3.2.7. Pirellulaceae

The two methyltransferases of betaine biosynthesis (GMT, DMT) are also found in the chemotrophic *Rhodopirellula europaea* and *Rubripirellula obstinata* (*Pirellulaceae*). Phylogenetically, these GMT sequences form a distinct branch distantly associated with *Ectothiorhodospiraceae*, though with low confidence. In *Rhodopirellula europaea*, the *betT* gene forms a cluster with the two methyltransferases (*GMT-DMT-betT*).

#### *3.3. Phylogeny of EctB*

The ability for ectoine biosynthesis is found in several phylogenetic distant lineages of anoxygenic phototrophic bacteria. Two major phylogenetic branches (type-1 and type-2 EctB sequences) can be distinguished, which also show differences in the *ect* gene cluster structure (Figure 2).


**Table 4.** Ectoine and betaine biosynthesis of selected chemotrophic bacteria a.

a Gene clusters are given and "/" denotes a separate locus of the genes in the genome. Color shades indicate different types of gene and gene associations, also marine (blue) and halophilic (rose) growth response of the bacteria.

**Figure 2.** The phylogeny of ectoine biosynthesis on the basis of sequences of diaminobutyrate-pyruvate aminotransferase EctB is shown in a neighbor-joining tree. Sequences and gene bank accession numbers are shown in the Supplementary Table S2. Bootstrap values expressed as percentages of 1000 replications are given at the branches. The bar indicates an evolutionary distance of 0.05. The following color code highlights different systematic groups.

*Gammaproteobacteria* are found in both branches—representatives of *Ectothiorhodospiraceae* in the EctB type-1 branch and representatives of *Oceanospirillales*, *Cellvibrionales*, and *Vibrionales* in the EctB type-2 branch. According to genome analysis of the selected species and also BLAST search with representative EctB sequences, ectoine biosynthesis is absent from all tested *Cyanobacteria*, from *Heliobacteria*, *Chlorobi*, *Chloroflexi*, *Chloracidobacterium*, as well as from all studied phototrophic *Betaproteobacteria*, *Rhizobiales*, *Chromatiaceae*, and the *Ectothiorhodospira* species (Tables 1–3).

As shown in Figure 2, a first major phylogenetic branch (EctB type-1) contains four clearly separated groups: (1) the *Ectothiorhodospiraceae*, (2) *Rhodospirillaceae* including the non-phototrophic relatives, (3) *Bacillales*, and (4) *Actinobacteria*. This branch is characterized by a gene cluster lacking the regulatory *ectR* gene and the *ask\_ect* gene.


A second major phylogenetic branch (EctB type-2) is characterized by the presence of an extended gene cluster for ectoine biosynthesis, often including the regulatory gene (*ectR*) and a specific isoenzyme of aspartokinase (*ask\_ect*). The EctB type-2 group shows considerable deeper branching points and is phylogenetically even more diverse compared to the EctB type-1 group (Figure 2).

The *Alphaproteobacteria* form two related subbranches. One of these is represented by *Rhodobacterales*, with numerous anaerobic and aerobic phototrophic bacteria, as well as chemotrophic relatives that are unable to perform photosynthesis and has an extended *ask\_ect\_ectABCR* gene cluster. Species of *Rhodovulum* and the related genera *Roseivivax*, *Roseovarius*, *Rhodosalinus*, and *Roseisalinus* are included (Table 2, Figure 2). However, in the related *Rhodobacter* species, ectoine biosynthesis and the *ectABCR* genes are absent. One distinct separate lineage is formed by *Rhodobaca barguzinensis* and *Rhodobaculum claviforme*. The marine *Roseospirillum parvum* is the only representative of *Rhodospirillaceae* within the type-2 group. Its gene cluster includes the regulatory gene (*ectABCR*) but lacks the *ask\_ect* gene. However, this species and *Thioclava pacifica* (*Rhodobacteraceae*) are outsider of this subbranch.

A second subbranch of *Alphaproteobacteria* is distantly associated with the *Rhodobacterales* branch. It includes the aerobic phototrophic *Acidiphilium multivorum* (*Acetobacteraceae*, *Rhodospirillales*) and *Acidiphilium cryptum* (non-phototrophic), which also encode the formation of hydroxyectoine and have *ectD* included in the gene cluster *ectRABCDask\_ect* (Table 2). The related phototrophic *Acetobacteraceae Rhodopila globiformis* and *Paracrauro-* *coccus ruber* are unable to produce ectoine (Table 2). A second lineage of this subbranch contains *Nitrobacter winogradskyi* (*Rhizobiales*) and representatives of the *Sphingomonadales*.

Although none of the phototrophic *Betaproteobacteria* included in this study was able to synthesize ectoine, in several chemotrophic *Betaproteobacteria* (*Burkholderiales*) EctB is present and a distinct branch of type-2 EctB sequences is formed. Two separate lineages include representatives of *Alcaligenaceae* (*Achromobacter* and *Bordetella* species) in one, and the *Paucimonas* (*Burkholderiaceae*) and *Herminiimonas* (*Oxalobacteraceae*) species in another lineage (Figure 2). An *ectRABCD* gene cluster was found in the genomes of *Achromobacter xylosoxidans*, *Bordetella avium*, *Herminiimonas arsenicoxydans*, and *Paucimonas limoigeni* (Table 4).

Two branches of *Gammaproteobacteria* are found among the EctB type-2 sequences. One is represented by *Halomonadaceae* with *Halomonas elongata* and *Chromohalobacter salexigens* and related species (*ectABC*/*ectD*), and associates distantly with the sequences of *Betaproteobacteria*. The second one is poorly associated and is quite distant to the branch of *Alphaproteobacteria*. It includes representatives of *Oceanospirillaceae* (*Marinomonas* and *Amphritea* species) and deeply branching lines with representatives of other families (Table 4, Figure 2). There is a considerable variation of the composition of the *ect* gene cluster in this branch (*Marinomonas mediterranea*: *ectAB*C/*ectR*/*ask\_ect*; *Hahella chejuvensis*: *ectABCD*; *Haliea salexigens*: *ectRABCask\_ect*), as shown in Table 4.

#### *3.4. Betaine Synthesis from Choline—Distribution of Bet Genes in Phototrophic Bacteria*

An alternative and independent pathway of betaine synthesis that starts from choline exists in a number of bacteria [49]. It depends on an external source of choline that needs to be taken up by the cells and is then converted into betaine. Thus, the availability of choline is a crucial factor that eliminates this pathway from consideration, as an important and independent option to adapt to high salt concentrations. Though this pathway offers a good chance for bacteria living in eutrophic locations rich in biomass and provide choline as the source, such habitats with extremes of salt concentrations are almost devoid of higher developed eukaryotes that could produce and release choline. Here, mass developments of halophilic microorganisms that would be the primary colonizers could be a possible source under such conditions. If the presence of this pathway is compared with the ability to cope with even low (marine) salt concentrations, it is then obvious that it does not play a primary role in conquering marine and hypersaline habitats. Most marine species that can perform the choline pathway, are not dependent on this pathway, but have alternative options of compatible solute biosynthesis. Examples are *Rhodobaca barguzinensis* (produces also ectoine), *Rhodovibrio* species (also have the ectoine and glycine-dependent pathway), *Marichromatium gracile* (also has the glycine-dependent pathway), and *Ectothiorhodospira* species (have the glycine-dependent pathway) (see Tables 2 and 3). The only marine species that exclusively rely on the choline-dependent pathway of biosynthesis for betaine synthesis (in addition to several transport systems) are members of the *Rhizobiales* including *Fulvimarina pelagi*, *Hoeflea phototrophica*, *Rhodobium orientis*, and the *Afifella* species (Table 2). It is also found in some freshwater bacteria, such as *Rhodopseudomonas palustris*, *Rhodospirillum rubrum*, and *Rhodobacter* species.

Quite interestingly, the choline-dependent pathway of betaine synthesis is also an additional option in a few extremely halotolerant species, which are capable of betaine synthesis from glycine and of ectoine synthesis. These include a few *Rhodobacteraceae* (species of *Roseivivax*, *Rhodosalinus*, and *Roseovarius*), as well as the *Rhodovibrio* species (Table 2).

#### *3.5. Transport System for Uptake of Glycine Betaine and Choline*

All marine and halophilic phototrophic bacteria have one or more transport systems for betaine or choline. This underlines the importance of transport to gain or regain osmolytes for successful adaptation to elevated salt concentrations. In contrast, freshwater phototrophic bacteria not only lack possibilities of biosynthesis but also of common

transport systems for betaine, ectoine, and related osmolytes. While there are no obvious correlations between these transport systems and the biosynthetic pathways present, various phylogenetic groups have characteristic sets of transport systems for betaine, choline, and possibly ectoine and related osmolytes.

The most common and highly variable system is the ABC transporter ProU, which is present in different distinct variants and in different association with other genes. Often multiple forms are found in one and the same bacterium (Figure 3, Tables 1–3).

**Figure 3.** The phylogeny of the permease protein ProW of the L-proline/glycine betaine transport system ProU is shown in a neighbor-joining tree. The ProW1-ProW4 gene clusters are defined with Table 1. Sequences and gene bank accession numbers are shown in Supplementary Table S3. Bootstrap values expressed as percentages of 1000 replications are given at the branches. The bar indicates an evolutionary distance of 0.1. The following color code is used for the different ProW sequence types.

> According to the ProW sequences, at least four distinct types of this transporter can be distinguished (Figure 3). One separate *proVWX* gene cluster is related to the *E. coli proU* cluster (type ProW1). Another separate *proXWV* cluster represents a second phylogenetic group (type ProW2). Phylogenetically distinct is a third group (type ProW3), which has *proXWV* genes associated with a cluster of genes for choline-dependent betaine biosynthesis (*betABI-proXWV*), e.g., in *Rhodosalinus sediminis*. A fourth group with a particularly large ProW sequences (type ProW4) is clearly distant to the first three groups. Finally, a gene cluster including a YehZ betaine binding protein is especially found in some species restricted to freshwater habitats (data not shown). In fact, YehZYXW mediated activity in *E. coli* was inhibited at increased salinity and it was concluded that this transport system is not relevant to osmotic protection [50]. A branch of sequences of ProW of this type of transport system is distantly related to the four others and includes sequences of the *E. coli* YehW (data not shown). This was not considered in the phylogenetic tree of ProW sequences (Figure 3). In this tree of selected ProW sequences, the presence of several sequence types

is demonstrated in *Rhodobacter veldkampii* (types W2, W3, W4), *Rhodobaca barguzinensis* (types W2, W3), *Roseovarius halotolerans* (types W1, W3, W4), and others. Though sequence variations might sugges<sup>t</sup> possible differences in catalytic properties, gene regulation or substrate specificities between the 4 types of ProU transport systems, these are ye<sup>t</sup> to be demonstrated in future studies.

Common to most marine and halophilic phototrophic bacteria, no matter whether they synthesize betaine or ectoine, is the BetT secondary transporter, which is known as a specific choline transport system in *E. coli*, but is shown to specifically transport betaine in *Aphanothece halophytica* [18].

Another secondary transporter for glycine betaine, OpuD, is found in the freshwater Betaproteobacteria *Rhodoferax fermentans* and *Rubrivivax gelatinosus*, as only a transport system for betaine.Whereas, it is present in the marine *Caenispirillum salinarum* as well as in the extreme halophilic *Halorhodospira* species, both of which synthesize betaine and ectoine, (not in *Hlr. neutriphila*) together with BetT and ProU (type W1) (not in *Hlr. halochloris*). The only species of *Ectothiorhodospira* that encodes OpuD is *Ectothiorhodospira haloalkaliphila*.

The ABC transporter OpuA was found together with BetT in a small group of marine Chromatiaceae producing betaine (*Thiocapsa*, *Thiocystis*, *Marichromatium*, *Imhoffiella*, and *Thiorhodococcus* species), and also in all *Ectothiorhodospira* species. In all of these latter species, *betT* clusters with the *opuA* genes (*betT*-*opuAA,AB,AC*) and in a few Chromatiaceae, a copy of *betT* also specifically associates with the *opuAC* gene (Table 3). In a few cases, a chimera of the *proU* and *opuA* genes occur. In marine *Prosthecochloris vibrioformis*, this is *proVopuAB,AC*, in *Rhodovibrio salinarum* and also in two Chromatiaceae (*Rhabdochromatium marinum* and *Thiorhodovibrio winogradskyi*) this is *proVWopuAC*. Furthermore, a chimera between *opuC* and *proU* (*opuCB-proXV*) is annotated in the *Roseiflexus* species.

## *3.6. Evolutionary Considerations*

From an evolutionary point of view, the primary adaptation to high solute concentrations clearly requires a biosynthesis of compatible solutes. Already in the early prokaryote era the ability for biosynthesis and accumulation of organic solutes was a prerequisite for archaic eubacteria to conquer saline and hypersaline environments. Inorganic ions and common simple organic metabolites might have allowed a basic osmotic adaptation in freshwater and marine bacteria. Potassium very likely is a suitable candidate and is responsible for a kind of basic osmotic adaptation in many bacteria in which it accumulates together with glutamic acid, and thereby contributes to the overall osmotic balance [7]. In the extremely halophilic Archaea, the *Halobacteria*, it is even the primary osmolyte and accumulates to several molar concentrations [51,52]. In freshwater bacteria such as *E. coli*, its uptake or release makes possible rapid responses to small changes in osmotic conditions. For osmotic adaptation of marine and halophilic eubacteria, however, it is not important.

In marine eubacteria, a number of non-charged neutral organic molecules contribute to different degrees in achieving osmotic balance. As was shown for *Cyanobacteria*, the degree of salt tolerance depends on the kind of osmolytes that are accumulated. Only limited osmotic adaptation to lower ranges of salt concentrations is possible by carbon compounds such as sugars (sucrose, trehalose), glucosylglycerol, and others [1,6,7,11–13]. The advantage of accumulating these compounds is the comparable cheap biosynthesis and the absence of nitrogen as one of the most severe limiting elements in the environment. The disadvantage is their comparatively low compatibility or solubility. For example, trehalose has a solubility of approximately 2 M, ectoine of 3.87 M, and glycine betaine of 13.6 M. The extreme compatibility of betaine, which goes hand in hand with its excellent solubility, and the almost perfect ability to protect macromolecules against denaturation even at very high concentrations, make betaine the top compatible solute. The excellent osmotic protection of betaine is also reflected in the finding, that the phototrophic bacteria that have most successfully adapted to extremely high salt concentrations accumulate betaine as compatible solute through de novo biosynthesis from glycine. In fact, a striking dependence on betaine synthesis for adaption, to live in extreme and even moderate salt

concentrations as found in this study, points out that the kind of osmolytes accumulated in a critical way determine the success of salt adaptation. Therefore, the establishment of betaine biosynthesis is expected to represent an evolutionary breakthrough in osmotic adaptation of eubacteria to high salt concentrations, and is considered to be a prerequisite to conquer highly saline environments.

Nonetheless, the synthesis and accumulation of high concentrations of betaine is a costly process. Though glycine, the immediate biosynthetic precursor of betaine, is a common intermediate of central metabolic processes and can be transformed to betaine with a few enzymatic steps, the three methylation steps have a high energy demand. Additionally, the accumulation of several molar concentrations can be easily limited by the availability of nitrogen. Therefore, bacteria that have no shortage in energy and nitrogen supply, such as phototrophic bacteria, which use sunlight as an energy source and are capable of fixing dinitrogen, are likely to be the first and also to be the most successful in conquering highly saline habitats. In fact, mass developments of phototrophic bacteria regularly occur in salt and soda lakes, as well as in coastal lagoons, and cause colored blooms [1,7,53–55].

Compared to betaine, the accumulation of ectoine has two major and significant disadvantages. The physiological disadvantage of ectoine is its low solubility compared to betaine, which limits its accumulation and protective action at hypersaline salt concentrations. Its ecological disadvantage is the content of 2 nitrogen atoms in the molecule, which doubles the requirement for nitrogen compared to betaine. This restricts accumulation to environments with a high content of combined nitrogen compounds, unless nitrogen can be supplied by nitrogen fixation. In this case, the energy requirement for osmotic adaptation is further increased, because twice as much nitrogen has to be fixed, compared to betaine synthesis. Phylogenetically, it appears as a late event compared to betaine biosynthesis. The most deeply branching points within both type-1 and type-2 EctB sequences are those of the Gammaproteobacteria, forming one branch in the type-1 group and two branches in the type-2 group. Quite remarkably, also the structure of the *ect* gene cluster is much more variable in Gammaproteobacteria compared to other groups. In addition to the *ectABC* genes, either *ectD* or *ask\_ect* or *ectR*, or combinations thereof might be present, but might not be part of the gene cluster (Table 3). Therefore, it is assumed that the two types of EctB sequences have their origin in ancestors of the *Gammaproteobacteria*, which might also represent the most ancient ectoine producers. Both phylogenetic lineages might have separated early and given rise to the two independent major branches. Today, the two most prominent and best studied representatives of the two branches that have evolved are found in the Gammaproteobacteria: *Halorhodospira halochloris* (type-1) and *Halomonas elongata* (type-2). Among phototrophic bacteria, ectoine biosynthesis is restricted to a few distinct groups, with representatives of *Gammaproteobacteria* (*Halorhodospira* species) and *Alphaproteobacteria* (*Rhodobacterales* and *Rhodospirillales*).
