sign in sign up
<<
Home , Bacteriophage, Microbial ecology

Microbial

Isolation and Genetic Analysis of Haloalkaliphilic in a North American Shereen Sabet, Weiping Chu and Sunny C. Jiang

Department of Environmental Health, Science, & Policy, University of California, 1367 SE II, Irvine, CA 92697-7070, USA

Received: 19 January 2006 / Accepted: 19 January 2006 / Online Publication: 6 May 2006

Abstract [7] and some of the highest recorded viral of any natural aquatic system [11], yet little is Mono Lake is a meromictic, hypersaline, soda lake that known about the viral that resides in this harbors a diverse and abundant microbial community. A extreme environment. previous report documented the high viral abundance in Bacteriophages, along with their hosts, make up the Mono Lake, and pulsed-field gel electrophoresis analysis largest on earth, residing mostly in aquatic of viral DNA from lake water samples showed a diverse . are believed to play a critical role in the population based on a broad range of viral sizes. mortality of aquatic bacteria [6], thereby affecting the To better understand the ecology of bacteriophages and microbial and biogeochemical processes, as their hosts in this unique environment, water samples well as affecting bacterial diversity by restructuring the were collected between February 2001 and July 2004 for microbial community [15, 30]. Bacteriophages can also of bacteriophages by using four indigenous affect their hosts via lysogeny and and by bacterial hosts. Plaque assay results showed a differential mutually providing that enhance host survival [6, seasonal expression of cultured bacteriophages. To reveal 13, 25]. the diversity of uncultured bacteriophages, viral DNA Viruses and viruslike particles have been detected from lake water samples was used to construct clone and isolated in hypersaline environments such as the libraries. Sequence analysis of viral clones revealed Dead Sea, a marine solar saltern in Australia, the Great homology to viral as well as bacterial . Further- Salt Lake in Utah, and the Little Salt in Yallahs, more, dot blot DNA hybridization analyses showed that Jamaica [16, 20, 23, 31]; however, the literature is sparse the uncultured viruses are more prevalent during most regarding the ecology of these extremophilic aquatic seasons, whereas the viral isolates (A7 and 72) were less bacteriophages. As a first step toward better understand- prevalent, confirming the belief that uncultured viruses ing these viruses, especially in an extreme environment, represent the dominant members of the community, this study aims to track the seasonal presence of Mono whereas cultured isolates represent the minority species. Lake’s representative community as well as uncultured phages. Fragments of viral from lake waters were cloned and analyzed, and their sequences Introduction provide insight as to the nature of the viruses that reside in a hypersaline, alkaline lake. Mono Lake is a closed basin, hypersaline, alkaline soda lake located in central California, USA, presently charac- terized by a salt concentration of 8% (80 g/L) and a pH Methods of 10. Mono Lake is a meromictic lake that undergoes Mono Lake Water Samples. Small-volume water holomixis approximately once every 10 years [19], with samples (50 mL) were collected monthly by researchers the most recent event recorded in October of 2003 (R. at the Sierra Nevada Aquatic Research Laboratory, Jellison, pers. comm.). The lake has a very productive University of California, from Mono Lake stations 2, 6, microbial community [9] with a high diversity of 7, and 11 (Fig. 1)[21] between March 2003 and July 2004. A single, upper 9-m integrated Correspondence to: Sunny C. Jiang; E-mail: [email protected] sample was taken from all of the stations, whereas a

DOI: 10.1007/s00248-006-9069-1 & Volume 51, 543–554 (2006) & * Springer Science+Business Media, Inc. 2006 543 544 S. SABET ET AL.: SEASONAL PRESENCE OF HYPERSALINE PHAGES

Figure 1. Map of Mono Lake showing locations of sampling stations. All stations were used for water sample collection for the duration of this study, except for stations 4, 5, and 9. detailed vertical profile at 2, 8, 10, 12, 16, 20, 24, 28, and sterilized by boiling. Plates were incubated for 24–72 h at 35 m was sampled at station 6 using a Niskin bottle room temperature and observed for colony growth. onboard a small boat. Integrated water samples were Individual colonies were picked and cultivated further in collected by lowering a 1-in.-diameter Tygon tubing their respective media. Frozen stocks were made and weighted at one end into the water down to the 9-m stored at _80-C. M12-2c and M12-chla hosts were depth and then lifting the weighted end out of the water identified by 16S rRNA sequence analysis in a previous and transferring the water to a bucket. Water samples study (Hollibaugh, pers. comm.) and are listed under were filtered through 0.22-mm pore-size syringe filters GenBank accession numbers AY730244 and AY730242, (Millex GP PES membrane; Millipore, Carrightwohill, respectively. M12-2c is 99% similar to Vibrio metschniko- Co. Cork, Ireland) to remove bacteria and , vii and M12-chla only has a close match with a bac- and viral fractions were shipped cold overnight to the terium previously isolated from Mono Lake sediment in University of California Irvine laboratory. the database. In this study, the MN12-2a and MN1-12.5a Large-volume water samples (5–10 L) were collected bacterial isolates were also characterized by 16S rRNA at sampling stations 1, 3, 6, 8, 10, 11, and 12 from both sequence analysis. Those sequences are listed under the oxygenic (2–10 m) and anoxic (35 m) layers during the accession numbers AY856383 and AY856384, field trips in February and August 2001, in August 2002, respectively. and in February and October 2003 for isolation of To isolate bacteriophages, 5–10 L of water samples bacterial hosts and viruses. All samples were processed were prefiltered through glass fiber and 0.45-mm filters on site within 4 h of collection. (Millipore) to remove large when plankton concentrations were high. The samples were then Bacterial Host and Viral Isolation. Four Mono sequentially filtered via TFF, first by using a 0.22-mm Lake bacterial hosts were isolated in February (designated pore-size filtration cartridge followed by a 30-kDa M12-2c, M12-chla) and in August (designated MN1- molecular weight cutoff cartridge. The first cartridge 12.5a, MN12-2a) of 2001 by first concentrating 10 L of removes bacteria, whereas the second cartridge concen- water using a tangential flow filtration (TFF) system with a trates viruses in the retentate. The final retentate volume 30-kDa molecular weight cutoff filtration cartridge (PALL is approximately 100 mL. Corp., Hauppauge, NY, USA). One hundred microliters Plaque assays were carried out by using the top agar of TFF retentate was then spread onto either marine agar overlay method [1]. Briefly, 1 mL of bacterial host plates (Difco, Becton, Dickinson and Co., Sparks, MD, (described above) in 3 mL of top agar was inoculated USA) supplemented with 1% NaCl (M12-2c and M12- with 1 mL of TFF-concentrated viral sample, then mixed chla), or onto agar plates made with Mono Lake water and poured over a bottom plate. Both top and supplemented with 0.5% peptone, which was then bottom agar contain the same ingredients as the bacterial S. SABET ET AL.: SEASONAL PRESENCE OF HYPERSALINE PHAGES 545 host isolation and growth media. Plates were incubated cut from the gel and stained with SYBR Gold (Molecular at room temperature for 48 h before the number of viral Probes) to visualize the DNA bands. The rest of the gel plaques was counted. Selective individual plaques were was not stained in order to protect the viral DNA from picked and suspended in 300 mL of either 0.5 M Tris (pH UV damage and to increase cloning efficiency. The 8.0) or MSM buffer (450 mM NaCl/50 mM MgSO4/50 stained lane was then used as a guide to cut out gel mM Tris, pH 8.0). Each isolate was purified by slices containing the 35- to 55-kb desired band from the reisolation of individual plaques four more times before remainder of the unstained gel. DNA was then purified it was considered to be a pure isolate. Viral lysates were from the gel by the GELase product (Epicentre Technol- harvested by plate elution using MSM buffer. Lysates ogies, Madison, WI, USA) according to the manufactur- were then filtered through 0.22-mm pore-size syringe fil- er’s protocol. The DNA and digested agarose were then ters to remove any remaining bacteria. Viral filtrates were transferred to Centricon 100-kDa filter tubes (Millipore) either frozen at _80-C in a 1:1 ratio of MSM buffer and and washed two times with 1 TE (1000 g, 15 min). 50% sterile glycerol or stored at 4-C for further analyses. After gel extraction, DNA concentration was again measured via the fluorometer as described above. Bacterial Artificial Chromosome Cloning. To Viral DNA was ligated into the pCC1 copy control understand the diversity and dynamics of uncultured BAC vector and cells were transformed according to the bacteriophages in Mono Lake, a bacterial artificial manufacturer’s protocol (Epicentre). Electrocompetent chromosome (BAC) clone library was constructed. First, transformations were carried out with the Eppendorf viruses were concentrated from 240 mL of 0.22-mm- Electroporator model 2510. Clones were screened for filtered lake water, collected from station 3 at 2 m in size and similarity by using BglII, HindIII, and NotI February 2003, to a final volume of 2 mL via the restriction (Promega). Four unique clones with Centricon Plus-80 100,000 NMWL centrifugal filter insert sizes ranging between 2.6 and 3.85 kb were (Millipore). The concentrate was then split into two identified. equal volumes and each volume was further concentrated and washed with 0.5 M Tris–HCl (pH 8.0) via the Sequencing and Analysis of BAC clones. The Centricon YM 100-kDa filter unit (Millipore). To remove following BAC vector oligonucleotides were synthesized any potentially contaminating extraviral DNA, DNase (Sigma Genosys, The Woodlands, TX, USA) and used treatment was carried out in each Centricon filter tube by for initial sequencing of the BAC clones: pCC1 forward adding 7 mL 0.5 M Tris–HCl (pH 8.0), 1 mL10 sequencing primer: 50-GGATGTGCTGCAAGGCGA reaction buffer, 2 mL(2U)RQ1DNaseenzyme TTAAGTTGG-30; pCC1; Reverse sequencing primer: 50- (Promega,Madison,WI,USA).Twohundred CTCGTATGTTGTGTGGAATTGTGAGC-30.Primer nanograms of l DNA and 150 ng of station 12 15-m walking was undertaken to complete the remaining (oxycline) sample were used as controls for the DNase length of the inserts. Sequencing was carried out with treatment. The mixture was incubated in a 37-C the ABI PRISM BigDye version 3.0 sequencing chemistry incubator for at least 30 min. Following DNase (Applied Biosystems, Foster City, CA, USA) by either digestion, viral samples were washed with 1 TE twice, slab gel (ABI Prism 377) or capillary (ABI Prism 3100) incubated with 5 TE for 1 h at 4-C, followed by three electrophoresis. BAC clone sequences were analyzed and more washes with 1 TE before heating at 65-C for contigs were assembled via the EditSeq and SeqMan 10 min to denature the proteins and DNase. programs (DNASTAR, Inc.; Madison, WI, USA). Open Viral DNA concentration was measured via the Turner reading frames (ORFs) were detected by using ATG, Quantech Digital Filter Fluorometer (Barnstead Inter- TTG, CTG, and GTG as start codons; TGA, TAA, national, Dubuqu, IA, USA) using the PicoGreen dsDNA and TAG as stop codons; and 34 codons (approximately Quantitation kit (Molecular Probes, Eugene, OR, USA) 100 bases) as a minimum ORF size. Homology searches according to the manufacturer’s instructions. were conducted against the nonredundant (nr) NCBI Five hundred nanograms of viral DNA sample was GenBank database and by using the blastx program; loaded into each of eight lanes of a 1% low-melting- significant homology was judged by using an E value of point agarose gel (Ultrapure L.M.P. Agarose, Invitrogen, less than 0.001, except in the case of clone 31. All BAC Carlsbad, CA, USA). A 5-kb DNA marker was loaded clone sequences were submitted to the NCBI GenBank into each of the remaining two lanes on each end of the database. gel. The DNA sample was resolved via pulsed-field gel electrophoresis (PFGE) in the CHEF-DR II model (Bio- Viral Probe Synthesis. Nonradioactive probes Rad, Hercules, CA, USA) with the following parameters: were made up of two viral isolates, A7 and 72, and 6 V/cm; 18 h; 1.9–7.1 second switch; running buffer final two BAC clones, clone 5 and clone 76. Both phages were temperature 14-C. One lane containing the viral sample isolated in the summer of 2002 by using bacterial host as well as one lane containing the 5-kb DNA marker were M12-2C and MN1-12.5, respectively. The random 546 S. SABET ET AL.: SEASONAL PRESENCE OF HYPERSALINE PHAGES primed labeling of viral genome DNA or BAC clone Prehybridization was carried out for at least 30 min, DNA was performed with the DIG High Prime and hybridization was done overnight in a glass DNA Labeling and Detection Starter Kit II (Roche, hybridization tube (Robbins Scientific-SciGene, Sunny- Penzberg, Germany) following the manufacturer’s vale, CA, USA) at 43-C in a roller hybridization oven protocol. Briefly, DNA was digested with the AluI (Hybaid Mini 10; National Labnet Co., Woodbridge, NJ, (Promega), denatured, and incubated USA). The membrane was washed with 0.5 SSC/0.1% with DIG-High Prime mix overnight at 37-C. Efficiency SDS at 70-C. All other washes and the immunological spot testing for each probe and immunological detection detection steps were done at room temperature on a were carried out according to the manufacturer’s rocker following the manufacturer’s instructions protocol (Roche). The sensitivity of each probe was (Roche). Membranes were exposed to Kodak Biomax measured to be at 1 pg of DNA, and all probes were used MS film (New Haven, CT, USA) either for 75 min or at the recommended concentration of 25 ng/mL. Probes overnight at room temperature before film development. were stored at _20-C, thawed at 42-C, and denatured at Membranes were stripped for reprobing according to the 68-C for 10 min before each use. manufacturer’s protocol (Roche).

Dot Blot Sample Preparation and Hybridization. Vi- Statistical Analysis. ANOVA was performed using ruses used in the hybridization assay were concentrated seasonal data of cultured bacteriophages on each host. Box by ultracentrifuging 10–20 mL of 0.22-mm-filtered water and whisker graphs were plotted to represent the seasonal sample at 41,000 rpm with a Beckman SW41 Ti rotor and between-host differences of culturable bacteriophages. for 1.5 h at 4-C. Most of the supernatant was discarded, All statistical analysis and graphs were created with SPSS leaving the viral pellets in õ200 mL of the original water. statistical software (SPSS Inc., Chicago, IL). Then 1 mL 5 TE was added and the sample was in- cubated at 4-C for 1 h before being transferred to GenBank Accession Numbers. For bacterial hosts Centricon YM 100-kDa MWCO centrifugal filter units MN12-2a and MN1-12.5a, the accession numbers are (Millipore Corporation; Billerica, MA, USA) and cen- AY856383 and AY856384, respectively. Accession num- trifuged at 1000 g room temperature for 20 min. bers for each BAC clone is as follows: clone 5, AY853713; The viruses were then washed two times with 100 mL5 clone 31, AY853715; clone 76, AY853716; and clone 122, TE. After the second wash, 25 mL1 TE was added to AY853714. each filter unit and an invert spin was done at 500 g for 1 min to collect the viruses. Viral samples were heated Results at 65-C for 10 min to denature the capsid proteins and release the DNA. DNA concentration was measured via Mono Lake Bacteria and Phage Isolates. 16S rRNA the fluorometer as described previously. sequence homology of the MN12-2a and MN1- A7 and 72 DNA was extracted by using the Wizard 12.5a bacterial hosts indicated that the MN12-2a Lambda Preps DNA Purification System (Promega) and host shows 98.5% homology to a Halomonas boliviensis used as controls to optimize the hybridization conditions species [24] (GenBank accession number AY245449), and stringency. Genomic DNA from the two bacterial whereas the MN1-12.5a host is 93% homologous to a hosts, M12-2c and MN1-12.5a, as well as from Aero- Marinospirillum alkaliphilum strain Z4, which was monas hydrophila (ATCC strain 7966) and E. coli was isolated from a soda lake in Inner Mongolia, China extracted [2] and used as controls. All control DNA [36] (GenBank accession number AF275713). concentrations were quantified spectrophotometrically To determine the seasonal dynamics of culturable with an Eppendorf BioPhotometer (Eppendorf AG, bacteriophages in Mono Lake, the four bacterial hosts— Hamburg, Germany). M12-2c, M12-chla, MN12-2a, and MN1-12.5a—were The DNA samples were blotted onto a positively used to isolate phages from the upper 10 m of the water charged nylon membrane (Magna Charge Nylon 0.45 column (the oxygenic layer) collected during different mm, GE Osmonics, Inc., Minnetonka, MN, USA) as seasons. Figure 2 shows box and whisker plots of seasonal previously described [11] by use of the Minifold I, phage counts for all four hosts. Phages infecting the Microsample Filtration Manifold apparatus (Schleicher M12-2c host had the lowest counts in the summer & Schuell, Keene, NH, USA). Controls were titrated onto (August 2001), with a slight increase in winter (February the blots with 1, 10, 100, and 500 ng of DNA. After 2003), and the highest counts in autumn (October 2003) blotting, the nylon membrane was then rinsed for 5 min (Fig. 2A), although overall there is no significant in 2 standard saline citrate (SSC), the DNA was fixed difference among all three seasons according to ANOVA by UV cross-linking (Spectrolinker XL-1000 UV cross- test (P = 0.119). Interestingly, the M12-chla phages seem linker; Spectronics Corp., Westbury, NY, USA), and the to be prevalent strictly during the winter season, as no membrane was stored at _20-C until further use. phage could be cultured in either the summer or autumn S. SABET ET AL.: SEASONAL PRESENCE OF HYPERSALINE PHAGES 547

months (Fig. 2B). In keeping with this trend, the Halomonas-like phages were far more abundant during February, whereas lower counts were observed during August and October (Fig. 2C). The ANOVA test yielded a P value of 0.003. However, the opposite result was observed for the MN1-12.5a host, which yielded an abundant number of phages in the summer season but fewer in the subsequent winter and autumn months (Fig. 2D). These data illustrate a differential seasonal pattern of phage prevalence infecting different species of Mono Lake haloalkaliphiles, suggesting that the phage community is dynamic and in a state of flux. To determine the vertical distribution of bacterio- phages infecting specific bacterial hosts, water samples collected from stations 3 and 6 from the anoxic layer (35 m) during October 2003 were also used for phage plaque assays (Table 1). Phages infecting the M12-2c and MN1-12.5a hosts were isolated from this depth, but no phages of the other two hosts could be found.

DNA Sequence Analysis of Cloned Viral DNA Fragments To characterize the uncultured Mono Lake phage population, we attempted to clone whole viral DNA genomes via BAC vector cloning technology. Sample preparation entailed the removal of any possible con- taminating DNA by carrying out a DNase digest of the whole viral sample before capsid denaturation. Figure 3 shows that DNase digestion was efficient at removing extraviral DNA (lane 1 vs untreated lane 2), whereas it had little or no effect on DNA encapsulated by the viral coat (lanes 3 and 4). Therefore, the DNase treatment ensured that the DNA sample that was subsequently purified from PFGE for cloning was indeed viral in origin. Figure 4A shows the PFGE of the viral community DNA collected from station 3, 2 m during February 2002. The 35- to 55-kb region (Fig. 4A) was targeted as the desired cloning region of interest because a relatively large amount of viral DNA was present there, which would favor cloning efficiency. The purity and size of the 35- to 55-kb DNA sample retrieved after gel purification is shown in Fig. 4B. After ligation and transformation, approximately 200 clones were screened and five clones were found to contain an insert (Fig. 4C) with insert sizes

Figure 2. Box and whisker plots of seasonal plaque counts on four Mono Lake bacterial hosts: (A) M12-2c, (B) M12-chla, (C) MN12- 2a, (D) MN1-12.5a. Outlier is indicated by the name of the station where water sample is assayed. Surface water was sampled from stations 1, 3, 6, 8, 10, 11, and 12 during three different seasons (*summer, from 2- to 10-m depths; **winter and autumn, from 2-m depth). 548 S. SABET ET AL.: SEASONAL PRESENCE OF HYPERSALINE PHAGES

Table 1. Plaque counts on four Mono Lake bacterial hosts using water samples collected from the anoxic layer (35 m) of stations 3 and 6 in October 2003 Plaque-forming units (pfu/L)

Bacterial host

Station M12-2c M12-chla MN12-2a MN1-12.5a 3 132 0 0 176 6 110 0 0 264 Average (pfu/L) 121 0 0 220 SD 15 0 0 62 ranging between 3 and 5 kb. These inserts were Clone 76 contained 14 ORFs between 123 and 654 significantly smaller than the expected whole viral bp, with an average length of 251 bp and a median size of genome size estimated by the PFGE DNA size markers. 216 bp. The highest scoring match was to a hypothetical Restriction analysis identified four unique viral clones, protein of Rhodobacter sphaeroides (Table 2). Other hits designated as clones 5, 31, 76, and 122. included sequence homology to a Mesorhizobium species, Selective sequence homology analysis results of the as well as to the Pas6 protein of 7Asp2, a that inserts are shown in Table 2. The most interesting insert infects the eubacterial Actinoplanes genus [8]. Another was clone 5, which contained the most number of ORFs (34) ranging in size from 133 to 684 bp. The average ORF size was 272 bp, whereas the median ORF size was 226 bp. Of the four clones, clone 5 ORFs showed the A B highest homology to phage-related sequences based on st 3 M M 1 the E value and score number (Table 2). The most 200 kb relevant and highest hits, shown in Table 2, were to 150 kb 50 kb several different species of single-strand DNA-binding 100 kb 35 kb protein from Chromobacterium violaceum, Nitrosomonas europaea, pyogenes phage 315.4, temperate 50 kb phage PhiNIH1.1, and LambdaSa1 phage. 35 kb 10 kb also showed significant hits. The rest of the ORFs 15 kb either returned no homology results from the database or yielded very low scores with genes of human or mouse C Bgl II Hind III Not I origin (data not shown). 1 2 3 4 5 1 2 3 4 5 1 2 3 4 5 M kb

21 M 1 2 3 4 M kb 5-3.5

2-1.9 200 1.6-1.4 150 0.9-0.8 100 0.5

50 Figure 4. Gel purification and subsequent cloning of DNA extracted from the station 3 2-m Mono Lake water sample. (A) Whole community viral DNA was resolved on a 1% PFGE low- 10 melting-point agarose gel. Arrows and brackets highlight bands of interest; the 35- to 55-kb region was cut out; lane M contains the 5 5-kb DNA ladder. (B) Relative purity of the 35- to 55-kb band of Figure 3. DNase treatment of phage l control DNA and of a interest in lane 1 after gel purification; lane M contains the 5-kb Mono Lake water sample before gel purification. Lane 1 contained DNA ladder. (C) Restriction digest analysis of five clones 200 ng of l DNA digested with 1 U DNase. Lane 2 contained l containing different-sized inserts. The enzyme used in each DNA minus DNase. Station 12 15-m DNA sample before (lane 3) reaction is noted above the gel image. Lanes 1, clone 31; lanes 2, and after (lane 4) DNase treatment. Lane M contains 5-kb DNA clone 5; lanes 3, clone 75; lanes 4, clone 76; lanes 5, clone 122; lane ladder mixed with 50-kb l marker. M = l+EcoRI+HindIII DNA marker. .S S. Table 2. Sequence homology analysis of the cloned Mono Lake viral DNA fragments BTE AL ET ABET No. ORFs % (larger than ORF hits Identity/

Clone no. Insert size (kb) 100 bp) [accession no.] E value Score Similarity Comments S .:

_ EASONAL 5 3.854 34 Single-strand DNA-binding protein 3.00E 16 87 39.1/54.7 ATG start; (C. violaceum ATCC 12472) ORF 10

[AAQ59563.1] P _ OF RESENCE Single-strand binding protein family 7.00E 11 69 33.1/50.7 ATG start; (N. europaea ATCC 19718) ORF 10 [NP_842444.1] _ Putative single-strand DNA-binding protein 2.00E 10 67 31.1/51.1 ATG start; H phage associated (S. pyogenes phage 315.4) ORF 10 YPERSALINE [NP_665053.1] _ Single-strand binding protein 2.00E 10 67 31.1/51.1 ATG start;

(temperate phage PhiNIH1.1) ORF 10 P [AAL15055.1] HAGES _ Prophage LambdaSa1, single-strand binding 2.00E 10 67 33.6/53.3 ATG start; protein (Streptococcus agalactiae 2603V/R) ORF 10 [NP_687595.1] _ Prophage LambdaSo, , putative 5.00E 10 65 38/46.7 ATG start; (Shewanella oneidensis MR-1) ORF 1 [NP_718538.1] _ Phage holin, putative 4.00E 07 55 37/53.1 ATG start; ( putida KT2440) ORF 1 [NP_746013.1] _ Hypothetical protein BQ11480 2.00E 05 50 43.1/50.8 GTG start; (Bartonella quintana strain Toulouse) ORF 6 [YP_032690.1] 31 3.318 8 Hypothetical protein Aaphi23p07 0.059 42 22.8/40.8 Not an ORF (bacteriophage Aaphi23) hit; entire [NP_852729.1] 3.3-kbp insert used in search Putative NTP-binding protein 1.5 33 42.9/66.7 ATG start; (bacteriophage phi AT3) ORF 1 [YP_025055.1] _ 76 2.604 14 Hypothetical protein Rsph020849 2.00E 20 101 47.1/67.3 ATG start; (R. sphaeroides) [ZP_00004944.1] ORF 1 _ Hypothetical protein MBNC02002420 7.00E 12 71 36.4/60.2 CTG start; (Mesorhizobium sp. BNC1) ORF 5 [ZP_00197163.1] _ Pas6 (Actinoplanes phage phiAsp2) 5.00E 05 48 42.0/53.1 CTG start; [AAT36754.1] ORF 5 Capsid protein gpC (virus PhiCh1) 0.001 43 36.4/48.9 CTG start; [AAM88682.1] ORF 5 Only ORFs with close match in the database are presented. 549 550 S. SABET ET AL.: SEASONAL PRESENCE OF HYPERSALINE PHAGES similar match was to the capsid protein of the haloalka- identical hybridization pattern and intensity (Fig. 5B, liphilic phage, 7Ch1. Again, as for clone 5, the majority clone 5 results; clone 76 data not shown) indicating that of ORFs had no match in the database. both clone 5- and clone 76-like phages were present in Clone 31 contained eight ORFs, ranging in size from the lake during all seasons sampled. However, there was a 108 to 153 bp, almost all of which did not show any difference overall in the intensity of hybridization among matches with the current database. Although none of the the months. Clone 5 and clone 76 probes showed the hits met the search parameters, one of the ORFs was strongest hybridization to the February and March found to be homologous to a putative NTP-binding samples (Figs. 5B, C), followed by the April samples protein from 7AT3, a temperate phage of Lactobacillus (Fig. 5B). The August samples showed relatively weak casei (Table 2). In addition, when the entire 3.3-kb insert hybridization overall, and the November samples showed was used in the search, the highest scoring homology for the weakest hybridization (Fig. 5B). These results suggest clone 31 (E value=0.059) was to a hypothetical protein of that seasonal variability of clone 5- and clone 76-like Aa723, a temperate phage of Actinobacillus actinomyce- phages exists at the same stations throughout the year, temcomitans. Clone 122 did not share any relevant homology with the database and is not included in the

ila final analysis and Table 2. h p o dr A y h a s Seasonal Dynamics of Cultured and Uncultured .5 a 6 2 ) 7 5 1 c k on e e - -2 n Bacteriophages by Dot Blot DNA Hybridization n n 1 2 a o o φ 1 l rom l l 2 (b e Analysis. To determine if there is any seasonal C C A φ MN M A variability of the Mono Lake phage community, dot Clone 5 probe blot hybridization was carried out by using genomic probes made from the cloned viral fragments, clones 5 Clone 76 probe and 76, as well as the phage isolates, A7 and 72. Figure 5A φ shows the hybridization results of each probe with A probe control DNA. Clones 5 and 76 hybridized to each other φ2 probe as well as weakly to the genome of the M12-2c host, but they did not hybridize to either of the viral isolates, to the MN1-12.5a host, or to the negative control genomic B DNA from A. hydrophila. The hybridization to the M12- st2 st6 st7 st11 2c host was unexpected, which implies that this bacterial August 2003 host shares some sequence homology to the viral fragments or to the clone vector. However, a control November 2003 probe made of empty vector alone did not hybridize with March 2004 any of the control DNA including the M12-2c host (data not shown) suggesting that homology exists between the April 2004 cloned viral fragments and the bacterial chromosomal DNA. Whether this is due to lysogeny or transduction is C not known. The cross hybridization between clones 5 and 76 could be the result of common vector sequence within 2m 8m 12m 16m 20m 24m 28m 35m the clone because there was no sequence homology between clone 5 and 76 viral inserts as determined by sequence analysis (see results in Table 2). Figure 5. Temporal and spatial distribution of specific bacteri- The phage isolates also hybridized to each other, and ophages among the Mono Lake viral community revealed by dot the A7 to the MN1-12.5a host, again unexpected results blot DNA hybridization. (A) Evaluation of the stringency and implying sequence homology; but they did not hybridize specificity of each probe used in the hybridization study using to the cloned viral fragments (Fig. 5A). The cross- control DNA extracted from bacteria, phage, and cloned viral hybridization of these isolates was surprising because fragments; target DNA in each well is labeled above image, while they were isolated from different hosts (A7 from M12- the probe used for each blot is identified on the right side of image. (B) Hybridization of the clone 5 probe with the Mono Lake 2c, a Vibrio spp.; and 72 from the MN1-12.5a, identified viral community DNA collected from the upper 9 m of the water as a Marinospirillum spp.), which apparently do not column at different stations and times. One microgram of DNA share any similarities and are not closely related. was blotted from August, November, and April; whereas 500 ng of Hybridization of clones 5 and 76 probes with Mono DNA was used from the March samples. (C) Hybridization of the Lake viral community DNA collected from various clone 5 probe to 500 ng of DNA collected from various depths of stations at different seasons and depths showed nearly station 6 during February and March, 2004. S. SABET ET AL.: SEASONAL PRESENCE OF HYPERSALINE PHAGES 551

Aφφ2 Discussion Several previous studies have reported the isolation and 25 µg, 2m phylogenic characterization of the Mono Lake bacterial May 2004 community [7, 14, 17, 18, 22, 27, 32]; however, very little 1 µg, 2m is known of the Mono Lake viral community. It is not surprising that four of the Mono Lake bacterial isolates µ 1 g, 8m that have successfully served as phage hosts resemble July 2004 1 µg, 35m culturable species from similar environments elsewhere in the world. H. boliviensis strain LC1, the closest match Figure 6. Dot blot hybridization of A7 and 72 probes with Mono for MN12-2a, is a moderately halophilic, psychrophilic, Lake viral community DNA. Water samples were collected from alkali-tolerant bacterium isolated from a soil sample station 6 at 2-, 8-, and 35-m depths in May and July of 2004. around lake Laguna Colorada in Bolivia [24]. M. Amount of target DNA blotted is labeled on the right of image. alkaliphilum Z4 strain, the match for MN1-12.5a, was Film was exposed overnight before development. isolated from Haoji soda lake in the Inner Mongolia region of China and optimally grows at 2% NaCl, pH 9.5, and 37-C[36]. M12-2c was identified as a Vibrio and that the phage sequences are more prevalent during spp., perhaps the most commonly isolated species from the late winter and spring seasons and less prevalent the brackish environment [5, 28]. However, M12-chla during the summer and autumn months. One possible only resembles a bacterium previously isolated from exception may be station 7, which showed weaker Mono Lake. hybridization compared to the other stations in April (Fig. 5B). This observation may reflect microenviron- A model describing the effect of viral infection on the mental changes at different stations during the same maintenance of bacterial diversity was put forth by season. No significant variability of clones 5- and 76-like Wommack and Colwell [35] in which specific host sub- phages was observed along a water column depth profile populations that underwent fast growth and reached at station 6 (Fig. 5C), despite the stratification of the high concentrations were selectively killed by their viruses. water column with oxygenic and anoxic layers. This model takes into account the presence of multiple Surprisingly, hybridization of the A7 and 72 probes phage–host systems in the same environment, each with with the same set of DNA blots under the same its own independent cycle of bacterial host growth and conditions yielded negative results (data not shown). reduction, and is generally known as the Bkill-the-winner^ This suggests that both A7- and 72-like phages were less theory. The observation of seasonal dynamics of cultur- abundant within the viral community than the uncul- able bacteriophages in Mono Lake partially supports this tured clones 5- and 76-like phages. A 25-fold increase in theory. For example, phages infecting host M12-chla were the amount of viral community DNA on the blots from found only during the winter season when their host was 1to25mg and an increase in the film exposure time initially isolated, suggesting the winter condition favors from 75 min to 18 h resulted in a positive hybridization the growth of the host and, subsequently, the proliferation signal between the A7/72 probes and the Mono Lake of its phages. Strong seasonal variability of MN12-2a and viral community DNA (Fig. 6). A7 hybridized most MN1-12.5a phage concentration (ranging from 50 to over strongly to the 1 mg viral DNA collected at the 8-m 8000 pfu/L) was also observed, suggesting the active depth, but much less to the 2- and 35-m depths, interaction between these phages and their hosts. Future implying that variability exists in the vertical distribution investigations that incorporate the temporal dynamics of of the viral community in Mono Lake and that the A7- specific host populations may provide additional insights like virus is most prevalent at 8 m. 72 also hybridized to support this theory. However, phages infecting M12-2c more strongly to the sample collected from 8 m relative did not display any temporal patterns that support the to the 2-m sample; however, the intensity of 72 kill-the-winner theory. They were present at constantly hybridization was very low relative to A7. 72 very low levels (100–200 pfu/L) at all seasons and depths sam- weakly hybridized to the 2-m sample, and is seemingly pled (Fig. 2 and Table 1), which suggests that the kill-the- absent altogether at 35 m. In addition, the amount of winner theory may not apply to all host–phage systems. target DNA was also a crucial factor since the A7 and 72 Temperate interaction between virus and host, where probes hybridized more intensely when 25 mg of viral virus and host concentrations reach equilibrium, may community DNA was blotted—a 25-fold increase. These also be an important mode of phage and host interaction data suggest that A7 and, in particular, 72 represent in the aquatic environment [12]. minority populations within the Mono Lake bacterio- All plaques observed and counted on their specific phage community since more DNA and a longer host lawn were clear or semiclear in this study, suggesting exposure time were needed to detect their presence. lytic infection. Turbid plaques resembling those formed 552 S. SABET ET AL.: SEASONAL PRESENCE OF HYPERSALINE PHAGES by temperate coliphage on their host lawn have not been bias. We are confident the cloned inserts are of viral observed or confirmed with host–phage systems isolated origin based on the follow reasons: (1) The Centricon-80 from Mono Lake. However, we cannot completely rule with 100,000 NMWL does not favor the recovery of out the presence of temperate phage among culturable dissolved extracellular DNA; (2) Dissolved DNA forms a phages found in Mono Lake. For example, although both smear on PFGE gels rather than DNA bands with defined A7 and 72 form clear plaques on their host lawns, we sizes, therefore, extracting DNA from a defined-size DNA also found cross-hybridization between the A7 genome band does not favor recovery of any significant amount and its host, MN1-12.5a. Future genetic studies focusing of dissolved DNA; and (3) DNase treatment before PFGE on identification of the homologous DNA sequence efficiently removed dissolved DNA from the samples. between this phage and its host will shed light on the Sequence homology analysis indicated that the question of Blytic^ or Blysogenic^ of this specific host– uncultured Mono Lake phages contain both viral and phage system. Assuming a portion of culturable phages bacterial genes (Table 2). The largest ORF exhibited observed in Mono Lake is temperate, then the increase in relatively high homology to single-strand DNA-binding specific phage density could reflect an event of lysogenic protein from several different bacterial and viral species induction that changes the host bacterial population (Table 2). Single-strand DNA-binding protein, a helix from steady state to the state of Bemergency response^ destabilizer, has high affinity to single-stranded DNA and under environmental stress conditions. Such an observa- is involved in replication, recombination, and repair. It is tion was noted for halophages isolated in a Jamaican not unreasonable to think that viruses that harbor this hypersaline . When the pond was in an evapora- would employ it in either their own replication or tive state with a robust bacterial density and high salinity perhaps contribute to the function of their host’s (5 M [Cl_]), relatively few phages were isolated, whereas genome. when the pool was diluted by rains and the Other sequence hits were to phage holin, Pas6 halophilic bacterial population was destroyed due to protein of 7Asp2, and capsid protein from 7Ch1. Along reduced salinity (1.3 M [Cl_]), there was a much higher with , holin is part of a dual protein system in lytic phage yield [31]. It was suggested that bacterial pop- double-stranded DNA phages that forms a pore in the ulations may escape viral by exploiting an bacterial cell’s membrane allowing lysin access to the cell environmental niche (in this case, high salinity). Hosts wall to degrade the layer ultimately for would maintain a stable relationship with their viral phage release. Pas6 is a hypothetical protein believed to parasites within a robust population under ideal host be involved in head , and is similar to conditions, but when environmental conditions were no bacteriophage SPP1 Gp7 (accession number X89721). longer ideal and the host population was Bdoomed^ to Bacteriophage 7Ch1 is a haloalkaliphilic phage that death, viruses would then act as scavengers by becoming infects Natrialba magadii, a bacterium isolated from an active lytic phages [31]. Therefore, the seasonal variabil- alkaline East African soda lake [29]. Interestingly, 7Ch1 ity of phages in Mono Lake may be indicative of the contains segmented nucleic acid composed of both DNA removal of certain predominant bacterial populations as and RNA molecules [34]. The same ORF that showed a result of environmental changes that produce a denser high homology to Pas6 and 7Ch1 also showed high host population (kill-the-winner theory); or it may be homology to a hypothetical protein from a Mesorhi- due to environmental changes that trigger lysogens to zobium species. become lytic. The absence of a more defined seasonal ORFs from clone 31 did not reflect any high expression of phages infecting host M12-2c may be due homology to current database sequences. In light of this to a continuous, stable presence of the M12-2c host in observation, a search was conducted using the sequence Mono Lake year-round, or simply to a chronic, nonlytic of the entire 3.3-kb insert, and a homology hit with an E infection. value of 0.059 was returned to a hypothetical protein Our attempt at cloning entire bacteriophage ge- from bacteriophage Aa723, isolated from an oral nomes has not been successful in spite of the promise of bacterium. This temperate phage expresses a 44-kb linear the BAC vector technology [4, 26]. This failure may be DNA molecule that has 1.6-kb terminal redundancy due to the expression of toxic phage genes in the E. coli signifying circular permutations [33]. Another finding host and/or the possibility that the lake viruses were from clone 31 is the sequence hit to a putative NTP- composed of circular DNA, or DNA that is otherwise binding protein from 7AT3, a temperate bacteriophage modified. The smaller than expected size inserts may be isolated from an anaerobic lactic-acid-forming bacteri- due to the shearing of the viral genomes during the um. The fact that clone 31 did not express any genes with cloning procedure. It was also interesting to observe that strong homologies to current database sequences implies two of the five clones containing viral fragments were that clone 31 ORFs are unique and belong to a phage identical based on initial sequencing, implying some viral that is quite distinct from any that has been described fragments are more clonable than others due to cloning thus far. This is not surprising because as much as 74% S. SABET ET AL.: SEASONAL PRESENCE OF HYPERSALINE PHAGES 553 of metagenomic sequences do not have hits to current Furthermore, we have illustrated a more detailed image database sequences [3], indicating that a large percentage of the genetic makeup of these phages. They contain of environmental viruses are still undescribed and unlike typical phage genes that encode for holins and capsid any of their laboratory isolate cousins. proteins, as well as for single-strand DNA-binding Hybridization results obtained by using probes made proteins. It is very likely that the phage community of cultured phage isolates and uncultured viral clones in contains temperate phages as well as lytic phages, linear this study led to the conclusion, long suspected, that DNA-containing viruses as well as viruses that contain laboratory isolates do not accurately represent the circular permutations. dominant or majority species in the environment, and that the uncultured members of the microbial commu- nity are those that truly reflect the dominant, majority groups. The phages represented by clones 5 and 76 were Acknowledgments present at a fairly constant level year-round. Perhaps their hosts are more similar in seasonal abundance to the We thank Dr. Grieg Steward and Dr. Robert Jellison for M12-2c host, which yielded phages year-round, and less their technical expertise and their collaboration in field like the other hosts, which only yielded phages seasonally sampling. We also thank Sandra Roll, Kimberly Rose, (Fig. 2). An interesting observation of the clone 5 and and other members of the SNARL team for collecting clone 76 seasonal hybridization pattern is the relatively water samples during this study. We thank Sam Choi for weak hybridization to the November 2004 DNA sample his help with statistical analysis and anonymous (Fig. 5B). This may relate to the physical mixing event reviewers for their suggestions for improving this that occurred in November 2003. The last time the lake manuscript. This study was supported by NSF awards underwent holomixis (lake turnover) was between DEB-01-30528, DEB-01-29174, and DEB-01-29160 to October 22 and November 14, 2003. The November S.C.J., G.F.S., and R.J., respectively. water sample used in this study was collected on November 14, the last day of holomixis (R. Jellison, pers. comm.). It is likely the microbial community was References affected during the holomixis such that the abundance 1. Adams, MH (1959) Bacteriophages. Interscience Publishers, New of the phage population as represented by clones 5 and York 76 was reduced compared to when the lake was stratified. 2. Ausubel, FM (1987) Current Protocols in Molecular . This could explain the relatively weaker hybridization Greene Publishing Associates and Wiley, Brooklyn, NY intensity of these clones to the November sample. 3. Breitbart, M, Salamon, P, Andresen, B, Mahaffy, JM, Segall, AM, Another possibility is that the phages as represented by Mead, D, Azam, F, Rohwer, F (2002) Genomic analysis of uncultured marine viral communities. Proc Natl Acad Sci USA clones 5 and 76 do indeed express a seasonal profile and 99: 14250–14255 are simply not as abundant during the autumn months. 4. Choi, S, Wing, RA (2003) The Construction of Bacterial Artificial Such a result would confirm our findings (Fig. 2)of Chromosome (BAC) Libraries 2003. Clemson University seasonal variability among phage isolates. Institute (CUGI), Clemson, SC The A7 and 72 probes hybridized with each other’s 5. Eilers, H, Pernthaler, J, Glockner, FO, Amann, R (2000) Cultur- ability and in situ abundance of pelagic bacteria from the North phage DNA (Fig. 5A), indicating that sequence homol- Sea. Appl Environ Microbiol 66: 3044–3051 ogy exists between these two viruses although they were 6. Fuhrman, JA (1999) and their biogeochemical and isolated from very distinct bacterial hosts. This result ecological effects. Nature 399: 541–548 contradicts the previous finding that only phages infect- 7. Humayoun, SB, Bano, N, Hollibaugh, JT (2003) Depth distribu- ing related host bacteria share genome similarities [10]. tion of microbial diversity in Mono Lake, a meromictic soda lake in California. Appl Environ Microbiol 69: 1030–1042 The results here suggest that bacteriophages from the 8. Jarling, M, Bartkowiak, K, Pape, H, Meinhardt, F (2004) The same environment may share some genetic homology as genome of phiAsp2, an Actinoplanes infecting phage. Virus Genes it relates to increased survivability and/or better infection 29: 117–129 in their shared respective , even though they infect 9. Jellison, R, Melack, JM (1993) Algal photosynthetic activity and its diverse hosts. Environmental viruses might then have response to meromixis in hypersaline Mono Lake, California. Limnol Oceanogr 38: 818–837 more in common with each other on the basis of their 10. Jiang, S, Fu, W, Chu, W, Fuhrman, JA (2003) The vertical geographical distribution, which may in turn lead to a way distribution and diversity of marine bacteriophage at a station off to classify them based on their shared Bgeographical^ Southern California. Microb Ecol 45: 399–410 genes, in addition to their host range. 11.Jiang,S,Steward,G,Jellison,R,Chu,W,Choi,S(2004) In summary, the findings in this report contribute to Abundance, distribution, and diversity of viruses in alkaline, hypersaline Mono Lake, California. Microb Ecol 47: 9–17 a better understanding of the seasonal population 12. Jiang, SC, Paul, JH (1998) Significance of lysogeny in the marine dynamics of both phage isolates and uncultured mem- environment: studies with isolates and a model of lysogenic phage bers of the Mono Lake bacteriophage community. production. Microb Ecol 35: 235–243 554 S. SABET ET AL.: SEASONAL PRESENCE OF HYPERSALINE PHAGES

13. Mann, NH, Cook, A, Millard, A, Bailey, S, Clokie, M (2003) 26. Shizuya, H, Birren, B, Kim, U, Mancino, V, Slepak, T, Tachiiri, Y, Marine : bacterial genes in a virus. Simon, M (1992) Cloning and stable maintenance of 300-kilobase- Nature 424: 741 pair fragments of human DNA in using an F- 14. Mason, DT (1967) of Mono Lake, California. Univ factor-based vector. Proc Natl Acad Sci USA 89: 8794–8797 Calif Publ Zool 83: 1–110 27. Switzer Blum, J, Burns Bindi, A, Buzzelli, J, Stolz, JF, Oremland, 15. Middelboe, M (2000) Bacterial growth rate and marine virus–host RS (1998) arsenicoselenatis, sp. nov., and Bacillus selenitir- dynamics. Microb Ecol 40: 114–124 educens, sp. nov.: two haloalkaliphiles from Mono Lake, California 16. Nuttall, SD, Dyall-Smith, ML (1993) HF1 and HF2: novel that respire oxyanions of selenium and arsenic. Arch Microbiol bacteriophages of halophilic . 197: 678–684 171: 19–30 17. Oremland, RS, Hoeft, SE, Santini, JM, Bano, N, Hollibaugh, RA, 28. Thompson, FL, Iida, T, Swings, J (2004) of vibrios. Hollibaugh, JT (2002) Anaerobic oxidation of arsenite in Mono Microbiol Mol Biol Rev 68: 403–431 Lake water and by a facultative, arsenite-oxidizing chemoauto- 29. Tindall, BJ, Ross, HNM, Grant, WD (1984) Natronobacterium gen troph, strain MLHE-1. Appl Environ Microbiol 68: 4795–4802 nov. and Natronococcus gen. nov., two new genera of haloalkali- 18. Oremland, RS, Miller, LG, Culbertson, CW, Robinson, SW, Smith, philic archaebacteria. Syst Appl Microbiol 5: 41–57 RL, Lovley, D, Whiticar, ML, King, GM, Kene, RP, Iversen, N, 30. van Hannen, EJ, Zwart, G, van Agterveld, MP, Gons, HJ, Ebert, J, Sargent, M (1993) Aspects of the of methane in Laanbroek, HJ (1999) Changes in bacterial and eukaryotic Mono Lake and the Mono Basin of California. In: Oremland, RS (Ed.) community structure after mass of filamentous Biogeochemistry of Global Change. Chapman & Hall, New York associated with viruses. Appl Environ Microbiol 65: 795–801 19. Oren, A (2002) Halophilic and their environ- 31. Wais, AC, Daniels, LL (1985) Populations of bacteriophage ments. In: Seckbach, J (Ed.) Cellular Origin and Life in Extreme infecting Halobacterium in a transient brine pool. FEMS Microbiol Habitats, Vol. 5. Kluwer Academic Publishers, Boston, pp 497–498 Ecol 31: 323–326 20. Oren, A, Bratbak, G, Heldal, M (1997) Occurrence of virus-like 32. Ward, BB, Martino, DP, Diaz, MC, Joye, SB (2000) Analysis of particles in the Dead Sea. 1: 143–149 ammonia-oxidizing bacteria from hypersaline Mono Lake, Cal- 21. Pelagos Corporation (1987) A bathymetric and geologic survey of ifornia, on the basis of 16S rRNA sequences. Appl Environ Mono Lake, California. Project report, San Diego, CA Microbiol 66: 2873–2881 22. Pikuta, EV, Hoover, RB, Bej, AK, Marsic, D, Detkova, EN, 33. Willi, K, Meyer, J (1998) DNA analysis of temperate bacteriophage Whitman, WB, Krader, P (2003) Tindallia californiensis sp. nov., Aa(phi)23 isolated from Actinobacillus actinomycetemcomitans. a new anaerobic, haloalkaliphilic, spore-forming isolated Mol Gen Genet 258: 323–325 from Mono Lake in California. Extremophiles 7: 327–334 34. Witte, A, Baranyi, U, Klein, R, Sulzner, M, Luo, C, Wanner, G, 23. Post, FJ (1981) of the Great Salt Lake north arm. Kruger, DH, Lubitz, W (1997) Characterization of Natronobacte- Hydrobiologia 81: 59–69 rium magadii phage phi Ch1, a unique archaeal phage containing 24. Quillaguaman, J, Hatti-Kaul, R, Mattiasson, B, Alvarez, MT, DNA and RNA. Mol Microbiol 23: 603–616 Delgado, O (2004) Halomonas boliviensis sp. nov., an alkali- 35. Wommack, KE, Colwell, RR (2000) Virioplankton: viruses in tolerant, moderate halophile isolated from soil around a Bolivian aquatic ecosystems. Microbiol Mol Biol Rev 64: 69–114 hypersaline lake. Int J Syst Evol Microbiol 54: 721–725 36. Zhang, W, Xue, Y, Ma, Y, Grant, WD, Ventosa, A, Zhou, P (2002) 25. Seaton, GG, Hurry, VM, Rohozinski, J (1996) Novel amplification Marinospirillum alkaliphilum sp. nov., a new alkaliphilic helical of non-photochemical chlorophyll fluorescence quenching follow- bacterium from Haoji soda lake in Inner Mongolia Autonomous ing viral infection in Chlorella. FEBS Lett 389: 319–323 Region of China. Extremophiles 6: 33–37