Specific chemical modification explores dynamic structure of the NqrB subunit in Na⁺-pumping NADH-ubiquinone oxidoreductase from Vibrio cholerae

Abstract

The Na⁺-pumping NADH-ubiquinone oxidoreductase (Na⁺-NQR) is a main ion transporter in many pathogenic bacteria. We previously proposed that N-terminal stretch of the NqrB subunit plays an important role in regulating the ubiquinone reaction at the adjacent NqrA subunit in Vibrio cholerae Na⁺-NQR. However, since approximately three quarters of the stretch (NqrB-Met¹_–Pro³⁷) was not modeled in an earlier crystallographic study, its structure and function remain unknown. If we can develop a method that enables pinpoint modification of this stretch by functional chemicals (such as spin probes), it could lead to new ways to investigate the unsettled issues. As the first step to this end, we undertook to specifically attach an alkyne group to a lysine located in the stretch via protein-ligand affinity-driven substitution using synthetic ligands NAS-K1 and NAS-K2. The alkyne, once attached, can serve as an “anchor” for connecting functional chemicals via convenient click chemistry. After a short incubation of isolated Na⁺-NQR with these ligands, alkyne was predominantly incorporated into NqrB. Proteomic analyses in combination with mutagenesis of predicted target lysines revealed that alkyne attaches to NqrB-Lys²² located at the nonmodeled region of the stretch. This study not only achieved the specific modification initially aimed for but also provided valuable information about positioning of the nonmodeled region. For example, the fact that hydrophobic NAS-Ks come into contact with NqrB-Lys²² suggests that the nonmodeled region may orient toward the membrane phase rather than protruding into cytoplasmic medium. This conformation may be essential for regulating the ubiquinone reaction in the adjacent NqrA.

1. Introduction

The Na⁺-pumping NADH-ubiquinone (UQ) oxidoreductase (Na⁺-NQR) is one of the main ion transporters as well as the first enzyme in the respiratory chain of many pathogenic bacteria (such as Vibrio alginolyticus, Vibrio cholerae, and Haemophilus influenzae) and, hence, a promising target for antibiotics [1–3]. Na⁺-NQR consists of six subunits (NqrA–F), encoded by the nqr operon, with a total molecular mass about 200 kDa. This enzyme couples electron transfer from NADH to UQ with Na⁺-pumping, generating an electrochemical Na⁺ gradient across the inner bacterial membrane. Earlier studies have led to the consensus that the electron transfer takes place along a pathway consisting of at least five redox cofactors: from NADH to a FAD, a 2Fe-2S center, two covalently bound FMNs, and a riboflavin, before finally reaching UQ [4–13]. However, the mechanism responsible for Na⁺-pumping driven by electron transfer remains unknown.

An X-ray crystallographic structure of V. cholerae Na⁺-NQR, with no bound UQ or inhibitor, has provided valuable information about the overall structure of the enzyme (Fig. 1A, [14]). Nevertheless, the spatial distances between several donor-acceptor pairs of redox cofactors in the crystallographic model (e.g. between FMN and riboflavin in NqrB) are too long (29–32 Å) to support physiologically relevant electron transfer [14,15]. In particular, the putative binding position of the UQ head-ring in the NqrA subunit is located ∼20 Å above the cytoplasmic membrane surface and too far (∼40 Å in a straight line) from its proximal electron donor, riboflavin, which is modeled to be located between the NqrB and NqrE subunits (Fig. 1B and C). It has, therefore, been suggested that the subunits harboring the cofactors and the binding cavity for the UQ head- ring are need to undergo large conformational rearrangements to reduce these spatial gaps during catalytic turnover [14].

Fig. 1.

X-ray crystallographic structure of V. cholerae Na⁺-NQR. (A) Overall structure of Na⁺-NQR, in which a part of the N-terminal stretch of NqrB (Met¹_–Pro³⁷) was not modeled ([14], PDB ID: 4P6V). (B) Enlargement of the area marked by a red square in panel A. The binding region of the UQ head-ring in NqrA (Leu³²_–Met³⁹ and Phe¹³¹_–Lys¹³⁸) identified in the previous work [16] is indicated in red. The binding sites for different inhibitors in NqrB (Trp²³_–Lys⁵⁴ on the N-terminal stretch (brown) and His¹⁵³_–Gly¹⁵⁸ in the cytoplasmic loop connecting TMHs 2–3 (black)), which were identified in the previous studies [16,17], are shown. The putative binding pocket for the UQ head-ring is marked by a red circle. The riboflavin modeled between NqrB and NqrE is also indicated. (C) The positions of NqrB-Lys⁴², Lys⁵⁴, and Gly¹⁴¹ is shown with the cytoplasmic loop connecting TMHs 2–3 (NqrB-His¹⁵³_–Gly¹⁵⁸, black).

Through photoaffinity labeling experiments, we previously demonstrated that the binding site for the UQ head-ring is located in a cavity in NqrA (NqrA-Leu³²_–Met³⁹ and Phe¹³¹_–Lys¹³⁸, Fig. 1B), as was originally predicted in the crystallographic model [14], and that specific inhibitors of Na⁺-NQR (korormicin and aurachin derivatives, Fig. S1) bind to the cytoplasmic surface of NqrB, at a region that includes a protruding N-terminal stretch (Try²³_–Lys⁵⁴ and His¹⁵³_–Gly¹⁵⁸, Fig. 1B) [16,17]. The efficiencies of the photoaffinity labeling by the inhibitors varied drastically between reduced and oxidized forms of the enzyme, suggesting that this stretch undergoes structural changes that depend on the redox states of the cofactors [17]. Although the binding positions of the UQ head-ring and the inhibitors do not overlap, short-chain UQs such as UQ₁ and UQ₂ competitively suppressed inhibitor binding through obstruction by their side chain moieties [17]. Based on these results, along with the findings obtained by mutagenesis of amino acid residues located in the N-terminal stretch, we proposed that this stretch is functionally critical for regulating the UQ reaction in the adjacent NqrA subunit. The inhibitors may block the structural rearrangements of the stretch that are required for regulating the UQ reaction [17]. However, since as much as three quarters of the stretch (Met¹_–Pro³⁷ from a whole stretch NqrB-Met¹_–Lys⁵⁴) was not modeled in the earlier crystallographic study [14], its spatial position and functional role still need to be explored by various approaches.

For this reason, if we are able to develop a new method that enables pinpoint chemical modification of an amino acid residue located in the N-terminal stretch by functional chemicals such as spin probes (for EPR study) and fluorophores (for single molecule analysis), this could lead to new ways of studying the structure and function of the stretch. In this context, it should be noted that we have succeeded in pinpoint modification of Asp¹⁶⁰ in the 49-kDa subunit in the bovine mitochondrial H⁺-pumping NADH-UQ oxidoreductase (complex I), a residue located inside the reaction cavity of UQ, with a range of functional chemicals via a two-step conjugation reaction [18–21]: first, tosyl chemistry which introduces a chemical “anchor” into 49-kDa-Asp¹⁶⁰ and second, click chemistry to conjugate functional chemicals onto the anchor (Fig. S2A). Tosyl chemistry is one of the protein-ligand affinity-driven substitution reactions which targets a nucleophilic amino acid residue in the protein of interest [22,23]. Click chemistry is a convenient and bioorthogonal method for binding two molecules in aqueous environment by, for example, azide-alkyne [3+2] cycloaddition [23,24]. Achieving pinpoint modification of 49-kDa-Asp¹⁶⁰ by functional chemicals has enabled us to conduct unique chemistry-based investigations for characterizing the structural dynamics of the UQ reaction cavity in complex I [25–27]. On the basis of this experience, here we have undertaken to specifically modify the N-terminal stretch of NqrB by attaching a chemical tag containing an alkyne group via a protein-ligand affinity-driven substitution reaction. The alkyne, once attached, to the stretch can serve as an anchor for connecting secondary functional chemicals under physiological conditions via click chemistry (Fig. S2A).

During the initial course of this study, we found that the tosyl chemistry, which is a slow reaction and, hence, needs a long incubation of the enzyme with the ligand molecule (over 8 h at room temperature), is unsuitable for the isolated V. cholerae Na⁺-NQR because the enzyme undergoes substantial aggregation during the incubation (An example of a ligand molecule synthesized for tosyl chemistry is shown in Fig. S2B). Therefore, we had to search for other protein-ligand affinity-driven substitution reactions, which complete the substitution within a short incubation period of time while maintaining specificity of substitution. Through trial-and-error, we found N-acyl-N-alkyl sulfonamide chemistry (NAS chemistry), a new protein-ligand affinity-driven substitution technique [28], to be suitable for our purpose. A reaction mechanism of NAS chemistry was schematically shown in Fig. S3. To carry out the NAS chemistry, we synthesized two ligand probes: NAS-K1 and NAS-K2 (Fig. 2), which were designed using korormicin A (a very potent inhibitor of V. cholerae Na⁺-NQR [16]) as a molecular template. Using these probes, we succeeded in pinpoint modification of NqrB-Lys²² by a chemical tag containing alkyne, as described hereafter. The modification of NqrB-Lys²² was almost complete within ∼10 min at room temperature. This method allowed us to achieve the specific modification initially aimed for, and also provided valuable information about positioning of the N-terminal stretch, including nonmodeled region (NqrB-Met¹_–Pro³⁷). For example, the fact that the hydrophobic ligands NAS-K1 and NAS-K2 bind to a place that is in contact with NqrB-Lys²² resided at the nonmodeled region strongly suggests that this region may orient toward, or be anchored into, the membrane phase rather than protruding to cytoplasmic medium. Such a positioning of the stretch could influence the conformation of the adjacent NqrA, which harbors the pocket where the UQ head-ring binds. This could result in reduction of the large spatial gap between UQ and riboflavin^NqrB/E, facilitating physiologically relevant electron transfer. In connection with this proposed positioning of the stretch, we will also discuss the mechanism responsible for korormicin A-resistance of the NqrB-G141A mutant.

Fig. 2.

The structures of korormicin A and its ligand probes (NAS-K1 and NAS-K2) synthesized in this study. The averaged IC₅₀ value of each inhibitor, which was determined with 0.90 nM Na⁺-NQR, is shown in the parenthesis.

2. Material and methods

2.1. Materials

UQ₁ and UQ₂ were kind gifts from Eisai (Tokyo, Japan). Protein standards (Precision Plus Protein Standards Dual Xtra) for SDS-PAGE were purchased from Bio-Rad. n-Dodecyl-β-D-maltoside (DDM) and lauryldimethylamine N-oxide (LDAO) were purchased from Dojindo (Kumamoto, Japan) and Sigma-Aldrich, respectively. Other reagents were all of analytical grade.

2.2. Syntheses of NAS-K1 and NAS-K2

The synthetic procedures for NAS-K1 and NAS-K2 are described in the Supplemental Information (Schemes S1 and S2). All compounds were characterized by ¹H and ¹³C NMR spectroscopy and mass spectrometry.

2.3. Plasmid construction and purification of Na⁺-NQR

Site-directed mutants were constructed using the QuikChange Lightning and Quik Change II XL mutagenesis kits (Agilent) as reported before [29]. The sequences of the forward primers are listed in Table 1.

Table 1.

Primers used in this study.

NqrB_K22A_F:	CCAGGCGGCAAACACGAGGCTTGGTTTGCCCTGTATGAA
NqrB_K22A_R:	TTCATACAGGGCAAACCAAGCCTCGTGTTTGCCGCCTGG
NqrB-K42A_F:	CTATACACCAGGTCTGGTAACCGCTAGAAGCTCGCACGTTCGTGAA
NqrB-K42A_R:	TATCACGAACGTGCGAGCTTCTAGCGGTTACCAGACCTGGTGTATG
NqrB-K54A_F:	ACGTTCGTGATAGCGTTGACCTAGCTCGTATCATGATCATGGTTTGGCTT
NqrB-K54A_R:	AAGCCAAACCATGATCATGATACGAGCTAGGTCAACGCTATCACGAACGT

Recombinant wild-type and mutant Na⁺-NQR strains were grown in LB (Miller) medium as reported before [29] in 30-liter cultures in fermenters (New Brunswick BF-5000, Microbiology Core Facility, CBIS, RPI) at 37 °C with constant aeration (20 l per minute) and agitation (300 rpm). The expression of the nqr operon was induced by adding arabinose. Cells were harvested, washed, and broken in buffer (50 mM Tris-HCl (pH 8.0), 300 mM NaCl, 10 mM MgCl₂) in the presence of DNase and a cocktail of protease inhibitors. Membranes were obtained by ultracentrifugation (100,000 ×g) and washed with a medium containing 5.0 mM imidazole, 300 mM NaCl, and 0.05% glycerol.

Wild-type and mutant Na⁺-NQRs were purified by Ni⁺-NTA affinity chromatography and anion exchange DEAE chromatography as reported before [16,30]. For the preparation of LDAO-washed Na⁺-NQR [11,16], the enzyme stock solution (50 μl) was diluted in a 20-fold volume with a washing buffer (50 mM Tris-HCl, 1.0 mM EDTA, 5% glycerol, and 0.05% LDAO, pH 8.0), concentrated with Amicon Ultra 100 K centrifugal filter (Merck-Millipore, Billerica, MA), diluted 20-fold again, and re-concentrated.

2.4. Measurement of the electron transfer activity of Na⁺-NQR

The NADH-UQ₁ oxidoreductase activity of purified Na⁺-NQR was determined by following reduction of UQ₁ at 282 nm (ε ₌ 14.5 mM^{− 1} cm^{− 1}) with a Shimadzu UV-3000 instrument at 30 °C [16]. The reaction medium (2.5 ml) contained 5% glycerol, 0.05% DDM, 1 mM EDTA, 100 mM NaCl, and 50 mM Tris-HCl (pH 8.0). The final enzyme concentration was set to 0.90 nM. The reaction was started by the addition of 100 μM NADH after the incubation of the enzyme with UQ₁ for 1 min. When the effects of inhibitors were examined, the enzyme was incubated with inhibitor for 1 min before the addition of UQ₁.

2.5. Chemical modification of Na⁺-NQR by NAS-Ks

LDAO-washed Na⁺-NQR was suspended in buffer (100 mM NaCl, 5% glycerol, 0.05% LDAO, and 50 mM Tris-HCl, pH 8.0). The enzyme solution (1.8–3.6 μM, 20 μl) was incubated in a 1.5 ml Eppendorf tube with NAS-K1 or NAS-K2 (0.1–7.2 μM) for different periods of times at 37 °C, as indicated in the text. When we conducted competition tests between NAS-Ks and various inhibitors, the inhibitor was added to the enzyme and incubated for 5 min at room temperature prior to the treatment with NAS-Ks. The enzyme treated with NAS-Ks was conjugated with a florescent TAMRA-azido tag via Cu⁺-catalyzed click chemistry using the Click-iT Protein reaction buffer kit (ThermoFisher Scientific) according to the manufacturer’s instructions.

The TAMRA-conjugated proteins were resolved by a Schägger-type 16% SDS gel (16% T, 3% C, [31] or by a Laemmli-type 15% SDS gel containing 6.0 M urea [32]. The migration pattern of fluorescent protein was visualized using bio-imaging analyzer FLA-5100 (Fuji Film, Tokyo, Japan) using a 532 nm light source and an LPG emission filter (575 nm). The data were processed and quantified using MultiGauge software (Fuji Film) and ImageQuant software (Cytiva), respectively. When we discriminate between the fluorescence from a bound FMN (at NqrB-Thr²³⁶) and TAMRA, the gel was monitored under two different scanning conditions: A: 473 nm laser and LPB filter (emission wavelengths shorter than 510 nm are cut off) and B: 532 nm laser and LPG filter (emission wavelengths shorter than 575 nm are cut off). Since the maximum emission wavelength of FMN (λ_ex 448 nm and λ_em 514 nm, [33]) is shorter than that of TAMRA (λ_ex 541 nm and λ_em 565 nm, [34]), the fluorescence from FMN and TAMRA is dominant under condition A and B, respectively. However, as the emission spectra of the two fluorophores are partially overlap (particularly under condition A), the two contributions cannot be completely separated.

When the NqrB modified by NAS-Ks is enriched, the Na⁺-NQR treated with NAS-Ks was biotinylated using a biotin-SS-azido tag (Fig. S5, [18,21]) under the same experimental conditions as for the conjugation of a TAMRA-azido tag above. The biotinylated NqrB subunit was partially purified by SDS-PAGE and enriched by streptavidin-agarose by the procedures reported previously [18,21].

2.6. Proteomic analyses

For partial and exhaustive digestions of the NqrB subunit modified by NAS-K1 or NAS-K2, NqrB was recovered from the SDS gel by direct diffusion. The slice of unfixed SDS gel was vigorously stirred for 12–16 h in a buffer containing 0.1 M NH₄HCO₃, 0.1% (w/v) SDS, and 1% 2-mercaptoethanol. The purified subunit was digested with lysylendopeptidase (Lys-C, Wako Pure Chemicals, Osaka, Japan) or endoprotease Asp-N (Roche Applied Science) in 50 mM Tris-HCl buffer containing 0.1% SDS or 50 mM NaPi buffer containing 0.01% SDS, respectively. These digests were separated on a Schägger-type SDS gel (16% T, 6% C containing 6.0 M urea, [30]), followed by the fluorescent gel imaging. The NqrB subunit was also digested in gel with V8-protease (Roche Applied Science) according to the procedures reported previously [16,35], and the digests were resolved by SDS-PAGE using a 20% Tris-EDTA mapping gel. The digests on the SDS gel were transferred onto a PVDF membrane and stained with 0.025% (w/v) CBB in 40% methanol [36]. Then, their N-terminal amino acid residues were determined with a Procise 491 cLC protein sequencing system (Applied Biosystems) at the Institute for Protein Research, Osaka University (Japan).

For the mass spectrometric analysis of proteins, the NqrB subunit modified by NAS-Ks were in gel digested with trypsin (Promega) or chymotrypsin (Promega) in a buffer containing 25 mM NH₄HCO₃ at 37 °C overnight. The digests were extracted from the gel using a solution containing 50% acetonitrile and 5% aqueous trifluoroacetic acid as reported previously [21,36]. The peptide mixture was characterized using an LTQ Velos Orbitrap mass spectrometer equipped with an Ultimate 3000 nano-LC instrument (LC-MS, Thermo Scientific) under the same experimental conditions reported previously [21]. Data were analyzed using Proteome Discoverer 2.1 (Thermo Scientific) with Mascot 2.3 (Matrix Science). For tryptic and chymotryptic digestions, carbamidomethylation (Cys) was set as a static modification, and oxidation (Met) and modification by NAS-Ks (at Lys, Tyr, and His) were set as dynamic modifications.

Alternatively, the digests were characterized with a Bruker Autoflex III Smartbeam instrument (MALDI-TOF/TOF, Bruker Daltonics), [36]. Peak detection and data processing were performed flex Analysis and Biotools (Bruker Daltonics), respectively. The MS and MS/MS spectra were compared against SwissProt (http://www.expasy.org/sprot) using Mascot.

3. Results

3.1. Molecular design of NAS-K1 and NAS-K2

Many chemical strategies based on a protein-ligand affinity-driven substitution have been used to specifically label a protein of interest with small functional molecules [23]. Although tosyl chemistry is one such strategy [22], this method turned out not to be suitable for the work with the isolated V. cholerae Na⁺-NQR for the reason described above. Tamura et al. [28] recently reported that NAS chemistry allows for rapid and specific modification of a lysine in close proximity to a ligand binding site (Fig. S3), though this method involves a slightly off-target substitution reaction due to the inherently high reactivity of the sulfonamide moiety. To specifically label the isolated Na⁺-NQR with a chemical tag via the shortest possible incubation time, in order to avoid aggregation of the enzyme, we decided to use NAS chemistry in place of tosyl chemistry.

To this end we synthesized two ligand probes: NAS-K1 and NAS-K2, by introducing the N-acyl-N-alkyl sulfonamide moiety into the right-side half of korormicin A (as shown in Fig. 2) because the structural framework of the left half of the molecule (methyl/ethyl branch at the 5S-position, OH group at the 3^′R-position, and epoxy group at 9^′S/10^′R) is essential to retain strong binding affinity to Na⁺-NQR [17]. The IC₅₀ values of NAS-K1 and NAS-K2, determined in the NADH-UQ₁ oxidoreductase assay using 0.90 nM Na⁺-NQR, are 130 (±16) and 15 (±3.0) nM, respectively, corroborating the importance of the epoxy group for inhibitory action. While the binding affinities of both ligands, in terms of the IC₅₀ values, were weaker compared to natural korormicin A (IC₅₀ = 5.0 nM), we used them in the following experiments.

3.2. Chemical modification of Na⁺-NQR by NAS-Ks

Isolated wild-type V. cholerae Na⁺-NQR (1.8 μM) was incubated with NAS-K1 or NAS-K2 (1.0 μM each) for 10 min at 37 °C to attach an alkyne group via NAS chemistry. This was followed by conjugation with a fluorescent TAMRA-azido tag (Fig. S2C) via Cu⁺-catalyzed click chemistry to visualize the labeled protein(s). The samples treated with NAS-K1 and NAS-K2 showed a strong fluorescent band at ∼30 kDa on the SDS-PAGE gel, corresponding to the NqrB subunit (Fig. 3A). Although other five subunits were also slightly modified due to the high electrophilic reactivity of the sulfonamide group, only the modification of NqrB was significantly suppressed in the presence of 40-fold molar excess of korormicin A or aurachin D-42 (Figs. 3A and S1). These results indicate that NAS-Ks predominantly modify lysine residue(s) located on NqrB, which harbors the inhibitor-binding site in Na⁺-NQR [16,17].

Fig. 3.

Chemical modification of Na⁺-NQR by NAS-K1 and NAS-K2. (A) The isolated Na⁺-NQR (1.8 μM) was treated with NAS-K1 or NAS-K2 (1.0 μM each) for 10 min at 37 °C in the absence or presence of korormicin A or aurachin D-42 (40 μM each). (B) Na⁺-NQR (3.6 μM) was modified by NAS-K1 or NAS-K2 (2.0 μM each) for different incubation periods of time (0.1–120 min) at 37 °C. (C) Na⁺-NQR (3.6 μM) was modified by different concentrations of NAS-K1 or NAS-K2 (0.1–7.2 μM) for 10 min at 37 °C. All data are representative of three independent experiments.

We examined the time dependence of the modification by NAS-Ks. The modification progressed rapidly and was almost complete within 10 min using a fixed concentration of NAS-Ks (2.0 μM) (Fig. 3B). The concentration dependence of the modification was also examined using a fixed incubation time (10 min) (Fig. 3C). The extent of modification of NqrB increased with increasing concentrations of NAS-Ks, but the modification of other off-target subunits also increased. Nevertheless, 80–85% of the total fluorescence of six subunits was detected in NqrB under the varying experimental conditions above. This means that the modification specificity for NqrB is ∼20-fold greater compared to any of the other five subunits, based on a simple calculation.

The fluorescence intensities in the left (NAS-K1) and right (NAS-K2) panels in Fig. 3B and C can be directly compared because they were separately presented from an identical imaging picture of SDS-PAGE gel. Although NAS-K2 exhibited stronger binding affinity to the enzyme than NAS-K1 (in terms of the IC₅₀ values), the modification efficiency of NAS-K1 was similar with or slightly better than that of NAS-K2 under varying experimental conditions. This result may not be peculiar because the modification efficiency is determined not only by the binding affinity of the whole ligand molecule but also by local spatial orientation of the electrophilic sulfonamide moiety and the nucleophilic lysine(s). Although the modification yields of NqrB by NAS-Ks varied according to the experimental conditions, the average yields of NqrB (with 1.8 μM Na⁺-NQR) by NAS-K1 and NAS-K2 (1.0 μM each) after 10 min incubation at 37 °C, which were estimated by the previous method [37], are 13 and 10%, respectively.

3.3. Localization of the modified sites by NAS-Ks in NqrB

To localize the site modified by NAS-Ks in NqrB, the modified NqrB was isolated from the SDS-PAGE gel. The protein was then treated with increasing amounts of lysylendopeptidase (Lys-C) for an hour, followed by resolution of the digests on a Schägger-type SDS gel (Fig. 4A). To discriminate between the fluorescence from a bound FMN (at NqrB-Thr²³⁶, [38]) and from TAMRA, we used two monitoring methods (two different lasers and two detection filters, see Experimental Procedures). Since the maximum emission wavelength of FMN (λ_ex 448 nm and λ_em 514 nm, [33]) is shorter than that of TAMRA (λ_ex 541 nm and λ_em 565 nm, [34]), the fluorescence from FMN and TAMRA is dominant in panel “FMN” and panel “TAMRA”, respectively. Note, however, that as the emission spectra of the two fluorophores partially overlap, two contributions cannot be completely separated.

Fig. 4.

Localization of the region modified by NAS-K1 and NAS-K2 in NqrB. (A) The wild-type Na⁺-NQR (3.6 μM) labeled by NAS-K1 or NAS-K2 (7.2 μM) was separated on a 16% Schägger-type SDS gel (16% T, 3% C), followed by isolation of NqrB by electroelution. The modified NqrB subunit was partially digested with Lys-C, and the digests were resolved by SDS-PAGE using a 16% Schägger-type SDS gel (16% T, 6% C containing 6.0 M urea). The SDS gel was scanned using a two-color method (see Experimental Procedures). (B) The NqrB modified by NAS-K1 was partially digested with V8-protease, and the digests were resolved by SDS-PAGE using a 20% Tris-EDTA mapping gel. The SDS gel was monitored as above. (C) The NqrB subunit labeled by NAS-K1 or NAS-K2 was exhaustively digested with Lys-C or Asp-N, and the digests were resolved by SDS-PAGE using a 16% Schägger-type SDS gel (16% T, 6% C containing 6.0 M urea). The SDS gel was monitored as above. (D) Schematic presentation of the digestion of the NqrB subunit by Lys-C, V8-protease, or Asp-N. The predicted cleavage sites are denoted by arrows and marked with their residue numbers in the sequences of the V. cholerae NqrB subunit (SwissProt entry: Q9KPS2). The N-terminal sequences of segments L2 (R⁴³SSHV), L3 (R⁴³SSHV), and L4 (M¹²²LLGA), which were determined by Edman degradation, are indicated in italics. All data are representative of three to four independent experiments.

The partial Lys-C digestion of NqrB modified by NAS-K1 and NAS-K2 gave three sequential partial digests containing FMN in the range ∼20–35 kDa (bands A1, A2, and A3 in Fig. 4A). A strong TAMRA fluorescence band appeared at ∼5 kDa (band A5) for NAS-K1 and NAS-K2 (Fig. 4A). An additional minor TAMRA fluorescence band was observed for NAS-K1 at ∼8 kDa (band A4) but not for NAS-K2. Note that the exhaustive Lys-C digestion of the NqrB modified by NAS-K1 reproducibly provided this minor band, as described below (band C3 in Fig. 4C), and this minor band contains modified Lys⁵⁴, as described in the next section. Based on the N-terminal sequences of bands A1 (R⁴³SSHV), A2 (M¹²²LLGA), and A4 (R⁴³SSHV) determined by Edman degradation, bands A1, A2, and A4 can be assigned to segments L3 (Arg⁴³_–Gln⁴¹⁵, 40.7 kDa), L4 (Met¹²²_–Gln⁴¹⁵, 31.8 kDa), and L2 (Arg⁴³_–Lys¹²¹, 8.7 kDa), respectively (“Lys-C” in Fig. 4D). Unfortunately, we have not been able to determine the N-terminal sequences of bands A3 and A5.

On the other hand, partial V8-protease digestion of the NqrB modified by NAS-K1 (i.e. Cleveland mapping) provided three fluorescent bands from FMN in the range ∼20–30 kDa (bands B1–B3 in Fig. 4B). Considering the theoretical cleavage sites for V8-protease (“V8-protease” in Fig. 4D), these bands may be assigned to V1 (Val¹⁴⁵_–Glu⁴⁰² (27.6 kDa) or Gly¹⁵⁸_–Glu⁴¹⁵ (27.6 kDa)), V2 (Val¹⁴⁵_–Glu³⁸⁰ (24.1 kDa) or Gly¹⁵⁸_–Glu⁴⁰² (24.9 kDa)), and V3 (Val¹⁹³_–Glu³⁸⁰ (22.4 kDa)), respectively. Two strong fluorescent bands derived from TAMRA were observed at ∼12 and ∼10 kDa (bands B7 and B8 in Fig. 4B). These bands should be peptides containing the Ala²⁹_–Glu¹⁰⁶ fragment (8.7 kDa) and including some additional amino acids resulting from one or two miss cleavage site(s); namely, fragments V7 (Met¹_–Glu¹⁰⁶ (12.1 kDa)) and V8 (Lys²²_–Glu¹⁰⁶ (9.7 kDa)), respectively (“V8-protease” in Fig. 4D). A fluorescent band from FMN at ∼12 kDa (band B4), which overlaps with that from TAMRA, may be the fragment V4 (Gly¹⁵⁸_–Glu²⁷⁴ (12.1 kDa)). Taken together, the results obtained by partial digestion using Lys-C and V8-protease strongly suggest that the site(s) modified by NAS-Ks are located at the N-terminal region of NqrB rather than the middle to C-terminal regions that hold the bound FMN (at Thr²³⁶).

Exhaustive digestion of the modified NqrB by Lys-C (Fig. 4C) gave comparable migration patterns with those observed for the partial digestion above (Fig. 4A); that is, a strong TAMRA fluorescence band was observed at ∼5 kDa for NAS-K1 and NAS-K2 (band C1), although NAS-K1 gave an additional weak band at ∼7 kDa (band C3). Exhaustive Asp-N digestion provided a ∼5 kDa-band (band C2) for both NAS-K1 and NAS-K2, which is consistent with the theoretical fragmentation pattern (Asp⁹_–Arg⁴⁸ (4.6 kDa), Asp⁹_–Val⁵¹ (4.9 kDa (band Asp1)), or Asp⁵²_–Gly⁸⁹ (4.3 kDa), “Asp-N” in Fig. 4C). Taking all of the results together, although we were unable to determine the N-terminal sequence of band A5 (equivalent to band C1) by Edman degradation, it is reasonable to conclude that NAS-K1 and NAS-K2 modify one or more lysines in the peptide Asp⁹_–Val⁵¹ (Lys¹⁹, Lys²² and/or Lys⁴²), which corresponds to the protruding N-terminal stretch. NAS-K1 also modifies a lysine in the peptide Arg⁴³_–Lys¹²¹ (Lys⁵⁴, band L2), albeit to a much lesser extent.

In the above peptide mapping, we excluded NqrB-Lys⁴ and -Lys⁵ as the candidate of modified residue. To verify this, the N-terminal sequences of the whole NqrB subunit modified by NAS-K1, which was enriched using biotin-SS-azido according to the procedures illustrated in Fig. S4 [18,21], were determined by Edman degradation. The sequences were found to be G²LKKF⁶ as standard phenylthiohydantoin derivatives; therefore, neither NqrB-Lys⁴ nor -Lys⁵ are modified by NAS-K1.

3.4. Identification of the lysine(s) modified by NAS-Ks

Since the lysine(s) modified by NAS-Ks could be localized to the regions Asp⁹_–Val⁵¹ (the major site containing Lys¹⁹, Lys²², and Lys⁴²) and Arg⁴³_–Lys¹²¹ (the minor site for NAS-K1 containing Lys⁵⁴), we characterized the tryptic and chymotryptic digests of the NqrB subunit, modified by NAS-K1, by Orbitrap LC-MS/MS. The analyses identified 4 tryptic and 28 chymotryptic digests with sequence coverages of 9 and 53%, respectively (Tables S1 and S2, Fig. S5). Among the tryptic digests, we detected a doubly-charged ion at m/z 456.743 (z = 2), which matches the calculated mass of the modified peptide D⁴⁹SVDLKR⁵⁵ (m/z 416.730 + 40.013 (z = 2)) (Fig. 5). The fragment spectra of the modified peptide confirmed the sequence D⁴⁹SVDLKR⁵⁵ with the modification at Lys⁵⁴ (Fig. 5), corresponding to the minor NAS-K1 modification site (a segment L2 in “Lys-C” of Fig. 4D).

Fig. 5.

Identification of the labeled lysine in NqrB. The tryptic digests of the NqrB modified by NAS-K1 were characterized by Orbitrap LC-MS/MS. The fragment ion spectra of the control and the modified peptide D⁴⁹SVDLKR⁵⁵. K* (in red) corresponds to the modified lysine with a +40.013 (z = 2) mass shift. The identified b- and y-fragment ions were mapped onto the amino acid sequence.

To analyze the major site, NqrB treated with NAS-K1 was enriched using biotin-SS-azido (Fig. S4, [18,21]). We detected a triply-charged ion at m/z 552.27 (z = 3) from the tryptic digests both of the NAS-K1 treated (and enriched) and nontreated NqrB, which matches the calculated mass of a nonmodified peptide Phe⁶_–Lys¹⁹ (Fig. S4). This result indicates that Lys¹⁹ is unlikely to be modified by NAS-K1. Unfortunately, no peptide containing Lys²² and/or Lys⁴² was found in either the tryptic or chymotryptic digests due to low sequence coverage (Fig. S5). Nevertheless, the likelihood of occurrence of modification at Lys⁴² could be low because the A4 band (i.e. a segment L2 (Arg⁴³_–Lys¹²¹) in Fig. 4D) may not be generated if the adjacent Lys⁴² is modified.

To clarify these issues, we constructed the mutants NqrB-K22A, -K42A, and -K54A and performed chemical modification by NAS-Ks. Before conducting the modification, we measured the NADH-UQ₁ oxidoreductase activities of the isolated mutant enzymes. Concentrations of NADH and UQ₁ were set to 100 and 50 μM, respectively, because these result in an apparently maximum activity. Although the enzyme activity of NqrB-K22A was comparable to that of the wild-type enzyme, the activities of mutants NqrB-K42A and -K54A decreased to ∼50 and ∼30% of the wild-type enzyme, respectively (Table 2). These results suggest that NqrB-K42A and -K54A mutation may induce structural changes of the cytoplasmic interfacial area between the NqrB and NqrA subunits.

Table 2.

Relative catalytic activity of the mutated enzyme.

Enzyme	Relative enzyme activity (%)a
Wild-type	100
NqrB-K22A	100 (±16)
NqrB-K42A	54 (±5)
NqrB-K54A	30 (±5)

^aThe catalytic enzyme activity was determined in the NADH-UQ₁ oxidoreductase assay. The averaged activity of the wild-type enzyme was 5.1 (± 0.3) mmol UQ₁/min/μmol of Na⁺-NQR. The concentrations of Na⁺-NQR, NADH, and UQ₁ were set to 0.20 μg/ml (0.90 nM), 100 μM, and 50 μM, respectively. Values in table are means ± S.E. (n = 3).

Next, each of the isolated mutant was treated with NAS-K1 or NAS- K2 for 10 min at 37 °C and the results of modification are summarized in Fig. 6. The extents of modification by NAS-Ks of the NqrB-K22A mutant, in which the predicted major target lysine was mutated (to Ala), significantly decreased but did not completely disappear since the minor target NqrB-Lys⁵⁴ remains. We cannot exclude the possibility that the modification of NqrB-Lys⁵⁴ was rather enhanced due to a lack of the nearby reactive NqrB-Lys²². The extent of modification of the NqrB-K54A mutant by NAS-K1 also significantly decreased (Fig. 6). Given that NqrB-Lys⁵⁴ is the minor target of NAS-K1, the decrease seemed to be greater than expected. This is probably because the modification of NqrB-Lys²² may be somewhat reduced by the structural changes of the N-terminal stretch induced by the mutation. The comparable decrease of the modification by NAS-K2, which does not react with NqrB-Lys⁵⁴ (Fig. 4C), may support this notion. For the NqrB-K42A mutant, the modification was hardly affected, indicating that NqrB-Lys⁴² is not the target lysine, as discussed above. Taken all of the results together, we conclude that NqrB-Lys²² is the target lysine modified by NAS-K1 and NAS-K2 in the major region NqrB-Asp⁹_–Val⁵¹ (a fragment “Asp1” in Fig. 4D).

Fig. 6.

The chemical modification of the mutants NqrB-K22A, -K42A, and -K54A. The isolated wild-type and mutated enzymes (3.6 μM each) were modified by NAS-K1 or NAS-K2 (3.6 μM each) for 10 min at 37 °C. The fluorescence intensities of the modified NqrB were compared. The extent of the modification of the wild-type enzyme by NAS-Ks was used as a control (100%). Values in graphs are means ± S.E. (n = 3). **P < 0.01 compared with control (one-way ANOVA followed by Dunnett’s test).

3.5. Chemical modification of the korormicin-resistant NqrB-G141A by NAS-Ks

We previously showed that the V. cholerae NqrB-G141A mutant confers significant resistance against korormicin A (∼160-fold) [16]. As this residue lies in transmembrane helix (TMH) 2 under the cytoplasmic loop connecting TMH2–3 (His¹⁵³_–Gly¹⁵⁸) (Fig. 1C), it is some distance from the N-terminal stretch (the binding site of korormicin A) and not actually part of the binding site of the UQ head-ring, which is located in NqrA [14,16]. Interestingly, the photoaffinity labeling experiments with this mutant enzyme using a photolabile korormicin derivative ([¹²⁵I] PKRD-1, Fig. S1) found that the labeling efficiency was essentially unchanged compared with the wild-type enzyme; that is, there was no loss of its binding affinity to the mutant [17]. Therefore, it remains to be elucidated why the ability of korormicin A to block electron transfer from riboflavin^NqrB/E to UQ decreases remarkably in the NqrB-G141A mutant. We cannot exclude the possibility that the mutation at the NqrB-141 position may influence the structure of the cytoplasmic interfacial area between NqrA and NqrB, including the N-terminal stretch, in spite of the fact that NqrB-141 is relatively far from the stretch (Fig. 1C).

To understand the effects of the mutation on the structure of the N-terminal stretch, we compared the modification efficiencies of NqrB by NAS-Ks between the wild-type and NqrB-G141A enzymes. With both NAS-K1 and NAS-K2, the modification efficiencies were significantly lower with NqrB-G141A compared with the wild-type enzyme (Fig. 7A). It is difficult to determine from only this result whether the decreases are due to lower binding affinities of the ligand molecules to the mutated enzyme or whether they are due to changes in microenvironment around the reaction point (NqrB-Lys²²) that could directly affect the reaction efficiency without significant effects on the binding affinities of the ligands. However, considering that our previous study found no loss of binding affinity of [¹²⁵I]PKRD-1 to NqrB-G141A compared to the wild-type enzyme [17], as mentioned above, the likelihood of latter is high.

Fig. 7.

The chemical modification patterns were compared between the wild- type and NqrB-G141A enzymes. (A) The isolated wild-type and NqrB-G141A enzymes (3.6 μM each) were modified by NAS-K1 or NAS-K2 (7.2 μM each) for 10 min at 37 °C. The fluorescence intensities of the modified NqrB were compared. The extent of the modification of the wild-type enzyme by NAS-Ks was used as a control (100%). Values in graphs are means ± S.E. (n = 3). **P < 0.01 compared with control (one-way ANOVA followed by Dunnett’s test). (B) The NqrB subunit modified by NAS-K1 was exhaustively digested with Lys-C, and the digests were resolved by SDS-PAGE using a 16% Schägger-type SDS gel (16% T, 6% C containing 6.0 M urea). The experimental conditions are identical to those shown in Fig. 4C.

To further examine the effects of the NqrB-G141A mutation on the modification of NqrB by NAS-K1, the purified mutant enzyme was first reacted with NAS-K1 and the NqrB subunit was isolated using a Schagger-type SDS gel. The subunit was then digested by Lys-C, and ¨ fluorescent intensities of the digests were compared with the case of the wild-type enzyme (Fig. 7B). Interestingly, although in wild-type, NqrB- K⁵⁴ (∼9 kDa-band, equivalent to band C3 in Fig. 4C) was modified more weakly than NqrB-K²² (∼5 kDa-band, equivalent to band C1 in Fig. 4C), in the NqrB-G141A mutant NqrB-K⁵⁴ is consistently modified more strongly than NqrB-K²². This indicates a change in relative reactivity between the sulfonamide moiety and the lysines in the N-terminal stretch in the mutant. Altogether, these results strongly suggest that the structural change due to the mutation at the NqrB-141 position dynamically propagates to the “upper” region in NqrB (i.e. the cytoplasmic surface region, Fig. 1C), resulting in some structural rearrangement in this region including the N-terminal stretch. This long-rage structural rearrangement due to the mutation is likely connected with the korormicin-insensitive electron transfer from riboflavin^NqrB/E to UQ in the NqrB-G141A mutant, as discussed below.

4. Discussion

We previously performed photoaffinity labeling experiments using UQ and inhibitor derivatives to identify the binding sites for these ligands in V. cholerae Na⁺-NQR [16,17]. Based on the results of these experiments in wild-type and site-specific mutants, we proposed that the protruding N-terminal stretch of NqrB is functionally critical for regulating the UQ reaction in the adjacent NqrA subunit [17]. Nevertheless, since approximately three quarters of the stretch (NqrB-Met¹_–Pro³⁷) was not modeled in the crystallographic study [14], its spatial position and functional role remain unknown. To develop new ways to address these unsettled issues, here we undertook to develop a new method that enables pinpoint modification of a lysine residue located on the N-terminal stretch with a small chemical tag containing alkyne via NAS chemistry [28]. NAS-K1 and NAS-K2 were synthesized as the high affinity ligands for conducting NAS chemistry. Detailed analyses of modification in the wild-type enzyme and mutants (NqrB-K22A, -K42A, and -K54A) revealed that the alkyne group of the ligands predominantly attaches to NqrB-Lys²², and to a lesser extent to NqrB-Lys⁵⁴ for NAS-K1 (a pattern reversed in the NqrB-G141A mutant as discussed later). Thus, we succeeded in the specific modification (alkynylation) of the NqrB-Lys²² residue, located in the N-terminal stretch, simply by incubating Na⁺-NQR with the ligands. It is, therefore, likely that secondary conjugation of varying functional reagents to the alkyne attached to NqrB-Lys²² via convenient click chemistry has come into our view. For example, if a nitroxide spin probe can be connected to the alkyne, we may be able to obtain the information about dynamic conformation of the N-terminal stretch by determining spin-spin interaction between the attached spin probe and the stable neutral radical of riboflavin^NqrB/E [39] using various EPR spectroscopic techniques. For such experiments to succeed, we need to further enhance the modification efficiency by the NAS chemistry ligands and also regulate the position of the target lysine on the N-terminal stretch (besides Lys²²). To this end, we are currently synthesizing NAS ligands with a wide range of structural alterations and studying their modification properties.

Based on the X-ray crystallographic structure of V. cholerae Na⁺-NQR [14], the spatial distance between the binding position of the UQ head- ring in NqrA and riboflavin^NqrB/E is too long (∼40 Å in a straight line) for physiologically relevant electron transfer [15] (Fig. 1B). Therefore, this large spatial gap must be reduced during catalytic turnover. For this to happen, the cytoplasmic interfacial area between NqrA and NqrB may need to undergo substantial structural rearrangement [14]. We previously proposed that the protruding N-terminal stretch of NqrB, including the nonmodeled region, may anchor the binding region of the UQ head- ring in NqrA to the membrane phase under physiological conditions, and that inhibitors may block the structural rearrangements necessary for the electron transfer [17]. The results of the present study corroborate this proposition. Namely, the findings that hydrophobic ligands NAS-K1 and NAS-K2 come into contact with NqrB-Lys²² located in the nonmodeled region (NqrB-Met¹_–Pro³⁷) and, additionally, that NAS-K1 comes into contact with both NqrB-Lys²² and NqrB-Lys⁵⁴ strongly suggest that the nonmodeled region may orient toward, or be anchored into, the membrane phase rather than protruding to the cytoplasmic medium. The hypothetical conformation of the N-terminal stretch under the physiological conditions is shown schematically in Fig. 8 (right hand side). This positioning of the stretch could involve structural changes of the region harboring the binding pocket for the UQ head-ring in NqrA. This notion is supported by the previous [17] and present mutagenesis experiments, which indicated that the mutations of NqrB-Lys⁴², -Asp⁴⁹, -Asp⁵², and -Lys⁵⁴ (to Ala), all of which are located in the N-terminal stretch, resulted in moderate decreases in the catalytic enzyme activity. Based on the hypothetical model, the spatial gap between the UQ head- ring and riboflavin^NqrB/E may be reduced and the highly hydrophobic natural UQ₈ may not need to extensively protrude from the membrane lipid phase into the cytoplasmic medium, which is energetically unfavorable [40].

Fig. 8.

A schematic model of the conformation of the protruding N-terminal stretch of NqrB. Left: Schematic presentation of the crystallographic structure [14], in which the region NqrB-Met¹_–Pro³⁷ was not modeled. Right: The hypothetical conformation of the stretch under the physiological conditions. Relative positions of key residues Lys²², Lys⁵⁴, and Gly¹⁴¹ are indicated. Q (red); the binding pocket for the UQ head-ring, RBF (red); riboflavin^NqrB/E, and I (blue lozenge); bound inhibitor (i.e. NAS-Ks).

We previously demonstrated that the V. cholerae NqrB-G141A mutation confers remarkable resistance against korormicin A and its derivatives, whereas the binding affinities of these inhibitors are essentially unchanged compared with the wild-type enzyme [16,17]. This finding raises the critical question of why korormicins do not block electron transfer from riboflavin^NqrB/E to UQ in this mutant in spite of the fact that they bind with same high affinity as in the wild type. To obtain a clue to this question, we compared the modification by NAS-Ks between the wild-type and NqrB-G141A enzymes (Fig. 7). Both the efficiencies and patterns (NqrB-Lys²² vs. -Lys⁵⁴) of modification were considerably different between the two enzymes. The results strongly suggest that the structural change due to the amino acid substitution at the NqrB-141 position propagates to the cytoplasmic surface area of NqrB (probably through the cytoplasmic loop connecting TMHs 2–3 (NqrB-His¹⁵³_–Gly¹⁵⁸), Fig. 1C), resulting in a structural rearrangement of the N-terminal stretch. The structural model proposed above (Fig. 8) could account for this long rage interaction between the NqrB-141 position and the N-terminal stretch. The structural rearrangement of the stretch could also influence the structure of the adjacent NqrA which harbors the reaction pocket for the UQ head-ring. Thus, the NqrB-G141A mutation may induce structural changes of the cytoplasmic interfacial area between NqrB and NqrA. We previously demonstrated that, although the binding positions of the UQ head-ring (in NqrA) and korormicins (in NqrB) do not overlap, binding of short-chain UQs and korormicins to the enzyme are competitive due to obstruction of their side chain moieties [17]. Taking all of these findings together, the long rage structural rearrangement induced by the NqrB-G141A mutation could enable the UQ head-ring to gain access to the reaction site even in the presence of bound korormicin A.

In conclusion, we succeeded in the pinpoint chemical modification of NqrB-Lys²², resided at the nonmodeled region of the N-terminal stretch, via NAS chemistry using NAS-K1 and NAS-K2. The findings that these hydrophobic ligands come into contact with NqrB-Lys²² and, additionally, that NAS-K1 comes into contact with both NqrB-Lys²² and NqrB- Lys⁵⁴ strongly suggest that the nonmodeled region of the stretch may orient toward the membrane phase rather than protruding to the cytoplasmic medium (Fig. 8). This predicted positioning of the stretch could be connected with the remarkable resistance of the NqrB-G141A mutant against korormicin A.

Abbreviations:

Asp-N

endoprotease Asp-N

CBB

Coomassie Brilliant Blue

IC50

the molar concentration needed to reduce the control electron transfer by 50%

DDM

n-dodecyl-β-D-maltoside

FMN

flavin mononucleotide

LDAO

lauryldimethylamine N-oxide

Lys-C

lysylendopeptidase

NADH

nicotinamide adenine dinucleotide

Na+-NQR

Na+-pumping NADH-ubiquinone oxidoreductase

TMH

transmembrane helix

UQ

ubiquinone

UQn

ubiquinone-n

SDS-PAGE

sodium dodecyl sulfate-polyacrylamide gel electrophoresis