Sensor I Threonine of the AAA+ ATPase Transcriptional Activator PspF Is Involved in Coupling Nucleotide Triphosphate Hydrolysis to the Restructuring of σ54-RNA Polymerase*

Transcriptional initiation invariably involves the transition from a closed RNA polymerase (RNAP) promoter complex to a transcriptional competent open complex. Activators of the bacterialσ54-RNAP are AAA+ proteins that couple ATP hydrolysis to restructure the σ54-RNAP promoter complex. Structures of the σ54 activator PspF AAA+ domain (PspF1–275) bound to σ54 show two loop structures proximal to σ54 as follows: the σ54 contacting the GAFTGA loop 1 structure and loop 2 that classifies σ54 activators as pre-sensor 1 β-hairpin AAA+ proteins. We report activities for PspF1–275 mutated in the AAA+ conserved sensor I threonine/asparagine motif (PspF1–275T148A, PspF1–275N149A, and PspF1–275N149S) within the second region of homology. We show that sensor I asparagine plays a direct role in ATP hydrolysis. However, low hydrolysis rates are sufficient for functional output in vitro. In contrast, PspF1–275T148A has severe defects at the distinct step of σ54 promoter restructuring. This defect is not because of the failure of PspF1–275T148A to stably engage with the closed σ54 promoter, indicating (i) an important role in ATP hydrolysis-associated motions during energy coupling for remodeling and (ii) distinguishing PspF1–275T148A from PspF1–275 variants involved in signaling to the GAFTGA loop 1, which fail to stably engage with the promoter. Activities of loop 2 PspF1–275 variants are similar to those of PspF1–275T148A suggesting a functional signaling link between Thr148 and loop 2. In PspF1–275 this link relies on the conserved nucleotide state-dependent interaction between the Walker B residue Glu108 and Thr148. We propose that hydrolysis is relayed via Thr148 to loop 2 creating motions that provide mechanical force to the GAFTGA loop 1 that contacts σ54.

Transcriptional initiation invariably involves the transition from a closed RNA polymerase (RNAP) promoter complex to a transcriptional competent open complex. Activators of the bacterial 54 -RNAP are AAA؉ proteins that couple ATP hydrolysis to restructure the 54 -RNAP promoter complex. Structures of the 54 activator PspF AAA؉ domain (PspF 1-275 ) bound to 54 show two loop structures proximal to 54 as follows: the 54 contacting the GAFTGA loop 1 structure and loop 2 that classifies 54 activators as pre-sensor 1 ␤-hairpin AAA؉ proteins. We report activities for PspF 1-275 mutated in the AAA؉ conserved sensor I threonine/asparagine motif (PspF 1-275 T148A , PspF 1-275 N149A , and PspF 1-275 N149S ) within the second region of homology. We show that sensor I asparagine plays a direct role in ATP hydrolysis. However, low hydrolysis rates are sufficient for functional output in vitro. In contrast, PspF 1-275 T148A has severe defects at the distinct step of 54 promoter restructuring. This defect is not because of the failure of PspF 1-275 T148A to stably engage with the closed 54 promoter, indicating (i) an important role in ATP hydrolysis-associated motions during energy coupling for remodeling and (ii) distinguishing PspF 1-275 T148A from PspF 1-275 variants involved in signaling to the GAFTGA loop 1,which fail to stably engage with the promoter. Activities of loop 2 PspF 1-275 variants are similar to those of PspF 1-275 T148A suggesting a functional signaling link between Thr 148 and loop 2. In PspF 1-275 this link relies on the conserved nucleotide state-dependent interaction between the Walker B residue Glu 108 and Thr 148 . We propose that hydrolysis is relayed via Thr 148 to loop 2 creating motions that provide mechanical force to the GAFTGA loop 1 that contacts 54 .
Regulation of transcription enables cells to adapt and differentiate through coordination of protein synthesis. Two major mechanisms exist to control gene transcription as follows: one through recruitment to or preventing the RNA polymerase (RNAP) 4 binding to promoter DNA and one through activation or inhibition of the RNA polymerase activity. Despite the variations in subunit numbers that constitute the basal transcription machinery, RNA polymerases are structurally and functionally highly conserved in all kingdoms of life (1). During the initiation of transcription, RNA polymerases invariably melt the promoter DNA and have to engage the single-stranded template DNA in the structurally conserved catalytic cleft (2).
The major variant bacterial 54 -RNAP is regulated by an activating mechanism with some resemblance to the operation of eukaryotic RNAP II. In both cases, passing from the closed to the open RNAP promoter complex requires activator proteins that hydrolyze nucleoside triphosphate to drive open RNAP promoter complex formation (3). The 54 -RNAP binds to specific promoter sites centered on positions Ϫ24 and Ϫ12 relative to the transcription start site and remains in a transcriptionally silent conformation. Open complexes of the 54 -RNAP promoter complexes are thermodynamically labile, and activation relies on the productive coupling of nucleoside triphosphate hydrolysis (energy coupling) from 54 activators to the 54 -RNAP promoter complex. The 54 activator-dependent initial events in open complex formation are mediated by structural changes between 54 and the promoter DNA, which can be melted from position Ϫ12 to Ϫ5 in a reaction that can occur independently from RNAP core determinants. The restructuring of promoter DNA by activated 54 has been termed 54 isomerization (4 -6). Following the initial events and requiring the presence of the RNAP core, the promoter DNA opening extends to position ϩ3 relative to the transcription start site (7).
Energy coupling primarily involves intimate interactions between the activator and the 54 promoter complex. 54 activators, also termed enhancer-binding proteins (EBP), belong to the versatile AAAϩ protein (ATPases associated with various cellular activities) family of molecular machines (for review see Refs. 8 -10). Members of EBPs include the well studied NtrC, PspF, DctD, XylR, NifA, and DmpR proteins (11). PspF is the phage shock protein F that activates transcription of psp genes involved in the phage shock response in Escherichia coli (12,13). The AAAϩ domain (see Fig. 1 for AAAϩ domain features) of the 54 activator PspF (PspF 1-275 ) is necessary and sufficient to activate transcription of the 54 -RNAP in vitro and in vivo (14,15). The 54 activator signature sequence GAFTGA loop 1 directly binds to the region I of 54 , which is involved in maintaining a closed 54 -RNAP promoter complex (16,17). Stable interactions between PspF 1-275 and 54 form when PspF 1-275 is bound to the nonhydrolyzable ATP transition state mimic ADP-AlF x (16). This transition state complex was deemed a structural and functional intermediate complex that is "trapped" in a conformation en route to open complex formation (18). Structures of PspF 1-275 in complex with 54 and PspF 1-275 structures bound to different nucleotides, combined with biochemical properties of PspF 1-275 variants, suggested a detailed molecular mechanism by which different nucleotide states would trigger motions of the GAFTGA motif within a PspF 1-275 subunit (19,20). Signaling between Walker motif B and the GAFTGA loop 1 involves tight interactions between the PspF Walker B residue Glu 108 and Asn 64 , nucleotide bindingdependent motions of linker regions 1 and 2, and a number of signaling residues that would position the GAFTGA loop 1 to contact 54 . Motions of the GAFTGA loop 1 appear synchronized with motions of the adjacent contacting loop 2. The AAAϩ sequence insertion that structurally forms loop 2 in EBP classifies this family as member of the pre-sensor I ␤-hairpin insertion superclade of AAAϩ proteins ( Fig. 1) (21). Loop 1 and loop 2 are positioned near the pore of the hexameric PspF 1-275 assembly (19).
AAAϩ proteins invariably contain the Walker A and Walker B motifs that structurally define them as P-loop ATPases. The second region of homology (overlapping with the previously described AAA minimum consensus sequence) or degenerate forms thereof is the major distinguishing primary sequence motif that distinguishes AAAϩ proteins from other P-loop ATPases (22). The SRH contains the sensor I motif, often a threonine/asparagine pair, and an arginine residue known as the R-finger that functions as a trans-acting catalytic residue between adjacent subunits that form the catalytic site (8,9). The R-finger provides a structural rationale for why oligomerization is required for efficient hydrolysis in AAAϩ proteins. The first sensor residue of the SRH is predominantly a threonine, and the second one is often an asparagine, although threonines and methionines are also commonly found at this position. Well studied members of the AAAϩ superfamily that share this motif include N-ethylmaleimide-sensitive fusion protein (TN), p97 (TN), RuvB (TT), ClpA (TT), FtsH (TN), and PspF (TN). It was suggested that the second residue of sensor I works in concert with the second acidic residue of the Walker B motif to coordinate a catalytic water molecule (23). Currently, the role of the first residue of sensor I is not clear.
We wished to identify specific roles for sensor I residues of PspF in the steps that are known to be required for open complex formation within the 54 -RNAP promoter complex. We tested activities of sensor I variants ( , only PspF 1-275 T148A is severely affected in transcriptional activity. We show that defects of PspF  T148A are because of defects in productive coupling nucleoside triphosphate hydrolysis to a restructuring of the 54 promoter complex and is not because of the failure to stably engage with the closed 54 promoter complex, suggesting that Thr 148 is critically important in the events of open complex formation subsequent to 54 promoter engagement and independent of the presence of the RNAP core. The 54 promoter binding by PspF  T148A distinguishes this variant from other PspF 1-275 variants proposed to alter the position the GAFTGA loop (20). This suggests that an additional important structural link between the nucleotide-binding site and the 54 proximal contacting sites exists. Thr 148 appears to enable hydrolysis-dependent changes of the second acidic Walker B residue (Glu 108 ) to propagate to other SRH residues that could mediate movements to the PspF loop 2. In support of this communication route, activities of PspF 1-275 variants that connect loop 2 with the adjacent GAFTGA loop are similar to those of PspF  T148A . We propose that Thr 148 structurally couples hydrolysis-dependent motions to movements of loop 2 that are needed to restructure the 54 -RNAP closed complex via the loop 1 GAFTGA adaptor sequence.
Gel Filtration Chromatography-Gel filtration chromatography on Superdex 200 10/300 (Amersham Biosciences) was carried out exactly as described previously (25). Briefly, PspF  proteins were introduced at high (63 M) and low (10 M) protein injection concentrations to assess concentration-dependent oligomer formation of apo-PspF 1-275 . In the presence of ATP or ADP, the column was pre-equilibrated with buffer (20 mM Tris-HCl, pH 8.0, 50 mM NaCl, and 15 mM MgCl 2 ) supplemented to give final ATP or ADP concentrations of 0.5 mM. Filtration was performed at 4°C at a flow rate of 0.5 ml/min. The column was calibrated with globular proteins as follows: apoferritin (443 kDa), alcohol dehydrogenase (150 kDa), bovine serum albumin (66 kDa), and carbonic anhydrase (29 kDa). All experiments were repeated at least twice.
ATP Binding and Hydrolysis-ATP binding assays of PspF 1-275 proteins by UV cross-linking and ATPase assays were essentially as described (24). Briefly, 20-l samples containing 3.4 g of purified protein in reaction buffer A (35 mM Tris acetate, pH 8, 70 mM potassium acetate, 5 mM magnesium acetate, 19 mM ammonium acetate, 0.7 mM dithiothreitol) contained 3 Ci of [␣-32 P]ATP (3000 Ci/mmol) Ϯ 1 mM ATP. Samples were illuminated (254 nm, UV lamp UVG; UVP Inc., CA) for 15 min on ice. PspF 1-275 was analyzed by 12% SDS-PAGE after staining with Coomassie Blue, and the gels were scanned and then dried, and radioactivity was determined by phosphorimaging (Fuji Bas-1500 with Tina 2.10g software). The amount of protein per band was estimated by light transmission scanning.
ATPase assays were in reaction buffer A (above) supplemented with MgCl 2 to give final Mg 2ϩ concentrations of 15 mM and incubated at 30°C with different ATP concentrations containing 0.06 Ci/l [␣-32 P]ATP. Reactions were stopped by adding 5 volumes of 2 M formic acid. Released 32 P i or [␣-32 P]ADP was separated from ATP by thin layer chromatography, and radiolabeled P i , ADP, and ATP amounts were measured by phosphorimaging.
PspF  Protein Binding to and Isomerization of the 54 Promoter Complex-54 promoter complexes were performed for binding and isomerization assays with PspF 1-275 , essentially as described (4). Briefly, the bottom strand of the nifH promoter oligomer (Ϫ60 to ϩ28 of the S. meliloti nifH promoter) was labeled with a fluorescent HEX tag, purchased from Operon AG and named WVC3-HEX. Duplex formation between the bottom strand and nonlabeled top strand occurred by mixing 5 l of Hex-labeled NifH bottom strand (200 nM) and 5 l of (400 nM) nonlabeled top strand in 10 mM Tris-HCl, pH 8.0, and 1 mM MgCl 2 . The mixture was then heated to 95°C and allowed to anneal slowly while cooling over 2.5 h to minimize the forma-tion of mismatch double-stranded DNA. For isomerization assays, the top strand carries a mismatch at positions Ϫ12/Ϫ11 that is thought to stabilize the melted 54 -DNA structure (4). Final concentrations in binding to and isomerization of the 54 promoter complexes were 20 nM promoter DNA, 2.3 M 54 , and 3 M PspF 1-275 . Binding of PspF 1-275 proteins to 54 promoter DNA was assessed by in situ formation PspF 1-275 -ADP-Al x in the presence of 0.4 mM ADP, 5 mM NaF, 0.4 mM AlCl 3 (16). Isomerization assays were performed in the presence of 4 mM dGTP.

Sensor I Residues Are Not Required for Nucleotide Binding or
Nucleotide-dependent Hexamer Formation-The process of transcription activation depends on a number of distinct functionalities of the PspF AAAϩ domain that are thought to occur in the following order: nucleotide binding, hexamer formation, contacting the 54 promoter complex, ATP hydrolysis coupled to restructuring of the RNAP promoter, and disengagement from the open RNAP promoter complex. Disruption at any stage in this process should prevent subsequent events and result in the failure to activate transcription. It is therefore important to test PspF 1-275 variants for each of those activities to attribute discrete functional roles in the multistep process of transcriptional activation.
We substituted the sensor I TN residues to Ala as well as the Asn to the polar Ser and purified the respective proteins to obtain PspF 1-275 T148A , PspF 1-275 N149A , and PspF 1-275 N149S in the same manner as PspF 1-275 . All proteins were soluble. We applied UV cross-linking of PspF 1-275 proteins to radiolabeled [ 32 P]ATP to detect changes in nucleotide binding. PspF  and other EBP appear to have a high off rate for ATP, in agreement with the high K d values (34 M) of PspF 1-275 for ATP␥S reported previously using isothermal calorimetry (25). The nonequilibrium UV cross-linking method allowed us to identify significant changes in binding affinities for PspF 1-275 Walker A and Walker B variants consistent with their established binding characteristics in other AAAϩ proteins, including the EBP NtrC (24,27,28). We found that none of the sensor I variants of PspF 1-275 was severely affected in ATP binding ( Fig. 2A).
We next tested the PspF 1-275 sensor I variants for their ability to form higher order oligomers in the absence and presence of nucleotide. PspF 1-275 was shown to form higher order oligomers, probably hexamers, in a strictly concentrationdependent fashion that correlates with ATPase activity (25). Defects in oligomer formation are therefore predicted to negatively affect hydrolysis because of loss of cooperativity in ATP hydrolysis between subunits. It is established that ATP and ADP binding induces structural changes within PspF   (20) and strongly promotes hexamer formation of PspF 1-275 , indicating nucleotide binding induces structural changes to increase self-association (25). To test if PspF 1-275 sensor I variants are defective in higher oligomer formation in their apo-, ADP-, and ATP-bound forms, we carried out gel filtration chromatography experiments in the presence and absence of nucleotides. When ADP or ATP was present, we chose the lowest protein concentrations that could still be detected with satisfactory confidence to maximize the sensitivity for detecting , and PspF 1-275 N149S in the presence of 0.5 mM ADP or ATP during gel filtration indicates that these variants have no ADP or ATP binding-dependent oligomer formation defects. Apparent apo-PspF 1-275 hexamers elute around 11.7 ml and apparent nucleotide-bound PspF 1-275 hexamers at around 11.2 ml. Apparent dimers elute at around 15 ml. Structural and functional data argue that the difference in elution volumes between higher order oligomeric apo-and nucleotide-bound forms of PspF 1-275 are because of nucleotide-induced conformational changes in the hexamer rather than to a different subunit composition (25). We conclude that substitutions of sensor I residues do not result in conformational changes at the subunit interface and that ADP or ATP binding-dependent conformational changes occur that are compatible with hexamer formation.
Sensor I Residues Are Required for High ATP Hydrolysis Rates-We determined the kinetic ATP hydrolysis parameters of mutant proteins at 30°C (Table 1). Sensor I variants showed 6 -15-fold decrease in ATP hydrolysis rates, and PspF 1-275 N149A and PspF 1-275 N149S have a significantly reduced K m value. A reduced K m value usually indicates an increase in nucleotide affinity and/or structural changes in  , ATPase activity was stimulated in the presence of subsaturation ADP concentrations and inhibited at Ͼ4 mM concentrations of ATP (data not shown). Therefore, no obvious role for sensor I residues in cooperative hydrolysis is evident.
Low ATP Turnover Rates of PspF 1-275 N149A Are Not Detrimental for Activating Transcription-ATP hydrolysis by PspF 1-275 is needed to restructure the stable 54 -RNAP promoter closed complex to an open thermodynamically unfavorable conformation (4,29). We determined whether the reduced turnover rates of PspF 1-275 T148A and PspF 1-275 N149A could still activate transcription in vitro, in single round transcription assays from the S. meliloti supercoiled nifH promoter (Fig.  3) (19,20). The substituted residues of these variants form part of a structural conformational signaling pathway that extends from the nucleotide-binding site to the GAFTGA containing loop 1 (Ser 75 , Glu 76 , His 80 , and Arg 95 ) or are part of loop 2 (Arg 131 and Val 132 ) that is adjacent to the GAFTGA loop 1 (see Fig. 6 for orientation). We proposed that the signaling to and the integrity of the GAFTGA loop 1 was critical for contacting 54 and energy coupling. Structures (19) and models (10,30) of the PspF 1-275 -54 complex suggested that the promoter DNA lies "sandwiched" between 54 and the upper face of the PspF 1-275 hexameric ring. Contact features of PspF subunits, notably the GAFTGA loop 1, the associated loop 2, and 54 , suggest that the GAFTGA loop 1 and loop 2 are proximal to promoter DNA. Proximity-based studies support this architecture, although direct PspF 1-275 promoter DNA contacts have not been identified (31). To directly assess the capacity of PspF 1-275 variants to restructure the 54 promoter complex independently from a possible contribution of the RNAP core, we carried out direct 54 promoter isomerization assays (4). The assay detects bona fide open 54 promoter complexes after isomerization by and disengagement from PspF 1-275 (4) permitting distinction between activities directly associated with energy coupling to the 54 promoter and those that may be RNAP core-dependent. Fig. 4  shows some residual isomerization activity in agreement with our in vitro transcription results. All other transcriptionally inactive PspF 1-275 variants tested also fail to isomerize the 54 promoter. These results underline the major mechanistic importance of energy coupling-dependent isomerization of the 54 promoter in initiating 54 -RNAP-dependent transcription. We infer an important role for Thr 148 at this step in nucleotide hydrolysis-dependent open complex formation of the 54 -RNAP promoter. 54 Promoter DNA Binding Activities Functionally Link PspF  T148A with Loop 2-Considering the likely spatial organization of PspF in complex with 54 bound to promoter DNA (see above), we reasoned that the 54 bound to promoter DNA would provide the best structural in vitro mimic to study PspF 1-275 54 promoter complex interactions. ADP-AlF x -dependent structural changes within the PspF 1-275 -54 -RNAP promoter complex have been reported and represent a structural and functional transition conformation that is trapped en route to open complex formation (5,32). To test if the failure of PspF 1-275 variants to isomerize the 54 promoter complex is a consequence of a reduced capacity to stably engage with the 54 promoter complex, we incubated a preformed 54 -nifH promoter probe complex with PspF 1-275 variants in the presence of ADP-AlF x reagents. Fig. 5  , and PspF 1-275 R95A fail to engage with the 54 promoter DNA (Fig. 3). This defect explains why PspF 1-275 S75A , PspF 1-275 E76A , PspF 1-275 H80A , and PspF 1-275 R95A fail to isomerize the 54 promoter complex, which relies on complex formation and coupling nucleoside triphosphate hydrolysis to the 54 promoter complex. Activity data are in full agreement with the suggested function of Ser 75 , Glu 76 , His 80 , and Arg 95 to help position the GAFTGA loop 1 to stably contact the 54 promoter. In marked contrast, PspF 1-275 T148A , PspF 1-275 N149A , and PspF 1-275 V132A represent a novel class of PspF 1-275 variants in that they show full 54 promoter binding activities but fail to isomerize the 54 promoter DNA structure. These results indicate that severe isomerization and transcription activation defects of PspF 1-275 T148A and PspF 1-275 V132A are caused by defects subsequent to engaging with 54 promoter DNA. PspF  R131A shows minor defects in contacting 54 promoter DNA. The functional defects of PspF 1-275 T148A and the loop 2 variants PspF 1-275 R131A and PspF 1-275 V132A that are directly associated with energy coupling ( 54 promoter isomerization) and not with 54 promoter DNA engagement suggest a structure/function link between Thr 148 and loop 2 (see below). Importantly, this is a distinct link from the nucleotide binding pocket to the GAFTGA loop 1 signaling pathway.

DISCUSSION
More generally, we wished to establish the functional roles of the AAAϩ conserved sensor I residues of AAAϩ proteins. In particular, we studied sensor I residue variants of the AAAϩ domain of the 54 activator PspF that is necessary and sufficient to activate transcription in vivo and in vitro. A systematic analysis of various PspF 1-275 activities that are required for functional output allowed assignment of specific roles of sensor I residues during the nucleotide binding initiated process that results in transcriptional activation. The wealth of structural and functional data for PspF 1-275 and the availability of methods to study basal and full transcriptional activation properties helped to identify specific mechanistic functions of the PspF sensor I residues and by extension their roles in other AAAϩ proteins.
The Asparagine of the PspF Sensor I Is Involved in ATP Hydrolysis per Se but Is Not Essential for Functional Output-We found that the conserved Asn residue of sensor I is required for high ATPase rates and affects the Michaelis-Menten kinetics. Gel filtration experiments with PspF 1-275 N149A (Fig. 2B) and PspF 1-275 N149S in the presence of ATP or ADP show that Asn 149 is not essential for nucleotide binding-dependent hexamer formation, excluding the possibility that reduced hydrolysis rates are caused by oligomerization defects. ADP was not a stronger binding competitor for ATP in PspF 1-275 N149A compared with PspF 1-275 as judged by ATPase rates in the presence of excess ADP, suggesting that the reduced k cat for PspF 1-275 N149A is not because of slower ADP release. Substitutions of sensor I Asn residue of the FtsH protease (33), the  p97/VCP homologue VAT (34), or the Lon protease (35) also resulted in strongly reduced hydrolysis rates in those proteins. Consistent with our findings with PspF 1-275 , a strong decrease in k cat and a 2-3-fold decrease in K m was reported for the NBD2 AAAϩ domain of Hsp104 carrying an Asn to Ala substitution of the sensor I (36). Hsp104 is involved in thermotolerance, and decreased high temperature survival rates were observed for this sensor I mutant. In FtsH the corresponding mutation resulted in reduced but not abolished in vivo protease activity. In PspF , Asn 149 is not essential for nucleotide-dependent interactions with 54 , isomerization of the 54 promoter complex, or transcriptional activation of the 54 -RNAP. We deduce that Asn 149 of PspF is involved in ATP hydrolysis per se but is not essential for functional output in vitro. We propose that the side chain of Asn 149 contributes to the catalytic site geometry that allows for the high turnover rates of PspF. Our functional assignment is in agreement with the previously suggested roles of the second sensor I residue in other AAAϩ proteins (37,38).
In contrast to PspF  (Fig. 6). In the ATP-bound state, Walker B residue Glu 108 clearly interacts with Asn 64 . The Glu 108 -Asn 64 interaction is required to position the distant GAFTGA loop 1 for engagement with 54 via linker 1 and a number of signaling residues, including Ser 75 , Glu 76 , His 80 , and Arg 95 (20) (Glu 76 and His 92 are shown in Fig. 6). In the ADP-bound state, Glu 108 pivots away from Asn 64 toward Thr 148 and then interacts with Thr 148 via a water molecule. The Thr 148 -interacting water molecule is present in all PspF 1-275 structures (apo, ADP, ATP, and AMP-PNP (20)), suggesting that it is an integral and significant structural element of PspF. The Glu 108 -Thr 148 connectivity can link the ADP-bound state to the SRH and thus via the SRH C-terminal helix to ␣-helix 4 that forms the structural base of loop 2. More directly, the Thr 148 -associated water molecule could structurally bridge between the Glu 108 side chain and the Glu 108 main chain. Main chain movements are important for linker 2 movements, which in turn positions ␣-helix 4. Although a detailed structural signaling pathway from Thr 148 to loop 2 cannot be described with complete confidence with the available structural and functional data, our results strongly suggest that two distinct signaling pathways from the P-loop Glu 108 exist. One signals toward the GAFTGA loop (via Asn 64 and linker 1), and one signals to loop 2 (via Thr 148 and possibly linker 2). Similarities of the linker 1 and linker 2 conformations in the ADP-and the ATP-bound structures between PspF 1-275 and the Ltag helicase (40) provided evidence for some conserved nucleotide hydrolysis-dependent dynamics between these two AAAϩ proteins (20). The functional similarities of PspF 1-275 T148A , PspF 1-275 V132A , and PspF 1-275 R131A support a structural linkage between Thr 148 and loop 2. Furthermore, properties of PspF 1-275 V132A and PspF 1-275 R131A suggest that loop 2 does not directly contact 54 , because neither variant fails to engage with the 54 promoter DNA complex. Nucleotide hydrolysis-induced changes in the ATP-binding pocket probably propagate toward loop 2. The loop 2 motion then drives the GAFTGA loop adaptor during the power stroke that remodels the 54 -RNAP promoter complex to an open, transcriptionally competent conformation, a coupling mechanism that critically involves Thr 148 .
The clear conservation of the sensor I motif suggests a generalizable role within the AAAϩ family. Interactions between the first sensor I residue (1stSI, Thr 148 in PspF) with the second acidic residue of the Walker B motif (2ndWB) are probably common in the wider AAAϩ family as judged by the available high resolution structures of AAAϩ proteins. 1stSI-2ndWB direct interactions exist in p97 ((41), PDB entry 1E32, sensor I motif TN), N-ethylmaleimide-sensitive fusion protein ((42), PDB entry 1D2N, TN), ClpA ((43), PDB entry 1R6B, TN), and judging from proximity also in the Ltag helicase ((40), PDB entry 1SVO), which has unconventional sensor I ((T)MN) and Walker B motifs (ED instead of DE). 1stSI-2ndWB interactions exist in FtsH ( (44,45), PDB entry 1IXZ, TN) and RuvB (PDB entry 2C90, SN) and also involve a water molecule, very similarly positioned as the Thr 148 -associated water molecule in PspF. The coupling role of the Thr 148 of PspF to loop 2 could also be of functional importance in other AAAϩ proteins. Insertions at sites corresponding to the position of loop 2 of PspF define a subclass of AAAϩ proteins termed pre-sensor I ␤-hairpin (loop 2 equivalent) super-clade (21). Members of this super-clade also include the HslU, Clp, and Lon proteases, RuvB, the Ltag helicase, and the MCM helicases. The loop 2 equivalent insertions form hairpin structures in RuvB ((46), sensor motif TT) and Ltag. In RuvB the ␤-hairpin is important for interactions with RuvA to cooperate in Holliday junction branch migration. In Ltag helicase, the ␤-hairpin appears directly involved in ATP-driven translocation of doublestranded DNA (40). Further functional studies that address the role of the conserved first residue of sensor I are needed to understand mechanistic similarities and differences of AAAϩ proteins.