J Biol Chem, Vol. 275, Issue 17, 12789-12798, April 28, 2000

The Functional Role of Basic Patch, a Structural Element of Escherichia coli Transcript Cleavage Factors GreA and GreB^*

Dmitry Kulish^§, Jookyung Lee, Ivan Lomakin, Beata Nowicka^¶, Asis Das^¶, Seth Darst, Kristjan Normet, and Sergei Borukhov^**

From the  Department of Microbiology and Immunology, State University of New York, Health Science Center at Brooklyn, New York 11203, the ^¶ Department of Microbiology, University of Connecticut Health Center, Farmington, Connecticut 06030, and the  Rockefeller University, New York, New York 10021

	ABSTRACT

TOP ABSTRACT INTRODUCTION MATERIALS AND METHODS RESULTS DISCUSSION REFERENCES

The transcript cleavage factors GreA and GreB of Escherichia coli are involved in the regulation of transcription elongation. The surface charge distribution analysis of their three-dimensional structures revealed that the N-terminal domains of GreA and GreB contain a small and large basic "patch," respectively. To elucidate the functional role of basic patch, mutant Gre proteins were engineered in which the size and charge distribution of basic patch were modified and characterized biochemically. We found that Gre mutants lacking basic patch or carrying basic patch of decreased size bind to RNA polymerase and induce transcript cleavage reaction in minimally backtracked ternary elongation complex (TEC) with the same efficiency as the wild type factors. However, they exhibit substantially lower readthrough and cleavage activities toward extensively backtracked and arrested TECs and display decreased efficiency of photocross-linking to the RNA 3'-terminus. Unlike wild type factors, basic patch-less Gre mutants are unable to complement the thermosensitive phenotype of GreA^-:GreB^- E. coli strain. The large basic patch is required but not sufficient for the induction of GreB-type cleavage reaction and for the cleavage of arrested TECs. Our results demonstrate that the basic patch residues are not directly involved in the induction of transcript cleavage reaction and suggest that the primary role of basic patch is to anchor the nascent RNA in TEC. These interactions are essential for the readthrough and antiarrest activities of Gre factors and, apparently, for their in vivo functions.

	INTRODUCTION

TOP ABSTRACT INTRODUCTION MATERIALS AND METHODS RESULTS DISCUSSION REFERENCES

Prokaryotic transcript cleavage factors GreA and GreB and eukaryotic transcription elongation factor TFIIS (SII)¹ affect the efficiency of transcription elongation in vitro by stimulating the intrinsic endonucleolytic activity of RNA polymerases (RNAPs) (1-8). The endonucleolytic hydrolysis of the nascent RNA has been observed in transcription elongation complexes (TECs) of multisubunit bacterial (9, 10) and viral RNAPs (11) and in eukaryotic RNAP I, II, and III (4-7, 12, 13). SII-like factors have been found in viruses (14), archaebacteria (15), yeast (16), insects (17), and mammals (18, 19), whereas the genes encoding GreA and GreB have been identified in more than 30 different bacterial organisms (see Fig. 1). The ubiquity of transcript cleavage factors in nature underscores their biological importance. Although gre and SII genes are not essential for cell viability, the deletion of these genes in Escherichia coli and S. cerevisiae renders them sensitive to certain growth conditions, such as elevated temperatures (10) or the presence of 6-azauracil (20).

The factor-stimulated cleavage of RNA occurs 2-18 bases upstream from the 3'-terminus, followed by dissociation of the 3'-proximal fragment from TEC. The 5'-proximal fragment of the transcript remains in TEC and can be extended in the presence of rNTPs. GreA induces hydrolysis of predominantly di- and trinucleotides (type A cleavage activity), whereas SII and GreB induce cleavage of 2-18-nucleotide-long RNA fragments (type B cleavage activity) (2, 3, 21, 22). In the absence of factors, the endonucleolytic activity of RNAP can be induced by pyrophosphate (23) or at alkaline pH (10), indicating that the same active center of RNAP is responsible for the cleavage reaction, as well as for pyrophosphorolysis and RNA synthesis reactions. The precise molecular mechanism by which Gre factors activate the endonucleolytic activity of RNAP is not known.

The cleavage reaction mediated by Gre and SII factors is believed to be an integral part of the mechanism that allows RNAP to overcome obstacles encountered during elongation in vivo, such as pausing and arresting sites in the DNA sequences, nucleosomes, or DNA-binding proteins and drugs (6, reviewed in Ref. 24). Although there are no direct in vivo data to support this, many in vitro studies have shown that SII and GreB can rescue arrested TECs (2-7, 9, 24) in which RNAP has irreversibly "backtracked" along the DNA template. Backtracking of RNAP occurs when it is forced to stall during elongation and results in backward translocation of RNAP with simultaneous repositioning of its catalytic center from the 3'-terminus to an internal site of the transcript (25-27). The cleavage of 3'-terminal segment of RNA induced by GreB and SII in arrested TEC allows RNAP to restart transcription from a newly generated 3'-terminus, which is now properly aligned with its catalytic center ("antiarrest activity") (2-6, 28). Unlike GreB, GreA is unable to induce cleavages in preformed arrested TECs and to reactivate them (1, 2, 21). However, GreA and GreB were shown to prevent elongating TECs from falling into an arrested conformation ("readthrough activity") (1, 2). The functional selectivity of GreA may be related to its ability to induce cleavage of only short RNAs. Besides readthrough/antiarrest function, the factor-induced transcript cleavage reaction may play a part in the mechanisms that regulate transcription fidelity (29-31) and transition of RNAP from initiation to elongation stage of transcription (32).

According to the established three-dimensional structures, GreA and GreB of E. coli share a similar overall structural organization (33, 34) that consists of an N-terminal coiled-coil domain (NTD) and a C-terminal globular domain (CTD). The NTD is responsible for the induction of type-specific nucleolytic activity by Gre factors (34, 35), whereas the CTD is responsible for the high affinity binding of Gre factors to RNAP (34-37). The charge distribution analysis of three-dimensional structures revealed a cluster of positively charged residues on NTD of GreA and GreB that form a distinct structural feature, described as basic "patch," on the side of the protein that is presumed to face RNAP in TEC. The basic patch residues of all known GreA and GreB molecules are highly conserved (Fig. 1). The photocross-linking experiments performed on various TECs have demonstrated that the 3'-terminus of RNA is located in the immediate vicinity of basic patch on both Gre proteins (33, 34), implicating basic patch in the functional interaction of Gre proteins with RNA in TEC. Additionally, the lengths of basic patch on GreA and GreB, ~7 and ~35 Å, respectively, were noted to correlate with the maximum lengths of RNA fragments that can be excised from TEC in the presence of Gre factors: 2-3 nucleotides for GreA and up to 18 nucleotides for GreB. Finally, it was observed that in the absence of factors, the nucleolytic activity of RNAP can be stimulated at a mildly alkaline pH, which mimics the effect of GreA (at pH 8.5-9.5) and GreB (at pH above 10.0) (10).

Based on these observations, two alternative hypotheses were proposed for the role of basic patch in Gre function: first, the basic patch residues increase the local pH near the active site of RNAP and directly affect its catalytic properties; second, they bind the nascent RNA by electrostatic interaction, thereby facilitating the cleavage reaction and determining the size of cleavage product. To test these hypotheses, we introduced a series of mutations in Gre proteins that would result in altered size and charge distribution of basic patch. The functional properties of mutant proteins were analyzed in vivo and in vitro. The results that we present rule out the first hypothesis and support the second, suggesting that the role of basic patch is to bind the 3'-terminal portion of the nascent RNA and to maintain it in proper orientation in TEC. Based on our results, we conclude that Gre-RNA interactions mediated through basic patch are essential for the readthrough and antiarrest functions of Gre factors in vitro and for their activity in vivo.

	MATERIALS AND METHODS

TOP ABSTRACT INTRODUCTION MATERIALS AND METHODS RESULTS DISCUSSION REFERENCES

Strains and Plasmids-- E. coli greA^-greB^- strain AD8571 ((ompR-greB), greA::KanR derivative of E. coli SG480 76 strain (10)) was used for overexpression of Gre proteins, for preparation of GreA/GreB-free RNAP, and for in vivo studies. pTRC99A (Amersham Pharmacia Biotech) was used as an expression vector for construction of plasmids overproducing the wild type (wt) GreA, wt GreB, and mutant basic patch-less (BPL) GreA factors. pET19B (Novagen) was used as an expression vector for GreA-large basic patch 3 (GreA-LBP3), GreA-LBP4, and all GreB mutants (Table I). The plasmid pMO1.1 carrying the wt greA gene and the plasmid pMO1.4His carrying the wt greB gene with C-terminal hexahistidine tag were used as templates for polymerase chain reaction preparation of the target greA and greB mutants, respectively (34). Because alteration of basic patch was expected to disrupt specific functions of Gre proteins, the cloning and selection of plasmids carrying mutant gre genes were performed in E. coli XL-Blue cells (Stratagene) in GreA⁺/GreB⁺-background to minimize the occurrence of spontaneous secondary null mutations. The DNA sequences were verified by dideoxy chain termination sequencing of the double stranded plasmid DNA.

Oligonucleotides-- The following oligonucleotides were used for polymerase chain reaction mutagenesis (underlined nucleotides define the restriction sites of the enzymes shown in parentheses; mutagenized residues of GreA and GreB are shown italicized): 1) 5'-GGTCGCCATGCTCAGCCGCTTCCGCGATAGC-3' (GreA-R37A, DdeI); 2) 5'-ACCCTGCTGTTCAGCAGCTGCGTGGTATTCGGC-3'(GreA-R52A, PvuII); 3) 5'-GCGCGTTCGACAGCTTTTTTTCGATATCTTTAATACGGCGTTCGCAGAA-3' (GreA: G60R/A67K, EcoRV); 4) 5'-TTCGCAGAAGCGCTGCTGTTTACGAGCTGCGTGG-3' (GreA-E53K/G56R, HaeII); 5) 5'-GATTTCACGCAGACGTGCTGCATTATACTGATAGTC-3' (GreB-K52/53A, MaeII); 6) 5'-GTGAGATAGCGCACGCGAGCGTCGATTTCAGCCAGACGTGCTGCATTATA-3' (GreB-R56A/R60A, HgaI); 7) 5'-GAGATTTTCCAGGCACGCAGTGAGATAGCGCAC-3' (GreB-K67A, DraIII); 8) 5'-GGCACGCAGTGAGATAGGCCACGCGATCGTCGATTTCATCCAGACGTGCTGC ATTATA-3' (GreB-R56D/R60D, PvuI). Four additional oligos corresponding to the 3'- and 5'-terminal portions of wt greA and wt greB with C-terminal His₆ tag were also used as described previously (34).

Preparation of Mutant GreA and GreB Proteins-- All gre mutants were obtained by a conventional two-step polymerase chain reaction procedure as described (34) using an appropriate combination of one of the mutagenic oligos (oligos 1-8) and one of oligos corresponding to the flanking region of greA and greB genes (2, 38). The resulting DNA fragments containing gre genes with a desirable mutation (or mutations) were inserted into either pTRC99A (GreA-37A, -52A, and -BPL) or PET-19b (GreA-LBP2, -LBP3, -LBP4, and all GreB mutants) vector using appropriate restriction sites (34, 35). The resulting expression plasmids (see Table I) carrying mutant gre genes were transformed into E. coli greA^-greB strain AD8571. The growth and isopropyl--D-thiogalactopyranoside (IPTG) induction of the cells carrying derivatives of pTRC99A was performed as described (34). The cells transformed with PET-19b derivatives were grown and infected with phage  (CE6) carrying phage T7 RNA polymerase gene according to the manufacturer's protocol (Novagen). The mutant GreA and GreB factors were purified to apparent homogeneity using the same purification scheme as described for wt GreA and His₆-tagged wt GreB (49) with minor modifications (35). Electrophoretically homogeneous Gre proteins were stored at 10 mg/ml in storage buffer (40 mM Tris-HCl, pH 7.9, 0.8 M NaCl, 50% glycerol, 0.1 mM EDTA, 5 mM 2-mercaptoethanol) at 20 °C.

In Vitro Transcription Assays-- The readthrough, antiarrest, and transcript cleavage assays were performed as described previously (49) using 202 bp E. coli rrnB P1 DNA fragment with the initial transcribed sequence +1CACCAUGACACGGA. The initial TEC carrying radiolabeled hexameric transcript CpApCpCpApC (6C-TEC) was prepared by incubation of 0.9 µg (7.3 pmol) of the DNA fragment with 12 µg (30 pmol) of GreA/GreB-free RNAP holoenzyme (10), 1 mg/ml bovine serum albumin, 0.5 mM CpA, 5 µM ATP, and 1 µM [-³²P]CTP (3000 Ci/mmol) in 35 µl of standard transcription buffer (40 mM Tris-acetate, pH 7.9, 30 mM KCl, and 10 mM MgCl₂) for 10 min at 37 °C (here and elsewhere in the text, boldface type symbolizes radioactive phosphates). The 6C-TEC was further purified by gel filtration on Quick-Spin G-50 column. The TEC carrying heptamer CpApCpCpApCpU (7U-TEC), nonamer CpApCpCpApCpUpGpA (9A-TEC), and arrested TECs carrying 12-meric and 13-meric transcripts CpApCpCpApCpUpGpApCpApC(pG) (12C/13G-TEC) were obtained as described (49) by extension of 6C-TEC in the presence of UTP (3 µM), UTP+GTP+ATP (5 µM each) and four rNTPs (30 µM each), respectively, followed by purification on Quick-Spin G-50 column. The TEC carrying decamer CpApCpCpApCpUpGpApC (10C-TEC) was obtained by extension of purified 9A-TEC in the presence of 10 µM CTP (2). For the analysis of the type of cleavage activity of Gre factors, 9A- and 10C-TECs carrying radioactive phosphate at position +9 were prepared from unlabeled 6C-TEC (2). First, 6C-TEC was extended in the presence of 2 µM each of UTP + GTP + [-³²P]ATP (1500 Ci/mmol) yielding 9A-TEC, which was purified by Quick-Spin G-50 column. Then, the radiolabeled 9A-TEC was extended in the presence of 10 µM CTP yielding 10C-TEC. For determination of specific transcript cleavage activity of Gre mutants (as well as their antiarrest/readthrough activities), we used procedures that are described in our previous publications (1, 2, 10, 34, 35, 49). Briefly, for each Gre protein, the specific cleavage activity was determined by measuring the percentage of cleavage of the initial 2.5 nM TEC after incubation with serial 3-fold dilutions of Gre factor in 10 µl of standard transcription buffer at 37 °C for 20 min (49). The titration curves were made for each Gre protein, and the minimum amount of factor that causes 50% hydrolysis of RNA in the corresponding TEC was calculated. Similarly, for determination of specific readthrough activity, the initial radiolabeled 2.5 nM 6C-TEC was incubated in the presence of 100 µM NTPs with serial dilutions of Gre proteins as described above, and the minimum amount of factor that is required to decrease the formation of arrested 12C/13G-TECs by 50% was calculated (49). The RNA products were analyzed by urea-23%-PAGE, autoradiographed and quantified using a PhosphorImager as described (34).

Quantitative RNA-Gre photocross-linking assay was performed using radiolabeled 9A-TEC carrying either 8-N₃AMP at the RNA 3'-terminus or N₃-NPA-UMP at position +7 of the transcript as described (34, 35). The samples were analyzed by Tris-Tricine SDS-16%-PAGE (50) followed by autoradiography and quantification using a PhosphorImager.

In Vivo Temperature Sensitivity (ts) Complementation Assay-- The greA^-/greB^- E. coli cells (strain AD8571) transformed with pTRC99A, pDKA37A, pDKA52A, pDKABPL, pMO1.1, or pMO1.4His were grown at 30 °C in LB medium containing 150 µg/ml ampicillin and 0.3% glucose to an absorbance of 0.35 at 600 nm. Five 10-fold dilutions of each strain were made, and 5 µl of each sample were plated (in triplicate) on LB-agar plates containing 100 µg/ml ampicillin and either 0.3% glucose or IPTG at various concentrations. The plates were incubated at 30 or 42 °C for 36 or 18 h, respectively. The remaining liquid culture of each strain was diluted three times with LB-medium, grown, and induced with different concentrations of IPTG (see below and in the legend to Fig. 5 for details). An aliquot of each culture was analyzed for Gre expression by SDS-PAGE.

	RESULTS

TOP ABSTRACT INTRODUCTION MATERIALS AND METHODS RESULTS DISCUSSION REFERENCES

Construction and Expression of Gre Mutants-- Among 34 prokaryotic organisms of which the partial or complete genomes have been sequenced to date and made available in the NCBI data bank, all but two possess the genes that encode homologs of GreA, GreB, or both. Fig. 1 shows the amino acid sequence alignment of the N-terminal half of 33 GreA and 13 GreB proteins. Although the overall homology shared between different Gre proteins varies from 21 to 95%, the residues that make up the basic patch in E. coli GreA (Arg³⁷ and Arg⁵²) and GreB (Lys⁵², Lys⁵³, Arg^56, Arg⁶⁰, and Lys⁶⁷) are highly conserved (Fig. 1). Specifically, three out of six basic patch residues of GreB are invariant among 13 organisms. A few exceptions reside in greA genes of Clostridium acetobutylicum, Deinococcus radiodurans, Chlamydia trachomatis, and Treponema pallidum, in which the highly conserved Arg³⁷ or Arg⁵² is changed to leucine or isoleucine, and in greB genes of Pseudomonas aeruginosa and Shewanella putrefaciens, in which the conserved Arg⁶⁰ is changed to serine. These observations suggest that not all of the residues making up the basic patch may be necessary for in vivo activities of Gre factors.

View larger version (117K): [in this window] [in a new window]   Fig. 1.   Sequence alignment of the N-terminal domain of Gre proteins from 33 different organisms (see text for details). The Gre gene sequences that were obtained from published sources are referenced (2, 38-48). Sequence data for all other Gre genes was obtained using the BLAST search engine from NCBI, The Institute for Genomic Research, and the Sanger Center. The groups that provided the sequence data are identified by a capital letter after each organism as follows: O, Oklahoma University Advanced Center for Genome Technology; S, Pathogen Sequencing Unit at the Sanger Center; T, The Institute for Genomic Research; G, Genome Therapeutics Corp.; P, University of Washington Genome Center and PathoGenesis Corp. GenBankTM accession numbers for the GreA genes of Rhizobium leguminosarum and Zymomonas mobilis and the GreB gene of Yersinia enterocolitica are G2944085, G2654150, and Y08950, respectively. The numbering above the sequences is with reference to E. coli GreA. The sequences in open red box (also indicated by horizontal red brackets) denote the fragments of E. coli GreA and GreB containing the cross-link with RNA 3'-terminus (33, 34). Identical amino acid residues are shown in gray. Conserved residues forming basic patches of GreA and GreB are shown in blue boxes.

To investigate the functional role of basic patch in E. coli GreA and GreB, and to assess the contribution of basic patch residues to the characteristic properties of Gre factors, the residues Arg³⁷ and Arg⁵² of GreA and Lys⁵², Lys⁵³, Arg⁵⁶, Arg⁶⁰, and Lys⁶⁷ of GreB were substituted by site-directed polymerase chain reaction mutagenesis to alanine individually, in groups of spatially adjacent residues, or altogether, yielding BPL GreA and GreB mutants (Table I). Four additional mutants were made by substitution of specific residues to generate GreA carrying increasingly large basic patch (GreA-LBP2, -LBP3, and -LBP4) and GreB carrying an acidic patch on the basic patch-less background (GreB-BPL5-DD) (Table I). Basic patch residues of GreA and GreB are located on the surface of NTD and are not involved in intramolecular and interdomain interactions. Therefore, mutations of these residues are expected not to affect the folding of Gre molecules. Indeed, the chromatographic behavior of purified mutant proteins during size-exclusion high pressure liquid chromatography and their apparent binding affinity to RNAP holoenzyme tested by Gre-RNAP competition binding assay (34) were indistinguishable from those of wt GreA and GreB (data not shown).

Quantitative in Vitro Analysis of BPL Gre Mutants-- To characterize the functional properties of mutant Gre proteins, we used three types of quantitative in vitro transcription assays based on E. coli ribosomal rrnB P1 transcription unit (51). First, a specific transcript cleavage assay was used to test the ability of Gre mutants to induce nucleolytic reactions in minimally backtracked 7U-TEC, extensively backtracked 9A- and 10C-TECs, and arrested (irreversibly backtracked) 12C/13G-TEC (1, 2). These TECs differ by the extent to which RNAP is backtracked from the RNA 3'-terminus (1-2, 2-4, 3-5, and 7-8 nucleotides, respectively). Second, the specific Gre-RNA interactions were analyzed by photocross-linking assay using a photoactive probe with a short (~2 Å) space arm, 8-azido-AMP, placed at the RNA 3'-terminus in 9A-TEC (33, 34). Third, a readthrough assay was used to examine the ability of mutant proteins to help RNAP to read through intrinsic arresting sites on rrnB P1 template at positions +12 and +13, and prevent the formation of arrested 12C- and 13G-TECs during elongation (49). The results of these experiments are summarized in Fig. 2.

View larger version (44K): [in this window] [in a new window]   Fig. 2.   Quantitative in vitro analysis of BPL Gre mutants. A, water-accessible surface and charge distribution of wt and mutant Gre factors generated using GRASP (52) based on published crystal structure of GreA (33) and model structure of GreB (34). The surface is colored by the electrostatic potential: white, uncharged; red, negative (Asp and Glu); blue, positive (Arg and Lys). The side of the protein that presumably faces the RNAP and RNA in TECs is shown. B, the transcript cleavage assay toward 7U-, 9A-, 10C-, and 12C/13G-TECs. The bar graphs, shown in logarithmic scale, represent the specific activities of Gre proteins expressed in relative units/µg (34, 51). Each value represents an average of at least three independent experiments. C, the photocross-linking and transcription readthrough assays. The light gray bars represent the efficiency of specific RNA-Gre cross-linking expressed in arbitrary units/µg. 1 unit is defined as 0.05 fmol of the radioactive 9A-TEC cross-linked to GreA and GreB under standard reaction conditions (34). All proteins were titrated over a wide range of concentrations (10 nM to 20 µM) to obtain the specific values of cross-linking efficiency within the linear range of concentrations. The black bars represent the specific readthrough activities of Gre proteins expressed in relative units/µg (34, 51). Each value represents an average of three independent experiments.

Transcript Cleavage Assays-- Single and multiple alanine substitutions of basic patch residues and the introduction of acidic patch (GreB-BPL5-DD) did not affect the specific transcript cleavage activity of mutant proteins toward 7U-TEC (Fig. 2B, open bars). Thus, the basic patches of GreA or GreB are not responsible for activation of the nucleolytic center in RNAP. However, compared with the wt factors, the cleavage activity of mutant proteins toward 9A- and 10C-TECs significantly decreased (Fig. 2B, light gray and dark gray bars, respectively). The effect of mutations was more pronounced for GreA, in which each individual Ala substitution caused a 10-fold decrease in cleavage activity toward both TECs and in which double mutation (GreA-BPL) resulted in 30- and 200-fold decreases toward 9A- and 10C-TECs, respectively. The basic patch mutations in GreB caused a moderate (3-8-fold) loss of cleavage activity for both TECs, except for GreB-BPL5-DD, in which the decrease was more substantial (15- and 50-fold, respectively). These results show that multiple mutations of basic patch residues in both Gre proteins have an additive effect on the overall loss of cleavage activity toward 9A- and 10C-TECs. The most striking effect of mutations was observed in the cleavage assay using arrested 12C/13G-TECs (Fig. 2B, black bars). As expected, unlike GreB, which was shown to be highly active toward arrested TECs (2, 21), the wt GreA and its mutant variants were completely inactive in this assay. For GreB, the Ala substitution of two basic patch residues at the tip of NTD (K52A and K53A: GreB-BPL2) resulted in a 15-fold decrease in cleavage activity toward arrested 12C/13G-TECs. Further mutations of Arg⁵⁶ and Arg⁶⁰ at the center of the basic patch (GreB-BPL4) caused an additional 7-fold decrease, leading to an overall loss of cleavage activity by a factor of ~100. The Ala substitution of the last basic patch residue, Lys⁶⁷, at the top of NTD (GreB-BPL5) resulted in an additional 2-fold decrease in cleavage activity, making this mutant 200 times less active than the wt GreB. Finally, GreB-BPL5-DD carrying two acidic residues, Asp⁵⁶ and Asp⁶⁰, in the center of basic patch-less NTD, showed no detectable cleavage activity (more than 5000-fold decrease in activity).

In addition to concentration-dependence assays, for each Gre protein, we also performed the time course analyses of transcript cleavage reaction toward all TECs studied. For the wild type Gre factors and their corresponding BPL mutants (all used at 1 nM), the rates of cleavage reactions toward 7U-TEC (used at 2.5 nM) were similar. However, the Gre-BPL mutants displayed progressively decreased rates of cleavage reactions toward 9A-, 10C-, and 12C/13G-TECs relative to the wt factors (data not shown). These results were consistent with conclusions drawn from analyses of the specific cleavage activities of Gre mutants using assays described above.

Gre-RNA Photocross-linking Assay-- GreA and GreB have been shown to cross-link to the nascent RNA 3'-terminus in 9A-TEC. Their cross-linking efficiencies differ by a factor of ~30, in favor of GreB (34). Compared with the wt factors, the efficiency of mutant protein cross-linking to the RNA 3'-terminal photoactive probe in 9A-TEC (Fig. 2C, light gray bars) decreased for GreA mutants in the order GreA GreA-37A  GreA-52A GreA-BPL, and for GreB mutants in the order GreB > GreB-BPL2 GreB-BPL-4  GreB-BPL5 GreB-BPL5-DD. The strongest negative effect was observed with GreA-BPL and GreB-BPL5-DD, the cross-linking efficiencies of which were 4 and 0.2% of the cross-linking efficiencies of wt GreA and GreB, respectively. To test whether this result reflects a general failure of Gre mutants to bind 9A-TEC, we performed a similar cross-linking experiment using a photoactive probe 5-(4-azido-2-nitrophenyl)-allylamino derivative of UMP (N₃-NPA-UMP) with a long spacer (~12 Å) incorporated at position +7 of the transcript in 9A-TEC (35). N₃-NPA-UMP has a wider radius of action than 8-N₃-AMP and allows us to detect the presence of Gre factors in TEC without being as discriminating as 8-N₃-AMP (33, 34). Gre mutants lacking basic patch cross-linked with 9A-TEC carrying N₃-NPA-UMP with the same efficiency as the wt factors (data not shown), indicating that the binding affinity of Gre proteins toward 9A-TEC was not affected by basic patch mutations. Therefore, the strength of specific RNA-Gre interactions in TEC appears to be a direct function of the size and the overall charge of the basic patch on the surface of NTD.

Specific Transcriptional Readthrough Assay-- The readthrough assay is based on the ability of wt GreA and GreB to suppress the transcriptional arrest of RNAP during elongation in several promoter systems in vitro (1, 2). When the initial radiolabeled 6C-TEC obtained on rrnB P1 template is allowed to elongate in the presence of NTPs, ~50% of the complexes that do not reach the end of the template becomes arrested at positions +12C and +13G. The presence of GreA or GreB at the beginning of transcription elongation significantly (~10-20-fold) reduces the formation of arrested TECs and increases the amount of the runoff product (1, 2). As reported earlier (34), the specific readthrough activity of GreB, expressed in relative units/µg of Gre protein, was ~10 times higher than that of GreA (Fig. 2C, black bars). Interestingly, in this assay, the decrease in the specific readthrough activities of mutants compared with the wt factors followed the same order as the decrease in their cross-linking efficiencies: GreA GreA-37A  GreA-52A GreA-BPL, and GreB > GreB-BPL2 GreB-BPL-4  GreB-BPL5 GreB-BPL5-DD. Again, the most severe defect was observed in proteins completely lacking basic patch: GreA-BPL and GreB-BPL5-DD (25- and 450-fold decreases, respectively). A similar loss of activity was observed for GreB mutants in their antiarrest activity (49), namely the ability of Gre factors to reactivate preformed arrested 12C- and 13G-TECs (data not shown). These observations imply a functional correlation between the property of Gre proteins to interact with RNA and their readthrough/antiarrest activities.

Analysis of the Type of Cleavage Activity of BPL Gre Mutants-- Next, we tested the hypothesis that basic patch functions as a molecular ruler to determine the maximum length of RNA to be excised from TEC, and is thus directly responsible for conferring A- or B-type specific cleavage activities to Gre factors. We analyzed the BPL-mutants in qualitative transcript cleavage assay (49) using 9A- and 10C-TECs carrying radiolabeled transcripts CpApCpCpApCpUpGpA and CpApCpCpApCpUpGpApC, respectively. In reversibly backtracked 9A and 10C-TECs, the wt GreA stimulates cleavage and release of dinucleotide pGpA and trinucleotide pGpApC, respectively (type A cleavage activity), whereas the wt GreB predominantly induces the cleavage of tetranucleotide pCpUpGpA and pentanucleotide pCpUpGpApC, respectively (type B cleavage activity) (2). Thus, in the assay using 9A- and 10C-TECs, the type of products generated define the GreA- or GreB-type activity. The results obtained from comparative analyses of BPL-mutants and wt Gre proteins using these criteria are shown in Fig. 3. The removal of one or both basic residues from GreA (GreA52A, GreA37A, and GreA-BPL) did not affect the type specificity of the cleavage reaction either in 9A-TEC (Fig. 3A, lanes 3-6) or in 10C-TEC (data not shown). However, the stepwise reduction of the basic patch in GreB (GreB-BPL2, -BPL4, and -BPL5 mutants) led to a gradual decrease in the length of excised 3'-terminal RNA fragment in 9A-TEC from tetranucleotide to tri- and dinucleotide (Fig. 3A, lanes 7-10), and in 10C-TEC from pentanucleotide to tetra- and trinucleotide (Fig. 3B, lanes 2-5). The most striking effect was observed in GreB-BPL5-DD mutant, which, like GreA, stimulated cleavages of only di- and trinucleotides (Fig. 3A, lane 11, and Fig. 3B, lane 6). These results demonstrate that basic patch residues of GreB determine the type of cleavage activity characteristic for this factor and that decreasing the size of its basic patch switches its activity from B to A type.

View larger version (40K): [in this window] [in a new window]   Fig. 3.   Analysis of the type of cleavage activity of BPL Gre mutants toward 9A-TEC (A) and 10C-TEC (B). Shown are the autoradiographs after urea-23%-PAGE. A, the starting 9A-TEC (2 nM) (lane 1) was incubated in transcription buffer for 10 min at 37 °C alone (lane 2) or in the presence of 20 nM Gre factors (lanes 3-11) as indicated. B, the initial 10C-TEC (1 nM) (lane 1) was incubated in transcription buffer for 15 min at 37 °C in the presence of 50 nM Gre factors (lanes 2-7) as indicated. The asterisk indicates radioactive phosphates.

In Vitro Analysis of GreA Mutants Carrying Large Basic Patch-- To test whether the size and charge of basic patch alone determine the functional properties of Gre factors, we analyzed GreA mutants carrying increasingly large basic patch in in vitro transcription assays (Fig. 4A). As shown in Fig. 4B, none of GreA-LBP mutants gained the ability of the wt GreB to induce transcript cleavage reactions in arrested 12C/13G-TECs (black bars). Moreover, whereas the introduction of two basic residues, Arg⁶⁰ and Lys⁶⁷, into GreA (GreA-LBP2) only slightly affected its functional activities, the additional substitutions of Glu⁵³ to Arg (GreA-LBP3) and Gly⁵⁶ to Lys (GreA-LBP4) caused a substantial (50-80-fold) decrease in transcript cleavage activity toward 7U, 9A-, and 10C-TECs (Fig. 4B, open, light gray, and dark gray bars, respectively). Furthermore, the readthrough activity of GreA-LBP3 and -LBP4 mutants decreased by factors of ~15 and ~50, respectively, making the latter mutant, which carries the largest basic patch, totally inactive in the readthrough assay (Fig. 4C, black bars). At the same time, the efficiency of GreA-LBP3 and -LBP4 cross-linking to the RNA 3'-terminal probe in 9A-TEC increased 3- and 15-fold, respectively, over that of wt GreA (Fig. 4C, light gray bars). Again, the stronger effect was observed in GreA-LBP4, the photocross-linking efficiency of which was similar to that of the wt GreB.

View larger version (45K): [in this window] [in a new window]   Fig. 4.   In vitro analysis of GreA mutants carrying large basic patch. All details of the figure are as described in Fig. 2.

These results demonstrate that the enlargement of basic patch on the surface of NTD in GreA leads to an increased attraction of RNA 3'-terminus, apparently through nonspecific electrostatic interactions. The increase in Gre-RNA binding by itself, however, is not sufficient to confer GreB-like properties to GreA. Moreover, the loss of transcript cleavage and readthrough activities by GreA-LBP mutants (especially LBP4) suggests that the basic residues introduced at positions 53 and 56 of GreA (E53R and G56K mutations) distort the local structure of NTD and disrupt its normal contacts with RNAP and RNA. Therefore, it appears that stronger but less specific Gre-RNA interactions are detrimental to the functional activity of a Gre protein. Similar functional defect has previously been observed in posttranslationally modified GreB. Under conditions of limited proteolysis, wt GreB can be selectively cleaved by endoproteinase AspN at Asp⁴⁷, which is located at the tip of NTD near the loop connecting two  helices of the coiled-coil domain (34). The nicked GreB displays ~10 times higher efficiency of photocross-linking to RNA 3'-terminal probe in 9A-TEC, but ~200-times lower transcript cleavage activity toward 7U- and 9A-TECs, than the intact factor and completely lacks antiarrest and readthrough activities (data not shown). As in the case of GreA-LBP4 mutant, the distortion of the coiled-coil structure of NTD in GreB, caused by a nick in the polypeptide chain, results in an aberrant RNA binding and loss of functional activity.

In Vivo Analysis of GreA and GreB Mutants with Decreased Basic Patch-- To assess the role of basic patch in the in vivo function of Gre factors and to correlate them with their in vitro activities, we subjected Gre mutants to an in vivo ts complementation assay. The assay is based on the ability of wt greA or greB genes, under conditions of low level expression from a pTRC99A expression vector, to complement the ts phenotype of the greA^-/greB^- E. coli strain (10). The results of in vivo assay for GreA mutants with decreased basic patch are shown in Fig. 5.

View larger version (85K): [in this window] [in a new window]   Fig. 5.   In vivo analysis of BPL-GreA mutants. A, Coomassie staining after Tris-Tricine SDS-16%-PAGE of the total cell lysates of greA-/greB- E. coli strain containing vector alone (lane 1) or plasmids expressing wt GreA (lanes 2-5), GreA-37A (lanes 6-9), GreA-52A (lanes 10-13), and GreA-BPL (lanes 14-17). Cells were induced by 0.03 mM (lanes 3, 7, 11, and 15), 0.1 mM (lanes 4, 8, 12, and 16), and 0.3 mM (lanes 1, 5, 9, 13, and 17) IPTG. The arrow indicates the position of expressed Gre protein on the gel. B, photograph of cultures grown on LB-agar medium under conditions specified in the figure and under "Materials and Methods." The numbers at the top of panels a-c indicate the number of cells/ml.

Without IPTG induction, there is low but detectable expression of wt and mutant GreA proteins from the leaky trc promoter of pTRC99A (Fig. 5A, lanes 2, 6, 10, and 14). Under these conditions, the intracellular concentration of Gre proteins was ~5 µM according to the quantitative immunoblotting performed with anti-GreA antibodies (data not shown), about 3-4 times the natural concentration of GreA in the parental wild type E. coli cells (35). At a permissive temperature (30 °C), growth of greA^-/greB^- E. coli cells was not affected by the presence of any Gre protein (Fig. 5B, panel a). However, at a nonpermissive temperature (42 °C), the basal level expression of wt GreA provided full viability to the cells, whereas double mutant GreA-BPL or vector alone did not (Fig. 5B, panel d). Under the same conditions, the expression of single mutants GreA-37A and GreA-52A resulted in moderately and weakly positive phenotypes, respectively.

The induction of cells with 0.03 mM IPTG resulted in about ~8-fold increase in Gre expression (Fig. 5A, lanes 3, 7, 11, and 15), reaching an estimated intracellular concentration of about 40 µM. In contrast to control cells carrying vector alone, the elevated expression of the wt GreA, as well as all mutant factors, was sufficient to complement the ts phenotype of greA^-/greB^- E. coli cells at 42 °C (Fig. 5B, panel e), without affecting the cell growth at 30 °C (panel b). Next, the expression of Gre factors in the presence of 0.1 mM or 0.3 mM IPTG resulted in an ~40-fold increase of Gre synthesis compared with uninduced cells (Fig. 5A, lanes 4, 5, 8, 9, 12, 13, 16, and 17). Under these conditions, the overexpression of wt GreA became highly toxic to the cells and led to a lethal phenotype at both permissive and nonpermissive temperatures (Fig. 5B, panels c and f, respectively). Unlike wt GreA, and to a lesser extent unlike each single GreA mutant, the overproduction of double mutant GreA-BPL was much less toxic for the cells grown at 30 °C, and moreover, this mutant was able to confer viability to cells grown at 42 °C. At the same time, GreA-37A and GreA-52A mutants displayed weak and moderate complementation phenotypes, respectively, conferring partial viability to the cells at a nonpermissive temperature. These observations suggest that complementation of the ts phenotype of greA^-/greB^- E. coli cells at 42 °C requires 24-32 times higher concentration of GreA-BPL mutant than the minimal working concentration of the wt factor. Mutations of each residue of the basic patch contribute to the loss of in vivo activity by GreA-BPL, although GreA-37A mutant appears to be more active than GreA-52A mutant under conditions of low expression and becomes much more toxic when overexpressed. The loss of in vivo activity by GreA-BPL mutants correlated with the general decrease in their in vitro readthrough and transcript cleavage activities toward backtracked TECs and their ability to cross-link to nascent RNA (see Fig. 3).

Due to the leakiness of trc promoter, only nontoxic Gre mutants with moderate functional defects could be analyzed using this system. The attempt to clone GreA-LBP3, GreA-LBP4, GreB-BPL4, GreB-BPL5, and GreB-BPL5-DD mutants into pTRC99A plasmid failed due to the appearance of secondary null mutations.² However, the fact that the low-level expression of mutant GreB factors was toxic even in the presence of chromosomal copies of wt gre genes suggests that mutations in basic patch of GreB caused severe functional defect, resulting in a dominant negative phenotype in vivo. Therefore, we conclude that the basic patch residues of GreA and GreB indeed play an essential role in the in vivo functions of these factors.

	DISCUSSION

TOP ABSTRACT INTRODUCTION MATERIALS AND METHODS RESULTS DISCUSSION REFERENCES

Previous biochemical studies have implicated the NTD of Gre proteins in the functional interactions of Gre with RNA in TECs (33, 34), the induction of nucleolytic activity of RNAP (35), and the antiarrest/readthrough activities of Gre proteins (34). In this work, we investigated the role of basic patch residues of NTD in biochemical activities of GreA and GreB in vitro and correlate it with the in vivo functions of Gre factors.

Our functional assays showed that the removal of basic patch, or even the introduction of acidic patch, does not affect the ability of Gre protein to induce cleavage in 7U-TEC but affects its cleavage activity in 9A- and 10C-TEC and significantly impairs the activity toward arrested 12/13-TEC (Fig. 2B). These results clearly demonstrate that the residues of basic patch are not responsible for directly influencing the catalytic residues of RNAP, the way alkaline pH might be thought to stimulate the endonuclease activity of RNAP. Apparently, some other amino acid residues of NTD are responsible for this function. However, the presence of basic patch residues is required in GreA and GreB for the induction of RNA cleavage in extensively backtracked and arrested TECs.

A gradual reduction in the size of basic patch results in a corresponding decrease in the efficiency of RNA-Gre cross-linking in TEC and in readthrough/antiarrest activities of Gre factors (Fig. 2C). The diminution of basic patch in GreB decreases the length of the excised 3'-terminal RNA fragment and results in gradual conversion of its activity from GreB type to GreA type (Fig. 3). These observations indicate that the basic patch residues are involved in specific electrostatic interactions with the exposed 3'-terminal portion of the nascent transcript in backtracked TECs. We propose that these interactions provide the internal phosphodiester bond of RNA a proper orientation and alignment toward RNAP catalytic center and facilitate the cleavage reaction. Increasing the size of basic patch on GreA, on the other hand, results in aberrantly strong RNA-Gre interactions and loss of cleavage and readthrough activity by GreA (Fig. 4). This result emphasizes the critical role of the specificity of RNA-Gre interactions required in the induction of nucleolytic activity of RNAP. It also implies that despite their apparent structural similarity, GreA and GreB may not interact with TEC in an entirely superimposable manner.

Above results are consistent with the "molecular ruler" hypothesis, which posits that the maximum length of cleaved RNA fragment is determined by the size of basic patch in Gre proteins. However, according to our results, the length of cleavage product and the efficiency of cleavage reaction are also governed by the extent of TEC backtracking. In minimally backtracked 7U-TEC, for instance, in which the catalytic center is 1-2 nucleotides in distance away from RNA 3'-terminus, the wt GreA, GreB, and all Gre BPL mutants, including GreB-BPL5-DD, induce cleavage of only dinucleotides and display similar specific cleavage activities (Fig. 2B). Thus, in minimally backtracked TECs, basic patch does not act as molecular ruler, and moreover, the basic patch residues are dispensable for the induction of cleavage reaction. We interpret these results to mean that in minimally backtracked TEC, tight Gre-RNA interactions are not required for efficient transcript cleavage reaction. As the extent of backtracking increases in 9A-, 10C-, and 12C/13G-TECs, the active center of RNAP moves further from the 3'-terminus of RNA (4-8 nucleotides), and an increasingly large 3'-terminal portion of RNA becomes extruded in TEC (25, 26).³ Under these conditions, the size of basic patch is shown to be a determining factor in the length of cleavage products (Fig. 3) and in the efficiency of cleavage reaction (Fig. 2B). In other words, in extensively backtracked and arrested TECs, the strength and specificity of Gre-RNA interactions play an essential role in the transcript cleavage reaction.

It should be noted that besides the basic patch residues, other surface-exposed residues of NTD may contribute to specific interactions with RNA. This follows from the observation that GreA-BPL and GreB-BPL mutants still display a low-level cross-linking to RNA and possess a residual antiarrest/readthrough function, whereas the introduction of acidic patch into GreB-BPL5 (GreB-BPL5-DD) abolishes both activities (Fig. 2, B and C).

Gre-RNA interactions most likely constitute a necessary step in the readthrough and antiarrest activities mediated by GreA and GreB. We envision that the finger-like NTD of Gre protein acts as a linker that holds the transcript in a fixed location relative to RNAP active site. This is possible because Gre itself is secured to RNAP through its CTD. By physically anchoring the RNA 3'-terminus to their basic patch, Gre factors prevent stalled (or paused) TEC from backtracking, or backtracking further. The specific NTD-RNA interactions serve another important function: they prime the internal phosphodiester bond of RNA for an efficient cleavage reaction. Thus, by promoting cleavage and resynthesis of RNA at the onset of backtracking, Gre factors increase the overall odds for elongation and, conversely, decrease the probability of continued backtracking, which may lead to transcriptional arrest. The Gre proteins lacking basic patch are unable to secure RNA and are therefore required at higher concentration than wt Gre to achieve the same efficiency of cleavage reaction. This explains why Gre BPL factors display low readthrough activity (Fig. 2C). Similar mechanism may be at work during abortive initiation of transcription cycle; by inducing cleavage, Gre factors may prevent RNAs from dissociating from the initiation complex. Increasing the time of RNA occupancy in the initiation complex may play a crucial role in facilitating the transition of RNAP from initiation to elongation stage of transcription (32). Preliminary data demonstrate that GreA-BPL indeed exhibits reduced ability to promote transition from initiation to elongation both in vivo and in vitro (data not shown).

Our views on Gre-RNAP and Gre-RNA interactions in TECs that resulted from in vitro studies of Gre-BPL mutants are generally consistent with recently published crystal structure of Thermus aquaticus RNAP (53) and with electron microscopy studies of the binding site of the Gre factor on E. coli RNAP reported earlier (36). On the one hand, the CTD of Gre factor binds to the solvent exposed surface of the RNAP molecule next to but not in the ~25-27-Å internal channel, which is thought to be occupied by double-stranded DNA and RNA-DNA hybrid, and at a distance of about 30 Å away from the location of the active site (36, 53). Furthermore, the interaction of Gre protein with RNAP appears to occur near the opening of the 10-12-Å-wide secondary channel, which is presumed to provide access to the NTP substrates to the active center (53). Because the minimum width of the NTD of GreA and GreB is about 14-16 Å (33), any insertion of the Gre N-terminal domain through secondary channel and direct interactions with the catalytic residues of the RNAP active center are unlikely on structural grounds. On the other hand, the 3'-terminal end of the nascent RNA may only be able to interact with Gre proteins if it protrudes through the secondary channel in the backtracked TEC (53). This interaction may be possible if the NTD of Gre factor occupies a corridor, near the opening of the secondary channel, formed by an extended coiled-coil subdomain of '-subunit and parts of conserved regions '-E and '-G. In this case, the basic patch of the NTD will be placed in close proximity to region '-G, in which the cross-linking to the nascent RNA 3'-terminus has been observed in backtracked TECs (54, 55).

Finally, we show that under low level of expression, BPL Gre factors are unable to complement the ts phenotype of GreA:GreB^- E. coli strain. Together with our in vitro data, this indicates that basic patch interactions with nascent transcript are important for the in vivo functions of Gre proteins. However, these interactions can be detrimental to the cell if TECs become oversaturated with Gre protein, because we observed toxic effects after overexpression of wt GreA. It is not the binding of Gre to TEC per se but rather its functional interactions with TEC that cause the toxic effect(s). This is based on the observation that the GreA-BPL mutant that binds TEC normally but have reduced RNA binding and readthrough activity did not confer a toxic effect when overexpressed (Fig. 5B). The observed toxicity may result from the reduced rate of transcription elongation and incomplete transcription of vital genes due to overstimulation of RNAP endonuclease activity. GreA lacking just one basic residue (Arg⁵² or Arg³⁷) conferred a partial or full complementation of ts phenotype under low and high levels of expression, respectively. These results are in keeping with the fact that in some bacterial organisms, GreA has only one basic residue in its basic patch (Fig. 1). In addition, an interesting correlation was observed between the readthrough activity of GreA-BPL mutants and their in vivo phenotypes. The impairment of readthrough activity of GreA mutants observed in vitro almost quantitatively parallels their loss of ability to complement the ts phenotype of greA^-/greB strain (Fig. 5). These results suggest that in vivo functions of Gre factors involve, at least in part, the readthrough activity of Gre proteins observed in vitro.

Interestingly, eukaryotic transcription elongation factor TFIIS, the functional homolog of Gre, lacks any structural or sequence homology with GreA or GreB (56, 57). However, similar to Gre factors, SII interacts with nascent RNA in TEC (58). Moreover, within the zinc ribbon domain of SII, which is functionally similar to NTD of Gre factors, basic and aromatic residues that are critical for SII functions are also implicated in electrostatic and stacking interactions with RNA or RNA-DNA hybrid (59, 60). These data suggest that specific interactions between transcript cleavage factors and nascent RNA play an important role in regulation of both prokaryotic and eukaryotic transcription elongation.

	ACKNOWLEDGEMENTS

We thank M. Anikin for helpful discussion and A. Mustaev for providing the photocross-linking reagent.

	FOOTNOTES

^* This work was supported in part by National Institutes of Health Grants GM54098 (to S. B.) and GM28946 (to A. D.).The costs of publication of this article were defrayed in part by the payment of page charges. The article must therefore be hereby marked "advertisement" in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.

^§ Present address: Dept. of Biochemistry and Molecular Pharmacology, Harvard Medical School, Boston, MA 02115.

^** To whom correspondence should be addressed: Dept. of Microbiology and Immunology, State University of New York, Health Science Center at Brooklyn, 450 Clarkson Ave., BSB 3-27, Brooklyn, NY 11203. Tel.: 718-270-3752; Fax: 718-270-2656; E-mail: serbor@asan.com.

² D. Kulish, unpublished data.

³ I. Lomakin and D. Kulish, unpublished observation.

	ABBREVIATIONS

The abbreviations used are: SII, transcription elongation factor TFIIS; RNAP, RNA polymerase; TEC, ternary elongation complex; CTD, C-terminal globular domain; NTD, N-terminal coiled-coil domain; BPL, basic patch-less; LBP, large basic patch; 8-N₃AMP, 8-azidoadenosine monophosphate; N₃-NPA-UMP, 5-(4-azido-2-nitrophenyl)-allylaminouridine monophosphate; Tricine, N-[2-hydroxy-1,1-bis(hydroxymethyl)ethyl]glycine; ts, temperature sensitivity; wt, wild type; PAGE, polyacrylamide gel electrophoresis; IPTG, isopropyl-- D-thiogalactopyranoside.

	REFERENCES

TOP ABSTRACT INTRODUCTION MATERIALS AND METHODS RESULTS DISCUSSION REFERENCES