doi:10.1016/j.jmb.2003.11.002

J. Mol. Biol. (2004) 335, 707–722

Characterization of a trp RNA-binding Attenuation Protein (TRAP) Mutant with Tryptophan Independent RNA Binding Activity Pan T. X. Li and Paul Gollnick* Department of Biological Sciences, State University of New York, Buffalo, NY 14260 USA

TRAP (trp RNA-binding attenuation protein) is an 11 subunit RNA-binding protein that regulates expression of genes involved in tryptophan metabolism (trp) in Bacillus subtilis in response to changes in intracellular tryptophan concentration. When activated by binding up to 11 tryptophan residues, TRAP binds to the mRNAs of several trp genes and downregulates their expression. Recently, a TRAP mutant was found that binds RNA in the absence of tryptophan. In this mutant protein, Thr30, which is part of the tryptophan-binding site, is replaced with Val (T30V). We have compared the RNA-binding properties of T30V and wild-type (WT) TRAP, as well as of a series of hetero-11-mers containing mixtures of WT and T30V TRAP subunits. The most significant difference between the interaction of T30V and WT TRAP with RNA is that the affinity of T30V TRAP is more dependent on ionic strength. Analysis of the hetero11-mers allowed us to examine how subunits interact within an 11-mer with regard to binding to tryptophan or RNA. Our data suggest that individual subunits retain properties similar to those observed when they are in homo-11-mers and that individual G/UAG triplets within the RNA can bind to TRAP differently. q 2003 Elsevier Ltd. All rights reserved.

Keywords: protein– RNA interaction; tryptophan binding; ANS fluorescence; nucleoside analog; heteromer

*Corresponding author

Introduction Expression of genes involved in L -tryptophan metabolism in Bacillus subtilis is regulated in response to changes in intracellular tryptophan concentration through several mechanisms, each mediated by the trp RNA-binding attenuation protein (TRAP). When activated by binding tryptophan, TRAP binds to RNA targets that contain 9 to 11 GAG and/or UAG triplet repeats optimally separated by two or three non-conserved “spacer” nucleotides.1 – 4 To date, four TRAP binding sites Present address: P. T. X. Li, Department of Chemistry, University of California, Berkeley, CA 94720, USA. Abbreviations used: TRAP, trp RNA-binding attenuation protein; WT, wild-type; CD, circular dichroism; Kglutamate, potassium glutamate; ANS, 1anilinonaphthalene-8-sulfonic acid; bIAA, 3-bindoleacrylic acid; NTPaS, nucleoside phosphorothioates; G2 -Ome, 20 -O-methyl riboguanosine. E-mail address of the corresponding author: [email protected] 0

have been characterized in the B. subtilis genome.5 TRAP binds to a site in the leader region of the trpEDCFBA operon transcript and regulates transcription of the operon by an attenuation mechanism.6 – 9 TRAP binding to the same site in the leader region of read-through trp mRNAs induces formation of an RNA structure that sequesters the trpE ribosome binding site, thus inhibiting translation initiation of trpE,7,10,11 as well as reducing expression of the downstream trp genes through translational coupling and transcriptional polarity.12 trpG is located in the folate operon and TRAP regulates translation of this gene by directly competing with ribosomes for binding to trpG mRNA.13,14 The product of the yhaG gene is a putative tryptophan transport protein, and translation of this gene is also regulated by TRAP, apparently through a similar mechanism as for trpG.15 The rtpA gene encodes a regulatory protein called Anti-TRAP (AT), which binds to tryptophan-activated TRAP and inhibits its ability to bind RNA, thus elevating expression of the trp genes.16,17 The rtpA-ycbK operon also

0022-2836/$ - see front matter q 2003 Elsevier Ltd. All rights reserved.

708

contains a series of triplet repeats characteristic of TRAP binding sites,18 although the role and mechanism of TRAP regulation of these genes has not yet been characterized. TRAP is a ring-shaped protein composed of 11 identical subunits that form 11 seven-stranded antiparallel b-sheets (Figure 1).1,19 Each b-sheet consists of four b-strands from one subunit and three strands from an adjacent subunit. TRAP is activated to bind RNA by binding up to 11 molecules of L -tryptophan in pockets formed between adjacent subunits.20 – 22 In the crystal structures of TRAP complexed with several different RNAs containing 11 tandem GAG repeats separated by various dinucleotide spacers, the bound RNA forms a circle around the outer perimeter of the protein ring23,24 (Figure 1). TRAP does not undergo any significant conformational changes upon binding RNA as seen by comparing the crystal structures of the protein alone and in complex with RNA (r.m.s. difference in main-chain ˚ ).23 atoms is 0.39A The detailed structure of TRAP in the absence of tryptophan (apo-TRAP) is not known. However, biochemical and biophysical studies have shown

The T30V Mutant TRAP Protein

that apo-TRAP is also an 11-mer with extensive (D. Scott, b-sheet secondary structure25 unpublished observations). The observation that the bound tryptophan residues are nearly completely buried within the protein suggests that the binding pockets must exist in a different, more open conformation prior to tryptophan binding.1,19,23 Tryptophan binding is required to activate TRAP to bind RNA,20,21 and the RNAbinding site is distinct from the tryptophan binding site on the protein23 (Figure 1B). These observations suggest that tryptophan binding induces a conformational change in TRAP that affects the RNA-binding site. A recent study supports this hypothesis by showing that tryptophan binding alters the environment of Asn20, which is located within the RNA-binding site on the protein.22 Furthermore, two recent NMR studies that investigated the dynamic properties of TRAP in both the non-liganded and activated states showed that apo-TRAP is more flexible than tryptophan-bound protein.26,27 Recently a mutant TRAP, T30V (Thr30 replaced with valine) was found to bind trp leader RNA in the absence of tryptophan, although with lower

Figure 1. Structure of TRAP – RNA complex. A, Ribbon diagram of TRAP– RNA complex. Each of the 11 subunits is shown in a different color. The b-strands of the protein are shown as arrows connected by loops. The bound L -tryptophan molecules are shown in van der Waals spheres and the bound RNA is displayed in yellow sticks. B, A portion of the structure containing three adjacent subunits. b-strands B, C, D and E on the center subunit as well as the connecting loops BC and DE are labeled. Thr30 (magenta), as well as residues E36, K37, K56 and R58 (red), which directly contact the RNA, are displayed in ball-and-stick figures and labeled.

The T30V Mutant TRAP Protein

709

affinity than tryptophan-activated WT TRAP.28 Here, we investigated the tryptophan and RNAbinding properties of T30V TRAP as compared to the WT protein. Examining the binding of tryptophan analogs to WT and T30V TRAP revealed structural requirements of tryptophan for binding to both proteins and for activating WT TRAP to bind RNA. Our studies show that both WT and T30V TRAP bind RNA similarly; however, the stability of the T30V TRAP– RNA complex is more sensitive to changes in ionic strength. We created hetero-11-mers consisting of different ratios of WT and T30V TRAP subunits using a subunit mixing approach.22,25 The RNA-binding activity of the WT subunits in the hetero-11-mers can be controlled by the presence of tryptophan, whereas T30V subunits do not bind tryptophan. Examining these hetero-11-mers allowed us to study how T30V subunits interact with either activated or non-liganded subunits in the hetero-11-mers with regard to RNA-binding.

Results Comparing the tryptophan binding properties of T30V and WT TRAP Circular dichroism spectra of WT and T30V TRAP The circular dichroism (CD) spectrum of WT TRAP in the absence of tryptophan is typical of a protein composed mainly of b-sheets34 showing a strong minimum near 215 nm (Figure 2(a)).25,35 As we have shown previously, tryptophan binding to WT TRAP results in a new positive peak centered at 228 nm.22,35 This signal results from the bound tryptophan being in an asymmetric environment, rather than a significant change in the secondary structure of the protein upon tryptophan binding. In the absence of tryptophan, the CD spectrum T30V TRAP is nearly identical with WT TRAP (Figure 2(b)), indicating that the secondary structure of the mutant protein is similar to that of the WT protein. In contrast to WT TRAP, the presence of tryptophan does not significantly change the CD spectrum of T30V TRAP. This observation is consistent with previous studies showing that T30V TRAP does not bind tryptophan.28 Using ANS fluorescence to measure tryptophan binding to TRAP Previously we have used both CD spectroscopy22 and equilibrium dialysis with 14C-lableled tryptophan23,36 to measure tryptophan binding to TRAP. We have developed a new assay for tryptophan binding to TRAP using the fluorescent probe 1-anilinonaphthalene-8-sulfonic acid (ANS), which allows a rapid and more sensitive measurement of tryptophan binding. This assay is particularly useful to study TRAP mutants with low affinity for tryptophan or tryptophan analogs with low affinity for TRAP. ANS is an environment-sensitive probe

Figure 2. Effect of tryptophan binding on the CD spectra of WT and T30V TRAP. A, The CD spectra of 12 mM WT TRAP apo-protein (W) and in the presence of tryptophan (X); B, CD spectra of 12 mM T30V TRAP in the absence (A) and presence (B) of tryptophan.

such that its quantum yield increases greatly when bound to hydrophobic surfaces of proteins.37 This property of ANS has been used previously to characterize the binding of tryptophan to the trp repressor.38,39 ANS is weakly fluorescent in sodium phosphate buffer but fluoresces strongly when WT TRAP is present (Figure 3(a)). Adding excess L -tryptophan (1.0 mM) virtually eliminates ANS fluorescence in the presence of TRAP (Figure 3(a)). ANS and tryptophan compete for binding to TRAP as shown by equilibrium dialysis using 14C-lableled tryptophan, and by showing that ANS inhibits the ability of tryptophan to activate TRAP to bind RNA (data not shown). This observation is similar to prior results, which demonstrated that ANS competes with tryptophan for binding to the Escherichia coli trp repressor protein.38,39 We also found that ANS has a weaker affinity for TRAP than does tryptophan (S0.5 . 500 mM versus 5 mM; data not shown), yet the presence of only 1 mM ANS with TRAP provides a strong fluorescence signal. Hence, these properties allow us to use changes in ANS fluorescence to follow tryptophan binding to TRAP without complication from considering the competing equilibria when analyzing the data.

710

The T30V Mutant TRAP Protein

By analyzing the decrease in ANS fluorescence as a function of tryptophan concentration (Figure 3(b)), we derived an apparent S0:5 of 9 mM and a Hill coefficient n ¼ 1.4 for tryptophan binding to WT TRAP, both of which agree well with values obtained previously by equilibrium dialysis (S0:5 ¼ 11 mM; n ¼ 1:5)36 or by CD measurements (S0.5 ¼ 24 mM; n ¼ 1:2).22 Similar results were obtained using 5-fluorotryptophan, a tryptophan analog known to have similar properties to tryptophan for binding to, and activating TRAP.21 Moreover, the binding of tryptophan and 5-fluorotryptophan to TRAP, as measured by ANS fluorescence, also correlates well with activation of TRAP to bind RNA (Figure 3(b)). In contrast, tryptophan analogs that do not activate TRAP to bind RNA did not affect ANS fluorescence under the conditions of this assay (1.0 mM analog) (Table 1). The fluorescence emission spectrum and intensity of ANS in the presence of T30V TRAP are similar to that observed with WT TRAP (Figure 3(c)), indicating that ANS binds to T30V TRAP similarly as to WT TRAP. However, in this case there is no effect when excess tryptophan (1.0 mM) is added, suggesting that tryptophan does not compete with ANS for binding to T30V TRAP, which is again consistent with previous findings that T30V TRAP does not bind tryptophan.28 Tryptophan analogs binding to WT and T30V TRAP

Figure 3. Changes in ANS fluorescence upon binding to TRAP. (a) Emission spectra of 1 mM ANS in 50 mM sodium phosphate pH 8.0 ( £ ), with 5 mM WT TRAP (B) and in the presence of 5 mM TRAP and 1 mM tryptophan (W). (b) L -Tryptophan (A) and 5-fluoro-L -tryptophan (W) binding to 1 mM WT TRAP as measured by the decrease of ANS fluorescence (DFluorescence) as a function of ligand concentration. The maximal DFluorescence was set as 100% and all values were normalized accordingly. Activation of WT TRAP (1 mM) to bind (GAGUU)11 RNA was measured by filter binding and the percentage of RNA bound (Activation%) was plotted as a function of tryptophan (B) and 5-fluoro-L -tryptophan (X) concentration. The data represent at least three independent experiments with standard deviations less than 10% of the mean. For tryptophan binding to WT TRAP, S0:5 was determined to be 9 mM and n ¼ 1:4: (c) Emission spectra of 1 mM ANS with 2 mM T30V TRAP (B) and with 2 mM T30V in 1 mM tryptophan (W).

Babitzke & Yanofsky21 employed tryptophan analogs to study the structural features of tryptophan required for activating TRAP to bind RNA. We used competition with ANS as an assay to compare the requirements for individual moieties on tryptophan for binding to WT and to T30V TRAP. As shown in Table 1, altering either the indole ring (tyrosine and phenylalanine) or removing the amino group (3-b-indoleacrylic acid and indole-3-acetic acid) of tryptophan eliminates binding to WT TRAP. Two analogs in which the carboxyl group is changed, tryptamine (lacking the carboxyl group) and tryptophanol (a hydroxyl group instead), also show no detectable binding. However, L -tryptophanamide with only the carbonyl moiety of the carboxyl group binds weakly to WT TRAP. Both of these results are consistent with the role of the carboxyl group as a hydrogen bond acceptor as seen in the crystal structure.1 Furthermore, the structural features of tryptophan required for binding to WT TRAP are nearly identical with those required for activating the protein to bind RNA (Table 1).21 T30V TRAP binds to different tryptophan analogs than the WT protein. Specifically, T30V TRAP binds tryptophan analogs that lack the amino group, including 3-b-indoleacrylic acid, indole-3-propionic acid, indole-3-DL -lactic acid and indole-3-butyric acid, which do not bind to

711

The T30V Mutant TRAP Protein

Table 1. Binding and activation of TRAP by various L -tryptophan analogs ANS fluorescence (%)a

Activation (%)b

Ligand

WT

T30V

WTc

T30V

No ligand

100

100

0

100

L -tryptophan

4

NCd

100

100

D -tryptophan

NC

NC

NBe

86

L -phenylalanine

NC

NC

NB

93

L -tyrosine

NC

NC

NB

102

5-fluoro-L -tryptophan

4

NC

94

90

3-b-indoleacrylic acid

NC

58

NB

84

Indole-3-propionic acid

NC

73

NB

87

Indole-3-butyric acid

NC

83

NB

80

Indole-3-DL -lactic acid

NC

77

NB

84

Tryptamine

NC

NC

NB

81

85

NC

35

95

NC

NC

NB

97

L -tryptophanamide

Tryptophanol

a ANS (1 mM) and TRAP (1 mM) were mixed with various L -tryptophan analogs (1 mM) and fluorescence was measured at 480 nm. ANS fluorescence with TRAP in the absence of ligand was set as 100%. Data are the average of two independent experiments with the standard deviation less than 10% of the mean. b The binding of 10 nM TRAP to (GAGUU)11 RNA was measured in the presence of 1 mM tryptophan analogs.21 The percentage of RNA bound (RNAbound%) obtained with 1 mM L -tryptophan was set as 100%. Each value represents the average of four independent experiments with standard deviation less than 20% of the mean. c Values of activation for WT TRAP agree well with those reported previously.21 d NC, less than 10% fluorescence change was observed. e NB, less than 5% RNA bound.

712

Figure 4. Binding of tryptophan and bIAA to WT and T30V TRAP. Binding of tryptophan (B) and bIAA (X) to WT TRAP as well as binding of tryptophan (A) and bIAA (W) to T30V TRAP as measured by ANS fluorescence as described in the legend to Figure 3. Each value represents the average of three independent experiments with standard deviations less than 20% of the mean.

WT TRAP. Since Thr30 forms a hydrogen bond with the amino group of the bound tryptophan1 (see also Figure 1B), substituting valine at residue 30 will disrupt this hydrogen bond and should also increase the local hydrophobicity. Both of these changes are unfavorable for interacting with the amino group of tryptophan and hence removing the amino group partially restores the affinity of these analogs for T30V TRAP. Among these analogs, 3-b-indoleacrylic acid (bIAA) competes most strongly with ANS for binding to T30V TRAP, with an apparent Kd of 80 mM (Figure 4). The affinity of T30V for RNA is, however, not affected by the presence of up to 1.0 mM bIAA (data not shown).

The T30V Mutant TRAP Protein

Figure 5. Temperature dependence of the TRAP– RNA interaction for WT and T30V TRAP. Apparent association constants (Ka-obs) for WT TRAP binding to (UAGUU)11 (B) and (UAGCC)11 RNA (X) as well as for T30V TRAP binding to (UAGUU)11 (A) and (UAGCC)11 RNA (W) were obtained by filter binding at temperatures between 25 8C and 65 8C in filter binding buffer with 250 mM potassium glutamate. Each Ka-obs value represents the average of four independent experiments with standard deviations less than 50% of the mean.

interaction. Increasing the temperature enhances the affinity of structured RNAs for WT TRAP, whereas changing temperature has little effect on the affinity of unstructured RNAs for TRAP.4,35 We compared the effect of temperature on T30V and WT TRAP binding to RNAs containing 11 UAG repeats separated by either CC or UU dinucleotide spacers (Figure 5). As seen previously,4 increasing the temperature from 25 8C to 65 8C increased the

RNA-binding properties of T30V and WT TRAP Our studies, as well as those of Yakhnin et al.28 have shown that T30V TRAP binds trp leader RNA in the absence of tryptophan with approximately 30-fold lower affinity than tryptophanactivated WT TRAP. To probe whether these two proteins bind RNA by similar mechanisms, we compared the properties of complex formation for both proteins. RNA structure Previous studies have shown that the presence of secondary structure in RNA inhibits binding to TRAP,3,35 and that RNAs with CC dinucleotide spacers between the G/UAG repeats form stable structure whereas those with UU spacers do not.4 Since both WT TRAP35 and T30V TRAP (data not shown) are thermostable, we can assess the influence of RNA structure on binding to TRAP by examining the effect of temperature on this

Figure 6. Ionic strength dependence of RNA binding to WT and T30V TRAP. Apparent association constants Ka-obs of WT (O) and T30V (B) TRAP binding to (GAGUU)11 RNA in 16 mM Hepes (pH 8.0) buffer containing different concentrations of potassium glutamate were determined by filter binding at 37 8C. Each Ka-obs value represents the average of four independent experiments with standard deviations less than 30% of the mean.

713

The T30V Mutant TRAP Protein

Phosphorothioate substitutions

Table 2. Energetic contribution of ionic and non-ionic interactions in TRAP– RNA interactions at 250 mM potassium glutamate Unit: kJ/(mol K) WT T30V

DGtotal

DGnon-ionic

DGionic

251.4 (100) 242.8 (100)

249.3 (96) 234.5 (81)

22.1 (4) 28.3 (19)

The observation that ionic interactions play a greater role in the T30V TRAP –RNA complex suggests that there may be different interactions between this protein and the RNA phosphate backbone as compared to WT TRAP. To investigate this possibility, we compared the effects of phosphorothioate substitution in (GAGUU)11 RNA on complex formation with WT and T30V TRAP. In nucleoside phosphorothioates (NTPaS) one of the two non-bridging phosphate oxygen atoms is replaced with sulfur.41 Phosphorothioates alter the electronic density around the phosphate group and in some cases such substitutions alter interactions between proteins and RNA.42 – 46 Substituting phosphorothioate analogs for the G residues in the first and third positions or for the UU spacer residues in all 11 repeats of (GAGUU)11 RNA had little effect on binding to WT TRAP (Table 3). Replacing the 11 A residues in (GAGUU)11 with phosphorothioAs resulted in an 11-fold decrease in the affinity of (GaSAGUU)11 for WT TRAP, which is somewhat surprising, since the crystal structure of the TRAP – RNA complex does not indicate any direct interactions between this phosphate and the protein.23 Phosphorothioate substitutions had similar effects on the interaction of RNA with T30V TRAP, although the effects were greater, particularly for phosphorothioA (42-fold; see Table 3). Since there are 22 phosphorothioG substitutions in each modified RNA, DDG for each phosphorothioG substitution is only 0.12 and 0.22 kJ/(mol K) for WT and T30V TRAP, respectively, whereas each phosphorothioA substitution results in DDG of 0.6 kJ/(mol K) for WT TRAP and of 0.9 kJ/(mol K) for T30V TRAP.

The following equations were used to derive values of free energy from data in Figure 6: DGtotal ¼ 2RT ln Ka-obs ¼ DGnon-ionic þ DGionic ¼ DGnon-ionic þ Dr ln½Kglutamate in which Dr is the slope of dln Ka-obs =dln½Kglutamate:40 This slope reflects the ratios of the DG values. The values in parentheses are percentages.

affinity of WT TRAP for the structured RNA (UAGCC)11 but had little effect on the interaction with (UAGUU)11. Changing temperature had similar effects on the interaction of these RNAs with T30V TRAP in the absence of tryptophan, although with lower affinity for both RNAs as compared to WT TRAP. These results show that RNA structure also inhibits binding to T30V TRAP to a similar degree as for WT TRAP.

Ionic strength There are no direct interactions with phosphate in the TRAP –RNA complex,23 and the stability of the complex shows little dependence on ionic strength (dlog Kapp/d 2 log[Kglutamate] < 0.6)35 (Figure 6). The stability of the T30V TRAP – RNA complex is approximately four times more dependent on ionic strength (dlog Kapp/d 2 log[Kglutamate] < 2.3) than the complex with the WT protein. Using the method described by Record,40 we estimated the contributions of ionic and non-ionic interactions in complex formation between (GAGUU)11 and both proteins (Table 2). As seen previously,35 non-ionic interactions dominate the WT TRAP complex (96% of DGtotal at 250 mM Kglutamate). In contrast, at 250 mM Kglutamate, ionic interactions contribute nearly 20% of DGtotal to the T30V TRAP –RNA complex.

Deoxyribonucleoside substitutions Consistent with the crystal structure of the TRAP –RNA complex,23 deoxyribonucleotide substitution studies have shown that a 20 -hydroxyl (20 -OH) group is only required on the sugar of the third G (G3) of each G/UAG repeat for highaffinity binding to TRAP; the remaining residues can be replaced with deoxyribonucleosides.32 Hence the DNA/RNA chimera (taGcc)11

Table 3. Effect of phosphorothioate substitutions of (GAGUU)11 RNA on TRAP –RNA interaction RNAa (GAGUU)11 (asGAasG UU)11 (GasAGUU)11 (GAGasUasU)11

WT Kd (nM)

Fold increaseb

DDG kJ/(mol K)

T30V Kd (nM)

Fold increase

DDG kJ/(mol K)

0.6 1.7 6.8 1

1 2.7 11 1.7

0 2.7 6.3 1.3

25.5 183 1078 30

1 7 42 1

0 5 9.6 0.4

a Nucleoside phosphorothiates (NTPaS) were incorporated into (GAGUU)11 RNA by in vitro transcription using T7 RNA polymerase. In each substitution, one nucleotide was completely replaced by the corresponding NTPaS. The phosphorothioates in each sequence are printed as bold and italic letters in the Table. Apparent Kd values were determined by filter binding assay. Each point represents the average of four independent experiments with standard deviations less than 20% of the mean. b The fold increase in the Kd was determined by dividing the dissociation constant found in the presence of the analog by that obtained with the unmodified RNA.

714

The T30V Mutant TRAP Protein

Table 4. Effect of base modifications of (GAGUU)11 RNA on TRAP – RNA interaction RNA ðGAGUUÞ11 ðUAGCCÞ11 ðtaGccÞ11 b ðtagccÞ11 ðtaG20 -Ome ccÞ11 d ðtiGccÞ11 pe ðUAICCÞ11 e ð2aG – Þ11f

WT Kd (nM)a

T30V Kd (nM)

1 3.5 0.7 NBc NB NB NB 5

30 56 235 NB NB NB NB 250

a Apparent Kd values were determined by direct or competition ( p ) filter binding assays at 55 8C in 250 mM potassium glutamate and 16 mM Hepes (pH 8.0).32 Each value represents the average of at least three experiments with standard deviations less than 20% of the mean. b Upper case letters represent ribonucleotides and lower case letters stand for dexoyribonucleotides. Triplet repeat recognition sequences are underlined. c NB, no detectable RNA binding with up to 3 mM TRAP. d Guanosine was substituted by 20 -O-methyl riboguanosine (G2 -Ome). e Adenosine or guanosine was substituted by riboinosine (I) or deoxyriboinosine (i). f Each “– ” indicates an abasic sugar phosphate. 0

(deoxynucleosides are represented by lower case letters (g, a, t, c) and ribonucleotides by upper case letters (G, A, U, C)), with deoxyribonucleotides in all positions except G3 of each repeat, binds WT TRAP with similar affinity and specificity as the corresponding RNA (UAGCC)11, whereas deoxyriboG substitutions in the third residue of all 11 repeats eliminated measurable binding to TRAP.32 To probe whether T30V TRAP makes similar contacts to the RNA backbone, we examined the effects of deoxyribonucleotide substitutions on RNA-binding to the mutant protein. T30V TRAP binds (taGcc)11 with only fourfold lower affinity than for (UAGCC)11 RNA (Table 4). Similar to WT TRAP,32 deoxyribonucleotide (or 20 O-methyl) substitutions at G3 eliminated measurable binding to T30V TRAP (Table 4). These results suggest that T30V, like WT TRAP, also recognizes the 20 -OH on the third G residue in each repeat. It is also possible that T30V TRAP also recognizes additional 20 -OH groups in the RNA, since we did observe that the mutant protein has lower affinity for (taGcc)11; however, if so, these interactions make a relatively minor contribution to the overall stability of the complex (DDG ¼ 3.7 kJ/(mol K).

Inosine substitution Nucleoside analog studies identified several functional groups on the RNA that contribute to the stability of the complex with WT TRAP, including the 6-amino group and N1 nitrogen of A2, as well as the 2-amino group, O6 and N1 nitrogen from G3 from each G/UAG repeat.32 The crystal structure of TRAP bound to (GAGAU)11 RNA further showed that each of these functional groups form hydrogen bonds with the protein.23

Comparing the effects of nucleoside analog substitutions on to T30V with WT TRAP provides a powerful tool to assess whether the mutant protein recognizes similar features of the RNA as the WT protein. Inosine lacks the 2-amino group of guanosine, and differs from adenosine by changing the 6-amino group to an oxygen as well as by altering the protonation state of N1. Substituting inosine for either the second A, or for the third G in each triplet repeat eliminated measurable binding of the resulting RNA or D/RNA chimera to either WT or T30V TRAP (Table 4). These findings further indicate that both proteins recognize similar functional groups on the RNA. Abasic nucleotide substitution Our previous studies indicated that the majority of the specificity and stability of the TRAP – RNA complex comes from interactions between the protein and the AG dinucleotides of the G/UAG repeats. This conclusion was based on the finding that a D/RNA chimera consisting of 11 deoxyAriboG repeats separated by three abasic sugar – phosphate residues, (2 aG – )11, bound TRAP with only fivefold lower affinity than (GAGUU)11 RNA32 (Table 4). Similarly, we found that T30V TRAP also binds (2 aG – )11 with only eightfold lower affinity than for (GAGUU)11 indicating that the majority of the interactions between T30V TRAP and the RNA are also to the AG dinucleotides of each triplet repeat. Properties of WT-T30V hetero-11-mers Upon binding 11 tryptophan molecules, TRAP is activated to bind RNAs with up to 11 triplet repeats. Studies of TRAP homo-11-mers provide information about the overall reaction but reveal little regarding the binding mechanism or the interactions between individual subunits in these binding reactions. One way to overcome this limitation is to introduce defective subunits into the TRAP 11-mer to generate hetero-11-mers and study how their binding to tryptophan and RNA are affected. We have previously characterized such hetero-11-mers composed of WT subunits mixed with either tryptophan-binding defective subunits22 or RNA-binding defective mutant subunits.25 To examine the interaction between subunits in the T30V –RNA interaction, we created heteromeric TRAP 11-mers composed of various ratios of fully active subunits and T30V subunits.25 As the source of active subunits, we used a TRAP protein in which Lys71 is replaced with Ala (K71A). Lys71 is located far from the tryptophan and RNA-binding sites,1,23 and K71A TRAP is fully functional both in vivo36 and in vitro.25 Moreover, hetero-11-mers composed of various ratios of K71A and WT subunits all have nearly identical tryptophan and RNA-binding activities with WT homo-11-mers.25 The difference in surface charge

The T30V Mutant TRAP Protein

Figure 7. Native polyacrylamide gel electrophoresis of WT-T30V hetero-11-mers. For lanes 1 and 2, untreated WT or T30V TRAP were used. For lanes 3 through 8, WT and T30V TRAP were denatured in 6.0 M GdnHCl for 30 minutes, mixed in various ratios, and dialyzed into 50 mM Na2HPO4 (pH 8.0) overnight. Samples were electrophoresed on a native 9% polyacrylamide gel. The gel was stained with Coomassie brilliant blue.

between K71A and T30V subunits alters the mobility of hetero-11-mers containing various subunit compositions on native polyacrylamide gels, which allows us to follow subunit mixing (Figure 7). For simplicity, we will refer to K71A subunits as WT here. We constructed several pools of heteromeric mixtures with different ratios of WT and T30V TRAP subunits. Neither substitution (K71A or T30V) alters residues near the subunit interface1 and all the resulting TRAP pools have similar secondary structure to WT TRAP based on circular dichroism (CD) spectroscopy (data not shown). When displayed on native polyacrylamide gels, these pools show a total of 12 different bands (Figure 7) with intensities correlating to a binomial distribution of the input ratio of the two types of subunits (data not shown). Hence, we suggest that the two types of subunits assemble randomly into 11-mers and that these 12 bands correspond to the two homo-11-mers (WT and T30V), and ten hetero11-mers with different numbers of each type of subunit (10WT –1T30V, 9WT – 2T30V, etc.). Similar patterns have been observed previously when K71A or WT TRAP subunits are mixed with several other mutant subunits.22,25 Tryptophan binding to WT-T30V hetero-11-mers Previous studies have shown that WT (and K71A) TRAP binds tryptophan with an S0:5 of approximately 10 mM and a Hill coefficient (n) of 1.5,22,36 whereas T30V TRAP does not bind tryptophan (Figure 2(b)).28 To examine tryptophan binding to WT-T30V heteromers, we measured the decrease in ANS fluorescence as a function of tryptophan concentration for heteromer pools composed of different ratios of WT and T30V subunits. As seen in Figure 3, tryptophan competes with ANS for binding to WT TRAP and the presence of excess tryptophan completely eliminates ANS fluorescence. In contrast, tryptophan does not compete with ANS for binding to T30V TRAP and has

715

Figure 8. Maximal decrease in ANS fluorescence ðDFÞ for WT-T30V TRAP heteromer pools. The maximal decrease in ANS fluorescence for WT and T30V TRAP and for WT-T30V TRAP heteromer pools in the presence of excess tryptophan (1 mM) is plotted as a function of the fraction of WT subunits in the pool. The remaining ANS fluorescence that cannot be competed with tryptophan indicates the fraction of tryptophan binding sites that are occupied by ANS but not tryptophan.

no effect on ANS fluorescence. For each WT-T30V heteromer pool, ANS fluorescence decreased as the concentration of tryptophan was raised, indicating that all the heteromeric mixtures bind tryptophan. We obtained a binding curve for each pool similar to that seen for WT TRAP in Figure 3(b), except that the maximal decrease in ANS fluorescence (DFluorescence%) obtained for the mixtures increased proportionally with the fraction of WT subunits in the pool (Figure 8). These results suggest that in the hetero-11-mers only the tryptophan binding sites containing Thr at position 30, contributed from the WT subunits, bind tryptophan, whereas those with Val at residue 30 bind ANS but not tryptophan. The overall apparent affinity (S0:5ðtotalÞ ) of the heteromeric pools for tryptophan decreased from 16 mM for WT homo-11-mer to 55 mM for the pool containing 20% WT subunits and 80% T30V subunits (Table 5). However, if we accept that only WT binding sites in the heteromers bind tryptophan, as indicated above, then the apparent affinity of these sites for tryptophan (S0:5ðWTÞ ) varies little with the composition of the mixtures (Table 5) and is nearly the same as seen with the WT homo-11mer. As the fraction of T30V subunits was increased in the heteromer pools, the positive cooperativity for tryptophan binding was disrupted and the observed Hill coefficient decreased from 1.2 for WT homo-11-mer to 0.6 for the heteromer pool consisting of 20%WT and 80%T30V subunits (Table 5). RNA-binding to the WT-T30V hetero-11-mers We can switch the WT subunits in the hetero-11mers from the non-liganded to the activated state

716

The T30V Mutant TRAP Protein

Table 5. Tryptophan binding of WT-T30V heteromeric mixtures WT%:T30V% Maximal DF%a S0:5ðtotalÞ c (mM) S0:5ðWTÞ c (mM) nd

100:0 100 16 16 1.2

80:20 80 16 13 1.1

60:40 61 22 13 1.0

50:50 56 25 12 0.9

40:60 44 44 18 0.7

20:80 25 55 11 0.6

0:100 NBb NB NB

a

Tryptophan binding was measured as the decrease of ANS fluorescence (DF) (see Figure 3). The maximal change in fluorescence for WT TRAP binding to tryptophan was set to 100% and the others values were normalized accordingly. b NB, no detectable binding up to 1 mM tryptophan. c Data obtained from ANS fluorescence assay using 5 mM TRAP were fit to the Hill equation. Each value represents as least two independent experiments with standard error less than 20% of the mean. S0:5ðtotalÞ is the apparent affinity for all TRAP 11-mers in the populations, and S0:5ðWTÞ is for only active tryptophan binding sites as determined by dividing the S0:5ðtotalÞ value by the fraction of WT subunits. d n is the Hill coefficient.

by adding tryptophan. This allows us to study how either form of the WT subunits interacts with T30V subunits with regard to RNA binding. Tryptophanactivated TRAP binds RNA mainly through interactions between the RNA bases and the protein23,32 and hence the stability of the complex shows little dependence on ionic strength (Figure 6).31 In the absence of tryptophan, WT (and K71A) TRAP binds RNA very weakly ðKd q 5:0 mMÞ: T30V TRAP binds RNA with moderate affinity ðKd < 30 nMÞ in the absence or presence of tryptophan and the stability of the T30V TRAP – RNA complex is approximately four times more dependent on ionic strength than tryptophan-activated WT TRAP (Figure 6). This difference in ionic strength dependence is a convenient tool to examine the contribution of each type of subunit to the affinity of WT-T30V heteromers for RNA under various conditions. Tryptophan-independent RNA-binding At low salt (250 mM potassium glutamate), the observed affinity of the WT-T30V hetero-11-mer pools for (GAGUU)11 RNA gradually decreased from 35 nM for T30V homo-11-mer to approximately 1500 nM for the pool containing 20% T30V subunits and 80% WT subunits (Table 6). WT apoTRAP binds this RNA very weakly under these conditions. Increasing the potassium glutamate

concentration to 1.0 M eliminated measurable RNA-binding to apo-TRAP and reduced the affinity of the T30V TRAP for (GAGUU)11 RNA by 12-fold. Similarly, the affinity of all WT-T30V heteromer pools for RNA was also decreased at the higher salt concentration, and in this case, RNA-binding was barely detectable for the pools with over 50% WT subunits. These data suggest that in the absence of tryptophan, RNA-binding of WT-T30V heteromers is largely determined by T30V subunits and the WT subunits contribute little to the stability of the complex. Tryptophan-dependent RNA binding In the presence of excess (1.0 mM) tryptophan, WT TRAP binds to (GAGUU)11 RNA with an apparent Kd of 1.4 nM (Table 6). The affinity of T30V TRAP for this RNA (Kd of 37 nM) is not significantly changed by the presence of tryptophan. Under these conditions, the affinities of all the WT-T30V heteromeric pools for (GAGUU)11 are similar to that of WT TRAP, although the observed Kd decreased from 1.4 nM to 8.8 nM as the percentage of WT subunits decreased from 100% to 20% in the pools (Table 6). These data are consistent with several previous studies, which showed that as few as one or two activated WT subunits within hetero-11-mers are sufficient to significantly stabilize the TRAP – RNA complex.22,25

Table 6. (GAGUU)11 RNA binding to WT-T30V heteromeric mixtures WT%:T30V% KGlu (mM)

100:0

80:20

60:40

40:60

20:80

0:100

Kd (nM)a

RNA binding in the absence of tryptophan 250 .5000 1000 NBb

,1500 .5000

687 ^ 138 .3000

475 ^ 120 ,1100

122 ^ 43 605 ^ 61

35 ^ 6 439 ^ 38

RNA binding in the presence of tryptophan 250 1.4 ^ 0.4 1000 5.3 ^ 1.8

0.6 ^ 0.4 11 ^ 3.8

4.2 ^ 0.6 17 ^ 5.7

4.8 ^ 1.0 29 ^ 3.1

8.8 ^ 1.2 46 ^ 4.9

37 ^ 9 569 ^ 104

a Apparent equilibrium dissociation constants (Kd ) for (GAGUU)11 RNA were determined by filter binding in 16 mM Hepes (pH 8.0) and 250 mM or 1000 mM potassium glutamate (KGlu) at 37 8C. Kd values represent the average ^ the standard deviation of at least four independent experiments. b No binding was detected with up to 5 mM TRAP.

717

The T30V Mutant TRAP Protein

To study the contribution of the T30V subunits in RNA-binding to heteromers containing tryptophan-activated WT subunits, we examined the ionic strength dependence of the affinity of several WT-T30V hetero-11-mer pools for (GAGUU)11 (Table 6). When the concentration of potassium glutamate was increased from 250 mM to 1.0 M, the affinity of WT TRAP decreased only twofold, whereas that of T30V TRAP decreased 18-fold. The affinities of all the heteromeric mixtures are closer to that of WT TRAP than to that of T30V TRAP under these conditions. However, at higher ionic strength, the affinities of heteromeric mixtures with higher percentages of T30V subunits have slightly lower affinity for RNA.

Discussion Relatively little is known about the mechanism by which tryptophan binding activates TRAP to bind RNA. Previous studies,28 as well as those presented here show that the T30V mutant TRAP binds RNA in the absence of tryptophan, albeit with < 30-fold lower affinity than the tryptophanactivated WT protein. Our results show that in general, T30V TRAP recognizes the same features of the (G/U)AG repeats as WT TRAP. The major differences between the complexes of T30V and WT TRAP with RNA are that the T30V complex is more dependent on ionic strength (Figure 6), as well as more sensitive to phosphorothioate substitution of the second A residue in each triplet repeat. Together these observations suggest that T30V TRAP forms a similar but not identical complex with RNA as WT TRAP. By studying how this single substitution of threonine to valine, which changes one hydroxyl group to a methyl group, changes properties of the protein, we hope to further our understanding of how tryptophan binding activates TRAP. Tryptophan binding and activation of TRAP to bind RNA Each tryptophan binds between adjacent TRAP subunits and interacts with amino acids on two loops, one from each subunit, formed by residues 25– 33 (BC) and 49 – 52 (DE)1 (Figure 1). Two types of interactions contribute to tryptophan binding to TRAP: hydrophobic interactions between the indole ring and aliphatic residues on the protein, and an array of hydrogen bonds formed between the amino and carboxyl groups, as well as the indole nitrogen atom of the ligand, and polar residues on the protein. In the TRAP –tryptophan complex, the conformations of the BC and DE loops appear to be maintained by interactions with the bound tryptophan because they do not interact with the rest of the protein. Moreover, the bound tryptophan residues are nearly completely buried within the protein, which implies that this

region of the protein must have a different conformation prior to binding tryptophan.1 The RNA-binding site on TRAP is distinct from and distant to the tryptophan-binding sites (Figure 1). Hence tryptophan binding must affect the protein beyond the residues immediately involved in its binding. Several of the key residues that contact the bound RNA, including Glu36, Lys37, Lys56 and Arg58, are located on b-strands C and E, which are directly connected to the two aforementioned loops in the tryptophan binding site23 (Figure 1B). Therefore, tryptophan binding between the BC and DE loops likely alters the conformation and/or dynamics of TRAP, particularly the b-strands connected to these loops. Consistent with this proposal, recent NMR studies indicate that tryptophan binding reduces the flexibility and dynamics of both loops involved in tryptophan binding as well as many residues on the b-strands attached to these loops.27 The T30V mutant TRAP WT TRAP with threonine at position 30 requires tryptophan binding in order to bind RNA. The crystal structure shows that the amino group of the bound tryptophan forms four hydrogen bonds with the protein including three to main-chain atoms of Gly27, Asp29 and Thr25, and one with the side-chain of Thr30.1 Changing Thr30 to alanine renders the protein inactive for binding tryptophan or RNA,28 even though only one hydrogen bond is disrupted. Substituting valine at position 30 also disrupts this hydrogen bond but in this case partially activates the protein to bind RNA. The differences between threonine, alanine and valine must be responsible for the different properties of these three proteins. The valine sidechain is larger than that of threonine (and alanine) and this substitution also increases the hydrophobicity of the tryptophan-binding pocket. Consequently the T30V mutant binding pocket does not accommodate tryptophan but instead binds several tryptophan analogs that lack the amino group, such as bIAA, that do not bind to WT TRAP. However, the interaction of T30V TRAP with RNA is not affected by these tryptophan analogs (data not shown). These findings show that the tryptophan binding pocket of T30V TRAP is open enough to accommodate bIAA and therefore the mechanism by which the T30V substitution partially activates TRAP is not to lock the BC and DE loops in the same “closed” conformation seen in the TRAP – tryptophan complex.23 Based on the observations described above, we suggest that tryptophan binding activates TRAP to bind RNA, at least in part, by reducing the dynamic motion of the protein.27 Hence the T30V substitution may partially activate TRAP by reducing the motion of the BC and/or DE loops and one or more of the attached b-strands. Thr30 is located on the BC loop, and the associated b-strand C contains two of the key residues, Glu36 and Lys37,

718 involved in RNA-binding23,36 (Figure 1); hence changing Thr30 to valine might affect this loop and strand. However, preliminary NMR spectra of T30V are more similar to that of apo-TRAP than to the TRAP – tryptophan complex, suggesting that the mutation does not induce a change in protein dynamics of the same magnitude as tryptophan binding to WT TRAP (C. McElroy, M. Foster & P.G., unpublished observations). Likewise preliminary studies show that T30V TRAP, like WT apo-TRAP, is susceptible to trypsin cleavage at residue 31 (V. Payal & P.G., unpublished observations) whereas the TRAP –tryptophan complex is resistant to cleavage.27 Together these observations suggest that although T30V TRAP forms a similar complex with RNA as tryptophanactivated WT TRAP, the mechanism of activation may be different for the two proteins, possibly related to the dynamic properties of both proteins. Given the greater contribution of ionic interactions to the RNA complex with T30V TRAP, it is also possible that solvation and solvent release are different for RNA complexes with WT and T30V TRAP. Preliminary studies indicate that slightly fewer water molecules are released when T30V TRAP binds RNA than when tryptophan-activated WT TRAP binds RNA (our unpublished observations). The interaction of T30V TRAP and RNA The observation that the affinity of T30V TRAP for RNA is more dependent on ionic strength than is the WT TRAP – RNA complex indicates that the mutant protein relies more on ionic interactions with the RNA than does WT TRAP. The DGnon-ionic, which reflects energetic contributions of hydrogen bonds and hydrophobic interactions to the stability of TRAP – RNA complex, is < 15 kJ(mol K)21 lower for the T30V complex than for the WT TRAP – RNA complex (Table 2). This decrease in DGnon-ionic could result from either fewer non-ionic interactions in the T30V RNA complex or from having the same number but weaker non-ionic interactions. There are no direct interactions with phosphate seen in several complexes of WT TRAP with RNA;23,24 however, phosphorothioA substitutions in the second position of all 11 repeats in (GAGUU)11 RNA decrease the affinity of the substituted RNA for WT TRAP by 11-fold. While the explanation for this effect is not obvious, it may be due to a subtle difference in the RNA backbone conformation or flexibility rather than due to a direct effect on an interaction with this phosphate. In the crystal structure of TRAP complexed with (GAGAU)11 RNA there is a sharp twist in the RNA backbone between G1 and A2,23 so perhaps a phosphorothioate at A2 is less able to adopt this conformation. The same phosphorothioA substitutions have a greater impact on T30V TRAP (42-fold), suggesting that T30V TRAP forms a different interaction with the RNA backbone. It is possible that the RNA in these complexes may

The T30V Mutant TRAP Protein

adopt slightly different conformations, which consequently change the electrostatic interactions with the protein. Together these data suggest that although T30V TRAP has similar specificity in RNA recognition to WT TRAP, the mutant protein forms a somewhat different complex with the RNA. This difference could originate from the absence of a particular interaction between the T30V protein and RNA. Alternatively, the T30V complex may involve all of the same interactions seen in the WT complex but, due to alterations in the angle or length of one or more of the bonds, the stability of the T30V complex is reduced. The disparity may also relate to differences in the dynamics of the two proteins. WT-T30V TRAP heteromers Our studies of TRAP hetero-11-mers containing mixtures of WT and T30V subunits indicate that the tryptophan binding sites containing Thr30 (contributed from WT subunits) bind tryptophan with similar affinity to that seen with the WT homo-11-mer, and that this binding activates the associated RNA-binding sites (Table 5). These studies also suggest that the tryptophan binding sites containing Val30 (from the T30V subunits) do not bind tryptophan, and that their associated RNA-binding sites are not affected by the WT subunits in the heteromers. Previously, we saw similar results with hetero-11-mers composed of WT and T25A subunits (defective in tryptophan and RNAbinding).22 In contrast, we did observe that the presence of the T30V subunits affected cooperativity of tryptophan binding to the WT subunits (Table 5). The Hill coefficient dropped from 1.2 for WT TRAP, which indicates a low degree of positive cooperativity, to 0.6 for the heteromer pool containing 20% WT subunits. The n value of 0.6 for the 20%WT– 80%T30V mixture suggests either that T30V subunits negatively influence tryptophan binding to WT subunits or that there are multiple types of binding sites on the heteromers with different affinities for tryptophan. We have previously studied TRAP hetero-11mers consisting of WT subunits and mutant subunits defective either in RNA-binding,25 or tryptophan binding.22 In all cases, hetero-11-mer pools with very low percentages of WT subunits bind RNA with similar affinity to WT TRAP. We have interpreted these findings in terms of our two-step model for RNA-binding to TRAP involving formation of an initiation complex followed by cooperative wrapping of the RNA around the protein.47 We suggest that at least one tryptophan-activated WT subunit is required to form the initiation complex. Once this complex is formed, mutant subunits that are defective in an amino acid that makes one hydrogen bond with the bound RNA (K37A, K56A and R58A) or are defective in tryptophan binding (T25A) are adequate for the second binding step of wrapping the tethered RNA around the protein ring.

719

The T30V Mutant TRAP Protein

By examining the RNA-binding properties of WT-T30V hetero-11-mers in the absence and presence of tryptophan, we can assess the ability of T30V subunits to function in either the initiation or the second step of RNA-binding, respectively. In the presence of tryptophan the activated WT subunits can function for initiation. Hence, by examining heteromers containing a high fraction of T30V subunits we can test the ability of the mutant subunits to function in the second step of wrapping the RNA around TRAP. We found that in the presence of tryptophan and 250 mM potassium glutamate, all WT-T30V hetero-11-mer pools, even that with only 20% WT subunits, bound RNA with virtually the same affinity as WT TRAP (Table 6). These observations suggest that T30V subunits function well in the second step of RNA-binding to TRAP. Increasing the potassium glutamate concentration from 250 mM to 1000 mM had little effect on the affinity of tryptophan-activated WT TRAP for RNA, but reduced the affinity of T30V TRAP for (GAGUU)11 by over 18-fold (Table 6). Therefore, examining the effect of ionic strength on the RNAbinding properties of WT-T30V heteromers in the presence of tryptophan, we can assess the contribution of each type of subunit in the complex. Increasing the ionic strength decreased the affinity of all of the WT-T30V heteromer pools for RNA, suggesting that the T30V subunits do contribute to the stability of the complexes. Moreover, at high salt (1000 mM Kglutamate), the Kd of the heteromer pools increase from 11 nM for the pool with 20% T30V subunits to 46 nM for the pool with 80% mutant subunits. Together these results suggest that when (GAGUU)11 RNA binds to a hetero-11mer composed of both WT and T30V subunits, the RNA interacts differently with each type of subunit such that the interaction with RNA-binding sites associated with T30V subunits is sensitive to ionic strength, whereas the interaction with sites associated with WT subunits is not. Recent crystallographic studies of TRAP complexed with either (UAGAU)11 or (GAGUUU)11 RNAs have shown that the nucleotides in the first position as well as the spacer residues between the triplet repeats can adopt different configurations (A. Antson & P.G., unpublished observations).24 Hence it is possible that when (GAGUU)11 binds to the heteromers, the RNA can adjust its conformation to interact differently with the WT and the T30V subunits. Studies of the WT-T30V hetero-11-mers binding RNA in the absence of tryptophan suggest that the T30V subunits are far less effective than tryptophan-activated WT subunits in forming the initiation complex with RNA. Pools composed of 20– 40% T30V subunits bind RNA weakly (Kd 680– . 1500 nM) and even the pool with 60% T30V subunits displays an apparent Kd of 475 nM, which is over 13-fold weaker than the T30V 11-mer (Table 6). Since our studies of WT-T25A hetero-11-mers indicated that non-liganded TRAP subunits function relatively well in the second

step of binding,22 the results with WT-T30V hetero11-mers in the absence of tryptophan suggest that the T30V subunits do not function well to initiate binding. T30V TRAP does not induce attenuation of the trpoperon in vivo (V. Payal, unpublished observations).28 This could be due simply to the lower affinity of T30V TRAP for trp leader RNA. Alternatively the reduced ability of T30V TRAP to form the initiation complex may specifically slow the rate of TRAP binding to trp leader RNA, which would interfere with attenuation control of transcription. Studies of the kinetics of WT and T30V TRAP may shed light on this issue.

Materials and Methods Purification of TRAP Wild-type, K71A and T30V TRAP were overexpressed in E. coli SG62052 (pGP1-2).29 WT TRAP was purified by ammonium sulfate precipitation followed by phenyl agarose chromatography as described.30 Purification of K71A and T30V TRAP was similar to that of the WT protein except that ammonium sulfate was added to 30% or 40% saturation, respectively. Each protein preparation was examined by SDS-PAGE and was quantified by UV absorbance (extinction coefficient of 1280 M21 cm21 at 280 nm), and confirmed by BCA protein assay (Pierce) using WT TRAP protein standards. Chemicals Tryptophan analogs were purchased from Sigma or Aldrich and were dissolved in either water or methanol to a final concentration of 2 - 20 mM. ANS was purchased from Molecular Probes and was dissolved in water. Circular dichroism CD spectra of TRAP (0.5 mg/ml) in 25 mM sodium phosphate buffer (pH 8.0) were obtained in 1-mm quartz cuvettes with a JASCO model J-700 spectropolarimeter at room temperature or at 37 8C.22 Spectra of TRAP with tryptophan were derived by subtracting the spectra of the same concentration of free tryptophan from the overall spectra. Fluorescence Fluorescence of protein samples in the presence of 1 mM or 10 mM ANS was measured with an SLMAMINCO 8000 spectrofluorometer at room temperature, with excitation at 372 nm and emission from 400– 600 nm. Both excitation and emission had 16 nm spectra band pass filters and a 420 nm long pass filter was also used for emission. RNA synthesis and filter binding assay RNAs were prepared by in vitro transcription from linearized plasmid DNA templates containing a T7 promoter as described.31 Reactions were incubated for 20 minutes at 37 8C and included [a-32P]UTP or

720

The T30V Mutant TRAP Protein

[a-32P]ATP (DuPont NEN). The RNA products were purified on 8 M urea/10% (w/v) polyacrylamide gels. For in vitro transcription reactions employed to prepare inosine (I)-substituted (UAICC)11 RNA, GTP was replaced with 2 mM ITP and 0.12 mM ApG (Sigma) was included to initiate transcription. These reactions were incubated for two hours at room temperature. Phosphorothioate substituted RNAs were also prepared by in vitro transcription with T7 polymerase, with one of the ribonucleotide triphosphates replaced by 0.2 mM of the corresponding phosphorothioate analog (Amersham). RNAs modified by deoxyribonucleotides or abasic nucleoside analogs were chemically synthesized at the CAMBI nucleic acid facility (University of Buffalo, State University of New York) as described.32 These modified RNAs were 50 -end labeled with [g-32P]ATP (DuPont NEN) using T4 polynucleotide kinase. Filter binding assays were performed as described previously, in filter binding buffer (250 mM potassium glutamate, 16 mM Hepes, pH 8.0).31 To study the ionic strength dependence of TRAP– RNA interactions, the concentration of potassium glutamate was varied between 250 mM and 1.0 M. Tryptophan binding Tryptophan binding to TRAP was measured by fluorescence spectroscopy using ANS. Background from free tryptophan was subtracted from all fluorescence intensities (F). Values for the decrease in ANS fluorescence ðDFÞ were normalized to the maximal spectral signal change at saturation of tryptophan binding to yield DF%: DF% values were fit to the Hill equation:22,33 DF% ¼ að½Trp=S0:5 Þn Þ=ð1 þ ð½Trp=S0:5 Þn Þ where a is the saturation level of bound tryptophan; S0.5 represents the binding affinity and is defined as the concentration of free tryptophan at 50% of binding saturation. The Hill coefficient, n; describes cooperativity: n is 1 for no cooperativity, . 1 for positive cooperativity and , 1 for negative cooperativity. Since the affinity of ANS for TRAP is much lower than that of tryptophan (see Results), the effects of ANS binding to TRAP was not considered in our calculations. Subunit mixing Heteromeric TRAP 11-mers were generated as described.22,25 Various ratios of T30V mutant and WT TRAP proteins were mixed together in $ 4:0 M guanidine hydrochloride (GdnHCl) for at least two hours at room temperature. The mixtures were then dialyzed against 50 mM sodium phosphate buffer (pH 8.0) at 5 8C overnight. The recovered samples were quantified by BCA assay (Pierce) and electrophoresed on native 9% polyacrylamide gels using a BioRad Mini-Protean II system at 30 mA for 4.5 hours at 5 8C. Gels were stained with Coommassie brilliant blue and photographed with a Kodak DC290 digital camera. Gel images were analyzed by GeneImager software (Scanalytics Inc).

Acknowledgements The authors thank Paul Babitzke for providing the plasmid for expression of T30V mutant protein.

We thank Amanda Manfredo for outstanding technical assistance, and Vandana Payal and Xiufeng Li for sharing results prior to publication. We thank Mark Foster and Craig McElroy for sharing NMR data prior to publication, and for critical reading of the manuscript. We also thank the Pharmaceutical Sciences Instrumentation Facility at SUNY for CD data collection. This work was supported by grants GM62750 from the National Institutes of Health and MCB 9982652 from the National Science Foundation.

References 1. Antson, A. A., Otridge, J. B., Brzozowski, A. M., Dodson, E. J., Dodson, G. G., Wilson, K. S. et al. (1995). The three dimensional structure of trp RNAbinding attenuation protein. Nature, 374, 693– 700. 2. Babitzke, P., Bear, D. & Yanofsky, C. (1995). TRAP, the trp RNA-binding attenuation protein of Bacillus subtilus, is a toroid-shaped molecule that binds trancripts containing GAG or UAG repeats separated by two nucleotides. Proc. Natl Acad. Sci. USA, 92, 7916– 7920. 3. Babitzke, P., Yealy, J. & Campanelli, D. (1996). Interaction of the trp RNA-binding attenuation protein (TRAP) of Bacillus subtilis with RNA: effects of the number of GAG repeats, the nucleotides separating adjacent repeats and RNA secondary structure. J. Bacteriol. 178, 5159– 5163. 4. Xirasagar, S., Elliott, M., Bartolini, W., Gollnick, P. & Gottlieb, P. (1998). RNA structure inhibits the TRAP (trp RNA-binding attenuation protein) – RNA interaction. J. Biol. Chem. 273, 27146 –27153. 5. Babitzke, P. & Gollnick, P. (2001). Post-transcription initiation control of tryptophan metabolism in Bacillus subtilis by TRAP, anti-TRAP, and RNA structure. J. Bacteriol. 183, 5795– 5802. 6. Kuroda, M., Shimotsu, H., Henner, D. & Yanofsky, C. (1986). Regulatory elements common to the Bacillus pumilus and Bacillus subtilis trp operons. J. Bacteriol. 167, 792– 798. 7. Kuroda, M., Henner, D. & Yanofsky, C. (1988). cisacting sites in the transcript of the Bacillus subtilis trp operon regulate expression of the operon. J. Bacteriol. 170, 3080– 3088. 8. Shimotsu, H., Kuroda, M., Yanofsky, C. & Henner, D. (1986). Novel form of transcription attenuation regulates expression of the Bacillus subtilis tryptophan operon. J. Bacteriol. 166, 461– 471. 9. Gollnick, P. & Babitzke, P. (2002). Transcription attenuation. Biochim. Biophys. Acta, 1577, 240–250. 10. Merino, E., Babitzke, P. & Yanofsky, C. (1995). trp RNA-binding attenuation protein (TRAP)-trp leader RNA interactions mediate translational as well as transcriptional regulation of the Bacillus subtilis trp operon. J. Bacteriol. 177, 6362– 6370. 11. Du, H. & Babitzke, P. (1998). trp RNA-binding attenuation protein-mediated long distance RNA refolding regulates translation of trpE in Bacillus subtilis. J. Biol. Chem. 273, 20494– 20503. 12. Yakhnin, H., Babiarz, J. E., Yakhnin, A. V. & Babitzke, P. (2001). Expression of the Bacillus subtilis trpEDCFBA operon is influenced by translational coupling and Rho termination factor. J. Bacteriol. 183, 5918– 5926. 13. Yang, M., Saizieu, A., Loon, A. P. G. M. & Gollnick, P.

721

The T30V Mutant TRAP Protein

14.

15.

16. 17.

18.

19.

20.

21.

22.

23.

24.

25.

26.

27.

28.

(1995). Translation of trpG in Bacillus subtilis is regulated by the trp RNA-binding attenuation protein (TRAP). J. Bacteriol. 177, 4272– 4278. Du, H., Tarpley, R. & Babitzke, P. (1997). The trp RNA-binding attenuation protein regulates TrpG synthesis by binding to the trpG ribosome binding site of Bacillus subtilis. J. Bacteriol. 179, 2582– 2586. Sarsero, J., Merino, E. & Yanofsky, C. (2000). A Bacillus subtilis gene of previously unknown function, yhaG, is translationally regulated by tryptophan-activated TRAP and appears to be involved in tryptophan transport. J. Bacteriol. 182, 2329– 2331. Valbuzzi, A. & Yanofsky, C. (2001). Inhibition of the B. subtilis regulatory protein TRAP by the TRAPinhibitory protein, AT. Science, 293, 2057– 2059. Valbuzzi, A., Gollnick, P., Babitzke, P. & Yanofsky, C. (2002). The anti-trp RNA-binding attenuation protein (Anti-TRAP), AT, recognizes the tryptophan-activated RNA binding domain of the TRAP regulatory protein. J. Biol. Chem. 277, 10608– 10613. Sarsero, J., Merino, E. & Yanofsky, C. (2000). A Bacillus subtilis operon containing genes of unknown function senses tRNATrp charging and regulates expression of the genes of tryptophan biosynthesis. Proc. Natl Acad. Sci. USA, 97, 2656– 26661. Chen, X., Antson, A. A., Yang, M., Li, P., Baumann, C., Dodson, E. J. et al. (1999). Regulatory features of the trp operon and the crystal structure of the trp RNA-binding attenuation protein from Bacillus stearothermophilus. J. Mol. Biol. 289, 1003– 1016. Otridge, J. & Gollnick, P. (1993). MtrB from Bacillus subtilis binds specifically to trp leader RNA in a tryptophan-dependent manner. Proc. Natl Acad. Sci. USA, 90, 128– 132. Babitzke, P. & Yanofsky, C. (1995). Structural features of L -tryptophan required for activation of TRAP, the trp RNA-binding attenuation protein of Bacillus subtilis. J. Biol. Chem. 270, 12452– 12456. Li, P. & Gollnick, P. (2002). Using hetero-11-mers composed of wild-type and mutant subunits to study tryptophan binding to TRAP and its role in activating RNA binding. J. Biol. Chem. 277, 35567– 35573. Antson, A. A., Dodson, E. J., Dodson, G. G., Greaves, R. B., Chen, X. & Gollnick, P. (1999). Structure of the trp RNA-binding attenuation protein, TRAP, bound to RNA. Nature, 401, 235– 242. Hopcroft, N. H., Wendt, A. L., Gollnick, P. & Antson, A. A. (2002). Specificity of TRAP – RNA interactions: crystal structures of two complexes with different RNA sequences. Acta Crystallog. sect. D, 58, 615– 621. Li, P., Scott, D. & Gollnick, P. (2002). Creating hetero11-mers composed of wild-type and mutant subunits to study RNA binding to TRAP. J. Biol. Chem. 277, 11838– 11844. Flynn, P., Wendt, A. & Gollnick, P. (2002). Influence of induced fit in the interaction of Bacillus subtilis trp RNA-binding attenuator protein and its RNA antiterminator target oligomer. Proteins: Struct. Funct. Genet. 49, 432– 438. McElroy, C., Manfredo, A., Wendt, A., Gollnick, P. & Foster, M. (2002). TROSY-NMR studies of the 91 kDa TRAP protein reveal allosteric control of a gene regulatory protein by ligand-altered flexibility. J. Mol. Biol. 323, 463– 473. Yakhnin, A., Trimble, J., Chiaro, C. & Babitzke, P. (2000). Effects of mutations in the L -tryptophan binding pocket of the trp RNA binding attenuation

29.

30.

31.

32. 33.

34.

35.

36.

37. 38.

39.

40. 41. 42.

43.

44.

protein of Bacillus subtilis. J. Biol. Chem. 275, 4519 –4524. Tabor, S. & Richardson, C. (1985). A bacteriophage T7 RNA polymerase/promoter system for controlled exclusive expression of specific genes. Proc. Natl Acad. Sci. USA, 82, 1074– 1078. Antson, A., Brzozowski, A., Dodson, E., Dauter, Z., Wilson, K., Kurecki, T. et al. (1994). 11-fold symmetry of the trp RNA-binding attenuation protein (TRAP) from Bacillus subtilis determined by X-ray analysis. J. Mol. Biol. 244, 1 – 5. Baumann, C., Otridge, J. & Gollnick, P. (1996). Kinetic and thermodynamic analysis of the interaction between TRAP of Bacillus subtilis and trp leader RNA. J. Biol. Chem. 271, 12269– 12274. Elliott, M. B., Gottlieb, P. A. & Gollnick, P. (1999). Probing the TRAP –RNA interaction with nucleoside analogs. RNA, 5, 1277– 1289. Segel, I. H. (1975). Multisites & allosteric enzymes. In Enzyme Kinetics: Behavior and Analysis of Rapid Equilibrium and Steady-state Enzyme Systems (Segel, I. H., ed.), pp. 360– 361, Wiley, New York. Venyaminov, S. & Yang, J. (1996). Determination of protein secondary structure. In Circular Dichroism and the Conformational Analysis of Biomolecules (Fasman, G., ed.), pp. 69 – 108, Plenum Press, New York. Baumann, C., Xirasagar, S. & Gollnick, P. (1997). The trp RNA-binding attenuation protein (TRAP) from B. subtilis binds to unstacked trp leader RNA. J. Biol. Chem. 272, 19863– 19869. Yang, M., Chen, X., Militello, K., Hoffman, R., Fernandez, B., Baumann, C. & Gollnick, P. (1997). Alanine-scanning mutagenesis ob Bacillus subtilis trp RNA-binding attenuation protein (TRAP) reveals residues involved in tryptophan binding and RNA binding. J. Mol. Biol. 270, 696– 710. Lakowicz, J. (1999). Principles of Fluorescence Spectroscopy, Kluwer Academic/Plenum Publishers, New York. Chou, W., Bieber, C. & Matthews, K. (1989). Tryptophan and 8-anilino-1-naphthalenesulfonate compete for binding to trp repressor. J. Biol. Chem. 264, 8309 –8313. Reedstrom, R., Martin, K., Vangala, S., Mahoney, S., Wilker, E. & Royer, C. (1996). Characterization of charge change super-repressor mutants of trp repressor: effects on oligomerization conformation, ligation and stability. J. Mol. Biol. 264, 32 – 45. Record, M., Lohman, T. & de Haseth, P. (1976). Ion effects on ligand – nucleic acid interactions. J. Mol. Biol. 107, 145–158. Eckstein, F. & Gindl, H. (1970). Polyribonucleotides containing a phosphorothioate backbone. Eur. J. Biochem. 13, 558–564. Connolly, B., von der Haar, F. & Eckstein, F. (1980). Structure of the metal nucleotide complex in the yeast phenylalanyl transfer ribonucleic acid synthetase reaction as determined with diastereomeric phosphorothioate analogs of ATP. J. Biol. Chem. 255, 11301– 11307. Schatz, D., Leberman, R. & Eckstein, F. (1991). Interaction of Escherichia coli tRNA(Ser) with its cognate aminoacyl-tRNA synthetase as determined by footprinting with phosphorothioate-containing tRNA transcripts. Proc. Natl Acad. Sci. USA, 88, 6132 –6136. Rudinger, J., Puglisi, J., Putz, J., Schatz, D., Eckstein, F., Florentz, C. & Giege, R. (1992). Determinant nucleotides of yeast tRNA(Asp) interact directly

722

The T30V Mutant TRAP Protein

with aspartyl-tRNA synthetase. Proc. Natl Acad. Sci. USA, 89, 5882– 5886. 45. Dertinger, D., Behlen, L. & Uhlenbeck, O. (2000). Using phosphorothioate-substituted RNA to investigate the thermodynamic role of phosphates in a sequence specific RNA – protein complex. Biochemistry, 39, 55 – 63.

46. Loverix, S., Winqvist, A., Stro¨mberg, R. & Steyaert, J. (2000). Mechanism of RNase T1: concerted triesterlike phosphoryl transfer via a catalytic three-centered hydrogen bond. Chem. Biol. 7, 651– 658. 47. Elliott, M., Gottlieb, P. & Gollnick, P. (2001). The mechanism of RNA binding to TRAP: initiation and cooperative interactions. RNA, 7, 85 – 93.

Edited by D. E. Draper (Received 5 September 2003; received in revised form 27 October 2003; accepted 2 November 2003)