*3.3. Nitrogen Metabolism*

*Hrs. convoluta* is strongly diazotrophic [21], and as in *Hbt. modesticaldum*, genes for nitrogen fixation are grouped into a single *nif* gene cluster containing *nifI1*, *nifI2*, *nifH*, *nifD*, *nifK*, *nifE*, *nifN*, *nifX*, *fdxB*, *nifB*, and *nifV* [6]. Each of these genes shows between 63% and 93% sequence identity and analogous gene synteny to corresponding genes in *Hbt. modesticaldum*. A study with *Paenibacillus* sp. WLY78—also an endospore-former within the phylum *Firmicutes*—concluded that nine genes (*nifB*, *nifH*, *nifD*, *nifK*, *nifE*, *nifN, nifX*, *hesA*, *nifV*), which were grouped into a single gene cluster, are essential to synthesize a catalytically-active nitrogenase for dinitrogen assimilation [45]. All of these nitrogen fixation genes, except for *hesA*, were identified in the *Hrs. convoluta* and *Hbt. modesticaldum* genomes. Since HesA is proposed to play a role in metallocluster biosynthesis [45], it is possible that a gene (FTV88\_2056) located outside of the *nif* gene cluster and encoding a putative dinitrogenase Fe/Mo cofactor biosynthesis protein fills this role in *Hrs. convoluta*. This encoded protein showed high (~64%) sequence identity to a corresponding protein in *Hrs. acidaminivorans* and over 50% sequence similarity to that from a variety of nonphototrophic *Firmicutes*, but it showed no significant similarity to proteins encoded by *Hbt. modesticaldum*.

Research on *Hbt. modesticaldum* revealed changes in expression levels of numerous genes essential for various metabolic, biosynthetic, and other cellular pathways when the organism was grown under N2-fixing conditions [46]. This diazotrophic effect likely exists in other heliobacteria as well, including *Hrs. convoluta*. In terms of regulation of nitrogen fixation genes, however, it is interesting that neither *orf1* nor *nifA*, which encode regulatory proteins for the expression of *nif* structural genes [47,48], could be identified in the *Hrs. convoluta* genome. In *Hbt. modesticaldum*, the *orf1* gene product likely regulates the expression of *nif* genes when levels of fixed nitrogen are too low to support non-diazotrophic growth of the organism [16,48]. It is possible that *Hrs. convoluta* lacks the *orf1* and *nifA* regulatory genes and instead employs only *nifI1* (FTV88\_2453) and *nifI2* (FTV88\_2454) to coordinate post-translational regulation of nitrogenase [49].

*Hrs. convoluta* and *Hbt. modesticaldum* both contain gene clusters (*hypABCDEF* and *hupCDLS*) that encode an uptake [NiFe] hydrogenase that can putatively catalyze the oxidation of H2 produced during nitrogen fixation [6] (Figure 3). The arrangemen<sup>t</sup> of these genes in *Hrs. convoluta* is identical to that reported for *Hbt. modesticaldum* [6], being organized into a single cluster instead of dispersed throughout different regions of the chromosome, as has been observed in the genomes of other *Firmicutes* (Figure 4).

**Figure 4.** Comparison of uptake [NiFe]-hydrogenase genes in related *Firmicutes*. The genes are concatenated within a single region in *Heliorestis convoluta*, but they are dispersed in different regions of the *Desulfitobacterium hafniense* chromosome. Colors: blue, [NiFe]-hydrogenase structural genes; purple, hydrogenase expression/formation; red, hydrogenase assembly/maturation. Adapted from Sattley et al. [6]. *J. Bacteriol.* **2008**, *190*, 4687–4696. Copyright 2008 American Society for Microbiology.

In addition to performing N2 fixation, cells of *Hrs. convoluta* strain HH<sup>T</sup> could assimilate ammonia, glutamine, and asparagine as nitrogen sources [21]. Accordingly, genes encoding the ammonium transporter protein Amt (FTV88\_2595) and enzymes of the glutamine synthetase-glutamate synthase pathway, which incorporates ammonia in the formation of glutamine from glutamate [50,51] (Figure 3), were identified in the *Hrs. convoluta* genome. Following transport, glutamine can then be used for purine biosynthesis or, through the activity of NADPH-dependent glutamate synthase, can be condensed with α-ketoglutarate to yield two molecules of glutamate for other biosynthetic pathways [50,51]. In addition, a gene encoding NADP-specific glutamate dehydrogenase (FTV88\_2506) enables *Hrs. convoluta* to assimilate ammonia when synthesizing glutamate directly from α-ketoglutarate (Figure 3). Finally, genes encoding a glutamine-hydrolyzing asparagine synthetase (FTV88\_1161 and FTV88\_3319), which converts asparagine and glutamate into aspartate and glutamine, respectively (Figure 3), allow for the use of asparagine as a nitrogen source. In contrast, aspartate and glutamate cannot serve as nitrogen sources for strain HH<sup>T</sup> [21]. Taken together, these findings sugges<sup>t</sup> that, although the reactions are generally considered reversible, the enzymes catalyzing the conversion of asparagine to aspartate and glutamine to glutamate are physiologically unidirectional, strongly favoring the formation of aspartate and glutamate, respectively (shown as bolded arrows in Figure 3).

#### *3.4. Assimilation of Sulfur*

Growth studies indicate *Hrs. convoluta* is capable of assimilatory sulfate reduction [21]. Consistent with these observations, genomic analyses revealed that the pathway of assimilatory sulfate reduction in *Hrs. convoluta* begins with sulfate uptake using a sulfate/thiosulfate ABC transporter (*cysAWTP*). Typically, an enzyme encoded by *cysD* and *cysN*, sulfate adenyltransferase, catalyzes the assimilation of sulfate as adenosine phosphosulfate (APS) [52,53]. *Hrs. convoluta* lacks *cysD*, but genes encoding the bifunctional enzyme CysN/CysC (FTV88\_1460 and FTV88\_1458, respectively), which can also perform this function [54], are present. As shown in Figure 3, adenylyl-sulfate kinase (*cysC*) and phosphoadenylyl-sulfate reductase (*cysH*, FTV88\_1461) catalyze the subsequent reaction to yield sulfite [52,53].

To produce sulfide for amino acid biosynthesis, sulfite must undergo further reduction through sulfite reductase [53]. However, a gene encoding sulfite reductase could not be identified, suggesting that *Hrs. convoluta* may employ an unusual reductase or an alternative mechanism to perform this reaction. Genes encoding all successive enzymes necessary to synthesize cysteine, homocysteine, and methionine from hydrogen sulfide were identified (data not shown). By comparison, the *Hbt.* *modesticaldum* genome lacked *cysN*, *cysH*, and sulfite reductase, supporting physiological studies indicating that *Hbt. modesticaldum* requires a reduced sulfur source for biosynthetic purposes [16].

Interestingly, cultures of *Hrs. convoluta* strain HH<sup>T</sup> were able to grow well in the presence of high levels of sulfide (10mM), with sulfide oxidation accompanied by the production of elemental sulfur globules during growth [21]. However, the pathway for this reaction remains unclear, as the *Hrs. convoluta* genome appears to lack genes encoding traditional sulfide oxidoreductases, such as the sulfide:quinone oxidoreductase (SQR) from the green sulfur bacterium *Chlorobaculum* (*Chlorobium) tepidum* that oxidizes H2S to S0 and reduces quinone [55], or sulfide:flavocytochrome *c* oxidoreductase from the purple sulfur bacterium *Allochromatium vinosum* that oxidizes sulfide to sulfur or polysulfides [56]. Thus, it is possible that *Hrs. convoluta* contains a novel sulfide oxidoreductase for this purpose.

#### *3.5. Photosynthesis Genes and Pigment Biosynthesis*

Heliobacteria synthesize bacteriochlorophyll (BChl) *g*, a pigment absorbing light maximally between 785 and 790 nm, for phototrophic growth [4]. Accordingly, genes encoding enzymes that catalyze the conversion of glutamic acid to divinyl protochlorophyllide (*gltX*, *hemALBCDEN,* and *bchIDHME*) for pigment biosynthesis (Figure 5) were annotated in *Hrs. convoluta*. However, as for *Hbt. modesticaldum*, neither of the genes encoding protoporphyrinogen oxidase (*hemY* or *hemG*), which catalyzes the oxidation of protoporphyrinogen to protoporphyrin, was identified in the *Hrs. convoluta* genome. Moreover, comparisons with *hemG* from *Escherichia coli* and *hemY* from *Bacillus subtilis* yielded no significant sequence identity to genes in the *Hrs. convoluta* genome. Due to the anaerobic nature of *Hrs. convoluta*, an alternative and unidentified enzyme likely acts as a dehydrogenase rather than an oxidase in this step of pigment biosynthesis. Studies with *Desulfovibrio gigas*, also a strict anaerobe, sugges<sup>t</sup> that electron carriers, such as flavins and pyridine nucleotides, or electron-transport complexes, such as nitrite and fumarate reductases, do not use O2 as the electron acceptor in the conversion of protoporphyrinogen to protoporphyrin [57,58]. More recently, however, an alternative pathway that does not use protoporphyrin to synthesize heme has been described in *Hbt. modesticaldum* [59], and a similar mechanism likely exists in *Hrs. convoluta*.

Following the synthesis of divinyl protochlorophyllide in *Hrs. convoluta*, genes encoding protochlorophyllide reductase (*bchLNB*), chlorophyllide reductase (*bchXYZ*), and bacteriochlorophyll synthase (*bchG*) are present to facilitate catalysis of subsequent reactions and produce BChl *g*. Previous work with *Hbt. modesticaldum* suggested the need for an isomerase in the interconversion between 8-vinyl bacteriochlorophyllide *a* and bacteriochlorophyllide *g* [6], but more recent experimental work with this species revealed the ability of chlorophyllide reductase to perform both reduction and isomerization of divinyl chlorophyllide *a* and circumvent the need for a separate isomerase in the biosynthesis of bacteriochlorophyllide *g* [60,61].

Heliobacteria also contain an alternative form of chlorophyll (Chl) *a*, 81-OH-Chl *a*, which was observed as a smaller absorption peak at 672 nm in spectrophotometric studies of *Hrs. convoluta* [21]. Whereas BChl *g,* a bacteriochlorin-type chlorophyll, is reduced at the C-7 and C-8 bond and has an ethylidene functional group at C-8 [5], 81-OH-Chl *a*, a chlorin, has a double bond connecting C-7 and C-8 with a hydroxyethyl group at C-8 [44]. BChl *g* and 81-OH-Chl *a* are putatively synthesized from a common precursor, divinyl chlorophyllide *a* [7,60].

Hydration of the C-8 vinyl group of divinyl chlorophyllide *a* is catalyzed by 8-vinyl chlorophyllide hydratase, and bacteriochlorophyll synthase catalyzes the addition of a farnesyl group to produce the mature 81-OH-Chl *a* [7,60]. However, a gene encoding chlorophyllide hydratase or an analogous enzyme was not identified in the genomes of *Hrs. convoluta* or *Hbt. modesticaldum* [6]. Hence, a possible alternative mechanism for 81-OH-Chl *a* synthesis includes steps of dehydrogenation and subsequent hydroxygenation of bacteriochlorophyllide *g* to produce 81-OH-chlorophyllide *a* [60], but genes encoding enzymes for this reaction were not identified in either *Hrs. convoluta* or *Hbt. modesticaldum*. Yet another possible mechanism for 81-OH-Chl *a* synthesis would require the irreversible conversion of BChl *g* into 81-OH-Chl *a* upon exposure to O2 and light [62]. However, as strict anaerobes, the viability of heliobacteria is compromised upon exposure to O2, and therefore this mechanism is unlikely as the major pathway for 81-OH-Chl *a* production [3,62].

**Figure 5.** Predicted biosynthetic pathway of major pigments in *Heliorestis convoluta*. The non-mevalonate pathway shows the synthesis of farnesyl diphosphate for either conversion into carotenoids (orange) or incorporation into the final chlorophyll (green) structures. The enzymes that catalyze each individual numbered reaction are (1) 1-deoxy-d-xylulose-5-P synthase, (2) 1-deoxy-d-xylulose-5-P reductoisomerase, (3) 4-(CDP)-2-C-methyl-d-erythritol synthase, (4) 4-(CDP)-2-C-methyl-d-erythritol kinase, (5) 2-C-methyl-d-erythritol 2,4-cyclo-PP synthase, (6) 4-hydroxy-3-methylbut-2-enyl-PP synthase, (7) 4-hydroxy-3-methylbut-2-enyl-PP reductase, (8) isomerase, (9) geranyl diphosphate synthase, (10) farnesyl diphosphate synthase, (11) 4,4--diapophytoene synthase, (12) diapophytoene dehydrogenase, (13) hydratase, (14) glucosyl transferase, (15) esterase, (16) enzymes encoded by *gltx* and *hemALBCDENYG* genes, (17) enzymes encoded by *bchIDHME* genes, (18) protochlorophyllide reductase (*bchLNB*), (19) chlorophyllide reductase (*bchXYZ*), and (20) bacteriochlorophyll synthase (*bchG*). Red, boxed numbers represent enzymes not ye<sup>t</sup> identified in the *Hrs. convoluta* genome but are proposed based on the predicted pathway. Adapted from Takaichi et al. [63], *Arch. Microbiol.* **2003**, *179*, 95–100. Copyright 2002 Springer Nature; Dubey et al. [64] *J. Biosci.* **2003**, *28*, 637–646. Copyright 2003 Springer Nature; Sattley et al. [6] *J. Bacteriol.* **2008**, *190*, 4687–4696. Copyright 2008 American Society for Microbiology; Sattley and Swingley [7], *Adv. Bot. Res.* **2013**, *66*, 67–97, Copyright 2013 Elsevier Ltd.; and Tsukatani et al. [60], *Biochim. Biophys. Acta* **2013**, *1827*, 1200–1204. Copyright 2013 Elsevier Ltd.

Due to high sequence identity between genes allowing for phototrophic growth (data not shown), a mechanism similar to BChl *g* and 81-OH-Chl *a* biosynthesis in *Hbt. modesticaldum* [6,7,60] is predicted for *Hrs. convoluta* (Figure 5). Many of the genes encoding enzymes required for pigment biosynthesis are grouped into a single photosynthesis gene cluster (PGC) in heliobacteria. The PGCs of *Hrs. convoluta* and *Hbt. modesticaldum* were nearly identical and displayed a shared gene synteny in all key genes, including those associated with pigment and cofactor biosynthesis, electron transport, and light harvesting, suggesting that a common genetic architecture—one that differs substantially from the PGCs present in the genomes of purple bacteria—defines the heliobacterial PGC (Figure 6).

**Figure 6.** Photosynthesis gene clusters from *Heliorestis convoluta* and the purple bacterium *Rhodobacter capsulatus*. Shared genes are outlined in bold. Lines indicate gene synteny: black, single gene rearrangements; red, inverted genes; and blue, inverted genes with a gene insertion. Dashed boxes show *Rba. capsulatus* photosynthesis genes absent from *Hrs. convoluta*. Colors: green, bacteriochlorophyll biosynthesis (bch); orange, carotenoid biosynthesis (crt); pink, proteobacterial reaction centers (puf) and light harvesting complexes (puh); olive, heliobacterial reaction center (psh); teal, regulatory proteins; light green, electron transport (pet); red, cofactor biosynthesis; purple, cell division and sporulation; light blue, nitrogen fixation; grey, transcription; light grey, other nonphotosynthesis genes; and white, uncharacterized genes. Adapted from Sattley et al. [6]. *J. Bacteriol.* **2008**, *190*, 4687–4696. Copyright 2008 American Society for Microbiology.

Like BChls *c*, *d,* and *e* of green sulfur bacteria, both BChl *g* and 81-OH-Chl *a* of heliobacteria are esterified with farnesol [63]. A non-mevalonate pathway is employed to synthesize the esterifying alcohol, farnesyl diphosphate, of heliobacterial pigments [64]. As was noted for *Hbt. modesticaldum*, *Hrs. convoluta* contained the complete complement of genes for this pathway, beginning with pyruvate and glyceraldehyde-3-phosphate and proceeding to an unidentified but predicted isomerase that could catalyze the interconversion of isopentenyl diphosphate and dimethylallyl diphosphate [6,64] (Figure 5). Following this, farnesyl diphosphate can either be incorporated into the final structures of BChl *g* and 81-OH-Chl *a* or further transformed into the major carotenoids found in *Hrs. convoluta* [63] (Figure 5). The high specificity of BchG for incorporation of a farnesol moiety over longer alcohol groups, such as phytol, has been demonstrated in studies of pigment biosynthesis in *Hbt. modesticaldum* [61], and the high sequence identity (68%) of BchG from *Hrs. convoluta* to that of *Hbt. modesticaldum* suggests a similar activity in the alkaliphile.

Experimental work and pigment extraction from *Hrs. convoluta*, *Hrs. daurensis*, and *Hrs. baculata* revealed that the major carotenoid in alkaliphilic heliobacteria is OH-diaponeurosporene glucoside C16:0 ester, followed by 4,4--diaponeurosporene, OH-diaponeurosporene glucoside C16:1 ester, and 8,8--zeta-carotene [21,63]. These novel glucoside esters in alkaliphilic heliobacteria were not found in neutrophilic heliobacteria, in which 4,4--diaponeurosporene was the major carotenoid [63,65]. The synthesis of these C30 carotenoids [66] is complicated by the apparent absence of a gene (*crtM*) encoding 4,4--diapophytoene synthase in both *Hbt. modesticaldum* [6] and *Hrs. convoluta*. Presumably, the presence of an enzyme with 4,4--diapophytoene synthase activity is essential in the proposed biosynthetic pathway for each of the carotenoids found in heliobacteria [63] (Figure 5). Although *crtM* was not identified, two nonidentical copies of *crtN* (FTV88\_2648 and FTV88\_3059) having a sequence identity of 71% were annotated in the *Hrs. convoluta* genome, and it is possible that one of their gene products exhibits CrtM-like activity.

In alkaliphilic heliobacteria, a proposed CrtC-like hydratase catalyzes the formation of OH-diaponeurosporene from 4,4--diaponeurosporene, followed by synthesis of OH-diaponeurosporene glucoside by a CrtX-like glucosyl transferase, with a putative esterase making the final conversion to the mature glucoside ester [63]. Genes encoding the enzymes catalyzing the final three steps of OH-diaponeurosporene glucoside ester synthesis were not identified in *Hrs. convoluta* (Figure 5), but genes encoding two carotenoid biosynthesis proteins (FTV88\_0301 and FTV88\_0302) were annotated. These genes showed no significant sequence similarity to genes of the neutrophilic *Hbt. modesticaldum*, and they may be candidates for encoding proteins to perform the final steps of carotenoid biosynthesis in alkaliphilic heliobacteria.

#### *3.6. Reaction Center and Electron Transport Chain*

Heliobacteria possess a type I (Fe–S type) photosynthetic reaction center (RC) imbedded in the cytoplasmic membrane [4,67,68]. As the simplest known and perhaps most ancient extant (bacterio)chlorophyll-binding photochemical apparatus [69], the heliobacterial RC is a symmetrical homodimer consisting of the PshA polypeptide and the novel, single-transmembrane helix PshX polypeptide [70]. PshA of *Hrs. convoluta* (encoded by *pshA*, FTV88\_2638) showed 71% sequence identity to PshA of *Hbt. modesticaldum* but nearly 96% identity to PshA of *Hrs. acidaminivorans* (GenBank accession WBXO01000000, unpublished). As is the case in *Hbt. modesticaldum*, *pshX* (FTV88\_2551) is situated outside of the PGC in *Hrs. convoluta* and encodes a protein consisting of just 31 amino acids. The PshX RC subunit from *Hrs. convoluta* showed a 74% sequence identity to that of *Hbt. modesticaldum*.

The crystal structure of the *Hbt. modesticaldum* RC revealed the presence of 54 BChl *g* molecules, two 81-OH-Chl *a* molecules, two carotenoids (4,4--diaponeurosporene), four BChl *g*- molecules (a C-13 epimer of BChl *g* that functions as the primary electron donor, P800) [68,71,72], two lipids, and one [4Fe–4S] cluster [70]. Experimental data on the structure of the *Hrs. convoluta* RC are not ye<sup>t</sup> available. However, with their highly similar PshA and PshX proteins, the geometry and pigment composition of the *Hrs. convoluta* RC should closely resemble that of the *Hbt. modesticaldum* RC [69]. Nevertheless, some distinctions may materialize considering the alternative carotenoids produced by alkaliphilic heliobacteria and their inherently alkaline habitat [63].

Proteins of the electron transport chain (ETC) of *Hrs. convoluta* exhibited high sequence similarity to those from *Hbt. modesticaldum*, and thus the overarching mechanism of light-driven energy conservation is likely to be highly conserved across all heliobacterial taxa. Although not experimentally confirmed, it is likely that electrons first enter the chain by the activity of either NADH:quinone oxidoreductase (Figure 3), a 14-subunit protein complex embedded in the cytoplasmic membrane and encoded by *nuoABCDEFGHIJKLMN*, or perhaps a complex having ferredoxin:menaquinone oxidoreductase activity. As observed in *Hbt. modesticaldum*, the *nuoEFG* genes in *Hrs. convoluta* are not co-localized within the same operon as the other *nuo* genes. However, unlike in *Hbt. modesticaldum*, in which *nuoEF* are fused, *nuoE* and *nuoF* exist as individual genes (present in duplicate copies) in *Hrs. convoluta* (Figure 7). With the exception of this distinction, all *nuo* genes show high sequence identity (62–79%) between the two species. As for *Hbt. modesticaldum*, menaquinone is predicted to shuttle electrons from Complex I to the cytochrome *bc* complex (PetABCD), and electron transfer through these complexes drives translocation of H<sup>+</sup> to the periplasmic space (Figure 3), forming a proton motive force (PMF) [6,73].

Cytochrome *bc*1 complexes, which are found in a variety of anoxygenic phototrophs and also in eukaryotic mitochondria, consist of a minimum of three protein subunits: cytochrome *b*, cytochrome *c*1, and the Rieske iron-sulfur protein [74]. In contrast, the related cytochrome *b*6*f* complex, which is present in cyanobacteria and chloroplasts, is comprised of cytochrome *f* (PetA), cytochrome *b*6 (PetB), the Rieske iron-sulfur protein (PetC), and subunit IV (PetD) [74]. Having similar functions but distinct structural properties, cytochrome *b* contains eight transmembrane helices, whereas cytochrome *b*6 and its associated subunit IV contain four and three transmembrane helices, respectively [74]. Cytochrome *b*6 shows homology to the *N*-terminal half of cytochrome *b*, and subunit IV is homologous to the *C*-terminal half of cytochrome *b* [74].

**Figure 7.** Comparison of *nuoEFG* genes in *Heliorestis convoluta* and *Heliobacterium modesticaldum*. Whereas in *Hbt. modesticaldum* (and *Heliobacillus mobilis*) *nuoE* and *nuoF* are fused, these genes are independent in *Hrs. convoluta*. In both species, however, *nuoEFG* are separated from other *nuo* genes on the chromosome. In addition, unlike in *Hbt. modesticaldum*, *Hrs. convoluta* contains two copies of *nuoEFG*, as well as the *hydEFG* maturase genes (FTV88\_1003–1005) that may impart [FeFe] hydrogenase activity. Colors: gold, NADH dehydrogenase subunits; orange, structural genes.

An analysis of the cytochrome *bc* complex of *Hrs. convoluta* indicated that it resembles a hybrid of the cytochrome *b*6*f* complex and the cytochrome *bc*1 complex. A comparison of cytochrome *b*6 and subunit IV proteins from *Hrs. convoluta* and the model cyanobacterium *Synechocystis* PCC 6803 showed 48% and 42% amino acid sequence identity, respectively. However, cytochrome *b*6 and subunit IV from *Hrs. convoluta* also showed 36% and 30% amino acid identity, respectively, to the *N*-terminal and *C*-terminal halves of cytochrome *b* from the purple bacterium *Rhodobacter sphaeroides*. Furthermore, whereas subunit IV from *Hrs. convoluta* is predicted to contain the usual three transmembrane helices, cytochrome *b*6 from *Hrs. convoluta* contained a predicted five transmembrane regions instead of the four typically observed in the *b*6*f* complex. Notably, cytochrome *b*6 from *Hbt. modesticaldum* is predicted to contain the conventional four transmembrane helices. Therefore, considering the above sequence analyses and their total of eight predicted transmembrane helices, the PetB and PetD proteins of *Hrs. convoluta* may represent a structural and evolutionary intermediate between cytochrome *b* and cytochrome *b*6/subunit IV proteins, a distinction perhaps not shared with neutrophilic heliobacteria.

The PetA protein in heliobacteria is also of interest because it functions as a diheme cytochrome *c* (as opposed to the typical monoheme protein) and shows no sequence or structural similarity to cytochrome *f* [74]. Although unusual among the *Firmicutes*, the diheme cytochrome *c* has been identified in all heliobacteria studied thus far and is likely a universal feature of these phototrophs. PetA from *Hrs. convoluta* showed high sequence identity with PetA from *Hrs. acidaminivorans* (79%), and sequence identities to PetA from neutrophilic heliobacteria (e.g., *Hbt. modesticaldum*, *Hbt. gestii*, *Heliobacillus mobilis*, and *Heliophilum fasciatum*) were all near 50%. Based on similarities in its N- and C-terminal domains, the heliobacterial diheme cytochrome *c* may have been the result of a past gene duplication and subsequent fusion [75].

A single operon containing all eight genes encoding the subunits of ATP synthase (*atpABCDEFGH*) was identified in the genome of *Hbt. modesticaldum* [6,26], and the encoded ATP synthase itself has since been biochemically characterized [76]. The composition and arrangemen<sup>t</sup> of ATP synthase genes in *Hrs. convoluta* was identical to that in the *Hbt. modesticaldum* genome. Kinetic studies with *Hba. mobilis* and *Hbt. modesticaldum* and physiological similarity to photosystem I of cyanobacteria sugges<sup>t</sup> that a PMF established by cyclic electron flow drives photophosphorylation in heliobacteria [73,77,78]. For overviews of electron transfer reactions in heliobacteria, see Sattley and Swingley [7], Kondo et al. [79], and, more recently, Kashey et al. [73].
