EPR characterization of the new Qrc complex from sulfate reducing bacteria and its ability to form a supercomplex with hydrogenase and TpI<i>c</i> ₃

Abstract

NiFe Hydrogenase physically interacts with Qrc by native page (View interaction). In the main group of sulfate-reducing bacteria (SRB), of the Deltaproteobacteria class, most organisms are characterized by the presence of a high number of multiheme cytochromes c, of which the most abundant and well characterized is the tetraheme Type I cytochrome c 3 (TpIc 3). In addition, this group has a considerable number of membrane-associated redox complexes containing a cytochrome c subunit, which are responsible for quinone reduction or transmembrane electron transfer for the cytoplasmic reduction of sulfate (reviewed in [1]). Recently, a new respiratory membrane complex present in SRB was described and named Qrc (for Quinone–reductase complex) [2]. Qrc was shown to transfer electrons from the TpIc 3 to the menaquinone pool, with the periplasmic hydrogenases (Hases) and/or formate dehydrogenases (Fdhs) as primary electron donors. The presence of Qrc is associated with the presence of Hases and/or Fdhs lacking a membrane subunit for direct quinone reduction. Qrc is the only described respiratory complex in SRB with a quinone reductase function, and it is essential for growth on H2 or formate, as recently shown in D. desulfuricans G20 [3]. Qrc comprises three subunits in the periplasm, namely a membrane-anchored hexaheme cytochrome c (QrcA), a periplasmic-facing large protein of the molybdopterin oxidoreductase family (QrcB), and a periplasmic iron–sulfur four cluster protein (QrcC), and also an integral membrane subunit of the NrfD family (QrcD) [2]. The three QrcBCD subunits are closely related to the bacterial respiratory enzymes of the complex iron-sulfur molybdoenzyme (CISM) family [4], although no molybdopterin is present in QrcB. The absence of Mo is also observed in the Nqo3/NuoG subunit of the respiratory complex I [5], and in the related protein of the alternative complex III (ACIII) [6-8]. ACIII, a functional substitute of the bc 1 complex, is closely related to Qrc [2, 6], but has a higher number of subunits and acts reversibly of Qrc, by oxidising menaquinol [9, 10]. Previous EPR studies of Qrc, allowed the detection of six low-spin hemes c, one [3Fe–4S]1+/0 cluster and at least three interacting [4Fe–4S]2+/1+ clusters [2]. Here, we report a more detailed study of the Qrc hemes and Fe–S centers through potentiometric titrations followed by EPR spectroscopy, providing a better characterization of the redox centers responsible for electron transfer in Qrc. In addition, we report that Qrc can form a supermolecular structure (supercomplex) with Hase and TpIc 3, its physiological electron donors. Qrc was purified from D. vulgaris Hildenborough as previously described [2]. A potentiometric titration was performed inside an anaerobic chamber (95% Ar, 5% H2) at 20 °C using 40 μM of Qrc and 70 μM of a mixture of redox mediators [2] in 100 mM MOPS pH 7, 5 mM EDTA. The redox potential spontaneously decreased slowly down to −200 mV, and sodium dithionite was used after that. Samples were transferred to EPR tubes under anaerobic conditions, capped and immediately frozen in liquid nitrogen upon removal from the chamber. EPR spectra were obtained using a Bruker EMX spectrometer equipped with an ESR-900 continuous flow of helium cryostat. Spectra were recorded under the following conditions: microwave frequency, 9.38 GHz; microwave power, 20.1 mW (for the Fe–S centers) and 2.4 mW (for the hemes); modulation frequency, 100 kHz; modulation amplitude, 1 mT; temperature, 10 K. Proteins were separated on 10% polyacrylamide Tricine–SDS–PAGE or 7–13% high resolution Clear Native (hr CN–PAGE) containing 0.02% (wt/vol) dodecyl maltoside (DDM), with running buffer also containing 0.02% DDM and 0.05% deoxycholic acid, as described in [11]. Gels were stained with Coomassie Blue, with heme-staining using 3,3′-dimethoxybenzidine dihydrochloride [12] or for hydrogenase activity [13]. Bands from the first-dimension CN–PAGE were excised and used for a second-dimension Tricine–SDS–PAGE, after denaturation by incubation with 5% SDS and 5% β-mercaptoethanol at 37 °C for 1 h, followed by a washing step, and thiol blockage with a 4% iodoacetamide solution. The supercomplex band from the CN–PAGE was subjected to in-gel tryptic digestion followed by peptide identification in a MALDI-TOF/TOF analyzer (Applied Biosystems 4800). The data were analyzed in a combined mode using Mascot search engine and NCBI database. EPR spectroscopy is a most useful tool to selectively characterize the redox properties of both hemes and Fe–S centers of respiratory proteins. A few detailed EPR studies on CISM enzymes have been reported, which were most informative regarding the properties of the Fe–S clusters (reviewed in [4]). These clusters provide a linear pathway to transfer electrons between the quinone pool and the catalytic site. To generate samples for the EPR study, we titrated Qrc in the presence of redox mediators, inside an anaerobic chamber containing 5% H2. It was observed that the redox potential of the sample slowly decreased without addition of any reductant down to ∼−200 mV. This decrease was due to trace amounts of Hase in the Qrc sample (see below), which slowly reduced the complex, but this decrease was slow enough to allow us to take several samples at defined redox potentials. The Qrc hemes were reported to give rise to three EPR signals with g max of 3.21, 2.93 and 2.82 (Fig. 1 A), which by spectral simulation account for 3, 2 and 1 hemes, respectively [2]. The intensity of these peaks was followed along the redox titration. The data points for the signal with g max = 3.21 were simulated by the sum of three Nernst equations with midpoint redox potentials (E m) of +60 and −100 mV in a 1:2 ratio approximately (Fig. 1B), in accordance with the fact that this signal corresponds to 3 hemes. The data points for the signal with g max = 2.93 could be simulated with two different redox potentials with 50% contribution each, with E m values of +100 and −130 mV (Fig. 1C). The low intensity of the signal with g max = 2.82, due to one heme, together with partial overlap with the 2.93 signal, did not allow the determination of a satisfactory redox titration curve. In the oxidized state a strong EPR signal at g ∼ 2.01 due to a [3Fe–4S]1+/0 center is also observed (Fig. 2 , inset), as described in other CISM and related complexes such as NarGHI nitrate reductase [14], FrdABCD fumarate reductase [15], or ACIII [16]. This signal titrates with E m value of +160 mV (Fig. 2), which is close to the +140 mV or +180 mV redox potential determined for the [3Fe–4S]1+/0 cluster of ACIII and NarGHI, respectively [14, 16]. In Qrc, the [3Fe–4S]1+/0 center is close to the menaquinone binding site [2]. At least three [4Fe–4S]2+/1+ clusters were reported in D. vulgaris Qrc [2]. These clusters give rise to a complex EPR signal due to magnetic interactions, as reported for other four cluster proteins [17]. Five signals with g-values of 2.07, 2.01, 1.98, 1.93 and 1.91 (Fig. 3 A) were monitored in the redox titration. The data points obtained from signals g 2.07, 1.98 and 1.91 can be fitted with a single component, with E m = −150 mV for the first two and −130 mV for the latter. For both signals with g 2.01 and 1.93, two components have to be considered with E m of −130 and −240 mV for g = 2.01 and −150 mV and −230 mV for g = 1.93 (Fig. 3). These data suggest the presence of three clusters, two of which with E m of −130 to −150 mV and the third with a lower E m of −240 mV. This is supported by a plot of the total signal intensity versus the potential, which can be fitted with these same three potentials (Fig. 3G). Thus, one of the Qrc [4Fe–4S]2+/1+ clusters has a lower potential than the other two, a similar situation to what is described for CISM enzymes [17]. The midpoint potentials of the Qrc cofactors have values that are suitable for electron transfer from the low redox potential TpIc 3 (−325 to −170 mV) to the menaquinone (E m = −70 mV). All the potentiometric results are summarized in Table 1 . ND: not determined. During the purification of Qrc, we observed that it was difficult to separate it from the [NiFe] Hase, as they eluted together in the first two ion-exchange columns, and also in a subsequent gel filtration. We could only separate these two proteins using a metal chelating IMAC column (Supplementary Fig. 1). This suggested the presence of a strong interaction between Qrc and Hase, which could result in the formation of a supercomplex. Analysis of the protein fractions by high resolution clear native gels (hr CN–PAGE) [11] stained for H2 uptake activity, indeed revealed the presence of a high molecular mass band, besides the band for the isolated Hase at 87 ± 9 kDa (Fig. 4 A). This band, with an estimated molecular mass of 360 ± 10 kDa, was found in fractions containing both Qrc and the Hase, but was no longer detected in the fractions from the IMAC column, which can separate the two proteins. The band for the isolated Qrc complex is present at 250 ± 24 kDa and stains with Coomassie and for hemes (Fig. 4B), but does not stain for H2 uptake activity, while the supercomplex band is very well detected by Hase activity staining (which is more sensitive), but stains weakly for hemes. To define the composition of the supercomplex, a two dimension Tricine–SDS–PAGE was performed on the 360 kDa band (Fig. 5 B). Several subunits from the Qrc complex were identified, as well as the large subunit of the Hase, by comparison with the 2D gel of the isolated Qrc and Hase native gel bands. Furthermore, the 360 kDa band was excised from the hr CN–PAGE gel, digested with trypsin and analyzed by mass spectrometry. Two proteins were clearly identified to be present: the large subunit of Qrc (QrcB) and the large subunit of the soluble periplasmic [NiFe] Hase (HynA1) (Supplementary Table 1), showing that this high molecular mass band is indeed formed by a supercomplex between Qrc and the [NiFe] Hase. The identification of this Hase was also confirmed by Western blot (Supplementary Fig. 1). The electron transfer between Hase and Qrc was shown to be functional only in the presence of the TpIc 3 cytochrome [2], the Hase physiological electron acceptor. Like the Hase, the TpIc 3 was detected in the Qrc fraction from the second Q-Sepharose column as a heme-staining band of 15 kDa (Supplementary Fig. 1). Additionally, a tenuous band of this molecular mass was also observed in the 2D gel of the supercomplex, but was absent in the Qrc and Hase controls of the 2D gel (Fig. 5B). Moreover, the molecular weight difference between the supercomplex band and the sum of the Hase plus the Qrc native bands, estimated from several CN gels, is about 22 kDa, which likely corresponds to the TpIc 3. Altogether, these results indicate that the 360 kDa supercomplex is constituted of three proteins: Qrc, [NiFe] Hase and TpIc 3. The formation of organized respiratory enzyme complexes in supermolecular structures has been well described in mitochondria [18] (sometimes called respirasomes), and also in bacteria [19]. In summary, the redox studies reported here provide the first detailed characterization of a new bacterial respiratory complex, and support the function of Qrc in electron transfer between the low redox potential TpIc 3 and menaquinone. The assembly of the [NiFe] Hase, TpIc 3 and Qrc in a supermolecular structure is likely to optimize the oxidation of H2, a key step in energy metabolism of SRB. We thank João Carita for cell growth and Ana Coelho for mass spectrometry data from the Analytical Services Unit of ITQB/IBET. This work was supported by research grants (PTDC/BIA-PRO/68486/2006 and PTDC/QUI-BIQ/100591/2008) and S.S.V. is recipient of a fellowship (SFRH/BD/30648/2006) both from Fundação para a Ciência e Tecnologia, Portugal. Supplementary data associated with this article can be found, in the online version, at doi:10.1016/j.febslet.2011.05.054. Supplementary data 1. Please note: The publisher is not responsible for the content or functionality of any supporting information supplied by the authors. Any queries (other than missing content) should be directed to the corresponding author for the article.