international Journal of Systematic Bacteriology (1 999), 49, 1 63 1-1 643 Printed in Great Britain

Psychrophi Iic sulf ate-reducing isolated from permanently cold Arctic marine sediments : description of oceanense gen. nov., sp. nov., sp. nov., gelida gen. nov., sp. nov., Desulfotalea psychrophila gen. nov., sp. nov. and Desulfotalea arctica sp. nov.

Christian Knobtauch, Kerstin Sahm and Bo B. Jcbrgensen

Author for correspondence: Christian Knoblauch. Tel: +49 421 2028 653. Fax: +49 421 2028 690. e-mail : [email protected]

Max-PIa nc k-I nst it Ute for Five psychrophilic, Gram-negative, sulfate-reducing bacteria were isolated Mari ne M icro bio logy, from marine sediments off the coast of Svalbard. All isolates grew at the in Celsiusstr. 1, 28359 Bremen, Germany situ temperature of -1.7 "C. In batch cultures, strain PSv29l had the highest growth rate at 7 "C, strains ASV~~~and LSv54l had the highest growth rate at 10 "C, and strains LSv21Tand LS~514~had the highest growth rate at 18 "C. The new isolates used the most common fermentation products in marine sediments, such as acetate, propionate, butyrate, lactate and hydrogen, but only strain ASv26' was able to oxidize fatty acids completely to CO,. The new strains had growth optima at neutral pH and marine salt concentration, except for LSv54l which grew fastest with 1O/O NaCI. Sulfite and thiosulfate were used as electron acceptors by strains ASv26'. PSv29l and LSv54l, and all strains except PSv29' grew with Fe3+(ferric citrate) as electron acceptor. Chemotaxonomy based on cellular fatty acid patterns and menaquinones showed good agreement with the phylogeny based on 165 rRNA sequences. All strains belonged to the 6 subclass of but had at least 9% evolutionary distance from known sulfate reducers. Due to the phylogenetic and phenotypic differences between the new isolates and their closest relatives, establishment of the new genera Desulfotalea gen. nov., Desulfofaba gen. nov. and Desulfofrigus gen. nov. is proposed, with strain ASv26' as the type strain of the type Desulfofrigus oceanense sp. nov., LSv21Tas the type strain of Desulfofrigus fragile sp. nov., PSv2gTas the type strain of the type species sp. nov., LSvW as the type strain of the type species Desulfotalea psychrophila sp. nov. and LSv514l as the type strain of Desulfotalea arctica sp. nov.

Keywords: sulfate-reducing bacteria, psychrophiles, chemotaxonomy, Arctic sediment, Svalbard

INTRODUCTION organic carbon remineralization in marine sediments (Jargensen, 1982; Canfield et al., 1993; Nedwell et al., Sulfate reducers are responsible for up to 50% of the 1993). Acetate, propionate, lactate, butyrate and hydrogen, which are the major end-products of fer- ., ...... , , , .. . , . , . . , , , ...... , ,. , ...... , ., ., . , , , , ., . , , , , ., , , , , , ., . . , , , , , , ., . , , ., .. , ., ...... mentation, constitute their most important carbon and Abbreviations: ECL, equivalent chain-length; MK, menaquinone. energy substrates (Sorensen et al., 198 1 ; Christensen, The GenBank accession numbers for the 165 rDNA sequences of Desulfotalea arctica LSv5 14T, Desulfotalea psychrophila LSV~~~,Desulfo- 1984; Parkes et al., 1989). According to their nutrition, faba gelida PSv2gT, Desulfofrigus oceanense ASV~~~and Desulfofrigus sulfate-reducing bacteria can be separated into two fragile LSv2l are AF099061-AF099065, respectively. distinct groups. Lactate, hydrogen and propionate are

01129 0 1999 IUMS 1631 C. Knoblauch, K. Sahm and B. B. Jerrgensen the typical substrates for incompletely oxidizing medium (Widdel & Bak, 1992). The medium contained sulfate-reducing bacteria, which are mainly repre- (g 1-I) : NaC1, 26.4; MgSO, .7H,O, 6.8 ; MgC1,. 6H,O, 5.7; sented by Desulfovibrio and Desulfobulbus species. The CaC1,.2H,O, 1.5; KBr, 0.09; and KC1, 0.7. After auto- main end-product of their catabolism is acetate which claving, the medium was cooled under a gas mixture of CO,/N, (10/90, v/v) and the following components were they oxidize The major of do not further. substrates added: 50 ml of a NH,Cl (5 g 1-I) and KH,PO, (4 g 1-I) completely oxidizing sulfate-reducing bacteria like solution, 1 ml nonchelated trace element solution, 1 ml Desulfobacter, Desulfobacterium, Desulfococcus and selenite/tungstate solution, 1 in1 vitamin solution (modified Desulfosarcina strains, are fatty acids which are solution 6, with an additional 4 mg folic acid and 1.5 mg oxidized to CO, (Holt et al., 1994). Phylogenetically, lipoic acid per 100 ml), 1 ml thiamin solution, 1 ml vitamin most sulfate reducers belong to the 6 subclass of B,, solution, 1 ml riboflavin solution (25 mg 1-1 in 25 mM Proteobacteria. phosphate buffer, pH 3.2), 30 ml bicarbonate solution (1 M), 1 ml resazurin solution (1 g 1-l) and 1 ml sodium sulfide The natural environment of most sulfate reducers is solution (1 M). If necessary, the pH was adjusted with HCl cold, since 90% of the sea floor has temperatures or NaOH to 7.1-7.3. The medium was dispensed under an below 4 "C (Levitus & Boyer, 1994). Like other benthic atmosphere of COJN, (10/90, v/v) into sterile serum bottles bacteria, sulfate reducers must therefore be able to that were closed with black butyl rubber stoppers or into grow at low temperatures. However, nearly all the sterile 15 ml Hungate tubes. Before inoculating the medium, known isolates are mesophiles with a temperature dithionite (final concentration 150 pM) and the desired optimum at or above 30 "C and unable to grow below carbon source were added from sterile stock solutions. The dilution series were inoculated on board ship and trans- 4°C (Widdel & Bak, 1992). It was unclear whether ported back at 4 "C. In our laboratory, they were incubated those sulfate reducers active at low in situ temperatures at five different temperatures between 0 and 20 "C. For the are closely related to the known mesophiles or whether isolation of pure cultures, the modified deep agar dilution they represent members of new genera and species. technique (Isaksen & Teske, 1996) was applied, which Pure cultures were needed to understand their metab- protects temperature-sensitive organisms from overheating. olism and temperature adaptation as well as their Agar (Noble; DIFCO) was washed five times with distilled phylogeny. The first moderately psychrophilic sulfate- water (Widdel & Bak, 1992) before use. After three to four reducing species, Desulforhopalus vacuolatus, was iso- subsequent agar dilution series, 30 different pure cultures lated by Isaksen & Teske (1996) from a temperate were isolated from the 0, 4 and 10 "C enrichments. Stock estuary. In polar environments with permanent cultures were kept at the temperature used for isolation and temperatures around 0 "C, low-temperature-adapted transferred every 3-4 weeks to fresh medium. For the characterization of pure cultures, the saltwater medium should be the dominant organisms. The aim bacteria (Widdel & Bak, 1992) with a lower concentration of major the study isolate of present was to and describe the salts was used. This medium contained (g 1-I): NaCl, 20; most abundant low-temperature-adapted sulfate redu- Na,SO,, 4; MgC1,. 6H,O, 3 ; CaCl, .2H,O, 0.1 5; KBr, 0-09; cers from polar sediments. Special attention was paid and KCl, 0.5. After autoclaving, the medium was prepared to organisms oxidizing acetate, propionate, lactate and as described above. To prevent damage of temperature- butyr ate . sensitive cells, great care was taken to protect enrichments and pure cultures from temperatures above those used for isolation. METHODS Physiology and metabolism. The salt requirement for growth was monitored in media with 15 different NaCl concen- Sources of organisms. Arctic marine sediments at Svalbard trations between 0.2 and 5.8% (w/v) or 16 different were sampled in 1995 on a cruise with the RV 'Jan Mayen'. MgC1,. 6H,O concentrations between 0-0 and 7.0 Oh (w/v). Strains LSv21T, ASV~~~and PSV~~~ originated from The concentrations of all other salts, except the one being Hornsund sediment (76'58.2' N, 15'34.5' E) with a bottom tested, were kept constant. The vitamin demand of the water temperature of 2.6 'C. Strains LSV~~~and LS~514~ different strains was tested for at least ten subsequent were isolated from Storfjord sediment (77'33.0' N, transfers on medium without vitamins. The pH optimum 19'05.0' E) with a bottom water temperature of - 1.7 "C. was tested using media adjusted to 12 different pH values Further information about sampling sites are given in Glud between 4.9 and 9-1. The pH was adjusted in triplicate tubes et al. (1998). Desulforhopalus vacuolatus strain 1tklF' (= with HCl or NaOH and the tubes were inoculated. The DSM 9700T) was kindly provided by Kai Finster, Arhus, initial pH was measured in one tube and the remaining tubes Denmark; Desulfovibrio giganteus (DSM 4123) was were incubated. Sulfide was measured periodically during obtained from the Deutsche Sammlung von Mikro- the following six months. Growth with different electron organismen und Zellkulturen (DSMZ), Braunschweig, donors was tested with sulfate as electron acceptor. Tubes Germany. without an electron donor were inoculated and served as negative controls. Sulfide was measured periodically during Enrichment, isolation and cultivation. Sediment samples were the following year. Growth tests on different electron collected with a multicorer and subsampled directly on the acceptors were made in sulfate-free medium which was deck of the ship at an ambient temperature of 2-7 "C. supplied with the carbon source used for isolation of the Subcores were sliced in an anaerobic glove bag and samples tested strain and either thiosulfate (10 mM), elemental from five sediment depths between the surface and 30cm sulfur, nitrate (5 mM), nitrite (2 mM), iron(II1) oxy- were transferred to 90 ml sterile artificial seawater medium. hydroxide or iron(II1) citrate (30 mM). Amorphous iron(II1) These samples were suspended for 2 min with a vortex mixer oxyhydroxide was prepared by titration of an acidic FeC1, and further diluted in 15 ml Hungate tubes containing 10 ml solution (0.5 M) with NaOH (2 M) to pH 7.0. Elemental

1632 International Journal of Systematic Bacteriology 49 Psychrophilic sulfate-reducing bacteria sulfur was added with a spatula from a sterilized suspension; bacterial cell numbers in the cultures (P < 0.005). If direct all other electron acceptors were added from sterilized stock cell counts or OD were needed, cultures were carefully solutions. Tubes without an electron donor served as homogenized with an ultrasonic probe (HD200 ; Bandelin) negative controls. Disproportionation of thiosulfate and applying the lowest intensity. Microscopic controls revealed elemental sulfur was tested in sulfate-free medium. Either that cells were not damaged by this treatment. Direct cell 20 mM thiosulfate or elemental sulfur plus iron(II1) oxy- counts were made under the microscope in an improved hydroxide was added to the tubes. Additionally 2mM Neubauer chamber. The OD was measured in a spectro- acetate was added as a carbon source. All test-tubes were photometer (UV-1202) at 580 nm. Growth rates were inoculated with a sulfate-free preculture. The same pre- calculated in exponentially growing cultures by a linear cultures were used as inocula for fermentation tests. Test regression of ln(OD,,,) as a function of time. tubes contained no electron acceptor but did contain lactate, pyruvate, fumarate, malate or propionate at a final con- DNA isolation and DNA-DNA hybridization. DNA of strains centration of 10 mM. Growth was measured by direct counts LSv5 14T, Ls~54~and Desulforhopalus vacuolatus ltk10' under the light microscope. All tests were incubated at least was isolated according to Marmur (1961). DNA-DNA in duplicate at the temperature used for isolation of the hybridization was performed using the hydroxyapatite different strains, i.e. 4 or 10 "C. method as described by Ziemke et al. (1998) except that DNA was labelled with [32P]dCTP by nick-translation Fatty acids, lipoquinones and polar lipids. Cellular fatty acids as described by Rossello-Mora et al. (1994). After de- were determined at the DSMZ, Braunschweig, Germany by naturation, DNA-DNA mixtures were incubated at 30 "C R. M. Kroppenstedt. Fatty acid methyl esters were obtained below the melting temperature of homologous DNA, which by saponification and separated by GC as described by in our case was 62 "C for all hybridization pairs. The degree Vainshtein et al. (1992). Respiratory lipoquinones and polar of reassociation (binding ratio) was calculated by dividing lipids were extracted from freeze-dried cell material and the counts for double-stranded DNA by the total counts for analysed by TLC (Tindall, 1990). The analyses were carried double- and single-stranded DNA. The relative binding out by B. J. Tindall at the DSMZ. ratio for the heterologous pairs is expressed as the percentage Pigments and Gram-staining. The desulfoviridin test was of homologous binding (Lind & Ursing, 1986). carried out as described by Postgate (1984). Desulfovibrio 165 rDNA amplification. Cells were harvested from 2 ml giganteus (DSM 4123) served as positive control. To culture samples by centrifugation and resuspended in 100 pl determine the Gram-reaction of the strains, heat-fixed cells 1 x PBS. A subsample of 1 pl was used directly as a template were stained with Crystal Violet as described by Murray for the amplification of 16s rDNA. PCR reactions were ef al. (1994). performed as described by Buchholz-Cleven et aE. (1 997). To G+C content of genomic DNA. The G+C content of the amplify the nearly complete 16s rDNA, primers 8F/1492R genomic DNA was determined by HPLC (Mesbah et al., (Buchholz-Cleven et al., 1997) were used in a 35 cycle PCR 1989) at the DSMZ. with an annealing temperature of 40 "C. Chemical analysis. Sulfide was measured according to the 165 rDNA sequencing. PCR products were purified with quick method described by Cord-Ruwisch (1985). If higher the QIAquick PCR Purification kit (Qiagen). The Taq precision was needed, the methylene blue method of Cline DyeDeoxy Terminator Cycle Sequencing kit (Applied Bio- (1969) was applied. Volatile fatty acids and lactate were systems) was used to directly sequence the purified PCR determined by ion-exclusion chromatography with an products. Sequencing reactions were analysed on the Ap- HPLC system (Sykam) and a refractometer (ERC-7515) as plied Biosystems 373s DNA sequencer. Both strands of the detector. The components were separated on a Sarasep WA1 amplification products were sequenced using primers 8F, column (300 x 7.8 mm) at 60 "C with H,SO, (15 mM) as 787F, 787R, 1175R, 1099F and 1492R (Buchholz-Cleven eluent. The flow was adjusted to 0.6 ml min-'. Fifty et al., 1997). Primer nomenclature refers to the 5' ends of microlitres of a 0.45-pm-filteredsample (Acrodisc 4 ;Gelman the respective target sites on the 16s rDNA according to Sciences) was injected onto the column. Dissolved Fe2+was the Escherichia coli numbering of 16s rRNA nucleotides determined according to Stookey (1970) with a Ferrozine (Brosius et al., 1981). solution (1 g 1-1 in 50 mM HEPES buffer, pH 7.0). A 100 p1 Phylogenetic analysis. The ARB program package and the sample was diluted in 5 ml HCl (0-5 M) and after 15 min, ARB database (Strunk et al., 1999) were used for phylogenetic 50 p.1 were removed and added to 2.5 ml Ferrozine solution analysis. Sequences were aligned to the 16s rRNA primary and measured in a spectrophotometer (UV- 1202; Shimadzu) structures present in the ARB database by using the automatic at a wavelength of 562 nm. Organic carbon was determined aligning tool and the results were corrected manually where in a CHNS analyser (Cutter & Radford-Knoery, 1991). A necessary. Pairwise distance matrix analysis was performed known culture volume was filtered on two GF/F-filters with the 16s rRNA sequences taking only those positions (1-28 cm; Frisenette) placed in two filter holders on top of into account that were present in both sequences. Evol- each other. The lower filter was used as a blank. The filters utionary distances were calculated using the Jukes-Cantor were flushed with a marine salt solution and then dried in a correction. Phylogenetic trees were reconstructed with repre- stream of sterile-filtered air. The filters were placed in tin sentatives of most genera from the 6 subclass of Proteo- capsules and 50 p1 distilled water and 50 p1 HC1 (50 mM) bacteria. A selection of representatives for major groups were added to dissolve the bicarbonate. After 2 h, the filters outside this subclass was used as an outgroup. Only were dried overnight at 105 "C and analysed on a CHNS sequences with at least 1400 nt were used. Tree topology was analyser (Fisons). evaluated by using neighbour-joining, maximum-parsimony Growth determination and growth rates. Since most of the and maximum-likelihood algorithms on either the full set of isolates tended to grow in aggregates, the determination of data or on a selected subset. Furthermore, filters were growth via an increase in the OD was difficult. Growth was applied that excluded positions with less than 50% con- therefore measured routinely by the increase of sulfide in the servation within the 6 subclass. Branching orders that were cultures. Sulfide concentrations correlated significantly with not supported by all methods are shown as mwltifurcations. lnternational Journal of Systematic Bacteriology 49 1633 C. Knoblauch, K. Sahm and B. B. Jargensen

Naming of strains. The isolated strains were named with a containing up to several hundred cells. In old cultures code indicating the carbon source used for isolation (L, of strains ASV~~~,LSv21T, PSV~~~and LSv54*, some lactate; A, acetate; P, propionate), the sampling site (Sv2, cells were motile. All isolated strains stained Gram- Hornsund; Sv5, Storfjord) and a number. negative.

Growth conditions RESULTS Enrichment and isolation All strains had pH optima in the neutral range (Table 1). Marine sodium concentrations supported optimum Enrichment cultures at 4 and 10 "C started to produce growth of ASV~~~,LSv21T, PSV~~~and LS~514~ but sulfide after 4 weeks if lactate was used as carbon the lower range of optimal sodium concentrations source, and after 8-10 weeks if acetate or propionate differed slightly (Table 1). Although strain LSV~~~was was used. Sulfide-producing enrichments were trans- enriched and isolated on medium with a marine salt ferred to fresh medium until stable enrichments were concentration of 2-5% NaCl and 1-1 YOMgC1,. 6H,O, obtained and then agar dilution series were inoculated. it had a remarkably lower salt optimum of 1 YONaCl After 68weeks, all dilution series were dominated by and 0.034-0.700 Yo MgCl, . 6H,O. Marine magnesium brownish to blackish, disk-shaped, smooth colonies. If concentrations supported optimum growth of strains lactate was used as carbon source, white, fluffy colonies ASV~~~,LSv21T, PSV~~~and LS~514~. All strains were also present in the lower dilutions. Single colonies grew well at a temperature of - 1-8 "C, but optimum were picked and directly transferred to new dilution growth temperatures were up to 20 "C higher (Table series until pure cultures were obtained. Only the 1). LSv21T and LS~514~had highest growth rates (p) brownish colonies contained sulfate reducers. Thirty at about 18 "C with values of 0-036 and 0.021 h-I, pure cultures were isolated from different sediment respectively (I,= 19 and 33 h, respectively), when samples and dilution steps. Based on a preliminary grown on lactate. A lower optimum temperature of physiological and phylogenetic characterization, the 10 "C was characteristic for strains ASV~~~(p = five psychrophilic strains presented here were selected 0.0041 h-l; t, = 169 h on acetate) and LSV~~~(p = for further description. Strains LSv21T, LSV~~~and 0.026 h-l; t, = 27 h on lactate). PSV~~~grew fastest at LS~514~were isolated on lactate, whereas acetate was 7 "C with p of 0.0048 h-l (t, = 144 h on propionate). used for the isolation of ASV~~~and propionate was At the next highest temperature tested (10 "C), this used for the isolation of PSV~~~.The temperature isolate could only reduce sulfate but not grow. A during isolation was 4 "C for ASV~~~,LSv21T, PSV~~~ detailed description of the temperature response of the and LS~514~and 10 "C for strain LSV~~~. strains will be presented elsewhere (Knoblauch & Jorgensen, 1999). Added vitamins were not required by any of the strains. Cultures were transferred usually Purity controls with 10 YOinoculum since some of the strains exhibited All strains were transferred on media containing yeast a very long and unpredictable lag phase if smaller extract (O-lY, w/v) and either formate (20 mM), inocula were used. Strain LSv21T lysed rapidly in the pyruvate (10 mM), glucose (5 mM), fructose (5 mM) stationary phase and was therefore transferred every or the carbon source used for isolation of the respective second week. strain. Microscopic examinations revealed that no morphologies different from those of the tested strains Substrate spectra could be detected in any culture. Growth on yeast Strain ASV~~~grew on a wide variety of carbon extract, glucose or fructose was never observed. The sources (Table 1) and oxidized fatty acids such as different strains were also transferred to agar dilution formate, acetate, butyrate and valerate completely to series. Only the strain-typical colony morphologies CO,. Although LSv21T was isolated on lactate, it also developed in all dilution steps. grew on longer-chain saturated fatty acids (C6, C10, C16) which were only incompletely oxidized to acetate. Morphology Fast growth was also found with different alcohols. PSV~~~was the only strain, that could grow on Cells of the acetate-oxidizing strain ASV~~~(Fig. la) propionate, which was incompletely oxidized to acet- were thick rods with rounded ends. Growth in loose ate. Growth was also possible on alcohols and clumps was often observed. Cells of strain LSv21T dicarboxylic acids. The substrate spectra of strains (Fig. lb) were slightly curved rods with rounded ends. LS~514~and LSV~~~were similar but distinct from PSV~~~(Fig. lc) was a large, slightly curved, peanut- those of strains LSv21T,ASV~~~ and PSV~~~ (Table 1). shaped rod. The cells grew exclusively in dense With the exception of formate, no straight-chain fatty aggregates and single cells could rarely be observed at acids were oxidized but both strains grew fast on any growth stage. Cells of strain LSV~~~(Fig. Id) hydrogen when acetate was added as carbon source. often appeared in pairs. In exponentially growing Lactate was incompletely oxidized to acetate. In cultures, one or two shorter cells between two long comparison to LS~514~,strain LSV~~~had a wider cells frequently could be observed. Cells of LS~514~ substrate spectrum, also growing on various alcohols (Fig. le) were short rods often growing in clumps and amino acids.

1634 lnterna tional Journal of Systematic Bacteriology 49 Psychrophilic sulfate-reducing bacteria ~.

...... , .. , , ,. , , , , ,. . . , . , .. . . , .. . . , . . . . , . , .. . , ...... , . . . ., , , ., , ,, , ...... , , , . Fig. 1. Phase-contrast photomicrographs of (a) Desulfofrigus oceanense strain ASV~~~, (b) Desulfofrigus fragile strain LSv21T, (c) Desulfofaba gelida strain PSv2gT, (d) Desulfotalea psychrophila strain LSV~~~ and (e) Desulfotalea arctica strain LS~514~. Bar, 10 pm.

The electron acceptors used by the different strains are fermented pyruvate but none disproportionated thio- listed in Table 1. Besides sulfate, strains ASV~~~,sulfate or elemental sulfur. PSV~~~and LSV~~~ also used thiosulfate and sulfite an electron acceptor. LS~514~also reduced iron(II1) Desulfoviridin was not detected in any of the strains. oxyhydroxide and sulfur very slowly but could not The major polar lipids of all strains were phosphatidyl- grow on these electron acceptors. All isolated strains ethanolainine and phosphatidylglycerol (Table 1 ); in

Interna tional lourna I of Systematic Bacteriology 49 1635 C. Knoblauch, K. Sahm and B. B. Jmgensen

Table 7. Some characteristics of new psychrophilic sulfate-reducing bacteria

Characteristic Desulfofrigus Desulfofrigus Desulfofaba Desulfotalea Desulfotak oceanense fragile gelida psychrophila arctica ASV~~~ LSV21T PSv29T Lsv54T LSv514T

Cell size (pm) : Width 2.1 0.8 3.1 0.6 0.7 Length 4-24'1 3.2-4.2 5.4-6.2 4.5-7.4 1-6-2-7 pH optimum 7.0-7.5 7.0-7.4 7.1-7.6 7.3-7.6 7.2-7.9 Optimum salt requirement (Yo): NaCl 1.5-2.5 1.0-2.5 14-2.5 1.0 1.9-2.5 MgCl,. 6H,O 0.00 3-2.0 0'3-2.0 0.0 15-25 0.03-0.7 0.3-1'4 Temperature 10/ - 1.8-16 18/ - 1.8-27 7/ - 1-8-10 10/ - 1.8-19 18/ - 1.8-: optimum/range ("C) Growth rate (h-')/doubling 0.0041 / 169 0*036/19 0-0048/ 144 0*026/27 0.02 1/3 3 time (h) at optimum temperature Electron donors (mM) :* Formate (20) ++ + + ++ ++ Acetate (10) ++ - - - - Propionate (1 5) - - ++ - - Butyrate (5) ++ + + - - Valerate (5) + - - - - Caproate (3) - ++ - - - Caprate (2) - ++ - - - Palmitate (2) - +- - - - Lactate (1 0) ++ ++ ++ ++ ++ Pyruvate (1 0) + ++ + + +t ++ Malate (10) ++ ++ + +- - Succinate (1 0) - - + - - Fumarate (1 0) - + + ++ - Ethanol (10) ++ ++ ++ ++ ++ Propanol ( 10) ++ ++ ++ ++ - Butanol (10) ++ ++ ++ ;+ - Glycerol (10) +- ++ +- - + Glycine (1 0) +- - +- + - Alanine (1 0) - + + + - Serine (10) +- + - + +- HJCO, +acetate (2) +- - - ++ ++ Electron acceptors (mM):f Sulfate (28) + + + + + Thiosulfate (10) + - + + - Sulfite (2) + - + + - - Sulfur - - - -§ Iron(II1) citrate (30) + + - + + - - - - Iron(II1) oxyhydroxide - /I Fermentable compounds (mM) : Pyruvate (1 0) + + + + + Malate (10) + + - - - Lactate (10) + - - - - Fumarate (10) - - + + - Polar lipids1 PE, PG PE, PG PE, PG PE, PG, PE, PG, DPG DPG Major menaquinones MK-9 MK-9 MK-8 MK-6H, MK-6 G + C content (mol YO) 52.8 52-1 52-5 46.8 41.8 *Sulfate (28 mM) was used as electron acceptor. Substrates tested (mM) but not oxidized by any strains: formate (autotrophic), isovalerate (5), methanol (lo), glutarate (lo), betaine (lo), choline chloride (lo), L-proline (lo), D-sorbitol(5), D-mannitol(5), benzoate

1636 International Journal of Systematic Bacteriology 49 Psychrophilic sulfate-reducing bacteria

Table 2. Fatty acid composition of psychrophilic sulfate-reducing bacteria

Abbreviations exemplified by: 14:0, tetradecanoic acid; 14: lc9, 9-tetradecenoic acid, double bond cis-standing; 15 : 0 3-OH, 3- hydroxy-heptadecanoic acid; 16: 0 alde, hexadecanal. Major fatty acids are printed in bold.

Fatty acid Desulfofrigus Desulfofrigus Desulfofaba Desulfotalea Desulfotalea Desulforhapalus oceanense fragile gelida psychvophila avctica vacuolatus ASV~~~ LSV21T PSv29T LSv54T LSv514T ltklOT

lo:o - 12:o - 13:O 2.0 14:O 4.9 14: lc9 0.5 15:O 38-1 15:O 3-OH 0.3 15: lc9 24.7 16:O alde - 16:O 5.4 16:O 3-OH - 16: lc7 0.5 16: lc9 6.5 16: lcll 1.5 17:O 1.7 17: lcll 4-4 18:O 0.3 1810 3-OH 0.3 18:O alde - 18: lc9 - 18: lcll - 18: lc13 - 20: lc13 - strains LSV~~~and LSv5 14T, diphosphatidylglycerol of the fatty acids that are characteristic for known also was present in lower amounts. As indicated in sulfate-reducing bacteria, namely the branched fatty Table 1, all strains contained menaquinones (MK). In acids for Desulfovibrio species (Vainshtein et al., 1992), strains ASV~~~and LSv21T, only MK-9 was present, the 16 :0 10-me for Desulfobacter species (Dowling which is uncommon for sulfate-reducing bacteria and et al., 1986), or the 17: lc9 for Desulfobulbus species has only been found in Desulfonema magnum (Collins (Taylor & Parkes, 1983). The two strains ASV~~~and & Widdel, 1986). PSV~~~contained MK-8 as the sole LSv21T had a similar fatty acid pattern dominated by menaquinone, which is also rare in sulfate-reducing 16: lc9, 18: lcll and 16:O. Only even-numbered fatty bacteria but is characteristic for sulfur-reducing acids were present, an indication of the use of acetyl- Desulfuromonas strains (Collins & Widdel, 1986). In CoA as precursor for chain elongation during the LSV~~~,the major menaquinone was MK-6H2 but synthesis of fatty acids (Taylor & Parkes, 1983). Both traces (z1 YO)of MK-5H, also were present. LS~514~ strains contained a fatty acid (4.0 and 5.6%, re- only contained MK-6. spectively) with an equivalent chain-length (ECL) of 15-49 that could not be identified. A completely different pattern was found in strain PSV~~~which was Cellular fatty acids dominated by the 15 : 0 fatty acids. This acid was also The cellular fatty acids of the different strains are found (23%) in Desulfobulbus species grown on presented in Table 2. The new isolates contained none propionate (Taylor & Parkes, 1983), which was also

(I), nicotinate (I), H, (autotrophic), glucose (5), fructose (5). + + , Substrate oxidized after 6 weeks; + , substrate oxidized after 4 months; + -, substrate oxidized after 8 months or more; -, substrate not oxidized. No sulfide produced, only fermentative growth. 1Nitrate and nitrite were not reduced by any of the strains. $Reduction of elemental sulfur up to a concentration of 3 mM but no growth. 11 4 mM iron(II1) reduced after 1 year. 7 PE, Phosphatidylethanolamine ; PG, phosphatidylglycerol ; DPG, diphosphatidylglycerol. lnternationa I Journal of Systematic Bacteriology 49 1637 C. Knoblauch, K. Sahm and B. B. Jarrgensen

Desulfofaba gelida PSv2gT Desulfofrigus oceanense ASV~~~ DesulfosarcinalDesulfonema Desulfo frigus fragile LSv21 Desulfotalea psychrophila LSV~~~ Desulfotalea arctica LS~514~

Desulfomonile tiedjei Desulforhopalus vacuolatus

Desulfofustis glycolicus PelobacterlGeobacterIDesulfuromonas

10 Yo

Desulfovibrionaceae

Fig. 2. Distance tree based on 165 rRNA sequences. The tree shows the 6 subclass of Proteobacteria and was constructed by the neighbour-joining method using a 50 % conservation filter. It depicts the phylogenetic distances using Jukes-Cantor correction with the bar representing 10 % estimated sequence divergence. Areas of interest where the branching order changed when different treeing methods were applied are shown as multifurcations. The arrow indicates the position of the outgroup (a selection of bacterial 165 rDNA sequences from a wide range of phyla). In addition to the Svalbard isolates, which are shown here in bold, 37 known species from the 6 subclass of Proteobacteria were selected to reconstruct the tree. These reference species are shown here as groups for a better overview. the carbon source of PSv29'. The second most used as precursor for chain elongation, since odd- abundant fatty acid, 15 : 1c9, is uncommon in sulfate- numbered fatty acids dominated, being 87 % of the reducing bacteria and was until now found only in low identified acids. amounts in a Desulfobacter spp. grown on a mixture of fatty acids (Dowling et al., 1986). In strain PSV~~~, 80 % of the identified fatty acids were odd-numbered, G + c content of genomic DNA an indication that propionyl-CoA was precursor for chain elongation. The same result was also found in The G+C contents of strains ASV~~~,LSv21T and different species (Taylor & Parkes, PSv29' were very similar, 52-53 mol% (Table 1). Desulfobulbus Lower G+C contents were found in LSV~~~ 1983). Two fatty acids with an ECL of 14-80 (5.5%) and 19.47 (7.4%) could not be identified. Strains (47 mol%) and LSv5 14T (42 mol %). LSv54' and LS~514~exhibited very similar fatty acid patterns which were clearly different from those of the Phylogenetic affiliation other new strains. The dominant fatty acids were the monounsaturated 16: lc9 and 16: lcl 1 comprising The 16s rDNA sequences showed that all isolates more than 80 % of the total identified fatty acids. As in belonged to the 6 subclass of Proteobucteriu. Although ASV~~~and LSv2 lT, the strains have preferentially they share a common characteristic in being psychro- even chain fatty acids. Desulforhopalus vacuolatus philic, they do not form a cluster within the 6 subclass strain ltklOT was included in our study since it is, as but are distributed between groups of mesophilic yet, the only known moderately psychrophilic sulfate- sulfate-reducing bacteria (Fig. 2). All isolates had at reducing bacterium. It is most closely related to strains least 9% evolutionary distance from 16s rDNA LSv54' and LSv514'. The fatty acid pattern of sequences of known mesophilic sulfate-reducing bac- Desulforhopalus vacuolatus was clearly dominated by teria. Based on their 16s rDNA sequence, strains the 17 : 1cl 1 fatty acid. Although Desulforhopalus ASV~~~and LSv21' are closely related to each other vacuolatus was grown on lactate, propionyl-CoA was with an evolutionary distance of 3-2%. We regard

1638 InternationalJournal of Systematic Bacteriology 49 Psychrophilic sulfate-reducing bacteria them as different species since they showed distinct such as propionate, butyrate, caproate, caprate and physiological differences with different substrate even palmitate incompletely to acetate which can be spectra and ASV~~~is a complete oxidizer whereas further oxidized by strain ASV~~~to CO,. Besides LSv2 1 is an incomplete oxidizer. The closest relative lactate and hydrogen, volatile fatty acids are the most to both strains was PSV~~~with evolutionary distances important end-products of fermentation and are the of 10.4 and 103YO, respectively, followed by Desuljo- major organic substrates for sulfate reducers in sarcina variabilis with a distance of 1 1.6 YOto ASV~~~temperate environments (Ssrensen et al., 1981 ; and 12.2% to LSv21T. PSV~~~was not closely related Christensen, 1984; Parkes et al., 1989). The fact that all to any known strain but shares the highest 16s rDNA these compounds were also oxidized at subzero similarity with Desuljosarcina variabilis (9.5 YOevol- temperatures by the new psychrophilic isolates is a utionary distance). Strains LSV~~~and LSvS 14T were further indication that these bacteria occupy the same closely related having 3.3 YO evolutionary distance. ecological niche in cold sediments as the mesophiles in Since their physiological and chemotaxonomic proper- temperate sediments. ties were quite similar (Table l), we performed DNA- DNA hybridization to establish whether they belonged Chemotaxonomy to the same species. The relative binding ratio for DNA-DNA hybridization of these two isolates was According to their cellular fatty acid pattern, the new below 20 %, which is well below the threshold value of strains can be assigned to three distinct groups. The 70% accepted for the distinction of different species major fatty acids of both ASV~~~and LSv21T were (Wayne et al., 1987). Based on these results, strains 16:lc9, 18:lcll and 16:O acids (Table 2). This LSV~~~and LS~514~ can be regarded as two dif- combination is unique among the known sulfate ferent species. The closest relatives of both strains reducers and supports the assignment of these two were Desuljofustis glycolicus and Desulforhopalus isolates to a new genus. PSV~~~,the closest relative to vacuolatus with an evolutionary distance of 9-0-9.3 YO. ASV~~~and LSv2 1T, had a completely different pattern with the 15:O and the 15: lc9 fatty acids dominating. Both fatty acids were absent in Desulfosarcina DISCUSSION variabilis (Kohring et al., 1994), which is phylo- Ecology genetically the closest related strain to PSV~~~.The third group comprised strains LSV~~~and LSvS 14T. The psychrophilic isolates from permanently cold Both have a very similar fatty acid pattern dominated Arctic sediments are the first known sulfate reducers by 16: 1c9 and 16: lcl 1, a combination which, to our that are able to grow below 0 "C. The existence of low- knowledge, has not been found in any other sulfate temperature-adapted sulfate reducers was evident, reducer. Desulforhopalus vacuolatus strain ltk 1OT was since previous studies demonstrated that rates of included in our study since it is moderately psychro- sulfate reduction in polar sediments were comparable philic and most closely related to strains LSV~~~and to those in temperate environments (Nedwell et al., LS~514~.Its fatty acid pattern was dominated by 1993; Sagemann et al., 1998). The new strains LSV~~~17: lcl 1 acid (Table 2) but it also contained 15: lc9 and LS~514~grew fastest on lactate, pyruvate, acid (12 %). Therefore, in the Euclidian distance tree alcohols and hydrogen, which are the characteristic (data not shown), ltklOT was more closely related to carbon and energy substrates of Desulfovibrio species. PSV~~~than to LSV~~~ and LS~514~. These results In the present study, no organisms morphologically cannot be due to differences in the growth conditions, resembling Desulfovibrio strains were detected in any because ltklOT, LSV~~~and LS~514~ were all cul- enrichment incubated between 0 and 10°C and no tivated with lactate on the same medium at 4 "C. The Desulfovibrio strain was isolated at low temperatures. percentage of unsaturated fatty acids in strains LSV~~~ These results are unusual, since Desulfovibrio strains and LS~514~(89 and 91 %, respectively; Table 2) is are generally dominant in lactate-containing enrich- remarkably high. Among the mesophilic sulfate-re- ment cultures (Postgate, 1984). The major difference ducing bacteria, the highest content of unsaturated between previous enrichments and ours was the fatty acids was 62% in Desulfovibrio africanus temperature used for incubation, 28-36 "C vs 0-10 OC, (Vainshtein et al., 1992). The high degree of un- indicating that growth temperature might affect the saturated fatty acids in the new strains may be an outcome of competition between different groups of adaptation to low temperature. Unsaturated fatty sulfate reducers. This was also supported by the results acids lower the gel-liquid&crystalline phase transition of Isaksen & Teske (1996) who enriched sulfate temperature of membranes (Russell, 1990), thereby reducers with lactate from a temperate estuary at maintaining the necessary fluidity at low growth 10 "C. They isolated the moderate psychrophilic strain temperatures. High concentrations of unsaturated 1tk1OT,which they described as the type strain of a new fatty acids were also found in ASv26' and LSv21T genus Desulforhopalus vacuolatus. However, it might (Table 2) but not in PSV~~~.Another way to increase also be that Desuljovibrio species are not abundant in membrane fluidity is to decrease the fatty acyl chain the investigated habitat and that their ecological niche length (Bhakoo & Herbert, 1979). This is the case in is taken by species related to LSV~~~arid 123~514~. PSV~~~,which contains more than 70% fatty acids Strains LSv21T and PSV~~~oxidize various fatty acids with a chain length of 15 carbon atoms or less. It is

International Journal of Systematic Bacteriology 49 1639 C. Knoblauch, K. Sahrn and B. B. Jsrgensen not known if sulfate-reducing bacteria alter their fatty acid pattern and MK content, we propose the membrane composition when grown at different establishment of a new genus, Desulfofrigus, for temperatures. ASV~~~and LSv21T. On the basis of an evolutionary distance of 3.2 Yo,distinct morphologies, temperature The MK analyses revealed similar relationships be- responses of growth and substrate spectra, we classify tween the new isolates as the fatty acid profiles. ASV~~~ASV~~~ and LSv21T as two species of the genus and LSv21T contained only MK-9, which was not Desulfofrigus. ASV~~~is the type strain of the type found in any of the other new isolates and, out of 45 species Desulfofrigus oceanense and LSv2 1 is the type sulfate-reducing bacteria strains investigated by strain of Desulfofrigus fragile. Collins & Widdel (1986), it only occurred in Desulfonema magnum. The sole MK in PSV~~~was Due to the evolutionary distance of 9.5% between MK-8, which is absent in its closest relative, PSV~~~and Desulfosarcina variabilis and the dif- Desulfosarcina variabilis (Collins & Widdel, 1986). ferences in their physiology and chemotaxonomic MK with six isoprenoid subunits were dominant in properties, we propose the establishment of the new strains LSV~~~and LS~514~ (Table 1). These MK are genus Desulfofaba with PSV~~~as the type strain of the characteristic for most Desulfovibrio strains (Collins & type species Desulfofaba gelida. Widdel, 1986). Unfortunately, MK data are not Considering the phylogenetic distance between strains available for their closest relatives, Desulfofustis LSV~~~,LSv5 14T and Desulforhopalus vacuolatus as glycolicus and Desulforhopalus vacuolatus. The well as their physiological, morphological and chemo- chemotaxonomic data were in good agreement with taxonomic differences, we place strain LSV~~~and the 16s rRNA sequence data, which showed the same LS~514~in a new genus Desulfotalea. Since both affiliation between the different isolates (Fig. 2). More strains have distinct physiological properties and a conflicting results arose from the substrate spectra of DNA-DNA similarity of less than 70%, we have strains ASV~~~and LSv21T. Although most closely defined two new species of the genus Desulfotalea, with related to ASV~~~(evolutionary distance 3-2YO), strain LSV~~~as the type strain of the type species LSv2 lT only incompletely oxidized fatty acids to and LSv5 14T as the type acetate. Complete or incomplete substrate oxidation Desulfotalea psychrophila was traditionally used as a criterion to distinguish strain of Desulfotalea arctica. genera of sulfate-reducing bacteria (Devereux et al., 1989; Holt et al., 1994). These distinctions have been Description of Desulfofrigus gen. nov. supported by phylogenetic data. The lowest evo- (De.sul.fo.fri'gus. L. prefix off; L. n. lutionary distance between a completely oxidizing Desulfofrigus de sulfur sulfur; L. neut. n. frigus cold; M.L. neut. n. species (Desulfococcus multivorans) and an incom- sulfate reducer living in the cold). pletely oxidizing sulfate-reducing bacteria (Desulfo- Desulfofrigus vibrio sapovorans) was 11 YO.Our results demonstrate Members are Gram-negative, obligately anaerobic that complete substrate oxidation to CO, is not always, bacteria and belong to the 6 subclass of Proteobacteria. phylogenetically, a deep branching property. Chemo- Sulfate is used as terminal electron acceptor and taxonomic parameters such as fatty acid pattern and reduced to sulfide. Iron(II1) can also be used as electron MK content of the different isolates were in closer acceptor. Fermentative growth on pyruvate or other agreement with the phylogenetic data than the sub- carbon substrates. Chemoorganotrophic growth on strate spectra. fatty acids and alcohols that are either completely Strains LSV~~~and LSv5 14T were phylogenetically oxidized to CO, or incompletely to acetate. No most closely related to and chemoautotrophic growth. Major cellular fatty acids Desulfofustis glycolicus are even, mono-unsaturated and unbranched. MK-9 is Desulforhopalus vacuolatus. The new strains shared only a few general characteristics with the genus the dominant menaquinone. The type species of this genus is A second member of Desulfofustis, e.g. the absence of desulfoviridin and Desulfofrigus oceanense. the incomplete oxidation of lactate. The most con- this genus is Desulfofrigus fragile. spicuous feature, which the new strains shared with Desulforhopalus vacuolatus, is psychrophily. On Description of Desulfofrigus oceanense gen. nov., sp. the other hand, the new strains could be easily nov. differentiated from by Desulforhopalus vacuolatus (o.ce.a.nen'se. L. adj. the absence of gas vacuoles, by their inability to grow Desulfofrigus oceanense on propionate and by their completely different cellular oceanensis, -e belonging to the ocean). fatty acid pattern. Cells are 2.1 pm wide and 4.2-6-1 pm long. The pH optimum for growth is 7-0-7.5. Requires sodium Taxonomic affiliation chloride; optimum growth occurs at marine salt concentrations. Temperature optimum for growth is Phylogenetically, the closest relative to ASv26* and 10 "C and growth also occurs at - 1.8 "C.Contains the LSv2IT was PSV~~~with evolutionary distances of polar lipids phosphatidylethanolamine and phosph- 10.4 and 10.5 %, respectively. Due to this phylogenetic atidylglycerol ; MK-9 is the sole menaquinone. Major distance, as well as dissimilarities in physiology, the cellular fatty acids are 16 : 1c9 and 18 : 1cl 1. The G + C

1640 Interna tionaI lo urna I of Systematic Bacteriology 49 Psychrophilic sulfate-reducing bacteria content is 52.8 mol %. Sulfate, thiosulfate and sulfite Description of Desulfofaba gelida gen. nov., sp. nov. are used as electron acceptors and are reduced to sulfide. Iron(II1) serves as an electron acceptor when Desulfofaba gelida (ge'kda. L. adj. gelidus, -a, -urn ice- added as iron(II1) citrate. Elemental sulfur, iron(II1) cold, referring to the low temperature optimum). oxyhydroxide, nitrate and nitrite are not reduced. Cells are 3.1 pm wide and 5-46.2pm long. The pH Formate, acetate, butyrate, valerate, lactate, pyruvate, optimum for growth is 7-1-76. Requires sodium malate, ethanol, propanol, butanol, glycerol, glycine chloride and magnesium chloride ; optimum growth and serine serve as carbon substrates. Fatty acids are occurs at marine salt concentrations. The temperature oxidized completely to CO,. Growth on H, plus acetate optimum for growth is 7 "C; growth also occurs at (2mM) is slow. Fermentative growth on pyruvate, - 1.8 "C. Contains the polar lipids phosphatidyl- malate and lactate. Vitamins are not required for ethanolamine and phosphatidylglycerol ; MK-8 is the growth. Desulfoviridin is absent. Elemental sulfur and sole menaquinone. Major cellular fatty acids are 15 :0 thiosulfate are not disproportionated. Habitat is per- and 15 : lc9. The G + C content is 52.5 mol YO.Sulfate, manently cold marine sediments. Type strain is ASV~~~thiosulfate and sulfite are used as electron acceptors (= DSM 12341T). which are reduced to sulfide. Elemental sulfur, nitrate, nitrite, iron(II1) oxyhydroxide and iron(II1) citrate are Description of Desulfofrigus fragile sp. nov. not reduced. Sulfur and thiosulfate are not dis- proportionated. Formate, propionate, butyrate, lac- Desulfoofigus fragile (fra'gi.le. L. adj. fragilis, -e re- tate, pyruvate, malate, succinate, fumarate, ethanol, ferring to the rapid lysis of the type strain in the propanol, butanol, glycerol, glycine and alanine serve stationary phase). as carbon sources. Fatty acids are oxidized incom- Cells are 0.8 pm wide and 3.2-4.2 pm long. The pH pletely to acetate. Fermentative growth on pyruvate optimum for growth is 7.0-7.4. Requires sodium and fumarate. Vitamins are not required for growth. chloride; optimum growth occurs at marine salt Desulfoviridin is absent. Their habitat is permanently concentrations. Temperature optimum for growth is cold marine sediments. Type strain is PSV~~~(= DSM 18 "C and growth also occurs at - 1-8 "C. Contains 12344T). the polar lipids phosphatidylethanolamine and phosphatidylglycerol ; MK-9 is the sole menaquinone. Description of genus Desulfotalea gen. nov. Major cellular fatty acids are 16: lc9, 16:0 and 18 : lcl 1. The G + C content is 52.1 mol Yo. Sulfate is Desulfotalea (De.sul.fo.ta'1e.a. L. prefix de off; L. n. used as electron acceptor and is reduced to sulfide. sulfur sulfur; L. fem. n. talea a rod; M.L. fem. n. Elemental sulfur, sulfite, thiosulfate, nitrate, nitrite Desulfotalea a sulfate-reducing rod). and iron(II1) oxyhydroxide are not reduced. Iron(II1) Members are Gram-negative obligately anaerobic serves as electron acceptor when added as iron(II1) bacteria belonging to the 6 subclass of Proteobacteria. citrate. Formate, butyrate, caproate, caprate, palmi- Sulfate is used as electron acceptor and reduced to tate, lactate, pyruvate, malate, fumarate, ethanol, sulfide. Fermentative growth on pyruvate. Iron(II1) propanol, butanol, glycerol, alanine and serine serve as can be used as electron acceptor. Major carbon or carbon sources and electron donors. Fatty acids are energy sources are lactate, alcohols and hydrogen. oxidized incompletely to acetate. Fermentative growth Cellular fatty acids comprise even-numbered, mono- on pyruvate and malate. Vitamins are not required for unsaturated acids. The dominant menaquinones have growth. Desulfoviridin is absent. Cells lyse rapidly in six isoprenoid units. Desulfotalea psychrophila is the the stationary growth phase. Elemental sulfur and type species of the genus. A second member of this thiosulfate are not disproportionated. Habitat is genus is Desulfotalea arctica. permanently cold marine sediments. Type strain is LSv21T (= DSM 12345T). Description of Desulfotalea psychrophila gen. nov., Description of Desulfofaba gen. nov. sp. nov. Desulfofaba (De.sul.fo.fa'ba. L. prefix de off; L. n. Desulfotalea psychrophila (psy.chro'phi.la. Gr. adj. sulfur sulfur; L. fem. n. faba a bean; M.L. fem. n. psychros cold ; philos loving ; M .L. adj. psychrophilus, DesuZfofaba a sulfate-reducing bean). -a, -urn cold-loving). Members are Gram-negative, obligately anaerobic Cells are 0.6 pm wide and 4.5-7-4 pm long. The pH bacteria belonging to the 6 subclass of Proteobacteria. optimum for growth is 7-3-7.6. Requires sodium Sulfate is used as electron acceptor and reduced to chloride and magnesium chloride for growth; op- sulfide. Fermentative growth on pyruvate or other timum sodium chloride concentration is 1 %. The carbon substrates is possible. Major carbon sources temperature optimum for growth is 10 "C and growth and electron donors are fatty acids and alcohols that also occurs at - 1.8 "C. Cells contain the polar lipids are oxidized incompletely to acetate. Major cellular phosphatidylethanolamine, phosphatidylglycerol and fatty acids are odd-numbered and unbranched. MK-8 diphosphatidylglycerol. MK-6H2 is the major mena- is the dominant menaquinone. The "ype species is the quinone but traces of MK-SH, are also present. Major only species in the genus, Desulfofaba gelida. cellular fatty acids are 16: lc9 and 16: lcl 1. The G + C

International Journal of Systematic Bacteriology 49 1641 C. Kiioblauch, K. Sahm and B. B. Jlargensen content is 46.8 mol Yo. Sulfate, thiosulfate and sulfite Brosius, J., Dull, T. J., Sleeter, D. D. & Noller, H. F. (1981). Gene are used as electron acceptors and are reduced to organization and primary structure of a ribosomal RNA operon sulfide. Elemental sulfur, nitrate, nitrite and iron(II1) from Escht.richia coli. J Mol Biol 148, 107-127. oxyhydroxide are not reduced. Growth by reduction Buchholz-Cleven, 6. E. E., Rattunde, 6. & Straub, K. L. (1997). of iron(III), if added as iron(II1) citrate, is possible. Screening for genetic diversity of isolates of anaerobic Fe(l1)- Elemental sulfur or thiosulfate are not dis- oxidiing bacteria using DGGE and whole-cell hybridization. proportionated. Formate, lactate, pyruvate, malate, Syst Appl Microhiol20, 301-309. fumarate, ethanol, propanol, butanol, glycine, alanine Canfield, D. E., Jrargensen, B. B., Fossing, ti. & 7 other authors and serine serve as carbon sources. They grow on (1993). Pathways of organic carbon oxidation in three con- hydrogen plus acetate (2 mM). Organic substrates are tinental margin sediments. Mnr Geol 113, 2740. oxidized incompletely to acetate. Fermentative growth Christensen, D. (1 984). Determination of substrates oxidized by on pyruvate and fumarate. Vitamins are not required. sulfate reduction in intact cores of marine sediments. Limnol Desulfoviridin is absent. Habitat is permanently cold Oceanogr 29, 189-192. marine sediments. Type strain is LSV~~~(= DSM Cline, J. D. (1969). Spectrophotomctric determination of hy- 12343T). drogen sulfide in natural waters. Lirnnol Uccurzogr 14.454-458. Collins, M. D. & Widdel, F. (1986). Respiratory quinones of Description of Desuffutalea arctica sp. nov. sulphate-reducing and sulphur-reducing bacteria : a systematic investigation. S~!sfAppl Microhiol8, 8-18. (arc'ti.ca. L. adj. Desulfotalea arctica arcticus, -a, -urn Cord-Ruwisch, R. (1985). A quick method for the determination from the Arctic, referring to the site were the type of dissolved and precipitated sulfides in cultures of sulfate- strain was isolated). reducing bacteria. J Microbiol Methods 4, 33-36. Cells are 0.7 pm wide and 1-&2-7 pm long. The pH Cutter, G. A. & Radford-Knoery, J. (1991). Determination of optimum for growth is 7-2-7-9. Sodium chloride and carbon, nitrogen, sulfur, and inorganic sulfur species in marine magnesium chloride are required for growth ; optimum particles. In Murine Pcirticles: Anctlvsis and Ciicirnc~tt.ri,-utioii, growth occurs at marine salt concentrations. The pp. 57-63. Edited by D. C. Hurd. Washington, DC: American temperature optimum for growth is 18 "C and growth Geophysical Union. also occurs at - 1.8 "C. Cells contain the polar lipids Devereux, R., Delaney, M., Widdel, F. & Stahl, D. A. (1989). phosphatidylethanolamine, phosphatidylglycerol and Natural relationships among sulfate-reducing bacteria. diphosphatidylglycerol. MK-6 is the sole mena- J Bacterioll71, 6689-6695. quinone. Major cellular fatty acids are 16: lc9 and Dowling, N. 1. E., Widdel, F. & White, D. C. (1986). Phospholipid 16:lcll. The G+C content is 41-8molYO. Sulfate ester-linked fatty acid biomarkers of acetate-oxidizing sulphate- serves as electron acceptor and is reduced to sulfide. reducers and other sulphide-forming bacteria. J Gen Microbiol Thiosulfate, sulfite, nitrate and nitrite are not reduced. 132, 1815-1825. Growth by reduction of iron(III), if added as iron(II1) Glud, R. N., Holby, O., Hoffmann, F. & Canfield, D. E. (1998). citrate. Elemental sulfur and iron(1II) oxyhydroxide Benthic mineralization and exchange in Arctic sediments are slowly reduced without growth. Elemental sulfur (Svalbard, Norway). Mar Ecol Prog Ser 173, 237-251. and thiosulfate are not disproportionated. Formate, Holt, 1. G., Krieg, N. R., Sneath, P. H. A,, Staley, 1. T. & Williams, lactate, pyruvate, ethanol, glycerol and serine serve as 5. T. (editors) (1994). Bergey's Manual of Determinative Bac- carbon sources and electron donors. Hydrogen plus teriology. Baltimore : Williams & Wilkins. acetate (2 mM) allows rapid growth. Organic sub- Isaksen, M. F. & Teske, A. (1996). Desulforhopalus vacuolatus gen. strates are oxidized incompletely to acetate. Fermenta- nov., sp. nov., a new moderately psychrophilic sulfate-reducing tive growth on pyruvate. Vitamins are not required bacterium with gas vacuoles isolated from a temperate estuary. for growth. Desulfoviridin is absent. Habitat is Arch Microbioll66, 160-168. permanently cold marine sediments. Type strain is Jsrgensen, B. B. (1982). Mineralization of organic matter in the LS~514~(= DSM 12342T). sea bed - the role of sulphate reduction. Nature 296, 643-645. Knoblauch, C. & Jsrgensen, B. B. (1999). Effect of temperature on ACKNOWLEDGEMENTS sulphate reduction, growth rate, and growth yield in five psychrophilic sulphate-reducing bacteria from Arctic sedi- We thank Ramon Rossello-Mora for help with the DNA- ments. DNA hybridization and helpful discussions on , Env Microbiol (in press). Jcns Harder for introduction to anaerobic cultivation Kohring, L. L., Ringelberg, D. B., Devereux, R., Stahl, D. A., techniques, and Jan Kuver for helpful hints and for critically Mittelman, M. W. & White, D. C. (1994). Comparison of phylo- reading thc manuscript. Fricdrich Widdcl is thanked for genetic relationships based on phospholipid fatty acid profiles many inspiring discussions. Donald E. Canfield is thanked and ribosomal RNA sequence similarities among dissimilatory for leading the cruise to Svalbard and Svantjc Fleischer ;tnd sulfate-reducing bacteria. FEMS Microhiol Lett 119, 303-308. Birgit Rattunde f'or tcchnical assistance. This work was Levitus, 5. & Boyer, T. (1994). World Ocean Atlas, vol. 4, financially supported by thc Max Planck Society, Germany. Temperature. Washington, DC : US Department of Commerce. Lind, E. & Ursing, J. (1986). Clinical strains of Enterobacter REFERENCES agglomerans (synonyms : Erwin ia herbicolu, Erwin iu millet iue) Bhakoo, M. & Herbert, R. A. (1979). The effects of temperature on identified by DNA-DNA hybridization. Acta Pathol Microbiol the fatty acid and phospholipid composition of four obligately Immunol Scand Sect B Microbiol94, 205-2 13. psychrophilic Vibrio spp. Arch Microbiol121, 121-127. Marmur, J. (1961). A procedure for the isolation of deoxy-

1642 International Journal of Systematic Bacteriology 49 Psychrophilic sulfate-reducing bacteria ribonucleic acid from micro-organisms. J Mol Biol3, 208-21 8. Ssrensen, J., Christensen, D. & Jsrgensen, B. B. (1981). Volatile Mesbah, M., Premachandran, U. & Whitman, W. B. (1989). Precise fatty acids and hydrogen as substrates for sulfate-reducing measurement of the G + C content of deoxyribonucleic acid by bacteria in anaerobic marine sediment. Appl Environ Microbiol high-performance liquid chromatography. Int J Syst Bacteriol 42, 5-1 1. 39, 159-167. Stookey, L. L. (1970). Ferrozine - a new spectrophotometric reagent for iron. 779-78 1. Murray, R. G. E., Doetsch, R. N. & Robinow, C. F. (1994). De- Anal Chem 42, terminative and cytological light microscopy. In Methods .for Strunk, O., Gross, O., Reichel, B. & 11 other authors (1999). ARB: General and Molecular Bacteriology, pp. 21-41. Edited by a software environment for sequence data. http :// P. Gerhardt, R. G. E. Murray, W. A. Wood & N. R. Krieg. www.mikro.biologie.tu-muenchen.de. Department of Micro- Washington, DC : American Society for Microbiology. biology, Technische Universitat Miinchen, Munich, Germany. Nedwell, D. B., Walker, T. R., Ellis-Evans, 1. C. & Clarke, A. (1993). Taylor, 1. & Parkes, R. 1. (1983). The cellular fatty acids of the Measurements of seasonal rates and annual budgets of organic sulphate-reducing bacteria, Desulfobacter sp., Desulfobulbus carbon fluxes in an Antarctic coastal environment at Signy sp. and Desulfovibrio desulfuricans. J Gen Microbiol 129, Island, South Orkney Islands, suggest a broad balance between 3303-3309. production and decomposition. Appl Environ Microbiol 59, Tindall, B. 1. (1990). A comparative study of the lipid com- 3989-3995. position of Halobacterium saccharovorum from various sources. Syst Appl Microbiol 12, 128-130. Parkes, R. J., Gibson, G. R., Mueller-Harvey, I., Buckingham, W. J. & Herbert, R. A. (1989). Determination of the substrates for Vainshtein, M., Hippe, H. & Kroppenstedt, R. M. (1992). Cellular sulphate-reducing bacteria within marine and estuarine fatty acid composition of Desulfovibrio species and its use in sediments with different rates of sulphate reduction. J Gen classification of sulfate-reducing bacteria. Syst Appl Microbiol Microbioll35, 175-1 87. 15, 554-566. Wayne, L. G., Brenner, D. J., Colwell, R. R. & 9 other authors Postgate, J. R. (1984). The Sulplzate-reducing Bacteria. Cambridge: Cambridge University Press. (1987). International Committee on Systematic Bacteriology. Report of the ad hoc committee on reconciliation of approaches Rossello-Mora, R. A,, Caccavo, F., Jr, Osterlehner, K. & 7 other to bacterial systematics. In[ J Syst Bacteriol37, 463464. authors (1994). Isolation and taxonomic characterization of a Widdel, F. & Bak, F. (1992). Gram-negative mesophilic sulfate- halotolerant, facultatively iron-reducing bacterium. Syst Appl reducing bacteria. In The Prokaryotes, pp. 3352-3378. Edited Microbiol 17, 569-573. by A. Balows, H. G. Triiper, M. Dworkin, W. Harder & K.-H. Russell, N. J. (1990). Cold adaptation of microorganisms. Philos Schleifer. New York: Springer. Trans R SOCLond B Biol Sci 326, 595-6 1 1. Ziemke, F., Hafle, M. G., Lalucat, J. & RossellB-Mora, R. (1998). Sagemann, J., Jsrgensen, B. B. & Greef, 0. (1998). Temperature Reclassification of Shewanella putrefaciens Owen’s genomic dependence and rates of sulfate reduction in cold sediments of group I1 as Shewanella baltica sp. nov. Int J Syst Bacteriol48, Svalbard, Arctic Ocean. Geomicrobiol J 15, 85-1 00. 179-186.

International Journal of Systematic Bacteriology 49 1643