Allosteric Regulation of SecA

In bacteria, the SecA protein associates with a ubiquitous protein channel SecYEG where it drives the post-translational secretion of pre-proteins across the plasma membrane. The high-resolution structures of both proteins have been determined in their resting states; however, the mechanism that couples ATP hydrolysis to active transport of substrate proteins through the membrane is not well understood. An analysis of the steady-state ATPase activity of the enzyme reveals that there is an allosteric binding site for magnesium distinct from that associated with hydrolysis of ATP. We have demonstrated that this regulation involves a large conformational change to the SecA dimer, which exerts a strong influence on the turnover and affinity for ATP, as well as the affinity for ADP. The strong inhibitory influence of magnesium on the ATPase activity can be countered by cardiolipin and conditions that promote protein translocation.

In bacteria, the SecA protein associates with a ubiquitous protein channel SecYEG where it drives the post-translational secretion of pre-proteins across the plasma membrane. The high-resolution structures of both proteins have been determined in their resting states; however, the mechanism that couples ATP hydrolysis to active transport of substrate proteins through the membrane is not well understood. An analysis of the steady-state ATPase activity of the enzyme reveals that there is an allosteric binding site for magnesium distinct from that associated with hydrolysis of ATP. We have demonstrated that this regulation involves a large conformational change to the SecA dimer, which exerts a strong influence on the turnover and affinity for ATP, as well as the affinity for ADP. The strong inhibitory influence of magnesium on the ATPase activity can be countered by cardiolipin and conditions that promote protein translocation.
Chemo-mechanical protein machines have evolved to couple energy derived from chemical reactions to mechanical motion. They have been exploited to drive a number of processes that are central to cell biology, such as protein translocation.
Membrane and secretory proteins are targeted to specific membranes before they pass through or into the bilayer. A ubiquitous protein channel is found in the plasma membrane of archaea and bacteria, and also in the endoplasmic reticulum of eukaryotes. Co-translational and post-translational targeting systems converge on the Sec complex whereupon the unfolded protein substrate is driven through the membrane (1). The structure of the protein channel has been determined in its resting state to high resolution (2, 3), and its active and membrane-bound state is most likely a dimer (3)(4)(5)(6)(7).
In bacteria, post-translational translocation of secretory proteins requires that they are maintained in an unfolded conformation, before they are passed on to a soluble ATPase SecA (8). Upon receiving the substrate, SecA associates with the proteinconducting channel, the SecYEG complex, where it harnesses the energy from ATP binding and hydrolysis to actively push it across the membrane (9 -11). A model has been proposed for a multistep reaction for the transport of proteins across the membrane (12), regulated by distinct ATP binding and hydrolysis events, translocating ϳ25 amino acids (2.5 kDa) in each step (13). However, another study estimates that 5 ATP molecules are required to send each amino acid across the membrane (14).
Unliganded SecA exists in solution in a monomer-dimer equilibrium (15,16), and is predominantly dimeric (16). However, the precise oligomeric state of SecA that binds to the protein channel and drives the translocation reaction has formed the subject of conflicting accounts. Several studies reported that a dimer with two active subunits was required for active translocation (17) and in support, a later study demonstrated that covalently cross-linked dimers were able to sustain ATP hydrolysis and protein translocation (18,19). Other investigators showed that acidic phospholipids, detergent, signal peptide, and SecYEG apparently bring about a dissociation of the dimer (20). Following on from this, a mutant of SecA (⌬11/N95) was constructed with truncations to the N and C termini (⌬2-11 and ⌬832-901), which was shown not to form inter-subunit cross-links, and to sediment through sucrose cushions as if it were a monomer (21). An analysis of complexes formed between SecYEG and SecA by analytical ultracentrifugation showed that in the absence of nucleotide one SecA bound to the channel, whereas two bound in the presence of the ATP analogue AMPPNP 2 (7).
Several different structures of SecA have now been determined, five dimers (22)(23)(24)(25)(26) and one monomer (27). Four of the dimers revealed the SecA protomers packed head-to-tail in the crystal lattice, all in different antiparallel conformations (22)(23)(24)26), another has a parallel arrangement of the subunits (25). These differences obviously add further confusion to our picture of the nature of the oligomeric state and the conformation of the active form of the protein. The monomer structure, and subsequently the latest dimer were found to exhibit a large domain movement, exposing a cleft that has been proposed to form a polypeptide binding site (26,27). Interestingly, the conformation of the nucleotide binding fold does not seem to be affected by the presence of bound nucleotide (22-24, 26, 27).
The ATPase activity of SecA has to be regulated in some fashion, to prevent futile hydrolysis of ATP during moments of rest. The basal ATPase activity of SecA is partially increased by the presence of acidic phospholipids and then fully stimulated by addition of vesicles containing SecYEG and precursor protein (28,29). The mechanism for this regulation is not clear.
There have been several observations with respect to magnesium and its effects on the regulation of SecA, which have not been adequately explained. Removal of the cation from a binding site, distinct from the nucleotide binding folds, can result in a 10-fold stimulation of ADP release (30). The activity of SecA and the elevated activity found in a complex of SecA and SecYEG were both stimulated by the removal of Mg 2ϩ , which also served to stabilize the latter complex in the presence of ATP (or ATP␥S) (5). Mg 2ϩ has also been shown to modulate fluorescence anisotropy measurements, as well as the fluorescence observed from intrinsic and extrinsic probes of the enzyme (27,31,32).
Despite the fact that the structures of both SecA and SecYEG have been determined, we are still a long way from understanding their concerted mechanism of action. To reach this point, we need to understand the nature and timing of the conformational changes within SecA, and how they are affected by the ATPase cycle. In the light of the conflicting and fragmented accounts of the reaction cycle, we decided to conduct a thorough analysis of the kinetics of the protein translocation process. As a first approach to this characterization we have performed a steady-state analysis of SecA in isolation, together with a study of its oligomeric state by equilibrium and velocity sedimentation using analytical centrifugation. An analysis of the wild-type and proposed monomeric mutant SecA-⌬11/N95 (21) in this way generated some new findings on the oligomeric state of SecA, its ATPase activity, and the influence of magnesium and lipids.

EXPERIMENTAL PROCEDURES
Chemicals and Biochemicals-Escherichia coli polar lipid and E. coli cardiolipin were purchased from Avanti, and were prepared at 10 mg/ml in 50 mM triethanolamine, pH 7.5, 50 mM KCl. Polyoxyethylene(9)dodecyl ether (C 12 E 9 ) was purchased from Anatrace, the EnzChek TM kit from Molecular Probes (Invitrogen), Chelating Sepharose fast flow from GE Healthcare, and all other reagents from Sigma.
ATPase Assays-Steady-state ATPase assays were carried out and monitored at 25°C, either by the linked pyruvate kinase/lactate dehydrogenase assay, or by using the Enz-Chek TM Kit (Invitrogen), employing a Lambda 25 spectrophotometer (PerkinElmer Life Sciences). The pyruvate kinase/lactate dehydrogenase assay was used to measure extremely slow turnover accurately by virtue of a regenerating system, and the EnzChek TM kit to investigate the effects of Mg 2ϩ , where the former would be inappropriate due to the dependence of pyruvate kinase on this cation.
EnzChek TM Assay-The standard assay components were 0.2 mM 2-amino-6-mercapto-7-methylpurine riboside and 1 unit of purine nucleoside phosphorylase in TK buffer (50 mM triethanolamine, pH 7.5, 50 mM KCl). 0.3 M SecA or SecA-⌬11/N95 was used (monomer concentration) in each reaction, and ADP, Mg 2ϩ , and E. coli lipids were added to the concentrations indicated in the text, prior to the initiation of the reaction by addition of ATP.
Determination of Residual Magnesium Concentration-The amount of Mg 2ϩ contained within the standard assay conditions (TK buffer) was calculated using inductively coupled plasma-atomic emission spectroscopy, using a Jobin Yvon Horiba Ultima 2 sequential spectrometer fitted with a Bergener Mira Mist Nebulizer and automated with a Jobin Yvon AS421 auto-sampler.
Curve fitting and analysis for all kinetic data, including that of the global fits is shown under supplemental materials.
Analytical Ultracentrifugation-SecA stock solutions of wild-type and SecA-⌬11/N95 were dialyzed extensively against either TK or TKM buffers. These were subjected to sedimentation velocity and sedimentation equilibrium centrifugation experiments using either an XL-A or XL-I centrifuge (Beckman Instruments) equipped with standard Epon centerpieces and an An-60Ti rotor. SecA or SecA-⌬11/N95 (0.4 -13 M) were used and scans were recorded at either 230 or 280 nm. Sedimentation velocity experiments were carried out at 25°C and 130,000 ϫ g. The scans were fitted to a continuous c(s) model, using Ultrascan (34). The data were then extrapolated to zero concentration to yield s 20,w 0 . Equilibrium scans were taken at 16 and 20 h, at speeds of 3,000, 5,800, and 8,000 ϫ g at 25°C. The data were analyzed using a variety of models, the data shown are that of a 1-component ideal species model, which yielded the most appropriate fit using Ultrascan (34). The data were then subjected to a Monte Carlo analysis, to reveal the statistical values for all parameters.
In Vitro Translocation Assays-SecYEG was reconstituted into E. coli polar lipid proteoliposomes as described previously (35). Translocation of proOmpA into SecYEG proteoliposomes was assayed essentially as described (36), except that proteaseprotected proOmpA was detected by Western blot using an antibody raised against proOmpA.

SecA Activity Is Affected Dramatically by Magnesium Acting at a Location Distinct from the Hydrolytic Site-Steady-state
ATPase assays were used in a range of conditions to characterize the kinetics of ATP turnover by both wild-type SecA and the proposed monomeric mutant SecA-⌬11/N95 (21). In 2 mM Mg 2ϩ , ATP binding is tight and substrate turnover is slow (Fig.  1A, Table 1). However, when no Mg 2ϩ was included in the buffer (thus in the presence of only residual magnesium), the behavior of SecA is dramatically altered, the affinity for ATP being weakened by ϳ150 fold, and the turnover increased by a factor of 30 (Fig. 1B, Table 1); both wild-type and the SecA-⌬11/ N95 behave essentially in an identical manner.
The residual magnesium concentration in the assay buffer was calculated using inductively coupled plasma-atomic emission spectroscopy to be 1 M, ϳ3 times greater than the SecA concentration in the experiment (0.3 M). Unsurprisingly, addition of EDTA to the reaction inhibited the activity completely (data not shown). Characteristic of all ATPases, SecA uses Mg 2ϩ as a cofactor coordinating the ␥-phosphate of ATP required for hydrolysis.
Accordingly, the apparent inhibitory effect of Mg 2ϩ was examined in more detail and it was shown to be relatively unaffected by a 10-fold increase in the concentration of ATP (Fig. 2, Table 2). The affinity of Mg 2ϩ for ATP is relatively high (K d -[Mg 2ϩ -ATP] ϭ 20 M) (37), therefore the fact that the inhibition is not strongly coupled to the concentration of ATP indicates that the inhibition occurs via Mg 2ϩ , and not by Mg 2ϩ -ATP. In other words, it is acting at a location distinct from the nucleotide binding site and the effect is allosteric. Moreover, it can be deduced that Mg 2ϩ -ATP is not the substrate for SecA, as the Mg 2ϩ required for catalysis must be always bound at the NBD, as hydrolysis can occur when ATP is in great excess over Mg 2ϩ (1 mM ATP to 1 M Mg 2ϩ , Fig. 1B). The actual binding affinity for Mg 2ϩ at this second allosteric site must be much higher than the affinity Mg 2ϩ has for ATP, given the fact that it remains inhibitory when ATP is more than 500 times in excess (Fig. 1B).
The data were fitted according to a global fit, which incorporates a matrix of Mg 2ϩ and ATP concentrations and the respective steady state rate kinetics ( Fig. 2B and supplemental materials). This procedure made an adjustment for the chelating effect of ATP on Mg 2ϩ and also incorporated the data for both weak and tight binding of ATP with residual and 2 mM Mg 2ϩ , respectively. The fit matched the data well, and a simple model has been proposed to explain the allosteric influence of Mg 2ϩ on the reaction cycle of SecA (Fig. 3).
The Competitive Inhibition of SecA by ADP Is Affected by Magnesium-The effects of product inhibition were studied both in the presence and absence of Mg 2ϩ ( Fig. 4 and Table 3). ADP exhibited a classic competitive inhibition effect with residual Mg 2ϩ ; there was a 10-fold increase in K i(app) at a 10-fold higher ATP concentration ( Table 3). The affinity of SecA for ADP, given by K d [ADP], was calculated at different concentrations of ATP and Mg 2ϩ , taking into account the increased affinity for ATP at the higher Mg 2ϩ concentration (see supplemental materials and Table 3). As with affinity of SecA for ATP, in the presence of 2 mM Mg 2ϩ the affinity for ADP was greatly increased, by ϳ200-fold (Tables 1 and 3).
The Inhibitory Effect of Magnesium Can Be Alleviated by Cardiolipin-Preliminary experiments indicated that cardiolipin binds to SecYEG and stabilizes the dimeric conformation,  Table 1.

TABLE 1
Calculated values of the K m ͓ATP͔ and V max for wild-type SecA and SecA-⌬11/N95 with and without added 2 mM Mg 2؉ The data were collected according to Fig. 1, A and B, and fitted to the Michaelis-Menten equation as described under supplemental materials. S.E. from the fitting procedure is shown. which is proposed to be the active form. 3 Cardiolipin is also known to be a chelator of divalent cations (38). In addition, acidic phospholipids have been shown to be important for protein translocation (39). Therefore, we reasoned that the influence of cardiolipin on protein translocation might be an indirect effect mediated by Mg 2ϩ , and decided to determine the effect of cardiolipin on SecA activity. The experiment was conducted with and without added 2 mM Mg 2ϩ ; the detergent C 12 E 9 was used to solubilize the lipid, and was shown to change neither the affinity of SecA for ATP nor the turnover rate (  shown. An allowance of 10% was allowed for enzyme concentration between the two separate data sets in the fit. Calculated values for K i(app) are shown in Table 2.    Table 3.

TABLE 3 Calculated values for the K i(app) ͓ADP͔ determined with different concentrations of Mg 2؉ and ATP
The data were collected according to Fig. 4 . In low concentrations of Mg 2ϩ , cardiolipin was found to inhibit the ATPase activity (Fig. 5A), presumably due to chelation of the Mg 2ϩ required for the hydrolysis of ATP. At higher concentrations of Mg 2ϩ (100 M) there was a striking increase in activity peaking at ϳ2.4 M cardiolipin, followed by a gradual reduction (Fig. 5A), indicating that the effect cannot be simply explained by the chelation of Mg 2ϩ . The effects of increasing cardiolipin in a yet higher concentration of Mg 2ϩ resulted in a less pronounced rise and fall of ATPase activity. The inhibitory phase of cardiolipin in the presence of added Mg 2ϩ may have been the result of the aggregation of SecA-lipid complexes, because it cannot be the result of Mg 2ϩ chelation at this high a concentration. Next, we determined the K m value for ATP binding with high (2 mM) Mg 2ϩ and high (36 M) cardiolipin (asterisk in Fig. 5, A and B, and Table 4); both the K m and V max were characteristic of conditions having only a low concentration of Mg 2ϩ . Therefore, we can infer from this data that cardiolipin counteracts the inhibition caused by Mg 2ϩ . Total E. coli polar lipids had a less dramatic effect (Fig. 5C). An ϳ10-fold higher concentration was required to reach an equivalent enhancement. Therefore, we assumed that this was a result of the cardiolipin (9.8%, w/w) contained in this fraction.
Magnesium Alters the Conformation of the SecA Dimer, but Not the Oligomeric State-To determine whether the observed change in activity was affected by a change in the oligomeric state of SecA, sedimentation equilibrium and sedimentation velocity experiments were carried out by analytical ultracentrifugation. Sedimentation equilibrium revealed unambiguously that SecA is a dimer within a 0.4 -13 M concentration range in equivalent buffers to those used in the ATPase assays ( Fig. 6 and Table 5), and also in the presence of 2 mM EDTA (data not shown). Rather surprisingly, SecA-⌬11/N95, previously reported to be a monomer (21), was found to be a dimer as well. In both cases, the monomers were in such rarity that they could not be detected with sufficient accuracy to determine a dissociation constant, an indication that this value must therefore be in the low nanomolar range.
Subsequently, sedimentation velocity also revealed that there is essentially only a dimeric species present ( Fig. 7 and Table 6). However, it could be shown for SecA and SecA-⌬11/N95 that the sedimentation properties, and hence shape, were sensitive to Mg 2ϩ . An increase of 0.54 (SecA) and 0.11 (SecA-⌬11/N95) s 20,w 0 units (the sedimentation coefficient S, expressed in terms of a water solvent at 20°C and extrapolated to zero concentra-  Table 4. C, the ATPase activity wild-type SecA was measured in the presence 100 M ATP and 100 M Mg 2ϩ with increasing concentrations of cardiolipin and total E. coli polar lipids, which contain 9.8% cardiolipin. Error bars represent S.D. from four to six replicates.

(control)
These are compared to the values in ϩ/ϪMg 2ϩ conditions. The data were collected according to Figs. 1, A and B, and 5B, and fitted to the Michaelis-Menten equation as described under supplemental materials. S.E. from the fitting procedure is shown. tion) was observed in the presence of 2 mM Mg 2ϩ , compared with residual Mg 2ϩ , indicative of a more extended shape. The Influence of Magnesium Is Countered by Conditions That Promote Protein Translocation-The protein translocation reaction driven by wild-type SecA and SecA-⌬11/N95 through SecYEG were compared and the mutant was found to have a significant proportion of the wild type activity. The amount of proOmpA translocated by wild type and by ⌬11/N95 is approximately the same after 5 and 15 min, respectively (Fig. 8A). Mg 2ϩ was titrated into the protein translocation assay and its effects monitored (Fig. 8B). There is a difference in the dependence of Mg 2ϩ in translocation compared with the ATPase activity of isolated SecA, indicating that conditions that promote translocation also alleviate the inhibitory effect of Mg 2ϩ . This may, in part, be due to the influence of cardiolipin present in the SecYEG proteoliposomes, but possibly also due to interactions of SecA with substrate and the SecYEG complex.

DISCUSSION
Our understanding of the molecular mechanism of Sec-dependent translocation and the specific role of SecA is in its infancy. There are now six structures available of this protein, five dimers and one monomer. There is, for instance, no consensus with respect to the active oligomeric form of the protein, the processive nature of the reaction and the stoichiometry of ATP/amino acid translocated (see Introduction). These uncertainties are understandable in view of the fact that there is little data to describe the interactions that occur between SecYEG, SecA, and substrate. There is also a paucity of information on the kinetics and the hydrolytic cycle of SecA and the timing and nature of the reactions that bring about the conformational changes that must ultimately drive the vectorial passage of proteins through the membrane.
The experiments described in this paper address this fundamental aspect of the system, the reaction cycle of SecA. We begin at the simplest level of the system and elucidate the basic ATPase mechanism. We have characterized SecA by steadystate ATPase assays and related the observed kinetics to structural changes by analytical ultracentrifugation.
The divalent metal cation Mg 2ϩ was found to have a powerful effect on the binding affinity and turnover of ATP: tight and slow in the presence of high concentrations, loose and fast with trace amounts. The simplest way of explaining this FIGURE 6. Sedimentation equilibrium centrifugation demonstrates that both SecA and SecA-⌬11/N95 are dimers, irrespective of the Mg 2؉ concentration. 1-13 M SecA or SecA-⌬11/N95 were centrifuged at three speeds at 25°C in either TK or TKM buffers. Equilibrium scans were taken at 16 and 20 h, and the data analyzed using a 1-component ideal species model using ultrascan (34). The data shown are an example fit from an experiment containing 1.5 M SecA in TK buffer at 3,000 (filled circles), 5,800 (open triangles), and 8,000 ϫ g (open circles) and the corresponding residual values. Calculated values for molecular weight are shown in Table 5.

TABLE 5
Molecular weight determined by equilibrium velocity centrifugation for wild-type SecA and SecA-⌬11/N95 with and without added 2 mM Mg 2؉ The data were collected according to Fig. 6, and fitted to a 1-component ideal species model as described under "Experimental Procedures." The S.D. from the Monte Carlo analysis is shown, following the converged value from the fit. The molecular weight of a monomer of each species, based on the amino acid sequence, is shown in parentheses.

Molecular mass (kDa)
Oligomeric   Table 6. inhibition is by binding of Mg 2ϩ at a site distinct from the nucleotide binding cleft, which leads to an increase in ATP affinity but a decrease in the rate of turnover. This allosteric effect of Mg 2ϩ is likely to be mediated by a rearrangement of the SecA dimer, which we have observed by the change in sedimentation behavior described below. In the presence of Mg 2ϩ , the affinity for ADP was also found to be increased by about 200-fold, very similar to the effect on the ATP affinity (150-fold increase). This indicates that the binding of ADP is similarly sensitive to this conformational change induced by Mg 2ϩ binding. The fact that Mg 2ϩ is both an inhibitor and an essential cofactor to the enzyme has meant that many of the previous results on the activity of SecA have been difficult to interpret. Moreover, the fact that the cation concentration varies according to the concentration of ATP will have added further confusion.
It has been proposed that Zn 2ϩ binds to the C-terminal 22 residues of SecA (40) to stabilize the interaction with the molecular chaperone SecB (41). The divalent cation-binding site determined in this study cannot be the same one that co-ordinates Zn 2ϩ , as SecA-⌬11/N95 has a 70-amino acid deletion at the C terminus (42), and remains the subject of Mg 2ϩ inhibition. IRA elements have been identified as "internal regulators" of ATP hydrolysis (43,44). Deletion of IRA1 (residues 783-795) results in an elevated rate of ATP turnover (44). These observations may have been due to the disruption of the allosteric site identified in this study.
The observed change in affinity for ATP caused by Mg 2ϩ indicates that a conformational change occurs upon binding. This prediction was tested by analytical ultracentrifugation to determine the extent of the rearrangement. In potentially oligomeric proteins, such rearrangements could take the form of changes in the state of assembly, but sedimentation equilibrium data showed that the molecular mass of the protein (200 kDa) was unchanged by Mg 2ϩ and was consistent with a dimeric quaternary structure. However, the results of sedimentation velocity reveal that structural rearrangements of the dimer are quite large, with the Mg 2ϩ -bound form of the protein having a significantly larger (ϳ6%) Svedberg constant. Interpreted empirically this is equivalent to the difference in sedimentation behavior between bovine carboxypeptidase A and bovine superoxide dismutase (45). These proteins have the same molecular mass and partial specific volumes but the latter is 40% longer in its longest axis (from 50 to 72 Å) (45). Hence, the change in sedimentation coefficient observed in SecA could represent a considerable opening of the structure.
The mutant SecA-⌬11/N95 is also a dimeric species but does not undergo such a large conformational change, probably due in part to its smaller size, and possibly also to the participation of the C terminus in this rearrangement, which has been truncated in the mutant form. This could explain why the mutant behaves in an identical manner in terms of ATPase activity, but has a reduced activity in terms of translocation. One slightly surprising aspect of this analysis was the indication that the predicted monomeric mutant SecA-⌬11/N95 (42) is a dimer, in contrast to previous findings. However, this is not to say that a monomeric form of SecA plays no part in the transport process.
The different structures of SecA all reveal essentially the same nucleotide binding fold. Insofar as we can tell most crystals were grown in the presence of inhibitory Mg 2ϩ concentrations. Therefore the activated state of SecA with a dramatically reduced affinity for ATP that we have identified kinetically, may not be represented in the structural data base.
The concentration of free Mg 2ϩ in bacterial cells is ϳ1 mM (46), sufficient to effectively inhibit the reaction and prevent futile ATP consumption by SecA. The curious observation that cardiolipin is able to counteract the Mg 2ϩ -induced inhibition might reflect a natural function or regulatory process, important for protein translocation. Conditions that promote protein translocation are insensitive to high concentrations of Mg 2ϩ .
SecA is water-soluble, but due to the nature of its activity has an intimate association with the lipid bilayer. It has been reported that acidic lipids and those that do not form lamellae increase SecA ATPase activity (28,29). Moreover, reports show that acidic phospholipids are required for protein translocation (39). There have also been accounts of lipid binding sites on SecA (33,47). The ability of lipids such as cardiolipin to chelate divalent cations (38) might indicate that the activation in ATPase activity is not directly affected by the lipid, but by its potential to extract Mg 2ϩ from the allosteric binding site. This effect cannot be by a simple bulk chelation of Mg 2ϩ , as we observe the activation with only 2.4 M cardiolipin in the presence of 100 M Mg 2ϩ . Instead, the effect is likely to occur upon the formation of a SecA-Mg 2ϩ -cardiolipin ternary complex, in which the Mg 2ϩ may be shifted from the allosteric site by local interaction with the cardiolipin. It is tempting to propose that this kind of regulation might be important in the activation of the enzyme as it delivers substrate protein in the vicinity of the membrane and the Sec complex. Preliminary experiments show that cardiolipin stabilizes the dimeric and active form of the SecYEG complex in E. coli, 3 raising the possibility that this activation could be promoted by a specific cardiolipin encountered only during the initiation of protein translocation (Fig. 9).
The results presented here describe a comprehensive analysis of the steady-state kinetics of the isolated motor domain of the major bacterial protein translocation apparatus. A conformational change in the dimer regulated by Mg 2ϩ brings about a large shift in the properties of the nucleotide binding affinity and turnover of the enzyme, and it is likely that these events are coupled to the protein translocation reaction. More detailed mechanistic understanding requires an analysis of the transient kinetics of ATP binding and hydrolysis that can then be extended to include the protein substrate and SecYEG. FIGURE 9. Schematic overview depicting alleviation of Mg 2؉ -induced inhibition, mediated by cardiolipin. Dimers of SecA and SecYEG are shown as a single object. SecA with Mg 2ϩ bound to an allosteric binding site has low catalytic activity, denoted by the smaller arrow. When SecA is targeted to the membrane, cardiolipin (CL), possibly bound to SecYEG, can alleviate this inhibition, probably by formation of a SecA-Mg 2ϩ -cardiolipin complex. In this ternary complex the cardiolipin shifts the Mg 2ϩ from the inhibitory site, the resulting species having a lower affinity for ATP and an increase in catalytic activity, denoted by the larger arrow. This model does not rule out subsequent dissociation of SecA into monomers, proposed elsewhere (21).