THE JOURNAL OF BIOLOGICAL CHEMISTRY © 2002 by The American Society for Biochemistry and Molecular Biology, Inc.

Vol. 277, No. 38, Issue of September 20, pp. 35567–35573, 2002 Printed in U.S.A.

Using Hetero-11-mers Composed of Wild Type and Mutant Subunits to Study Tryptophan Binding to TRAP and Its Role in Activating RNA Binding* Received for publication, June 13, 2002, and in revised form, July 9, 2002 Published, JBC Papers in Press, July 19, 2002, DOI 10.1074/jbc.M205910200

Pan T. X. Li‡ and Paul Gollnick§ From the Department of Biological Sciences, State University of New York, Buffalo, New York 14260

Tryptophan biosynthesis in Bacillus subtilis requires the products of seven trp genes (1). Expression of these genes, which are located in the trpEDCFBA and folate operons, is negatively regulated by TRAP (trp RNA-binding attenuation protein) in response to changes in the intracellular levels of L-tryptophan (2– 4). TRAP regulates both transcription of the trp operon and translation of its initial structural gene, trpE. In the presence of excess tryptophan, TRAP binds to the 5⬘ leader region of the nascent trp mRNA, facilitating formation of a transcription terminator, which halts transcription prior to the structural genes. Under conditions of limiting tryptophan, TRAP does not bind to the trp leader RNA, thus allowing it to form an alternative antiterminator structure that permits transcription to continue into the structural genes (2). TRAP also binds to the leader region of nonterminated trp mRNAs * This work was supported by Grants GM62750 from the National Institutes of Health and MCB 9982652 from the National Science Foundation (to P. G.) and funds from the Mark Diamond Research Fund of State University of New York at Buffalo (to P. T. X. L.). The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked “advertisement” in accordance with 18 U.S.C. Section 1734 solely to indicate this fact. ‡ Present address: Dept. of Chemistry, University of California, Berkeley, CA 94720. § To whom correspondence should be addressed. Tel.: 716-645-2887; Fax: 716-645-2975; E-mail: [email protected]. This paper is available on line at http://www.jbc.org

that have extended through the attenuator region, and this binding alters the RNA structure so as to sequester the trpE Shine-Dalgarno sequence and inhibit translation initiation of this gene (3–5). Translation of trpG, which is located within the folate operon, is also regulated by TRAP; in regulating expression of this gene, TRAP competes directly with ribosomes for binding to the mRNA (3, 6, 7). TRAP also regulates translation of yhaG (8) and ycbK (9). The former gene is predicted to encode a tryptophan transport protein (8), whereas the function of the latter gene product is unknown. TRAP is composed of 11 identical subunits arranged in a symmetric ring (10, 11). The secondary structure mainly consists of 11 seven-stranded antiparallel ␤-sheets, each formed from ␤-strands from two adjacent subunits. 11 tryptophanbinding sites are formed between adjacent subunits, and each tryptophan interacts with amino acid residues of both subunits (see Fig. 1). When activated by 11 tryptophan molecules, TRAP binds to RNAs containing multiple (up to 11) (G/U)AG triplet repeats optimally separated by two or three nonconserved nucleotides (10, 12, 13). The crystal structure of TRAP complexed with an RNA containing 11 GAG repeats shows that the singlestranded RNA wraps entirely around the outer perimeter of the protein ring forming specific interactions between the bases of the RNA and the protein (14). The observations that TRAP is composed of 11 identical subunits that create 11 binding sites for L-tryptophan between adjacent subunits and that TRAP binds RNAs with up to 11 (G/U)AG repeats (with restricted spacing between repeats) raise several issues with regard to how this protein functions to regulate gene expression. TRAP must avoid binding to RNAs with only several (G/U)AG repeats, and it must only bind its various RNA targets at the appropriate level of intracellular tryptophan. Little is known about how tryptophan binding activates TRAP to bind RNA. The structure of TRAP in the absence of tryptophan is not known; however, previous studies have shown that the apo-protein remains an 11-mer (15). In particular, the relationship between the number of bound tryptophan molecules and the number of RNA-binding sites activated is not known. Because the binding of 11 molecules of tryptophan to TRAP is positively cooperative (10, 16, 17), it has been difficult to directly analyze the properties of TRAP 11mers containing defined subsaturating levels of bound tryptophan. Recently we have developed a method to generate TRAP hetero-11-mers composed of two different types of subunits, such as mutant and wild type (WT).1 Our data show that the mutant subunits can assemble randomly with WT subunits yielding a total of 12 different species of 11-mers (the two 1 The abbreviations used are: WT, wild type; 5-IAF, 5-iodoacetamidofluorescein.

35567

Downloaded from www.jbc.org at STATE UNIV OF NEW YORK ALBANY on March 5, 2007

Expression of genes involved in tryptophan metabolism in Bacillus subtilis is regulated by the TRAP protein in response to changes in L-tryptophan levels. TRAP binding to several RNA targets that contain between 9 and 11 (G/U)AG repeats regulates transcription and/or translation of these genes. TRAP consists of 11 identical subunits and is activated to bind RNA by binding up to 11 molecules of tryptophan. To investigate the mechanism by which tryptophan binding activates TRAP, we generated hetero-11-mers containing different proportions of subunits from wild type (WT) TRAP that bind tryptophan and from a mutant TRAP (Thr25 to Ala) defective in tryptophan binding. Studies of these hetero11-mers show that tryptophan-binding sites created from active subunits bind tryptophan with similar affinity to those in WT homo-11-mers, whereas sites containing the T25A substitution do not bind tryptophan. Hetero-11-mers with very few (one or two) bound tryptophans show only 10-fold lower affinity than WT TRAP for an RNA with 11 GAG repeats, whereas TRAP with no bound tryptophan shows no detectable binding to this RNA. We also demonstrate that tryptophan binding induces a conformational change in TRAP in the vicinity of the RNA-binding site, suggesting a possible mechanism for activation of RNA binding.

35568

Tryptophan Binding to TRAP and Its Role in RNA Binding MATERIALS AND METHODS

homo-11-mers and 10 different possible hetero-11-mers) (15). By varying the ratio of the two types of subunits that are mixed, we can change the abundance of each species within the mixture of renatured hetero-11-mers. In this paper, we created TRAP hetero-11-mers composed of subunits from WT TRAP and subunits from a mutant protein defective in tryptophan binding. Residues from two adjacent subunits contribute to the formation of each tryptophan-binding site on TRAP (10) (Fig. 1). Hence within the hetero-11-mers there are two different types of tryptophan-binding sites. In one case the binding site contains the altered residue from the mutant subunit. Our data indicate that these sites do not bind tryptophan. The second type of binding site is created from the juxtaposition of two fully functional regions from the two adjacent active subunits, either two wild type subunits or the unaltered portion of a mutant subunit combined with that from a wild type subunit. Our results show that these sites bind tryptophan with an affinity similar to that of those in the wild type homo-11-mer. We also show that binding only one or two tryptophan molecules to hetero-11-mers (those with only one or two active binding sites) dramatically stabilizes the TRAP-RNA complex as compared with TRAP 11-mers with no bound tryptophan. The affinity of these TRAP hetero-11-mers for RNA depends on both the number of tryptophan molecules bound and the number of (G/U)AG triplet repeats in the target RNA, suggesting that nonliganded subunits within the hetero-11-mers contribute to the stability of complexes with RNA. Our studies also indicate that tryptophan binding induces a conformational change in only the subunits to which it is bound. Together, our results suggest that binding one or two tryptophan molecules activates RNA-binding sites associated with these tryptophanbinding sites and that this activation plays a crucial role in nucleating the TRAP-RNA interaction. Once this initial RNAbinding complex is formed, the remaining RNA-binding sites on the protein can interact with the RNA even though they have not been activated by bound tryptophan.

⌬␪% ⫽ a*共关Trp兴/S0.5)n)/(1⫹([Trp]/S0.5)n)

(Eq. 1)

where a is the saturation level of bound tryptophan ([Trp]b) and S0.5 represents the binding affinity and is the defined as the concentration of free tryptophan ([Trp]f) when [Trp]b is at 50% of saturation. The Hill coefficient (n) reflects the cooperativity of the binding. Binding is noncooperative for n ⫽ 1.0, positively cooperative for n ⬎1.0, and negatively cooperative for n ⬍1.0. When S0.5 values were normalized based on the number of WT subunits in the heteromers, the Hill coefficient was set to 1.0 for fitting because we found no cooperativity for tryptophan binding to these heteromers. RNA binding to TRAP in the presence of 100 mM tryptophan was determined using a nitrocellulose filter binding assay at 37 °C (15, 20). The reactions were incubated at 37 °C for 1 h before filtering. For each assay, at least one set of reactions was performed in the absence of tryptophan, as a control. The data were analyzed using a nonlinear least squares fitting to single binding site equation (20) (Prizm, Graphpad Software Inc., San Diego, CA). Fluorescence Labeling Assay—200 ng of 5-iodoacetamidofluorescein (5-IAF; Molecular Probes) in 50 mM sodium phosphate (pH 7.5) was quickly mixed with 10 ␮g of TRAP in a total of 100 ␮l. The reaction was stopped after 0.5–20 min by adding SDS-PAGE loading buffer containing 0.1 M ␤-mercaptoethanol. The samples were run on 15% SDS-PAGE to separate free dye from labeled proteins. Digital images of the gel were taken using a 312-nm transilluminator and a Kodak DC290 digital camera. Fluorescence intensities of bands were quantified using 1D Kodak software, version 3.5. The number of 5-IAF molecules bound per TRAP 11-mer was determined by comparison with intensities of free fluorophore at known concentrations. To adjust the loading errors, the gels were subsequently stained with Coomassie Brilliant Blue to determine the protein concentration. RESULTS AND DISCUSSION

Constructing WT-T25A Heteromeric TRAP Mixtures—We created heteromeric TRAP 11-mers composed of various ratios of subunits derived from wild type 11-mers and from mutant 11-mers that are inactive for tryptophan binding, so as to mimic situations in which the wild type TRAP protein has different numbers of bound tryptophan molecules. As the source of inactive subunits, we used the mutant TRAP protein T25A (Thr25 in the tryptophan-binding pocket changed to Ala).

Downloaded from www.jbc.org at STATE UNIV OF NEW YORK ALBANY on March 5, 2007

FIG. 1. Ribbon diagram of three adjacent TRAP subunits bound to RNA. Each subunit is shown in a different color: yellow, olive, and green. The ␤-strands are shown as arrows, and the bound 25 L-tryptophan molecules are shown as van der Waals’ spheres. The Thr side chain on each subunit is shown in ball-and-stick models. Each tryptophan-binding site is created from elements on two adjacent subunits, and each subunit contributes to two tryptophan-binding sites. The bound RNA is shown in ball-and-stick format. The native TRAP protein contains 11 subunits arranged in a ring.

RNA Synthesis and TRAP Purification—Wild type and mutant mtrB genes encoding TRAP were subcloned on EcoRI/HindIII fragments following the T7 promoter in pBlueScript (Stratagene) and were expressed in the Escherichia coli strain BL21 as described previously (18). We followed a previously published protocol to purify TRAP by immunoaffinity chromatography (19). The concentration of each protein was determined by UV absorbance (extinction coefficient of 1280 M⫺1 cm⫺1 at 280 nm) and confirmed by BCA protein assay (Pierce) and SDSPAGE. (GAGAU)n RNA, where n indicates the number of tandem GAGAU repeats, was transcribed in vitro using T7 RNA polymerase and labeled with [␣-32P]UTP (PerkinElmer Life Sciences) as described previously (19). Subunit Mixing—Subunit mixing was performed as described previously by Li et al. (15). Various ratios of WT and mutant TRAP proteins (0.5–5.0 mg/ml) were denatured in 4 M guanidine hydrochloride (Angus, Niagara Falls, NY) at room temperature for 1 h. The mixtures were then dialyzed against 50 mM phosphate buffer (pH 8.0) overnight. The concentrations of renatured proteins were quantified by BCA protein assay (Pierce) and SDS-PAGE based on comparison with known TRAP standards. Tryptophan and RNA Binding—Tryptophan binding to TRAP was measured at 37 °C by circular dichroism spectroscopy using a JASCO model J-715 spectrapolarimeter. The spectra were recorded between 190 and 300 nm for 5 ␮M TRAP 11-mers in the presence of various concentrations of tryptophan. At each tryptophan concentration, the spectrum of free tryptophan was subtracted. The CD at 225 nm, which shows the maximal difference between spectra of apo- and liganded TRAP, was used to measure tryptophan binding. CD225 (␪) of free tryptophan increases linearly with concentration, whereas the net CD signal (⌬␪) from bound tryptophan to TRAP saturates at high concentrations of tryptophan. The values of ⌬␪ were normalized to give ⌬␪% by comparing to maximal CD signal change at saturation and ⌬␪% were fit to following the Hill equation (10, 18, 21).

Tryptophan Binding to TRAP and Its Role in RNA Binding

35569

FIG. 2. Native polyacrylamide gel electrophoresis of WT-T25A hetero-11-mers. WT and T25A TRAP were denatured in 4.0 M guanidine HCl at room temperature for 1 h, mixed in various ratios, dialyzed in 50 mM Na2HPO4 (pH 8.0) overnight, and then run on a 9% native polyacrylamide gel. The gel was stained with Coomassie Brilliant Blue.

FIG. 3. Tryptophan binding measured by circular dichroism spectroscopy. A, CD spectra of 12 ␮M WT TRAP in 50 mM sodium phosphate (pH 8.0) (solid line) and with the addition of 100 ␮M Ltryptophan (dashed line) at 37 °C. Free tryptophan background was subtracted from the spectrum. B, tryptophan binding of 12 ␮M WT (f) or T25A (⽧) TRAP as well as heteromeric mixtures of 75% WT and 25% T25A (Œ), 50% WT and 50% T25A (●), and 25% WT and 75% T25A () as measured by CD at 228 nm. The data represent the averages of three experiments with standard deviation less than 10% of the mean.

T25A (data not shown), which are defective in tryptophan binding. These observations establish that this change is due to tryptophan binding to TRAP. Examining the CD228 of WT TRAP as a function of tryptophan concentration at 37 °C (Fig. 3B) yielded an apparent S0.5 of 24 ␮M and a Hill coefficient (n) of 1.2 (Table I), both values are similar to those derived previously from equilibrium dialysis of wild type TRAP at 5 °C (S0.5 ⫽ 5–10 ␮M and n ⫽ 1.5–2.0) (10, 17, 18), although the S0.5 value is somewhat higher, and the Hill coefficient is slightly lower. These differences may be due to the different temperature used in the current studies. We next examined the tryptophan binding properties of heteromeric mixtures composed of various ratios of WT and T25A subunits (Fig. 3B and Table I). We measured the number of tryptophan molecules/TRAP protein bound in these pools. To do so, we compared the saturation level of tryptophan binding to each pool to that of WT TRAP, which is known to bind 11 tryptophan molecules/11-mer (10). Because the heteromer pools contain mixtures of hetero-11-mers with different ratios of WT and mutant subunits, the values we measured represent weighted averages (and thus cannot be directly ascribed to any one member of the pool). We found that the saturation of tryptophan binding to the heteromeric mixtures depends

Downloaded from www.jbc.org at STATE UNIV OF NEW YORK ALBANY on March 5, 2007

This protein is defective in tryptophan binding and does not bind RNA (18). Because the T25A protein contains all the residues shown to be involved in complex formation with RNA (14, 18), we assume that the RNA-binding sites on this protein are structurally identical to those in WT TRAP. The active subunits were from a mutant TRAP protein in which Lys71 is replaced by Ala (K71A). The advantage of using this protein is that because of the charge change (Lys to Ala) on the surface of each subunit, K71A TRAP displays faster mobility on native polyacrylamide gels than the T25A protein. This property allows us to distinguish hetero-11-mers with different subunit compositions using these gels (Fig. 2). Lys71 is distant from the tryptophan and RNA-binding sites on TRAP (10, 14), and K71A TRAP is fully active both in vivo (18) and in vitro (15). Moreover, all hetero-11-mers composed of WT and K71A subunits are fully active in tryptophan binding and in RNA binding (15). For simplicity, we will refer to K71A subunits as WT in this paper. To construct WT-T25A TRAP hetero-11-mers, we denatured WT and T25A TRAP 11-mers into unfolded monomers using guanidine hydrochloride, mixed the two types of subunits in various ratios, and regenerated 11-mers by dialysis in sodium phosphate buffer (15). All pools of heteromers have similar secondary structure as WT TRAP homo-11-mers, based on CD analysis, and all have assembled as 11-mers based on size exclusion chromatography (data not shown). When displayed on native polyacrylamide gels, these pools show a total of 12 different bands corresponding to the two homo-11-mers (WT and T25A) and 10 hetero-11-mers with different numbers of each type of subunit (10WT-1T25A, 9WT-2T25A, etc.) (Fig. 2). The distribution of heteromers on the gel is nearly identical to that predicted based on random association of mutant and WT monomers. Thus there is no evidence suggesting preferential association of either type of monomer. We previously observed similar random assembly of 11-mers when monomers from several other TRAP mutant proteins were mixed with WT subunits (15). Tryptophan Binding of WT-T25A Heteromeric Mixtures— Assembly of WT and T25A subunits into hetero-11-mers creates two different types of tryptophan-binding sites, those with Thr at position 25 (from the WT subunits) and those with the Ala substitution at position 25 (from the T25A subunits) (Fig. 1). To characterize the tryptophan binding properties of WTT25A heteromers, we developed an assay based on changes in the CD spectrum of TRAP upon tryptophan binding (Fig. 3). The CD spectrum of WT TRAP in the absence of tryptophan (apo-TRAP) is typical of a protein composed predominately of ␤-sheet secondary structure, showing a negative peak near 215 nm (Fig. 3A) (15, 20). In the presence of tryptophan, a new positive peak appears near 228 nm. This spectral change occurs for all TRAP proteins that we tested that bind tryptophan but does not occur in the spectra of mutant TRAP proteins, such as

35570

Tryptophan Binding to TRAP and Its Role in RNA Binding

TABLE I Tryptophan binding of WT-T25A heteromeric mixtures as measured by CD225 NA, not applicable; NB, no measurable binding with up to 100 ␮M protein.

n Saturationa S0.5 (total)b S0.5 (WT)b

nm

100% WT and 0% T25A

75% WT and 25% T25A

50% WT and 50% T25A

25% WT and 75% T25A

1.2 ⫾ 0.1 100% 24 ⫾ 2 ␮M 24 ⫾ 2 ␮M

0.96 ⫾ 0.06 69 ⫾ 1% 27 ⫾ 2 ␮M 20 ⫾ 2 ␮M

0.94 ⫾ 0.07 52 ⫾ 2% 37 ⫾ 4 ␮M 18 ⫾ 2 ␮M

0.89 ⫾ 0.09 21 ⫾ 2% 67 ⫾ 12 ␮M 17 ⫾ 3 ␮M

0% WT and 100% T25A

NA 0% NB NB

a

Maximal CD228nm for binding by WT TRAP is normalized to 100%. S 0.5, the apparent affinity for tryptophan, was determined by fitting data into the Hill equation with the Hill coefficient (n) fixed at 1.0. S0.5 (total) is determined for the total TRAP concentration, and S0.5 (WT) is normalized to the percentage of WT TRAP in the mixture. The data represent the averages of at least two independent experiments. b

absence of tryptophan shows measurable binding to this RNA (at protein concentrations up to 5 ␮M). T25A TRAP has all the residues shown to be involved in RNA binding in the crystal structure of the TRAP RNA complex (14) as well as by biochemical studies (18). However, T25A TRAP is inactive because it does not bind tryptophan, which is necessary to activate TRAP for RNA binding. We measured the affinity of several pools of hetero-11-mers containing various ratios of WT to T25A subunits for the RNA (GAGAU)11, in the presence of excess tryptophan (Table II). The affinity of the hetero-11-mer pools for this RNA decreased from a Kd of 1.6 to 12 nM as the percentage of WT subunits in the mixture dropped from 100 to 5% (Table II). In every instance, RNA binding was tryptophan-dependent (data not shown). The presence of RNA did not alter the saturation level of tryptophan binding (data not shown), indicating that bound RNA does not influence the ability of binding sites containing the T25A substitution to bind tryptophan. We have also found that bound RNA does not alter the tryptophan binding properties of WT TRAP.2 Several lines of evidence suggest that hetero-11-mers of WT and T25A TRAP form similar complexes with RNA as does WT TRAP. The heteromers form complexes with (GAGAU)11 that migrate with similar mobility on native gels as that with WT TRAP (data not shown). In addition, near-UV CD spectroscopy studies (data not shown) indicate that the RNA bound to these hetero-11-mers undergoes similar conformational changes as occur upon complex formation with WT TRAP (20, 23). Surprisingly, the hetero-11-mer pools containing only 5–10% WT subunits bound (GAGAU)11 RNA with Kd values of 12 and 6.5 nM, respectively, which are only 7.5- and 4-fold weaker than for WT TRAP, even though the hetero-11-mers in these pools were predicted to contain an average of only approximately one WT subunit. In contrast, we detect virtually no specific RNA binding to T25A homo-11-mers (Table II). One explanation for these findings could be that the observed RNA binding affinity is due to the presence of a fraction of WT homo-11-mers or hetero-11-mers with large numbers of WT subunits within our heteromer pools. The measured affinity (as Kd) for any pool is the weighted average of the dissociation constants for each of the TRAP hetero-11-mer species comprising the pool. Our results show that subunit mixing to generate these heteromers is based on random assortment (Fig. 2), and hence the composition of each pool is predicted by a binomial distribution (15). Hence WT homo-11-mers are predicted to represent less than 1:1014 of the population in the mixture with 5% WT subunits (WT11 ⫽ 0.0511 %) and less than 1.5% of the total 11-mers in this pool will have more than two WT subunits. Native gel analysis of these pools is consistent with these predictions because all the visible bands run just adjacent to the T25A homo-11-mer, i.e. there are no visible bands corresponding to hetero-11-mers with large numbers of WT subunits (data not

2

P. Li, C. Baumann, and P. Gollnick, unpublished observations.

Downloaded from www.jbc.org at STATE UNIV OF NEW YORK ALBANY on March 5, 2007

directly on the percentage of WT subunits used in the pool (Table I). These findings were confirmed by equilibrium dialysis (data not shown). The simplest interpretation of these results is that assembly of hetero-11-mers composed of WT and T25A mutant subunits is random and that tryptophan-binding sites containing Thr25 contributed from WT subunits retain their ability to bind tryptophan, whereas those with the alanine substitution at residue 25 do not bind tryptophan under these conditions. We refer to the former type of binding sites as active binding sites and the latter as inactive binding sites. Thus in the presence of excess tryptophan, the hetero-11-mers created with different ratios of WT and T25A subunits can be considered to represent randomly assembled TRAP 11-mers with various numbers of tryptophan molecules bound, depending on their subunit composition. The observed affinity of the heteromeric pools for tryptophan (S0.5 total)) decreased from 24 ␮M for the WT homo-11-mer to 67 ␮M for the pool composed of 25% WT and 75% T25A (Table I). However, if we assume that only the active binding sites within the hetero-11-mers bind tryptophan (as indicated above) and normalize the observed affinities to the fraction of active binding sites in the mixtures, we find that the affinity of the active sites (S0.5 (WT)) does not change significantly within the hetero11-mer pools (Table I). The Hill coefficient (n) decreases slightly from 1.2 to 0.89 as the percentage of WT subunit in the pool drops from 100 to 25%, suggesting that introduction of the mutant subunits interferes with the cooperativity of tryptophan binding. The Hill coefficient of 1.2–2.0 for WT TRAP binding 11 tryptophan molecules as determined by equilibrium dialysis (10, 17, 18) or by CD measurements (Fig. 3) indicates a low degree of positive cooperativity. This is because the maximal possible value of n for fully cooperative binding in this system is 11, the total number of binding sites on the protein, and n equals 1.0 for noncooperative binding (21, 22). Our results are consistent with a model in which, following tryptophan binding to one or mores sites on the TRAP 11-mer, two types of unoccupied tryptophan-binding sites are generated: “empty” and “activated” sites. The simplest model for the generation of cooperativity is that tryptophan binding to the first site activates only the neighboring site(s) and increases their affinity for tryptophan. However, we cannot rule out a model in which tryptophan binding to one site on TRAP moderately increases the affinities of the remaining 10 unoccupied sites. The most significant implication of our findings is that there may be conditions in vitro or in vivo where wild type TRAP binds to less than its full complement of 11 tryptophan molecules. RNA Binding of WT-T25A Heteromeric Mixtures—We next addressed the question of how many bound tryptophans are required to activate the TRAP 11-mer for RNA binding? WT TRAP binds (GAGAU)11, an RNA with 11 tandem GAGAU repeats, with a Kd of 1.6 nM in the presence of excess tryptophan (Table II). However, neither WT TRAP in the absence of tryptophan nor T25A TRAP homo-11-mer in the presence or

Tryptophan Binding to TRAP and Its Role in RNA Binding

35571

TABLE II The affinity (Kd) of WT-T25A mixtures for RNAs with different numbers of (GAGAU) repeat The equilibrium dissociation constants (Kd) were determined by filter binding in the presence of 100 ␮M L-tryptophan. Each value is the average of at least four individual experiments. The standard deviation is generally less than 10% of the mean, but for Kd values less than 10 nM, the standard deviation is less than 50% of the mean. Each assay was also repeated in the absence of tryptophan as control. NB, no detectable binding up to 5 ␮M TRAP. Kd 0% WT and 100% T25A

5% WT and 95% T25A

10% WT and 90% T25A

20% WT and 80% T25A

30% WT and 70% T25A

40% WT and 60% T25A

60% WT and 40% T25A

80% WT and 20% T25A

100% WT and 0% T25A

11 9 8 7 6 5 4 3

NB NB NB NB NB NB NB NB

12 36 39 57 176 600 1023 5886

6.5 9.7 13 29 29 45 194 3653

nM 6.9 8.4 11 19 24 34 291 864

6 9.4 9.8 16 21 22 126 462

2.9 4.8 3.4 4.9 9 7.7 54 224

3.1 1.3 2 3.4 4.2 4.4 40 162

1.9 2 1.5 2.4 3.7 2.4 29 124

1.6 2 2.7 1.8 3.3 3.7 22 94

shown). In the mixture with 5% WT subunits, if all hetero-11mers containing three or more WT subunits bind (GAGAU)11 RNA with the same affinity as WT homo-11-mer (and the remaining heteromers do not contribute to the observed affinity), the predicted apparent Kd for the mixture would be 106 nM (1.6 nM/1.5%). This value is much higher than that observed (12 nM). Hence these data suggest that the binding of as few as one or two tryptophan molecules sufficiently activates a TRAP 11mer so that its complex with RNA is significantly stabilized. It is possible that the observed affinities of (GAGAU)11 RNA for the members of the hetero-11-mer pools containing 5 or 10% WT subunits (Kd values of 12 and 6.5 nM) are the results of only the tryptophan-bound subunits (WT and possibly the adjacent T25A subunits) in the heteromers interacting with the RNA. If this were the case, most of the T25A subunits in the heteromers would not contribute to the complex. However, several lines of evidence argue against this hypothesis. We have previously shown that RNAs with three or fewer GAG repeats bind to WT TRAP very weakly (Kd ⱖ 800 nM) (24). Therefore, it seems unlikely that the high affinities that we observed for heteromers with low percentages of WT subunits are due solely to the interactions between the few WT subunits and a small number of GAG repeats. Thus our results suggest that in these hetero11-mers, many (or all) of the T25A subunits contribute to the stability of the complex with RNA, even though they have not bound tryptophan. To test this hypothesis, we measured the affinity of these heteromeric TRAP pools for RNAs containing 3–11 GAG repeats (Table II). For all of the heteromeric pools, the affinity for RNA increased as the number of triplet repeats increased. This effect is particularly evident for pools that contain low (5–10%) fractions of WT subunits. Because the hetero-11-mers in these low WT percentage pools contain an average of approximately one WT subunit each, these results strongly suggest that RNA-binding sites associated with defective tryptophan-binding sites contribute to the complex with RNA in these hetero-11-mers. This conclusion is further supported by nuclease protection studies that demonstrate that all 11 trinucleotide repeats are protected from digestion by the heteromeric TRAP pools.3 In previous studies, we also obtained several lines of evidence suggesting that mutant subunits (15) or partially defective triplet repeats (24) contribute to the stability of the TRAP-RNA complex when accompanied by at least one WT subunit and one fully functional RNA triplet repeat. Tryptophan Binding Induces a Conformational Change in TRAP—The mechanism by which bound tryptophan activates TRAP to bind RNA is not known. Several studies have shown 3

P. Li and P. Gollnick, unpublished observations.

that the protein remains an 11-mer in the absence of bound tryptophan (15),4 and hence tryptophan binding does not activate TRAP by assembling the 11-mer. Furthermore, neither the bound tryptophan nor any of the amino acid residues that contact tryptophan directly interact with the RNA (14) (Fig. 1). Hence we have proposed that tryptophan binding induces a conformational change in TRAP that activates its RNA binding ability (10, 14, 18). Here we report the first direct evidence for this hypothesis. RNA binds to activated TRAP by wrapping around the outside of the protein ring. Each GAG repeat in the bound RNA interacts with residues on two adjacent protein subunits including Lys37 and Glu36 from one subunit and Phe32, Lys56, and Arg58 from the adjacent subunit (14). Asn20 lies in the RNAbinding region between residues Lys37 and Arg58 but does not interact with the RNA. Asn20 is also distant from the tryptophan-binding sites and forms no direct contact with tryptophan. Virtually any other amino acid can be substituted for Asn20 without affecting TRAP function (18). In particular, TRAP with cysteine at position 20 (N20C) binds tryptophan and RNA similarly to WT TRAP (data not shown). Because there are no other cysteine residues in TRAP (25), it is possible to specifically fluorescently label position 20 of N20C TRAP with 5-IAF, a thiol-reactive fluorophore. After incubation with 5-IAF, N20C TRAP shows strong green fluorescence (Fig. 4), whereas there is virtually no fluorescence from similarly treated WT TRAP (data not shown). We found that the rate of labeling N20C TRAP is greatly reduced by the presence of L-tryptophan but is not affected by D-tryptophan (Fig. 4), suggesting that Cys20 is more accessible to 5-IAF in apo-TRAP than in the TRAP-tryptophan complex. These findings support the hypothesis that bound tryptophan induces conformational changes in the vicinity of the RNA-binding sites of TRAP. There are two obvious interpretations of the observation that binding of just one tryptophan molecule to a TRAP 11-mer significantly activates TRAP to bind RNA. The first possibility is that the binding of one tryptophan molecule to these hetero11-mers induces a conformational change in only the subunits to which it binds but that this dramatically stabilizes the TRAP-RNA complex as compared with an entirely unliganded 11-mer (either defective in tryptophan binding or WT TRAP in the absence of tryptophan). The second possibility is that binding of one tryptophan molecule to a TRAP 11-mer induces a conformational change in many or all of the 11 subunits. To distinguish between these two possibilities, we examined 5-IAF labeling of hetero-11-mers created from 10% WT and 90% 4

D. Scott, personal communication.

Downloaded from www.jbc.org at STATE UNIV OF NEW YORK ALBANY on March 5, 2007

No. of GAGAU repeats

35572

Tryptophan Binding to TRAP and Its Role in RNA Binding

N20C/T25A double mutant TRAP subunits. These hetero-11mers contain, on average, one WT subunit, which is active in binding tryptophan but contains no Cys residues (and thus is not labeled by 5-IAF), and 10 N20C/T25A subunits, which can be labeled at Cys20 with 5IAF but contain the Thr25 to Ala substitution that eliminates tryptophan binding. If tryptophan binding to the WT subunits induces a conformational change over the entire hetero-11-mer, then the rate of labeling of this protein should be inhibited. However, we found that the rate of 5IAF labeling of both N20C/T25A mutant TRAP as well as the hetero-11-mer pool containing 10% WT and 90% N20C/T25A subunits was not reduced by added tryptophan (Fig. 4A). These data suggest that tryptophan binding does not induce conformational changes throughout the T25A subunits in these hetero-11-mers. These results thus favor the model in which tryptophan binding induces a conformational change only in the subunits to which it binds. Moreover, these findings suggest that the significant activation of RNA binding observed when only one or two tryptophan molecules are bound to TRAP is due to conformational changes in just a few subunits. Implications—Based on our previous studies using nucleoside analogs (24) as well as with TRAP hetero-11-mers containing WT subunits together with those from a mutant protein defective in RNA binding (15), we have proposed the following model for RNA binding to tryptophan-activated TRAP. Binding initiates between several subunits in the protein and one or two (G/U)AG repeats in the RNA target to form a binding initiation complex. This complex tethers the RNA to TRAP, thereby reducing its degrees of freedom, as well as partially aligns the remaining triplet repeats with the potential binding sites on the protein. Hence following formation of the initiation complex, the remaining triplet repeats in the RNA bind to TRAP cooperatively, possibly because of an increased local concentration of triplet repeats near the RNA-binding sites on the protein. We have shown that formation of the initiation complex requires interactions between one or two WT subunits and one

5

C. McElroy, M. Foster, and P. Gollnick, submitted for publication.

Downloaded from www.jbc.org at STATE UNIV OF NEW YORK ALBANY on March 5, 2007

FIG. 4. Accessibility of Cys20 for 5-IAF. A, SDS-polyacrylamide gel of N20C and N20C/T25A mutant TRAP proteins as well as a 10% WT and 90% N20C/T25A heteromeric mixture labeled with 5-IAF in the absence and presence of 1 mM L-tryptophan or D-tryptophan for 5 min. The gel was photographed on a UV transilluminator operating at 312 nm and photographed with a Kodak DC290 digital camera. B, time course of 5-IAF labeling of N20C TRAP in the presence (f) and absence (E) of tryptophan. The amounts of 5-IAF/protein were estimated by comparing fluorescence intensities of the proteins with that of the free fluorophore at known concentrations. The gel was subsequently stained with Coomassie Brilliant Blue to determine the protein concentration.

or two fully functional (G/U)AG repeats containing all the functional groups involved in the complex (15, 24). In contrast, our studies showed that after the initiation complex has formed, the remaining repeats in the RNA can interact with TRAP and contribute to the stability of the complex, even if they lack one of the functional groups that forms a critical hydrogen bond to the protein (24). Similar results were obtained with TRAP hetero-11-mers composed of mixtures of subunits from WT TRAP and a mutant protein lacking one of the amino acid side chains that interacts with the RNA. In this work we further studied the role of tryptophan binding in the formation of the TRAP-RNA complex. The low degree of cooperativity (n ⫽ 1.5–2.0 for 11 binding sites) of tryptophan binding to TRAP suggests that WT TRAP may bind fewer than 11 tryptophans under subsaturating conditions. Here we show that binding of as few as one or two tryptophans dramatically activates RNA binding to TRAP (Kd of 6 –12 nM for (GAGAU)11), as compared with the situation in which there is no tryptophan bound (Kd ⬎⬎ 5 ␮M). In view of our model presented above, these results suggest that tryptophan binding is required for initiation of RNA binding but is less essential for subsequent binding steps. Moreover, our results suggest that tryptophan binding induces a conformational change in TRAP in the vicinity of the RNA-binding site on the liganded subunit(s) (Fig. 4). Together, our findings suggest that tryptophan binding activates TRAP by inducing a conformational change in an RNA-binding site, which is essential for initiation of RNA binding, after which, the tethered RNA can also interact, although with reduced affinity, with the remaining binding sites on the protein, even if they have not been activated by bound tryptophan. Thus, these results also suggest that the RNA-binding sites on TRAP have weak affinity for (G/U)AG containing RNA even in the absence of tryptophan and that tryptophan binding serves to increase the affinity enough to allow formation of the initiation complex. In the crystal structure, RNA binds to tryptophan-activated TRAP by wrapping around the outside of the protein ring with each GAG repeat interacting with residues on two adjacent protein subunits (14). From one subunit, Lys37 interacts with the first G and the second A, whereas Glu36 hydrogen bonds with the third G in the RNA. From the adjacent subunit Phe32, Lys56, and Arg58 form hydrogen bonds with the third G (14). The results from a recent NMR study suggest that tryptophan binding to TRAP reduces the conformational dynamics of the protein in the region of the tryptophan-binding site as well as in the area where RNA binds, suggesting a mechanism for tryptophan activation of RNA binding.5 Our studies of hetero-11-mers composed of WT and T25A subunits that mimic partially saturated TRAP 11-mers show that the affinity of TRAP for RNA depends on both the number of tryptophan molecules bound to the protein as well as the number of triplet repeats in the RNA (Table II). WT TRAP binds weakly to RNAs with fewer than five GAG repeats with 14 – 60-fold lower affinity than for the RNA with 11 GAG repeats (Table II). The influence of the number of repeats on the affinity for TRAP is even greater for UAG repeats (24). Moreover, proteins with subsaturating numbers of bound tryptophan molecules show an even greater influence of the repeats on the stability of the complex (Table II). Hence this feature may be an important property of TRAP to prevent it from interacting with cellular RNAs that contain a few (G/U)AG but are not bona fide sites for regulation. In B. subtilis, TRAP-mediated regulation involves at least four different RNA targets (26). These contain between 9 and 11 triplet repeats, and in some of these there is rather subop-

Tryptophan Binding to TRAP and Its Role in RNA Binding

Acknowledgments—We thank Sathyamangalam V. Balsubramanian for assistance with instruments. We thank Jim Stamos, Alan Siegel, and Alfred Antson for preparation of figures and Gerald Koudelka and Charles Yanofsky for critical reading of the manuscript. We also thank the Pharmaceutical Sciences Instrumentation Facility at SUNY for CD data collection. REFERENCES 1. Henner, D., and Yanofsky, C. (1993) in Bacillus subtilis and Other Gram-positive Bacteria: Biochemistry, Physiology and Molecular Genetics (Sonenshein, A.,

2. 3. 4. 5. 6. 7. 8. 9. 10.

11. 12. 13. 14. 15. 16. 17. 18. 19. 20. 21.

22. 23. 24. 25. 26. 27. 28. 29. 30.

Hoch, J., and Losick, R., eds) pp. 269 –280, American Society of Microbiology, Washington, D.C. Kuroda, M., Shimotsu, H., Henner, D., and Yanofsky, C. (1986) J. Bacteriol. 167, 792–798 Kuroda, M., Henner, D., and Yanofsky, C. (1988) J. Bacteriol. 170, 3080 –3087 Du, H., and Babitzke, P. (1998) J. Biol. Chem. 273, 20494 –20503 Merino, E., Babitzke, P., and Yanofsky, C. (1995) J. Bacteriol. 177, 6362– 6370 Du, H., Tarpey, R., and Babitzke, P. (1997) J. Bacteriol. 179, 2582–2586 Yang, M., Saizieu, A., Loon, A. P. G. M., and Gollnick, P. (1995) J. Bacteriol. 177, 4272– 4278 Sarsero, J., Merino, E., and Yanofsky, C. (2000) J. Bacteriol. 182, 2329 –2331 Sarsero, J., Merino, E., and Yanofsky, C. (2000) Proc. Natl. Acad. Sci. U. S. A. 97, 2656 –26661 Antson, A. A., Otridge, J., Brzozowski, A. M., Dodson, E. J., Dodson, G. G., Wilson, K. S., Smith, T. M., Yang, M., Kurecki, T., and Gollnick, P. (1995) Nature 374, 693–700 Chen, X., Antson, A. A., Yang, M., Li, P., Baumann, C., Dodson, E. J., Dodson, G. G., and Gollnick, P. (1999) J. Mol. Biol. 289, 1003–1016 Babitzke, P., Bear, D., and Yanofsky, C. (1995) Proc. Natl. Acad. Sci. U. S. A. 92, 7916 –7920 Babitzke, P., Yealy, J., and Campanelli, D. (1996) J. Bacteriol. 178, 5159 –5163 Antson, A. A., Dodson, E. J., Dodson, G. G., Greaves, R. B., Chen, X., and Gollnick, P. (1999) Nature 401, 235–242 Li, P., Scott, D., and Gollnick, P. (2002) J. Biol. Chem. 277, 11838 –11844 Babitzke, P., and Yanofsky, C. (1995) J. Biol. Chem. 270, 12452–12456 Yakhnin, A., Trimble, J., Chiaro, C., and Babitzke, P. (2000) J. Biol. Chem. 275, 4519 – 4524 Yang, M., Chen, X., Militello, K., Hoffman, R., Fernandez, B., Baumann, C., and Gollnick, P. (1997) J. Mol. Biol. 270, 696 –710 Otridge, J., and Gollnick, P. (1993) Proc. Natl. Acad. Sci. U. S. A. 90, 128 –132 Baumann, C., Xirasagar, S., and Gollnick, P. (1997) J. Biol. Chem. 272, 19863–19869 Segel, I. H. (1975) in Enzyme Kinetics: Behavior and Analysis of Rapid Equilibrium and Steady-state Enzyme Systems (Segel, I. H., ed) pp. 360 –361, John Wiley & Sons Inc., New York Cantor, C., and Schimmel, P. (1980) in Biophysical Chemistry, Vol. III, pp. 864, W. H. Freeman and Company Xirasagar, S., Elliott, M., Bartolini, W., Gollnick, P., and Gottlieb, P. (1998) J. Biol. Chem. 273, 27146 –27153 Elliott, M., Gottlieb, P., and Gollnick, P. (2001) RNA 7, 85–93 Gollnick, P., Ishino, S., Kuroda, M., Henner, D., and Yanofsky, C. (1990) Proc. Natl. Acad. Sci. U. S. A. 87, 8726 – 8730 Babitzke, P., and Gollnick, P. (2001) J. Bacteriol. 183, 5795–5802 Yakhnin, H., Babiarz, J., Yakhnin, A., and Babitzke, P. (2001) J. Bacteriol. 183, 5918 –5926 Baumann, C., Otridge, J., and Gollnick, P. (1996) J. Biol. Chem. 271, 12269 –12274 Valbuzzi, A., and Yanofsky, C. (2001) Science 293, 2057–2059 Valbuzzi, A., Gollnick, P., Babitzke, P., and Yanofsky, C. (2002) J. Biol. Chem. 277, 10608 –10613

Downloaded from www.jbc.org at STATE UNIV OF NEW YORK ALBANY on March 5, 2007

timal spacing between the repeats (13). Moreover, transcription attenuation control of the trp operon probably requires that TRAP initially recognize and bind to the trp leader RNA before all 11 triplet repeats are synthesized (26). Regulation of expression of each of these genes or operons in response to the tryptophan level depends, in part, on the affinity of TRAP for each site as well as the abundance of each site. Hence they may be regulated differently in response to variations in tryptophan concentration. Consistent with this hypothesis, Yakhnin et al. (27) recently showed that regulation of translation of trpE requires a higher concentration of tryptophan in the growth medium than is required for transcription attenuation control of the trp operon. In addition to affecting the affinity of TRAP for its various RNA targets, the number of tryptophan molecules bound to TRAP may affect the kinetics of the interaction between TRAP and RNA. For example, the dissociation rate of the TRAP-RNA complex is dependent on the concentration of tryptophan (28). This rate is important in vivo because it affects the number of molecules of free TRAP available for binding to its target mRNAs. Another regulatory protein, AT (anti-TRAP), has been shown to bind the TRAP-tryptophan complex and prevent RNA binding (29, 30). Expression of AT is induced by uncharged tRNATrp (9). The availability of several regulatory mechanisms that can influence expression of genes involved in tryptophan biosynthesis in B. subtilis allows this organism to fine-tune tryptophan synthesis in response to changes in the intracellular level of free and tRNA-charged tryptophan.

35573