged proteins. Cells were harvested by centrifugation at 8200g for 10 min at 4 °C and were lysed using French Press. Cell-free extracts was prepared by centrifugation of the lysate at 38,700g for 20 min at 4 °C. The supernatant containing soluble protein was purified by successive chromatography on HisTrap HP (affinity), Mono Q-5/50 GL (ion-exchange), and Hiload Superdex 200 prepgrade (size exclusion) columns using AKTA-PRIME and AKTA-FPLC systems (GE Healthcare) (Kanamaru et al., 2004). For single cysteine mutants, since most of the protein aggregated by forming nonspecific inter-molecular disulfide bonds, the Mono Q fractions were treated with immobilized TCEP gel [the reducing agent tris(2-carboxyethyl) phosphine, Thermo scientific] at 0.5 mg/ml for 30 min at 4 °C. The immobilized TCEP gel was removed by centrifugation followed by filtration through a sterile 0.2 μ filter before proceeding to size-exclusion chromatography. The purified proteins were concentrated by Amicon Ultra centrifugal filters (Millipore), aliquoted, and stored at −70 °C. The proteins were ~95% pure as judged by SDS-polyacrylamide gel electrophoresis and Coomassie blue R staining.

Bulk in vitro DNA packaging assay

In vitro DNA packaging experiments were carried out as previously described (Kondabagil, Zhang, and Rao, 2006). The purified WT gp17 or the cysteine mutants (1 μM) and purified 17am18amrll proheads (~2 × 10⁹ particles) were incubated in a reaction mixture (20 μl) containing 50 mM Tris–HCl (pH 7.5), 0.1 M NaCl, 5 mM each of MgCl₂, spmeridine, and putriscine, 5% (w/v) polyethylene glycol, 1 mM ATP, and 300 ng of ladder DNA (Fermentas) for 45 min at 37 °C. DNase I (Sigma) was added to a final concentration of 0.5 μg/μl and incubated for 30 min at 37 °C. Reaction was stopped and deproteinized by the addition of 50 mM EDTA, and 0.5 μg/μl of proteinase K (Fermentas) plus 0.2% SDS and incubation for 30 min at 65 °C. The samples were subjected to 0.8% (w/v) agarose gel electrophoresis for 2–3h at 100 V. DNA was stained with ethidium bromide and the amount of DNA packaged was quantified by Eagle Eye II (Biorad) imaging system.

ATPase assay

ATPase assays were performed according to Leffers and Rao (2000). The purified gp17 proteins either alone or with gp16 (at a ratio of one gp17 monomer to one gp16 oligomer) were incubated in a reaction mixture (20 μl) containing 1 mM cold ATP and 5 μCi of [γ-³²P] ATP (Sp. Act. 3000 Ci/mmol) at 37 °C in ATPase buffer (50 mM Tris–HCl pH 7.5, 0.1 M NaCl and 5 mM MgCl₂) for 20 min. The reactions were terminated by 50 mM EDTA followed by thin layer chromatography and phosphorimaging (Storm 820, Molecular dynamics).

Nuclease assay

E. coli BL21 (DE3) pLys-S cells containing gp17 clones were grown at 30 °]C to 4 × 10⁸ cells/ml in Moore's medium containing ampicillin (50 μg/ml) and chloramphenicol (37 μg/ml). One ml samples at 0, 60, and 120 min after induction with 1 mM IPTG were withdrawn and the plasmid DNA was isolated by the alkaline lysis procedure. An aliquot of the plasmid DNA was electrophoresed on a 0.8% (w/v) agarose gel (Alam et al., 2008). K577 is a truncated form of gp17 in which the C-terminal 33 amino acids were deleted. This protein retains all the activities of the full-length gp17 and was used as another positive control in the nuclease assays.

In vitro nuclease activity was determined by incubating the purified WT, K577, and M15 mutant proteins (1 or 2 mM) with phage λ DNA for 30 min at 37 °C. The samples were electrophoresed on a 0.8% (w/v) agarose gel.

Single molecule DNA packaging

Single molecule fluorescence experiments were performed on a wide-field prism-type total internal reflection microscope with a 630 nm laser (Melles Griot) for Cy5 excitation (Vafabakhsh et al., in press). Clean quartz slides and glass cover slips were surface-passivated with PEG and 3% biotinylated PEG (Laysan Bio). The slide was treated in succession with NeutrAvidin (Thermo Scientific) (0.01 mg/ml), biotinylated protein-G (Rockland Immuno-chemicals) (25 nM), and polyclonal anti-T4 antibody (15 nM). Preassembled packaging complexes were prepared by mixing empty phage heads (~5 × 10⁹ particles)(Zhang et al., 2011), 1 μM gp17, 1 mM ATPγS, and 200 nM priming DNA (120 bp dsDNA) in 1 × PEG buffer and incubating the mixture for 20 min on ice (5 μL reaction volume). The complexes were then incubated with the T4 antibody-coated slide for 30 min and then the chamber was washed with T50-BSA buffer (10 mM Tris–HCl (pH 7.5), 50 mM NaCl and 0.1 mg/mL BSA) containing 1 mM ATPγS. The assembled packaging machines were imaged in the imaging buffer containing 1 mM ATP and 2.5 nM dsDNA (5′–CA/iCy5/C TCG TCC AGC AGATAG AAGTCA CAG CGG ATC CTATAG ACA GAG-3′). Each spot in Fig. 3f represents packaging by a single packaging machine. Single molecule movies were acquired and analyzed to yield the single molecule traces (Fig. 3h–j), using an in-house software. Packaging efficiency was quantified by determining the average number of fluorescent spots per area (70 μm × 35 μm) from 15 different imaging areas for each sample (Fig. 3g).

Abbreviations

FRET

fluorescence resonance energy transfer

TIRF

total internal reflection fluorescence

SOE

splicing by overlap extension

WT

wild-type