The Unique C-Terminal Extension of Mycobacterial F-ATP Synthase Subunit α Is the Major Contributor to Its Latent ATP Hydrolysis Activity

Mycobacterial F₁F_o-ATP synthases (α₃:β₃:γ:δ:ε:a:b:b′:c₉) are incapable of ATP-driven proton translocation due to their latent ATPase activity. This prevents wasting of ATP and altering of the proton motive force, whose dissipation is lethal to mycobacteria. We demonstrate that the mycobacterial C-terminal extension of nucleotide-binding subunit α contributes mainly to the suppression of ATPase activity in the recombinant mycobacterial F₁-ATPase.

ABSTRACT

Mycobacterial F₁F_o-ATP synthases (α₃:β₃:γ:δ:ε:a:b:b′:c₉) are incapable of ATP-driven proton translocation due to their latent ATPase activity. This prevents wasting of ATP and altering of the proton motive force, whose dissipation is lethal to mycobacteria. We demonstrate that the mycobacterial C-terminal extension of nucleotide-binding subunit α contributes mainly to the suppression of ATPase activity in the recombinant mycobacterial F₁-ATPase. Using C-terminal deletion mutants, the regions responsible for the enzyme’s latency were mapped, providing a new compound epitope.

INTRODUCTION

F₁F_o-ATP synthase is essential in Mycobacterium tuberculosis isolates for the formation of ATP (1, 2) and has become a drug target to fight tuberculosis (3–7). The F₁ domain contains subunits α₃:β₃:γ:ε, the proton-translocating F_o domain (a:c₉), and subunits b:b′:δ holding both domains together (8–10). Rotation of γ:ε connects H⁺ conduction and ATP formation within subunits α₃:β₃. Interestingly, mycobacterial F₁F_o-ATP synthase does not perform ATP hydrolysis-driven proton translocation because of latent ATPase activity (11, 12). Different structural features in the mycobacterial nucleotide-binding subunit α (13, 14) and γ:ε have been proposed to be linked to suppress ATPase activity (11, 12, 15, 16). These features include the extended 3.5-kDa C terminus of subunit α (α_CTD) (14) (Fig. 1), the extra 14-amino-acid γ-loop (12), and the C terminus of subunit ε (15, 16). Understanding these mycobacterial entities resulted in discovery of the mycobacterial F₁F_o-ATP synthase inhibitors GaMF1 (17), epigallocatechin gallate (2) and EpNMF1 (16).

FIG 1.

Amino acid sequence alignment of subunit α of different mycobacterial organisms in comparison with Homo sapiens, Escherichia coli, and G. stearothermophilus. The sequence alignment of subunit α of the following organisms: H. sapiens (UniProt ID P25705-2), E. coli (UniProt ID P0ABB0), G. stearothermophilus (UniProt ID P42005), M. tuberculosis (UniProt ID P9WPU7), M. smegmatis (UniProt ID A0R202), and Mycobacterium bovis (UniProt ID A1KI96) were obtained from the UniProt database (30) and imported into Jalview (31). Alignment of the sequences was performed using ClustalWS (32). Thereafter, the calculation of the percentage of identity was performed and presented in darker to lighter shades of blue, representing the most homologous to the least homologous. As highlighted in red, the C-terminal extension was observed specifically in mycobacteria and not in other species. As previously studied, the α-helix is present from V525 to V538 (according to M. tuberculosis amino acid numbering). For reference, the α-helix present in the C terminus is presented by a green cylinder, and the region showing no secondary structure is denoted by a single black line.

The mycobacterial α_CTD was unresolved in the crystallographic structure (PDB ID 6FOC) (18). Residues 514 to 549 and 540 to 549 of the M. tuberculosis α_CTD were predicted to form a random coil, whereby residues 526 to 539 were determined to form an α-helix (13). The chromosomal deletion mutation of the α_CTD mutant Δα(514–548) stimulated ATP hydrolysis of inverted membrane vesicles (IMVs) (13), whereas fusing the M. tuberculosis α_CTD at the C terminus of subunit α of the Geobacillus stearothermophilus (formerly Bacillus PS3) F₁-ATPase decreased ATPase activity of the hybrid enzyme (13). These data suggest that the mycobacterial α_CTD may play a role in latency.

The 8-fold decrease of recombinant Mycobacterium smegmatis F₁-ATPase compared with its ε-free form α₃:β₃:γ demonstrated an inhibitory effect of subunit ε in ATP hydrolysis (16). However, the α₃:β₃:γ ATPase activity is still significantly lower than the nonlatent α₃:β₃:γ complex, e.g., of G. stearothermophilus [4.9 ± 0.04 μmol min⁻¹ (mg of protein)⁻¹] (13), highlighting a significant contribution of another mycobacterial F₁-ATPase element to latency. To fully comprehend the mechanical system of latency and the role of the α_CTD and its three regions, a systematic assessment using recombinant M. smegmatis F₁-ATPase mutants at the α_CTD was performed.

First, the α_CTD-deleted M. smegmatis F₁-ATPase mutant, MsF₁-α_Δ514-549βγε, was engineered using the recently generated template of the atp genes AGDC, encoding subunits α:β:γ:ε within the pYUB1049 vector (16, 18) and the primers listed in Table S1 in the supplemental material. The linearized pYUB1049 vector was amplified (19), and the two DNA fragments were incorporated as previously published (16). To ease purification, a His₆ tag was added to the N terminus of the β subunit (18). Protein purification was performed as mentioned previously published (16), with an MsF₁-α_Δ514-549βγε in proper stoichiometry and an α_Δ514-549 band running faster than its wild-type (WT) counterpart, revealing the successful deletion (Fig. 2A and B). Subsequently, continuous ATP hydrolysis assay was performed according to previously published methods (16, 20, 21). ATPase activity of 0.05 ± 0.001 μmol min⁻¹ (mg of protein)⁻¹ was calculated for WT MsF₁-ATPase (Fig. 2C, Table 1) and 3.31 ± 0.2 μmol min⁻¹ (mg of protein)⁻¹ for mutant MsF₁-α_Δ514-549βγε, reflecting a >60-fold increase in ATP hydrolysis. The 60-fold increase compared to the 1.7-fold increase of the chromosomal deletion Δα(514–548) mutant in IMVs (13) underlines the need for a defined enzyme to be presented, since endogenous MsF₁F_o ATP synthase and ATP-driven translocators within IMVs effect accurate measures. Interestingly, the ATP hydrolysis rate of MsF₁-α_Δ514-549βγε was significantly higher than that of the ε-free complex [0.63 ± 0.003 μmol min⁻¹ (mg of protein)⁻¹] (16) but comparable to that of the nonlatent G. stearothermophilus F₁-ATPase [4.9 ± 0.04 μmol min⁻¹ (mg of protein)⁻¹] (13). The results suggest that subunit α is a major contributing factor in latent ATP hydrolysis of mycobacterial F₁-ATPase.

FIG 2.

Characterization of the recombinant MsF₁-α_ΔCTD mutants. (A) Fractions from ion exchange were pooled and subjected to size-exclusion chromatography. The recombinant proteins showed consistency in elution at ∼11.6 ml, and their integrity and constituents were confirmed on a 12% SDS-PAGE gel (inset). The subunits are labeled, where α* refers to subunit α and its mutants β, γ, and ε, which correspond to ∼60, 54 , 35, and 10 kDa, respectively. The corresponding proteins are as labeled: lane 1, MsF₁-ATPase; lane 2, MsF₁-α_Δ514-549βγε; lane 3, MsF₁-α_Δ523-549βγε; and lane 4, MsF₁-α_Δ538-549βγε. The purification protocol and 12% SDS-PAGE gel were replicated at least three times, and results represented in the elution diagram and gel remained consistent. (B) Densitometric analysis of the γ to ε ratio of MsF₁-α_Δ514-549βγε revealed a 1:0.3 ratio, identical to that of the WT enzyme (16) and demonstrating the correct stoichiometric subunit ratio. (C) Recombinant mutants were tested for their ATP hydrolysis rate. The decrease in NADH absorption at 340 nm is plotted against the progressing time. MsF₁-α_Δ514-549βγε showed a significant increase in ATP hydrolysis (red triangle). On the other hand, MsF₁-α_Δ523-549βγε (purple diamond) and MsF₁-α_Δ538-549βγε (green asterisk) showed lesser ATP hydrolysis than MsF₁-α_Δ514-549βγε. To calculate the specific activity, the initial rate was used (solid lines), and their calculated specific activities and standard error of regression slope (S_b1) were 3.31 ± 0.18, 1.54 ± 0.03, and 1.33 ± 0.01 μmol min⁻¹ (mg of protein)⁻¹.

TABLE 1.

Summary of specific enzyme activities of the MsF₁-ATPase and its mutants

ATPase	ATP hydrolysis rate [μmol min−1 (mg of protein)−1]	Reference
MsF1-ATPase	0.05 ± 0.001	This study
MsF1-ATPase trypsin treated	1.75 ± 0.018	(16)
MsF1-αβγ	0.63 ± 0.003	(16)
MsF1-αΔ514–549βγε	3.31 ± 0.18	This study
MsF1-αΔ538–549βγε	1.33 ± 0.01	This study
MsF1-αΔ523–549βγε	1.54 ± 0.03	This study

MsF₁-α_Δ523-549βγε and MsF₁-α_Δ538-549βγε were designed to identify whether the random coil regions 514 to 522 and 538 to 549 (18) or the α-helix region 523 to 537 (13) are/is critical for latent ATP hydrolysis (Table S1). Protein purification and ATP hydrolysis assay were performed as described previously. MsF₁-α_Δ523-549βγε and MsF₁-α_Δ538-549βγε (Fig. 2A) showed a similar ∼30-fold increase to that of the WT enzyme (Fig. 2C, Table 1). Compared to the ∼60-fold ATP hydrolysis increase of MsF₁-α_Δ514-549βγε, the 30-fold increase in ATP hydrolysis of MsF₁-α_Δ538-549βγε suggests that the α_CTD residues 538 to 549 and the 514-to-522 region contribute to the suppression of ATPase activity. In contrast, the comparable enzymatic increase in MsF₁-α_Δ523-549βγε reflects that the 523-to-537 region has no major impact on latency. The two-step increase (30- to 60-fold) of MsF₁-α_Δ538-549βγε and MsF₁-α_Δ514-549βγε suggests that both regions may interact with two different mechanistic epitopes of the enzyme. The F₁-ATPase structure of the pathogen Trypanosoma brucei (22), also consisting of an extended subunit of the α C terminus, might illustrate these aspects. As shown in Fig. 3, residues 536 to 539 of this extension form one α-helical turn, followed by a random region (540 to 544) and an α-helix (546 to 558) that come within 7.1 Å of the ADP within the nucleotide binding site. Either a small conformational change may bring R558 in close proximity to ADP or one of the remaining C-terminal residues not resolved in the structure may interact with the nucleotide, thereby stabilizing the ADP-inhibiting state (23). In analogy, we propose that the very C-terminal residues 538 to 549 of mycobacterial subunit α, whose deletion led to a 30-fold ATPase activity increase, come close to the ADP and trap the nucleotide.

FIG 3.

A proposed mechanism of ATP hydrolysis inhibition. (A) Part of the T. brucei F₁-ATPase crystal structure (PDB ID 6F5D ) (22) and a further zoom to highlight the proximity of its extended subunit α C terminus and ADP. The T. brucei C-terminal residues 536 to 539 (red) form an α-helical turn, followed by a random region (540 to 544) and an α-helix (546 to 558) that come close to the ADP. A conformational alteration could bring R558 closer to ADP to generate a hydrogen bond with ADP, or one of the C-terminal residues, not resolved in the structure, could come in proximity to the nucleotide. We predict that the C-terminal residues 538 to 549 of mycobacterial subunit α may come in close proximity to the ADP to stabilize the inhibited state. Subunits α, β, and γ and the T. brucei-specific p18 are shown in green, orange, yellow, and cyan, respectively. The figure was generated via PyMOL (33).

Concerning the second epitope interaction leading to the final 60-fold ATPase activity increase, residue 522 of the mycobacterial stretch 514 to 522 was described to come in proximity with polar residues of γ of the hybrid α^chi₃:β₃:γ complex, thereby decreasing the angular velocity of the power stroke after ATP binding (13).

In conclusion, during evolution, F-ATP synthases have evolved various mechanisms regulating ATP hydrolysis inhibition, including additional features, such as the inhibitory protein (24), subunit ζ (25), the extended C terminus of subunit ε (26–28), or the species-specific extra loop in γ (12, 29). Recent studies proposed cumulative effects of mycobacterial subunits α, γ, and ε to be responsible for suppressed ATP hydrolysis (12, 13, 15). Using defined enzyme complexes, the data presented demonstrate that the mycobacterial α_CTD of subunit α is the major regulator of latent ATP hydrolysis activity, preventing wastage of ATP. Together with the inhibitory mechanisms proposed, the data may contribute to the design of molecules disrupting the interactions of subunit α’s unique C terminus to activate ATPase hydrolysis.