Cell Biology of a Purple Sulfur Bacterium: Fluorescence Micros-Copy of Intracellular Sulfur Globules in Allochromatium vinosum

1. Introduction

A large proportion of sulfur-oxidizing bacteria form conspicuous sulfur deposits as intermediates during the oxidation of sulfide, polysulfides or thiosulfate [1,2,3,4]. The sulfur globules are deposited either extracellularly or intracellularly [1,5,6]. The formation of extracellular sulfur globules is characteristic of green sulfur bacteria and sulfur oxidizers of the gammaproteobacterial family Ectothiorhodospiraceae, while intracellular sulfur globules are typical of magnetotactic sulfur oxidizers, purple sulfur bacteria of the family Chromatiaceae, and sulfur-oxidizing bacterial endosymbionts [1]. The general target compartment for intracellular sulfur storage is the bacterial periplasm where the sulfur globules are encased by a protein envelope consisting of one or more structural proteins similar to cytoskeletal keratins or plant cell wall proteins [1,7].

A prominent and well-studied example of intracellular sulfur deposition are the sulfur globules of the anoxygenic phototrophic purple sulfur bacterium Allochromatium vinosum (class Gammaproteobacteria, family Chromatiaceae). In A. vinosum, the sulfur globules reside in the periplasm, i.e. in the same cellular compartment as the periplasmic thiosulfate- and sulfide-oxidizing enzymes. This is evidenced by the presence of signal peptide coding sequences in the genes for the three structural proteins of the sulfur globule envelope that have been studied so far [5,8,9]. In A. vinosum, the so-called rDsr system with the reverse-acting dissimilatory sulfite reductase rDsrAB as the key enzyme is essential for further oxidation of sulfur in the cytoplasm [10,11,12].

The sulfur globules of the Chromatiaceae, including those of A. vinosum, reach up to 1 μm in diameter [5] and account for up to 34% of the total cell dry weight [13,14]. The three established A. vinosum sulfur globule proteins SgpA (Alvin_1905), SgpB (Alvin_0358) and SgpC (Alvin_1325) have in common that they are extremely hydrophobic. SgpA and SgpB are very similar in amino acid sequence and can functionally replace each other. SgpC participates in sulfur globule expansion [7,9]. In 2014, we re-evaluated the A. vinosum sulfur globule proteome and identified the protein Alvin_2515 as a new putative component, which we named SgpD [15]. It resides next to genes for an hydrophobic/amphiphilic exporter and a probable two-component transcriptional regulator. Transcript levels for the sgpD gene dramatically increase on sulfide and thiosulfate compared to growth on malate in the absence of oxidizable sulfur compounds (28-fold and 6-fold, respectively) [16]. SgpD is synthesized with a cleavable Sec-dependent N-terminal signal peptide predicted to mediate transport to the periplasm. The protein could have a coil-coil structure typical of structural proteins, such as bacterial cytoskeletal proteins [15]. However, it should be noted that a recent re-evaluation of coil-coil prediction tools revealed a high number of incorrect predictions, seriously questioning their informative value [17].

At present, direct experimental evidence for a function of SgpD as a sulfur globule protein is lacking. This would require definitive information on the intracellular localization of the protein and its possible association with sulfur deposits in vivo. Fluorescent reporter proteins, such as green fluorescent protein (GFP), are valuable non-invasive molecular tools for real-time in vivo imaging of living specimens and have the greatest potential to address these questions. One limitation of fluorescent proteins in pigmented phototrophic bacteria is signal quenching when they emit light at a wavelength absorbed by the pigments, as discussed for the anoxygenic phototrophic Alphaprotebacteria Rhodobacter capsulatus and Rhodopseudomonas palustris [18,19]. GFP is an example of a fluorescent protein that is incompatible with bacteriochlorophyll a and carotenoids, the pigments that drive photosynthesis not only in R. capsulatus and R. palustris but also in A. vinosum. In addition, commonly used variants of GFP are not suited for investigating the subcellular localization of periplasmic proteins. When exported to the periplasm in an unfolded conformation through the Sec system, they fail to fold properly and do not fluoresce [20]. However, very good alternatives to GFP, such as mCherry, mStrawberry, and tdTomato, derived from screening serial mutants and genetic modification of the original GFP [21,22,23], are now available and widely used for gene expression measurement, protein localization, in situ screening, and multi-omic profiling. Among them, the fluorescent protein mCherry stands out because of its bright signal, rapid maturation, high photostability, high N-terminal fusion tolerance, and excellent pH tolerance [24]. Importantly, red fluorescent protein derivatives such as mCherry do not share the transport difficulties of their GFP relatives and can be effectively transported through the Sec system [25,26].

Unfortunately, despite all its advantages, mCherry shares with almost all fluorescent proteins a strict requirement for molecular oxygen for the maturation of the fluorophore [27], making it inherently difficult to use under anaerobic conditions, such as during the phototrophic growth of A. vinosum. While there is no general limitation to aerobic systems, there is still not much known about the exact conditions that allow full maturation of the fluorophore after exposure of anaerobically grown cells to oxygen [28,29,30]. Flavin-based fluorescent protein presents an alternative, but all available derivatives have two major limitations. They produce only cyan-green fluorescence incompatible with the A. vinosum pigments and the fluorescence emitted is significantly dimmer compared to GFP [31] or newer anaerobic fluorescent reporters such as Fluorescence-Activating Absorption-Shifting Tag, FAST [32]. The latter has so far been used primarily in eukaryotic systems and, when used in bacteria, is best suited for secretion studies [31] and has only been applied under a limited number of conditions, e.g. in E. coli during nitrate or fumarate respiration [31,33]. SNAP-tag and Halo-Tag are further promising reporters for fluorescent labeling in the absence of oxygen, but so far have only been adapted to few anaerobic bacteria, i.e. Clostridium species [34], which belong to the phylum Bacillota, and Bacteroides thetaiotaomicron (phylum Bacteroidota) [35].

After assessment of the available methods, we chose mCherry coupled to aerobic fluorescence recovery of the anaerobically produced protein as the most promising method for in situ fluorescence labeling and protein localization in the anoxygenic phototroph A. vinosum. Using mCherry we localized SgpD to the sulfur globules of the purple sulfur bacterium and collected evidence that its presence is not essential for sulfur globule formation.

2. Materials and Methods

2.1. Bacterial Strains, Plasmids, PCR Primers and Growth Conditions

The bacterial strains, plasmids and primers used in this study are listed in Table S1. All fluorescence experiments were carried out in a ΔsgpD strain derived from a spontaneous rifampicin-resistant mutant of the sequenced A. vinosum DSM 180^T. A. vinosum was cultivated photoorganoheterotrophically in RCV medium [36] or photolithoautotrophically in Pfennig medium [37] without reduced sulfur compound, referred to as “0 medium”. Sulfide or thiosulfate were added as electron sources at the desired concentration. The cultivation was performed under anoxic conditions and continuous illumination at 30°C in completely filled screw-capped culture bottles or on agar plates. Escherichia coli was cultivated in Luria Bertani medium [38]. Antibiotics used for mutant selection were applied at the following concentrations (in µg ml^-1): for E. coli: ampicillin 100, kanamycin 50, gentamycin 25, for A. vinosum: rifampicin 50, ampicillin 10, kanamycin 10, gentamycin 2.

2.2. Recombinant DNA Techniques

Standard methods were used for molecular biological techniques. Chromosomal DNA of A. vinosum strains was obtained by a modified sarcosyl lysis [39]. The genotypes of the A. vinosum recombinants used in this study were confirmed by Southern hybridization or PCR. Southern hybridization was performed over night at 68°C. PCR amplifications with Taq DNA polymerase and Pfu DNA polymerase were done essentially as described previously [40]. DNA probes for Southern hybridization were digoxygenin labelled by PCR. Plasmid DNA from E. coli was purified using the GenJet^TM Plasmid Miniprerp Kit (Thermo Scientific).

2.3. Construction of Plasmids for Production of mCherry in A. vinosum

We wanted to clone mCherry under the strong A. vinosum dsr promoter and used the replicative plasmid pBBR1MCS2-L [41] as the basis. This plasmid contains a XbaI-HindIII fragment of the PCR-amplified dsr promoter fused to gene dsrL in XbaI-HindIII of pBBR2MCS-2. The plasmid was first supplied with a gentamycin resistance cartridge which had been amplified from pBRRMCS-5 with primers Gent-CpoI-Fw and Gent-CpoI-rev (Table S1). The amplicon was digested with RsrII and inserted into the RsrII site present within the kanamycin resistance cartridge of pBBRMCS-2, resulting in plasmid pBBR1-L-Gm. The mCherry gene was amplified via PCR with primers for-mCherry-NdeI and rev-mCherry-SalI with plasmid pmCherry [24] as the template. After digestion with NdeI/SalI, the amplicon was cloned between the SalI and NdeI sites of pBBR1-L-Gm, giving plasmid pBBR_dsr_mCherry_Gm that carries the gene for mCherry under the control of the dsr promoter. Fusion of mCherry with the potential signal peptide for SgpD and its first amino acid (tryptophan) a s well as a methionine was achieved by cloning a PCR product generated with primers for-mCher-sig and rev-mCherry-SalI with plasmid pmCherry [24] as the template between NdeI and SalI of pBBR1-L-Gm. The complete sgpD gene was amplified with primers fAlvin2515-NdeI and rAlvin2515-NdeI cloned into the NdeI site of pBBR_dsr_mCherry_Gm, resulting in a fusion of SgpD including its signal peptide and mCherry.

2.4. Fluoresence Microscopy

For microscopy, cells were transferred to a microscopy slide coated with a thin film of 1% (w/vol) agarose. Fluorescence microscopy was carried out at room temperature using a Zeiss Axio Observer Z1 microscope (Zeiss, Jena, Germany) equipped with an HXP 120 V light source, an αPlan-APOCHROMAT 100×/1.46 oil immersion objective and an Axio Cam MR3 camera. Visualization of mCherry was achieved using Carl Zeiss filter set 64 HE (574–599 nm excitation, 605 nm beam splitter and 612-682 nm emission). Image acquisition and analysis were performed with Zen 2 software (Zeiss).

2.5. Construction of A. vinosum ΔsgpD::ΩKm

For the substitution of sgpD (Alvin_2515) in the genome of A. vinosum by an kanamycin cassette, SOE PCR [42] fragments were constructed using primer pairs Del-Alvin2515-fw1/Del-Alvin2515-rev1 and Del-Alvin2515-fw2/Del-Alvin2515-rev2 (Table S1). The resulting fragment was inserted into the mobilizable plasmid pSUP301 by HindIII restriction sites resulting in plasmid pSUP301_ΔsgpD. After digestion with EcoRI, the kanamycin cassette from pHP45ΩKm was ligated into the EcoRI site of pSUP301_ΔsgpD. The final mobilizable construct pSUP301_ΔsgpD_ ΩKm was transferred from E. coli S17.1 to A. vinosum Rif50 by conjugation [43]. Transconjugants were selected on RCV plates containing the appropriate antibiotics under anoxic conditions in the light. Double cross-over recombinants lost the vector encoded ampicillin resistance.

2.6. Characterization of Phenotypes, Detection of Sulfur Species and Protein Determination

A. vinosum wild-type and the ΔsgpD_ΩKm strains were characterized in batch culture experiments essentially as described before [16]. Cells of A. vinosum, grown photoorganoheterotrophically on malate (RCV medium [36]) for three days were used as an inoculum for experiments concerned with transformation of sulfide and thiosulfate. The culture volume of the precultures was 500 ml. Inoculum cells were harvested by centrifugation (10 min, 2680 x g) and washed once in “0” medium. The culture volume for phenotypic characterization was 250 ml. In these experiments, the starting optical density at 690 nm was set to 0.8 to 0.9. Sulfide, polysulfides, elemental sulfur and sulfate were either determined by HPLC [44] or using classical colorimetric or turbidometric methods as previously described [16,40].

3. Results

3.1. Suitability of mCherry for the Purple Sulfur Bacterium Allochromatium Vinosum

Here, we describe the production and validation of a direct fluorescence reporter for cell biological studies in an anoxygenic purple sulfur bacterium of the family Chromatiaceae (class Gammaproteobacteria). First, we show that intrinsic background fluorescence in the applied wavelength range is negligible (Figure 1A, middle panel). The next step was to determine whether mCherry is functional in A. vinosum, i.e. correctly expressed, translationally modified and folded. For this purpose, a reporter plasmid (pBBR_dsr_mCherry_Gm) was constructed in which the strong A. vinosum dsr promoter, which is inducible by sulfide, thiosulfate and elemental sulfur [16,41], was fused to the gene for the red fluorescent protein, mCherry [24,45], directed to the cytoplasm. The plasmid was transferred into A. vinosum via conjugation and the cells were then grown photolithoautotrophically in the light with 4 mM sulfide as electron donor. Figure 1B shows that the cells were fully and homogeneously fluorescent after 70 to 80 min of exposure to air. Our results are fully consistent with experiments in the anoxygenic phototroph R. palustris, in which mCherry also obtained fluorescence when switched from anoxic to oxic conditions [19]. The spectroscopic properties of mCherry are the same irrespective of whether it is originally produced under aerobic or anaerobic conditions [19].

3.2. Fusions with mCherry Reveals Intracellular Localization of SgpD in A. vinosum

The putative sulfur globule protein SgpD is equipped with a probable Sec-dependent signal peptide proposed to direct its transport across the cytoplasmic membrane into the periplasm of the Gram-negative bacterium A. vinosum [15]. The predicted periplasmic localization of SgpD would have been a further difficulty when using the classical GFP chromophore which is known not to be correctly assembled in the bacterial periplasm [20]. However, this is not the case for mCherry and accordingly brightly fluorescing cells were obtained when mCherry was fused to the signal peptide of SgpD (Figure 1C). A. vinosum is packed with intracytoplasmic membranes that are organized in vesicles (the so-called “chromatophores”). The interior of these chromatophores is extracytoplasmic, i.e. the entire cell body and not only its outermost layer as in other Gram-negative bacteria such as E. coli is filled with periplasmic space [8]. This fully explains the homogeneous distribution of mCherry when exported via the signal peptide (Figure 1C).

In the next step, mCherry was fused to full-length SgpD including its signal peptide. Now, fluorescence was no longer evenly distributed in the cells but concentrated to circular structures inside of the cells (Figure 1C, middle panel). Overlays of the fluorescence micrographs with images obtained by phase contrast microscopy showed that the circular structures exactly matched the position of sulfur globules in the cells and thus demonstrated that SgpD is attached to sulfur globules in vivo. Further evidence for this conclusion was obtained by microscopy of sulfur globules isolated from the A. vinosum strain with the SgpC-mCherry fusion (Figure 2). The sulfur globules were brightly fluorescent due to attachment of mCherry via SgpD.

3.3. Stability of mCherry Fluorescence in Allochromatium Vinosum

We sought to obtain more information about the development and stability of mCherry fluorescence in the purple sulfur bacterium and followed it via fluoresce microscopy over a period of 95 min. The fluorescence intensity gradually increased with oxygen exposure time and reached the maximum after a 60-min-exposure at room temperature. This maximum fluorescence was stably maintained for at least half an hour.

Figure 3. Development of fluorescence over time in A. vinosum Rif50 carrying pBBR_dsr_SgpD_mCherry. Left, middle and right panels show phase contrast microscopy, fluorescence microscopy and merged phase contrast and fluorescence images, respectively. Excitation times for phase contrast and for the mCherry channel (587 nm) were 49.2 s and 1 s, respectively. Size bar: 2 µm.

Figure 3. Development of fluorescence over time in A. vinosum Rif50 carrying pBBR_dsr_SgpD_mCherry. Left, middle and right panels show phase contrast microscopy, fluorescence microscopy and merged phase contrast and fluorescence images, respectively. Excitation times for phase contrast and for the mCherry channel (587 nm) were 49.2 s and 1 s, respectively. Size bar: 2 µm.

3.2. In Vivo Role of SgpD

The analysis of the A. vinosum sulfur globule proteome revealed SgpB as the second most abundant sulfur globule protein in this organism while SgpA and SgpC were less frequently detected [15]. The most abundant A. vinosum sulfur globule protein is SgpD (Weissgerber et al. 2014). The presence of a protein envelope around the sulfur inclusions in sulfur-oxidizing bacteria suggests an important structure–function relationship. Indeed, mutants of A. vinosum lacking SgpB and SgpC are no longer able to oxidize sulfide and thiosulfate and to form sulfur inclusions (Prange et al. 2004). SgpA and SgpB can replace each other in the presence of SgpC. SgpC probably plays a role in sulfur globule expansion. Information about the in vivo role of SgpD has so far not been available and we therefore set out to close this knowledge gap.

First, we re-evaluated its distribution within the bacteria and found that it is almost exclusively restricted to purple sulfur bacteria of the family Chromatiaceae with a threshold of e=10^-05 in Protein Blast searches. The only exceptions to this rule are a Gammaproteobacteria bacterium and a Chromatiales bacterium as well as the bacterial endosymbiont of Escarpia laminata, all of which cannot be more strictly taxonomically grouped, but could also be members of the Chromatiaceae.

With the other three sulfur globule proteins SgpA, SgpB and SgpC, A. vinosum SgpD shares a somewhat repetitive amino acid sequence with regularly spaced proline residues exemplified by the pattern P-X₂-P-X₂-P-X₄-P-X₂-P-X₂-P-X₂-P-X₅-P-X₂-PX₂-P-X₂-X₅P-X₂P in the central part of the protein. Scanning for repetitive sequences by RADAR [46] revealed four repeats (Figure 4).

An AlphaFold prediction for SgpD is available at https://alphafold.ebi.ac.uk/entry/D3RP35 and shown in Figure 5. Large predicted loop portions are modelled with low confidence (70 > pLDDT > 50). Two alpha-helices (amino acids 13 to 58 and 126 to 142) are predicted with very high confidence (pLDDT > 90). The first helix consists of the repeats 1 and 2 shown in Figure 4, while the second helix is made up of the residues in repeat 3. Clearly, experimental data is needed to gain further insight into the structural features of the proteins that make up the envelope of the sulfur globules.

Figure 5. Alphafold [47] prediction of A. vinosum SgpD. The signal peptide (amino acids 1-22) that is cleaved off after transport to the cytoplasm is hidden here.

Figure 5. Alphafold [47] prediction of A. vinosum SgpD. The signal peptide (amino acids 1-22) that is cleaved off after transport to the cytoplasm is hidden here.

In another approach to obtain information on the in vivo function of SgpD, we generated an A. vinosum ΔsgpD strain in which the gene was inactivated by insertion of a kanamycin Ω interposon [48]. The nucleotide sequence between the ATG and the TAA stop codons of sgpD was completely replaced by a kanamycin Ω interposon. Just as the other sgp genes, sgpD is preceded by a promoter sequence predicted by BPROM [49] (TTGATA-N₁₃-TTATATACT, ending 79 nucleotides upstream of the ATG start codon). An inverted repeat (CGGCCCCATCGACGATGGGGCCG) corresponding to the features of rho-independent transcription terminators [50] is found 45 nt downstream of the TAA stop codon. Thus, sgpD appears to form a separate transcription unit. Downstream polar effects of the Ω cassette should therefore be minimal. It should also be noted that the gene Alvin_2516, which follows sgpD at a distance of 216 nucleotides in the same direction of transcription, is under a completely different transcriptional control, as its transcript abundance is not affected by the presence of sulfide or thiosulfate in the medium [16].

Photoorganoheterotrophic growth of the sgpD-deficient mutant strain on malate-contianing medium with sulfate as the sole sulfur source was unaffected, thus excluding general growth defects. In addition, the ability of the mutant to grow photolithoautotrophically on sulfide-containing media and to store sulfur as intracellular sulfur globules was not affected (Figure 5). Similar to the wildtype, the mutant oxidized sulfide first to polysulfide and then to “zero-valent” polysulfane sulfur stored in sulfur globules, which was finally oxidized to sulfate (Figure 5A–C). The pattern of sulfur compounds in the growth experiments remained essentially the same regardless of the initial sulfide concentration, 2, 4 or 8 mM (Figure 5A–C,). Lack of SgpD neither influenced the size nor the quantity of sulfur globules (Figure 5D). This result is similar to findings for SgpB and SgpA [5,9].

Figure 5. Consumption of 4 mM (A), 6 mM (B) and (8 mM) sulfide by A. vinosum wildtype (filled symbols) compared to A. vinosum ΔsgpD (open symbols). Representative experiments are shown. The upper panels give sulfide (diamonds) and sulfate concentrations in culture supernatants. The lower panels show formation and degradation of the combination of the two main polysulfides (triangles) and of intracellular sulfur (red circles). Part D shows phase contrast micrographs of A. vinosum ΔsgpD at the indicated time points after addition of 4 mM sulfide.

Figure 5. Consumption of 4 mM (A), 6 mM (B) and (8 mM) sulfide by A. vinosum wildtype (filled symbols) compared to A. vinosum ΔsgpD (open symbols). Representative experiments are shown. The upper panels give sulfide (diamonds) and sulfate concentrations in culture supernatants. The lower panels show formation and degradation of the combination of the two main polysulfides (triangles) and of intracellular sulfur (red circles). Part D shows phase contrast micrographs of A. vinosum ΔsgpD at the indicated time points after addition of 4 mM sulfide.

4. Discussion

The anoxygenic phototrophic bacterium A. vinosum is widely distributed in sulfide-rich, light-penetrated environments and serves as one of the very few genetically accessible model organisms of the purple sulfur bacteria [43,51,52]. Here, we have expanded the genetic toolbox for A. vinosum by establishing a basic cell biological method that now allows protein localization in the living cell using a fluorescent protein reporter, mCherry. We show that mCherry produced in A. vinosum, originally grown in the absence of oxygen under photolithoautotrophic conditions, is fully fluorescent within one hour of exposure of the cells to air. Interference from intrinsic background fluorescence is negligible. When exported to the periplasm, mCherry also develops full fluorescence and specifically accumulates at the cellular target of the protein to which it is fused. In the experiments reported here, mCherry linked to a potential protein of the proteinaceous sulfur globule envelope did indeed appear exclusively in immediate vicinity of the intracellular sulfur deposits. To our knowledge, tagging with fluorescent proteins has so far not been established for any purple sulfur bacterium and we hope that the availability of the technique for A. vinosum will promote studies of the basic cell biology not only of this but also of other bacteria with the same physiology.

In the light, A. vinosum forms sulfur globules only under anaerobic conditions [8], while the red fluorescence protein mCherry needs molecular oxygen for the maturation of its chromophore. As pointed out by Jiang and coworkers [19], oxygen access into the fully folded barrel protein is of great importance, but may also trigger irreversible photobleaching and reduced photostability of a fluorescent protein [53]. In our experiments, the fluorescence signal of A. vinosum was fully recovered after exposure to oxygen for 60 min and was stably maintained at the maximum fluorescence intensity for at least 30 min. For comparison, fluorescence signals in the purple non-sulfur bacterium R. palustris required four hours oxygen exposure for full recovery [19]. In any case, our results, as well as those obtained by others with R. palustris, show that mCherry produced under anaerobic conditions matures and develops fluorescence in the presence of air.

Coupling mCherry to SgpD allowed us to show that SgpD is tightly bound to sulfur globules and represents a novel component of their proteinaceous envelope, although it is neither essential for growth on sulfide nor for sulfur globule formation during its oxidation. Organisms such as A. vinosum, which live in sulfide-rich and oxygen-poor environments, appear to be equipped with an at least partially redundant set of proteins [1,9] that ensure the compatibility of massive sulfur deposition within the cell boundary with energy metabolism and the formation of new biomass. Having a variety of sulfur globule proteins that can compensate for each other to some extent may allow for fine-tuned adaptation to environmental conditions. Possession of SgpD may be advantageous at either low or very high sulfide concentrations that do not allow photolithotrophic growth of pure batch cultures due to insufficient electron supply or sulfide toxicity, respectively, but may be necessary or tolerated in an environment where other members of the community compete for the substrate or remove it rapidly enough to minimize toxic effects. Without question, a lot of further work is needed to elucidate the interaction of all sulfur globule proteins and to gain insights into the structure of these exciting proteins.