HOME HELP FEEDBACK SUBSCRIPTIONS ARCHIVE SEARCH TABLE OF CONTENTS		QUICK SEARCH:   [advanced] Author: Keyword(s): Year:  Vol:  Page:

J. Biol. Chem., Vol. 277, Issue 49, 47420-47427, December 6, 2002

This Article Abstract Full Text (PDF) All Versions of this Article: 277/49/47420    most recent M208363200v1 Alert me when this article is cited Alert me if a correction is posted Citation Map Services Email this article to a friend Similar articles in this journal Similar articles in PubMed Alert me to new issues of the journal Download to citation manager Citing Articles Citing Articles via HighWire Citing Articles via Google Scholar Google Scholar Articles by Lee, S. J. Articles by Gralla, J. D. Search for Related Content PubMed PubMed Citation Articles by Lee, S. J. Articles by Gralla, J. D. Social Bookmarking                      What's this?

Promoter Use by ³⁸ (rpoS) RNA Polymerase

AMINO ACID CLUSTERS FOR DNA BINDING AND ISOMERIZATION^*

Shun Jin Lee and Jay D. Gralla

From the Department of Chemistry and Biochemistry and the Molecular Biology Institute, University of California, Los Angeles, California 90095-1569

Received for publication, August 15, 2002, and in revised form, September 23, 2002

	ABSTRACT

TOP ABSTRACT INTRODUCTION EXPERIMENTAL PROCEDURES RESULTS DISCUSSION REFERENCES

³⁸ is a non-essential but highly homologous member of the ⁷⁰ family of transcription factors. In vitro mutagenesis and in vivo screening were used to identify 22 critical amino acids in the promoter interaction domain of Escherichia coli ³⁸. Electrophoretic mobility shift assay studies showed that residues involved in duplex DNA binding largely segregated into distinct regions that coincided with those of ⁷⁰. However, the majority of these amino acids were in non-conserved positions. Analysis indicates that this region of the two s probably has a common overall organization but differs in how its amino acids are used to form functional open complexes. Placement of the mutations on the known ⁷⁰ holoenzyme structure shows two clusters; one appears to be used for duplex DNA recognition and the other for the subsequent isomerization events. Permanganate assays for DNA melting support this view.

	INTRODUCTION

TOP ABSTRACT INTRODUCTION EXPERIMENTAL PROCEDURES RESULTS DISCUSSION REFERENCES

The alternative  factor, ³⁸ (rpoS), is the principle regulator of the general stress response in Escherichia coli. The ³⁸ regulon controls 50 to 100 genes (1). Subsets of these genes are induced during starvation for various nutrients and in response to various stresses such as the accumulation of reactive oxygen species, changes in pH, and osmolarity. The highest activity of ³⁸ occurs during stationary phase when these and other stresses are present. Many factors contribute to this activity, especially the increased stability of the ³⁸ protein (2).

³⁸ is highly homologous to ⁷⁰, the vegetative  factor responsible for the transcription of most of the housekeeping genes. The regions of ⁷⁰ that recognize promoter DNA, conserved region 4.2 of the protein (recognizes the 35 element), and conserved regions 2.3-2.5 (10 element) are over 70% similar (60% identical) to those of ³⁸ (3). However, ³⁸ promoters generally contain only a single DNA recognition element, a 10 sequence centered between nucleotides 14 and 7 (4, 5). The four most conserved of these nucleotides, 13C, 12T, 11A, and 7T, are involved in directing promoter selectivity and play a dominant role in setting promoter strength (4-6). Three of these positions, 12T, 11A, and 7T, are also critical for ⁷⁰ function, although they are utilized at different steps during ⁷⁰ transcription initiation (7).

³⁸ and ⁷⁰ do not respond to regulators and the physiological state of the cell in the same manner. It is clear that such regulators as Lrp, CRP, H-NS, and many others can differentially effect transcription by ³⁸ and ⁷⁰ (8, 9). At certain promoters, a low supercoiled state of the DNA, which is present in stationary phase (10), seems to favor ³⁸ transcription (11). Increased concentrations of trehalose (12), as well as potassium glutamate (13), also preferentially stimulate ³⁸-dependent transcription at certain promoters.

The basis for these diverse properties between the two highly homologous holoenzymes have remained largely unknown. Some differences in ³⁸ function have been attributed to differential recognition of nucleotides 14 and 13 (5) and a C-terminal "tail" that helps in sensing potassium glutamate (14). The ³⁸ amino acids responsible for use of nucleotides 12 to 7 have not been characterized. Presumably, they would lie in the same part of ⁷⁰ that recognizes the 10 element, conserved regions 2.3-2.5. Because ³⁸ and ⁷⁰ promoters behave differently their interactions with region 2 may not be identical. There have been several studies identifying ⁷⁰ amino acids that use the 10 element (15, 16), but information concerning ³⁸ recognition is sparse.

One purpose of this study is to identify functionally important residues in the homologous region 2 of ³⁸. Another purpose is to begin to understand how these amino acids function. The results will be interpreted using the known structure of the DNA interaction region of ⁷⁰, and models for promoter usage by both s will be discussed.

	EXPERIMENTAL PROCEDURES

TOP ABSTRACT INTRODUCTION EXPERIMENTAL PROCEDURES RESULTS DISCUSSION REFERENCES

Site-directed Mutants and DNA-- ³⁸ site-directed mutants were constructed using the Stratagene site-directed mutagenesis kit. All oligonucleotides were synthesized by Operon Technologies and were gel-purified and prepared as described (17). Annealed DNA probes were prepared as described (17).

Strains and Plasmids-- pRpoS was created by inserting the ³⁸ gene (rpoS) into the isopropyl-1-thio--D-galactopyranoside-inducible ptrc vector (Invitrogen). Two unique restriction sites, AvrII and XmaI, were inserted into pRpoS at amino acids 124 and 183, respectively, by use of the Stratagene site-directed mutagenesis kit. E. coli RJ4099 (CAG4000 proP-104::TnphoA'-4 katF13::Tn10) lacks the ³⁸ gene and carries the lacZ gene under the control of a ³⁸ promoter, proP (18). E. coli RJ4095 (CAG4000 aldB-731::TnphoA'-4 katF13::Tn10) lacks the ³⁸ gene but carries the lacZ gene under the control of a ³⁸ promoter, aldB (19).

PCR Mutagenesis-- The primers used for amplification overlapped the AvrII and XmaI site. Error-prone PCR was conducted as follows: 50 µl of reaction mixtures contained 5 ng of pRpoS, 1 µg of each primer, 1× Taq polymerase buffer (Promega), 5 units of Taq polymerase, 25 mM MgCl₂, 0.425 mM MnCl2, and 0.2 mM dNTPs. After 30 cycles, the reaction mixtures were digested with dpnI (New England BioLabs) to remove the parental vector and then purified with the Qiagen PCR purification kit. Fragments were next digested with AvrII and XmaI. Full-length pRpoS was also digested with AvrII and XmaI, and the large fragment was purified by agarose gel electrophoresis.

Plating Test-- The mutated insert and the large vector fragment were ligated and transformed into RJ4099. The bacteria were plated onto LBN plates (1% tryptone, 0.5% yeast extract) containing 100 µg/ml ampicillin, 30 µg/ml kanamycin, and 0.5% glucose and grown at 30 °C for 16-18 h. LBN plates were used for RJ4099, because the proP promoter contains both a ⁷⁰- and a ³⁸-dependent promoter, but only the ³⁸ promoter (proP2) is active under these low salt conditions (18).

Transformants were replica-plated onto nitrocellulose and reincubated, colony side up, onto LBN plates containing 100 µg/ml ampicillin, 30 µg/ml kanamycin, 4 mM isopropyl-1-thio--D-galactopyranoside, and 80 µg/ml 5-bromo-4-chloro-3-indolyl--D-galactopyranoside (X-gal). The plates were incubated at 30 °C until blue colonies became distinguishable (from 3 to 6 h). Occasionally, the plates were incubated an additional 4 to 6 h at 4 °C to enhance blue/white color distinctions. White colonies were chosen and sequenced. The point mutants that were identified through the screen were transformed into a second E. coli strain, RJ4095. They were treated as described above with the exception that all plates were LB.

Statistical Analysis-- 383 white colonies were sequenced, and of these, 127 gave sequences that did not contain frameshifts, insertions, deletions, or stop codons. A statistical analysis was performed by comparing the frequency of mutagenesis at a position to the expected number of changes at that position by standard deviation analysis. The analysis was done as follows. The total number of changes in the nonfunctional mutant library was calculated by adding all of the mutations from the 127 colonies (417 changes). The expected number of changes/position was determined by dividing the total number of changes (417) by the number of amino acids in the mutated stretch (50) yielding a number of 8.3. The expected number of changes/position, 8.3, was compared with the actual number of changes in the library at a particular residue (e.g. Phe-140 was changed 14 times in the library). The difference between the actual and expected frequency at each residue (e.g. Phe-140, 14  8.3 = 5.7) was then used to calculate a standard deviation of 4.3. Any position with a number of changes over a standard deviation above the mean was taken to be a preferential target in the non-functional library. For these positions a pair-wise analysis was conducted and showed that none were correlated strongly with mutations at other positions. Accounting for codon usage by multivariate analysis did not alter the results significantly.

RNA Analysis-- The point mutants were grown overnight in 3 ml of LB plus 1% glucose at 37 °C. The cells were diluted 1:500 into 20 ml of LB (without glucose), grown until an optical density of 1.0, and induced with 1 mM isopropyl-1-thio--D-galactopyranoside for 2 h. Half of the culture was used for RNA analysis, and the remaining 10 ml were used for EMSA¹ with crude extract (below). Cells were harvested, and RNA was extracted with the Qiagen RNeasy kit. Mixtures for primer extension analysis were as follows: 15-µl reactions contained 5 µg of RNA, 10 nM labeled amp primer, 10 nM either labeled dps or aldB primer, 1× reverse transcriptase buffer (Promega), 5 units of reverse transcriptase, and 0.2 mM dNTPs. Urea stop dye was added, and samples were run at 32 watts on a 6% sequencing gel. The radioactive bands were visualized and quantified by phosphorimager analysis. RNA samples for each mutant were prepared at least three separate times. Experiments were conducted five times, and the average percentage was taken.

EMSA with Crude Extract-- Cells were prepared as described for primer extension. Once harvested, the cells were resuspended in 100 µl of TGED-NaCl without glycerol. The cells were lysed through sonication, and protein levels were checked by SDS-PAGE. The samples were then spun at maximum speed in a microcentrifuge at 4 °C for 20 min. The supernatants were taken, supplemented with 15% (w/v) glycerol, and were frozen immediately. One time use aliquots were used in band shift analysis. Samples were diluted with TGED-NaCl (15% glycerol) from 2- to 5-fold to standardize protein levels. T153K gave consistently low amounts of protein.

Mobility shift assays were conducted as follows: 20 nM core RNA polymerase (from Epicenter Technologies) and 0.5 µl of crude extract were added to a 9-µl reaction mixture with 1× Buffer B (50 mM Tris-HCL, pH 7.9, 200 mM potassium glutamate, 3 mM MgCl₂, 0.1 mM EDTA, 1 mM dithiothreitol, 100 µg/ml bovine serum albumin, 6 ng/µl dI-dC). This mixture was incubated at room temperature for 5 min to allow holoenzyme formation. The reaction was then preincubated for 10 min on ice before addition of 1 nM of annealed DNA probe. Reactions were further incubated for 20 min on ice. All samples were run on 7% PAGE with cold 1× Tris-buffered EDTA, packed in ice at 300 V. After electrophoresis, the radioactive bands were visualized and quantified by phosphorimager analysis. The crude extract for each of the mutants was prepared at least three separate times. Each band shift was repeated at least four times, and the average percent binding was taken.

For competition binding experiments, wild-type ³⁸ and a mutated form of ³⁸ were added at the same time, allowed to bind core polymerase for 5 min at room temperature, and were followed by the addition of labeled DNA. Each reaction contained a total of 2 µl of crude extract. The remaining conditions were the same as for the EMSA experiments described above.

KMnO4 Footprinting-- Proteins were purified as described (15). Plasmid pFic was constructed by inserting a 343-bp fragment containing the fic promoter into pTH8 through the unique restriction sites of BamHI and HindIII. pFicCon was created by changing pFic by the Stratagene site-directed mutagenesis kit. KMnO₄ footprinting was conducted as follows: 200 nM  and 40 nM core polymerase were incubated for 10 min at room temperature in 1× Buffer B (described above). The reactions were then preincubated for 10 min at 30 °C before addition of 2.5 nM pFicCon linearized at the BamHI site. Reactions were further incubated for 20 min at 30 °C. A final concentration of 2 mM KMnO₄ was then added to each sample and incubated for 15 s and quenched with -mercaptoethanol. Samples were processed for primer extension analysis as described (20). Each experiment was repeated three times, and results were similar for the template and nontemplate strands.

	RESULTS

TOP ABSTRACT INTRODUCTION EXPERIMENTAL PROCEDURES RESULTS DISCUSSION REFERENCES

Regions 2.3, 2.4, and 2.5 of ⁷⁰ have been identified as the critical components that recognize and melt the 10 element (21, 22). To understand their role in ³⁸ (rpoS), this DNA segment was mutated by error-prone PCR and then ligated into a receiver rpoS vector. The mutant plasmid collection was transformed into an rpoS minus strain that contained a lacZ fusion to the rpoS-dependent proP2 promoter (18). Blue colonies in low salt media indicate ³⁸ function, and white colonies are expected to contain nonfunctional ³⁸. These were subject to DNA sequence analysis.

127 white colonies contained changes consisting solely of point substitutions (no frameshifts, insertions, deletions, or stop codons). Each mutant had from one to eight changes, with 21 colonies containing a single substitution and one colony yielding a plasmid with eight changes (Table I). Every position in the 50-amino acid stretch was changed at least once. The screen may have approached saturation as several of the mutants appeared more than once.

View this table: [in this window] [in a new window]   Table I Statistical analysis of a nonfunctional 38 (rpoS) library Total number of colonies, 127; total number of changes, 417; number of amino acid positions, 50; expected number of changes/positions, 8.3 ± 4.2.

Two approaches were taken to analyze the collection of nonfunctional mutants. In the first, a statistical analysis identified residues that were mutated most frequently in the library. This includes many plasmids that had lost ³⁸ function because of multiple mutations. Second, we simply identified single point substitutions that appeared in the library; only these were tested further using in vitro assays.

Positions Commonly Mutated in Multiple Mutants in Region 2-- A statistical analysis of this non-functional library of substitutions was conducted (see "Experimental Procedures" and Table I) to identify amino acids that were mutated most frequently. Eight positions were found to be changed far more frequently than others (one standard deviation above the mean), suggesting they were selectively important for function (Fig. 1). We note that all of these eight positions have been shown to be important in ⁷⁰ function (Table II, top). For these positions a pair-wise analysis was conducted and showed that none were correlated strongly with mutations at other positions (data not shown). Accounting for codon usage ("hot spots") by multivariate analysis did not alter the results significantly (data not shown).

View larger version (15K): [in this window] [in a new window]   Fig. 1.   38 regions 2.3 to 2.5. A linear model based on the T. thermophilus 70 structure aligned with the 38 sequence is represented on the top. Conserved regions of the 70 family (middle) and the 38 amino acid numbers are also shown. S38 is the 38 sequence, and S70 is the analogous sequence in 70. Positions that were overrepresented, point mutations, site-directed mutants, or positions identified as being critical for DNA binding are indicated with an X.

View this table: [in this window] [in a new window]   Table II 38 positions non-randomly represented in the nonfunctional library The residue in 38 (column one) and the analogous residue S.D. in 70 (column two) are shown. The total number of times a particular residue was mutated in the library (column 3) and any previously identified phenotype associated with that residue is also shown (column 4). Overrepresented positions (top) are mutated more than one above the mean. Underrepresented positions (bottom) are mutated more than S.D. below the mean.

Eight different positions were identified that appeared far less frequently than expected (one standard deviation below the mean; see Table II, bottom). Seven of these have been characterized previously with six having only minor to moderate defects, in either ³⁸ or ⁷⁰. This contrast supports the legitimacy of the screen and suggests that many of the same positions are important for both s.

The most significant exception is Tyr-145 (Tyr-430 in ⁷⁰), which was identified previously as being critical for DNA melting and enzyme isomerization in ⁷⁰ (15, 16). It is interesting to note that the Tyr-430 defects were identified in vitro and that mutations in this residue have not appeared in ⁷⁰ nonfunctional screens (23). Previous data also suggested that the defects were not as severe at high temperatures or on supercoiled DNA, both conditions present during the screen conducted here (15, 24).

Single Point Substitutions-- The importance of certain individual residues is derived more directly from white colonies containing plasmids with single substitutions. 18 such mutants (Fig. 1) were identified as non-functional (3 of the 21 colonies of this type contain redundant sequences). These mutations are spread fairly uniformly throughout the region. A minority of these correspond to positions of significant importance in ⁷⁰. The collection also includes three of the eight residues identified as overrepresented in the library (Phe-140, Trp-149, Gln-152). The lack of appearance of the other five residues could be because the screen did not reach saturation, or because they need to be coupled with other mutations to cause loss of function.

These mutations depend on the loss of function of the proP2 promoter. This promoter is activator-dependent and reaches its highest expression levels during the end of exponential phase (18). The 18 point mutants were transferred to a different rpoS strain containing an aldB promoter-lacZ fusion. This promoter does not require an activator for strong expression and is active during stationary phase (19). 16 of the 18 mutants remained white on this screen indicating that these mutated residues were important for both promoters. The exceptions were R163H and R166H. We do not know the reason for the activity of these mutants at the aldB promoter. They also gave wild-type activity in subsequent primer extension and band shift experiments and are not discussed further. The remaining 16 point mutants identify amino acids important for ³⁸ function at both an activator-dependent and an activator-independent promoter.

RNA Levels-- These mutants were analyzed further by isolating RNA from transformed cells and conducting primer extension using promoter-specific probes. None of the 16 point mutants gave a detectable signal for the aldB promoter (data not shown), consistent with the lack of function in the genetic screen. The same preparations were analyzed using a probe specific for dps RNA. Upon entry into stationary phase, dps levels increase dramatically and are controlled at the level of transcription by ³⁸ (25, 26). dps is a strong ³⁸ promoter, and mutated versions of ³⁸ might be expected to show some function.

Fig. 2 shows the results of comparing dps expression with that of the ⁷⁰ ampicillin RNA encoded by the plasmid. RNA levels were judged by comparing the strength of these dps and amp signals (Fig. 2B). E132G, K133R, and M159T the only three mutants in which the dps signal was stronger than the amp signal (Fig. 2A), had nearly wild-type activity. The defects associated with these mutant forms of ³⁸ can therefore be masked at a strong promoter.

View larger version (65K): [in this window] [in a new window]   Fig. 2.   A, primer extension of RNA from the 38-dependent dps promoter. Vector (cells lacking 38), wild-type 38, and point mutants identified from the screen are labeled on the top. AMP (arrow) represents the control signal from RNA expressed by the ampicillin gene. DPS (brackets) represents the signal from the 38-dependent dps promoter. B, primer extension signals were quantified and standardized by taking the ratio of dps/ampicillin. Wild-type (W.t.) was set to the arbitrary value of 1.

The remaining mutants showed a different pattern, and analysis indicated that their RNA levels were at least ~4-fold down. The majority of the point mutants, therefore, remained significantly defective even at a strong promoter. This suggests that these 13 positions play a critical role in ³⁸-dependent transcription in vivo.

Duplex DNA Binding-- Duplex DNA binding by these mutants was assessed using protein extracts and closed complex conditions. One complication is that ³⁸ holoenzyme binding to double strand DNA oligonucleotides in vitro is very weak (4). However, the presence of a 35 element can increase duplex binding significantly even though ³⁸ promoters do not typically contain 35 elements (6). The addition of a 35 element to the synthetic fic con promoter strongly increased binding by ³⁸ holoenzyme in vitro (4). Using this template, crude protein extracts were used in EMSA experiments to test for duplex recognition (Fig. 3). The proteins were overexpressed, and levels were checked by SDS-PAGE (data not shown). All proteins migrated to the same position, and the amounts of protein were adjusted where necessary (see "Experimental Procedures").

View larger version (64K): [in this window] [in a new window]   Fig. 3.   Closed complex binding to a duplex probe by holoenzymes containing mutant forms of 38 as detected by EMSA. In typical experiments, 13% of the probe was bound by wild-type 38; 18% binding by K133R; 4-11% by R129C, E132G, K133M, F140L, S143T, Q152R, and Y145A; 1-2% by W149R, M159T, I169T, and W148A; and undetectable levels by R141S, T153K, A157T, H170L, I171N, and L175P. The unbound probe is not shown.

This assay showed high specificity for ³⁸ RNA polymerase (RNAP) binding as a strong band appeared using a vector overexpressing wt ³⁸ (top panel, lane 2) but did not appear from a vector without ³⁸ (lane 1). This band runs at the same position as that seen by purified ³⁸ RNAP (data not shown). Binding by endogenous s, holoenzymes, or other DNA-binding proteins was not detected as no band appeared with the cells lacking ³⁸ (lane 1). Occasionally, a band corresponding to nonspecific binding by core polymerase would appear (Fig. 3, left bottom panel, band below arrow), but this would be present in all lanes.

The EMSA experiments were conducted at 4 °C to minimize the open conformation of the DNA and to limit nuclease and protease activity. None of the mutants showed an increase in nuclease or protease activity as compared with wild-type (data not shown). Thus the experiment primarily assays for closed complex formation. The use of crude cell extracts to test for DNA binding has also been used in the ⁵⁴ system (27).

The 16 point mutants were overexpressed in E. coli, and crude cell extracts were prepared, mixed with purified core polymerase, and used in EMSA with a duplex fic con 35 probe (Fig. 3). Six mutants (R141S, T153K, A157T, H170L, I171N, L175P) did not give detectable levels of binding, likely accounting for their lack of function. Three mutants (W149R, M159T, I169T) were down ~7-fold or more (Fig. 3), a fairly severe defect. The remaining seven mutants bound within 3-fold of wild-type. The only mutant with a wild-type level of binding was the conservative change K133R.

We note that of the nine mutants that were down at least 7-fold, eight were in the 30-amino acid stretch C-terminal to Trp-149. By contrast, of the seven mutants with milder DNA binding defects, six were in the 20-amino acid stretch N-terminal to this residue. Thus it appears that the primary determinants for forming closed complexes with ³⁸ are in regions 2.4 and 2.5 with region 2.3 playing a different role.

Properties of Two Site-directed Mutants-- Two positions important for ⁷⁰ function, the conserved aromatics Y145A and W148A (Y430A and W433A in ⁷⁰) were not identified as non-functional in this screen. These play a major role in ⁷⁰ in vitro (15, 16), although their role in vivo is not known. Tyr-145 was also underrepresented in the library (Table II). Each of these was changed to alanine in ³⁸ (Fig. 1) and then characterized.

Y145A and W148A were transferred to rpoS strains carrying either a proP2 or aldB promoter-lacZ fusion to test for ³⁸ function in vivo. They were then screened using the blue/white colony test. Both site-directed mutants gave a light blue phenotype, indicating that they had partial function in vivo. Light blue colonies were not included in the screen of nonfunctional white colonies and thus Y145A and W148A would not have been identified. EMSA with duplex DNA (Fig. 3, right bottom panel), as described above, showed that both mutants led to a moderate defect in binding in vitro, with Trp-148 showing the greater defect (4-8-fold).

Core Polymerase Binding-- The DNA binding experiments measure the interaction between holoenzyme and DNA. However, ³⁸ requires core polymerase for binding, and so mutants that fail to bind core polymerase will also fail to bind DNA. We assayed the non-DNA binders for core binding using a  competition protocol.

In this assay radioactive DNA is mixed with wild-type ³⁸ in the presence of a limiting amount of core polymerase. An excess of various mutant forms of ³⁸ are also present. If a mutant  binds to core then it will titrate away the limiting core, leaving little of it associated with wild-type . Because neither mutant holoenzyme nor wild-type  without core binds DNA (4), the radioactive DNA band should be diminished. To enhance sensitivity we used a very tight binding DNA fork junction probe; the weak binding mutants W149R, M159T, and I169T bind this probe, but the others do not.² The protein competition was applied to the other mutants, and some examples are shown in Fig. 4.

View larger version (61K): [in this window] [in a new window]   Fig. 4.   Competition binding experiments for various mutants on a single stranded fork probe that contains the start site as detected by EMSA. Lane 1 (W.t.) bound at 46% and lane 2 (4× W.t.) at 61%. Lane 3 and lane 5 bound at 20-32%; lanes 8, 10, and 12 bound at 48-59%; lanes 4 and 6 bound at 5-8%; lane 7 bound at 38%; lanes 9 and 11 bound at 15-21%; and lane 13 bound at 64%.

The four innermost pairs of lanes show that a 4-fold excess of different mutant proteins diminishes the signal, as expected from a protein that binds core but not DNA. An excess of two proteins that do bind DNA in holoenzyme form, wild-type and K133R, do not lead to a diminished signal (Fig. 4, outer pairs of lanes). Two of the proteins, T153K and L175P, gave signals that were less diminished, indicating that they bind polymerase but not as well as wild-type. We infer that all the non-DNA binders can bind core RNA polymerase, with two having a partial defect.

Permanganate Assay for DNA Opening-- A significant number of non-functional mutants showed relatively normal levels of binding DNA and core polymerase. These are expected be defective in steps subsequent to forming a closed promoter complex between holoenzyme and DNA. As open complex formation cannot be assayed with the short DNA probes used above, we turned to a plasmid-based system to explore potential defects in DNA melting by these mutants. Permanganate was used to assess whether the mutant holoenzymes were capable of forming open promoter complexes (15, 20).

Fig. 5 shows that all of these mutant holoenzymes exhibit defects in opening linearized plasmid promoter DNA. In all cases the melting signal is significantly less than that of wild-type, with most being only slightly higher than the background signal associated with DNA alone. These defects in opening appear to be significantly greater than the minor decreases seen in duplex binding by these same mutants (Fig. 3). We infer that the residues within this region play a significant role in melting the DNA after polymerase forms a closed complex.

View larger version (75K): [in this window] [in a new window]   Fig. 5.   KMnO4 probing of Escherichia coli mutant 38 holoenzymes on the fic con promoter template strand. Controls with core alone and wild-type 38 holoenzyme are on the far left side. Reactions were performed at 30 °C on linearized plasmid DNA.

	DISCUSSION

TOP ABSTRACT INTRODUCTION EXPERIMENTAL PROCEDURES RESULTS DISCUSSION REFERENCES

Very little is known about how ³⁸ accomplishes promoter recognition. A prior study implicated region 2.5 of ³⁸ as being important for recognition of sequences upstream from the 7 to 12 DNA sequence element (5). Other inferences about recognition use analogies with ⁷⁰, which is 60% identical to ³⁸ in its potential DNA recognition regions 2.3, 2.4, and 2.5 (see Fig. 2). The role of the conserved and non-conserved amino acids is unknown as mutations in regions 2.3 and 2.4 of ³⁸ have not been identified. In this work we identify 22 positions (20 screened and 2 site-directed mutants) in region 2 that affect ³⁸ function. Below we use these data to locate the important regional determinants of ³⁸ function and use the large base of available structural and mutational data on ⁷⁰ to draw additional conclusions about the two proteins.

Binding of ³⁸ to Duplex DNA

Important positions were identified as sites of single substitutions or from a statistical analysis of clones containing multiple mutations. The single substitutions were spread fairly evenly throughout most of regions 2.3, 2.4, and 2.5.

The single substitutions were assayed for DNA binding using duplex DNA under closed complex conditions. When the results were interpreted in terms of location an interesting pattern emerged. Nearly all (eight of nine) of the most defective mutants were C-terminal to position Trp-149. Those with milder defects (six of seven) were N-terminal to Trp-149. From this distribution we infer that the determinants of closed complex formation are primarily in regions 2.4 and 2.5. Region 2.3 is also clearly important, as judged from the mutational analysis, but its main function does not appear to be recognition of duplex DNA to form closed complexes. Some of the positions in these and other regions may not be involved directly in ³⁸ function but could instead alter the local structure of the protein. Gross misfolding is unlikely as all mutants appear to bind core to form holoenzyme.

Comparison with ⁷⁰ and Implications

Organization and Residues-- ³⁸ and ⁷⁰ recognize very similar but not identical DNA sequences near the downstream 10 promoter element (4-6). The proteins are highly homologous in the regions just discussed. Therefore comparison of what is known about the two s should reveal new information about both.

The overall functional organization appears to be similar in the two s. Several studies of ⁷⁰ place Trp-148, Trp-149 and Gln-152 at or near the 12/11 fork junction boundary between single and double strand DNA (15, 28). The C-terminal segment is thought to interact with double strand DNA, and the N-terminal segment is thought to interact largely with non-duplex DNA. This is essentially the same arrangement inferred above for ³⁸.

Despite this similarity of arrangement, the residues important for DNA recognition by ³⁸ appear to be significantly different from those of ⁷⁰. Even though the proteins are 60% identical in these regions only two of the nine positions most important for DNA binding contain identical residues. The seven non-identical pairs of ³⁸/⁷⁰ amino acids were Arg-141/Lys-426, Thr-153/Ala-438, Ala-157/Ser-442, Met-159/Ala-444, Ile-169/Val-454, Ile-171/Met-456, and Leu-175/Ile-460. This indicates that ³⁸ includes an extensive set of determinants for DNA binding that differs from those used by ⁷⁰.

Other data indicate that there are also functional determinants in these regions used by both proteins. We identified eight positions in ³⁸ that were important when mutated in conjunction with other residues. All of these positions have been changed previously in ⁷⁰ and were found to have defects of various types (Table II). A previous study in ⁷⁰ identified five residues important for duplex recognition (Tyr-425, Trp-434, Arg-436, Arg-441, Arg-451) (15). The first four of these correspond to positions overrepresented in the ³⁸ nonfunctional library, and the aromatics were also sites of ³⁸ point mutations (Fig. 1).

In a number of positions the function appears to have changed between the two s. In the adjacent non-identical ³⁸/⁷⁰ positions, Phe-140/Tyr-425 and Arg-141/Lys-426, it was the former position (Tyr-425) reported to be important for ⁷⁰ duplex binding (15) and the latter position (Arg-141) reported to be important for ³⁸ duplex binding. Two other ³⁸ residues, Tyr-145 and Trp-148 (Tyr-430 and Trp-433 in ⁷⁰), were not identified as important in the screen conducted here, although in ⁷⁰ the positions play critical roles subsequent to duplex recognition (15, 16). Study of two site-directed mutants in these positions showed that both are partially functional in vivo; Tyr-145 was important for DNA melting, and Trp-148A has a moderate defects in both melting and duplex binding.

Placing Mutants on the Structure-- Recently a structure of Thermus thermophilus ⁷⁰ holoenzyme has been determined, making it possible to interpret the mutagenesis data in terms of structure (29). One caveat is that the structure is without DNA, and one expects that this will induce some changes. Another is that there is no guarantee that the ³⁸ structure will be the same, although the high degree of conservation argues for this.

Fig. 6A shows the placement of the nine amino acids determined to be important for binding duplex DNA. The view shows that duplex binding determinants are primarily in helix 15 and the C-terminal part of helix 14. The only exception among these is Arg-141. Of these nine amino acids, seven are solvent-exposed and two face helix 13 (Thr-153 and Ala-157).

View larger version (16K): [in this window] [in a new window]   Fig. 6.   Crystal structure of Thermus thermophilus 70 holoenzyme regions 2.2 to 2.5 (29) with mutations shown. A, residues involved in duplex recognition. Black side chains are positions found to be important for 38 duplex binding. The T. thermophilus 70 side chains are shown. The red ribbon backbone denotes positions that are not identical between E. coli 70 and 38. The helix numbering is from Malhotra et al. (32). B, addition of residues shown to be important primarily for isomerization. Aqua side chains are positions with strong defects subsequent to duplex binding, including site-directed mutants. Black side chains are positions important for strong duplex binding, as in part A. Gray ribbons are areas that were not part of the mutated region in this study.

In the case of ⁷⁰, duplex recognition is thought to occur using residues from these same two helix sections. For both s, the helix 15 segment includes amino acids thought to interact with DNA just upstream from the 12 to 7 "consensus" element (5, 30). The two most N-terminal of these positions (Ile-169 and His-170) are apparently not very important for duplex recognition by ⁷⁰ (30) and contain different amino acids at these positions. The four point mutations found in helix 15 are quite severe, however, and could disrupt the local structure of the helix rather than make direct contacts with the DNA. It has been shown that the nearby Lys-173 plays a critical role in recognizing duplex DNA (5).

Fig. 6B shows all the positions identified in this study as being important for ³⁸ function (15 screened and 2 site-directed mutants). These functional determinants are far more extensive than the DNA binding determinants. In fact, only the two loops are largely excluded from containing functional residues, suggesting that the loops are less likely to be contact points to DNA or protein motifs. The lack of mutations in parts of helices 13 and 15 (gray backbone) is expected, because these were outside the segment mutated to make the library. Helix 13 is the only of the three helices that is necessary for function (16) but is not a locus of mutations that lower duplex DNA binding. Overall, the data demonstrate that all helical elements within the probed region are important for function, with the primary DNA binding determinants in the C-terminal part of the structure.

The current data indicate that the mutations in helix 13 and the N-terminal part of helix 14 act primarily at steps subsequent to DNA duplex binding (Fig. 6B, aqua side chains). Extensive studies are required to identify the step(s) at which these act. Initial permanganate experiments (Fig. 5) show that this region is required for forming open promoter complexes on a ³⁸ promoter. These positions form a cluster in space that is adjacent to the elements that bind duplex DNA. This cluster includes the aromatic residues known to be important for the melting of the ⁷⁰ promoter DNA from 11 downstream (15, 24). Thus, it seems likely that this N-terminal cluster of amino acids is involved in the isomerization of the DNA and the enzyme, subsequent to duplex binding via the adjacent determinants. This cluster is close to the single stranded DNA and fork junction in the recent 6.5-Å resolution structure of ⁷⁰ holoenzyme bound to fork DNA (31) and is near the melted DNA consensus that controls isomerization by both s (4, 7).

The two types of holoenzymes are distinct in some mechanistic characteristics. ³⁸ holoenzyme seems to bind duplex poorly and to isomerize fairly readily. ⁷⁰ holoenzyme catalyzes these steps at widely varying rates at different promoters. The two holoenzymes also do not respond to small molecule effectors and protein activators in the same manner. Thus, although the arrangement of regions critical for these steps is organized similarly, the differences in the detailed interactions likely contributes to the mechanistic diversity seen for these and perhaps other types of bacterial holoenzymes.

	ACKNOWLEDGEMENT

We thank Dr. Reid C. Johnson for use of strains RJ4095 and RJ4099.

	FOOTNOTES

^* This work was supported by National Institutes of Health Grant GM35754.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.

To whom correspondence should be addressed: Dept. of Chemistry and Biochemistry and the Molecular Biology Inst., University of California, P. O. Box 951569, Los Angeles, CA 90095-1569. Tel.: 310-825-1620; Fax: 310-267-2302; E-mail: gralla@chem.ucla.edu.

Published, JBC Papers in Press, September 25, 2002, DOI 10.1074/jbc.M208363200

² S. J. Lee and J. D. Gralla, unpublished data.

	ABBREVIATIONS

The abbreviation used is: EMSA, electrophoretic mobility shift assay.

	REFERENCES

TOP ABSTRACT INTRODUCTION EXPERIMENTAL PROCEDURES RESULTS DISCUSSION REFERENCES

CiteULike

Complore

Connotea

Del.icio.us

Digg

Reddit

Technorati

This article has been cited by other articles:

				C. H. Eggers, M. J. Caimano, and J. D. Radolf Analysis of Promoter Elements Involved in the Transcriptional Initiation of RpoS-Dependent Borrelia burgdorferi Genes J. Bacteriol., November 1, 2004; 186(21): 7390 - 7402. [Abstract] [Full Text] [PDF]

				C. Checroun, P. Bordes, O. Leroy, A. Kolb, and C. Gutierrez Interactions between the 2.4 and 4.2 regions of {sigma}S, the stress-specific {sigma} factor of Escherichia coli, and the -10 and -35 promoter elements Nucleic Acids Res., January 2, 2004; 32(1): 45 - 53. [Abstract] [Full Text] [PDF]

				R. Hernandez-Castro, M. C. Rodriguez, A. Seoane, and J. M. Garcia Lobo The aquaporin gene aqpX of Brucella abortus is induced in hyperosmotic conditions Microbiology, November 1, 2003; 149(11): 3185 - 3192. [Abstract] [Full Text] [PDF]