Mapping of subunit-subunit contact surfaces on the beta' subunit of Escherichia coli RNA polymerase.

The RNA polymerase core enzyme of Escherichia coli with the catalytic activity of RNA polymerization is assembled sequentially under the order: 2alpha --> alpha(2) --> alpha(2)beta --> alpha(2)betabeta'. The core enzyme gains the activities of promoter recognition and transcription initiation after binding the sigma subunit. The subunit-subunit contact surfaces of beta' subunit (1407 residues) were analyzed by testing complex formation between various beta' fragments and either the alpha(2)beta complex or the sigma(70) subunit. Results indicate that two regions, one central region between residues 515 and 842 and the other COOH-terminal proximal region downstream from residue 1141, are involved in binding the alpha(2)beta complex; and the NH(2)-terminal proximal region between residues 201 and 345 plays a major role in binding the sigma(70) subunit. However, both alpha(2)beta binding sites have weak activity of the sigma(70) subunit; likewise, the sigma(70) subunit-contact surface has weak binding activity of the alpha(2)beta complex. The sites involved in the catalytic function of RNA polymerization are all located within two spacer regions sandwiched between these three subunit-subunit contact surfaces.

The RNA polymerase holoenzyme of Escherichia coli is composed of the core enzyme with the subunit composition of ␣ 2 ␤␤Ј and one of seven different species of the subunit (for review, see Ref. 1). The core enzyme carries the catalytic activities for RNA polymerization, but the subunit is required for promoter recognition and transcription initiation from the promoters. The core enzyme is assembled sequentially both in vitro and in vivo under the order: 2␣ 3 ␣ 2 3 ␣ 2 ␤ 3 ␣ 2 ␤␤Ј (premature core) 3 E (active core) (for review, see Ref. 2).
Genetic and biochemical studies indicated that the subunitsubunit contact sites on ␣ including the sites for ␣ dimerization and the contact sites with the ␤ and ␤Ј subunits are all located within the amino (NH 2 )-terminal domain down to residue 235 (3)(4)(5), whereas the carboxyl (COOH)-terminal domain is involved in transcription regulation through direct interactions with class I (or ␣ contact) transcription factors and DNA UP elements (1,(5)(6)(7). The NMR structure has been determined for the ␣ COOH-terminal domain (8); the structure of the ␣ NH 2terminal domain was determined by x-ray crystallography (9), and the two domains are connected by a long flexible linker (10). Detailed mapping of the ␣-␣, ␣-␤, and ␣-␤Ј contact sites on the ␣ subunit have been carried out by making a number of contact-defective ␣ mutants with deletion, insertion, and Ala substitution mutations (11)(12)(13)(14) or by mapping the cleavage sites in ␣ by a chemical protease conjugated at various positions of the ␣ subunit (15).
On the contrary, relatively little is known on the subunitsubunit contact sites on the two large subunits, ␤ and ␤Ј. The mapping of ␣ subunit contact sites on the ␤ subunit was carried out using two approaches: analysis of the proteolytic cleavage pattern of the unassembled free ␤ subunit and the intermediate subassembly ␣ 2 ␤ complex (16,17), and analysis of complex formation between various ␤ fragments and the hexahistidine (His 6 )-tagged ␣ subunit or between various His 6 -tagged ␤ fragments and the intact ␣ subunit (17). The results altogether indicate that the primary and tight contact site of the ␣ subunit is located in the central portion of the ␤ polypeptide (16,17); but in addition, the COOH-terminal proximal region is needed as the secondary and probable regulatory site for either efficient binding of the ␣ subunit or stabilization of ␣-␤ contact (17). All of the ␤ fragment-␣ binary complexes isolated were able to bind the ␤Ј subunit, leading to formation of pseudo-core complexes, suggesting that the ␤Ј subunit contact site(s) on the ␤ subunit is(are) located near the ␣ contact sites. In addition to contact with two core subunits, ␣ and ␤Ј, the ␤ subunit seems to interact with the subunit at its COOH-terminal proximal region (18).
At present, our knowledge of the location of subunit-subunit contact sites on the ␤Ј subunit is limited. The isolated ␤Ј subunit is able to form binary complexes with the 70 subunit (19). The contact site of the 70 subunit on ␤Ј has been analyzed using different methods. (i) The assay of 70 subunit binding to pseudo-core enzymes containing various ␤Ј fragments indicates that the deletion mutant ␤Ј lacking amino acid residues 201-477 is unable to bind to the 70 subunit (20). (ii) A short ␤Ј fragment containing residues 260 -309 forms complexes with 70 , as measured by co-immobilization assay using His 6 -tagged ␤Ј fragments (21). (iii) Essentially the same NH 2 -terminal proximal region of ␤Ј is cleaved by a chemical protease conjugated to 70 (22). The isolated ␤ and ␤Ј subunits do not form stable binary complexes under isolated states (19), but these two subunits contact each other in the assembled RNA polymerase, suggesting that the ␣ subunit plays a role in supporting the contact between the ␤ and ␤Ј subunits. In the assembled core enzyme, one ␣ subunit contacts the ␤ subunit while the other ␣ makes direct contact with the ␤Ј subunit (23).
In this study, we tried to map the subunit-subunit contact sites on the ␤Ј subunit with not only the 70 subunit but also the ␣ 2 ␤ complex using two approaches. First, limited proteolysis by trypsin was carried out to generate ␤Ј fragments for analysis of protease-sensitive sites exposed on the unassembled free ␤Ј subunit. To identify the subunit-subunit contact sites on ␤Ј, the tryptic fragments were tested for complex formation with preformed ␣ 2 ␤ complex or purified 70 subunit.
For detailed mapping, various ␤Ј fragments with or without the His 6 tag were expressed, purified, and analyzed for complex formation with the ␣ 2 ␤ complex and the 70 subunit. Results indicate that the NH 2 -terminal proximal region between residues 201 and 345 is central for binding of the 70 subunit, and at least two regions between 515 and 842, and 1141 and 1407 are involved in binding of the ␣ 2 ␤ complex.

EXPERIMENTAL PROCEDURES
Construction of Expression Plasmids-Plasmid pGETC for expression of the intact ␤Ј subunit was constructed in two steps. First, pGETBC containing both the ␤ and ␤Ј subunit genes (rpoBC) was treated with NheI and MluI to remove the entire ␤ subunit coding sequence and a short NH 2 -terminal proximal sequence of the ␤Ј subunit. The missing region of rpoC was prepared by PCR 1 using pGETBC as template and two primers, 5Ј-AGCGGATTGTGCTAGCTCCGGCTC-GAGTTTGTCC-3Ј and 5Ј-GCCTTCGATAACCACATAGG-3Ј, and ligated, after treatment with NheI and MluI, with the pGETBC fragment carrying the sequence for NH 2 -terminal truncated rpoC. The resulting plasmid pGETC(ϩN) produced the ␤Ј subunit with an additional peptide of MASSDGSKS sequence at the NH 2 terminus because of the presence of ATG codon on the vector (as a part of the NdeI site) near the junction with the rpoC insert (note that the original initiation codon for the ␤Ј subunit is GTG (24)).
To remove this NH 2 -terminal extra sequence, an additional NdeI site was introduced at the initiation codon of rpoC by oligonucleotide-directed mutagenesis of pGETC(ϩN). The resulting plasmid was digested with NdeI to isolate three fragments, two of which are religated to prepare plasmid pGETC encoding the intact ␤Ј subunit without the extra sequence.
For construction of expression plasmids for various ␤Ј fragments, we first constructed a modified expression vector pGET1 from pGET by removing one of two NdeI sites, located upstream of the f1 replication origin. The intact rpoC gene was inserted into this plasmid between XbaI and HindIII to construct pGET1C (Fig. 1). Various kinds of rpoC fragment were synthesized by PCR using a set of 5Ј (C501-508) and 3Ј (C301-313) primers (for the location of the primer sequences, see Fig.  1). The PCR-amplified CP fragments (PCR fragments of rpoC) (for the location of each CP fragment along the rpoC gene, see Fig. 1) were treated with various restriction enzymes as shown in Table I, and the resulting CF fragments (fragments of rpoC) were cloned into pGET1C or its shorter derivative between the respective restriction sites (for the location of each CF fragment along the rpoC gene, see Fig. 1).
For construction of the expression plasmid pGEMDCH of the 70 subunit with a His 6 tag at the COOH terminus, a PCR product was prepared using pGEMD (5) as template and a pair of primers, 5Ј-GCAACCTGGTGGATCCGTCAG-3Ј and 5Ј-CCCGGAATTCAAGCTTT-TAGTGGTGGTGGTGGTGGTGGTGATCGTCCAGGAAGC-3Ј, treated with BamHI and HindIII, and substituted for the corresponding segment of pGEMD.
For construction of plasmid pGET21-␣C-His 6 for high level expression of the ␣ subunit with His 6 tag at the COOH terminus, a XbaI-BamHI fragment of plasmid pLAW2-H6, which was constructed for moderate expression of the His 6 -tagged ␣ subunit (14), was substituted for the corresponding segment of pET21a (Novagen).
Preparation of the His 6 -tagged ␣ subunit was carried out according to Kimura and Ishihama (14). For preparation of the His 6 -tagged 70 1 The abbreviations used are: PCR, polymerase chain reaction; NTA, nitrilotriacetic acid; PAGE, polyacrylamide gel electrophoresis.
FIG. 1. Construction of the expression plasmids of ␤ fragments. Fragments of the rpoC gene were PCR amplified using the primers shown below the rpoC gene (C500 series oligonucleotides were used as 5Ј primers, and C300 series oligonucleotides were used as 3Ј primers (for the primer set used in each PCR, see Table I). The PCR products (CP01-CP20) were treated with pairs of the restriction endonucleases as described in Table  I. One of the enzymes cuts the site, shown by an arrow on each CF fragment, which are present in the rpoC gene as illustrated on the top of rpoC.
Either BamHI (open arrowheads) or NdeI (filled arrowheads) was used as the second enzyme to cut the sites within the primers. The resulting CF fragments (CF01-CF22), shown by open bars, were cloned into pGET1C between the restriction sites as indicated in Table I. The plasmid map of pGET1C is inserted in the figure; bp, base pairs. Note that the restriction sites other than NdeI and BamHI are not necessarily unique on the pGET1C plasmid.
subunit and His 6 -tagged ␤Ј fragments, a batch procedure of Ni 2ϩ -NTA affinity purification was employed. In the case of His 6 -tagged ␤Ј fragments, the proteins in precipitate fractions of expressed cell lysates were solubilized with a dissociation buffer (50 mM Tris-HCl, pH 8.0, at 4°C, 0.2 M KCl, 10 mM MgCl 2 , 1 mM EDTA, 10 mM dithiothreitol, 20% (v/v) glycerol, and 6 M urea) (25). After centrifugation, the supernatant was applied onto Ni 2ϩ -NTA agarose (QIAGEN) previously equilibrated with buffer D (50 mM Tris-HCl, pH 7.9, at 4°C, 0.1 mM EDTA, 5% (v/v) glycerol) plus 5 mM imidazole and 6 M urea and incubated for 1 h at 4°C. Immobilized proteins were washed with buffer D plus 5 mM imidazole and 6 M urea and eluted with buffer D plus 200 mM imidazole and 6 M urea. In the case of 70 C-His 6 , the renatured 70 C-His 6 in TGED buffer (10 mM Tris-HCl, pH 8.0, at 4°C, 0.1 mM EDTA, 1 mM dithiothreitol, and 20% (v/v) glycerol) was fractionated by chromatography on a DEAE-TOYOPEARL 650M (TOSOH, Japan) column prior to affinity purification with Ni 2ϩ -NTA agarose. After incubation for 1.5 h on ice in buffer D plus 5 mM imidazole, immobilized proteins were eluted with buffer D plus 200 mM imidazole.
Trypsin Cleavage-Limited proteolysis by trypsin was performed according to Negishi et al. (26). In brief, the isolated ␤Ј(ϩN) subunit (0.27-0.30 g/l) containing an extra sequence at the NH 2 terminus was preincubated in a cleavage buffer (40 mM Tris-HCl, pH 7.8, 40 mM KCl, and 5% glycerol) at 20°C for 5 min and then subjected to trypsin cleavage at an input trypsin:substrate ratio of 1:500 (w/w). After incubation for 30 -90 min at 20°C, the reaction was terminated by adding phenylmethylsulfonyl fluoride at a final concentration of 5 mM.
Determination of NH 2 -terminal Amino Acid Sequences-NH 2 -terminal amino acid sequences of the tryptic fragments of ␤Ј(ϩN) subunit were determined as described by Negishi et al. (26). In brief, proteins in gels were blotted onto polyvinylidene difluoride supports (Nippon Genetics, Japan), and the membranes were stained with Coomassie Brilliant Blue R-250. Stained protein bands were cut out from the membranes and subjected directly to Edman degradation analysis using an Applied Biosystems model 491 protein sequencer.
Simultaneous Assembly of Two ␤Ј Fragments to the Core Enzyme-Reconstitution of the core complex was carried out using 1.87 nmol of His 6 -tagged ␣ subunit, 0.85 nmol of the intact ␤ subunit, and various combinations of two ␤Ј subunit fragments, one at a fixed concentration (3.40 nmol) and the other at various concentrations. The reconstituted core complexes were purified by Ni 2ϩ -NTA affinity chromatography and subjected to SDS-polyacrylamide gel electrophoresis (PAGE) on 10% uniform gels. Proteins were stained with Coomassie Brilliant Blue R-250, and the band intensity was measured with an LAS-1000 image analyzer (Fuji Photo Film Co., Japan).
Competitive Reconstitution of the Core Enzyme-Reconstitution of the core enzyme was carried out using 2.01 nmol of the His 6 -tagged ␣ subunit and 0.914 nmol each of the intact ␤ and ␤Ј subunits in the presence of various amounts of different ␤Ј fragments. The reconstituted core complexes were purified and analyzed as above.
Ni 2ϩ -NTA Agarose Binding Assay-Batch mode Ni 2ϩ -NTA affinity chromatography was carried out as described by Tang et al. (27). The His 6 -tagged ␤Ј subunit or His 6 -tagged ␤Ј fragments were used for isolation of both core complexes (␣ 2 ␤␤Ј or ␣ 2 ␤␤Ј(fragment)) and ␤Ј 70 complexes (␤Ј 70 or ␤Ј(fragment) 70 ). In brief, 1.08 nmol of isolated His 6 -tagged ␤Ј subunit or His 6 -tagged ␤Ј fragments in 1 ml of the standard dissociation buffer (25) containing 1 mM dithiothreitol was mixed with 160 l of Ni 2ϩ -NTA agarose equilibrated with the same dissociation buffer. After a 1-h incubation at 4°C, agarose beads were washed six times with 1 ml each of the reconstitution buffer to remove urea. The agarose-bound ␤Ј subunit or ␤Ј fragments were mixed with 1.08 nmol of the ␣ 2 ␤ subcomplex or 2.16 nmol of 70 subunits in 1 ml of the reconstitution buffer. After a 1-h incubation at 4°C, the agarose beads were washed once with 1 ml of the reconstitution buffer, six times with 1 ml each of buffer D plus 25 mM imidazole, and then the proteins were eluted with 160 l of buffer D plus 400 mM imidazole. Each fraction was analyzed by SDS-PAGE on 5-20% gradient gels (Bio-Rad Ready Gels J) or on 7.5, 10, and 12.5% uniform gels.
The core complex formation was also analyzed using the His 6 -tagged ␣ subunit. In this case, 2.60 nmol of His 6 -tagged ␣ subunit, 1.18 nmol of ␤, and 1.18 nmol of untagged ␤Ј subunit or ␤Ј fragments were mixed together in the dissociation buffer and then subjected to the reconstitution according to Fujita and Ishihama (25). To the reconstituted subunit mixture, 160 l of Ni 2ϩ -NTA agarose previously equilibrated in reconstitution buffer was added and incubated for 1 h at 4°C. The agarose beads were washed with 1 ml of the reconstitution buffer and then three times with buffer D plus 5 mM imidazole, and finally the Ni 2ϩ -NTA agarose-bound proteins were eluted with 160 l of buffer D plus 400 mM imidazole.
The vectors used were treated with the indicated restriction enzymes; the DNA fragments for expression of ␤Ј fragments carried the indicated restriction enzyme sites at termini.

Limited Proteolysis of the Isolated ␤Ј Subunit
The ␤Ј subunit of E. coli RNA polymerase consists of 1407 amino acid residues and contains a total of 186 potential cleavage sites by trypsin. Limited digestion of the isolated ␤Ј(ϩN) subunit with a short extra sequence of MASSDGSKS at its NH 2 terminus was carried out at various trypsin concentrations and for various times. Fig. 2A shows the time-dependent cleavage pattern at a trypsin:substrate ratio of 1:500 (w/w), as analyzed by SDS-PAGE. Essentially the same cleavage patterns were obtained for the intact ␤Ј subunit without the extra NH 2 -terminal sequence (data not shown). The tryptic fragments in the gels were blotted onto polyvinylidene difluoride membranes, and each visible band was subjected to analysis of its NH 2 -terminal amino acid sequence (Fig. 2B). The sites of tryptic cleavage were determined for all major cleavage prod-ucts, and the COOH termini of these fragments were estimated from the fragment sizes estimated from the migration distance on SDS-PAGE and the location of potential cleavage sites along ␤Ј by trypsin.
During the initial stage of trypsin digestion for less than 30 min, three NH 2 -terminal fragments, bpn105, bpn90, and bpn88 (NH 2 -terminal fragments of beta-prime are designated as bpn), were the major cleavage products ( Fig. 2A). In addition, five COOH-terminal fragments, bpc73, bpc70, bpc61, bpc57, and bpc55 (COOH-terminal fragments of beta-prime are designated as bpc) could be identified. Upon further incubation, however, these COOH-terminal fragments were all degraded within 90 min of the incubation, whereas the three NH 2 -terminal fragments remained at least until 180 min of the trypsin treatment. Fig. 2B shows the location of all of these cleavage fragments along the ␤Ј polypeptide. Fragments bpn105, bpn90, and bpn88 consist of the NH 2 -terminal proximal two-thirds of ␤Ј, and include the conserved regions A to G or F. Fragments bpc73, bpc70, bpc61, bpc57, and bpc55 consist of the COOH-terminal one-third of ␤Ј and include the conserved regions G and H and a dispensable region between region G and H. From the location of the major tryptic fragments, shown in Fig. 2B, it was predicted that the initial trypsin cleavage took place at arginine positions 798, 799, 838, 842, 905, 933, and 943 between the regions F and G (Fig. 2B).
After prolonged incubation in the presence of trypsin, the high molecular mass cleavage products disappeared, and concomitantly a number of low molecular mass fragments smaller than 50 kDa were identified, which might be generated after secondary cleavage of the primary products. All of the visible bands, as indicated in Fig. 2A, were subjected to NH 2 -terminal sequencing, and the results indicated that some bands contained more than two different fragments of similar sizes ( Fig.  2A). Three fragments, bpn42, bpn41, and bpn39, were found to retain the original NH 2 -terminal sequence, indicating that these came from the initial NH 2 -terminal fragments (for the location along the ␤Ј polypeptide, see Fig. 2B). Two fragments, bpm50 and bpm47, carried the NH 2 -terminal sequence starting from residues 346 and 353, respectively, indicating that they contained the sequences corresponding to the central part of the ␤Ј subunit (fragments derived from the middle part of the beta-prime subunit are designated as bpm) (Fig. 2B). Fragments bpc55, bpc50, bpc47, bpc42, and bpc39 all carried the NH 2 -terminal sequence starting from residue Thr-934 or Ala-944.
Taking all of the sequence data together, the secondary cleavage sites could be identified as follows. (i) The initial NH 2 -terminal fragments were cleaved at residue Lys-345 or Arg-352 between the conserved regions B and C, leading to generation of the NH 2 -terminal proximal fragments (bpn42, bpn41, and bpn39) including the regions A and B, and the middle fragments (bpm50 and bpm47) including the regions C, D, and E. (ii) The initial COOH-terminal fragments were processed from both NH 2 and COOH termini, i.e. at residues Arg-933 or Arg-943 from NH 2 terminus and at residues approximately downstream from residue 1300 (see Fig. 2B).

Identification of ␤Ј Fragments with Binding Activity to the
Use of His 6 -tagged ␣ Subunit for Complex Isolation-To identify the contact site(s) of the ␤Ј subunit with ␣ and ␤ subunits, the core enzyme reconstitution experiment was carried out using a His 6 -tagged ␣,␤ subunit and mixtures of trypsintreated ␤Ј subunit fragments. The addition of the His 6 tag at the COOH terminus of wild-type ␣ subunit does not interfere with the assembly of core enzyme (however, some ␣ mutants Blue R-250. The stained bands of ␤Ј fragments are designated on the basis of the NH 2 -terminal sequences and the molecular masses estimated from the migration distance on SDS-PAGE. Panel B, the ␤Ј fragments thus isolated after a 30-min treatment with trypsin were blotted onto polyvinylidene difluoride membranes followed by staining with Coomassie Brilliant Blue R-250. Stained protein bands on the polyvinylidene difluoride membranes were cut out and subjected to NH 2 -terminal amino acid sequencing according to Negishi et al. (26). The NH 2 terminus of each tryptic fragment was determined from the resulting sequence; the COOH-terminal end was estimated from the fragment size and the location of potential cleavage sites by trypsin. The major products are shown by shaded bars. The same analysis was carried out for other tryptic digests treated with trypsin for various times.
become inactive in the assembly by the addition of His 6 tag (14)). Subunits were mixed in the dissociation buffer; after dialysis against the reconstitution buffer to remove urea, the subunit mixtures were subjected to Ni 2ϩ -NTA affinity chromatography. Almost all major ␤Ј subunit fragments, including those not overlapping each other, were recovered in the complex fractions, which were eluted only in the presence of high concentrations of imidazole (data not shown). These ␤Ј subunit fragments did not bind to the Ni 2ϩ -NTA agarose in the absence of His 6 -tagged ␣ and ␤ subunits. The binding of nonoverlapping ␤Ј fragments with the ␣ 2 ␤ complex indicates the presence of multiple contact sites on the ␤Ј subunit with ␣ and/or ␤ subunits. However, it is not excluded yet that some ␤Ј fragments bind to either ␣ or ␤ subunits in a nonspecific manner.
Use of His 6 -tagged ␤Ј Fragments for Complex Isolation-To confirm the results obtained above, we next carried out the core enzyme reconstitution experiment using ␣ subunit, ␤ subunit, and His 6 -tagged ␤Ј fragments and analyzed the proteins associated with each ␤Ј fragment. For this purpose, a series of expression plasmids was constructed, which overexpressed a total of 13 species of the ␤Ј subunit fragment, each with a His 6 tag at the COOH terminus. All of these His 6 -tagged ␤Ј fragments were purified to apparent homogeneity under denatured conditions in the presence of urea from inclusion bodies of the expressed cell extracts using Ni 2ϩ -NTA agarose (data not shown). Isolated ␤Ј fragments were bound to Ni 2ϩ -NTA agarose and, after washing with the reconstitution buffer, mixed with the ␣ 2 ␤ complex for reconstitution. The core complexes were eluted using buffer D containing imidazole and analyzed by SDS-PAGE. The elution patterns are shown in Fig. 3.
Almost all of the ␤Ј fragments, except for ␤Ј(346 -515)C-His 6 and ␤Ј(842-1140)C-His 6 , formed core complexes with the ␣ 2 ␤ binary complex. Because ␤Ј(346 -515)C-His 6 and ␤Ј(842-1140)C-His 6 fragments did not form the core enzyme, the contact sites on the ␤Ј subunit involved in core enzyme assembly could be mapped in three regions: the NH 2 -terminal region (residues 1-345), the middle region (residues 515-842), and the COOH-terminal region (1141-1407). Among the three candidate regions for binding of the ␣ 2 ␤ complex, the involvement of the NH 2 -terminal region was not yet clear because the His 6tagged ␤Ј fragments from the NH 2 -terminal proximal region did not associate with Ni 2ϩ -NTA agarose, presumably because the His 6 tag was not exposed on the protein surface. In addition, most of the NH 2 -terminal fragments tended to aggregate after isolation, thereby preventing the accurate estimation of complex formation with the ␣ 2 ␤ complex (data not shown).
Use of Recombinant ␤Ј Fragments and His 6 -tagged ␣ Subunit for Complex Isolation-To avoid the confusion arising from the nonspecific aggregation of some ␤Ј fragments, we next added the His 6 tag to the COOH terminus of the ␣ subunit and repeated the core enzyme reconstitution experiments. For this purpose, the same set of 13 different ␤Ј fragments without the His 6 tag were purified to apparent homogeneity (data not shown). Purified ␣C-His 6 , ␤ subunit, and either intact ␤Ј subunit or one of the ␤Ј fragments were mixed under the denatured condition at the ␣:␤:␤Ј molar ratio of 2.2:1:1ϳ4. To prevent aggregation, the concentration of ␤Ј fragments was kept below 0.14 mg/ml. If there are multiple contact sites on the ␤Ј subunit, the affinity of fragments with only one contact site may be weaker than the intact ␤Ј subunit. Thus, ␤Ј fragments were added up to 4-fold molar excess over the stoichiometric molar ratio (2:1:1 for ␣:␤:␤Ј). Core complexes (or pseudo-core enzymes) were reconstituted after removing urea by dialysis against the reconstitution buffer, isolated by Ni 2ϩ -NTA agarose chromatography, and analyzed for the subunit composition by SDS-PAGE. The elution patterns from Ni 2ϩ -NTA agarose are shown in Fig. 4.

Specificity of the ␤Ј Fragment Binding to the ␣ 2 ␤ Complex
Competition Assay of the Core Enzyme Reconstitution by the ␤Ј Fragments-As an attempt to examine the specificity of ␤Ј fragment binding to the ␣ 2 ␤ complex, we next carried out a competition assay of core enzyme reconstitution by various ␤Ј fragments constructed as above. Purified ␣C-His 6 and intact ␤ and ␤Ј subunits were mixed in the dissociation buffer at an input molar ratio of 2.2:1.0:1.0 and in the presence of various amounts of one of the ␤Ј fragments. To prevent protein aggre-FIG. 3. Reconstitution of core complexes using His 6 -tagged ␤ fragments. His 6 -tagged intact ␤Ј subunit or each of the His 6 -tagged ␤Ј fragments (1.08 nmol each) were bound to Ni 2ϩ -NTA agarose in the dissociation buffer. Immobilized proteins were renatured by soaking the protein-bound agarose in the reconstitution buffer and then mixed with 1.08 nmol of the isolated ␣ 2 ␤ complex. Proteins were eluted with increasing concentrations of imidazole in the elution buffer. Lanes I 1 and I 2 , the applied samples; lanes F 1 and F 2 , unbound fractions; lane W 6 , wash fraction with the elution buffer containing 25 mM imidazole; lane E, an elution fraction with the buffer containing 400 mM imidazole. Aliquots of 32 l were analyzed by SDS-PAGE on either 5-20% gradient gels or 7.5, 10, and 12.5% uniform gels. The arrowheads indicate the migration positions of His 6 -tagged ␤Ј subunit or ␤Ј fragments; the open and closed circles indicate the migration positions of the ␤ and ␣ subunit, respectively. gations, the upper limit of the addition of ␤Ј fragments was set at a 1:1 weight ratio of intact ␤Ј subunit and ␤Ј fragment (the concentration of total proteins added in the reconstitution mixture was less than 0.26 mg/ml), which corresponded to the molar ratio ranging from 1:4.32 to 1:9.28 depending on the molecular weight of the ␤Ј fragment. After reconstitution, the core complexes were isolated by Ni 2ϩ -NTA affinity chromatography, and the amount of intact ␤Ј subunit assembled into the core complexes was analyzed. Fig. 5A shows one example of the competition assay. Among six different ␤Ј fragments tested, two fragments, ␤Ј(346 -515) and ␤Ј(842-1140), did not show significant inhibition of the core enzyme assembly. The three ␤Ј fragments, ␤Ј(1-150), ␤Ј(515-842), and ␤Ј(1141-1407), showed significant inhibition of the core enzyme assembly. The addition of a 4.32-fold molar excess of ␤Ј(515-842) over the intact ␤Ј inhibited the assembly of intact ␤Ј to 75% of the control level in the absence of ␤Ј fragments, whereas the addition of an 8.94fold molar excess of ␤Ј(1-150) over intact ␤Ј inhibited the core enzyme assembly down to 61%. The fourth fragment, ␤Ј(201-345), also showed the activity of core enzyme assembly albeit at lower levels than the other three fragments. The competition assay confirmed our conclusion that the three separate segments of the ␤Ј subunit bind the ␣ 2 ␤ complex independently.
Next we carried out the competition binding assay between different ␤Ј fragments. For all of the combinations of ␤Ј fragment mixtures noted above, the input level of the NH 2 -terminal proximal ␤Ј fragment was increased up to a 4-fold molar excess over the COOH-terminal ␤Ј fragment, keeping the NH 2 -terminal ␤Ј fragment at a fixed concentration. In all of the combinations examined, the assembled level of the NH 2 -terminal ␤Ј fragment stayed at a plateau even after addition of a 4-fold FIG. 5. Simultaneous assembly of two different ␤ fragments into the core enzyme. Panel A, competitive inhibition of the core enzyme assembly by the ␤Ј fragments. To a standard reconstitution mixture consisting of His 6 -tagged ␣ (2.01 nmol) and intact ␤ (0.914 nmol) and ␤Ј(0.914 nmol) subunits, 142 g each of the indicated ␤Ј fragments was added, which corresponded to the 1:1 weight ratio with 0.914 nmol of the intact ␤Ј subunit added. The core complexes containing the His 6 -tagged ␣ subunit were isolated by Ni 2ϩ -NTA affinity chromatography, and the protein composition was analyzed by SDS-PAGE (5%). The gel was stained with Coomassie Brilliant Blue R-250, and the band intensity was measured with an LAS-1000 image analyzer. The amounts of intact ␤Ј subunit relative to that of ␤ subunit were measured for each sample, and the levels of ␤Ј subunit bound in the presence of the indicated ␤Ј fragments were compared with that in the absence of ␤Ј fragment addition. Panel B, simultaneous assembly of two different ␤Ј fragments. Reconstitution of the core complex was carried out using 1.87 nmol of His 6 -tagged ␣ subunit, 0.85 nmol of the intact ␤ subunit,  (1141-1407) (lane 4). The reconstituted core complexes were purified by Ni 2ϩ -NTA affinity chromatography and subjected to SDS-PAGE on 10% uniform gels.
FIG. 4. Reconstitution of core complexes using His 6 -tagged ␣ subunit and ␤ fragments without His 6 tag. Either intact ␤Ј or one of the ␤Ј fragments (1.08 -4.32 nmol) was mixed with His 6tagged ␣ subunit (2.38 nmol) and ␤ subunit (1.08 nmol) in the dissociation buffer. Core complexes were reconstituted after dialysis of urea against the reconstitution buffer. The core complexes formed were bound to Ni 2ϩ -NTA agarose and then eluted with increasing concentrations of imidazole. Lane I, the original sample mixture; lane D, the sample mixture after dialysis of urea; lane F, the unbound fraction; lane W 6 , a wash fraction with the elution buffer containing 5 mM imidazole; lane E, an eluate fraction with the elution buffer containing 400 mM imidazole. Aliquots of 32 l were analyzed by SDS-PAGE on either 5-20% gradient gels or 7.5, 10, and 12.5% uniform gels. The arrowheads indicate the migration positions of ␤Ј or ␤Ј fragments; the open and closed circles indicate the positions of the ␤ and ␣ subunit, respectively. molar excess of the COOH-terminal ␤Ј fragment (data not shown). The same type of experiment was performed by increasing the COOH-terminal ␤Ј fragment up to 4-fold over the NH 2 -terminal ␤Ј fragment, which was kept at a fixed level. Again the level of assembled NH 2 -terminal ␤Ј fragment stayed constant irrespective of the increase of COOH-terminal fragment. Results altogether indicate that two different ␤Ј fragments, each containing a different ␣ 2 ␤ contact domain, can be assembled into the core enzyme complex simultaneously and independently.

Identification of ␤Ј Fragments with Binding Activity to the 70 Subunit
Use of His 6 -tagged 70 Subunit for Complex Isolation-Isolated ␤Ј subunit forms complexes with the 70 subunit (19). Previously we identified some ␤Ј subunit fragments with the binding activity of the 70 subunit (20). The 70 subunit contact site on ␤Ј was further confirmed after mapping the contact-dependent cleavage sites in the ␤Ј subunit by 70 subunit-conjugated FeBABE (22). To confirm the previous results and to test the reliability of the contact site mapping method employed, we also carried out complex formation between tryptic cleavage fragments of the ␤Ј subunit and the His 6 -tagged 70 subunit.
The His 6 -tagged 70 subunit was purified from expressed E. coli cells in soluble form. For complex formation, the purified His 6 -tagged 70 subunit was mixed with each of the trypsintreated ␤Ј subunits under nondenatured conditions, and the ␤Ј(fragment) 70 complexes formed were isolated using Ni 2ϩ -NTA agarose. As in the case of core complex formation, almost all of the ␤Ј subunit fragments were recovered in the complex fractions with the 70 subunit (data not shown). Using mixtures of ␤Ј fragments, it was difficult to separate 70 subunitassociated ␤Ј fragments from those bound indirectly through formation of complexes between ␤Ј fragments.
Use of His 6 -tagged ␤Ј Fragments for Complex Isolation-To avoid this confusion, we next carried out the ␤Ј(fragment) 70 complex formation experiment using the intact 70 subunit and His 6 -tagged ␤Ј fragments and analyzed the proteins associated with each ␤Ј fragment. The denatured ␤Ј fragments were first bound to Ni 2ϩ -NTA agarose, and after renaturation by soaking with the reconstitution buffer, mixed with 70 subunit. The ␤Ј(fragment) 70 complexes were eluted with buffer D containing 400 mM imidazole.
Among 13 different species of the ␤Ј fragment, one NH 2terminal proximal fragment, ␤Ј(1-150)C-His 6 , did not bind to the Ni 2ϩ -NTA agarose, presumably because the His 6 tag was buried within the protein molecule, or this His 6 -tagged ␤Ј fragment formed aggregates. Almost all ␤Ј fragments except for two, ␤Ј(346 -515)C-His 6 and ␤Ј(842-1140)C-His 6 , were found to bind the 70 subunit (Fig. 6). The basic ␤Ј protein tends to form nonspecific complexes with the acidic subunit (19), but these complexes may be dissociated upon increase in salt concentrations. To test this possibility, we then washed the Ni 2ϩ -NTA agarose-bound ␤Ј(fragment) 70 complexes with a high salt buffer (Fig. 7). All of the fragments derived from the middle portion and the COOH-terminal region of ␤Ј subunit were eluted by washing with the high salt elution buffer (buffer D, 25 mM imidazole plus 0.3 M KCl), indicating that the binding affinity is weak and presumably nonspecific for these fragments. On the other hand, the NH 2 -terminal fragments, ␤Ј(1-345)C-His 6 , ␤Ј(151-345)C-His 6 , ␤Ј(201-515)C-His 6 , and ␤Ј(201-345)C-His 6 , were retained tightly onto the Ni 2ϩ -NTA agarose even after washing with the high salt elution buffer, but they could be eluted by washing with the elution buffer containing 400 mM imidazole. Thus, we concluded that the NH 2 -terminal portion of the ␤Ј subunit is involved in the specific binding of the 70 subunit.

DISCUSSION
Proteolytic Cleavage Map of the ␤Ј Subunit-The ␤Ј subunit of E. coli RNA polymerase is a large basic protein consisting of 1407 amino acid residues and forms, together with the ␣ and ␤ subunits, the core enzyme with the subunit composition ␣ 2 ␤␤Ј. In the assembled core enzyme, the ␤Ј subunit plays a major role in initial nonspecific binding of the RNA polymerase to template DNA. The isolated ␤ subunit is totally inactive, and its intrinsic activities are exposed, in a stepwise manner, after formation of the ␣ 2 ␤ complex and the core enzyme (for review, see Ref. 2). In contrast, the isolated ␤Ј subunit retains the binding activities of DNA and the 70 subunit (19), suggesting that the conformation of isolated ␤Ј subunit is close to that assembled in the RNA polymerase core enzyme. In agreement with this interpretation, the initial proteolytic cleavage sites of isolated ␤Ј subunit analyzed in this study are close to the cleavage sites of the ␤Ј subunit in the core enzyme (28).
The initial cleavage sites are located between Arg-798 and Arg-943 near one of the two split sites of archaebacterial ␤Ј homologs (29,30). This region corresponds to the spacer region between two subunit-subunit contact domains with the 70 subunit and the ␣ 2 ␤ complex (see below), suggesting that this region is exposed on the surface of RNA polymerase. As illustrated in Fig. 8, a number of functional sites are located within or near this trypsin-sensitive region, including the binding site for streptolydigin, a potent inhibitor of transcription elongation (31), the epitope for inhibitory monoclonal antibodies (32) and one of the binding sites of the RNA 3Ј end (33,34). One of the termination-altering mutations is also located in the narrow region between residues 933 and 936 (35). A highly variable sequence exists downstream of the trypsin-sensitive region of ␤Ј subunit, and mutations in this region affect transcription elongation and transcript cleavage (36). These observations altogether suggest that the proteolytic cleavage-sensitive region of the ␤Ј subunit is exposed on the surface of core enzyme and is involved in the catalytic functions of RNA polymerase.
Subunit-Subunit Contact Sites on the ␤Ј Subunit with the ␣ 2 ␤ Complex-Two separate fragments of the ␤Ј subunit showed strong affinity to the ␣ 2 ␤ complex, one ␤Ј(515-842) and the other ␤Ј(1141-1407) (see Fig. 8). The fragment ␤Ј(515-842) derived from the central portion of ␤Ј corresponds to the segment between the split sites of ␤Ј homologs of chloroplasts (37)(38)(39) and those of archaebacteria (29,30). In agreement with the concept that the core polypeptides within multisubunit complexes such as the RNA polymerase ␤ subunit have mosaic structures consisting of two functional domains, one for protein-protein contacts and the other for functions such as the catalytic activities for RNA polymerization, each being comprised of noncontiguous segments on the primary structure (17), the ␤Ј subunit is also composed of two distinct functional domains. As noted above, the downstream sequence is involved in the catalytic functions of RNA polymerase. In addition, several lines of evidence indicate the involvement of upstream sequence in RNA synthesis. (i) The binding sites of Mg 2ϩ (residues 457-464) (40,41), located upstream of the central ␣ 2 ␤ complex binding site, form a part of the catalytic site of RNA polymerization. (ii) The 3Ј end of nascent RNA can be crosslinked to both upstream (400 -467) and downstream (932-1020) residues of this ␣ 2 ␤-contact site (33,34). (iii) The termination-altering mutations are located in both upstream and downstream regions of the subunit assembly domain (35). The second tight contact site of the ␣ 2 ␤ complex is located at the extreme COOH-terminal end of the ␤Ј subunit. Again, this region is involved in protein-protein interactions with not only the ␣ 2 ␤ complex but also the 70 subunit (see below) and single-stranded DNA-binding protein of phage N4 (42), which plays a key role in specificity modulation of host RNA polymerase in phage-infected E. coli.
In addition to these two tight contact sites on ␤Ј with the ␣ 2 ␤ complex, the extreme NH 2 -terminal proximal fragment, ␤Ј(1-345), showed a weak activity of ␣ 2 ␤ binding. However, both NH 2 -(␤Ј(1-150)) and COOH-terminal (␤Ј(201-345)) subfragments derived from this region showed higher activities of the ␣ 2 ␤ binding, indicating that the internal segment between residues 151 and 201 interferes with the interaction of this NH 2 -terminal region with the ␣ 2 ␤ complex. One possibility is that the interaction between the NH 2 -terminal region of ␤Ј and the ␣ 2 ␤ complex varies depending on the conformational change of ␤Ј during RNA polymerase assembly or discrete steps of transcription. For instance, the major contact site of the ␤Ј subunit with the 70 subunit is located in the COOH-terminal proximal segment (residues 201-345), and thus, the ␣ 2 ␤ complex binding activity of the extreme NH 2 -terminal segment (residues 1-150) could be largely affected depending on whether the 70 subunit is associated (holoenzyme form) or not (core enzyme form). In the NH 2 -terminal proximal segment, there is a putative zinc binding motif (see Fig. 8) which is believed to play a role in recognition of some rho-independent termination signals because mutations in this region affect transcription termination (43).
We also tested the reconstitution of core subunit complexes from pairwise combinations of the ␤Ј fragments. Preliminary experiments suggested that two different ␤Ј fragments with ␣ 2 ␤ binding activity can bind to the ␣ 2 ␤ complex simultaneously and independently (data not shown).
Subunit-Subunit Contact Sites on the ␤Ј Subunit with 70 Subunit-The ␤Ј 70 binary complex is formed in vitro from isolated individual components (19). Mapping of the contact sites on ␤Ј subunit with the 70 subunit was therefore performed by measuring complex formation of ␤Ј fragments and the intact 70 subunit (20). Here we extended this approach using a larger collection of ␤Ј fragments with or without the His 6 tag. Luo et al. (20) reported that the region of ␤Ј between residues 201 and 477 is required for binding of the 70 subunit, and here we identified a smaller fragment covering residues 201-345 which is still capable of binding the 70 subunit. We also carried out mapping of the 70 subunit contact site by analysis of the contact-dependent cleavage sites on ␤Ј with the FIG. 7. Reconstitution and isolation at high salt of ␤ 70 complexes using His-tagged ␤ fragments. Intact ␤Ј or one of ␤Ј fragments (1.08 nmol each) with His 6 tag were fixed onto Ni 2ϩ -NTA agarose in the dissociation buffer. After renaturation of ␤Ј proteins by soaking with the reconstitution buffer, the Ni 2ϩ -NTAbound ␤Ј proteins were mixed with the 70 subunit (2.16 nmol), and the ␤Ј 70 complex formed was eluted by the high salt elution buffer containing 0.3 M KCl. Lanes I 1 and I 2 , the applied samples; lanes F 1 and F 2 , unbound fractions; lane W 6 , a wash fraction with the elution buffer containing 25 mM imidazole; lane E, an elution fraction with the buffer containing 400 mM imidazole. Aliquots of 32 l were analyzed by SDS-PAGE on either 5-20% gradient gels or 7.5, 10, and 12.5% uniform gels. The arrowheads indicate the migration positions of His 6 -tagged ␤Ј fragments; the open circles indicate the migration position of the 70 subunit. chemical protease, FeBABE, tethered at various positions of the 70 subunit (22). This completely different approach revealed that the 70 -conjugated FeBABE generates the cleavage of ␤Ј between residues 276 and 344, in good agreement with the above result. Arthur and Burgess (21) also analyzed complex formation of the intact 70 subunit with varieties of ␤Ј fragment and found that a small sequence of ␤Ј between residues 260 and 309 is involved in binding the 70 subunit. Among the predicted 70 contact sequences determined in these independent experiments using different approaches, a small sequence consisting of 34 amino acids between residues 276 and 309 is always included (see Fig. 8). This minimum 70 contact site on ␤Ј corresponds to the conserved region B.
Among 13 different fragments of the ␤Ј subunit examined in this study, ␤Ј(201-345) showed the strongest activity of 70 subunit binding. In addition to this NH 2 -terminal proximal strong contact site, however, we identified in this study two weak binding sites with the 70 subunit, one in the middle of ␤Ј between residues 515 and 842 and the other in the COOHterminal proximal region downstream from residue 1141. The weak 70 contact site in the central part of ␤Ј overlaps the region including residues 581, 613, and 728, which are protected by binding of the 70 subunit, from the cleavage by hydroxyl radicals generated by Fe-EDTA (44). On the other hand, the COOH-terminal proximal weak 70 binding site overlaps the contact site with phage N4 single-stranded DNAbinding protein (43). Likewise, a truncation of the COOH terminus region affects an as yet unidentified function(s) of the RNA polymerase, resulting in the loss of ability to support phage P2 growth (45). One possibility is that phage factors such as N4 single-stranded DNA-binding protein modulate the interaction between the core enzyme and the host 70 subunit.
Interestingly, the two weak 70 binding sites correspond to the strong binding sites with the ␣ 2 ␤ complex (see above).
Thus, it appears that the protein-protein contact sites are clustered within three narrow regions that are separated along the primary structure of the ␤Ј subunit. The finding described here supports the concept that the ␤Ј subunit is composed of two kinds of functional domain, one for protein-protein interactions and the other for functions related to transcription such as RNA polymerization and binding of nascent RNA.
FIG. 8. The subunit-subunit contact surfaces on the ␤ subunit. The binding activities of 13 different kinds of the ␤Ј fragment with ␣ 2 ␤ complex and 70 subunit are summarized. The levels of binding activity represent: ϩϩ, strong binding; ϩ, weak binding; Ϫ, no binding. ND, not determined. The contact surfaces for the ␣ 2 ␤ complex and the 70 subunit were estimated from the results of binding assays. The primary and strong contact sites are indicated by black bars; the secondary and weak contact sites are shown by gray bars. Both the eight conserved sequences among ␤Ј homologs of prokaryotic and eukaryotic RNA polymerases (46,47) and the functional sites of E. coli RNA polymerase ␤Ј subunit so far identified are shown along the primary sequence of E. coli ␤Ј subunit.