Cell sorting analysis of geographically separated hypersaline environments
- First Online:
- Cite this article as:
- Zhaxybayeva, O., Stepanauskas, R., Mohan, N.R. et al. Extremophiles (2013) 17: 265. doi:10.1007/s00792-013-0514-z
- 829 Downloads
Biogeography of microbial populations remains to be poorly understood, and a novel technique of single cell sorting promises a new level of resolution for microbial diversity studies. Using single cell sorting, we compared saturated NaCl brine environments (32–35 %) of the South Bay Salt Works in Chula Vista in California (USA) and Santa Pola saltern near Alicante (Spain). Although some overlap in community composition was detected, both samples were significantly different and included previously undiscovered 16S rRNA sequences. The community from Chula Vista saltern had a large bacterial fraction, which consisted of diverse Bacteroidetes and Proteobacteria. In contrast, Archaea dominated Santa Pola’s community and its bacterial fraction consisted of the previously known Salinibacter lineages. The recently reported group of halophilic Archaea, Nanohaloarchaea, was detected at both sites. We demonstrate that cell sorting is a useful technique for analysis of halophilic microbial communities, and is capable of identifying yet unknown or divergent lineages. Furthermore, we argue that observed differences in community composition reflect restricted dispersal between sites, a likely mechanism for diversification of halophilic microorganisms.
KeywordsHaloarchaea Cell sorting Genome amplification Biogeography Nanohaloarchaea Prokaryotic speciation
Solar salterns are industrial sea-salt manufacturing facilities found in coastal regions typically with hot and dry climates. Seawater is pumped into a series of shallow ponds, and increased salinity concentrations are achieved through evaporation. Salts, such as calcium carbonate and calcium sulfate, precipitate leaving mainly sodium and magnesium chlorides in the brine solution. When these industrial waters approach NaCl saturation (i.e., become ‘crystallizer ponds’), it creates a niche for specialized microbial communities of halophiles (Oren 1994).
Decades-long examinations of saturated NaCl brines, especially from solar salterns located in Spain, Australia, and Israel (e.g., Rodriguez-Valera et al. 1981; Oren et al. 1995; Burns et al. 2004), led to the conclusion that these microbial communities are ‘simple’. PCR amplification of 16S rRNA genes and visualization of communities by the fingerprinting method (denaturing gradient gel electrophoresis) showed that crystallizer ponds in Santa Pola (Spain) had limited microbial diversity compared to lower salinity ponds (Casamayor et al. 2002). Quantification of genotypes from the same saltern using florescence in situ hybridization analysis demonstrated that two species, the archaeon Haloquadratum walsbyi and the bacterium Salinibacter ruber, accounted for approximately 75 % of the total cells (Anton et al. 1999, 2002). Remaining archaeal representatives span the class Halobacteria (Anton et al. 1999) and the recently proposed class Nanohaloarchaea (Ghai et al. 2011; Narasingarao et al. 2011). Representatives from the Eukaryotes and Bacteria were also found in these systems, albeit at much lower frequencies: Among the most studied are photosynthetic protists (Oren 2005), fungi (Gunde-Cimerman et al. 2009), and the bacterial groups of Bacteroidetes/Chlorobi (Anton et al. 2002) and gamma-proteobacteria (de la Haba et al. 2011). Furthermore, hypersaline environments, and saturated brines specifically, have clear boundaries from their surrounding environments and are patchily distributed across the globe. Therefore, analogous to places like the Galapagos Islands, hypersaline settings can be viewed as microbial islands of limited biodiversity where restricted dispersal fosters the evolution of unique lineages and communities.
Studies of microbially dominated island-like environments provide evidence for limited dispersal. For example, thermophilic cyanobacterial mats from Zerka Ma’in hot springs in Jordan are dominated by the Synechococcus C1 types and lack Synechococcus A/B types (Ionescu et al. 2010), while in Yellowstone they are dominated by the A/B types (Ferris et al. 1996). Synechococcus spp. genotyping in hot springs from around the globe demonstrates the presence of distinct site-specific cyanobacterial communities that were likely assembled by chance (e.g., Papke et al. 2003; Hongmei et al. 2005; Finsinger et al. 2008; Lau et al. 2009; Ionescu et al. 2010).
The evidence for endemic microbial populations in isolated environments suggests that allopatric speciation might be common there. For example, distribution of Sulfolobus islandicus strains from acidic hot springs around the world was correlated with the geographic locations from where they were cultivated, indicating that geographic isolation was driving the divergence of these populations (Whitaker et al. 2003). However, a recent study (Oh et al. 2010) suggested that hypersaline environments might be different: an archaeon Haloquadratum walsbyi was found in Australia, Israel, Peru, Spain, Tunisia, and Turkey.
To improve our understanding of microbial diversity in hypersaline environments, we extracted genomic content of individual cells from two geographically separated brine communities, the South Bay Salt Works saltern in Chula Vista (California, USA) and Santa Pola saltern near Alicante (Spain). Individual cells were separated by high-throughput fluorescence-activated cell sorting and deposited into microtiter plate wells (Raghunathan et al. 2005; Zhang et al. 2006; Stepanauskas and Sieracki 2007; Swan et al. 2011). For each well, whole genome multiple displacement amplification (Dean et al. 2002) and 16S rRNA gene PCR and sequencing were performed. 16S rRNA amplification after cell sorting has been previously demonstrated as a useful technique for analyzing hypersaline environments (Trigui et al. 2011). Our additional step of whole genome amplification enables sequencing of multiple genes or genomes from individual, uncultured cells (Woyke et al. 2009; Swan et al. 2011). Indeed, the genome of a nanohaloarchaeaon identified in this study was successfully sequenced (Ghai et al. 2011). In this study, we present analysis of 16S rRNA genes from 207 individual cells, which reveal that two environments harbor significantly different communities of bacteria and archaea.
Materials and methods
Approximately 50 mL of saturated thalassohaline brines (32–35 % NaCl) were collected from the South Bay Salt Works in Chula Vista (CV) near San Diego, CA, USA (32°36′, 117°6′) and from the Santa Pola Saltern (SP), near Alicante, Spain (38°11′, 0°33′) in mid September 2009. Samples were shipped overnight directly to the Bigelow Laboratory in West Boothbay Harbor, Maine for the cell sorting, genome amplification and 16S rRNA sequencing.
Cell sorting and molecular analyses
Prokaryotic cell abundances were estimated at 108 cells per mL. Water sample was incubated for 10–60 min with SYTO-9 (5 μM final concentration; Invitrogen) and high nucleic acid content prokaryote cells were sorted with a MoFlo™ (Beckman Coulter) flow cytometer using a 488-nm argon laser for excitation, a 70-μm nozzle orifice, and a CyClone™ robotic arm for droplet deposition into microplates. High nucleic acid content was utilized because Halobacteria are known to be polyploid (Breuert et al. 2006) and theoretically would allow us to bias our analysis toward them. The cytometer was triggered on side scatter (see Supplemental Material Fig. S1). The “single 1 drop” mode was used for maximal sort purity, which insured the absence of non-target particles within the target cell drop and the adjacent drops. Under these sorting conditions, sorted drops contain a few tens of pL of sample surrounding the target cell (Sieracki et al. 2005), resulting in low or absent non-target DNA. The accuracy of 10 μm fluorescent bead deposition into the 384-well plates was verified by microscopically examining the presence of beads in the plate wells. Of the 2–3 plates examined on each sort day, <2 % wells were found not to contain a bead and only <0.5 % wells were found to contain more than one bead, indicating very high purity of single cells. In addition, we verified the lack of DNA contamination in the sheath fluid and in sheath fluid lines by performing real-time multiple displacement amplification with the processed sheath fluid as the template.
Single bacterial cells were deposited into 384-well plates containing 0.6 μL per well of TE buffer. Plates were stored at −80 °C until further processing. Of the 384 wells, 315 were dedicated for single cells, 66 were used as negative controls (no droplet deposition) and 3 received 10 cells each (positive controls). The cells were lysed and their DNA was denatured using cold KOH (Raghunathan et al. 2005). Genomic DNA from the lysed cells was amplified using multiple displacement amplification (MDA) (Dean et al. 2002; Raghunathan et al. 2005) in 10 μL final volume. The MDA reactions contained 2 U/uL Repliphi polymerase (Epicentre), 1× reaction buffer (Epicentre), 0.4 mM each dNTP (Epicentre), 2 mM DTT (Epicentre), 50 mM phosphorylated random hexamers (IDT) and 1 μM SYTO-9 (Invitrogen) (all final concentration). The MDA reactions were run at 30 °C for 12–16 h, and then inactivated by 15 min incubation at 65 °C. The amplified genomic DNA was stored at −80 °C until further processing. We refer to the MDA products originating from individual cells as single amplified genomes (SAGs).
The instruments and the reagents were decontaminated for DNA prior to sorting and MDA setup, as previously described (Stepanauskas and Sieracki 2007). High molecular weight DNA contaminants in all MDA reagents were cross linked by a UV treatment in a Stratalinker (Stratagene). An empirical optimization of the UV exposure was performed to remove all detectable contaminants without inactivating the reaction. Cell sorting and MDA setup were performed in a HEPA-filtered environment. As a quality control, all MDA reaction kinetics were monitored by measuring the SYTO-9 fluorescence using FLUOstar Omega (BMG). The critical point (Cp) was determined for each MDA reaction as the time required to produce half of the maximal fluorescence. The Cp is inversely correlated to the amount of DNA template (Zhang et al. 2006). The Cp values were significantly lower in 1-cell wells compared to 0-cell wells (p < 0.05; Wilcoxon Two-Sample Test) in each microplate. The MDA products were diluted 50-fold in sterile TE buffer. Then 0.5 μL aliquots of the dilute MDA products served as templates in 5 μL real-time PCR screens targeting the SSU rRNA gene using bacterial primers 27F (Page et al. 2004) and 907R (Casamayor et al. 2000), archaeal primers Arch_344 and Arch_915R (Casamayor et al. 2000) and prokaryote-wide primers Prok_340F and Prok_806R (Martinez-Garcia et al. 2011). Forward (5′-GTAAAACGACGGCCAGT-3′) or reverse (5′-CAGGAAACAGCTATGACC-3′) M13 sequencing primer was appended to the 5′ end of each PCR primer to aid direct sequencing of the PCR products. All PCRs were performed using LightCycler 480 SYBR Green I Master mix (Roche) in a LightCycler® 480 II real-time thermal cycler (Roche) following previously described cycling conditions (Martinez-Garcia et al. 2011). The real-time PCR kinetics and the amplicon melting curves served as proxies detecting successful SAG target gene amplification. New, 20 μL PCR reactions were set up for the PCR-positive SAGs and the amplicons were sequenced from both ends using M13 targets and Sanger technology by Beckman Coulter Genomics. Single cell sorting, whole genome amplification and PCR were performed at the Bigelow Laboratory Single Cell Genomics Center (http://www.bigelow.org/scgc). Previous studies and recent publications using this single cell sequencing technique demonstrate the reliability of the Center’s methodology with high purity of single cell MDA products (Woyke et al. 2009; Fleming et al. 2011; Heywood et al. 2011; Martinez-Garcia et al. 2011; Swan et al. 2011; Yoon et al. 2011). Cloned sequences were edited using Geneious bioinformatics and alignment software (http://www.geneious.com/). 16S rRNA sequences are available in GenBank under accession numbers JN839733-JN839939.
Ribosomal RNA classification
128 sequences from Chula Vista (CV) sample site and 79 16S rRNA sequences from Santa Pola (SP) sample site were placed taxonomically using Ribosomal Database Project’s (RDP) Naïve Bayesian Classifier (Wang et al. 2007) with 80 % confidence threshold and visualized as heat maps in RDP Taxomatic (http://rdp.cme.msu.edu/taxomatic). Sequences on heatmaps’ axes were arranged according to the RDP classification scheme. RDP database Release 10, Update 26 was used in these analyses (Cole et al. 2009).
16 archaeal sequences from a ‘high-intracellular DNA’ fraction of the solar saltern sample from Sfax, Tunisia (SFX) (Trigui et al. 2011) were downloaded from GenBank. 76 SP and 55 CV sequences classified by RDP classifier as archaeal, as well as the 16 SFX sequences, were uploaded to myRDP and aligned to the RDP rRNA alignment. These alignments were further used in community composition and phylogenetic analysis (see below).
Community composition analysis of archaeal and bacterial sequences
Sequences within SP, CV and SFX samples were grouped into operational taxonomic units (OTUs) in MOTHUR v.1.20.0 (Schloss et al. 2009) at distance cutoffs of 0.01, 0.03 and 0.05, using average neighbor distance method on uncorrected pairwise distances. Beta diversity between the three samples was calculated with these OTU assignments. Communities were compared using P test (Martin 2002), as implemented in the Unifrac program (Lozupone et al. 2006).
Phylogenetic analysis of archaeal sequences
Using RDP Sequence Match tool, similar sequences were added to the dataset of CV and SP Archaeal 16S rRNA sequences using the following criteria: for each CV and SP sequence, a maximum of two type strain matches, one isolate match and one uncultured match were extracted. The alignment of selected sequences was downloaded from RDP and a phylogenetic tree was reconstructed in the FastTree program version 2.1.3 (Price et al. 2009), with 100 bootstrap samples.
In a second sequence data set, all available matches to the putative archaeal class of Nanohaloarchaea, as well as one type strain representative per genus from Halobacteria class and one type strain representative per family in the rest of Euryarchaea, were retrieved from RDP. A crenarchaeote Aeropyrum pernix was added as an outgroup. The alignment of the selected sequences was downloaded from RDP and phylogenetic tree was reconstructed in RAxML version 7.0.4 (Stamatakis 2006) under GTR+Gamma model, with 100 bootstrap samples.
Archaeal class-level divergence analysis
For eight euryarchaeal classes represented in RDP database by at least some 16S rRNA isolate sequences ≥1200 nt long, pairwise Jukes–Cantor distances of all sequences within each class were downloaded from the RDP. Distances were also obtained for 39 sequences within a novel uncultured clade of Nanohaloarchaea, and for broader groups of ‘Halobacteria + Nanohaloarchaea’ and ‘Halobacteria + Methanomicrobia’. The class Methanomicrobia was chosen because, after Halobacteria, it was the euryarchaeal class best represented in the RDP.
Analyses of Bacteroidetes sequences
Using RDP Sequence Match tool, similar sequences were added to the dataset of 42 CV and SP 16S rRNA sequences classified as Bacteroidetes using the following criteria: for each CV and SP sequence a maximum of two type strain or isolate matches, and three uncultured matches were extracted. The alignment of selected sequences was downloaded from the RDP, and a phylogenetic tree was reconstructed with RAxML version 7.0.4 (Stamatakis 2006) under GTR+Gamma model, with 100 bootstrap samples.
Analyses of proteobacterial sequences
Using RDP Sequence Match tool, 30 CV and SP 16S rRNA sequences classified as proteobacteria were complemented by similar sequences using the following criteria: for each sequence, a maximum of one type strain match, one isolate match, and three uncultured matches were extracted. The dataset was enhanced by 44 proteobacterial 16S rRNA sequences previously observed in salterns by Maturrano et al. (2006) and Jiang et al. (2006). The alignment was downloaded from RDP, and a phylogenetic tree was reconstructed in the FastTree program version 2.1.3 (Price et al. 2009), with 100 bootstrap samples. Based on this initial phylogenetic reconstruction, the dataset was pruned from redundant sequences. The phylogenetic tree for the reduced dataset was reconstructed in RAxML version 7.0.4 (Stamatakis 2006) under GTR+Gamma model, with 100 bootstrap samples.
Distances within and between selected Euryarchaeal classes and the proposed novel Archaeal class of ‘Nanohaloarchaea’
Number of sequencesb
Distance observed within the group
39 (16 are from this study)
Halobacteria + ‘Nanohaloarchaea’
1101 (16 are from this study)
Halobacteria + Methanomicrobia
Halobacteria + ‘Nanohaloarchaea’
1101 (16 are from this study)
Halobacteria + Methanomicrobia
Notorious difficulty in microbial culturability (Rollins and Colwell 1986), bias in 16S rRNA primers (Suzuki and Giovannoni 1996; Lueders and Friedrich 2003) and limitations to assembling complete genomes from metagenomic sequence data (Rusch et al. 2007) makes the cell sorting technique followed by genome amplification an attractive approach to examine microbial diversity in halophilic environments. Our study focused only on the fraction of the community that displayed high nucleic acid content. Given a limited number of cells that could be collected and analyzed, we were concerned that without a sorting criterion, we will examine only the numerically dominant cells, and hence underestimate microbial diversity. High DNA content due to polyploidy was suggested to be common among Halobacteria (Breuert et al. 2006), and we conjectured that this sorting criterion would allow us to preferentially focus on a halobacterial fraction of the community. The screen was only partially successful in that respect, since a large bacterial fraction was recovered in one of the analyzed salterns (Fig. 1a). However, it demonstrates that high DNA content sorting may help highlighting less dominant (and therefore less studied) members of a halophilic community, such as yet unclassified bacteria and archaea. The cell sorting technique will allow future studies of gene content of these unclassified lineages (such as Nanohaloarchaea) without extra effort of gathering and assembling fragmented metagenomic sequence data.
Our observation of different prokaryotic communities in two geographically separated hypersaline environments is qualitatively in agreement with other studies of the same sites: while in SP Archaea comprised 87 % of the community (Ghai et al. 2011), in the CV, they constituted only 54 % (Rodriguez-Brito et al. 2010); in CV, Haloquadratum spp. were either co-dominant among halobacterial genera (Rodriguez-Brito et al. 2010), or not observed (Bidle et al. 2005), while they were described as dominant in SP (Anton et al. 1999; Ghai et al. 2011). Both CV and SP salterns are considered stable in their community composition (Rodriguez-Brito et al. 2010; Ghai et al. 2011). Quantitatively, however, our study produces different relative abundance of commonly found bacterial and archaeal genera, which may reflect differences in methodology (e.g., we screened for cells with high DNA content, while the above-mentioned studies did not) and in the extent of sampling across the studies.
Observed compositional difference in geographically separated hypersaline environments is discordant with the evidence for extensive wind-driven atmospheric dispersal of microorganisms (e.g., Kellogg and Griffin 2006). The immense sizes of prokaryotic populations combined with their small cellular dimensions are hypothesized to promote dispersal between similar environments, and to prevent endemism and localized extinctions (Finlay 2002; Fenchel 2003). Furthermore, genera including Haloquadratum, Halorubrum and Haloarcula are found worldwide in hypersaline systems with different ionic composition (e.g., Ward et al. 2000; Benlloch et al. 2001; Radax et al. 2001; Burns et al. 2004; Purdy et al. 2004; Bidle et al. 2005; Sorensen et al. 2005; Walsh et al. 2005; Papke et al. 2007; Mutlu et al. 2008; Oh et al. 2010; Dyall-Smith et al. 2011), suggesting unencumbered dispersal and niche invasion of halophiles.
How do we reconcile the seemingly contradictory observations that site composition is unique, yet similar halophilic OTUs are widely dispersed? To explain the disagreement, we hypothesize that the rate of evolution (i.e., rates of mutation, gene gain/loss, and recombination) is faster than the combined rate of dispersal and invasion. Three studies support our hypothesis (Whitaker et al. 2003, 2005; Dyall-Smith et al. 2011). Highly recombinogenic populations of Sulfolobus islandicus strains within geographically isolated hot springs are ~0.1 % divergent, yet strains between geographically distant sites accumulated larger amounts of divergence (~1 %) over short geological time periods (Whitaker et al. 2003, 2005). This indicates that population diversity is maintained by dispersal but reduced through recombination. Analogously, genomes of two closely related Haloquadratum walsbyi strains from Spain and Australia have variable gene content and are on average 1.4 % nucleotide sequence divergent across shared genes (Dyall-Smith et al. 2011). We suggest that these two strains have been geographically separated for a long time, since co-occurring Halobacteria are highly recombinogenic in both the laboratory (Naor et al. 2012) and nature (Boucher et al. 2004; Cuadros-Orellana et al. 2007; Rhodes et al. 2011; Andam et al. 2012) and can undergo genetic homogenization that keeps divergence below 1 % (Papke et al. 2007; Papke 2009). This hypothesis emphasizes that dispersal is an ongoing and frequent aspect of halophile biology, and that historical contingency (i.e., colonization order; Jousset et al. 2011; Langenheder and Szekely 2011) might be an important determinant of halophilic community composition.
This work was supported by grants to R.T.P from the National Science Foundation (award numbers 0919290 and 080024), the U.S.–Israel Binational Science Foundation (award number 2007043), and NASA Astrobiology: Exobiology and Evolutionary Biology Program (award number NNX12AD70G). We extend special thank you to Forest Rohwer (San Diego State University, USA) and Francisco Rodríguez-Valera (University Miguel Hernández, Spain) who sampled the Chula Vista and Santa Pola salterns, respectively.