Keywords

Introduction

Sulfur exhibits high reactivity in reduced forms and occurs in several stable oxidation states. Sulfate or sulfide in water and soil and sulfur dioxide in the atmosphere constitute the majority of sulfur in nature (Middelburg 2000). Smaller but significant roles are played by polysulfide, polythionates, thiosulfate, sulfoxides, as well as elemental sulfur. Sulfur is the element with the highest number of allotropes but only a few occur in biological systems. Sulfur appears in all organisms in many different organic compounds such as amino acids, enzyme cofactors, (poly)peptides, sulfolipids, vitamins, or carbohydrates. The biological roles of inorganic sulfur compounds are comparatively restricted: (1) They can serve as sources for sulfur assimilation and incorporation into the abovementioned organic compounds. (2) They can be employed as donors or acceptors of electrons for energy-generating electron transport. Dissimilatory sulfur-based energy generation goes along with mass transformations and occurs almost exclusively among prokaryotes, while assimilatory sulfur metabolism is not only very common in prokaryotes but also occurs in plants, algae, and fungi.

In oxygenic phototrophic organisms, the redox properties of sulfur-containing metabolites and of sulfur in proteins are very important for the interplay between the reductive assimilative processes of photosynthesis and reactive oxygen species that are formed as side products of photosynthetic electron transport (Dahl et al. 2008b). In anoxygenic phototrophic bacteria, reduced sulfur compounds can play a particularly important role as electron donors for photosynthetic carbon dioxide fixation. In fact, the utilization of sulfur compounds is common to almost all groups of phototrophic prokaryotes: certain species of the cyanobacteria can perform anoxygenic photosynthesis at the expense of sulfide as an electron donor (Arieli et al. 1991, 1994; Shahak and Hauska 2008). A few representatives of the strictly anaerobic Gram-positive heliobacteria as well as members of the filamentous anoxygenic phototrophic (FAP) bacteria of the phylum Chloroflexi are able to oxidize reduced sulfur compounds, thiosulfate oxidation is widespread among the photoheterotrophic aerobic anoxygenic phototrophic bacteria, and many of the classical purple non-sulfur bacteria can use thiosulfate and/or sulfide as electron donors. Utilization of reduced sulfur compounds is best known and studied for the purple (families Chromatiaceae and Ectothiorhodospiraceae ) and green sulfur bacteria (phylum Chlorobi).

In this chapter the sulfur-oxidizing capabilities of the various groups of phototrophic bacteria will be only briefly described. The reader is referred to a number of previous reviews that still provide valuable sources of information on sulfur compounds used by the various groups as well as on sulfur oxidation patterns (Brune 1989; Brune 1995b; Dahl 2008; Frigaard and Bryant 2008a, b; Frigaard and Dahl 2009; Gregersen et al. 2011; Sander and Dahl 2009). Here, I will focus on new developments arising from studies performed during the past 8–10 years that substantially broadened our knowledge of the biochemical details of the different sulfur oxidation pathways. In addition, a substantial number of additional genome sequences for purple sulfur bacteria became available that allows to draw additional information from comparative analyses of gene arrangements and occurrence. Transcriptomic profiling and comparative proteome analyses for phototrophic model organisms provide further crucial information resources (Eddie and Hanson 2013; Falkenby et al. 2011; Weissgerber et al. 2013, 2014a). A brief overview of assimilatory sulfate reduction metabolism will also be given. Organosulfur compound metabolism will not be dealt with here, and the reader is referred to information provided by others (Baldock et al. 2007; Denger et al. 2004, 2006; Kappler and Schäfer 2014; Visscher and Taylor 1993).

Sulfur Oxidation Capabilities of Phototrophic Bacteria

In the following section, the sulfur oxidation capabilities of the various groups of anoxygenic phototrophic bacteria are very briefly summarized.

Purple Sulfur Bacteria

Purple sulfur bacteria of the families Chromatiaceae and Ectothiorhodospiraceae preferentially use reduced sulfur compounds as electron donors during photolithoautotrophic growth. The most important difference between the two families is that Chromatiaceae produce intracellular sulfur globules when growing on sulfide, thiosulfate, polysulfides, or elemental sulfur, while the Ectothiorhodospiraceae accumulate extracellular sulfur. For one member of the Ectothiorhodospiraceae, Thiorhodospira sibirica , extra- as well as intracellular sulfur deposition has been reported (Bryantseva et al. 1999). All phototrophic members of the Chromatiaceae use sulfide and sulfur of the oxidation state zero as photosynthetic electron donors. Several species are limited to these compounds while a range of more versatile species uses several reduced sulfur compounds including thiosulfate and sulfite. Polysulfide oxidation is probably ubiquitous. This does not appear astonishing because polysulfides are formed as intermediates of the oxidation of sulfide en route to sulfur globules (Prange et al. 2004). Polysulfides are especially stable intermediates of sulfide oxidation by members of the Ectothiorhodospiraceae because these thrive under alkaline growth conditions which are essential for longer-term stability of polysulfides. Utilization of sulfide, elemental sulfur, and thiosulfate is common to the species of the genus Ectothiorhodospira, while species of the genera Halorhodospira and Thiorhodospira oxidize sulfide to sulfur which is further oxidized to sulfate by some species. Thiosulfate is used only by Halorhodospira halophila (Raymond and Sistrom 1969).

Green Sulfur Bacteria

GSB exhibit very little variation in their ability to oxidize sulfur compounds. Almost all members of this group oxidize sulfide and elemental sulfur to sulfate. The only exception is Chlorobium ferrooxidans for which only Fe2+ and hydrogen are suitable photosynthetic electron donors. In general, GSB have a very high affinity for sulfide, and it is the preferred sulfur substrate even in the presence of other reduced sulfur compounds. Typically, sulfide is first transformed into zero-valent sulfur which is deposited as extracellular sulfur globules. Some strains of the genera Chlorobaculum and Chlorobium can oxidize thiosulfate (Imhoff 2003), and one strain has been reported to be capable of tetrathionate utilization (Khanna and Nicholas 1982).

Purple Non-sulfur Bacteria

Purple non-sulfur bacteria are to a much lesser extent capable of tolerating and using toxic sulfur compounds such as sulfide than the PSB. The phototrophic Betaproteobacteria of the orders Rhodocyclales and Burkholderiales have not been reported to use reduced sulfur compounds as electron donors. Sulfide inhibits growth at low concentrations (Imhoff et al. 2005). In the genome of Rubrivivax gelatinosus , sox genes are present indicating the potential for thiosulfate oxidation (Sander and Dahl 2009). Sulfate can be reductively assimilated. Within the alphaproteobacterial purple non-sulfur bacteria, the ability to use reduced sulfur compounds is widespread. Intermediates and final products formed vary considerably between species. Complete oxidation of sulfide to sulfate has been described for several species (Frigaard and Dahl 2009; Imhoff et al. 2005; Sander and Dahl 2009). Thiosulfate is used by many species and either completely oxidized to sulfate or transformed into tetrathionate. Sulfur is also used as a substrate by some species (Sander and Dahl 2009).

Aerobic Bacteriochlorophyll-Containing Bacteria

Aside from cyanobacteria and proteorhodopsin-containing bacteria, aerobic anoxygenic phototrophic (AAP) bacteria are the third most numerous group of phototrophic prokaryotes in the ocean. This functional group represents a diverse assembly of species which taxonomically belong to various subgroups of Alpha-, Beta-, and Gammaproteobacteria . AAP bacteria are facultative photoheterotrophs which use bacteriochlorophyll-containing reaction centers to harvest light energy under fully oxic in situ conditions (Koblizek 2015). Almost 60 strains of AAP are currently fully genome sequenced (tabulated in Koblizek 2015).

In general AAP bacteria cannot grow photolithoautotrophically on reduced sulfur compounds. However, many representatives of this physiological group can oxidize sulfur compounds as additional sources of electrons and grow as sulfur-oxidizing lithoheterotrophs. The ability for thiosulfate oxidation appears to be especially widespread (Sorokin et al. 2000; Yurkov et al. 1994). The genomes of many AAP bacteria contain the genes soxB, soxAX, soxYZ, and soxCD encoding a periplasmic thiosulfate-oxidizing multienzyme complex (Friedrich et al. 2005; Sander and Dahl 2009). A recent study furthermore revealed that sox genes are present mainly in those members of the widespread and ecologically very important OM60/NOR5 clade that also encode genes enabling aerobic anoxygenic photoheterotrophy, like Congregibacter litoralis (C. litoralis) DSM 17192T, Congregibacter sp. strain NOR5-3, or Luminiphilus syltensis DSM 22749T (Spring 2014). However, a stringent correlation of genes encoding Sox proteins and subunits of the photosynthetic apparatus was not apparent, because some bacteriochlorophyll a-containing strains do not encode Sox proteins.

Acidobacteria

The phylum Acidobacteria , a sister clade to the δ-Proteobacteria in the domain Bacteria, encompasses a large and physiologically diverse group of microorganisms (Ciccarelli et al. 2006). Recently, a phototrophic member of this group was described, Chloracidobacterium thermophilum (Bryant et al. 2007; Tank and Bryant 2015a, 2015b), the first aerobic chlorophototroph that has a type I, homodimeric reaction center (RC). Key genes for all known carbon fixation pathways are absent as are genes for assimilatory sulfate reduction. Cab. thermophilum is unable to use sulfate as a sulfur source and instead relies on reduced sulfur sources such as thioglycolate, cysteine, methionine, or thiosulfate. Cultures containing sodium sulfide did not show sustained growth, but microscopic analyses revealed that sulfur globules were produced. Similar to green sulfur bacteria, these globules remained associated with the outer surfaces of cells and suggested that sulfide oxidation occurred. The genome lacks any known enzymes for the oxidation of sulfide, so how sulfide oxidation occurs is not clear (Tank and Bryant 2015b).

Phototrophic Gemmatimonadetes

Very recently a BChl a-producing, semiaerobic anoxygenic photoheterotroph from the phylum Gemmatimonadetes, Gemmatimonas phototrophica, has been described (Zeng et al. 2014, 2015). Sulfur oxidation capabilities have not been reported. None of the genome-sequenced members of the Gemmatimonadetes contain sox genes.

Sulfur Oxidation Pathways

With regard to their sulfur metabolism, phototrophic bacteria are characterized by a great variability of sulfur substrates used and pathways employed. On a molecular genetic and biochemical level, sulfur oxidation is best described in the purple sulfur bacterium Allochromatium vinosum and in the green sulfur bacterium Chlorobaculum tepidum . An overview of the currently proposed model of sulfur oxidation in A. vinosum is shown in Fig. 1. The figure is based on a combination of biochemical evidence, genome sequence information, as well as whole genome transcriptomic profiling and comparative quantitative proteomics (Weissgerber et al. 2011, 2013, 2014a).

Fig. 1
figure 1

Current model of sulfur oxidation in Allochromatium vinosum (Figure taken from Weissgerber et al. 2014a) (Copyright © American Society for Microbiology, Applied and Environmental Microbiology 80, 2014, 2279–92, doi: 10.1128/AEM.04182–13). The proteomic profiles (circles) and transcriptomic profiles (boxes) are depicted next to the respective proteins. Relative fold changes in mRNA levels above 2 (green) were considered significant enhancement. Relative changes smaller than 0.5 (red) were considered to indicate significant decreases in mRNA levels. Relative fold changes between 0.5 and 2 (yellow) indicated unchanged mRNA levels. Fig. 1 (continued) The same color coding is applied to changes on the protein level. Here, values above 1.5 (green) and below 0.67 (red) were considered significant. Those cases where transcriptomic data were not available or the respective protein was not detected in the proteomic approach are indicated by white squares or circles. Changes are depicted that occurred upon a switch from photoorganoheterotrophic growth on malate to photolithoautotrophic growth on, from left to right, sulfide, thiosulfate, elemental sulfur, and sulfite. Changes on sulfite were not determined on the proteome level

Full size image

Many enzymes involved in sulfur metabolism can readily be identified in genome sequences by sequence homology with known enzymes. The genome sequences of 15 strains of GSB have already been available for several years, and the occurrence of genes related to sulfur oxidation in these organisms has already been extensively tabulated and discussed (Frigaard and Bryant 2008b; Frigaard and Dahl 2009; Gregersen et al. 2011; Venceslau et al. 2014). A greater number of genome sequences for purple sulfur bacteria only became available over the last few years (Table 1). This chapter will therefore focus on analyzing this comparatively new set of sequence information.

Table 1 Genome-sequenced purple sulfur bacteria
Full size table

Oxidation of Thiosulfate

Thiosulfate (S2O3 2−) oxidation is conducted by a large number of photo- and chemotrophic sulfur-oxidizing bacteria. In general, two completely different pathways can be differentiated. In the first, two thiosulfate anions are oxidized to tetrathionate. In the second, catalyzed by the periplasmic Sox multienzyme system (Dahl et al. 2008a; Friedrich et al. 2001), multiple steps lead to complete oxidation to sulfate. In some bacteria including A. vinosum both pathways coexist (Hensen et al. 2006; Smith and Lascelles 1966). The occurrence of genes related with the two pathways in purple sulfur bacteria is summarized in Table 2.

Table 2 Genes related to thiosulfate oxidation in genome-sequenced purple sulfur bacteria a
Full size table

Oxidation of Thiosulfate to Tetrathionate

The ability to perform the very simple oxidation of two molecules of thiosulfate to tetrathionate according to the equation 2 S2O3 2− → S4O6 2−+2e− is widespread among prokaryotes. The reaction is not only well-established intermediate step in the oxidation of reduced sulfur compounds to sulfate in many obligately chemolithoautotrophic bacteria (Lu and Kelly 1988; Müller et al. 2004; Wentzien et al. 1994) but also known for some purple non-sulfur bacteria like Rhodomicrobium vannieli and Rhodopila globiformis and purple sulfur bacteria including A. vinosum (Frigaard and Dahl 2009; Hensen et al. 2006; Then and Trüper 1981).

Despite the well-documented significance of tetrathionate formation in aquatic and terrestrial habitats (Barbosa-Jefferson et al. 1998; Podgorsek and Imhoff 1999; Sorokin 2003), the membrane-bound doxDA encoding thiosulfate/quinone oxidoreductase from the thermoacidophilic archaeon Acidianus ambivalens was the only tetrathionate-forming enzyme characterized on a molecular level for a long time (Müller et al. 2004). Genes homologous to doxDA do not occur in phototrophic prokaryotes. Instead, a gene (tsdA) encoding a novel periplasmic 27.2 kDa diheme cytochrome c thiosulfate dehydrogenase was identified in A. vinosum (Denkmann et al. 2012). The crystal structure of the enzyme revealed two typical class I c-type cytochrome domains wrapped around two hemes. Heme 1 exhibits His/Cys iron coordination and constitutes the active site of the enzyme (Brito et al. 2015). His/Cys heme iron ligation is rare among prokaryotes, usually leads to a low redox potential of the corresponding heme (Bradley et al. 2012; Kappler et al. 2008; Pires et al. 2006; Reijerse et al. 2007), and appears to be of special importance in sulfur-based energy metabolism. In the oxidized state, Heme 2 iron is axially ligated by a histidine and a lysine residue (Fig. 2). Upon reduction, a switch occurs at this heme from Lys to Met axial ligation. This change probably affects the redox potential of Heme 2 and may be an important step during the reaction cycle (Brito et al. 2015).

Fig. 2
figure 2

Heme coordination of A. vinosum TsdA (Brito et al. 2015). Left: Heme 1 is coordinated by His53 and Cys96. Right: Heme 2 is coordinated by His His165 and Lys208. Upon reduction, a ligand switch from Lys208 to Met209 occurs. Sulfur atoms are shown in green

Full size image

TsdA enzymes of various source organisms exhibit different catalytic bias (Kurth et al. 2015). While the enzyme from the sulfur oxidizer A. vinosum is strongly biased toward catalyzing thiosulfate oxidation (Brito et al. 2015), TsdA from Campylobacter jejuni acts primarily as a tetrathionate reductase and enables the organism to use tetrathionate as alternative electron acceptor for anaerobic respiration (Liu et al. 2013).

Currently it is largely unclear which redox carriers mediate the flow of electrons arising from thiosulfate oxidation into respiratory or photosynthetic electrons chains. In several organisms including Thiomonas intermedia, Sideroxydans lithotrophicus, and Pseudomonas stutzeri, tsdA is immediately preceded by a gene encoding another diheme cytochrome, TsdB (Denkmann et al. 2012). TsdB is not itself reactive with thiosulfate but accepts electrons from TsdA even when TsdA and TsdB do not originate from the same source organism (Denkmann et al. 2012). In the anoxygenic phototrophic purple sulfur bacterium Marichromatium purpuratum , TsdA and TsdB form a fusion protein with TsdB constituting the amino-terminal domain. TsdBA fusion proteins are also encoded in other members of the Chromatiaceae , e.g., Thiorhodococcus sp. AK35, Thiocystis violascens, Thiorhodococcus drewsii, and Thioflaviococcus mobilis (Table 2). However, TsdBA fusions are not a common trait in purple sulfur bacteria. In A. vinosum , a tsdB gene is not present (Denkmann et al. 2012). In A. vinosum, the protein with the closest relationship to T. intermedia or P. stutzeri TsdB is Alvin_2879. This cytochrome c 4 (previously cytochrome c- 553(550)) is membrane bound possibly via the hydrophobic protein Alvin_2880 and has a positive redox potential of +330 mV (Cusanovich and Bartsch 1969).

Another candidate for accepting electrons from TsdA in purple anoxygenic phototrophic bacteria is the high potential iron-sulfur protein (HiPIP). A. vinosum and M. purpuratum produce HiPIP, and as this protein has a quite positive reduction potential (+350 mV) (Bartsch 1978), it would be well suitable as an electron acceptor for TsdA .

Oxidation of Thiosulfate to Sulfate

The Sox pathway of thiosulfate oxidation is a prime example for the oxidation of protein-bound sulfur atoms in the bacterial periplasm (Friedrich et al. 2001; Zander et al. 2010). Among the many organisms pursuing this pathway, some store sulfur globules as intermediates (e.g., A. vinosum ), whereas others do not form sulfur deposits (e.g., Paracoccus pantotrophus ). The Sox pathway in these two physiological groups appears to have one fundamental difference, and this is the involvement of the hemomolybdoprotein SoxCD (Fig. 3).

Fig. 3
figure 3

Model of Sox-mediated thiosulfate oxidation in Paracoccus pantotrophus (left) and (a) A. vinosum (right). Adapted from (Sander and Dahl 2009). All reactions take place in the periplasm

Full size image

In non-sulfur-storing organisms, the proposed mechanism for sulfur oxidation requires four different proteins: SoxB, SoxXA, SoxYZ, and SoxCD (Friedrich et al. 2001). The heterodimeric SoxYZ protein acts as the central player and carries pathway intermediates covalently bound to a cysteine residue located near the carboxy-terminus of the SoxY subunit (Appia-Ayme et al. 2001; Quentmeier and Friedrich 2001; Sauvé et al. 2007). The c-type cytochrome SoxXA(K) catalyzes the oxidative formation of a disulfide linkage between the sulfane sulfur of thiosulfate and the cysteine of SoxY (Bamford et al. 2002; Ogawa et al. 2008). The sulfone group is then hydrolytically released as sulfate in a reaction catalyzed by SoxB (Sauvé et al. 2009). The next step is oxidation of the SoxY-bound sulfane sulfur to a sulfone by the hemomolybdoprotein SoxCD and again hydrolytic release of sulfate (Zander et al. 2010).

In those organisms forming sulfur as an intermediate, SoxCD is not present and the SoxY-bound sulfane sulfur is transferred to zero-valent sulfur stored in sulfur globules residing in the periplasm by an unknown mechanism, possibly involving the rhodanese-like protein SoxL (Welte et al. 2009). In A. vinosum , sox genes are present in two clusters (soxBXAKL, Alvin_2167 to 2171, and soxYZ, Alvin_2111 and 2112) with soxBXA and soxYZ being indispensable for thiosulfate oxidation (Hensen et al. 2006). The protein encoded by soxK has been identified as a subunit of a SoxXAK complex in the green sulfur bacterium Chlorobaculum tepidum (Ogawa et al. 2008) and probably fulfills the same function in purple sulfur bacteria.

Oxidation of Sulfide

Different enzymes are candidates for sulfide oxidation: sulfide/quinone oxidoreductases (SQR ) (Schütz et al. 1997) and a flavocytochrome c sulfide dehydrogenase (FccAB) (Chen et al. 1994; Meyer and Cusanovich 2003) (Table 3). In Rhodovulum sulfidophilum , a member of the Rhodobacteraceae, the Sox system is not only essential for thiosulfate oxidation but also indispensable for the oxidation of sulfide in vivo (Appia-Ayme et al. 2001). The same might well be the case for other non-sulfur bacteria containing sox genes. In A. vinosum mutants deficient in either flavocytochrome c (Reinartz et al. 1998), sox genes (Hensen et al. 2006), or both (D. Hensen, B. Franz, C. Dahl, unpublished), sulfide oxidation proceeds with wild-type rates indicating that SQR plays the major role.

Table 3 Genes related to sulfide oxidation in genome-sequenced purple sulfur bacteria a
Full size table

All characterized SQRs are single-subunit flavoproteins associated with the cytoplasmic membrane (Marcia et al. 2009, 2010b; Shahak and Hauska 2008). Based on the protein structure, six distinct SQR types were identified (Marcia et al. 2010a). Here, the nomenclature suggested by Frigaard and coworkers is followed (Gregersen et al. 2011) to clearly identify the multiple types of sqr genes often found in the same organism (Table 3). Members of types SqrA, SqrB, SqrC, SqrE, and SqrF have been biochemically characterized (Arieli et al. 1994; Brito et al. 2009; Cherney et al. 2010; Griesbeck et al. 2002; Marcia et al. 2009; Shuman and Hanson 2016; Zhang and Weiner 2014). The SqrA type exemplified by the functionally well-characterized enzyme from the cyanobacterium Oscillatoria limnetica (Bronstein et al. 2000) and the purple non-sulfur bacterium Rhodobacter capsulatus (Schütz et al. 1999) does neither occur in green (Gregersen et al. 2011) nor in purple sulfur bacteria (Table 3). The same holds true for SqrE. The SqrD and SqrF types appear to be especially widespread in the family Chromatiaceae , while members of the Ectothiorhodospiraceae all contain a gene encoding SqrB. The SqrF-type enzyme from C. tepidum has recently been shown to have a low affinity for sulfide and a high enzymatic turnover rate consistent with a function as a high sulfide adapted SQR (Chan et al. 2009; Eddie and Hanson 2013). The primary reaction product of the SQR reaction is soluble polysulfide (Griesbeck et al. 2002).

In a variety of sulfide-oxidizing species, flavocytochrome c is present as a soluble protein in the periplasm or as a membrane-bound enzyme (Kostanjevecki et al. 2000). The protein consists of a larger flavoprotein (FccB) and a smaller hemoprotein (FccA) subunit. The proteins show sulfide/cytochrome c activity in vitro (Bosshard et al. 1986). FccAB occurs in many purple and green sulfur bacteria but there are also species that lack it (Frigaard and Dahl 2009; Sander and Dahl 2009). It is possible that FccAB is advantageous under certain growth conditions, and it has been speculated that it might represent a high-affinity system for sulfide oxidation especially suited at very low sulfide concentrations (Brune 1995b).

Oxidation of Polysulfides

Polysulfides occur as the primary reaction product of the oxidation of sulfide in purple (Franz et al. 2009; Prange et al. 2004) and green (Marnocha et al. 2016) sulfur bacteria. It is still unclear how polysulfides are converted into sulfur globules (Fig. 1). Theoretically this could be purely chemical spontaneous process as longer polysulfides are in equilibrium with elemental sulfur (Steudel et al. 1990).

Oxidation of External Sulfur

Many green and purple sulfur bacteria are able to oxidize externally supplied elemental sulfur. Sulfur of oxidation state zero mainly consists of S8 rings and chain-like polymeric sulfur. Traces of S7 rings are also present. Elemental sulfur is virtually insoluble in water, and it is still unclear how exactly phototrophs are able to bind, activate, and take up this substrate. A. vinosum uses only the polymeric sulfur fraction of commercially available sulfur (Franz et al. 2007). Soluble intermediates like sulfide, polysulfides, or polythionates do not appear to be formed. It therefore seems unlikely that mobilization of elemental sulfur by purple sulfur bacteria involves excretion of soluble sulfur-containing substances that would be able to act on substrate distant from the cells (Franz et al. 2009). Instead, direct cell-sulfur contact appears to be necessary for uptake of elemental sulfur by A. vinosum (Franz et al. 2007).

Properties of Sulfur Globules

In anoxygenic phototrophic sulfur bacteria, sulfur formed as an intermediate is never deposited in the cytoplasm. Green sulfur bacteria and members of the Ectothiorhodospiraceae form extracellular sulfur globules, and the globules of the members of the family Chromatiaceae are located in the periplasmic space (Pattaragulwanit et al. 1998). Independent of the site of deposition, the sulfur appears to be of similar speciation, i.e., long sulfur chains that might be terminated by organic residues in purple sulfur bacteria (Prange et al. 2002). While proteinaceous envelopes have never been reported for extracellular sulfur globules, the sulfur globules in the Chromatiaceae are enclosed by a protein envelope (Brune 1995a). In A. vinosum this envelope is a monolayer of 2–5 nm consisting of four different hydrophobic sulfur globule proteins, SgpABCD (Brune 1995a; Pattaragulwanit et al. 1998; Weissgerber et al. 2014a). All of these proteins are synthesized with cleavable N-terminal peptides mediating Sec-dependent transport to the periplasm and share a highly repetitive amino acid sequence rich in regularly spaced proline residues. They are predicted to act purely as structural proteins. A covalent attachment of sulfur chains to the proteins is unlikely as none of the Sgps contains any cysteine residues. The envelope is indispensable for formation and deposition of intracellular sulfur in A. vinosum. The 10.5 kDa SgpA and SgpB proteins resemble each other and are in part able to replace each other. SgpC is important for expansion of the globules (Prange et al. 2004). SgpD was only recently detected by investigating the sulfur globule proteome and proved to be the most abundant of the four sulfur globule proteins (Weissgerber et al. 2014a). The relative mRNA levels for the corresponding gene increased drastically with addition of sulfide or thiosulfate to the growth medium (Weissgerber et al. 2013). Genes encoding sulfur globule proteins occur in all genome-sequenced purple sulfur bacteria of the family Chromatiaceae but are absent in Ectothiorhodospiraceae. The combination of sulfur globule proteins appears to be variable (Table 4).

Table 4 Genes for sulfur globule proteins in genome-sequenced purple sulfur bacteria a
Full size table

Oxidation of Stored Sulfur to Sulfite

The oxidative degradation of sulfur deposits in phototrophic sulfur bacteria is still a major subject of research. Besides the comparatively well-characterized Dsr (dissimilatory sulfite reductase) system, a completely new pathway of sulfur oxidation involving a heterodisulfide reductase-like enzyme system is currently emerging (Dahl 2015; Venceslau et al. 2014) and appears to be implemented in several phototrophic members of the family Ectothiorhodospiraceae (Table 5).

Table 5 Genes related to stored sulfur oxidation in genome-sequenced purple sulfur bacteria a
Full size table

The Dsr System of Sulfur Oxidation

Currently, the best studied of the sulfur oxidation pathways operating in the cytoplasm is the so-called Dsr pathway (Fig. 4) involving the enzyme reverse dissimilatory sulfite reductase (DsrAB) (Dahl et al. 2005; Pott and Dahl 1998). Low-molecular-weight organic persulfides such as glutathione amide persulfide have been proposed as carrier molecules transferring sulfur from the periplasmic or extracellular sulfur deposits into the cytoplasm (Frigaard and Dahl 2009). It is not yet known how exactly the proposed persulfidic carrier molecules are generated and whether specific enzymes are involved in this process nor have transporters for such molecules be characterized from any sulfur-oxidizing prokaryote. An extensive Cys-SSH-based sulfur relay system exists in A. vinosum (Figs. 1 and 4) that traffics sulfur atoms stemming ultimately from sulfur stored in sulfur globules, through a cascade of protein persulfide intermediates hosted on a rhodanese, TusA, possibly DsrE2A, DsrE, and DsrC (Stockdreher et al. 2012) to the active site of the enzyme sulfite reductase (Cort et al. 2008; Dahl et al. 2008c; Dahl 2015), the enzyme that catalyzes formation of sulfite.

Fig. 4
figure 4

Model of Dsr-mediated sulfane sulfur oxidation in A. vinosum integrating a sulfur-mobilizing function for Rhd, sulfur transfer functions for TusA and DsrEFH, and a substrate-donating function for DsrC. As detailed in the text, the model is based on biochemical as well as on molecular genetic evidence

Full size image

A rhd-tusA-dsrE2 or at least a tusA-dsrE2 arrangement occurs in all currently genome-sequenced sulfur oxidizers harboring the Dsr system (Venceslau et al. 2014) (Table 5). In A. vinosum the tusA and the rhd and the dsrE2 gene follow the same pattern of transcription as observed for the established cytoplasmic sulfane sulfur-oxidizing proteins (i.e., the Dsr system) (Stockdreher et al. 2014; Weissgerber et al. 2013). A rhd-tusA-dsrE2-deficient A. vinosum mutant strain, although not viable in liquid culture, was clearly sulfur oxidation negative upon growth on solid media containing sulfide (Stockdreher et al. 2014). TusA is one of the major proteins in A. vinosum, and the rhd and possibly also the dsrE2A encoded protein were identified as entry points for sulfur delivery to this protein (Stockdreher et al. 2014). The rhodanese-like Rhd protein (Alvin_2599) catalyzes sulfur transfer from thiosulfate or glutathione persulfide (GSSH) to cyanide in vitro, and the TusA protein was clearly established as a protein accepting sulfane sulfur from the A. vinosum rhodanese (Stockdreher et al. 2014). The DsrE2A protein is less well characterized and its role remains elusive at present (Stockdreher et al. 2014). It is firmly established that A. vinosum TusA is an interaction partner of DsrEFH, a hexameric protein arranged in a α2β2γ2 structure (Dahl et al. 2008c). Sulfur transfer between TusA and DsrEFH is reversible in vitro (Stockdreher et al. 2014). From DsrEFH sulfur is transferred to DsrC (Stockdreher et al. 2012).

The eminently important DsrC protein works as the physiological partner of the DsrAB sulfite reductase not only in sulfur-oxidizing but also in sulfate-reducing prokaryotes (Venceslau et al. 2014). DsrC is a member of the DsrC/TusE/RpsA superfamily and contains two strictly conserved redox active cysteines in a flexible carboxy-terminal arm (Cort et al. 2008): CysA is the penultimate residue at the C-terminus and CysB is located ten residues upstream (Venceslau et al. 2014). When combined in solution in their native, non-persulfurated state, DsrEFH and DsrC form a tight complex (Stockdreher et al. 2012), and each DsrE2F2H2 heterohexamer associates with either one or two DsrC molecules. Interaction of DsrEFH with DsrC is strictly dependent on the presence of DsrE-Cys78 and DsrC-CysA (Cort et al. 2008; Stockdreher et al. 2012).

In Fig. 1 the concept is implemented that persulfurated DsrC serves as the substrate for DsrAB and oxidation of DsrC-CysA-S− by this enzyme is thought to result in persulfonated DsrC (DsrC-CysA-S03 −) from which sulfite is possibly released by the formation of a disulfide bridge between CysA and CysB (Stockdreher et al. 2014; Venceslau et al. 2014). However, this proposal is challenged by the very recent finding that a DsrC trisulfide, in which a sulfur atom is bridging the two conserved cysteine residues , is released as the product of the reverse reaction, i.e., sulfite reduction, upon catalysis by DsrAB from a sulfate reducer (Santos et al. 2015). An alternative model is represented in Fig. 4 integrating formation of a DsrC trisulfide possibly by the action of the membrane-bound DsrMKJOP electron-transporting complex that contains the heterodisulfide reductase-like subunit DsrK which could well characterize the suggested reaction (Grein et al. 2010a, 2010b, 2013; Sander et al. 2006). Electrons released during oxidation of the DsrC trisulfide to DsrC and sulfite may be transferred to the protein DsrL. This iron-sulfur flavoprotein is essential for sulfur oxidation in A. vinosum (Dahl et al. 2005; Lübbe et al. 2006). It bears striking sequence similarity to the electron-bifurcating subunit of the NfnAB complex from Thermotoga maritima (Demmer et al. 2015) and would have the theoretical capacity for reduction of NAD+; however, experimental evidence substantiating this idea is currently completely lacking. Understanding the exact mechanistic details of the interaction of DsrC, DsrAB, and the other Dsr proteins is, in fact, one of the most challenging points in research on sulfur-oxidizing prokaryotes.

The Hdr-Like System of Sulfur Oxidation

The rhd-tusA-dsrE2 arrangement does not only occur in all currently genome-sequenced sulfur oxidizers harboring the Dsr system but also in a wide array of chemo- and also phototrophic sulfur oxidizers that do not contain the Dsr pathway (Venceslau et al. 2014) (Table 5). In these sulfur oxidizers, a gene cluster hdrC1B1AhyphdrC2B2 encoding an array of proteins resembling different subunits of archaeal heterodisulfide reductases is inevitably present (Venceslau et al. 2014). As shown in Table 5, genes encoding a putative hdr-like complex occur in several phototrophic representatives of the family Ectothiorhodospiraceae . The typical arrangement of the hdr-like gene cluster is shown in Fig. 5 for a chemotrophic sulfur oxidizer ( Acidithiobacillus caldus ) and the phototroph Halorhodospira halochloris . In almost all cases, the hdr-like gene set is immediately linked with rhd-tusA-dsrE2 or rhd-dsrE2 arrangements promoting the notion that an Hdr-like protein complex is involved in the generation of sulfite from disulfide or even more likely protein-bound persulfide intermediates formed during sulfur oxidation. Heterodisulfide reductases (Hdr) are enzymes present in methanogenic archaea and catalyze the reduction of the heterodisulfide, CoM-S-S-CoB, formed in the last step of methanogenesis (Hedderich et al. 2005; Thauer et al. 2008). The general idea of an involvement of a Hdr-like complex and probably also specialized sulfurtransferases (Rhd, DsrE, TusA) in sulfite formation was first put forward by Quatrini and coworkers on the basis of microarray transcriptome profiling and quantitative RT-PCR analyses performed with A. ferrooxidans ATCC 23270 (Quatrini et al. 2006, 2009). The suggestion found support in further transcriptional regulation studies not only on several Acidithiobacillus species (Chen et al. 2012; Ehrenfeld et al. 2013; Latorre et al. 2016) and the Gram-positive Sulfobacillus thermosulfidooxidans (Guo et al. 2014) but also on the thermoacidophilic archaeon Metallosphaera sedula (Auernik and Kelly 2010). In addition, proteomic studies showed high levels of Hdr-like proteins in the presence of reduced sulfur compounds (Mangold et al. 2011; Osorio et al. 2013; Ouyang et al. 2013). In several of the cited studies, upregulation in the presence of reduced inorganic sulfur compounds affected the hdr-like genes as well as the sulfur transferase genes. Tight functional interaction of the encoded proteins is further indicated by the observation that genes dsrE to hdrB2 constitute a single, distinct transcriptional unit in A. ferrooxidans ATCC 16786 (Ehrenfeld et al. 2013). The whole concept is further substantiated by the recent purification of a Hdr-like complex from membranes of Aquifex aeolicus (Boughanemi et al. 2016).

Fig. 5
figure 5

Comparison of the hdr-like gene cluster in Acidithiobacillus caldus and Halorhodospira halophila . The soeABC genes encode a membrane-bound cytoplasmically oriented sulfite-oxidizing enzyme. Rhd, TusA, and DsrE are sulfur-mobilizing and sulfur-transferring proteins, respectively. HdrB1, HdrB2, HdrC1, HdrC2, and HdrA bear similarity to the HdrABC subunits of soluble heterodisulfide reductases from methanogens. Hyp indicates a gene for a hypothetical protein. LbpA1 and LbpA2, lipoate-binding proteins, and LplA, single-domain protein lipoate ligase or more probably octanoylate transferase (Christensen et al. 2011; Christensen and Cronan 2010). radSAM1 and radSAM2 are annotated as radical SAM proteins and could insert sulfur into octanoylated LbpA. Several of the hdr-like gene clusters in sulfur oxidizers encode a protein putatively involved in fatty acid transport which could play a role in import of lipoate precursors GGred, similarity to geranyl geranyl reductase, and could be involved in modification of imported fatty acids before they are channeled into the specific lipoylation pathway

Full size image

Ehrenfeld et al. 2013 were the first to point out the presence of a gvcH-like gene encoding a lipoate-binding protein (Ehrenfeld et al. 2013). On the basis of striking sequence similarity and the presence of a strictly conserved lysine residue known to be required for lipoate attachment (Spalding and Prigge 2010), the name LbpA (lipoate-binding protein A) is suggested for this single lipoyl domain protein. Many sulfur oxidizers carry two copies of the gene indicating a functional dimer. Furthermore, genes encoding another DsrE-like sulfurtransferase and proteins with the potential to act in biosynthesis of protein-bound lipoic acid (two radical SAM proteins, a lipoate-protein ligase, and geranyl geranyl reductase-like protein) are inevitably found in organisms containing hdr-like genes but not in sulfur oxidizers pursuing the Dsr pathway (Fig. 5, Table 5). In most cases these genes are immediately linked with the hdr genes as shown in Fig. 5, and in some cases they are located at other places in the genome (e.g., in the purple sulfur bacterium Thiorhodospira sibirica , Table 5).

Overall, the present circumstantial evidence is quite overwhelming in the argument that a Hdr-like enzyme system including dedicated sulfur transferases (Rhd, DsrE, TusA) and also a dedicated lipoate-binding protein is a central and key element in the bioenergetics of sulfur-oxidizing prokaryotes devoid of the Dsr system (Bobadilla Fazzini et al. 2013; Chen et al. 2012; Dahl 2015; Guo et al. 2014; Mangold et al. 2011; Quatrini et al. 2009; Venceslau et al. 2014). However, genetic experiments that would finally prove this omics-derived concept have so far not been published for any organism, and biochemical studies that would shed light on the underlying reaction mechanism(s) are completely lacking.

Currently, it appears premature to suggest a more detailed model of the Hdr-like mechanism. The LbpA protein is a prime candidate as a sulfur substrate-binding entity that presents the sulfur substrate to different catalytic entities. However, further functions can at present not be assigned.

Oxidation of Sulfite to Sulfate

The last step in the oxidation of reduced sulfur compounds is the oxidation of sulfite yielding sulfate as the final product. Sulfate formation from sulfite is energetically favorable and carried out by a wide range of organisms (Simon and Kroneck 2013). In addition, many purple sulfur bacteria can even use externally available sulfite as photosynthetic electron donor. Two fundamentally different pathways for sulfite oxidation have been well characterized in chemotrophic and phototrophic sulfur-oxidizing bacteria: (1) direct oxidation and (2) indirect, AMP-dependent oxidation via the intermediate adenylylsulfate (adenosine-5′-phosphosulfate).

Oxidation of Sulfite in the Periplasm

Many sulfite-oxidizing enzymes catalyzing direct oxidation of sulfite are located outside the cytoplasmic membrane (in the periplasm in Gram-negative bacteria). The best characterized enzyme belonging to this group, SorAB, stems from the chemotroph Starkeya novella and consists of a molybdopyranopterin (Mo-PPT) cofactor-carrying subunit (SorA) and a monoheme cytochrome c (SorB) (Kappler et al. 2000; Kappler and Bailey 2005). SorA-type molybdoproteins without a SorB subunit have been termed SorT (D’Errico et al. 2006; Wilson and Kappler 2009), but recently this discrimination has been questioned (Simon and Kroneck 2013). Neither genes closely related to sorAB nor those encoding SorT sulfite dehydrogenases occur in the currently available genomes of anoxygenic phototropic bacteria.

A second option for oxidation of sulfite in the periplasm is the Sox system. It has been shown that sulfite is accepted in vitro as a substrate of the reconstituted Sox system from the chemotroph Paracoccus denitrificans (Friedrich et al. 2001; Frigaard and Dahl 2009; Sander and Dahl 2009). Notably, Friedrich and coworkers proved this reaction to be independent on the presence of SoxCD , a molybdohemoprotein catalyzing the six-electron oxidation of SoxY-cysteine-bound persulfide to sulfone sulfur. Purple bacteria that form sulfur globules during thiosulfate oxidation contain the Sox system albeit without the SoxCD proteins (Frigaard and Dahl 2009; Hensen et al. 2006; Meyer et al. 2007). Notably, the presence of SoxB and SoxXA is not essential for sulfite oxidation in A. vinosum (Hensen et al. 2006).

However, the periplasmic sulfur substrate-binding protein SoxYZ is needed in parallel to cytoplasmic enzymes for effective sulfite oxidation in A. vinosum (Dahl et al. 2013). Genes for this protein are present in purple sulfur bacteria irrespective of the organisms’ substrate range with only one exception (Thioflaviococcus mobilis), while the presence of SoxXA(K) and SoxB appears to be strictly linked to the ability of the cells to utilize thiosulfate (Table 2).

Oxidation in the Cytoplasm

Indirect Pathway via Adenosine 5′-Phosphosulfate

It is firmly established that a number of purple as well as green anoxygenic phototrophic sulfur bacteria oxidize sulfite in the cytoplasm using an indirect pathway via adenosine-5′-phosphosulfate (APS) catalyzed by APS reductase (AprBA) and ATP sulfurylase (Sat) (Dahl 1996; Frigaard and Dahl 2009; Parey et al. 2013; Rodriguez et al. 2011; Sanchez et al. 2001) (Fig. 1).

In A. vinosum , the sat gene encoding ATP sulfurylase (Alvin_1118) is located immediately upstream of the aprMBA genes encoding membrane-bound APS reductase (Alvin_1119–1121) (Hipp et al. 1997; Weissgerber et al. 2011). AprM is predicted to contain five transmembrane helices with no sequence similarity to any currently known conserved domain or cofactor binding site in the databases. An essential function of AprM as a membrane anchor that allows spatial and functional association of this type of oxidative APS reductase with the membrane has been postulated, and it has been suggested that AprM serves as an entry point into the membrane for the electrons released during formation of APS from sulfite and AMP (Meyer and Kuever 2007). In the currently available complete genome sequences of phototrophic members of the family Chromatiaceae , the same gene arrangement is present in Thiorhodovibrio sp. 970 (Table 6). In Thiocapsa marina 5811, Thiorhodococcus drewsii AZ1, Thiocystis violascens DSM 198T, and Thioflaviococcus mobilis DSM 8321T, sat and aprMBA are not linked on the chromosome (Table 6). The occurrence of aprMBA has also been reported for Thiococcus pfennigii 4520 (Gregersen et al. 2011).

Table 6 Genes related to sulfite oxidation in genome-sequenced purple sulfur bacteria a
Full size table

The QmoABC complex was first identified in the dissimilatory sulfate-reducing bacterium Desulfovibrio desulfuricans (Pires et al. 2003). The complex consists of one membrane (QmoC) and two cytoplasmic subunits (QmoAB). The two QmoC hemes b are reduced by quinols, and experimental evidence strongly indicates that the Qmo complex participates in electron flow between the quinone pool and the cytoplasm, i.e., that it acts as the electron-donating unit for APS reductase in sulfate reducers (Frigaard and Dahl 2009; Ramos et al. 2012). The qmoABC genes are not only present in sulfate-reducing prokaryotes (Ramos et al. 2012) but occur also in many chemotrophic sulfur-oxidizing bacteria as well as in green sulfur bacteria (Frigaard and Dahl 2009; Rodriguez et al. 2011) and in one further purple sulfur bacterium (Thiodictyon sp. Cad16 (Gregersen et al. 2011)). In sulfur oxidizers, QmoABC is thought to act as electron acceptor for the electrons released during formation of APS and would thus have a function analogous to that of AprM. It is thus conceivable to state that the electrons generated by the oxidative formation of APS from sulfite and AMP are fed into the photosynthetic electron transport chain on the level of menaquinone either by AprM or by the much better characterized QmoABC complex (Grein et al. 2013; Meyer and Kuever 2007; Ramos et al. 2012; Rodriguez et al. 2011). It may be especially advantageous to be equipped with the Qmo-related electron-accepting unit for APS reductase. The presence of the HdrA-like QmoA in the Qmo complex opens the possibility that—in reverse to the mechanism suggested for sulfate reducers (Grein et al. 2013; Ramos et al. 2012)—an electron bifurcation occurs that could result in simultaneous reduction of low potential electron acceptors like ferredoxin or NAD+. Such a process would be of significant energetic advantage especially for chemolithoautotrophic growth because it would result in a lower energy demand for reverse electron flow.

Direct Pathway via SoeABC

Notably the APS reductase pathway is neither generally present in purple sulfur bacteria (Table 6) nor is it essential in A. vinosum (Dahl 1996; Sanchez et al. 2001). The sat and aprBA genes are not present in some members of the Chromatiaceae (Meyer and Kuever 2007) and generally absent in Ectothiorhodospiraceae (Table 6). Recently the membrane-bound iron-sulfur molybdoprotein SoeABC was identified as a major enzyme catalyzing direct oxidation of sulfite to sulfate in the cytoplasm of A. vinosum (Dahl et al. 2013). The function of SoeABC was proven by strongly reduced specific oxidation rates for externally supplied sulfite and by massive excretion of sulfite into the medium during oxidation of sulfide in A. vinosum SoeABC-deficient strains. Crude extract of a SoeABC-deficient A. vinosum lacked AMP-independent sulfite-oxidizing activity. Further indication for an involvement of SoeABC in dissimilatory sulfur oxidation in A. vinosum was gathered during recent genome-wide transcriptional profiling (Weissgerber et al. 2013). Relative transcription of all three A. vinosum soe genes was found to be increased about threefold during photolithoautotrophic growth on sulfide or thiosulfate than during photoorganoheterotrophic growth on malate (2.99-, 2.77-, and 2.93-fold increase on sulfide and 1.96-, 1.98-, and 3.00-fold increase on thiosulfate, for soeA, soeB, and soeC, respectively). Changes in the same range were observed for the genes encoding the enzymes of the APS reductase pathway when thiosulfate replaced malate, while relative transcript levels for the sat-aprMBA genes were 7.6–9.7-fold higher in the presence of sulfide compared to the presence of malate

In A. vinosum , SoeABC is encoded by genes Alvin_2491 (soeA), Alvin_2490 (soeB), and Alvin_2489 (soeC). The protein consists of the 108.95 kDa molybdoprotein SoeA carrying one [Fe4S4] cluster at the N-terminus; the 26.995 kDa iron-sulfur protein SoeB, which upon comparison with related structurally characterized proteins (Jormakka et al. 2008) is predicted to bind four [Fe4S4] clusters; and a 35.715 kDa NrfD-/PsrC-like membrane protein (Simon and Kern 2008) with eight transmembrane helices. Neither AvSoeA and AvSoeB nor any of the other purple sulfur bacterial SoeA or SoeB proteins listed in Table 6 are synthesized with cleavable TAT signal peptides that are usually present on the active site subunits of the biochemically well-characterized periplasmic sulfur-metabolizing complex iron-sulfur molybdoproteins, i.e., polysulfide and sulfur reductase (PsrABC, SreABC), thiosulfate reductase (PhsABC), or tetrathionate reductase (TtrABC) (Heinzinger et al. 1995; Hensel et al. 1999; Krafft et al. 1992; Laska et al. 2003). SoeA and SoeB are thus located in the cytoplasm and attached to the cytoplasmic membrane by interaction with SoeC. The holoprotein is therefore well suited for oxidation of sulfite generated in the cytoplasm. It should be noted that SoeABC and the periplasmic Sor-type sulfite dehydrogenases belong to completely different families of molybdoenzymes .

Genes encoding proteins related to SoeABC are present in purple as well as green sulfur bacteria and have in the past years repeatedly been speculated to be involved in the oxidation of sulfite generated by the Dsr system in the cytoplasm (Frigaard and Bryant 2008b; Frigaard and Dahl 2009) (Table 6). Notably, soeABC-like genes co-localize with dsr genes in several green sulfur bacteria and in Halorhodospira halophila (Dahl 2008; Frigaard and Dahl 2009).

The possession of the APS reductase pathway in addition to or instead of SoeABC may be advantageous because additional energy is gained by substrate phosphorylation in the ATP sulfurylase catalyzed step by transferring the AMP moiety of APS onto pyrophosphate (Parey et al. 2013).

Sulfate Assimilation

Sulfate assimilation by in anoxygenic phototrophic bacteria has been extensively covered in previous reviews (Frigaard and Dahl 2009; Sander and Dahl 2009). Some anoxygenic phototrophic bacteria are very much specialized for living in habitats with reduced sulfur compounds and such bacteria usually completely lack a sulfate reduction pathway. On the other hand, very many versatile purple sulfur and non-sulfur bacteria and even a few green bacteria are able to assimilate and reduced sulfate in the absence of a reduced source of sulfur. Among the filamentous anoxygenic bacteria, the ability to assimilate sulfate may or may not be present.

Here, the assimilatory sulfate reduction pathway in A. vinosum is presented as an example (Fig. 6). The pathway commences with the uptake of sulfate via the membrane-bound components of a periplasmic substrate-binding transport system similar to the situation in E. coli (Kredich 1996). Once inside the cell, sulfate is activated to adenosine-5′-phosphosulfate by the enzyme ATP sulfurylase (Leustek and Saito 1999). Assimilatory ATP sulfurylases occur in two different forms: a heterodimeric CysDN type as in E. coli (Leyh 1993) and a homo-oligomeric Sat-related type as found in other bacteria, plants, and fungi (Foster et al. 1994; MacRae et al. 2001). Both types occur in anoxygenic phototrophic bacteria (Frigaard and Dahl 2009; Sander and Dahl 2009). The sulfate reduction pathway in A. vinosum does not involve formation of phosphoadenosine- 5′-phosphosulfate (Neumann et al. 2000). Instead, a CysH-type iron-sulfur cluster binding APS reductase catalyzes reductive cleavage of APS yielding sulfite and AMP. Sulfite is finally reduced to sulfide by an assimilatory sulfite reductase. In the case of A. vinosum, this enzyme is a ferredoxin-dependent CysI-type siroheme-[4Fe-4S] cluster-containing protein as it also occurs in cyanobacteria, algae, and higher plants (Dhillon et al. 2005). This enzyme type is common in anoxygenic phototrophic bacteria (Frigaard and Dahl 2009). Biosynthesis of cysteine requires the formation of O-acetyl-l-serine, which is then further transformed to cysteine catalyzed by cysteine synthase B (CysM) in a reaction that is dependent on the availability of sulfide (Fig. 6) (Hensel and Trüper 1976). It is well established that the CysTWA ABC-type transporter in conjunction with the periplasmic binding protein CysP transports not only sulfate but also thiosulfate into the cytoplasm (Sirko et al. 1995). In Salmonella typhimurium and E. coli, cysteine synthase B (CysM) also accepts thiosulfate as a substrate and hooks it up to O-acetyl-serine resulting in the formation of S-sulfocysteine (Kredich 1992). S-sulfocysteine is then reduced to cysteine resulting in the release of sulfite (Nakatani et al. 2012; Sekowska et al. 2000). Glutathione, thioredoxins, or glutaredoxins have been discussed as possible reductants in this reaction (Funane et al. 1987; Nakatani et al. 2012; Woodin and Segel 1968). A similar reaction sequence is also probable for the assimilation of thiosulfate in A. vinosum (Fig. 6). In fact, thiosulfate was previously detected intracellularly in A. vinosum (Franz et al. 2009).

Fig. 6
figure 6

Current model of assimilatory sulfate reduction in A. vinosum . CysE serine O-acetyltransferase (Alvin_0683), CysM cysteine synthase B (Alvin_2228), GshA glutamate/cysteine ligase (Alvin_0863), CysM cysteine synthase B (Alvin_2228); GshA glutamate/cysteine ligase (Alvin_800), GshB glutathione synthetase (Alvin_0197), γ-GluCys γ-glutamylcysteine, GSH glutathione, XSH glutathione, reduced thioredoxin or glutaredoxin, XSSX oxidized glutathione, thioredoxin or glutaredoxin (see text for further explanation). The transcriptomic (boxes) (Weissgerber et al. 2013), proteomic (circles) (Weissgerber et al. 2014a), and metabolomic profiles (triangles) (all relative to growth on malate) are depicted next to the respective protein/metabolite. Relative fold changes in mRNA levels above 2 (red) were considered significantly enhanced. Relative changes smaller than 0.5 (blue) were considered as indicating significant decreases in mRNA levels. Relative fold changes between 0.5 and 2 (gray) indicated unchanged mRNA levels. The same color coding is applied to changes on the protein and metabolome levels. Here, values above 1.5 (red) and below 0.67 (blue) were considered significant. Those cases, where transcriptomic data was not available or the respective protein or metabolite was not detected in the proteomic or metabolomic approach, respectively, are indicated by white squares, circles, or triangles. Sulfur compounds added, from left to right, sulfide, thiosulfate, elemental sulfur, and sulfite. Changes on sulfite were not determined on the proteome and metabolome levels. Figure reproduced from (Weissgerber et al. 2014b)

Full size image

During photoorganoheterotrophic growth of A. vinosum on organic acids like malate, sulfide for biosynthesis of sulfur-containing cell constituents is provided by the assimilatory sulfate reduction pathway in an energy-consuming process (Fig. 6) (Neumann et al. 2000), while sulfide is readily available without any input of energy under sulfur-oxidizing conditions. Accordingly, the presence of reduced sulfur compounds results in elevated relative mRNA and protein levels for genes/proteins of central enzymes of oxidative sulfur metabolism, while transcript and protein levels for genes/proteins involved in assimilatory sulfate reduction are negatively affected (Weissgerber et al. 2013, 2014a). These responses are positively correlated to the concentration changes of the metabolites of the affected metabolic pathways (Weissgerber et al. 2014b) (Fig. 6). It is conceivable to assume that the interplay between the processes of dissimilatory sulfur oxidation and assimilatory sulfate reduction is regulated in a similar manner in other anoxygenic phototrophic bacteria capable of pursuing both pathways.