Two uptake hydrogenases differentially interact with the aerobic respiratory chain during mycobacterial growth and persistence

To persist when nutrient sources are limited, aerobic soil bacteria metabolize atmospheric hydrogen (H2). This process is the primary sink in the global H2 cycle and supports the productivity of microbes in oligotrophic environments. H2-metabolizing bacteria possess [NiFe] hydrogenases that oxidize H2 to subatmospheric concentrations. The soil saprophyte Mycobacterium smegmatis has two such [NiFe] hydrogenases, designated Huc and Hhy, that belong to different phylogenetic subgroups. Both Huc and Hhy are oxygen-tolerant, oxidize H2 to subatmospheric concentrations, and enhance bacterial survival during hypoxia and carbon limitation. Why does M. smegmatis require two hydrogenases with a seemingly similar function? In this work, we resolved this question by showing that Huc and Hhy are differentially expressed, localized, and integrated into the respiratory chain. Huc is active in late exponential and early stationary phases, supporting energy conservation during mixotrophic growth and transition into dormancy. In contrast, Hhy is most active during long-term persistence, providing energy for maintenance processes following carbon exhaustion. We also show that Huc and Hhy are obligately linked to the aerobic respiratory chain via the menaquinone pool and are differentially affected by respiratory uncouplers. Consistently, these two enzymes interacted differentially with the respiratory terminal oxidases. Huc exclusively donated electrons to, and possibly physically associated with, the proton-pumping cytochrome bcc–aa3 supercomplex. In contrast the more promiscuous Hhy also provided electrons to the cytochrome bd oxidase complex. These results indicate that, despite their similar characteristics, Huc and Hhy perform distinct functions during mycobacterial growth and survival.

groups has revealed that this net H 2 consumption is mediated by aerobic soil bacteria (3)(4)(5)(6)(7)(8). Based on this work, it has been established that gas-scavenging bacteria are the major sink in the global H 2 cycle, responsible for the net consumption of ϳ70 million tons of H 2 each year and 80% of total atmospheric H 2 consumed (6,(9)(10)(11). In addition to its biogeochemical importance, it is increasingly realized that atmospheric H 2 oxidation is important for supporting the productivity and biodiversity of soil ecosystems (12)(13)(14)(15)(16)(17)(18)(19)(20). This process is thought to play a key role under oligotrophic conditions, where the majority of microbes exist in a nonreplicative, persistent state (14,21). As the energy requirements for persistence are ϳ1000-fold lower than for growing cells (22), the energy provided by atmospheric H 2 can theoretically sustain up to 10 8 cells per gram of soil (23).
The genetic basis of atmospheric H 2 oxidation has largely been elucidated. Two distinct subgroups of hydrogenase, group 1h and 2a [NiFe] hydrogenases, are known to oxidize H 2 to subatmospheric concentrations (3,5,24). The operons for these hydrogenases minimally encode the hydrogenase large subunit containing the H 2 -activating catalytic center, the hydrogenase small subunit containing electron-relaying ironsulfur clusters, and a putative iron-sulfur protein hypothesized to have a role in electron transfer (23,(25)(26)(27). Additional operons encode the maturation and accessory proteins required for hydrogenase function (13,27). Increasing evidence suggests that hydrogenases capable of atmospheric H 2 oxidation are widely encoded in soil bacteria. Representatives of three dominant soil phyla, Actinobacteriota, Acidobacteriota, and Chloroflexota, have been shown experimentally to oxidize atmospheric H 2 (3-5, 8, 24, 28, 29). Moreover, genomic and metagenomic studies indicate that at least 13 other phyla possess hydrogenases from lineages known to support atmospheric H 2 oxidation (13,14,19,30).
The saprophytic soil actinobacterium Mycobacterium smegmatis has served as a key model organism for these studies (24,27,31). In M. smegmatis, H 2 oxidation has been shown to be solely mediated by two oxygen-tolerant hydrogenases: the group 2a [NiFe] hydrogenase Huc (also known as Hyd1 or cyanobacterial-type uptake hydrogenase) and the group 1h [NiFe] hydrogenase Hhy (also known as Hyd2 or actinobacterial-type uptake hydrogenase) (13,27). These enzymes belong to distinct phylogenetic subgroups, and their large subunits share less than 25% amino acid identity (13). Despite this, Huc and Hhy display striking similarities. Both enzymes oxidize H 2 to subatmospheric concentrations under ambient conditions (24), and both appear to be membrane-associated despite a lack of predicted transmembrane regions (24). Both Huc and Hhy are reported to be up-regulated during stationary phase in response to both carbon and oxygen limitation (27). Consistently, Huc and Hhy deletion mutants show reduced growth yield and impaired long-term survival, suggesting that atmospheric H 2 oxidation supports energy and redox homeostasis (24,31,32). Nevertheless, some evidence suggests that these enzymes are not redundant. For reasons incompletely understood, significant survival phenotypes are observed for both single and double mutants (24,27). In whole cells, the enzymes also exhibit distinct apparent kinetic parameters, with Hhy having higher affinity but lower activity for H 2 compared with Huc (24).
It remains to be understood whether and how the hydrogenases of M. smegmatis are integrated into the respiratory chain. As an obligate aerobe, M. smegmatis depends on aerobic heterotrophic respiration to generate proton-motive force and synthesize ATP for growth (33). M. smegmatis possesses a branched respiratory chain terminating in one of two terminal oxidases, the cytochrome bcc-aa 3 supercomplex or the cytochrome bd oxidase (34). The proton pumping cytochrome bccaa 3 oxidase is the more efficient of these two complexes, leading to the efflux of six H ϩ ions per electron pair received, and is the major complex utilized during aerobic growth (34,35). The nonproton pumping cytochrome bd complex is less efficient, mediating the transport of two H ϩ ions per electron pair, but is predicted to have a higher affinity for O 2 and is important during nonreplicative persistence (36 -39). In actively growing M. smegmatis, electrons entering the respiratory chain are derived from heterotrophic substrates and are donated to the respiratory chain by NADH largely via the nonproton pumping type II NADH dehydrogenase NDH-2 and succinate via the succinate dehydrogenase SDH1 (34). Although M. smegmatis is strictly heterotrophic for replicative growth, it was recently demonstrated that it is able to aerobically respire, using CO at atmospheric concentrations during carbon-limited persistence through the actions of a carbon monoxide dehydrogenase (40). It has likewise been predicted that Huc and Hhy support survival during persistence by providing electrons derived from H 2 to the respiratory chain (24,27). However, these studies were correlative, and it remains to be definitively demonstrated that H 2 serves as a respiratory electron donor in this organism.
In this work, we addressed these knowledge gaps by comprehensively studying Huc and Hhy during different stages of mycobacterial growth and persistence. We show that Huc and Hhy are differentially expressed throughout growth and persistence and form distinct interactions with the membrane. In addition, we show that both Huc and Hhy are obligately linked to the respiratory chain via the menaquinone pool but form distinct interactions with the terminal oxidases. These data demonstrate that H 2 oxidation in M. smegmatis provides electrons to the respiratory chain for mixotrophic growth via Huc and to energize persistence via Hhy. These findings represent a significant advance in our under-standing of the role of high-affinity hydrogenases in bacterial metabolism.

Mycobacterial hydrogenases are differentially expressed and active during growth and persistence
Previous work investigating the activity of Huc and Hhy in M. smegmatis showed that they are induced in batch culture upon exhaustion of carbon sources (24). However, we lack a high-resolution understanding of the expression and activity of Huc and Hhy during mycobacterial growth and persistence. To address this question, we quantified Huc and Hhy gene expression using qRT-PCR 2 at different growth phases in batch liquid cultures. In exponentially growing, carbon-replete cells (OD 600 0.3 and OD 600 1.0 cultures), transcript levels for hucL (the Huc large subunit) were relatively low; however, as carbon sources became exhausted, hucL expression increased, with maximum expression observed 1 day post-A max (OD 600 ϳ3.0). Subsequently, 3 days post-A max (OD 600 ϳ3.0), as cells endured prolonged carbon limitation, the expression levels of hucL declined significantly (Fig. 1a). In contrast, expression of hhyL (the Hhy large subunit) remained low during exponential growth before rapidly increasing by 57-fold when cells reached carbon-limited stationary phase and remained at high levels into late stationary phase (Fig. 1b).
Next we determined the rate of H 2 oxidation of WT M. smegmatis and mutant strains containing only Huc or Hhy at different stages of growth and persistence in liquid batch culture. H 2 oxidation rates in the Huc-only strain correlated well with gene expression levels; levels of H 2 oxidation were relatively low during early exponential growth (OD 600 0.3) and increased during late exponential phase (OD 600 1.0), before peaking 1 day post-A max and declining rapidly thereafter (Fig. 1c). The rapid decline in transcript levels and activity of Huc during stationary phase suggests tight regulation of this enzyme. In contrast, the activity of Hhy was low during exponential growth (OD 600 0.3 and 1.0), increased slightly 1 and 3 days post-A max , and increased markedly during prolonged persistence, with high levels of activity observed 3 weeks post-A max (Fig. 1d). A notable lag was observed between the increase in transcript levels and Hhy activity during stationary phase, suggesting posttranscriptional regulation of this hydrogenase. The H 2 oxidation activity profile of the WT strain in these assays are the same as the sum of the activity of Huc-and Hhy-only mutants, confirming that Huc and Hhy are functioning normally in the mutant background (Fig. 1e). Additionally, a mutant strain lacking both Huc and Hhy did not consume H 2 , confirming that Huc and Hhy are solely responsible for H 2 oxidation (Fig. 1f).
These data provide a clear picture of the differential regulation of Huc and Hhy, hinted at by previous studies (27). Huc is expressed by M. smegmatis during the transition from growth to persistence, allowing cells to grow mixotrophically on atmospheric H 2 and, where available, higher con-centrations produced through abiotic or biotic processes (e.g. fermentation, nitrogen fixation) (41). Subsequently, as cells commit to persistence because of carbon starvation, Hhy is expressed and supplies energy from atmospheric H 2 to meet maintenance needs.

Mycobacterial hydrogenases differentially associate with the membrane, with Huc potentially forming a supercomplex with the cytochrome bcc-aa 3 oxidase
To directly attribute the H 2 oxidation activity in our cellular assays to Huc and Hhy, we separated cell lysates of WT and hydrogenase mutant strains using native PAGE and detected hydrogenase activity by zymographic staining (Fig. 2a). A highmolecular-mass species and a low-molecular-mass weight band exhibiting H 2 oxidation activity was detected 1 day post-A max in WT and Huc-only cultures but not in the Hhy-only strain. We determined the size of the high-molecular-mass species to be Ͼ700 kDa and that of the low-molecular-mass band to be Ͼ66 kDa via blue native PAGE (Fig. S1). In contrast, 3 days post-A max , a mid-sized molecular mass H 2 oxidizing species was present in WT and Hhy-only cultures but absent from the Huconly strain (Fig. 2a). These high-and mid-sized molecular mass

Hydrogenases of the mycobacterial aerobic respiratory chain
bands from the WT strain were excised, and proteins present were identified by MS. The high-molecular-mass band yielded peptides corresponding to Huc (Tables S1 and S2), whereas the mid-sized molecular mass band yielded peptides corresponding to Hhy (Tables S1 and S3). These data correlate well with the activity of Huc and Hhy observed in our cellular assays, confirming that Huc is the dominant hydrogenase during the transition from growth to dormancy and that Hhy is more active during prolonged persistence.
The difference in size between Huc and Hhy activity observed on the native gel is striking. The slow migration of Huc may be due to formation of an oligomer containing multiple Huc subunits or other unidentified proteins. To test this hypothesis, we interrogated the MS data for likely Huc interaction partners. Intriguingly, components of the cytochrome bccaa 3 oxidase supercomplex were detected in the Huc sample with a high probability and coverage, demonstrating that they are prevalent in this region of the gel alongside various other proteins (Table S1 and S2 and Fig. S2). It has been shown previously that H 2 oxidation in M. smegmatis is oxygen-dependent, suggesting that Huc and Hhy activity are obligately linked to the respiratory chain (24). As cytochrome bcc-aa 3 is a large supercomplex (42), association with Huc could account for the high molecular mass of Huc activity on the native gel while placing the hydrogenase in an ideal position to donate electrons to this complex.
Previous work indicated that Huc and Hhy are membraneassociated despite lacking obvious transmembrane regions or signal peptides (24). To interrogate the nature of this membrane association, we fractionated cells into lysates, cytosols, and membranes and detected the hydrogenases by western blotting chromosomally StrepII-tagged variants of Huc and Hhy. Two bands corresponding to Huc were detected by western blotting in M. smegmatis whole cells, with sizes of ϳ60 and more than 200 kDa. Upon cell fractionation, the ϳ60-kDa band was observed in both cytoplasmic and membrane fractions, whereas the more than 200-kDa band was only observed in the membrane fraction (Fig. 2b). Interaction of Huc with the membrane was disrupted by 5% sodium cholate, with both bands partitioning to the soluble phase (Fig. 2b). In contrast, a single ϳ60-kDa band corresponding to Hhy was observed in the whole-cell lysate and membrane fractions but was absent from the cytoplasmic fraction. The interaction between Hhy and the cell membrane was not disrupted by addition of 5% sodium cholate, suggesting that it forms a strong interaction with the membrane relative to Huc and implies that different mechanisms are responsible for their membrane association (Fig. 2b). Both Huc and Hhy are predicted to form heterotetramers consisting of two large and small subunits with a molecular mass of more than 200 kDa (26,27). This is consistent with the bands observed via western blotting, with the more than 200-kDa species observed for Huc representing an intact tetramer and the ϳ60 kDa species representing partial disassociation of this complex.

Mycobacterial hydrogenases are coupled to the respiratory chain and interact differentially with the terminal cytochrome oxidases
Although it is known that O 2 is required for H 2 oxidation by Huc and Hhy in M. smegmatis (6,24), it had not been determined whether these enzymes support the reduction of O 2 through coupling to the respiratory chain. To resolve this question, we amperometrically monitored the H 2 and O 2 consumption in carbon-limited M. smegmatis cells (3 days post-A max ) following sequential spiking with H 2 and O 2 saturated buffer (Fig. 3a). In H 2 oxidation measurements, upon spiking cells with H 2 only, oxidation (0.31 M min Ϫ1 ) was observed because of ambient levels of O 2 present in solution. When O 2 -saturated buffer was subsequently added, the rate of H 2 oxidation increased markedly (1.2 M min Ϫ1 ) (Fig. 3a). In O 2 consump-

Hydrogenases of the mycobacterial aerobic respiratory chain
tion measurements, cells that were initially spiked with O 2 minimally consumed O 2 (0.01 M min Ϫ1 ). However, with the subsequent addition of H 2 , the cells consumed O 2 at approximately half the rate observed for H 2 oxidation (0.51 M min Ϫ1 ) (Fig.  3a). This rate is consistent with the expected stoichiometry of H 2 -dependent aerobic respiration (H 2 ϩ 1 ⁄ 2 O 2 3 H 2 O). These data directly link H 2 oxidation to O 2 consumption, providing strong experimental evidence that electrons derived from H 2 support respiratory reduction of O 2 in M. smegmatis.
Having established that H 2 oxidation directly supports O 2 reduction in M. smegmatis, we next sought to determine which of the two terminal oxidases were utilized for this process. To achieve this, we monitored the rate of H 2 oxidation and O 2 consumption in WT, Huc-only, and Hhy-only strains as well as mutant strains possessing either cytochrome bcc-aa3 or bd oxidase as the sole terminal respiratory complex. Given that loss of both terminal oxidases is lethal in M. smegmatis (43), we utilized zinc azide, a selective inhibitor of cytochrome bcc-aa 3 oxidase (44), to assess the effects of loss of both terminal oxidases on H 2 oxidation. As expected from our initial experiments 3 days post-A max (Fig.  1), the H 2 oxidation rate of the Hhy-only mutant was 3-fold higher than that of the Huc-only strain (Fig. 3b). H 2 oxidation was also observed in the cytochrome bcc-aa 3 oxidaseonly strain, showing that the complex receives electrons from H 2 oxidation. However, this activity was 3-fold lower than what was observed for WT cells, suggesting that the cytochrome bd complex also receives electrons from H 2 oxidation 3 days post-A max (Fig. 3b). In striking contrast, H 2 oxidation in the cytochrome bd-only strain was 6.3-fold greater than in the WT (Fig. 3b). This may be due to an increase in the amount of hydrogenases present in the cells or deregulation of their activity because of metabolic remodeling to cope with the loss of the proton-pumping cytochrome bcc-aa 3 oxidase (35). The O 2 consumption for the WT and Huc-and Hhy-only strains, when spiked with H 2 , fit approximately with the 2:1 stoichiometry observed in our initial experiment (Fig. 3c). However, O 2 consumption of either oxidase mutants in the presence of H 2 was significantly higher than in the WT and did not conform to a 2:1 ratio (Fig. 3c), suggesting that H 2 is co-metabolized with other substrates (e.g. carbon reserves). Taken together, these data show that both terminal oxidase complexes accept electrons from H 2 oxidation.
Next we probed the specifics of coupling between Huc and Hhy and the terminal oxidases by inhibiting the cytochrome bcc-aa 3 complex with zinc azide. In WT cells, addition of azide led to a 4.6-fold reduction in H 2 oxidation, demonstrating that hydrogenase activity is primarily coupled to the cytochrome bcc-aa 3 complex 3 days post-A max . For the Huc-only strain, addition of zinc azide largely abolished H 2 oxidation (8.4-fold reduction), suggesting that Huc is obligately coupled to the cytochrome bcc-aa 3 complex (Fig. 3b). In contrast, only a 1.5-fold reduction in Hhy activity was observed; this demonstrates that, although Hhy utilizes the cytochrome bcc-aa 3 oxidase, it is promiscuous and can also donate electrons to the alternative cytochrome bd complex (Fig. 3b). Addition of azide to the cytochrome bcc-aa 3 only strain led to nearly complete inhibition of H 2 oxidation (10.5-fold decrease) and O 2 consumption (22.9-fold decrease) (Fig. 3c). This confirms that Huc and Hhy require an

Hydrogenases of the mycobacterial aerobic respiratory chain
active terminal oxidase to oxidize H 2 and, thus, are obligately coupled to the respiratory chain. The high level of H 2 oxidation observed in the cytochrome bd-only strain was unchanged by azide treatment, which is expected, given that this complex is unaffected by azide inhibition (44).

Huc and Hhy transfer electrons into the respiratory chain via the quinone pool
Having firmly established that Huc and Hhy activity is coupled to terminal oxidase activity under the tested conditions, we sought to better understand this relationship. To do so, we measured H 2 oxidation of the WT, Huc-only, and Hhy-only strains in the presence of selective respiratory chain inhibitors and uncouplers. First we tested whether the electrons generated by Hhy and Huc are transferred to the electron carrier menaquinone, which donates electrons to both terminal oxidases in mycobacteria (45). To do so, we tested the effect of HQNO, a competitive inhibitor of quinone-binding (44,46), on H 2 oxidation in our WT and Huc-and Hhy-only strains. Addition of HQNO led to an 8.6-fold and 10.4-fold decrease in H 2 oxidation by the Huc-and Hhy-only mutants, respectively, demonstrating that transport of electrons generated by these enzymes occurs via the menaquinone pool (Fig. 4, a  and b). There was a 2.5-fold decrease in the activity of WT cells treated with HQNO, confirming that H 2 oxidation is also menaquinone-dependent in a nonmutant background (Fig. 4c).
Next we tested the effect of valinomycin on H 2 oxidation. Valinomycin is an ionophore that binds K ϩ , forming a positively charged complex, specifically transporting K ϩ ions across the cellular membrane along the electrical gradient, collapsing the electrical potential component of the proton-motive force (PMF) in respiratory bacteria (47). In M. smegmatis, at an external of pH above 5, the majority of the PMF is driven by electrical potential (48); thus, addition of valinomycin under our assay conditions (pH 5.8) leads to a dramatic reduction in PMF. Huc and Hhy exhibited strikingly different responses to valinomycin. Valinomycin reduced H 2 oxidation by 20-fold in the Huconly strain but increased oxidation by 1.3-fold in the Hhy-only and WT strains compared with untreated cells (Fig. 4). The nearly complete loss of Huc activity because of valinomycin treatment indicates that this enzyme is energy-dependent, requiring the largely intact PMF to function; this may indicate that the complex is obligately associated with the proton-translocating cytochrome bcc-aa 3 supercomplex. Conversely, the increase in Hhy activity demonstrates that this enzyme does not require the PMF, with the increase in H 2 oxidation possibly

Hydrogenases of the mycobacterial aerobic respiratory chain
resulting from increased metabolic flux as the cells attempt to maintain their membrane potential.
Next we tested the effect of nigericin on H 2 oxidation. Nigericin is an ionophore that acts as an antiporter of K ϩ and H ϩ ions and is uncharged in its ion-bound forms (47). In our assay, nigericin led to net efflux of K ϩ ions from and influx of H ϩ ions into the cell, dissipating the proton gradient but not affecting membrane potential. The experimentally determined pH of the external medium in our assay was 5.8. Under these conditions, ⌬pH across the membrane accounts for approximately one-third of the PMF in M. smegmatis (48). Thus, addition of nigericin will lead to a significant net influx of protons and acidification of the cytoplasm, shifting the equilibrium for H 2 toward reduction. Addition of nigericin inhibited H 2 oxidation to a moderate degree in our assay. WT and Huc-only strains exhibited a ϳ1.5-fold decrease in activity, whereas Hhy activity was reduced 2.3-fold (Fig. 4). This inhibition of both hydrogenases may be attributed to cytoplasmic acidification or direct inhibition of these enzymes by nigericin. The inhibitory effect of nigericin toward Hhy is in contrast with the stimulatory effect of valinomycin on this enzyme. This possibly results from the fact that valinomycin directly diminishes the PMF but nigericin does not, or it may be a consequence of a more complicated secondary effect on cellular metabolism.

A model of the integration of hydrogenases in the mycobacterial respiratory chain
Based on the findings from our work, we propose a model of integration of Huc and Hhy into the mycobacterial respiratory chain, which we outline in Fig. 5. Our data show that both Huc and Hhy are obligately coupled to O 2 reduction via the terminal oxidases of the respiratory chain. Huc preferentially donates electrons to the cytochrome bcc-aa 3 complex, whereas Hhy donates electrons both the cytochrome bcc-aa 3 and cytochrome bd complexes. Both enzymes are membrane-associated, positioning them for the transfer of electrons produced by H 2 oxidation to the respiratory chain. The size of Huc and its comigration with the cytochrome bcc-aa 3 complex on a native gel suggest that Huc association with the membrane may be mediated by protein-protein interactions, possibly with the terminal oxidase. Inhibition of Huc and Hhy activity because of blocking of menaquinone binding by HQNO shows that electrons from these enzymes are transferred to the terminal oxidases via this cofactor. It remains to be resolved whether electron transfer to menaquinone is directly mediated by the hydrogenases or through an intermediate protein; for example, the FeS proteins cotranscribed with the hydrogenase large and small subunits in the huc and hhy operons (27). Collapse of the PMF by valinomycin treatment leads to nearly complete inhibition of Huc activity but enhancement of Hhy activity. This suggests that a distinct relationship exists between these enzymes and the PMF, with Huc requiring an intact PMF, potentially because of its obligate coupling to the cytochrome bcc-aa 3 complex.
Our data also demonstrates that Huc and Hhy are differentially regulated during mycobacterial growth and persistence. The tightly controlled expression and activity of Huc during the transition between growth and dormancy suggests that it oxi-

Hydrogenases of the mycobacterial aerobic respiratory chain
dizes H 2 mixotrophically as heterotrophic energy sources become scarce. The proton-motive force generated by Huc, through obligate interaction with the cytochrome bcc-aa 3 complex, may help to energize the cell during the transition to dormancy. Expression of hhyL, the gene encoding the large Hhy subunit, is up-regulated at the cessation of cell division. However, high levels of Hhy-mediated H 2 oxidation are only observed several weeks into dormancy. This suggests that the enzyme primarily functions to meet maintenance needs during persistence, a role that is further supported by its promiscuous utilization of cytochrome bd oxidase. The observed lag between hhyL transcription and Hhy activity suggests that regulation of this hydrogenase also occurs downstream of transcription. This possibly provides flexibility to M. smegmatis by not committing to synthesis of this resource-intensive protein immediately upon exhaustion of carbon-derived energy sources but allowing rapid deployment of Hhy if resources remain scarce.
Although this work provides the basis for understanding how Huc and Hhy are regulated and integrated into cellular metabolism, further investigation is required to fully understand the mechanisms that regulate these enzymes and the biochemistry of their H 2 oxidation. For example, what are the regulatory pathways that allow the cell to rapidly switch on Huc when resources become scarce and then off again as M. smegmatis commits to dormancy? Analogously, how do nonreplicating M. smegmatis cells regulate transcription and then activity of Hhy during a state of resource limitation? In addition, the data we present suggest physical interactions between Huc and respiratory chain components. Purification of both Huc and Hhy from their native context in the M. smegmatis cell will likely provide insight into the protein-protein interactions that mediate electron transfer from these enzymes. Furthermore, purification of these complexes, combined with structural and spectroscopic analysis, will likely provide insight into the mechanisms that underpin the high H 2 affinity and O 2 tolerance of Huc and Hhy. In conjunction with this study, these data will provide a richer picture of how mycobacteria consume H 2 during growth and persistence.

Bacterial strains and growth conditions
M. smegmatis mc 2 155 and its derivatives were routinely grown in lysogeny broth (LB) agar plates supplemented with 0.05% (w/v) Tween 80 (LBT) (49). In broth culture, the strain was grown in either LBT or Hartmans de Bont minimal medium supplemented with 0.2% (w/v) glycerol, 0.05% (w/v) tyloxapol, and 10 mM NiSO 4 . Escherichia coli was maintained in LB agar plates and grown in LB broth (50). Selective LB or LBT medium used for cloning experiments contained gentamycin at 5 g ml Ϫ1 for M. smegmatis and 20 g ml Ϫ1 for E. coli. Cultures were routinely incubated at 37°C with rotary shaking at 150 rpm for liquid cultures unless otherwise specified. The strains of M. smegmatis (51,52) and its derivatives and E. coli are listed in Table S4.

Insertion of StrepII tags
To facilitate visualization of the hydrogenases in western blots, a StrepII tag was inserted at the C-terminal end of the small subunits of Huc (MSMEG_2262, hucS) and Hhy (MSMEG_2720, hhyS) through allelic exchange mutagenesis as described previously (33). Two allelic exchange constructs, hucS StrepII (2656 bp) and hhyS StrepII (3000 bp), were synthesized by GenScript. These were cloned into the SpeI site of the mycobacterial shuttle plasmid pX33 to yield the constructs pHuc-StrepII and pHhy-StrepII (Table S4). The constructs were propagated in E. coli TOP10 and transformed into WT M. smegmatis mc 2 155 cells by electroporation. Gentamycin was used in selective solid and liquid medium to propagate pX33. To allow permissive temperature-sensitive vector replication, transformants were incubated on LBT gentamycin plates at 28°C until colonies were visible (5-7 days). The resultant catechol-positive colonies were subcultured onto fresh LBT gentamycin plates and incubated at 40°C for 3-5 days to facilitate integration of the recombinant plasmid, via flanking regions, into the chromosome. The second recombination event was facilitated by subculturing catechol-reactive and gentamycin-resistant colonies onto LBT agar plates supplemented with 10% sucrose (w/v) and incubating at 40°C for 3-5 days. Catechol-unreactive colonies were subsequently screened by PCR to distinguish WT revertants from Huc-StrepII and Hhy-StrepII mutants. The primers used for screening are listed in Table S5.

Cellular fractionation for detection of Huc and Hhy
The untagged and StrepII-tagged Huc and Hhy were constitutively produced by growing M. smegmatis WT and the Huconly, Hhy-only, Huc-StrepII, and Hhy-StrepII strains in Hartmans de Bont minimal medium with 0.2% glycerol as the sole carbon source (24,49). Cells were grown at 37°C with agitation and harvested by centrifugation (15 min, 10,000 ϫ g, 4°C) 1 day post-A max (ϳ3.0) for the WT, Huc-only, and Huc-StrepII or 3 days post-A max (ϳ3.0) for the WT, Hhy-only, and Hhy-StrepII. They were washed in lysis buffer (50 mM Tris-Cl (pH 8.0), 2 mM MgCl 2 , 1 mM PMSF, 5 mg ml Ϫ1 lysozyme, and 40 g ml Ϫ1 DNase) and resuspended in the same buffer in a 1:5 cell mass to buffer ratio. The cell suspension was homogenized using a Dounce homogenizer and passed through a cell disruptor (40,000 p.s.i., four times; Constant Systems Ltd, model: One Shot series). After removal of unbroken cells by low-speed centrifugation (20 min, 8000 ϫ g, 4°C), the whole-cell lysates of the WT, Huc-only, and Hhy-only strains were used for activity staining. Cell lysates of the Huc-StrepII and Hhy-StrepII strains were separated by ultracentrifugation (60 min, 150,000 ϫ g, 4°C) into cytosol and membrane fractions. The membrane was washed in lysis buffer and analyzed by western blotting. Protein concentrations were measured by the bicinchoninic acid method against BSA standards.

Hydrogenase activity staining
Twenty micrograms of each whole-cell lysate was loaded onto two native 7.5% (w/v) BisTris polyacrylamide gels prepared as described elsewhere (53) and run in a Tris (25 mM)glycine (192 mM) buffer (pH 8.3) alongside a protein standard Hydrogenases of the mycobacterial aerobic respiratory chain (NativeMark Unstained Protein Standard, Thermo Fisher Scientific) for 3 h at 25 mA. For total protein determination, gels were stained overnight at 4°C with gentle agitation using AcquaStain Protein Gel Stain (Bulldog Bio). To determine hydrogenase activity, a duplicate gel was incubated in 50 mM potassium phosphate buffer (pH 7.0) supplemented with 500 M nitro blue tetrazolium chloride in an anaerobic jar amended with an anaerobic gas mixture (5% H 2 , 10% CO 2 , and 85% N 2 (v/v)) overnight at room temperature. Bands present after incubation corresponded to hydrogenase activity. For blue native PAGE, lysates loaded onto native 7.5% (w/v) BisTris polyacrylamide gels were run at 25 mA in a commercial buffer system consisting of anode NativePAGE Running Buffer (50 mM BisTris, 50 mM N-[2-hydroxy-1,1-bis(hydroxymethyl)ethyl]glycine (pH 6.8)) and NativePAGE Light Blue Cathode buffer (50 mM BisTris, 50 mM N-[2-hydroxy-1,1-bis(hydroxymethyl)ethyl]glycine, and 0.002% Coomassie Brilliant Blue G-250 (pH 6.8)) (Thermo Fisher Scientific). After the run, gels were incubated either for total protein staining or hydrogenase activity staining, as described above.

Membrane solubilization and western blots
Membrane solubilization was performed by resuspending washed membranes (to a final protein concentration of 1 mg ml Ϫ1 ) in solubilization buffer containing 50 mM Tris-Cl (pH 8.0), 1 mM PMSF, and 5% (w/v) sodium cholate (54). The solutions were incubated at room temperature with gentle agitation for 3 h. Detergent-soluble proteins were separated from the insoluble material by ultracentrifugation. As a control, the membrane was suspended in buffer without sodium cholate. Total proteins in the fractions were visualized in SDS-PAGE and Huc_StrepII and Hhy_StrepII by western blotting. For western blotting, 20 g of total protein was loaded and run onto Bolt 4 -12% SDS-polyacrylamide gels after boiling in Bolt SDS sample buffer and 50 mM DTT. The proteins in the gels were then transferred onto a PVDF membrane using Trans-Blot SD Semi-Dry Transfer Cell (Bio-Rad) set at 15 V for 60 min. Following transfer, the protein-containing PVDF membrane was blocked with 3% (w/v) BSA in PBS (pH 7.4) with 0.1% (v/v) Tween 20 (PBST). The PVDF membrane was washed three times in 20 ml of PBST and finally resuspended in 10 ml of the same buffer. Strep-Tactin HRP conjugate was then added at a 1:100,000 dilution. Peroxide-mediated chemiluminescence of luminol catalyzed by the HRP was developed according to the manufacturer's specifications (Amersham Biosciences ECL Prime detection reagent, GE Life Sciences), and the Strep-Tactin (HRP conjugated)-Stre-pII-tag complex was visualized in a Fusion Solo S (Fischer Biotech) chemiluminescence detector.

Gene expression analysis
For qRT-PCR analysis, five synchronized sets of biological triplicate cultures (30 ml) of WT M. smegmatis were grown in 125-ml aerated conical flasks. Each set of triplicates was quenched either at OD 600 0.3, OD 600 1.0, 1 day post-A max (OD 600 ϳ3.0), 3 days post-A max (OD 600 ϳ3.0), or 3 weeks post-A max (OD 600 ϳ3.0) with 60 ml of cold 3:2 glycerol:saline solution (Ϫ20°C). They were subsequently harvested by centrifugation (20,000 ϫ g, 30 min, Ϫ9°C), resuspended in 1 ml of cold 1:1 glycerol:saline solution (Ϫ20°C), and further centrifuged (20,000 ϫ g, 30 min, Ϫ9°C). For cell lysis, pellets were resuspended in 1 ml TRIzol reagent, mixed with 0.1-mm zircon beads, and subjected to five cycles of bead beating (4000 rpm, 30 s) in a Biospec Mini-Beadbeater. Total RNA was subsequently extracted by phenol-chloroform extraction according to the manufacturer's instructions (TRIzol Reagent User Guide, Thermo Fisher Scientific) and resuspended in diethylpyrocarbonate-treated water. RNA was treated with DNase using the Turbo DNA-free kit (Thermo Fisher Scientific) according to the manufacturer's instructions. cDNA was then synthesized using the SuperScript III First-Strand Synthesis System for qRT-PCR (Thermo Fisher Scientific) with random hexamer primers according to the manufacturer's instructions. qPCR was used to quantify the levels of the target genes hucL (Huc) and hhyL (Hhy) and the housekeeping gene sigA against amplicon standards of known concentration. All reactions were run in a single 96-well plate using the PowerUp SYBR Green Master Mix (Thermo Fisher Scientific) and LightCycler 480 Instrument (Roche) according to each manufacturers' instructions. A standard curve was created based on the cycle threshold values of hucL, hhyL, and sigA amplicons that were serially diluted from 10 8 to 10 copies (R 2 Ͼ 0.98). The copy number of the genes in each sample was interpolated based on each standard curve, and values were normalized to sigA expression. For each biological replicate, all samples, standards, and negative controls were run in technical duplicate. The primers used in this work are summarized in Table S5.

Microrespiration measurements
Rates of H 2 oxidation or O 2 consumption were measured amperometrically according to previously established protocols (27,30). For each set of measurements, either a Unisense H 2 microsensor or Unisense O 2 microsensor electrode was polarized at ϩ800 mV or Ϫ800 mV, respectively, with a Unisense multimeter. The microsensors were calibrated with either H 2 or O 2 standards of known concentration. Gas-saturated PBS (137 mM NaCl, 2.7 mM KCl, 10 mM Na 2 HPO 4 , and 2 mM KH 2 PO 4 (pH 7.4)) was prepared by bubbling the solution with 100% (v/v) of either H 2 or O 2 for 5 min. In uncoupler/ inhibitor-untreated cells, H 2 oxidation was measured in 1.1-ml microrespiration assay chambers. These were amended with 0.9-ml cell suspensions of M. smegmatis WT or derivative strains either at OD 600 0.3, OD 600 1.0, 1 day post-A max (OD 600 ϳ3.0), 3 days post-A max (OD 600 ϳ3.0), or 3 weeks post-A max (OD 600 ϳ3.0), They were subsequently amended with 0.1 ml of H 2 -saturated PBS and 0.1 ml of O 2 -saturated PBS. Chambers were stirred at 250 rpm at room temperature. For cells at midstationary phase, following measurements of untreated cells, the assay mixtures were treated with either 10 M nigericin, 10 M valinomycin, 40 M N-oxo-2-heptyl-4-hydroxyquinoline (HQNO), or 250 M zinc azide before measurement. In O 2 consumption measurements, initial O 2 consumption without addition of H 2 was measured in microrespiration assay chambers sequentially amended with 0.9-ml cell suspensions of M. smegmatis WT or derivative strains at mid-stationary phase (3 days post-A max ) and 0.1 ml O 2 -saturated PBS (0.1 ml) with stirring at 250 rpm at room temperature. After initial measure-Hydrogenases of the mycobacterial aerobic respiratory chain ments, 0.1 ml of H 2 -saturated PBS was added to the assay mixture, and changes in O 2 concentrations were recorded. Additionally, O 2 consumption was measured in cytochrome bcc-aa 3 -only strains treated with 250 M zinc azide. In both H 2 and O 2 measurements, changes in concentrations were logged using Unisense Logger software. Upon observing a linear change in either H 2 or O 2 concentration, rates of consumption were calculated over a period of 20 s and normalized against total protein concentration.

Mass spectrometry analysis
After hydrogenase activity staining, bands corresponding to Huc and Hhy activity were excised from the gel and destained with 50% acetonitrile (ACN) in 100 mM NH 4 HCO 3 (ABC) (pH 8.5). The proteins were subsequently reduced with 10 mM DTT (Astral Thermo Fisher Scientific) and carbamidomethylated with 25 mM chloroacetamide (Sigma). The gel was then dehydrated with 100% ACN and rehydrated with digestion solution containing 10 ng l Ϫ1 trypsin (Promega), 100 mM ABC, and 5% ACN. The gel was digested overnight at 37°C, and tryptic peptides were extracted from the gel with 50% ACN-5% formic acid solution lyophilized in a vacuum concentrator, and purified using OMIX C18 Mini-Bed tips (Agilent Technologies) prior to LC-MS/MS analysis. Using a Dionex UltiMate 3000 RSLCnano system equipped with a Dionex UltiMate 3000 RS autosampler, an Acclaim PepMap RSLC analytical column (75 m ϫ 50 cm, nanoViper, C18, 2 m, 100 Å; Thermo Scientific), and an Acclaim PepMap 100 trap column (100 m ϫ 2 cm, nanoViper, C18, 5 m, 100 Å; Thermo Scientific), the tryptic peptides were separated by increasing concentrations of 80% ACN/0.1% formic acid at a flow rate of 250 nl min Ϫ1 for 60 min and analyzed with a Orbitrap Fusion mass spectrometer (Thermo Scientific). The instrument was operated in data-dependent acquisition mode to automatically switch between full-scan MS and MS/MS acquisition. Each survey full scan (m/z 375-1575) was acquired in the Orbitrap with 70,000 resolution (at m/z 200) after accumulation of ions to a 1 ϫ 10 6 target value with a maximum injection time of 54 ms. Dynamic exclusion was set to 60 s. The 20 most intense multiply charged ions (z Ն 2) were sequentially isolated and fragmented in the collision cell by higher-energy collisional dissociation with a fixed injection time of 54 ms, 30,000 resolution, and automatic gain control target of 2 ϫ 10 5 . For assignment, the raw files were searched against the M. smegmatis database, containing 6717 entries, using Byonic v3.0.0 (ProteinMetrics). Database searching was performed with the following parameters: cysteine carbamidomethylation as a fixed modification, methionine oxidation and N-terminal acetylation as variable modifications, up to two missed cleavages permitted, mass tolerance of 10 ppm, and 1% protein false discovery rate for protein and peptide identification based on a decoy database. The M. smegmatis protein sequence database was downloaded from GenBank NCBI 2018. The raw mass spectrometry data have been deposited in Figshare (https://monash.figshare.com/projects/ Hydrogenase/69578). 3 File F1CH20190327_WT_Top.raw contains the raw data for the high-molecular-mass band corresponding to Huc activity. File F1CH20190327_WT_down. raw contains the raw data for the mid-sized molecular mass band corresponding to Hhy activity.