Characterization of the Prohead-pRNA Interaction of Bacteriophage 429"

The small prohead RNA (pRNA) of the Bacizzus subtizis bacteriophage 429 is essential for ATP-dependent packaging of viral DNA. The 174-, 124-, and 120-residue forms of pRNA produced in uitro using T7 RNA polymerase were equivalent in prohead binding and DNA packaging activity to pRNAs produced in 429-infected cells. pRNA binding to proheads, characterized by the use of North- ern hybridization and filter binding assays, was specific, rapid, and irreversible in the presence of 10 m~ M e . Proheads produced in phage-infected cells carried 5.8 2.7 copies of pRN& and proheads assembled in Escherichia coli in the absence of pRNA bound 6.0 * 3.5 copies of pRNA. Footprints of proheads on pRNA generated with the ribonucleases& T1, and V1 showed that nucleo- tides 22-84, 5' to 3', were protected from ribonuclease attack. Enhanced cleavage at nucleotides 3740 with ri- bonuclease V1 suggested a conformational change of pRNA upon prohead binding.

heads as a 120-residue molecule, can be detached from proheads and reattached with concomitant loss and restoration of DNA.gp3 packaging activity in the defined in vitro system (Guo et al., 1987c;Wichitwechkarn et al., 1989). The full-length molecule is 174 residues, and 54 residues are removed from the 3' end by adventitious nucleases during the prohead purification (Wichitwechkarn et al., 1989). The prohead is composed of the major head protein (gp8), the removable scaffolding protein (gp7), the head fibers (gp8.5), and the portal protein or connector (gpl0) that serves as the attachment site for pRNA and as a unique vertex for DNA.gp3 packaging. Purified proheads contain approximately six copies of the 174-residue pRNA by Northern hybridization analysis, and in vitro DNA packaging activity in the defined system is maximal when RNA-free proheads are reconstituted with six copies of pRNA (Wichitwechkarn et al., 1989). The purified 120-residue pRNAmolecule also binds to the purified connector in vitro and is sufficient to ' * This work was supported by grants 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.
The nucleotide sequence(s1 reported in this paper has been submitted to the GenBank-IEMBL Data Bank with accession number(s1 X05973.
pRNA secondary structure has been determined by phylogenetic analysis and is organized into two domains (Bailey et al., 1990). The larger 5' domain (domain I), composed of 117 residues and containing four helices, is necessary and sufficient for DNA packaging.
The ATPase activity of gp16 is dependent on DNA.gp3 and proheads in the defined in vitro DNA packaging system (Guo et al., 1987b). pRNA or purified proheads with pRNA stimulate the ATPase activity of gp16 in the absence of DNA.gp3, and the stimulation by proheads is pRNA-dependent (Grimes and Anderson, 1990).
Proheads with truncated pRNAs have been isolated and used to correlate pRNA size with DNA.gp3 packaging activity (Grimes and Anderson, 1989b). Cleaving pRNA alters the specificity to package left-and right-end restriction fragments of DNA.gp3. Residues 1-25 of pRNA contribute to a domain for DNA.gp3 interaction, while residues 26-49 contribute to a domain for prohead binding.
The role of pRNA in $29 DNA packaging remains an intriguing question. To facilitate the study of pRNA structure and function, an in vitro T7 transcription system was developed that produces pRNA in quantity. The binding of pRNA to RNAfree proheads was characterized by Northern hybridization and nitrocellulose filter binding, and RNase footprinting was used to determine pRNA contacts with proheads.
Alternatively, proheads were produced in the absence of pRNA in E. coli strains HMS174(DE3) or BL2UDE3) containing the isopropyl-p-Dthiogalactopyranoside-inducible plasmid pAR7-8-8.5-10 that encodes prohead structural proteins (Guo et al., 1991a). Cells were grown in Luna-Bertani broth (Maniatis et al., 1982) containing 100 pg/ml ampicillin at 37 "C with shaking to an A,,, of 1.0. Isopropyl-~-~-thiogalactopyranoside was added to 0.5 mM, and incubation was continued for 3 h at 37 "C. Cells were concentrated 100-fold by centrifugation and resuspended in 50 m~ Tris-HC1 (pH 7.5),100 m~ NaCl and 10 m~ MgC1, (TMS buffer) containing 30 unitdml RNase-free DNase I (Boehringer Mannheim). The cells were disrupted by sonication for 6-8 min at 0 "C, the extract was clarified twice by centrifugation at 12,000 x g for 10 min a t 4 "C, and the proheads were purified by centrifugation in a 10-30% linear sucrose density gradient in a Beckman SW 28 rotor at 25,000 rpm for 5 h a t 4 "C. Proheads in sucrose were diluted with TMS buffer, pelleted by centrifugation in a Beckman 50.2 rotor a t 35,000 rpm for 5 h, and resuspended in TMS buffer. The proheads were purified further on a second 10-3092 sucrose gradient, and the prohead concentration was determined by protein assay (Bio-Rad) and the use of a prohead mass of 12.1 MDa.Z Preparation of pRNA-pRNA was purified from B. subtilis 12A(pUM102) by electrophoresis in denaturing urea-acrylamide gels as described (Wichitwechkam et al., 1992). Individual pRNA bands were located by W shadowing over polyethyleneimine/UV,,,-cellulose plates (Alltech Associates, Inc.) with a 254-nm light source. pRNA bands were excised and the pRNAeluted twice by diffusion into 500 m~ ammonium acetate (pH 6), 0.1 m~ EDTA, and 0.1% SDS for 3-6 h at 37 "C.
Alternatively, pRNA was produced by in vitro T7 transcription from plasmid pRT72. The transcription template was constructed by sitedirected mutagenesis (Kunkel, 1985(Kunkel, , 1987 of plasmid pBluelO2 (Wichitwechkam et al., 1992). pBluelO2 is pBluescript KS(+) with the +29 pRNA gene on a 320-base pair fragment and the pRNA promoter in the same orientation as the pBluescript T7 promoter. The mutagenic oligonucleotide 5'-GCACTCACTATAGGAATGGTACGGTACTTC-3' was used to delete 123 base pairs of plasmid DNA between the pBluescript T7 consensus promoter and the pRNA transcription start site. The 30-mer juxtaposed the -12 to +2 sequence of the T7 promoter, numbering relative to the start of transcription, and the +3 to +18 sequence of the pRNA gene. The resulting plasmid, pRT7De1, was mutagenized with the 38-mer 5'-GTGCACGCTAC?TTCCTAAGATCTl'ACATGCGA-CACAG-3' to insert a DdeI site a t +120 and a BglII site at +124 with respect to the T7 promoter and produce pRT71. This also changed 116C-+G and 117T+G in the pRNA gene to complement the changes 1A-C and 2G.C from the first round of mutagenesis and maintain pRNA secondary structure (Bailey et al., 1990). Then pRT71 was mutagenized with the 30-mer 5'-CTGTGTCG'ITTl'AAA"TG'll'CATl" GACAAA-3' to insert a DraI site at +174 and create pRT72. Clones were identified aRer each round of mutagenesis by DNA sequencing or restriction mapping. Restriction digests with DraI (Boehringer Mannheim) and with DdeI and BglII (Life Technologies, Inc.) were performed according to the manufacturers' instructions.
Mertens, and D. Anderson, unpublished results. tRNA. Proheads bound 2.3 and 1.9 pRNA in TM buffer without and with tRNA, and 1.2 and 0.1 pRNA in TE buffer without and with tRNA, respectively.
pRNA was 5'4abeled with T4 polynucleotide kinase (Life Technologies, Inc.) and [y-"P]ATP (DuPont NEN) as described (Wichitwechkam et al., 1989) after removal of the 5'4erminal phosphate with calf intestine alkaline phosphatase (Boehringer Mannheim). Alternatively, pRNA was labeled by incorporating [a-"PIATP during in vitro T7 transcription. Reactions were prepared as above except that 25 pCi of [C~-~~P]ATF' (DuPont NEN), 0.18 pmol of template, and 50 units of T7 RNA polymerase were added to a 50-1. 11 reaction containing 10 1.1~ ATP and 500 p~ each of GTP, CTP, and UTP. Incubation was for 30 min at 37 "C. pRNA was purified by gel electrophoresis as described above, and bands were located by autoradiography.
Binding ofpRNA to Proheads-Binding of labeled pRNA to proheads was determined quantitatively by nitrocellulose filter binding as described (Carey et al., 1983). Briefly, lo5 cpm of ["PIpRNA was diluted into 50 m~ Tris-HC1 (pH7.5) and 10 mM MgClz (TM buffer), heated to 75 "C for 3 min, cooled quickly to 0 "C, and mixed with RNA-free proheads in 50 pl of TM buffer. After incubation for 10 min a t 25 "C, samples were diluted to 500 pl in TM buffer and loaded onto prewetted 25-mm diameter cellulose nitrate filters (Micro Filtration Systems) under constant suction in a filtration apparatus. After three washes of 2 ml each with TM buffer, the radiolabel retained on the filter was measured by scintillation counting.
For dot blot hybridization, 140 pg of RNA-free proheads were reconstituted with 8 pg of 174-residue pRNA (12 pRNNprohead) in TMS buffer for 15 min at room temperature. Proheads were separated from unbound pRNA by centrifugation in a linear 5-20% sucrose gradient, concentrated by centrifugation in the SW 55 rotor a t 35,000 rpm for 5 h at 4 "C, and quantified by protein assay (Bio-Rad). Hybridizations were performed as described (Wichitwechkam et al., 1989).
Proheads were then diluted to 5 ml with TM buffer a t 4 "C, separated from unbound pRNA by centrifugation to pellet the particles as described above, and the fraction of pRNA bound determined by filter binding assay.
DNA Packaging Reactions-In vitro DNA packaging assays were performed as described (Grimes and Anderson, 1989a). Briefly, purified proheads or proheads reconstituted with pRNA were mixed with purified DNA.gp3 in reaction buffer containing 10 m~ ATP, 6 m~ spermidine, and 3 m~ 2-mercaptoethanol in TMS buffer. The ATPase gp16 in 6 M guanidinium chloride was renatured by a 25-fold dilution into 2 m~ Tris (pH 7.5), 0.4 m~ CHAPS, and 5 m~ dithiothreitol for 40 min on ice. Reaction mixtures containing proheads, DNA.gp3, and gp16 (2:1:50) were incubated for 30 min a t room temperature. Unpackaged DNA was digested with 5 pg/ml DNase I, the DNase inactivated with 10 m~ EDTA, and the packaged DNA extracted from filled heads for 60 min at 65 "C and quantified by agarose gel electrophoresis.
RESULTS AND DISCUSSION In Vitro pRNA Synthesis and Activity-In vitro transcription by T7 RNA polymerase can provide large quantities of RNA for biochemical and biophysical analyses (Lowary et al., 1986;Studier et al., 1990). Three rounds of oligonucleotide-directed mutagenesis on plasmid pBluel02rev, which contains the pRNA gene, produced the in vitro transcription template pRT72 (see "Experimental Procedures"). First, the pRNA promoter was deleted to align the T7 promoter for pRNA expression. Second, changes of nucleotides 1A-4 and 2G+C in the pRNA gene were made to accommodate requirements of the T7 polymerase (Milligan et al., 1987;Chapman and Burgess, 1987;Dunn and Studier, 1983). Nucleotides 1 and 2 of pRNA are at the end of a helix (see Bailey et al. (1990) and Fig. 51, so the compensating mutations 116C+G and 117T-+G were made in the pRNA gene to restore base pairing in the transcripts. Finally, because a 3"truncated 120-nucleotide form of the 174residue pRNA is sufficient for packaging (Guo et al., 1987a(Guo et al., , 1987c, three restriction sites were inserted to permit the production of 174-, 124-, or 120-nucleotide pRNA. The prohead binding and in vitro DNA packaging activity of each of these pRNAs on proheads was indistinguishable from wild-type pRNA (data not shown). Moreover, the changes in pRNA nucleotides 1 and 2, along with the compensating changes at 116 and 117 to maintain secondary structure, did not affect pRNA function.
Characterization of pRNA Binding to Proheads-Northern hybridization and filter binding assays were used to determine the extent, specificity, rate, and ionic requirements of [32P]pRNA binding to proheads stripped of pRNA with EDTA or to proheads assembled in E. coli in the absence of pRNA. Northem hybridization was used to determine the extent of pRNA binding to proheads, and the results of a typical experiment are illustrated in Fig. 1. Proheads produced in phage-infected cells carried 5.8 2 2.7 ( n = 8) copies of pRNA. Proheads treated with EDTA retained 2.6 2 1.0 ( n = 6) copies of pRNA, and after reconstitution these proheads carried 5.4 2 3.2 ( n = 2) pRNM prohead. Proheads assembled in E. coli bound 6.0 2 3.5 ( n = 5 ) pRNMprohead. The pRNA content or binding per prohead was more variable than the value of 5.5 2 0.9 pRNMprohead determined by electron microscopic counts of particles (Wichitwechkarn et al., 1989). Possibly proheads of some preparations can bind more than the mean of 5-6 pRNA. The demonstration that proheads produced in E. coli in the absence of pRNA and pro- [32P]pRNA-prohead complexes (+) were treated with no enzyme (lunes 1 and 2), 5 x pg/ml RNase A (lunes 3 and 4), 1 x pg/d RNase   A (lanes 5 and 6), 2 x lo-* pg/ml RNase A (lanes 7 and 8). or 4 x lo-* pg/ml RNase A(lunes 9 and 10). Markers were generated by sequencing 5' end-labeled pRNA as per the manufacturer's instructions (Pharmacia).
Between 60 and 80% of the input pRNA was competent to bind to proheads. Similar variability in the fraction of RNA that can bind a protein has been observed with other small RNAs, such as the R17 coat protein binding sequence (Carey et al., 1983). The unbound pRNA was intact, and therefore structural variations may exist among pRNA molecules.
Varying amounts of proheads were incubated with a constant amount of [32P]pRNA in TM buffer or 50 mM Tris-HC1 (pH 7.5), 10 mM EDTA (TE buffer) to determine the specificity of pRNA binding in the presence or absence of Mg2+, with or without competitor tRNA (Fig. 2). In the presence of 10 mM Mg2+ each prohead bound 2.3 pRNA without tRNA and 1.9 pRNA with tRNA as determined by linear regression analysis. The relationship between proheads added and pRNA bound was linear, demonstrating stoichiometric binding, although the extent of binding was generally lower than that observed in the hybridization experiments, perhaps due to lability of the RNA-free proheads upon storage. Proheads in TE buffer bound 1.2 pRNN prohead without tRNA and 0.1 pRNNprohead with excess tRNA. Thus proheads have a general affinity for RNA in the absence of Mg2' but bind pRNA specifically in the presence of To test for reversibility of binding, [32P]pRNA (116 pmoVml) was incubated with proheads (23.2 pmoVml), and the filter assay showed 1.5 mol pRNA boundmol of proheads. Unlabeled pRNA (1.16 nmoVml) was then added, and the amount of labeled pRNA bound was assayed over 1 h. No decrease in the amount of labeled pRNA bound was observed, showing that bound pRNA did not exchange with unbound pRNA (data not shown). Additionally, the prohead-pRNA complex was stable upon dilution and ultracentrifugation.
To determine the rate of pRNA binding, proheads produced in E. coli (48 pmol/ml) were incubated with pRNA (510 pmoV ml) and tRNA (59 nmoVml), and the amount of pRNA bound was determined at 3-min intervals. Proheads bound approximately 1.3 pRNA within 3 min at 0, 25, or 37 "C, with no increase thereafter, and approximately 80% of the ultimate binding occurred within 1 min (data not shown). The amount of pRNA bound per prohead varied over time of storage of the proheads at -20 "C and also among prohead preparations.
Ionic requirements for pRNA binding were determined by varying NaCl or MgC12 concentrations in binding mixtures containing proheads (54 pmol/ml), pRNA (420 pmoVml), and tRNA (8.2 nmoVml). Varying NaCl concentration from 0 to 300 mM with 10 mM MgC12 had no effect on the amount of pRNA bound (data not shown). When MgCI2 concentration was varied from 0 to 25 mM, the amount of pRNA bound increased with increasing MgCI2 concentration to 10 mM and then remained constant.

'
tected residues, and the hatched box rep-

U-A
and irreversibly in the presence of 10 mM M e . Mg2' has both structural and catalytic roles for RNA enzymes (for review see Yarus (1993)). Mg2+ likely stabilizes pRNA secondary and tertiary structure to allow for specific binding to proheads.
pRNA bound to proheads was protected from RNase A digestion a t residues 25-84,5' to 3', and residues 4-19 and 110-120 were not protected (Fig. 3). Footprints produced with RNase V1 showed protection of nucleotides 22-69, with no protection of nucleotides 12-16 or from nucleotide 93 to the 3' end. Enhanced cleavages of pRNA with RNase V1 occurred at nucleotides 37-40. RNase T1 digests of pRNA-prohead complexes showed protection of residues 51-82, supporting the results shown with RNases A and V1 (Fig. 4). Cleavages by the singleor double-strand specific RNases agreed generally with previous digests of pRNA alone that supported the model of secondary structure (Bailey et al., 1990). The composite ribonuclease footprint presented in Fig. 5 shows that a 63-nucleotide region that subtends residues 22-84 within the 120-residue domain of pRNA interacts with the prohead. The footprinting results are consistent with the results of Grimes and Anderson (1989b), which showed that a 95-residue pRNA molecule that included nucleotides 26-120 bound to proheads and that further truncations of the 5' end to nucleotide 50 abolished prohead binding. Since nucleotides 26-49 constitute approximately one-half of the prohead footprinted region, deletion of these residues would be expected to prevent prohead binding.
The enhanced cleavage at residues 37-40 by RNase V1 might be due to enrichment of cleavable pRNA from a pool of heterogeneous molecules by prohead binding, since unbound pRNA was removed prior to ribonuclease treatment. However, similar enhanced cleavages at residues 37-40 by RNase V1 were observed when unbound pRNA was not removed (data not shown). Therefore, irreversible binding of pRNA to proheads apparently stabilized pRNA secondary structure or resulted in a conformational change of pRNA to allow enhanced cleavage by RNase V1. Some background cleavage a t nucleotides 17, 18, 39, and 40, likely due to contaminating nucleases, was noted upon addition of pRNA to proheads (Figs. 3 and 4, lane 2 ); this cleavage was reduced by a second gradient purification of the prohead preparations (data not shown).
The 63-residue prohead binding region of 429 pRNA seems relatively large when compared to protein interactive sites of certain other RNAs, for example, HIV TAR RNA (Churcher et al., 1993) and the bacteriophage R17 coat protein binding sequence (Carey et al., 1983). This may be due in part to the fact that ribonuclease probes are subject to steric constraints for cleavage near the prohead binding site. Other evidence sug-gests that much of the sequence is essential for prohead binding. A number of oligonucleotide-directed mutants within the 63-nucleotide prohead footprint region of pRNA have been pro-d~c e d .~ pRNAs of several deletion mutants, which together cover 44 residues of the footprinted region, do not compete with wild-type [32P]pRNA for binding to proheads when added in a 10-fold molar excess.
Our goal is to analyze the higher order structure of pRNA as it interacts with the 429 portal protein and the ATPase gp16, both of which have RNA recognition motifs (Grimes and Anderson, 19901, to constitute the DNA packaging machine.