**1. Introduction**

Phototrophic bacteria are widely distributed at suitable habitats in the marine and hypersaline environment. They are exposed to sometimes dramatically changing salt concentrations and some are found in saturated brines of salt and soda lakes, where they regularly develop massive blooms, often forming patches and pinkish-red layers, even within deposits of crystalized salts [1,2]. One of the prerequisites to cope with high salt and solute concentrations is the ability to keep an osmotic balance, i.e., a positive turgor pressure inside the cells through the accumulation of solutes in the cytoplasm that are compatible with the metabolic processes, even at high concentrations, and preserve active structures of proteins and nucleic acids [3]. In consequence, these bacteria need proper mechanisms of osmotic adaptation and ways to accumulate osmotically active compatible solutes up to several molar concentrations at the extremes.

Limitation to protect cell structures and metabolism is given by the compatibility of the solutes and the ability to accumulate to high or extremely high, several molar concentrations inside the cell. Glycine betaine (in some cyanobacteria also glutamate betaine) and ectoine (also hydroxyectoine) are the top candidates for this function in bacteria. Glycine betaine (hereafter "betaine") accumulation is widespread among phototrophic and

**Citation:** Imhoff, J.F.; Rahn, T.; Künzel, S.; Keller, A.; Neulinger, S.C.Osmotic Adaptation and CompatibleSolute Biosynthesis of Phototrophic Bacteria as Revealed from Genome Analyses. *Microorg* **2021**, *9*, 46. https://doi.org/10.3390/ microorganisms9010046

Received: 3 December 2020 Accepted: 23 December 2020 Published: 26 December 2020

**Publisher's Note:** MDPI stays neutral with regard to jurisdictional claims in published maps and institutional affiliations.

**Copyright:** © 2020 by the authors. Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (https:// creativecommons.org/licenses/by/ 4.0/).

chemotrophic eubacteria [4–8]. Ectoine was first identified as a compatible solute in the extreme halophilic *Halorhodospira halochloris* [9] and was later shown to be widely distributed among marine and halophilic eubacteria [10]. Though a number of other solutes are accumulated in bacteria in response to osmotic stress, these can provide protection only at low to moderate osmotic stress. Such compounds include sugars such as trehalose and sucrose, amino acids in particular glutamate, glucosyl glycerol, N-acetyl-glutaminyl glutamine amide, N-carbamoyl-glutamine amide, and others [6,7,11–13]. Basically, the accumulation can be achieved by uptake from the environment or by biosynthesis.

Betaine biosynthesis can be achieved by three consecutive methylation steps from glycine and includes the formation of monomethylglycine (sarcosine) and dimethylglycine as intermediates. In most of the bacteria studied, these enzymes are encoded by two genes and have overlapping enzymatic activities. In *Halorhodospira halochloris*, the first enzyme (glycine and sarcosine methyltransferase GMT) catalyzes the formation of monomethylglycine and dimethylglycine and the second one (dimethylglycine methyltransferase DMT) catalyzes the methylations to dimethylglycine and betaine [14,15]. In the halophilic cyanobacterium *Aphanothece halophytica*, the second enzyme specifically catalyzes the methylation of dimethylglycine [16]. In *Actinopolyspora halophila*, the two genes are fused, showing corresponding sequence homologies to the two genes in *Halorhodospira halochloris* [14].

An alternative route of betaine biosynthesis starts from choline and oxidizes this compound in two steps to betaine, catalyzed by choline dehydrogenase (BetA) and betaine aldehyde dehydrogenase (BetB). This route is widely distributed among bacteria but requires the external presence and uptake of choline. One possible uptake system is the high-affinity secondary transporter BetT, considered to be a specific choline transporter in *E. coli* [17] and a betaine transporter in *Aphanothece halophytica* [18]. In *E. coli*, the *betT* gene, together with a regulatory *betI* gene, is included in the *bet* gene cluster.

The biosynthesis of ectoine, which was first identified in the extreme halophilic phototrophic bacterium *Halorhodospira halochloris* [9], starts from aspartate. Aspartate is activated to L-aspartate-phosphate (Ask\_ect) and then reduced to L-aspartate-β-semialdehyde (Asd), followed by a transamination (with glutamate or alanine as donor of the amino group) to L-diaminobutyric acid (EctB), acetylation of the amino group to N-acetyldiaminobutyric acid (EctA), and finally ring closure (EctC) to form ectoine, as shown for *Halorhodospira halophila* and *Halomonas elongata* [19,20]. Genes responsible for ectoine biosynthesis and their osmotically regulated expression were first identified in *Marinococcus halophilus* [21]. The oxidation of ectoine to hydroxyectoine, which was first demonstrated in *Streptomyces parvulus* [22], turned out to be common to many Actinobacteria. The hydroxylation of ectoine (EctD) is strongly dependent on the presence of molecular oxygen and is accompanied by the oxidative decarboxylation of oxoglutarate forming CO2 and succinate [23]. While the *ask\_ect* gene, which encodes a specific aspartate kinase not underlying the feedback control of threonine [24,25], is often included in the *ect* gene cluster, *asd* usually is at a different locus in the genome. The coexpression of *ask\_ect* together with the osmotically induced gene cluster *ectABC*, ensures optimal supply of the precursor L-aspartate-β-semialdehyde under osmotic stress conditions. Aspects of ectoine and hydroxyectoine biosynthesis were recently reviewed by Czech et al. [10].

Osmotic adaptation can also be achieved by uptake of osmolytes rather than biosynthesis. Provided that such solutes are available in the environment, uptake generally is the favored way, because it is far less energetically expensive than de novo synthesis. Laboratory culture media with complex carbon sources (e.g., proteose peptone and yeas<sup>t</sup> extract), often contain these compounds. In such media, betaine was accumulated to high levels exceeding 1 M concentrations by a number of salt-tolerant chemoheterotrophic bacteria isolated from hypersaline soils [5]. Additionally, a number of phototrophic green and purple sulfur bacteria are able to take up and accumulate betaine but are incapable of its biosynthesis [13]. In nature, such situations can occur, whenever large decaying biomass is accumulated and betaine or choline becomes available, leaking from living cells or released upon cell lysis. Under such conditions, which are likely found in microbial mats,

uptake might become an important strategy to accumulate compatible solutes, although dependent on biosynthesis as their primary producers. Uptake systems for betaine, choline, and ectoine are found in numerous bacteria, many of which are unable to biosynthesize these molecules. Best known is a widely distributed transport system first identified in the transport of proline ProU (ProVWX), which has low affinity to choline and also transports glycine betaine, proline betaine, carnitine, and ectoine [26,27]. Another betaine transport system is OpuA. Both of these systems are ABC type transport systems and include binding proteins for the substrate and for ATP and a permease (ProVWX, OpuAA,AB,AC). Several more other transport systems for osmolytes are known (see [27]). Of these, OpuD is a single-component secondary transporter for betaine. BetT might also be active in betaine transport as found in *Aphanothece halophytica* [18].

In the present study, we focused on the biosynthetic capability of phototrophic bacteria to produce betaine and ectoine, using genomic information available from new genome sequences and from databases. The genomic repertoire and the distribution of the studied biosynthesis pathways are related to salt responses of the bacteria to conclude on their requirements for environmental adaptation. In addition, the phylogeny of ectoine and betaine biosynthesis was studied by comparing sequences of glycine–sarcosine–methyltransferase (GMT) and diaminobutyrate–pyruvate aminotransferase (EctB) of phototrophic bacteria, together with selected chemotrophic bacteria.

#### **2. Materials and Methods**

#### *2.1. Cultivation and DNA Extraction*

Cells were grown in the appropriate media, as described for the purple sulfur bacteria [28,29] and several groups of phototrophic nonsulfur purple bacteria [30]. Extraction of DNA was done as described earlier [31]. DNA from 2 mL of a freshly grown culture was extracted with the DNeasy® Blood&Tissue Kit, according to the manufacturer's instructions (QIAgen, Hilden, Germany), including the pretreatment for Gram-positive bacteria (consisting of enzymatic lysis buffer, proteinase K, and RNAse) and then dissolved in the TE-buffer. The extracted DNA was checked for quantity and quality by agarose gel electrophoresis with linear and double-stranded Lambda DNA used as control (Thermo Fisher Scientific, Waltham, MA, USA, Cat.No. SD0011). DNA was purified using a gel extraction procedure with the MoBio Ultra DNA Purification Kit (Cat.No. 12100-300).

#### *2.2. Sequencing and Assembly*

Sequencing of DNA and the assembly of sequences were done, as described earlier [31]. Samples were prepared with the Nextera® XT DNA Sample Preparation kit from Illumina, following the manufacturer's protocol. Afterwards, the samples were pooled and sequenced on the MiSeq using the MiSeq® Reagent Kit v3 600 cycles sequencing chemistry. The library was clustered to a density of approximately 1200 K/mm2.

Read quality filtering was performed with Trimmomatic v0.36 [32]. Reads were scanned for residues of Illumina Nextera XT adapters. Quality trimming was conducted with a 5-base pairs (bp) sliding window, trimming the read once the average Phred quality score within this window dropped below 30. Reads with a minimum length of 21 bp after quality trimming were retained. Single reads (i.e., reads with their mate deleted) were retained and included into downstream analysis. Reads were further checked for ambiguous base calls as well as for low complexity, employing the DUST algorithm [33]. They were filtered accordingly with an in-house R script in Microsoft R Open v3.3.2 (R Core Team 2016). Retained reads are referred to as 'filtered reads'. Filtered reads were preassembled with SPAdes v3.10.0 [34,35], using default k-mer lengths. Scaffolds ≥500 bp of this pre-assembly were subjected to extension and second-round scaffolding with SSPACE standard v3.0 [36].

#### *2.3. Genome Annotation and Submission to GenBank*

Genome sequences were annotated by the "Rapid Annotation using Subsystem Technology" (RAST) [37]. Sequences of EctB, GMT, and ProW were retrieved from the annotated genomes using the RAST and The SEED Viewer provided by this platform [38,39], which also offered the option to search with gene/protein sequences within the annotated genomes. In addition, standard protein BLAST of the NCBI database was used with EctB and GMT sequences to retrieve additional protein sequences. All genome sequences were deposited in the GenBank database (Supplementary Table S4). Accession numbers of gene and genome sequences, together with species and strain designations as well as the corresponding higher taxonomic ranks, are included in Supplementary Tables S1–S4.

#### *2.4. Phylogenetic Sequence Analyses*

For phylogenetic analysis, protein sequences of GMT, EctB, and ProW were aligned using ClustalX version 2.1 [40] and the trees were calculated by the neighbor-joining (NJ) method with correction for multiple substitutions, according to ClustalX [41]. NJ plot was used to draw the phylogenetic trees expressed in the Newick phylogenetic tree format [42]. The tree topologies were evaluated with bootstrap analyses, based on 1000 replicates and the values are indicated in the trees.

#### **3. Results and Discussion**

#### *3.1. Osmotic Adaptation in Phototrophic Bacteria*

According to genome analysis of representative phototrophic bacteria, the genetic repertoire and kind of possible responses to osmotic stress of these bacteria varies widely between the different groups and between freshwater, marine, and halophilic species. The genetic repertoire of all studied phototrophic bacteria is shown in Tables 1–3. While almost all marine and halophilic phototrophic bacteria can synthesize either betaine or ectoine or both, true freshwater bacteria lack the ability to synthesize betaine and ectoine. Often, they also lack the possibility of uptake of these osmolytes or their biosynthetic precursors, while marine and halophilic bacteria generally have this option. Obviously, biosynthesis or uptake of betaine more than that of ectoine is a prerequisite for their ability to thrive in marine and hypersaline habitats and to tolerate high salt concentrations.

**Table 1.** Genes and gene clusters of betaine and ectoine biosynthesis as well as relevant transport systems of phototrophic *Acidobacteria* (*Chloracidobacterium*), *Chlorobi*, *Chloroflexi*, *Cyanobacteria*, *Heliobacterium modesticaldum*, *Gemmatimonas phototrophica* and *Betaproteobacteria* are shown together with salt responses and their systematic affiliation a,b,c.




88

and H = moderate to extreme halophilic species (>6–25%/>15–>25%). As far as possible, the gene clusters are given and "/" denotes a separate locus of the genes in the genome. The genome accession numbers of GenBank are shown in the Supplementary Table S4. b Ectoine biosynthesis absent, c proW3, and ProW4 are absent. Color shades indicate different types of gene and gene associations: betaine synthesis from glycine (blue) and from choline (green), transport with betT (light-lila), opuA,C and D (shades of beige-brown), proU-W1 and W2

(shades of green); also marine (blue) and halophilic (rose-pink) growth response of the bacteria.

*Microorg* **2021**, *9*, 46






#### *Microorg* **2021**, *9*, 46

**Table 2.** *Cont.*


a see footnote a in Table 1; Color shades indicate different types of gene and gene associations, also marine (blue) and halophilic (rose) growth response of the bacteria. Color shades indicate different types of gene and gene associations: ectoine synthesis in rose, betaine synthesis from glycine in blue and from choline in green (in association with proU-W3 in pink; transport with betT (light-lila), opuA and D (shades of beige-brown), proU-W1 and proU-W2 shades of green, proU-W3 pink, proU-W4, yellow; also marine (blue) and halophilic (rose-pink) growth response of the bacteria.

#### *Microorg* **2021**, *9*, 46

**Table 2.** *Cont.*





**Table 3.** *Cont.*

 46

Ectothiorhodospiraceae*Ectothiorhodospira mobilis* DSM 237 M o

GMT-DMT-MAT /GMT-B

o betT/ betT-

opuAA,AB,AC o





a footnote as a in Table 1; b proW2, proW3, and proW4 are absent from Gammaproteobacteria of this study; Color shades indicate different types of gene and gene associations: ectoine synthesis (rose), betaine synthesis from glycine (blue) and from choline (green); transport with betT (light-lila), opuA and D (shades of beige-brown), proU-W1 (green); also marine (blue) and halophilic (rose-pink) growth response of the bacteria.

The ability to produce betaine or ectoine from the currently known routes is absent from the freshwater bacteria examined here, which include *Heliobacteria*, *Chloracidobacterium*, *Chloroflexi*, and the majority of *Cyanobacteria* and *Chlorobiaceae*, as well as phototrophic *Betaproteobacteria* (Table 1). Most of these bacteria also lack corresponding transport systems. Both pathways are also absent from the phylogenetic groups of freshwater phototrophic species of *Rhizobiales*, *Acetobacteraceae*, *Rhodospirillaceae*, *Betaproteobacteria*, and *Chromatiaceae* (Tables 1–3). In all freshwater bacteria that lack any of the options to accumulate betaine or ectoine, a limited osmotic adaptation might be achieved by accumulation of sugars, in particular trehalose and sucrose or glucosylglycerol, but also N-acetyl-glutaminylglutamine amide and N-carbamoyl-L-glutamine amide (Severin et al., 1992). Even potassium glutamate to some extent might contribute to osmotic adaptation [1,6,7].
