INTRODUCTION

Top Abstract Introduction Materials and Methods Results Discussion References

Expression of the Bacillus subtilis tryptophan biosynthetic genes is regulated in response to changes in the intracellular level of tryptophan by TRAP, the trp RNA-binding attenuation protein (5, 9, 17, 24, 30, 35), which is the product of the mtrB gene (17). (For a recent review, see reference 4.) Examination of the crystal structure of TRAP has revealed that it consists of 11 identical subunits arranged in a single ring structure termed the -wheel (2, 3). Cooperative TRAP activation occurs by binding of one tryptophan molecule between every two adjacent subunits (3, 6).

The 203-nucleotide (nt) untranslated trpEDCFBA operon leader transcript contains inverted repeats that allow folding of the transcript to form three RNA secondary structures (Fig. 1). Two of these structures, the antiterminator and an intrinsic terminator, overlap by 4 nt and therefore are mutually exclusive (Fig. 1). The TRAP binding target in the trp leader transcript consists of 11 closely spaced (G/U)AG repeats, 6 of which are present within the antiterminator (3, 7). When activated by tryptophan, 11 KKR motifs that outline the periphery of the TRAP complex interact with the 11 triplet repeats, presumably due to extensive hydrogen bond formation (42). Thus, TRAP binding to the nascent trp leader transcript blocks formation of the antiterminator structure, thereby allowing formation of the overlapping terminator, and hence promoting transcription termination upstream of the trp structural genes. In the absence of TRAP binding, the antiterminator structure forms, which results in transcription of the entire operon (5, 35).

View larger version (18K): [in this window] [in a new window]   FIG. 1.   Nucleotide sequence of the B. subtilis trp leader transcript showing the 5' stem-loop and the mutually exclusive antiterminator and terminator structures. Boxed nucleotides mark overlapping segments of the competing secondary structures. The (G/U)AG repeats known to be involved in TRAP-RNA recognition are boldfaced. Numbering is from the start of transcription. RNA secondary-structure predictions were performed by using MFOLD (40, 43).

In addition to the transcription attenuation mechanism described above, TRAP also regulates translation of trpE and the unlinked trpG gene. TRAP-mediated translational control of trpE can occur by a novel RNA conformational switch mechanism (15). TRAP binding to trp operon readthrough transcripts promotes formation of an RNA hairpin that sequesters the trpE Shine-Dalgarno sequence (15, 24, 27). The trpG gene of B. subtilis is involved in the biosynthesis of both tryptophan and folic acid (23, 37). TRAP regulates TrpG synthesis by binding to nine trinucleotide repeats that surround and overlap the trpG Shine-Dalgarno sequence (7, 16, 41). Thus, TRAP binding directly blocks ribosome access to the trpG ribosome binding site (16).

In addition to the antiterminator and terminator structures, an RNA hairpin is predicted to form at the extreme 5' end of the B. subtilis trp leader transcript (Fig. 1). Similar 5' stem-loop structures appear to be conserved in Bacillus pumilus, Bacillus caldotenax, and Bacillus stearothermophilus (Fig. 2). The trp operon leader transcripts of these bacilli also contain multiple triplet repeats, as well as overlapping antiterminator and terminator structures, suggesting that all four organisms control expression of the trp operon by similar transcription attenuation mechanisms (12, 20, 27).

View larger version (17K): [in this window] [in a new window]   FIG. 2.   Comparison of the 5' stem-loop structures predicted to form in the trp operon leader transcripts of B. subtilis (18), B. pumilus (20), B. caldotenax (36), and B. stearothermophilus (12). Numbering is from the start of transcription. The number below each structure indicates the calculated free energy of formation (G) in kilocalories per mole. Arrows associated with the B. subtilis structure indicate the single-nucleotide substitutions introduced into the 5' stem-loop. Asterisks mark the double compensatory changes that restore the predicted secondary structure. The 5' stem-loop deletion removed nt 3 to 32.

Since the 5' stem-loop and the attenuation mechanisms appear to be evolutionarily conserved, we were interested in determining if the 5' stem-loop affects expression of the B. subtilis trp operon. By examining the effect of a 5' stem-loop deletion and several point mutations within the 5' hairpin, we found that disruption of the structure resulted in a dramatic increase in trp operon expression and a substantial reduction in the ability of B. subtilis to regulate expression of this operon. Moreover, our results indicate that the 5' stem-loop functions in the transcription attenuation mechanism by increasing the affinity of TRAP for trp leader RNA. Our results suggest that TRAP-5' stem-loop interaction increases the probability that TRAP will bind to the (G/U)AG repeats before the antiterminator can form, thereby increasing the likelihood that transcription termination will occur before RNA polymerase can reach the trp operon structural genes.

	MATERIALS AND METHODS

Top Abstract Introduction Materials and Methods Results Discussion References

Bacterial strains and plasmids. All of the B. subtilis strains used in this study are listed in Table 1. Plasmids pTZ18R and pTZ18U were obtained from United States Biochemical Corp. ptrpBGI-PLK (27) was used to construct all of the translational fusion integration plasmids for B. subtilis. This plasmid is a derivative of ptrpBGI (34) in which the EcoRI-HindIII fragment containing the B. subtilis trp promoter and leader region has been replaced by the polylinker of pTZ18R, leaving a promoterless lacZ gene. ptrpBGI-PLK is designed to allow integration of essentially any trp promoter and leader construct into the amyE locus of the B. subtilis chromosome such that lacZ expression is under the control of the trp leader construct. pHY300PLK is an Escherichia coli-B. subtilis shuttle vector (22). Plasmid pSI45 is a derivative of pHY300PLK that carries the mtrAB operon under the control of its natural promoter (17).

Bacterial growth, transformation, and DNA isolation. B. subtilis and E. coli strains were routinely grown on L agar, L broth, or minimal acid casein hydrolysate (ACH) medium (38). Plasmid (11) and chromosomal (25) DNA was isolated by standard procedures. E. coli (32) and B. subtilis (1) transformations were performed as described previously. Appropriate antibiotics were added, as needed, to the following concentrations: ampicillin, 100 µg/ml; tetracycline, 10 µg/ml; chloramphenicol, 5 µg/ml.

-Galactosidase assays. Bacterial cultures were grown in minimal ACH medium in the presence or absence of 50 µg of L-tryptophan/ml supplemented with the appropriate antibiotics. Cells were harvested during late-exponential growth (110 Klett units, filter no. 54; Klett Manufacturing Co., Inc.). Aliquots were then assayed for -galactosidase activity by the method of Miller (28).

Gel mobility shift assay. TRAP purification was performed as described previously (5). Gel-purified transcripts used in this analysis were synthesized by using the Ambion MEGAscript in vitro transcription kit. 5'-end-labeled RNAs were generated by treating in vitro-generated transcripts with calf intestinal phosphatase and subsequently with polynucleotide kinase and [-³²P]ATP. The unlabeled and labeled RNA was gel purified as described previously (15). The binding affinity between TRAP and trp leader RNA was estimated by using gel mobility shift assays designed by modifying a published procedure (30). The binding data was fit to a nonlinear least-squares algorithm. Transcripts used in the analysis were generated from pPB77 (wild type) or pPB310 (5' stem-loop deletion) that had been linearized with BamHI. Binding reaction mixtures (40 µl) containing 0.2 nM 5'-end-labeled RNA, various concentrations of TRAP (TRAP excess), and 1 mM L-tryptophan in TKM buffer (40 mM Tris-HCl [pH 8.0], 250 mM KCl, 4 mM MgCl₂) were incubated at 25°C for 20 min. Aliquots of reaction mixtures were fractionated through 6% native polyacrylamide gels. Electrophoresis was performed at 4°C and in 0.5× Tris-borate-EDTA. Gels were dried, and the bound and free RNA bands were quantified by using a Phosphorimager (Molecular Dynamics) and the ImageQuant software package. Modifications of the standard reaction are described in the text or the appropriate figure legend.

In vitro transcription attenuation assay. In vitro transcription attenuation assays followed a previously published procedure (5). Briefly, transcription reaction mixtures contained B. subtilis vegetative (^A) RNA polymerase, 20 nM DNA template, TRAP (various concentrations), 1 mM L-tryptophan, and ribonucleoside triphosphates (2.7 mM ATP, 0.7 mM CTP, 1.1 mM GTP, 1.4 mM UTP) of which UTP was radioactively labeled. EcoRI-HindIII restriction fragments that contained the B. subtilis trp promoter and leader region from plasmids pPB22 and pJY2 were used as DNA templates in transcription reactions. Reactions were carried out at 30°C for 30 min. Samples were fractionated through 6% polyacrylamide gel electrophoresis (PAGE) gels containing 7 M urea. Radiolabeled RNA bands were quantified with a Phosphorimager (Molecular Dynamics, Inc.) and the ImageQuant software package. Modifications of the standard assay are described in the text or figure legends.

	RESULTS

Top Abstract Introduction Materials and Methods Results Discussion References

The 5' stem-loop is involved in regulating expression of the trpEDCFBA operon. A stem-loop structure is predicted to form at the 5' end of trp operon transcripts in B. subtilis, B. pumilus, B. caldotenax, and B. stearothermophilus (Fig. 2). Since the trp operon transcription attenuation mechanism appears to be conserved in all four of these organisms (12, 20, 27), we were interested in determining if the 5' stem-loop is involved in regulating expression of the B. subtilis trp operon. Accordingly, we deleted the DNA corresponding to the 5' stem-loop from the trp operon leader. We constructed B. subtilis strains containing trpE'-'lacZ translational fusions that were controlled by the wild-type (trpL_WT) or 5'S-L trp leader and analyzed -galactosidase expression when each strain was grown in the presence and absence of exogenous tryptophan. We observed minimal expression in the trpL_WT strain PLBS44 grown in the presence of tryptophan (Table 2). The effect of exogenous tryptophan on expression of the trpL_WT trpE'-'lacZ can be assessed from the ratio of expression in the absence of tryptophan to expression in the presence of tryptophan (Trp/+Trp ratio), which was 140. Comparable experiments were performed with the 5'S-L strain PLBS104. In this case the Trp/+Trp ratio was only 8.4, significantly lower than that observed for the trpL_WT strain. The 17-fold (140/8.4) reduction in the ability of the 5'S-L strain to regulate expression was primarily due to the dramatic increase in expression found to occur when this strain was grown in the presence of tryptophan. When cultures were grown in the presence of tryptophan, -galactosidase activity was approximately 200-fold higher in the deletion strain than in the wild type, while -galactosidase activity was only 13-fold higher in the deletion strain when cultures were grown without tryptophan (Table 2). The dramatic increase in expression, as well as the 17-fold reduction in regulation observed for the 5'S-L strain, demonstrated that the 5' stem-loop influences regulation of expression of the B. subtilis trp operon.

Effect of various 5' stem-loop point mutations on trp operon expression. Since deletion of the entire 5' stem-loop had such a dramatic effect on trp operon expression, we wanted to determine if the primary sequence, the secondary structure, or both were responsible for 5' stem-loop function. We examined the effects of several point mutations that altered the primary sequence and/or the predicted secondary structure of the 5' stem-loop (Fig. 2). Leader regions bearing these 5' stem-loop mutations were integrated as a single copy into the B. subtilis amyE locus as trpE'-'lacZ translational fusions. The effect of the point mutations in the 5' stem-loop of the trp leader on trp operon expression was assessed by comparing -galactosidase expression levels of these mutant strains to those of the wild-type and 5'S-L strains (Table 2). A substantial increase in -galactosidase activity was observed with all of the mutations predicted to disrupt the 5' stem-loop (C3G, U5A, C13G, C15G, G22C, G24C, and G31C). Furthermore, compared to that in the wild-type strain, the fold change in -galactosidase activity when cultures were grown in the absence versus the presence of tryptophan was reduced considerably in these 5' stem-loop mutants (Table 2). Interestingly, each of these single-nucleotide substitutions had an effect similar to that of the 5' stem-loop deletion.

The 5' stem-loop functions in TRAP-dependent regulation of the trp operon. Since it is well established that TRAP is responsible for regulating expression of the B. subtilis trp operon, we expected that the 5' stem-loop would function in the TRAP-dependent regulatory pathway. To test this hypothesis, we constructed two new strains by integrating the wild-type or 5'S-L trpE'-'lacZ fusion into a TRAP-deficient (mtrB) background. As expected, both strains lost the ability to regulate trp operon expression in response to tryptophan (Table 2). Interestingly, we observed a reproducible two- to threefold increase in expression in the 5'S-L mtrB strain (PLBS252) relative to the mtrB control (PLBS251). Thus, while it appears that the 5' stem-loop functions primarily through the TRAP-dependent regulatory mechanism, these results indicate that the 5' hairpin has a small effect on trp operon expression that is independent of TRAP.

Overexpression of mtrB partially suppresses the defect associated with the 5' stem-loop deletion. The results described above indicate that the 5' stem-loop primarily functions in TRAP-dependent regulation of the B. subtilis trp operon. One possible explanation for these results is that the 5' stem-loop participates in TRAP-RNA recognition. To test this possibility, we examined the effect of overexpressing mtrB (TRAP) in the 5' stem-loop deletion strain by measuring -galactosidase activity when cells were grown in the presence and absence of tryptophan. We found that mtrB overexpression largely suppressed the defect associated with the 5' stem-loop deletion (Table 3). Compare the expression levels of PLBS256 [trpL_5'S-L trpE'-'lacZ/pHY300PLK (vector)] and PLBS255 [trpL_5'S-L trpE'-'lacZ/pSI45 (mtrB⁺)] (Table 3) with that of PLBS44 (trpL_WT trpE'-'lacZ) (Table 2). These results suggest that the 5' stem-loop assists TRAP in binding to the (G/U)AG repeats present in the trp leader transcript by increasing the affinity of TRAP for trp leader RNA.

The 5' stem-loop structure increases the affinity of TRAP for trp leader RNA. The results described above suggested that the 5' stem-loop increases the affinity of TRAP for trp leader RNA. We performed TRAP-trp leader RNA gel mobility shift experiments to determine if the 5' stem-loop does in fact increase the affinity of TRAP for the trp leader transcript. We used in vitro-generated transcripts that contained nucleotides +1 to +111 of the trp leader (wild type) or a similar transcript containing nucleotides +32 to +111 in which the 5' stem-loop had been deleted (5'S-L). Note that both of these transcripts contained the 11 (G/U)AG repeats that are known to interact with TRAP (Fig. 1). Binding to the wild-type trp leader transcript was detectable at 2.5 nM TRAP and saturated at approximately 80 nM TRAP (Fig. 3). With the 5'S-L transcript, binding was detected at 10 nM TRAP and saturated at approximately 320 nM TRAP (Fig. 3). Note that the relative intensity of the TRAP-dependent band that is visible just above the free RNA band fluctuated from experiment to experiment. Since this species was found to increase with increasing TRAP concentrations and never saturated, and the shift was too small for it to represent a distinct TRAP-RNA complex, it is likely that this band resulted from complex dissociation soon after gel loading. Nonlinear least-squares analysis of these data yielded estimated K_d values of 17 ± 5.3 nM TRAP for the wild-type transcript and 79 ± 9.2 nM TRAP for the 5'S-L transcript. Our in vitro binding data indicate that deletion of the 5' stem-loop results in a four- to fivefold decrease in the affinity of TRAP for trp leader RNA. This result is consistent with the previous finding that overexpression of TRAP partially suppressed the defect associated with a 5' stem-loop deletion in vivo (Table 3).

View larger version (45K): [in this window] [in a new window]   FIG. 3.   Gel mobility shift analysis of TRAP complexed with wild-type or 5' stem-loop deletion trp leader transcripts. 5'-32P-labeled wild-type (WT) or 5' stem-loop deletion (5'S-L) trp leader transcripts (0.2 nM) were incubated with 1 mM tryptophan and TRAP at the concentrations indicated above the gel. Each transcript contained the 11 (G/U)AG repeats between +36 and +91. Control reactions without tryptophan were performed with each RNA at the highest TRAP concentration used. Bands corresponding to free RNA (Free) and shifted TRAP-RNA complexes (Bound) are indicated at the left.

The 5' stem-loop participates in the trp operon transcription attenuation mechanism. Since TRAP binding to trp leader RNA is responsible for controlling trp operon expression by transcription attenuation and translational control mechanisms, we were interested in determining if the 5' stem-loop participates in the transcription attenuation mechanism. To test this possibility, we performed in vitro transcription attenuation assays using wild-type (pPB22) and 5' stem-loop deletion (pJY2) templates in a reaction mixture containing B. subtilis vegetative (^A) RNA polymerase, TRAP, tryptophan, and the four ribonucleoside triphosphates. The major transcript produced from the wild-type template in the absence of TRAP was the 320-nt readthrough transcript (Fig. 4). Essentially identical results were observed when TRAP was present but tryptophan was omitted from the reaction mixture (data not shown). In the presence of increasing concentrations of TRAP (0.1, 0.3, and 1 µM), we observed a decrease in the readthrough transcript and an increase in the 139-nt terminated transcript (Fig. 4). Thus, as demonstrated previously, transcription termination in the wild-type trp leader in vitro is dependent on the presence of TRAP and tryptophan (5, 27). When the 5' stem-loop deletion template was used in the reaction in the absence of TRAP, the major transcript was the 290-nt readthrough transcript (Fig. 4). In the presence of increasing concentrations of TRAP (0.1, 0.3, and 1 µM), we observed a decrease in the readthrough transcript and an increase in the 109-nt terminated transcript (Fig. 4). At each concentration of TRAP tested, we observed a higher percentage of readthrough transcripts with the 5' stem-loop deletion template than with the wild-type template. The effect of the deletion was most pronounced at the highest concentration of TRAP used. In this case, the percentage of readthrough transcripts was approximately threefold higher than that observed with the wild-type template (Fig. 4). While the observed difference in readthrough efficiency is modest compared to the effect observed in vivo, these results demonstrate that deletion of the 5' stem-loop leads to an increase in transcriptional readthrough, consistent with the expression and gel shift results described above. Moreover, these results demonstrate that the 5' stem-loop influences the efficiency of the transcription attenuation mechanism.

View larger version (44K): [in this window] [in a new window]   FIG. 4.   Deletion of the 5' stem-loop increases transcription readthrough in vitro. The concentration of TRAP used in each reaction is indicated. Arrows on the left indicate the positions of the 320-nt readthrough transcript (RT) and the 139-nt terminated transcript (T) generated from the wild-type template. Arrows on the right indicate the positions of the 290-nt readthrough and the 109-nt terminated transcripts generated from the 5' stem-loop deletion template. The molar percentage of readthrough transcripts is shown below each lane.

5'S-L

View larger version (55K): [in this window] [in a new window]   FIG. 5.   Oligonucleotide competition in vitro transcription assay. The DNA oligonucleotide was complementary to nt 2 to 16 of the trp leader transcript. The concentrations of TRAP and the oligonucleotide added to each reaction mixture are given above the gel. Arrows indicate the positions of the 320-nt readthrough (RT) and the 139-nt terminated (T) transcripts. The molar percentage of readthrough transcripts is shown below each lane.

	DISCUSSION

Top Abstract Introduction Materials and Methods Results Discussion References

The B. subtilis trp operon is regulated by transcription attenuation in response to changes in the intracellular level of tryptophan by the mtrB gene product, TRAP. TRAP interaction with the 11 (G/U)AG repeats present within the nascent trp leader transcript promotes formation of the intrinsic terminator by blocking formation of the overlapping antiterminator structure (Fig. 1). An essentially identical transcription attenuation mechanism was shown to regulate expression of the B. pumilus trp operon (20). In addition, the trp operon leaders of B. caldotenax and B. stearothermophilus contain multiple triplet repeats, as well as overlapping antiterminator and terminator structures. Thus, it appears that all four of these bacilli regulate trp operon expression by a conserved attenuation mechanism.

In this study, the function of an RNA secondary structure predicted to form at the 5' end of the B. subtilis trp leader transcript was examined. The conservation of similar structures in the trp operon leaders of B. pumilus, B. caldotenax, and B. stearothermophilus suggested that it may have a regulatory role in trp operon expression (Fig. 2). We found that deletion of the B. subtilis 5' stem-loop resulted in a dramatic increase in expression of a trpE'-'lacZ translational fusion (PLBS104) compared to a similar fusion containing a wild-type trp leader (PLBS44). The effect of the deletion was most pronounced when cells were grown in the presence of tryptophan in the growth medium (Table 2). The Trp/+Trp -galactosidase ratios of these two strains demonstrate that the absence of the 5' stem-loop results in a partial loss of trp operon regulation in response to tryptophan.

To determine the features of the 5' stem-loop involved in trp operon regulation, we examined the effects of single point mutations that altered its sequence, the predicted secondary structure, or both. All of the point mutations that were predicted to disrupt the secondary structure (C3G, U5A, C13G, C15G, G22C, G24C, and G31C) had effects on trp operon expression that were similar to deletion of the entire hairpin. In fact, C15G and G22C were more deleterious than the deletion itself (Table 2). The finding that the 5' stem-loop increases the affinity of TRAP for trp leader RNA four- to fivefold (Fig. 3) suggests that TRAP interacts with the 5' stem-loop. Thus, disruption of the 5' structure by various point mutations probably prevents this interaction, resulting in higher expression levels. However, it should be pointed out that in addition to disruption of the 5' structure, the optimal and/or suboptimal structures predicted to form in several of these mutant transcripts (C3G, C13G, C15G, G22C, and G31C) sequester several of the normally single-stranded (G/U)AG repeats previously shown to function in TRAP binding (data not shown). Since it is known that RNA secondary structures that sequester (G/U)AG repeats inhibit TRAP binding (8), it is possible that these optimal and/or suboptimal structures contribute to the regulatory defects associated with these mutations. Despite this, the finding that the C3G G31C and C15G G22C compensatory changes resulted in the partial restoration of wild-type-like expression levels, especially when cultures were grown in the presence of tryptophan, clearly indicates that the structure of the hairpin is important for proper regulation of the trp operon (Table 2). Interestingly, while the expression levels of the compensatory mutants are approximately 10-fold higher than those of the wild-type strain in both the presence and the absence of tryptophan in the growth medium, the fold regulation (Trp/+Trp) was similar to that of the wild type in both cases. Results from in vitro transcription attenuation experiments are consistent with the in vivo trp leader mutation studies. Deletion of the 5' stem-loop or disruption of hairpin formation with a complementary oligonucleotide resulted in an increase of approximately threefold in transcriptional readthrough (Fig. 4 and 5).

The two point mutations that were predicted to alter only the primary sequence of the 5' stem-loop (G7A and A19U) had very different effects on trp operon expression. The free energies of the structures containing the G7A or A19U mutations are predicted to be the same as that of the wild-type hairpin. Both of these mutations alter potential UAG (G7A) or GAG (A19U) TRAP recognition sites. The A19U mutation had a minimal effect on trp operon regulation, suggesting that a KKR motif interaction with this GAG sequence is not involved in TRAP-trp leader RNA interaction. In contrast, the G7A mutation led to a substantial loss of regulation, comparable to that with the 5' stem-loop deletion. Thus, it is possible that TRAP recognizes this UAG sequence with a KKR motif. An alternative explanation is that this residue contributes to TRAP-RNA recognition by a non-KKR motif interaction. For example, G7 is part of a GAAA sequence that could form a tetraloop, a structure consisting of non-Watson-Crick base pairs known to stabilize RNA secondary structures (39). Thus, it is possible that G7 contributes to the structure of the 5' stem-loop.

Since TRAP is the only trans-acting factor known to regulate trp operon expression, we expected that the 5' stem-loop would function in TRAP-dependent regulation of this operon. We found that the effect of the 5' stem-loop deletion on trp operon expression was significantly reduced in a TRAP-deficient (mtrB) background (Table 2). Moreover, the observation that overexpression of mtrB partially suppressed the effect of the 5' stem-loop deletion (Table 3), combined with the finding that deletion of the 5' structure reduced the affinity of TRAP for trp leader RNA four- to fivefold (Fig. 3), suggests that the 5' structure increases the affinity of TRAP for trp leader RNA by a direct TRAP-5' stem-loop interaction. The results of the in vitro transcription studies are consistent with this interpretation (Fig. 4 and 5). Thus, the 5' stem-loop may tether TRAP to the nascent trp leader transcript in a manner not yet identified such that TRAP would be in position to bind to the (G/U)AG repeats as soon as they are transcribed. This multipartite binding mechanism may be important to increase the likelihood that tryptophan-activated TRAP would bind to the trp leader in time to block antiterminator formation.

In addition to TRAP-trp leader interaction, the 5' stem-loop may play a role in transcript stability. This could provide an explanation for the consistent two- to threefold increase in expression levels that was observed for strain PLBS252 (5'S-L mtrB) compared to PLBS251 (mtrB). The 5' stem-loop may be a target for an endonuclease leading to the decay of the trp operon transcript. Various studies regarding mRNA decay in B. subtilis have pointed to the importance of the 5' segment of mRNA in controlling the stability of the transcript under different growth conditions (10, 13, 26, 31). Some messages are thought to be degraded by a ribonucleolytic activity that begins at the 5' end and degrades the message in a 5'-to-3' direction (10), while processing of the thrS leader transcript plays a major role in the induction of thrS expression following threonine starvation in B. subtilis (13). Another study concludes that initiation of mRNA decay in B. subtilis generally occurs at or near the 5' terminus (14). Studies have also revealed a sequence that specifies a 5' stabilizer function which appears to be localized to a polypurine sequence that resembles a ribosome binding site in B. subtilis (14) and indicate that attack at the 5' end is a principal mechanism for initiation of mRNA decay in B. subtilis (21). In our study of the trp leader 5' stem-loop, loss of a potential endonucleolytic target would result in a two- to threefold increase in the stability of the trp operon transcript. It remains to be determined if the 5' stem-loop does in fact serve as an mRNA instability determinant.

Deletion of the 5' stem-loop affected regulation of the trp operon approximately 200-fold. Since only two- to threefold of this is TRAP independent, the remaining 67- to 100-fold effect of the deletion is due to a defect in TRAP-dependent regulation. While it is possible that the two- to threefold TRAP-independent effect is due to increased mRNA stability, the increase of approximately threefold in transcriptional readthrough observed in vitro cannot be attributed to mRNA stability, since the only proteins present in the transcription attenuation assay were RNA polymerase and TRAP. Taking into consideration the two- to threefold in vivo TRAP-independent and the approximately threefold in vitro TRAP-dependent effect of the 5' stem-loop deletion, we are left with a 20- to 30-fold TRAP-dependent effect that is unaccounted for. Thus, it appears that some feature or factor involved in TRAP-dependent trp operon regulation is missing from the in vitro transcription attenuation system.