In vitro reconstitution of transcriptional antitermination by the SacT and SacY proteins of Bacillus subtilis.

Expression of the sacPA and sacB genes of Bacillus subtilis is positively modulated by transcriptional regulatory proteins encoded by the sacT and sacY genes, respectively. Previous genetic studies led to the suggestion that SacT and SacY function as nascent mRNA binding proteins preventing early termination of transcription at terminators located in the leader regions of the corresponding genes. Here we report the overproduction, purification to near homogeneity, and characterization of the two antiterminators, SacT and SacY. Using mRNA band migration retardation assays and a reconstituted transcriptional antitermination system, the mRNA binding functions and antitermination activities of purified SacT and SacY are demonstrated under in vitro conditions. The results establish for the first time that members of the BglG family of antiterminators function in antitermination in the absence of other proteins in vitro. Purified SacT is shown to be phosphorylated by phosphoenolpyruvate in a phosphotransferase-catalyzed reaction dependent on Enzyme I and HPr. Unexpectedly, the purified SacT is shown to be functional in mRNA binding and in transcriptional antitermination independently of its phosphorylation state.

Expression of the sacPA and sacB genes of Bacillus subtilis is positively modulated by transcriptional regulatory proteins encoded by the sacT and sacY genes, respectively. Previous genetic studies led to the suggestion that SacT and SacY function as nascent mRNA binding proteins preventing early termination of transcription at terminators located in the leader regions of the corresponding genes. Here we report the overproduction, purification to near homogeneity, and characterization of the two antiterminators, SacT and SacY. Using mRNA band migration retardation assays and a reconstituted transcriptional antitermination system, the mRNA binding functions and antitermination activities of purified SacT and SacY are demonstrated under in vitro conditions. The results establish for the first time that members of the BglG family of antiterminators function in antitermination in the absence of other proteins in vitro. Purified SacT is shown to be phosphorylated by phosphoenolpyruvate in a phosphotransferasecatalyzed reaction dependent on Enzyme I and HPr. Unexpectedly, the purified SacT is shown to be functional in mRNA binding and in transcriptional antitermination independently of its phosphorylation state.
Induction of the sac (sucrose utilization) genes of Bacillus subtilis and the bgl (␤-glucoside utilization) operon of Escherichia coli is controlled by transcriptional antitermination (Crutz et al., 1990;Débarbouillé et al., 1990;Amster-Choder and Wright, 1993). Within the B. subtilis sac system is the sacB gene encoding levansucrase and the sacPA operon encoding the sucrose permease and a phosphosucrase. In these systems, transcription is initiated at constitutive promoters, and expression is regulated by controlled read-through at transcriptional terminators located between the promoters and the first structural genes of the operons. Gene-specific antiterminators are required to prevent early transcriptional termination.
Three genes within the bgl operon of E. coli, bglG, bglF, and bglB, are involved in aromatic ␤-glucoside utilization (Schnetz et al., 1987;Schnetz and Rak, 1988). BglG is a positive regulator that recognizes a specific RNA sequence located upstream of the terminator of the bgl operon. Binding of BglG to a specific secondary structural element on the bgl mRNA prevents transcriptional termination by blocking the formation of a terminator (Houman et al., 1990). BglF, the ␤-glucoside permease, is an Enzyme II of the phosphoenolpyruvate:sugar phosphotransferase system (PTS) 1 with a IIBCA domain structure . It also functions as a negative transcriptional regulator, controlling the activity of BglG by phosphorylation-dephosphorylation in response to the external level of inducer. The degree of phosphorylation of BglG in vivo is believed to be dependent on the cellular concentration of BglF which serves as the direct phosphoryl donor and acceptor in the antiterminator kinase-and phosphatase-catalyzed reactions. The antiterminator is active in the nonphosphorylated state because this form of the protein exhibits a dimeric structure while the phosphorylated form is monomeric and therefore inactive (Amster-Choder et al., 1989;Amster-Choder and Wright, 1992).
In B. subtilis two regulatory genes, sacT and sacY, encode proteins that respectively control transcription of the sacPA and sacB genes. Their protein products are homologous to BglG and apparently function by an analogous antitermination mechanism (Houman et al., 1990;Débarbouillé et al., 1990;Aymerich and Steinmetz, 1992). The PTS has been shown to regulate both the sacPA and sacB genes in vivo, and unphosphorylated SacY appears to be responsible for high level expression of these genes in mutants lacking Enzyme I (pstI) or HPr (ptsH). Conversely, in ptsH and ptsI mutants, SacT is nonfunctional, indicating that phosphorylation is required for its activity (Arnaud et al., 1992).
On the basis of these observations, we have postulated that SacT may possess two phosphorylation sites (Arnaud et al., 1992). HPr, the second phosphocarrier protein of the PTS, is phosphorylated by Enzyme I, and HPr(HisϳP) possibly phosphorylates the antiterminator SacT on one site as a prerequisite for its activity. Then, if SacT is phosphorylated at a second site in the absence of sucrose, inactivation may result as pro-posed for BglG and SacY (Schnetz and Rak, 1990;Amster-Choder et al., 1989;Crutz et al., 1990). This second phosphorylation event would depend on the presence of an as yet unidentified sucrose sensor protein, functionally similar to the sucrose-specific permease of the PTS. In the presence of sucrose, only this second site in SacT would become dephosphorylated, and SacT would thereby be activated when sucrose is translocated into the cell. Alternative mechanisms have been proposed (Arnaud et al., 1992).
In this report we describe the overproduction, purification, and characterization of SacT and SacY. We show that purified SacT and SacY bind to a specific sacPA mRNA sequence, and we describe a reconstituted SacY-or SacT-dependent transcriptional antitermination system which is functional in vitro. Finally, we demonstrate that SacT is phosphorylated by PEP in a process that depends only on Enzyme I and HPr as previously deduced from genetic studies. Altogether, the data presented herein serve to establish that both SacT and SacY serve as transcriptional antiterminators in the absence of auxiliary proteins.
Site-directed mutagenesis and introduction of restriction sites in the sacT gene was performed using the Muta-gene M13 in vitro mutagenesis kit (Bio-Rad). The oligonucleotide 5Ј-CTT TAT AGA TTT TCA TAT GGT TAT-3Ј was used to introduce an NdeI site (underlined) in the initiation codon of sacT, and the oligodeoxynucleotide 5Ј-CGT TAC GGT CGA CAA TCT GAC CCC GGG ATT-3Ј was used to introduce SalI and SmaI sites downstream of the translational stop codon of sacT. The modified sacT gene was then cloned between the NdeI and SalI sites of the overexpression vector pRE-1  to generate pRE1-T1, and between the NdeI and XhoI sites of pET19b (Novagen) to generate pET19b-T1. Site-directed mutagenesis of sacY was performed essentially as described previously (Sutrina et al., 1990) after cloning the ClaI-NruI fragment (ϳ1560 base pairs) of pMF20 (Arnaud et al., 1992) between the ClaI and EcoRV sites of pBluescript (Stratagene, La Jolla, CA). The oligodeoxynucleotide 5Ј-GGG GGA TGA AAG GAC AAA AAA CAT ATG AAA ATT AAA-3Ј was used to introduce an NdeI site in the initiation codon of the sacY gene, and the modified gene was then cloned between the NdeI and BamHI sites of pRE-1 and pET19b to generate plasmids pRE1-Y1 and pET19b-Y1, respectively.
Plasmid pPA2 was constructed as follows: an AccI-SspI fragment located upstream of the sacPA operon and containing the sacPA promoter, the palindromic terminator structure, and 19 codons of sacP was first isolated from plasmid pTP7 (Débarbouillé et al., 1990). This fragment was cloned between the AccI and SmaI sites of pAM18 (paired promoter T7 system, Amersham Corp.). Plasmid pPA3, similar to pPA2, was obtained by cloning the XmnI-Sau3A fragment containing the same promoter and the upstream half of the terminator between the SmaI and BamHI sites of pAM19. The XmnI-Sau3A DNA fragment was isolated from plasmid pBSG8 -34. In this plasmid, the original Sau3A restriction site was replaced by a BamHI site (Arnaud et al., 1992).
Previously described growth conditions of E. coli MZ1 containing the appropriate pRE1 derived overexpression vector were used to overproduce SacT, SacY, the PTS proteins, Enzyme I and HPr, and the IIA Glc protein domain of B. subtilis (Reizer et al., 1989Sutrina et al., 1990). For overproduction of SacT and SacY in E. coli BL21(DE3) bearing the overexpression vector pET19b-T1 or pET19b-Y1, 1 mM isopropyl-1-thio-␤-D-galactopyranoside was added at A 600 of 0.4 -0.7 to cultures grown in LB medium (at 37°C), and incubation was then continued for an additional 3 h. Cells were then harvested by centrifugation, washed twice with 50 mM Tris-HCl buffer, pH 8.5, containing 0.1 mM phenylmethylsulfonyl fluoride and 10% glycerol (TPG buffer), and ruptured by two passages through an Aminco French pressure cell at 10,000 p.s.i.
Protein Purification-Initial attempts to purify SacT and SacY from E. coli MZ1 containing pRE1-T1 or pRE1-Y1, respectively, resulted in precipitation of the partially purified antiterminators. Precipitation usually occurred during or after concentration by ultrafiltration (YM10, Amicon, Inc.) of the protein preparations that had been partially purified by ion-exchange chromatography using a DEAE-Sephacel column (see below). Because a similar concentrating step was not required during purification of the two antiterminators from extracts of E. coli BL21(DE3) bearing the pET19b-derived plasmids, subsequent purification of these proteins was carried out following overproduction of SacT and SacY in E. coli BL21(DE3) bearing pET19b-T1 and pET19b-Y1, respectively. The following purification protocol was applied to both SacT and SacY. A crude extract derived from cells grown in 6 -8 liters of culture medium was loaded onto a column of DEAE-Sephacel (200 ml bed volume) which had been pre-equilibrated with TPG buffer. The column was washed with 600 ml of the same buffer, and proteins were then eluted with a linear salt concentration gradient (2000 ml of 0 -1 M NaCl) in TPG buffer. Fractions (15 ml) were collected and assayed by SDS-polyacrylamide gel electrophoresis for the presence of SacT or SacY which eluted as a broad peak at about 0.35-1 M NaCl. Fractions containing His-tagged SacT or His-tagged SacY were pooled and loaded onto an immobilized Ni 2ϩ column (His-Bind resin, Novagen) pre-equilibrated with 20 mM Tris-HCl buffer, pH 7.9, containing 0.5 M NaCl, 10% glycerol, and 5 mM imidazole. The column was washed with 20 mM Tris-HCl buffer, pH 7.9, containing 0.5 M NaCl, 10% glycerol, and 60 mM imidazole, and the His-tagged antiterminator was eluted with the same buffer with the imidazole concentration increased to 1 M. The purified proteins were then dialyzed against 50 mM Tris-HCl buffer, pH 8.5, containing 20% glycerol, 0.5 mM EDTA, and 0.5 M NaCl and stored at Ϫ70°C until used.
Enzyme I, HPr, and the recombinant IIA Glc protein domain of the B. subtilis Enzyme II Glc were overproduced and purified as described previously (Reizer et al., 1989Sutrina et al., 1990).
In Vitro RNA Synthesis-The two plasmids, pPA2 and pPA3, which contain the sacPA promoter region (see Fig. 3), were used as templates for RNA probe synthesis. The plasmids were linearized with EcoRI or HindIII, respectively, and extracted with phenol before conducting in vitro RNA synthesis experiments. RNA synthesis was performed using the paired promoter T7 system kit (Amersham Corp.) following the manufacturer's instructions. Transcription mixtures used for probe synthesis contained 100 Ci of ␣-[ 33 P]UTP and 12.5 M UTP. The RNA probes were purified after migration on 8% denaturing polyacrylamide gels and resuspended in 10 mM Tris, pH 7.5, containing 1 mM EDTA and 25 g/ml yeast tRNA.
Phosphorylation of SacT-Assay mixtures for the PTS-dependent phosphorylation of SacT contained, unless otherwise indicated, 4 g of SacT, 2 g of HPr, 1.6 g of Enzyme I, and 12.5 mM PEP in a final volume of 20 l of binding buffer (10 mM HEPES, pH 7.6, containing 1 mM EDTA, 5 mM MgCl 2 , and 2 mM dithiothreitol). The reaction mixtures were incubated for 15 min at 37°C and transferred to an ice bath. They were used within a few minutes for binding assays.
RNA Binding Assays-The RNA transcripts (10 4 cpm/assay in 4 l of binding buffer) with 1 unit of human placenta RNase inhibitor) were incubated at 80°C for 3 min and transferred to an ice bath. The native or phosphorylated SacT or SacY (in 10 l of phosphorylation reaction mixture) was mixed with the RNA probes at 4°C, incubated at 16°C for 10 min, and finally incubated at 0°C for 10 min. Glass-distilled glycerol (final concentration 10%, Eastman Kodak Co.) was added before loading the samples onto an 8% nondenaturing polyacrylamide gel with a cross-linking ratio of 29:1 in 0.375 M Tris buffer (pH 8.8). Prior to loading, the gels were chilled for 1 h and prerun (300 V) for 1 h at 4°C. Electrophoresis for 2 h at 4°C was carried out at 300 V using 0.025 M Tris-HCl containing 0.2 M glycine as running buffer.
In Vitro Transcription Assay-Transcription was assayed in a reaction mixture containing 40 mM Tris-HCl, pH 7.9, 10 mM MgCl 2 , 100 mM KCl, 1 mM dithiothreitol, 5 nM template DNA, 0.2 M RNA polymerase holoenzyme of E. coli or B. subtilis, 200 M GTP, CTP, and ATP, 50 M UTP, and 5 Ci of ␣-[ 32 P]UTP (3000 Ci/mmol) in a final volume of 30 l. Assay mixtures were then incubated at 65°C for 10 min and loaded onto a 7% polyacrylamide sequencing gel (Sanger and Coulson, 1978). Radioactivity in bands was quantified using a PhosphorImager (Molecular Dynamics).
Other Methods-Polyacrylamide gel electrophoresis (SDS-polyacrylamide gel electrophoresis) was performed with a Pharmacia Phast System or as described before . [ 32 P]PEP was prepared by the method of Matoo and Waygood (1983). Proteins labeled with [ 32 P]PEP were separated by SDS-polyacrylamide gel electrophoresis and detected by autoradiography as described previously (Reizer et al., 1989). Protein was determined by the Lowry method (Lowry et al., 1951) with bovine serum albumin as the standard protein. Fig. 1A, the overproduced SacY or SacT represents the major protein band in an extract of E. coli BL21(DE3) bearing pET19b-Y1 or pET19b-T1, respectively. Following a two-step purification protocol that consisted of (a) ion-exchange chromatography on a DEAE-Sephacel column and (b) affinity chromatography of the His-tagged proteins on a His-Bind metal (e.g. Ni 2ϩ ) chelation resin (see "Experimental Procedures") nearly electrophoretically pure preparations of SacT and SacY were obtained (Fig. 1A). The single faint band that migrated faster than SacT or SacY is most likely a proteolytic product of the major protein band, since its mobility relative to SacT remained unchanged under modified polyacrylamide gel electrophoresis (Fig. 1B), and it was phosphorylated by the PTS as was SacT (see below). The purified SacY migrated on SDSpolyacrylamide gels as a band with an apparent molecular mass of ϳ34,000 Da, in agreement with the molecular mass of 32,465 Da calculated from the deduced amino acid sequence of the protein. Interestingly, purified SacT migrated on SDSpolyacrylamide gels as a band with an apparent molecular mass of ϳ30,000 Da, although the molecular mass of SacT calculated from the deduced amino acid sequence (32,074 Da) is similar to that of SacY. One additional prominent band, corresponding to the homodimeric form of SacT (ϳ68,000 Da), was detected in SDS-polyacrylamide gels if the purified SacT protein was not boiled before loading the gels and the reducing agent ␤-mercaptoethanol was omitted from the sample buffer (Fig. 1B). At increasing concentrations of dithiothreitol (1-10 mM) in the sample buffer the amounts of monomeric SacT gradually increased at the expense of its homodimeric form (Fig. 1B). These findings provide evidence for a dimeric SacT.

Overproduction and Purification of SacT and SacY-As shown in
PTS-dependent Phosphorylation of SacT-An autoradiogram showing the phosphorylation of SacT by purified Enzyme I and HPr in the presence of [ 32 P]PEP is shown in Fig. 2. In lane 1, the purified Enzyme I, HPr, and IIA Glc of B. subtilis were incubated with [ 32 P]PEP, and only these three proteins were labeled. In lane 2, Enzyme I and SacT were incubated with [ 32 P]PEP, and only the former protein was labeled. By contrast, SacT was readily phosphorylated when incubated in the presence of Enzyme I, HPr, and [ 32 P]PEP (lane 3). As

FIG. 2. Phosphorylation of SacT by HPr.
Autoradiogram of an SDS-polyacrylamide gel showing 32 P-labeled protein products obtained by phosphorylation reactions containing the proteins indicated at the bottom of the autoradiogram for the following lanes: lane 1, Enzyme I (0.6 g), HPr (2.6 g), and IIA Glc (1.5 g); lane 2, Enzyme I and SacT (4 g); lane 3, Enzyme I, HPr, and SacT. Reaction mixtures with unspecified amounts of the indicated proteins contained the amounts of proteins indicated above for previous lanes. The phosphorylation reaction (at 37°C for 15 min; 30 l final volume) contained 50 mM Tris-HCl buffer (pH 7.5), 5 mM MgCl 2 , 2 mM dithiothreitol, and 0.5 mM [ 32 P]PEP (specific activity, 10,000 cpm/nmol). expected for a histidyl or cysteyl phosphorylated protein, the 32 P-labeled SacT was labile under acidic conditions but stable under alkaline conditions (data not shown). These observations establish that SacT can serve as a phosphoryl acceptor with the general energy-coupling protein, HPr(HisϳP) serving as the phosphoryl donor. Although the phosphorylation of SacT by HPr(HisϳP) is in agreement with previously published genetic data, we do not ignore the possibility that an additional site in SacT may be phosphorylated by an unidentified sucrose-sensor protein of the PTS (Arnaud et al., 1992).
SacT and SacY Are mRNA Binding Protein-The binding of BglG to bgl mRNA has been studied by transferring protein extracts containing the unphosphorylated form of BglG onto a membrane and probing with mRNA transcripts. These experiments led to the conclusion that BglG dimers bind to a specific mRNA sequence (Houman et al., 1990). Similar mRNA sequences referred to as ribonucleic antiterminator sequences were identified as SacT and SacY binding sites (Aymerich and Steinmetz, 1992). Ribonucleoprotein complexes can be visualized on the basis of their altered mobility during gel electrophoresis compared to that of the corresponding nucleic acid alone. The binding of SacT to RNA was studied using this procedure. Since we previously showed that SacT needs the general components of the PTS for its activity in vivo (Arnaud et al., 1992), purified SacT was phosphorylated first with PEP in the presence of Enzyme I and HPr and then incubated with an RNA probe containing the promoter and terminator region upstream of sacPA (probe 1, Fig. 3). Employing electrophoresis in native polyacrylamide gels, no RNA-protein complex was detected under these conditions. Similar results were obtained with probe 2, which contains the upstream half of the termi-nator. These results may have been due to formation of the stem-loop structure that interferes with the binding of SacT as has been previously proposed for BglG (Houman et al., 1990). Indeed, the formation of mRNA-SacT complexes was observed after heat denaturation of the probes at 80°C as described under "Experimental Procedures." As shown in Fig. 4 (lanes  1-5), two distinct mRNA-protein complexes were clearly detected when a high concentration of SacT was incubated with probe 1. An mRNA-protein complex was also observed with probe 2 (lanes 6 -10), and heat denaturation of the probe was similarly required as had been observed with full-length probe 1. Competition experiments were then carried out by addition of unlabeled RNA probe 2 to a binding assay mixture containing the same radiolabeled probe. As shown in Fig. 4 (lanes  10 -15), when the concentration of the unlabeled competitor was increased, the labeled probe was released progressively from the RNA-protein complex. The same experiment was carried out with unlabeled probe 1 leading to the release of the corresponding labeled probe 1 (data not shown). In order to test whether the in vitro phosphorylation of SacT was required for RNA binding, gel shift experiments were performed using the native SacT. Results presented in Figs. 4 and 5 demonstrate that both the native and the phosphorylated SacT can bind both probes.
Previous studies suggested that SacT and SacY have similar mRNA targets and that the sacPA operon is constitutively expressed in a ⌬sacT,⌬pts(GHI) double mutant due to a SacYdependent function (Débarbouillé et al., 1990;Aymerich and Steinmetz, 1992). Similar binding assays were performed using purified SacY. Increasing amounts of SacY were incubated with either labeled probe 1 or labeled probe 2 (Fig. 6). Two  (Aymerich, 1992). The sacPA promoter was previously mapped into the XmnI-SspI 300-base pair DNA fragment (Débarbouillé et al., 1990). Boxed sequence denotes the position of the transcriptional start site that was estimated by comparing the lengths of the two transcripts with the DNA sequence of the sacPA promoter region. B, schematic diagram of the upstream region of the sacPA operon. RNA probe 1 and probe 2 correspond to the AccI-SspI DNA fragment and the XmnI-Sau3A fragment, respectively (see "Experimental Procedures"). The AccI-SspI DNA fragment was cloned into pAM18 generating pPA2, while the BamHI-XmnI fragment was cloned into pAM19 providing pPA3. These two plasmids served as templates for the synthesis of RNA probes used in the binding assays. Convergent arrows indicate the palindromic structure of the SacT-dependent terminator. ribonucleoprotein complexes were observed using either probe 1 (lanes 1-5) or probe 2 (lanes 6 -10). Retardation of the two RNA probes by SacT or SacY is due to specific RNA binding activities of the purified proteins, since no ribonucleoprotein complex was observed in the presence of irrelevant proteins, i.e. Enzyme I and HPr (Fig. 5, lanes 5 and 6), and all band migration retardation assays were performed in the presence of excess amounts of nonspecific tRNA. Altogether, the data demonstrate that the reported mRNA migration retardations represent specific recognition of the probes by SacT and SacY. They further indicate that PTS-catalyzed phosphorylation of SacT is not required for interaction in vitro with the two RNA probes used.
SacT and SacY Are Transcriptional Antiterminators in Vitro-SacY and SacT are homologous to BglG (see Introduction), and genetic evidence strongly suggests that all three proteins are transcriptional antiterminators (Aymerich and Steinmetz, 1987;Zukowski et al., 1990;Débarbouillé et al., 1990). The availability of purified SacT and SacY allowed direct determination of their transcriptional antitermination activities in vitro. Plasmid pTP7 contains a DNA fragment of 305 base pairs bearing the upstream region of the sacPA operon including the promoter and the SacT-regulated terminator (Débarbouillé et al., 1990). In this plasmid, a single BamHI restriction site is located about 80 base pairs downstream of the transcriptional terminator in the sacP coding sequence. The DNA of pTP7 was linearized with BamHI and then used as template in the in vitro transcription assays as described under "Experimental Procedures." As expected for RNA transcriptional termination at the palindrome, a short transcript of 121 bases was observed in the presence of the RNA polymerase core enzyme associated with A from B. subtilis. By contrast, addition of either SacT or SacY to the in vitro transcription mixture readily promoted the synthesis of a long, full-length transcript (212 bases) corresponding to the full length of the sacPA DNA template (Fig. 7, left panel). These experiments clearly show that both SacT and SacY are transcriptional antiterminators that can function under in vitro conditions independently of their modification by phosphorylation.
The latter conclusion was supported by findings demonstrating the lack of a significant effect on the antitermination activity following phosphorylation of SacT by Enzyme I, HPr, and PEP (data not shown). Since this finding is in apparent conflict with the previously published model deduced from in vivo genetic data (Arnaud et al., 1992), we considered the possibility that SacT had been phosphorylated during overproduction by the resident PTS of E. coli. To examine this possibility, the putative SacT-P was preincubated with approximately 20-fold excess of purified HPr and subsequently assayed for transcriptional antitermination activity. Since the characterized HPr- mediated phosphorylation reactions are all reversible, it is probable that excess HPr would effect removal of phosphate from the putative SacT-P. Nevertheless, the results obtained with the HPr-treated antiterminator were similar to those obtained with the untreated SacT (Fig. 7, right panel). While the apparent conflict between the presently demonstrated antitermination activity of unphosphorylated SacT and the previously deduced phosphorylation-dependent antitermination mechanism is not yet resolved, our data clearly establish that purified SacT and SacY are both transcriptional antiterminators that can function under in vitro conditions in the absence of auxiliary proteins.
Increasing concentrations of SacT were used in the in vitro transcription assay, and antitermination activity expressed as percentage of the full-length transcript was estimated using a PhosphorImager apparatus. As shown in Fig. 8, the amount of the full-length transcript increased with the amount of SacT used up to ϳ2.5 M. These results indicate that the purified SacT was saturating in the in vitro assay mixture.
The ability of SacT to antiterminate in the presence of a heterologous RNA polymerase, i.e. RNA polymerase from E. coli, was similarly tested in the in vitro transcription system. Increasing amounts of SacT were incubated in the presence of E. coli RNA polymerase, and the products were quantified as described above. As shown in Fig. 8, SacT suppressed transcriptional termination in the presence of either E. coli or B. subtilis RNA polymerase.

DISCUSSION
The data described in this work establish that both SacT and SacY are transcriptional antiterminators which can regulate expression of the sacPA operon due to interaction with a specific mRNA sequence element within the sacPA leader region. Reported herein are (a) the overproduction and purification of two homologous transcriptional regulatory proteins of B. subtilis, SacT and SacY; (b) biochemical evidence demonstrating that SacT is phosphorylated by PEP in a PTS-catalyzed reaction that is dependent on the general energy coupling proteins, Enzyme I and HPr; (c) binding assays demonstrating that both SacT and SacY are mRNA binding proteins; and (d) in vitro reconstitution of SacT-and SacY-dependent transcriptional antitermination activity. We focused our investigation on SacT because previously documented molecular genetic data suggested that the PTS exerts dual control over the antitermination activity of this protein as compared to PTS-mediated control of the two homologues, SacY and BglG (Arnaud et al., 1992). While evidence was previously presented demonstrating that the activities of SacY and BglG are negatively modulated by the PTS due to their phosphorylation by the respective sugar specific permeases, SacX and BglF, the activity of SacT was deduced to be under positive control by the general energy coupling protein HPr or a phosphorylated protein thereof (see Introduction and references cited therein). Our data establish that SacT is phosphorylated by HPr(HisϳP) by the sequential relay of phosphate through the PTS. We emphasize, however, that reversible phosphorylation of SacT by this general energy coupling protein of the PTS is not sufficient to confer sucrosespecific induction of the sacPA operon.
The N antiterminator protein of coliphage lambda, similarly to SacT and SacY, recognizes a sequence in the nascent RNA. This protein interacts with RNA polymerase to suppress transcriptional termination. In this case, control signals are transmitted to RNA polymerase via protein-protein interactions and RNA looping (Greenblatt et al., 1993;Friedman and Court, 1995). In addition to N, four host factors, NusA, NusB, NusE, and NusG, facilitate antitermination in vivo. Interestingly, only one factor, NusA, is required for in vitro antitermination at moderate N concentrations, while at high N concentrations NusA is not required (Rees et al., 1996). The results reported here contrast therefore with those obtained for the coliphage N protein in that no additional factors proved to be required for in vitro antitermination. To the best of our knowledge, this is the first report of a reconstituted in vitro antitermination system using a BglG homologue.
An unexpected finding reported in the present communica- tion is the apparent similar antitermination activity observed with unphosphorylated and phosphorylated forms of SacT. This finding is in apparent conflict with previously published in vivo data, suggesting that the effective antiterminator is SacT(HisϳP) rather than the unphosphorylated form of the protein (Arnaud et al., 1992). It is unlikely that purified SacT was active in antitermination due to its phosphorylation by the resident PTS of E. coli during overproduction, since this phosphorylation reaction is reversible as are all characterized HPrmediated phosphorylation reactions, and preincubation of the purified protein with a 20-fold excess of HPr did not affect antitermination. We are tempted to suggest that the comparable antitermination activity observed with the two forms of SacT is due to the oligomeric state of the purified protein. Thus, a previous study with BglG showed that PTS-catalyzed reversible phosphorylation of the protein regulates its activity by modulating its dimeric state and that the unphosphorylated dimer of BglG, rather than the phosphorylated monomer, binds to its RNA target sequence (Amster-Choder and Wright, 1992). While a similar phosphorylation-dependent mechanism may modulate the activity of SacT, we note that our purified (unphosphorylated) SacT and SacY preparations appear to contain both dimers and monomers (Fig. 1B). 2 Indeed, the distinct ribonucleoprotein complexes formed using SacT or SacY may be due to the presence of specific oligomeric forms of these antiterminators bound to the mRNA probes used . Consequently, if in vivo modulation of antitermination by SacT is dependent upon the dimeric state of the protein, it is conceivable that no significant difference in antitermination by SacT and the phosphorylated SacT would be detected under our in vitro assay conditions. While preliminary studies showed that the presence of high concentration of dithiothreitol (10 mM) did not abolish RNA binding activity by the unphosphorylated form of SacT, we cannot eliminate the possibility that residual dimeric species are responsible for this activity. Additionally, we do not ignore the possibility that SacT as well as the phosphorylated SacT both are capable of promoting antitermination in vivo, as reported here with the in vitro reconstituted system, although phosphorylation of the protein may enhance productive interaction with the ribonucleic antiterminator target sequence and/or facilitate antitermination activity. Indeed, recent in vivo studies have shown that the PTS dependence of the antitermination activity mediated by LicT is completely abolished upon overexpression of the licT gene (Krü ger and Hecker, 1995). Similarly, the PTS dependence of LevR activity is strongly suppressed upon overproduction of LevR (Stü lke et al., 1995). The two possibilities noted above are not mutually exclusive, since phosphorylation independent formation of efficiently antiterminating dimers can occur at high concentration of the protein. Other models can be entertained. Using the reconstituted antitermination assay system, our cur-rent studies are designed to examine these possibilities and to reconcile the genetic and biochemical data with respect to the physiologically relevant form of SacT.