**4. Discussion**

It is apparent that all three syntrophic mixtures analyzed, *Cp. ethylica* N2, N3, and 2-K, are virtually the same in both the green and colorless components. That is, the interaction between the two components of the mixtures appears to be highly specific, although no physical connection has been described and the mixtures were isolated from different habitats hundreds of miles apart. However, the finding that both the green component and the colorless component of the three syntrophic mixtures are virtually the same asks for some explanation. It would require to be quite a coincidence considering that the mixtures were isolated from habitats geographically separated by hundreds of miles in Russia

and that the type strain of *Desulfuromonas* was isolated from the Antarctic on the far side of the Earth. In support of coincidence, Biebl and Pfennig [18] reported that there is no specificity in the syntrophic growth of *Desulfuromonas* DSM 1675 with green bacteria. As far as growth yield is concerned, *Prosthecochloris* DSM 1685 could be replaced by *Chlorobium* strain DSM 258, and *Desulfuromonas* DSM 1675 could be replaced by *Desulfuromonas* DSM 684<sup>T</sup> in the syntrophic mixture. However, the doubling times were not measured, which might have shown a difference. A far better explanation of the identity of the three isolates, therefore, is that there is some undetected physical interaction between the two species in the mixtures.

A cell–cell complex, however fleeting, may require surface recognition proteins as previously postulated for the green bacterial consortia of *C. aggregatum* [13]. In that larger consortium, a single central, nonpigmented, heterotrophic bacterium forms close cell–cell interactions with multiple green-sulfur bacteria (15–40 cells) [13,14]. Even though the metabolic interactions in the *Cp. ethylica* symbiotic relationship are possibly different from the *C. aggregatum* and other large consortia, and are likely driven by sulfur cycling in the case of *Cp. ethylica*, the fact that similar large adhesin/agglutination proteins are detected in *Prosthecochloris ethylica* strains indicates that similar interactions may be formed in this more simple symbiont as those in the larger green sulfur consortia. The fact that this gene cluster appears to be present in other *Prosthecochloris* genomes as well (Figure 4) indicates that the capability to produce large outer membrane adhesin structures might be more widespread among green sulfur bacteria than anticipated.

The presence of unique Tad (tight adhesion) pili genes, only found in the *Prosthecochloris* strains that are known to form a symbiotic relationship, suggests that the structural formation of the cell–cell interactions occurs through specific pili and that large agglutination proteins are expected to help maintain this interaction. Both in the larger *C. aggregatum* and in a similar archaeal consortium with *Nanoarchaeum equitans*, fibers are observed that form periplasmic tubules surrounding the entire cell surface [14,20]. The exact nature of these fibers is still unresolved. However, tight adhesion pili, like the ones we found in the *Prosthecochloris ethylica* genomes, may be involved in the initial formation of the connecting tubules. Possibly, the pili are forming the initial connection before wider periplasmic tubules are established. This process would be similar to the better-known process of how conjugation pili (F pili) help to establish conjugation bridges during the process of DNA conjugation in many bacteria. It is interesting that both the Tad pili and the adhesion protein products described here are best known from studies of bacterial virulence factors. However, based on their presence in other nonpathogenic species, they seem to be more widespread amongs<sup>t</sup> bacteria and are likely involved in many environmentally important syntrophic interactions between bacteria.

In the *Chloropseudomonas ethylica* syntrophy, the colorless *Ds. acetoxidans* bacterium likely reduces sulfur compounds and returns the resulting H2S to the *Prosthecochloris* green sulfur component, as proposed by Biebl and Pfennig [18]. This would allow the green sulfur bacteria to use sulfide for their anoxygenic photosynthesis. Studies have shown that various cytochrome c7s can directly reduce metal ions or sulfur, and the *Desulfuromonas* C-551.5 is no exception in that respect [57,58]. Thus, this cytochrome may be responsible for sulfur cycling in the mixtures. In fact, it has previously been shown that *Ds. acetoxidans* cytochrome C-551 is capable of reducing polysulfides and is suggested to be the terminal reductase [59]. It was also shown that another strain of *Prosthecochloris* can grow by direct interspecies electron transfer with *Geobacter sulfurreducens* (a close relative of *Desulfuromonas*) as the electron donor and that they could form a cell–cell complex [54]. These results sugges<sup>t</sup> that the *Desulfuromonas* c7 (C-551.5) in the *Chloropseudomonas* mixtures may transfer electrons directly to *Prosthecochloris* without reducing sulfur to sulfide, although that too is a possibility, as shown above. Based on our findings of geopilin-type pili genes and of a larger multiheme cytochrome C gene found exclusively in the genomes of the strains that form these syntrophic mixtures, it is likely that a more complex system of electron transfer through e-pili and multiheme cytochromes is involved, with cytochrome C-551.5 functioning as the terminal sulfur reducing agen<sup>t</sup> in this complex.

Large multiheme cytochromes with monomeric molecular masses of 50 and 65 kDa (containing 6 and 8 hemes, respectively) were previously observed in *Ds. acetoxidans*, but no sequence data to compare with our current multiheme cytochrome C are available [59]. These multiheme cytochromes were found to cover an extremely wide range of reduction potential, but did not show any hydroxylamine oxidoreductase nor polysulfide reductase activity. Multiheme cytochromes of *Geobacter* and *Desulfuromonas* have been shown to play critical roles in the processing of many metals [57,60], although the monomers were much smaller in size (70–80 residues). A recent review on the role of multiheme cytochromes in anaerobic bacterial respiration [61] reports that multiheme cytochromes with a large molecular ratio of weight/heme ratio (7kDa/heme, or higher) appear to be most common in so-called electroactive organisms that are involved in the reduction of extracellular substrates, such as *Geobacteraceae* and *Shewanellaceae.* The mechanism used by *Geobacter* to transfer electrons onto solid extracellular substrates is still poorly understood, but involves both a pool of periplasmatic cytochromes and several outer membrane multiheme cytochromes. At least some of these form larger polymers [61,62]. The exact electron transfer process is still undetermined, but it is likely a complex multicomponent system where at least some of the multiheme cytochromes have overlapping functions covering the wide range of reduction potentials. By sequence homology alone we cannot state the function of the unique larger multiheme cytochrome C from the *Desulfuromonas acetoxidans* strains, but its high molecular weight/heme ratio (estimated to be 7.5 kDa/heme), N-terminal signal sequence, and unique homology to *Geobacter* large multiheme cytochromes point to a role in extracellular electron transfer.

Our findings from the simultaneously sequencing and comparison of the genomes in the *Cp. ethylica* syntrophic mixtures, in combination with observations described in related organisms, allow us to propose a basic model by which Tad pili and large agglutination proteins from the green sulfur *Prosthecochloris* are key elements in the formation of the syntrophic complex (Figure 9). This may be in conjunction with the formation of periplasmic tubules that were observed in larger photosynthetic consortia. The formation of larger consortia is dependent on cultivation conditions and is not easily reproduced under laboratory conditions. Although this has not been observed ye<sup>t</sup> in *Cp. ethylica*, given the detection of similar adhesion proteins and pili, it is possible that it could produce similar larger consortia under the right cultivation conditions. Once close cell–cell interactions are formed, the closed sulfur cycle can be established by electron transfer through specialized e-pili and several cytochromes produced by the *Desulfuromonas* component (Figure 9). Further biochemical studies will be needed to determine the exact electron transfer chain and the function of the cytochromes involved, but it is likely that cytochrome C-551.5 plays an essential role in the sulfur reduction. The sulfur metabolism of green sulfur bacteria involves the oxidation of sulfide and the deposition of elemental sulfur globules outside the cells [63]. The e-pili produced by *Desulfuromonas* are therefore expected to be involved in extracellular electron transfer to the sulfur acceptor deposited extracellularly by the green sulfur bacterium (Figure 9). The expression of both specific geopilin and MSHA pili (particularly the major pilin *msh*A gene) in the related organism *Pelobacter carbinolicus* was found to be substantially upregulated during ethanol oxidation, presumably for improved attachment or electron transfer to the extracellular electron acceptor S0 [50]. It requires further functional analysis, but the conserved ferredoxin protein directly upstream of the geopilin gene in *Ds. acetoxidans* may potentially be involved in the electron transfer through the e-pili.

**Figure 9.** Schematic overview of proposed interactions in the *Chloropseudomonas ethylica* syntrophic mixture. Unique protein structures identified in the genomes are indicated in green for *Prosthecochloris ethylica* and in brown/red for *Desulfuromonas acetoxidans*. Image created in BioRender.com.

Irrespective of the electron transfer proteins involved, the mutualistic metabolic benefits to each of the components are clear from the sulfur and electron cycling. In addition, the colorless *Desulfuromonas* also contains several genes for flagella production, and, therefore, provides the additional benefit of motility. The potential advantage of motility was previously proposed for the larger *C. aggregatum* consortium [14] and appears to be a part of the *Cp. ethylica* syntrophy as well. None of the green sulfur bacteria observed in these consortia produce flagella and obtaining motility provided by the nonpigmented component is likely to result in a competitive advantage to orient themselves much faster towards light and sulfide gradients in stratified or meromictic lakes.

**Author Contributions:** Conceptualization, J.A.K. and T.E.M.; methodology, J.A.K. and T.E.M.; software, J.A.K.; validation, J.A.K., J.J.V.B. and T.E.M.; formal analysis, J.A.K., J.J.V.B. and T.E.M.; investigation, J.A.K., J.J.V.B. and T.E.M.; resources, J.A.K. and T.E.M.; data curation, J.A.K., J.J.V.B. and T.E.M.; writing—original draft preparation, J.A.K. and T.E.M.; writing—review and editing, J.A.K., J.J.V.B. and T.E.M.; visualization, J.A.K.; project administration, J.A.K., J.J.V.B. and T.E.M.; funding acquisition, J.A.K. All authors have read and agreed to the published version of the manuscript.

**Funding:** This research received no external funding.

**Conflicts of Interest:** The authors declare no conflict of interest.
