Activity of Spore-Specific Respiratory Nitrate Reductase 1 of Streptomyces coelicolor A3(2) Requires a Functional Cytochrome bcc-aa₃ Oxidase Supercomplex

Streptomyces coelicolor forms spores that respire with either oxygen or nitrate, using only endogenous electron donors. This helps maintain a membrane potential and, thus, viability. Respiratory nitrate reductase (Nar) usually receives electrons directly from reduced quinone species; however, we show that nitrate respiration in spores requires a respiratory supercomplex comprising cytochrome bcc oxidoreductase and aa₃ oxidase. Our findings suggest that the Nar1 enzyme in the S. coelicolor spore functions together with the proton-translocating bcc-aa₃ supercomplex to help maintain the membrane potential more efficiently. Dissecting the mechanisms underlying this survival strategy is important for our general understanding of bacterial persistence during infection processes and of how bacteria might deal with nutrient limitation in the natural environment.

ABSTRACT

Spores have strongly reduced metabolic activity and are produced during the complex developmental cycle of the actinobacterium Streptomyces coelicolor. Resting spores can remain viable for decades, yet little is known about how they conserve energy. It is known, however, that they can reduce either oxygen or nitrate using endogenous electron sources. S. coelicolor uses either a cytochrome bd oxidase or a cytochrome bcc-aa₃ oxidase supercomplex to reduce oxygen, while nitrate is reduced by Nar-type nitrate reductases, which typically oxidize quinol directly. Here, we show that in resting spores the Nar1 nitrate reductase requires a functional bcc-aa₃ supercomplex to reduce nitrate. Mutants lacking the complete qcr-cta genetic locus encoding the bcc-aa₃ supercomplex showed no Nar1-dependent nitrate reduction. Recovery of Nar1 activity was achieved by genetic complementation but only when the complete qcr-cta locus was reintroduced to the mutant strain. We could exclude that the dependence on the supercomplex for nitrate reduction was via regulation of nitrate transport. Moreover, the catalytic subunit, NarG1, of Nar1 was synthesized in the qcr-cta mutant, ruling out transcriptional control. Constitutive synthesis of Nar1 in mycelium revealed that the enzyme was poorly active in this compartment, suggesting that the Nar1 enzyme cannot act as a typical quinol oxidase. Notably, nitrate reduction by the Nar2 enzyme, which is active in growing mycelium, was not wholly dependent on the bcc-aa₃ supercomplex for activity. Together, our data suggest that Nar1 functions together with the proton-translocating bcc-aa₃ supercomplex to increase the efficiency of energy conservation in resting spores.

IMPORTANCE Streptomyces coelicolor forms spores that respire with either oxygen or nitrate, using only endogenous electron donors. This helps maintain a membrane potential and, thus, viability. Respiratory nitrate reductase (Nar) usually receives electrons directly from reduced quinone species; however, we show that nitrate respiration in spores requires a respiratory supercomplex comprising cytochrome bcc oxidoreductase and aa₃ oxidase. Our findings suggest that the Nar1 enzyme in the S. coelicolor spore functions together with the proton-translocating bcc-aa₃ supercomplex to help maintain the membrane potential more efficiently. Dissecting the mechanisms underlying this survival strategy is important for our general understanding of bacterial persistence during infection processes and of how bacteria might deal with nutrient limitation in the natural environment.

INTRODUCTION

Streptomyces coelicolor A3(2) is a high-GC content, Gram-positive actinobacterium characterized by its saprophytic lifestyle and complex developmental cycle (1–3). S. coelicolor requires oxygen for growth of its substrate mycelium, and when nutrients become limiting, it produces hydrophobic aerial hyphae that develop into chains of dispersible spores (4, 5). The spores are less robust than endospores of firmicutes, but they nevertheless confer upon the bacterium a considerable survival advantage under adverse conditions. Spores have abundant supplies of storage compounds (4, 6), which are degraded to glucose and, thus, act as excellent endogenous electron donors. Endogenous electron donation is manifested by the ability of spores to reduce O₂ or nitrate continuously in aqueous buffer without the need to supply an exogenous source of electrons (e⁻) (7).

Energy conservation by spores is achieved mainly by coupling NADH, pyruvate, or succinate oxidation to O₂ reduction (8). It is imperative that spores, like all living cells, maintain a proton motive force (PMF) and, consequently, a membrane potential despite maintaining a reduced rate of metabolism. During oxidation of the reduced electron donor(s), the electrons are passed via one of two routes to O₂. When the O₂ concentration is low, they oxidize menaquinol via a high-affinity cytochrome bd oxidase, while at higher O₂ concentrations the electrons flow through a large proton-translocating supercomplex comprising a menaquinol-cytochrome c oxidoreductase (bcc complex) and an aa₃-type cytochrome oxidase, referred to henceforth as the bcc-aa₃ supercomplex (8, 9). Such a bcc-aa₃ supercomplex has been identified in all actinobacteria studied so far (10–13), and the structure of the supercomplex from Mycobacterium smegmatis has been resolved recently using cryo-electron microscopy (14, 15). For every two electrons that pass through the complex, a total of six protons are predicted to be translocated, with 4 H⁺ ions being translocated via the Q cycle of the bcc complex and 2 H⁺ ions being pumped by the aa₃ oxidase. Central to the function of this supercomplex is a membrane-associated diheme c-type cytochrome, QcrC, which, unlike the soluble cytochrome c in the mitochondrial respiratory chain (16), necessitates that complexes III and IV are in contact to allow electron transfer between them to occur (14, 15).

As well as being able to reduce O₂, S. coelicolor can also use nitrate as a respiratory electron acceptor (8). Surprisingly, although the bacterium synthesizes three nonredundant respiratory nitrate reductases (Nar), it cannot grow anaerobically by nitrate respiration, and, thus, it is assumed that these enzymes contribute to maintaining a PMF when O₂ becomes limiting or is depleted (17). Of the three Nar enzymes in S. coelicolor, Nar1 is exclusively found in spores and is always present in this cellular compartment (7), while Nar2 and Nar3 are mainly found in mycelium (18, 19). In bacteria such as Escherichia coli, Nar functions together with a formate dehydrogenase in a classical redox loop (20, 21), as originally formulated by Mitchell in his chemiosmotic theory (summarized in reference 22). Accordingly, although neither enzyme complex is proton translocating, the complexes are oriented facing opposite sides of the membrane. Upon passage of 2 e⁻ from the outside through the formate dehydrogenase, 2 H⁺ ions from the cytoplasm combine with them to reduce quinone, and these protons are then released at the outside of the membrane while the electrons pass through Nar to reduce nitrate in the cytoplasm. This results in generation of a proton gradient with a H⁺/e⁻ stoichiometry of 1 (23). While considerably less efficient at generating a PMF than the combination of proton pumps and the Q cycle, this is nevertheless sufficient to allow bacterial growth.

Recent studies have revealed that, like the Actinobacteria, the Epsilonproteobacteria Wolinella succinogenes (24) and Campylobacter jejuni (25) also possess a membrane-associated diheme cytochrome c, which allows the bcc complex of these bacteria to transfer electrons to a periplasmic nitrous oxide reductase and a nitrate reductase called Nap, respectively; unlike streptomycetes, these bacteria do not have a bcc-aa₃ supercomplex. Theoretically, therefore, a bcc complex could also deliver electrons to a respiratory nitrate reductase of the Nar type; however, this has never been documented. This would have the advantage of more efficient energy coupling than employing a redox loop-type mechanism. Here, we demonstrate that the Nar1 enzyme of S. coelicolor requires the bcc-aa₃ supercomplex to allow nitrate reduction in spores to occur.

RESULTS

Activity of Nar1 in spores is dependent on the bcc-aa₃ supercomplex.

Nar1 is exclusively synthesized and active in spores of wild-type S. coelicolor and is responsible for minimally 90% of the nitrate-reducing activity in this cell compartment, with the remainder being attributable mainly to Nar3 (7). During an analysis of Nar1 enzyme activity in spores of different respiratory mutants, we noted that a resting spore suspension of a mutant, COE192 (Fig. 1), lacking the complete genetic locus encoding the bcc-aa₃ supercomplex (9) exhibited a greater than 90% reduction in its ability to reduce nitrate to nitrite (Table 1). This phenotype was similar to that observed for a nar1 operon mutant (7), suggesting that Nar1 was the enzyme activity affected by the mutation. Spore suspensions of the triple nar mutant NM92 (Δnar1 Δnar2 Δnar3) (17) also exhibited essentially no reduction of nitrate to nitrite (Table 1). Nitrate-reducing activity could be restored to strain COE192 by introducing the complete qcr-cta gene locus on integrative plasmid pMS2148-56, yielding strain COE634 (Table 1), confirming that the deficiency in nitrate reduction resulted from deletion of the genes encoding the bcc-aa₃ supercomplex.

FIG 1.

Schematic representation of the qcr, cta, and cydAB operons and their corresponding gene products. Shown are the genes encoding the cytochrome bcc complex (green), the aa₃ oxidase (violet), and cytochrome bd oxidase (blue). Note that the cydAB genes are at a separate location on the genome (29). The products of the gray genes are not required for supercomplex function. The genotypes of the key strains used in this study are also depicted.

TABLE 1.

Nitrite production by resting spores

Strain	Nitrite production (mM)a
M145 (wild type)	0.79 ± 0.11
NM92 (Δnar123)	0.02 ± 0.01
COE192 (Δqcr-cta)	0.04 ± 0.02
COE634 (COE192/pMS2148-56)	1.18b
COE502A (ΔctaD)	0.09 ± 0.01
COE190 (ΔcydAB)	0.98 ± 0.03
COE640 (NM92/pMSnar1)	1.45 ± 0.14
COE648 (NM92/pNGnar1)	1.51 ± 0.04

^aSee Materials and Methods for details of the assay. All data were taken from a minimum of three biological replicates, each measured in triplicate, except as noted.

^bAverage taken from two biological replicates, with each sample assayed in triplicate.

Mutant COE502A carries an insertion mutation in the ctaD gene (SCO2155) (Fig. 1) and has recently been shown to lack activity of the cytochrome aa₃ oxidase (9). Analysis of a resting spore suspension derived from this mutant also exhibited a nitrate respiration phenotype indistinguishable from that of mutant COE192 (Δqcr-cta), i.e., a strongly reduced capacity to reduce nitrate to nitrite (Table 1). Notably, spores from the S. coelicolor mutant COE190 (ΔcydAB), which lacks the cytochrome bd oxidase but has a complete and functional qcr-cta locus (26), were unimpaired in their ability to reduce nitrate to nitrite (Table 1) (however, see below).

In the assay used here, nitrate respiration in resting spores requires only addition of the electron acceptor nitrate and is driven by endogenous electron donors from within the spores (7, 17). It is conceivable, therefore, that the inability of strain COE192 to reduce nitrate could be indirect and due to impaired nitrate uptake or nitrite export by the spores. To address this, we determined total Nar enzyme activity in spore extracts using the reduced redox dye benzyl viologen (BV) as an electron donor (Table 2). Crude extracts derived from spores of strains COE192 (Δqcr-cta) and COE502A (ctaD) exhibited a Nar activity that was reduced by 99% compared with the activity in extracts derived from the wild-type strain M145. Reintroduction of the complete qcr-cta gene locus on plasmid pMS2148-56 restored Nar activity to strain COE634 (COE192/pMS2148-56) (Fig. 1 and Table 2). Thus, the effect of the mutations in the qcr-cta gene locus on Nar activity reflects what was observed for nitrate reduction in resting spores, and the nitrate reduction-negative phenotype was not a consequence of impaired nitrate transport into spores. Analysis of Nar enzyme activity in spore extracts from strain COE190 (ΔcydAB) revealed that Nar activity was reduced by 70% compared to that of spore extracts of the wild type (Table 2). This suggests that cytochrome bd oxidase also influenced Nar activity in spores.

TABLE 2.

Nitrate reductase enzyme activity in extracts of spores and mycelium

Strain	Nitrate reductase activity (mU mg−1)a in:
	Spores	Mycelium
M145 (wild type)	79.66 ± 2.38	10.47 ± 3.61
NM92 (Δnar123)	0.6 ± 0.08	0.2 ± 0.4
COE192 (Δqcr-cta)	0.3 ± 0.04	6.98 ± 2.61
COE634 (COE192/pMS2148-56	74.92 ± 7.74	9.61 ± 3.16
COE502A (ΔctaD)	0.41 ± 0.23	ND
COE190 (ΔcydAB)	24.54 ± 0.66	15b
COE640 (NM92/pMSnar1)	132.66 ± 3.02	0.37 ± 0.26
COE648 (NM92/pNGnar1)	165.65 ± 63.69	2.17 ± 0.26
COE743 (COE192/pNGnar1)	0.42b	13.2 ± 0.7

^aAll data were taken from a minimum of three biological replicates, each measured in triplicate, except as noted. ND, not determined.

^bAverage taken from two biological replicates, with each sample assayed in triplicate.

Together, these data strongly suggest that loss of Nar1 enzyme activity was responsible for the absence of nitrate reduction in spores of the supercomplex mutant COE192. To demonstrate that Nar1 activity was indeed affected in spores of strain COE192, aliquots of spore extracts derived from different strains were separated by nondenaturing, clear-native PAGE, and the gel was subsequently stained for nitrate reductase enzyme activity (Fig. 2A). A diffuse “cloud” of enzyme activity characteristic of the migration pattern of Nar1 (7) was observed in the spore extract of the wild-type M145 and in that from strain NM68 (Δnar2 Δnar3), while an extract from the nar triple mutant NM92 (Δnar1 Δnar2 Δnar) showed no activity (Fig. 2A). Similarly, a spore extract derived from COE192 failed to show any Nar1 activity band, while in the extract from strain COE634, which represents strain COE192 complemented with the complete qcr-cta locus on the integrative plasmid pMS2148-56 (Table 3 and Fig. 1; see also Materials and Methods), the Nar1 activity band was restored. As a control to ensure that equal amounts of protein were loaded on the gel, we performed in parallel SDS-PAGE with aliquots of the same samples (Fig. 2A, bottom panel). A phenotype similar to that observed for spore extracts derived from COE192 was also observed for extracts of COE502A (Fig. 1), which lacked the CtaD subunit of the aa₃ oxidase (see Fig. S1 in the supplemental material). Together, these data indicate that Nar1 enzyme activity in spores is reliant on an active bcc-aa₃ supercomplex.

FIG 2.

Redox dye-dependent Nar1 enzyme activity is dependent on the bcc-aa₃ supercomplex in spore extracts. (A) Spore extracts (75 μg protein) of the indicated strains were separated in a nondenaturing polyacrylamide gel (10% [wt/vol] acrylamide) and subsequently stained for nitrate reductase enzyme activity, which appears as clear bands against a dark blue background (top panel), or were separated in a 3 to 16% SDS-PAGE gel and stained with Coomassie brilliant blue (bottom panel). The latter acted as a loading control for the upper gel. The locations of the Nar1 and Nar2 enzymes are indicated on the right side of the gel in the top panel. Strains are indicated at the top of the gel; extract derived from mycelium of the wild type was also applied to the gel (M145 mycel). (B) Western blot identifying NarG1. Polypeptides in the indicated spore extracts (60 μg of protein) were separated in a 7.5% (wt/vol) SDS-PAGE gel, and after transfer to nitrocellulose, the membrane was challenged with peptide antibodies raised against NarG1. The location of purified His-tagged version of NarG1-His (1 μg) (7) is shown on the right of the panel; the asterisk denotes an unidentified, cross-reacting polypeptide that acted as a loading control, and the molecular mass markers are shown on the left. WT, wild type.

TABLE 3.

Strains and vectors used in this study

Strain or vector	Genotype and/or characteristic(s)	Reference or source
Streptomyces coelicolor A3(2) strains
M145 (wild type)	SCP1− SCP2−	33
COE190	M145 ΔSCO3945–ΔSCO3946 (deletion of 2,525 bp removing cydAB)	26
COE192	M145 ΔSCO2148–ΔSCO2156 (deletion of 9,398 bp removing qcrCAB-ctaE-SCO2152-SCO2153-ctaCDF)	26
COE502A	M145 SCO2155::Tn5062	9
COE634	Like COE192, but with pMS2148-56 integrated in the chromosome	26
COE640	Like NM92, but with pMSnar1 integrated in the chromosome	This study
COE648	Like NM92, but with pNGnar1 integrated in the chromosome	This study
COE743	Like COE192, but with pNGnar1 integrated in the chromosome	This study
NM29	Lacking the nar1 and nar2 operons: ΔSCO6535–ΔSCO6532::aadA (deletion of 6,209 bp removing narG1-narH1-narJ1-narI1), ΔSCO0216–ΔSCO0219::aac(3)IV (deletion of 6,531 bp removing narG2-narH2-narJ2-narI2)	7
NM68	Lacking the nar2 and nar3 operons: NM3 SCO4947–SCO4950::aadA (deletion of 6,497 bp removing narG3-narH3-narJ3-narI3)	17
NM92	Lacking the nar1, nar2, and nar3 operons: NM90 SCO6532–SCO6535::aadA (deletion of 6,209 bp removing narG1-narH1-narJ1-narI1)	17
NM1821	M145 ΔmoaA::aac(3)IV	17
Escherichia coli strains
DH5α	λ− ϕ80dlacZΔM15 Δ(lacZYA-argF)U169 recA1 endA1 hsdR17(rK− mK−) supE44 thi-1 gyrA relA1	Laboratory stock
ET12567(pUZ8002)	dam mutant, dcm; with trans-mobilizing plasmid pUZ8002	35
Plasmids and cosmids
pMS82	ΦBT1 attP-int-derived integration vector for the conjugal transfer of DNA from E. coli to Streptomyces (Hygr)	30
pMS2148-56	pMS82 SCO2148–SCO2156 (with 200 bp of upstream and 200 bp of downstream sequence)	26
pMS2148-50	pMS82 SCO2148–SCO2150 (with 53 bp of upstream and 39 bp of downstream sequence)	9
pMS2148-51	pMS82 SCO2148–SCO2151 (with 150 bp of upstream and 39 bp of downstream sequence)	9
pMS2153-56	pMS82 SCO2153–SCO2156 (with 52 bp of upstream and 32 bp of downstream sequence)	This study
pMSnar1	pMS82 SCO6532–SCO6535 (with 200 bp of upstream of SCO6535 and 200 bp of sequence downstream of SCO2632)	This study
pNG2	PermE*, RBS, MCS, and fd-ter cloned into pNG1′/EcoRV/SpeI Hygra	31
pNGnar1	pNG2 SCO6532–SCO6535 (with 148 bp upstream of SCO6535 and 28 bp of sequence downstream of SCO2632)	This study

^aRBS, ribosome binding site; MCS, multiple cloning site; fd-ter, phage fd transcriptional terminator.

Finally, although a spore extract derived from strain COE190 (ΔcydAB) exhibited an activity band due to Nar1, this activity was qualitatively weaker than that observed for the wild type (Fig. 2A), which agreed with the reduction in Nar1 enzyme activity measured using reduced benzyl viologen (Table 2). Nevertheless, Nar1 retained activity and was detectable in the mutant background, which contrasted with what was observed for extracts of strain COE192 lacking the bcc-aa₃ supercomplex (Fig. 2A). We noted variability in the intensity of the Nar1 activity band in spore extracts of COE190, suggesting that this might be a consequence of impaired physiology of spores lacking the bd supercomplex or instability of the Nar1 enzyme.

In a reciprocal experiment designed to determine whether Nar1, or any other Nar, influenced cytochrome aa₃ oxidase activity, a spore extract of the triple nar operon mutant NM92 (Δnar1 Δnar2 Δnar) was analyzed for the presence of cytochrome aa₃ oxidase activity (Fig. S2). The result revealed that activity of the oxidase was qualitatively indistinguishable from that in spore extracts derived from the wild-type strain M145. This finding is also in agreement with the lack of an apparent growth phenotype of the triple nar operon mutant NM92 (17), which sharply contrasts with what was observed with strain COE192 (Δqcr-cta) (26). This result indicates that although the integrity and activity of the bcc-aa₃ supercomplex are essential for Nar1 activity in spores, a lack of Nar enzymes has no apparent negative impact on the activity of cytochrome aa₃ oxidase.

The NarG catalytic subunit of Nar1 is synthesized in spores of strain COE192.

Previous studies revealed that Nar1 is synthesized prior to completion of spore formation (7). How and when during development enzyme synthesis occurs are still unclear. In order to determine whether lack of Nar1 enzyme activity in spores of strain COE192 was due to lack of Nar1 enzyme synthesis, spore extracts were analyzed by Western blotting using peptide-specific antibodies raised against the catalytic subunit, NarG1 (7). The results revealed that synthesis of the NarG1 polypeptide was unimpaired in spore extracts of strain COE192 (Δqcr-cta) and was comparable to that in the spores of the wild-type strain M145 (Fig. 2B). Similar levels of NarG1 polypeptide were also detected in extracts of spores derived from NM68 (Δnar23) and COE190 (ΔcydAB). As expected, extracts of spores from the Nar-negative mutant NM92 (Δnar1 Δnar2 Δnar3) lacked a signal corresponding to NarG1 (Fig. 2B).

Typically, when Nar enzymes are synthesized, they usually exhibit enzyme activity in vitro using redox dyes except when they lack the bis-molybdenum guanidine dinucleotide (bis-MGD) cofactor (27, 28). To rule out that impaired molybdenum cofactor synthesis was the reason for the lack of Nar1 activity in spores of strains COE192 or COE502A, we analyzed qualitatively the activities of aldehyde dehydrogenase and xanthine oxidase/dehydrogenase, two further molybdoenzymes predicted from the genome sequence of S. coelicolor (29). While these enzymes belong to the molybdopterin (MPT) cofactor family (28), nevertheless, the majority of the biosynthetic steps on the pathways to MPT and bis-MGD are shared. Therefore, activity of these enzymes would provide an indirect assessment of whether bis-MGD was affected by the lack of the bcc-aa₃ supercomplex. Both aldehyde dehydrogenase and xanthine oxidase enzymes were detectable in spore extracts from both COE192 and COE502A (Fig. S3). As a control, neither activity could be detected in spore extracts of the moaA mutant NM1821, which is impaired in bis-MGD cofactor synthesis (17). These results ruled out that the general steps in molybdenum cofactor biosynthesis were affected by the mutations in strains COE192 and COE502A.

Manifestation of Nar1 activity requires the complete cytochrome bcc-aa₃ supercomplex.

The results of a recent study have revealed that the bcc complex and the cytochrome aa₃ oxidase function as a supercomplex in S. coelicolor in both spores and mycelium during aerobic growth (9). The lack of Nar1 enzyme activity in spores of strain COE502A (Table 2), where only the CtaD subunit of the aa₃ oxidase is missing, suggested that the cytochrome bcc complex alone is insufficient to support Nar1 activity and that the complete bcc-aa₃ supercomplex is required to support Nar1-dependent nitrate reduction in spores. To test this, plasmids carrying different portions of the qcr-cta locus were introduced into COE192, and these included pMS2148-50 (qcrBAC), pMS2148-51 (qcrBAC-ctaE), and, as a control, pMS2153-56 (ctaCDF) (Table 3) (9). Nar1 enzyme activity in spore extracts of the corresponding derivatives was determined (Fig. S4). Activity with pMS2148-50, encoding QcrB, QcrA, and QcrC of the cytochrome bcc complex, restored maximally 5% of Nar enzyme activity to spores of the mutant COE192. Neither pMS2148-51 nor the pMS2153-56 was able to restore Nar activity to the mutant. Only plasmid pMS2148-56 encoding the complete bcc-aa₃ supercomplex was capable of restoring Nar1 activity to spores (Fig. S4). These results demonstrate that a complete cytochrome bcc-aa₃ oxidase supercomplex is required for full Nar1-dependent nitrate reduction activity to be detected.

The bcc-aa₃ supercomplex is not essential for Nar2 activity or Nar3 synthesis in mycelium.

Nar2 is the main Nar enzyme active in exponentially growing mycelium, a cellular compartment where Nar1 is absent (7, 18). Nitrate reduction by intact exponentially growing mycelium of the wild-type M145 could be readily measured (Table 4) and attained a level of activity which is similar to previously reported values and which is due to Nar2 (17, 18). Intact mycelium of strain COE192 (Δqcr-cta) revealed an approximate 40% reduction in the level of Nar2-dependent nitrate reduction activity compared with the wild-type level (Table 4). Performing this experiment several times independently resulted in values varying between 30 and 90% of the wild-type level (data not shown), and this variability was reflected in the high standard deviation of the activity shown in Table 4. Analysis of Nar2 enzyme activity using the in vitro enzyme assay with reduced BV as an electron donor showed that extracts derived from mycelium of COE192 had an activity of roughly 65% of the level measured with extracts derived from M145 (Table 2).

TABLE 4.

Nitrite production by intact mycelium

Strain	Nitrite production (μmol/h/ml × 1,000 CAE)a
M145 (wild type)	7.31 ± 3.91
NM92 (Δnar123)	<0.001
COE192 (Δqcr-cta)	3.66 ± 2.05
COE634 (COE192/pMS2148-56)	ND
COE640 (NM92/pMSnar1)	0.02 ± 0.01
COE648 (NM92/pNGnar1)	0.33 ± 0.2
COE743 (COE192/pNGnar1)	1.44 ± 0.57

^aCAE, cell amount equivalents; ND, not determined.

To determine whether this reduction in Nar2-dependent nitrate reduction was a direct effect on the Nar2 enzyme, an in-gel activity stain with mycelium extracts derived from strains M145 (wild type), COE192 (Δqcr-cta), and NM92 (Δnar123) was performed (Fig. 3A). The results showed that Nar2 enzyme activity was readily detected in extracts of M145 and COE192 while, as expected, no Nar2 activity was observed in the nar triple mutant NM92. It could be ruled out that the weak activity band migrating just below the main Nar2 activity in the extract of COE192 was due to Nar1 enzyme activity because Western blotting failed to identify the catalytic subunit NarG1 in these extracts (Fig. S5). This suggests that this weak activity band is likely a degradation product of Nar2 that retains some enzyme activity.

FIG 3.

Nar2 enzyme activity is independent of the bcc-aa₃ supercomplex in extracts of mycelium. (A) Extracts of mycelium or spores (75 μg of protein) of the indicated strains were separated in a nondenaturing polyacrylamide gel (10% [wt/vol] acrylamide) and subsequently stained for nitrate reductase enzyme activity (top panel) or were separated in a 4 to 20% SDS-PAGE gel and stained with Coomassie brilliant blue (bottom panel). The latter acted as a loading control for the upper gel. The locations of the Nar1 and Nar2 enzymes are indicated. The strains included M145 (wild type), NM92 (Δnar123), COE192 (Δqcr-cta), COE648 (NM92/ pNGnar1), and COE743 (COE192/pNGnar1). (B) Western blot identifying the presence of NarG1 in the extracts of spores (65 μg of protein) or mycelium (180 μg of protein) that were separated by 4 to 20% (wt/vol) gradient SDS-PAGE; after transfer to nitrocellulose, the membranes were challenged with peptide antibodies raised against NarG1. The strains included M145 (wild type), NM92 (Δnar123), COE640 (NM92/pMSnar1), and COE648 (NM92/ pNGnar1). The location of purified His-tagged version of NarG1-His (1 μg) (7) is shown on the right, and the molecular mass markers are shown on the left.

Nar3 is synthesized in stationary-phase mycelium, and synthesis is repressed by a high phosphate concentration (19). Western blot analysis of extracts derived from stationary-phase mycelium of mutants COE192 (Δqcr-cta) and COE190 (ΔcydAB) (Fig. 1) revealed that both showed high levels of the catalytic subunit NarG3 and that synthesis was repressed by phosphate (Fig. S6). No NarG3 polypeptide could be detected in extracts of the triple nar mutant NM92.

Nar1 exhibits only low specific activity when constitutively synthesized in mycelium.

Nar1 is neither synthesized nor active in mycelium of wild-type strain M145 (7). To examine the effect of constitutively expressing the narG1-narH1-narJ1-narI1 genes in mycelium on nitrate reduction, the nar1 operon was cloned into two different integrative plasmid vectors. In the first plasmid, which acted as a control, the complete nar1 operon, along with 200 bp of upstream regulatory DNA sequence, was cloned into the ΦBT1-based integrative plasmid pMS82 (30). Expression of the nar1 operon was thus maintained under the control of its own promoter and delivered plasmid pMSnar1. In the second plasmid derivative, pNGnar1, the narG1-narH1-narJ1-narI1 genes were cloned into the expression vector pNG2 (31), placing the operon under the control of the strong, constitutive P_ermE* promoter (31) (Table 3; see also Materials and Methods). To demonstrate that the cloned narG1-narH1-narJ1-narI1 genes resulted in synthesis of a functionally active Nar1 enzyme, the triple nar deletion strain NM92 (Δnar1 Δnar2 Δnar3) was transformed with plasmid pMSnar1 or plasmid pNGnar1, and Nar1-specific nitrate reduction of the resting spores was analyzed (Table 1). The results clearly revealed that while spores of NM92 were incapable of reducing nitrate, introduction of either pMSnar1 or pNGnar1 restored nitrate reduction capability to spores of the transformed strain. Both plasmids restored nitrate reduction to a level similar to that observed for the wild type (M145) (Table 1). Measurement of the specific activity of Nar1 in spore extracts revealed that pMSnar1 restored activity to a level similar to that of the wild type and that pNGnar1 restored activity to a level approximately 2-fold higher than that of the wild-type M145 (Table 2). Analysis of in-gel Nar1 activity in spore extracts of NM92 transformed with pNGnar1 revealed a strong activity band of Nar1 compared to that of the wild type (Fig. 3A). As a control, pNGnar1 was also introduced into COE192 (Δqcr-cta), yielding strain COE743, and Nar activity was determined in spore extracts (Table 2). The Nar1-negative phenotype of the strain remained despite overexpression of nar1 operon.

Analysis of nitrate reduction by intact mycelium of the same strains revealed, however, that pMSnar1 failed to restore any nitrate-reducing activity to strain NM92, while pNGnar1 exhibited nitrite levels that were approximately 20-fold lower than the level measured for mycelium of wild-type strain M145 (Table 4). Because it is unknown how and whether nitrate for the Nar1 enzyme is transported into mycelium, we also analyzed Nar1 enzyme activity in vitro in extracts of mycelium from NM92 transformed with pNGnar1, strain COE648 (Table 2). The results showed that the specific activity of Nar1 in mycelium of COE648 was more than 80-fold lower than the specific activity of Nar1 in spore extracts of the same strain (Table 2). Finally, analysis of extracts of mycelium from strain COE648 failed to reveal a band corresponding to active Nar1 enzyme (Fig. 3A). Moreover, introduction of pNGnar1 into COE192 (Δqcr-cta) also failed to resolve any Nar1 activity band and revealed only the Nar2 enzyme activity band (Fig. 3A).

Western blot analysis of extracts of spores and mycelium from the transformed strains revealed that the NarG1 catalytic subunit of Nar1 was detected in spore extracts of strain NM92 transformed with either pMSnar1 or pNGnar1 (Fig. 3B, left). Analysis of extracts of mycelium revealed that the NarG1 polypeptide could only be detected in extracts of strain COE648 (NM92/pNGnar1) (Fig. 3B, right). Note that approximately 3-fold more protein was applied to the lane with the extract from mycelium than to that with the extract from wild-type spores (Fig. 3B), indicating that, despite being synthesized, the NarG1 polypeptide was present at a low level in mycelium compared with the level present in spores. Introduction of pNGnar1 into COE192 (Δqcr-cta) also revealed the presence of NarG1 in Western blotting of a mycelium extract (Fig. S4); however, despite overexpression of the nar1 operon in the strain, nitrate reduction could only be partially restored to mycelium of COE192 (Table 4).

DISCUSSION

Respiratory Nar enzymes generally function as quinol oxidases. In this study, we demonstrate that the spore-specific Nar1 enzyme of S. coelicolor is an exception and is dependent on a complete and functional bcc-aa₃ supercomplex. We also showed that spores of a mutant lacking the high-affinity cytochrome bd oxidase had reduced enzyme activity in vitro but surprisingly not in whole, resting spores. Whether this might be due to some stabilizing effect conferred upon the bcc-aa₃ supercomplex and, consequently, Nar1 by the cytochrome bd oxidase will require further detailed study.

We could rule out that the inability to reduce nitrate by spores of qcr-cta mutants was due to impaired nitrate uptake or to a defect in molybdenum cofactor biosynthesis. Moreover, because removal of only the catalytic CtaD subunit of the oxidase by deletion of the corresponding gene was sufficient to impede nitrate reduction, this supports our working hypothesis that the complete bcc-aa₃ supercomplex is necessary for nitrate reduction to be observed. This means that identification of the branch point within the supercomplex where electrons are diverted to Nar1 will require future detailed spectroscopic analysis of the purified complexes. Thermodynamically, the bcc complex component of the supercomplex should suffice to act as an electron conduit to the diheme-containing NarI membrane anchor subunit of Nar1. This hypothesis would be in accord with results of two recent reports where it was shown that W. succinogenes and C. jejuni, both members of the subphylum Epsilonproteobacteria and both of which have a membrane-associated diheme c-type cytochrome, QcrC, are capable of transferring electrons from a bcc complex to a periplasmic nitrous oxide reductase and nitrate reductase, Nap, respectively (24, 25). The requirement for an intact supercomplex in S. coelicolor may be simply a question of ensuring the stability and functionality of the bcc complex. Indeed, the recently proposed cryo-electron microscopy-based structure of the supercomplex from M. smegmatis (14, 15) reveals that the aa₃ oxidase complex packs tightly against the bcc complex and may afford enhanced stability to the intact supercomplex.

Due to the important role of the [2Fe-2S] cluster in the Rieske component, QcrA, for the Q cycle to function (16, 22), it is conceivable that the diheme QcrC is likely to be the earliest point at which electrons might be delivered to the heme b_D of the NarI membrane anchor (Fig. 4, model). This would also be commensurate with the role of the bcc complex in the Epsilonproteobacteria and with the conformational switch observed for QcrC in the recently solved structure of the mycobacterial supercomplex (15).

FIG 4.

Model of how Nar1 might be coupled to the bcc-aa₃ supercomplex. To aid clarity, the bcc (orange) and aa₃ (blue) components of the heterotetrameric supercomplex (bcc-aa₃)₂ are depicted with the left half transparent. The Nar1 enzyme (NarGHI) is superimposed on the complex with the membrane-spanning, heme b-containing (blue boxes) NarI subunit positioned in proximity to the dihemes of QcrC (blue boxes on an orange background). The dotted black line indicates the alternative paths of electron transfer through the complex. Proton translocation via the Q cycle is shown as solid red lines while proton-pumping activity of the aa₃ oxidase is indicated by the dotted red line. MK and MKH₂ signify menaquinone and menaquinol, respectively; bisMGD is the molybdenum cofactor of Nar1.

We could also show that while the mycelium-localized Nar2 enzyme exhibited moderately reduced activity in a qcr-cta mutant compared with the level in the wild-type strain, activity was nevertheless clearly detectable and, thus, independent of the bcc-aa₃ supercomplex. This suggests that the Nar2 enzyme probably functions as a more classical menaquinol oxidase.

Our attempts to synthesize active Nar1 enzyme in mycelium met with only partial success. Only low levels of nitrate-reducing activity could be determined for the enzyme in mycelium. This would be consistent with the Nar1 enzyme being unable to accept electrons directly from menaquinol but would imply at the same time that the bcc-aa₃ complexes are not identical in spores and mycelium, a suggestion made in another recent study (9). While further studies will be required to determine what makes the Nar1 enzyme dependent on the bcc-aa₃ supercomplex in order to be able to reduce nitrate, it will also be important to determine whether the composition of the bcc-aa₃ supercomplex is different in these cellular compartments.

What advantage would electron transfer to Nar1 from the bcc-aa₃ supercomplex, rather than directly from menaquinol, confer upon the spore? Assuming a proton motive force (Δp) of between 180 and 200 mV (16) across the spore cytoplasmic membrane, taking into consideration the redox span of approximately 500 mV for menaquinone/menaquinol (MK/MKH₂) (redox potential [E_m] = −75 mV) and NO₂⁻/NO₃⁻ (E_m = +430 mV), and coupling this with the Q cycle of the bcc complex (Fig. 4), this should allow translocation of 4 H⁺ ions per 2e⁻ compared with the scalar release of 2 H⁺ ions across the membrane when Nar functions as a menaquinol oxidase (27). While translocation of 4 H⁺ ions per nitrate reduced is less favorable in terms of PMF generation than the proposed 6 H⁺ ions translocated by the bcc-aa₃ supercomplex when coupled to O₂ reduction (14, 15, 32), this is nevertheless energetically beneficial for spores, which presumably have a considerably lower metabolic rate than mycelium to aid their decades-long survival (1). Combining nitrate reduction with Q-cycle-driven proton translocation would provide spores with a more efficient means of maintaining their membrane potential and PMF when O₂ becomes limiting.

MATERIALS AND METHODS

Bacterial strains and culture conditions.

The strains and plasmids used in this study are listed in Table 3. Growth media and culture conditions for S. coelicolor and E. coli have been described previously (33, 34). Strains of S. coelicolor A3(2) were grown on soya flour-mannitol (SFM) or on Difco nutrient broth (DNB) agar medium as indicated by Kieser et al. (33).

Spores were harvested from SFM agar plates after 8 days of growth at 30°C and filtered through absorbent cotton to separate aerial hyphae and spores. The resulting spore suspension was washed once using 100 mM potassium phosphate buffer, pH 7.2. Spores were resuspended in 200 to 400 μl of TE buffer (10 mM Tris-HCl, 1 mM EDTA, pH 8.0) and disrupted using a bead beater. Subsequently, cell debris and glass beads were separated from the crude extract by two centrifugation steps (4°C, 15,000 rpm, 5 min) as described previously (7).

For cultivation in liquid medium, Streptomyces strains were grown in tryptic soy broth (TSB; Oxoid) or in DNB broth supplemented with appropriate antibiotics to maintain selection. S. coelicolor A3(2) strains were grown as highly disperse liquid cultures in Duran-F tubes with morpholinepropanesulfonic acid (MOPS)-buffered TSB as described before (18).

E. coli DH5α (Stratagene) was used as a host for plasmid constructions. E. coli ET12567/pUZ8002 (35) is the nonmethylating plasmid donor strain used for intergeneric conjugation with S. coelicolor strain M145 (34). Apramycin (25 μg ml⁻¹), carbenicillin (100 μg ml⁻¹), chloramphenicol (25 μg ml⁻¹), kanamycin (25 μg ml⁻¹), spectinomycin (25 μg ml⁻¹), or hygromycin (Hyg; 25 μg ml⁻¹), all from Sigma, was added to growth medium when required.

Construction of plasmids.

Complementation plasmids pMSnar1 and pNGnar1 included the complete narG1-narH1-narJ1-narI1 operon (7, 29) but differed in that pMSnar1 included 200 bp of upstream nontranslated regulatory region. To construct pMSnar1, a 6,909-bp DNA fragment including the coding regions for SCO6532 through SCO6535 (plus 200 bp of upstream regulatory DNA) was amplified as two distinct but partially overlapping, DNA fragments (fragment 1, 3,895 bp; fragment 2, 3,014 bp) by PCR using oligonucleotides pMS_nar1_fw and pMSAnnealFrag1_rv for fragment 1 and pMSAnnealFrag2_fw and pMS_nar1_rv for fragment 2 (see Table S1 in the supplemental material). The two DNA fragments were cloned in series between the HindIII and NsiI restriction sites of plasmid pMS82 (30) using an NEBuilder kit, exactly as described by the manufacturer (New England Biolabs) (Table 3). A similar 6,685-bp DNA fragment including the complete nar1 operon was amplified to construct pNGnar1. For this insert, fragment 1 (3,300 bp) was amplified using oligonucleotides pNG2_nar1_fw and pNGAnnealFrag1_rv, and fragment 2 (3,385 bp) was amplified using oligonucleotides pNGAneealFrag2_fw and pNG2_nar1_rv (Table S1). The resulting DNA fragments were then cloned in series between the NdeI and SpeI restriction sites of plasmid pNG2 (31), to deliver pNGnar1.

The authenticity of the cloned fragments in both complementation plasmids was verified by DNA sequencing, and plasmids were introduced into the indicated strains of S. coelicolor via conjugation using the plasmid-containing E. coli strain ET12567/pUZ8002 (Table 3) (35).

Purification of His-tagged NarG1.

The His-tagged NarG1 large subunit of the Nar1 enzyme was purified according to Fisher et al. (7).

Measurement of enzyme activity.

Nitrate reduction, as determined by measuring excreted nitrite, in resting spores and mycelium was essentially determined colorimetrically as previously described (7, 17, 36).

Nitrate reductase enzyme activity in crude extracts was determined using reduced dithionite and benzyl viologen (0.4 mM) as described previously (17, 37). The protein concentration was determined according to the method of Lowry et al. (38) with bovine serum albumin as a standard. Generally, unless indicated otherwise, all enzyme assays were performed with a minimum of three biological replicates, each measured in triplicate.

Qualitative staining for nitrate reductase enzyme activity after nondenaturing PAGE was performed using 10% (wt/vol) polyacrylamide gels, pH 8.5, and included 0.1% (wt/vol) Triton X-100 in the gels (7). Occasionally, commercially purchased gradients gels (Eurogentec), including 4 to 20% polyacrylamide, were used. Samples (typically, 75 μg of protein) were incubated with 1% (wt/vol) Triton X-100 prior to application to the gels. Nar activity was detected in the gels after electrophoresis as described previously (17).

In-gel staining for cytochrome c oxidase enzyme activity was done according to Sabar et al. (39), as modified by Fisher et al. (26). Briefly, protein complexes in extracts or membrane fractions derived from spores or mycelium were separated by native PAGE (10% [wt/vol] polyacrylamide). Cytochrome aa₃ oxidase activity was visualized by incubating the gel in 10 ml of 50 mM potassium phosphate buffer, pH 7.2, containing 1 ml of cytochrome c as the substrate (stock solution, 10 mg ml⁻¹) and 0.5 ml of diaminobenzidine (DAB) reagent (stock solution of 10 mg ml⁻¹). Gels were incubated at room temperature (RT) usually for between 1 and 2 h.

In-gel staining for aldehyde dehydrogenase and xanthine oxidase enzymes activities were performed as described by Barabás et al. (40) and Özer et al. (41), with slight modifications. To detect xanthine oxidase activity, after separation of protein complexes by 10% (wt/vol) nondenaturing PAGE, the gels were developed in 50 mM Tris-HCl, pH 7.4, containing 0.5 mM xanthine, 1 mM 3-(4,5-dimethylthiazol-2-yl)-2,5-diphenyltetrazolium bromide (MTT), 0.1 mM phenazine methosulfate (PMS), and 30 mM tetramethylethylenediamine (TEMED) at RT for 30 min. Aldehyde dehydrogenase activity was detected by incubating the gel in 0.1 M Tris-HCl, pH 7.4, including 0.1 mM PMS, 1 mM MTT, and 20 mM acetaldehyde at RT for 30 min. All chemicals used for the enzyme assays were obtained from Sigma-Aldrich.

SDS-PAGE and Western blotting.

Aliquots (typically 25 to 75 μg of protein) from crude extracts derived from spores or fresh mycelium were separated by SDS-polyacrylamide gel electrophoresis (PAGE) using 7.5% or 10% (wt/vol) polyacrylamide gels (42) and transferred to nitrocellulose membranes as described previously (43). Occasionally, precast 4 to 20% gradient gels (Serva Electrophoresis, GmbH) were used. Affinity-purified peptide antibodies raised against NarG1 were generally used at a dilution of 1:50 unless specified otherwise. Secondary antibody conjugated to horseradish peroxidase was obtained from Bio-Rad. Visualization was done by the enhanced chemiluminescent reaction (Stratagene).