Phylogenetically driven sequencing of extremely halophilic archaea reveals strategies for static and dynamic osmo-response

Abstract

Organisms across the tree of life use a variety of mechanisms to respond to stress-inducing fluctuations in osmotic conditions. Cellular response mechanisms and phenotypes associated with osmoadaptation also play important roles in bacterial virulence, human health, agricultural production and many other biological systems. To improve understanding of osmoadaptive strategies, we have generated 59 high-quality draft genomes for the haloarchaea (a euryarchaeal clade whose members thrive in hypersaline environments and routinely experience drastic changes in environmental salinity) and analyzed these new genomes in combination with those from 21 previously sequenced haloarchaeal isolates. We propose a generalized model for haloarchaeal management of cytoplasmic osmolarity in response to osmotic shifts, where potassium accumulation and sodium expulsion during osmotic upshock are accomplished via secondary transport using the proton gradient as an energy source, and potassium loss during downshock is via a combination of secondary transport and non-specific ion loss through mechanosensitive channels. We also propose new mechanisms for magnesium and chloride accumulation. We describe the expansion and differentiation of haloarchaeal general transcription factor families, including two novel expansions of the TATA-binding protein family, and discuss their potential for enabling rapid adaptation to environmental fluxes. We challenge a recent high-profile proposal regarding the evolutionary origins of the haloarchaea by showing that inclusion of additional genomes significantly reduces support for a proposed large-scale horizontal gene transfer into the ancestral haloarchaeon from the bacterial domain. The combination of broad (17 genera) and deep (≥5 species in four genera) sampling of a phenotypically unified clade has enabled us to uncover both highly conserved and specialized features of osmoadaptation. Finally, we demonstrate the broad utility of such datasets, for metagenomics, improvements to automated gene annotation and investigations of evolutionary processes.

Figure 1. Sequencing depth of newly sequenced haloarchaeal genomes.

Histogram showing distribution of mean per base coverage for the 59 newly-sequenced and seven previously published draft genomes included in this study.

Figure 2. Updated haloarchaeal phylogeny.

Multi-marker concatenated phylogeny of the 80 genomes included in this study and other haloarchaeal genomes gathered from IMG. The unrooted tree was built from a concatenated alignment of 40 PhyEco markers using PHYML. Branch support values for top-level branches have been removed for ease of visualization. The full tree file can be accessed through Dataset S19. Grey boxes represent haloarchaeal clades as described in , expanded as described in text. Tree roots in the Methanocella (not shown).

Figure 3. Rarefaction analysis of sampled haloarchaeal protein space.

A rarefaction curve of protein diversity was created for the 80 haloarchaea included in this study using three alternative methods to define protein families: COG number (green), TRIBE-MCL clusters removing singletons (red), and TRIBE-MCL clusters without removing singletons (blue). The COG database is expected to represent an under-estimate of the true diversity. Similarly, using TRIBE-MCL clusters with all singleton genes excluded underestimates true diversity. The true diversity of the set is likely located between the blue and red curves.

Figure 4. Haloarchaeal osmoadaptation.

(A) Generalized model for osmoadaptation in the haloarchaea. During osmotic upshock, potassium is imported through Trk H⁺/K⁺ symporters and Na⁺ is expelled using a combination of NhaC H⁺/Na⁺ antiporters and/or YrbG Ca²⁺/Na⁺ antiporters. During osmotic downshock, excess potassium is removed through a combination of Kef H⁺/K⁺ antiporters and the non-specific mechanosensitive channel MscS. (B) Ion transport strategies not generally encoded by the haloarchaea. Sodium export and potassium import using ABC transporters (NatABC and KdpABC) may be less energetically efficient than secondary transport systems. Use of Ktr K⁺/Na⁺ symporters for potassium uptake would result in over-accumulation of sodium. (C) Compatible solute transport systems. Na+/proline symporters (OpuE and MM_2305) are ubiquitous in the haloarchaea. Glycine betaine uptake is mediated by OpuD through symport with sodium or, rarely, through the ABC transporter OpuC. The ectoine/sodium symporter EctP is also widespread. For transport systems with multiple substrates, a representative compound is shown. (D) Magnesium uptake is mediated by primary active transport (MgtA/B) or, rarely, by facilitated diffusion (MgtE/CorA). (E) Potassium accumulation is possible in all 80 haloarchaea via both secondary active transport (Tribe63) and facilitated diffusion (Tribe105). Some species also possess ATP-dependent potassium transporters (Cog3639/PhnD). Numbers in parenthesis represent the number of haloarchaeal species possessing the transporter gene. BR – bacteriorhodopsin, HR – halorhodopsin, GB – glycine betaine, Pro - proline, Ect – ectoine. Kef, Ktr and Trk each represent a class of transporters, rather than a single homolog. Tribe63 and Tribe105 refer to protein families defined in this study.

Figure 5. Proteome-wide isoelectric point distributions.

Histograms of computed pI values for (A) Haloferax mediterranei, (B) Halorubrum litoreum, (C) Haloarcula argentinensis, and (D) Natrialba aegyptia. Each of these four species is a representative from the four most populous genera in the set, which collectively contain 66% of the organisms in the study. For all sequenced haloarchaea, the pI histograms exhibited a bimodal distribution with a major mode at about 4.5 and much smaller minor mode around 10. For all isoelectric point plots see Data Dryad package .

Figure 6. Ribosomal subunit isoelectric points.

Structural mappings of the large and small ribosomal subunits, showing protein monomers with predicted low (red) and high (blue) pI. Subunits are oriented according to (A) complex formation, (B) cytoplasmic view of each subunit, and (C) internal view of each subunit. Subunits exposed to the cytoplasm tend to be acidified (red subunits, visible in B), while subunits buried within the ribosome tend to have an alkaline isoelectric point (blue subunits, visible in C). Structural models used were those for Thermus thermophilus (1FKA) and Haloarcula marismortui (1QVG) .

Figure 7. Sliding-window G+C content analysis.

Representative G+C content plots for (A) Haloferax mediterranei, (B) Halorubrum litoreum, (C) Haloarcula argentinensis, and (D) Natrialba aegyptia. Each of these four species is a representative from the four most populous genera in the set, which collectively contain 66% of the organisms in the study. Black line represents calculated G+C percent for each 100 bp window. Contig boundaries are represented as vertical red lines, contig mean G+C as horizontal blue lines, and genome mean G+C as a horizontal green line. The horizontal axis displays the number of 20 bp steps taken along the genome. For all G+C plots see Data Dryad package .

Figure 8. Genera-level comparisons of genomic features.

Assembled genome size (Mbp) (A), number of predicted protein coding sequences (B), %G+C (C), and mean protein isoelectric point (D) were extracted from each genome, organized by genus, and boxplots calculated using MATLAB's Statistics toolbox. Boxplots were generated using 25^th and 75^th percentile as box edges, with median demarcated with horizontal line within box. Genera are ordered by descending number of species sequenced, with the number of species shown in parentheses. Genera with only a single sequenced member are shown as horizontal lines.

Figure 9. Haloarchaeal transcription factor B phylogeny.

Unrooted phylogenetic tree of haloarchaeal transcription factor B (TFB) homologs. Green - TfbC/G, magenta - TfbB/D/F, blue - TfbE, red - TfbA, black - unassigned. Number of sequences and species represented in each clade are shown in parenthesis. Bootstrap support values over 30% are shown for major clades. Branch marked with asterisk is truncated. See Dataset S5 for tree file.

Figure 10. Haloarchaeal TATA-binding protein phylogeny.

Unrooted phylogenetic tree of haloarchaeal TATA-binding protein (TBP) homologs. Dark green - TbpC, light green - TbpD/B/F, magenta - TbpE/A, salmon - TbpW, dark purple - TbpX, blue – TbpY, light purple – TbpZ, black - unassigned. Number of sequences and species represented in each clade is shown in parenthesis. Bootstrap support values over 30% are shown for major clades. See Dataset S7 for tree file.