Azotobacter Salinestris Sp. Nov. , a Sodium-Dependent, Microaerophilic, and Aeroadaptive Nitrogen-Fixing Bacterium WILLIAM J

Home , Azotobacter, Azotobacter chroococcum, Azotobacter salinestris

INTERNATIONALJOURNAL OF SYSTEMATICBACTERIOLOGY, July 1991, p. 369-376 Vol. 41, No. 3 0020-77 13/91/030369-08$02.OO/O Copyright 0 1991, International Union of Microbiological Societies

Azotobacter salinestris sp. nov. , a Sodium-Dependent, Microaerophilic, and Aeroadaptive Nitrogen-Fixing Bacterium WILLIAM J. PAGE* AND SHAILAJA SHIVPRASAD Department of Microbiology, University of Alberta, Edmonton, Alberta T6G 2E9, Canada

Azotobacter salinestris sp. nov. was isolated from slightly saline soils of Western Canada, where strains of this species accounted for 55% of the aerobic nitrogen-fixing isolates. Also, one isolate was obtained from a saline Egyptian soil. These bacteria shared many physiological traits with Azotobacter chroococcum, but were absolutely dependent on Na+ ions for growth and frequently used melibiose as a carbon and energy source. Aerobically grown cells were melanized dark brown to black; the cells were large, gram negative, oval with pointed ends, and motile by means of peritrichous flagella and formed pairs and chains of six to eight cells during active growth. Capsule production was variable. Nitrogen fixation by A. salinestris occurred optimally at 35°C in the presence of molybdate or vanadate ions in a microaerophilic, aeroadaptive manner. The cells were very sensitive to H20z and catalase negative. However, a single weak catalstse electromorph was observed in cell extracts. This contrasted with a very active catalase represented by multiple electromorphs in A. chroococcum. Iron was absolutely required for growth and aeroadaptation. Other growth-promoting substrates included fructose, galactose, glucose, mannitol, starch, and sucrose. Acid was formed from growth- promoting sugars and also from the non-growth-promoting substrates arabinose, cellobiose, lactose, mannose, rhamnose, and xylose. Incubation in the presence of increased NaCl concentrations promoted acidification of the culture to inhibitory levels, and sufficient acid was released from nitrogen-fixing cells in the presence of 1.0 to 1.5% NaCl to solubilize CaCO, suspended in solid medium. The cells grew well in marine broth alone, producing an alkaline reaction. An acid reaction was produced both oxidatively and fermentatively in marine broth containing glucose. Nitrate was used in an assimilatory fashion, and there was no evidence of NO2- or N, formation. The type strain is strain 184 (=ATCC 49674), which was isolated from soils of Alberta, Canada.

Members of the genus Azotobacter are aerobic, free- strain 184T (T = type strain) has been used in growth and living, nitrogen-fixing bacteria that are found throughout the physiological studies (24, 25, 27, 34). Unlike other azotobac- world (1, 36, 37). The cells are typically large and gram ters, these bacteria, are not able to grow significantly in negative, and most species are motile. The cells usually iron-limited media (25). Iron is required by strain 184T and differentiate to form cysts and accumulate intracellular gran- has been shown to be necessary for catalase, peroxidase, ules of poly-P-hydroxybutyrate. These heterotrophic organ- and superoxide dismutase activities (27). Iron limitation isms characteristically fix atmospheric nitrogen in vigorously ameliorates oxygen toxicity. Observations that were surpris- aerated cultures. Protection of an oxygen-labile nitrogenase ing, although consistent with accumulated results, were that is achieved through rapid consumption of 0, and respiratory these strains are microaerophilic, very sensitive to H,02, protection of the enzyme (31). Life as an obligate aerobe and and catalase negative in the H,O, spot test (27). Although the use of respiration to consume oxygen during protection nitrogen-fixing growth normally starts in a microaerophilic of a nitrogenase necessitate active production of oxygen environment, these organisms can adapt to fix nitrogen in the protection enzymes. These bacteria have one of the mostly aerobic environments preferred by other azotobacters (27, highly active cytochrome oxidases known (15,16), as well as 34). This process of aeroadaptation can occur in the absence notably active superoxide dismutase and catalase systems of catalase and peroxidase activities and appears to be (14). dependent on the formation of catechol and melanin as Azotobacter chroococcum is the most common species in chemical traps for toxic oxygen products (34). the genus Azotobacter and is easily isolated from soils by In this study we further examined the physiological differ- simply sprinkling soil on nitrogen-free mineral salts medium ences between Na'-dependent and Na+-independent Azo- containing an organic carbon (and energy) source, such as tobacter chroococcum-like isolates. We propose a new starch or mannitol (1). However, as noted by Becking (l), species name, Azotobacter salinestris , for the Na+-depen- there is considerable variation among Azotobacter chroo- dent strains. coccum isolates; the cells tend to be very pleomorphic, capsules may or may not be formed, colony sizes vary, and MATERIALS AND METHODS brown to brownish-black pigmentation of the colonies may or may not occur. In 1981 Becking concluded that this Sources and isolation of bacterial strains. The strains of species was a complex group consisting of several species Azotobacter salinestris used in this study were previously whose delineations had not been sorted out (1). isolated from soils from Alberta, Canada, by F. D. Cook Strains conforming in many respects to the characteristics (Department of Soil Science, University of Alberta, Edmon- of Azotobacter chroococcum but exhibiting absolute depen- ton, Alberta, Canada) as described by Page (24). New dence on Na+ for growth were isolated recently from soils isolates were also obtained in this study by sprinkling from Alberta, Canada (24); representative Na+-dependent air-dried surface soil samples (approximately 0.1 g) onto the surface of Burk nitrogen-free mineral salts medium (8) which was modified as described by Page (24) and also contained * Corresponding author. 1% (wthol) glucose and 0.25 pg of CuCl per ml. All strains

369 370 PAGE AND SHIVPRASAD INT.J. SYST.BACTERIOL. were incubated at 30°C. Azotobacter salinestris strains were by Selander et al. (33). Peroxidase activity was detected in stored at 4°C and at room temperature as slant cultures the same gels by using the activity stain described by Butler grown in Burk nitrogen-free medium with the screw cap lids and Lachance (3). Standard horseradish peroxidase and beef tightly closed and sealed with Parafilm. Azotobacter salines- liver catalase (Sigma) were included (1 pg of enzyme protein tris strains appeared to lose viability when they were stored per well) as positive and negative controls in both gels. as slant cultures at 4"C, as reported previously for Azoto- Nitrogen fixation. Nitrogen fixation was estimated by bacter vinelandii (23). measuring growth and protein production (34) after incuba- Taxonomic tests. Taxonomic tests were performed on agar tion in liquid nitrogen-free Burk medium. In order to com- medium by using the methods of Thompson and Skerman pare the effects of molybdate, vanadate, and tungstate on (37). nitrogen-fixing growth, a culture was first starved for molyb- The oxidation of more than 100 carbon sources was also date by three serial transfers on solid Burk medium contain- lested by using multitest Biolog ES and GN plates (Biolog, ing 1.0% glucose or 1.0% sucrose and 0.1% ammonium Hnc., Hayward, Calif.). The plates were inoculated with an acetate, but no added molybdate (26). All glassware was acid active, washed cell suspension exactly as recommended by washed, and all media were made with deionized water. 1 he manufacturer. The plates were examined for positive Microscopic examination. Cell morphology and Gram re- (definite purple) reactions after 24 and 48 h of incubation at action were observed with a Leitz model HM-LUX light 30°C. microscope equipped with bright-field and phase-contrast Na'-dependent growth was defined as failure of a strain to optics and fitted with a micrometer stage slide and eyepiece. grow during 24 h of incubation with shaking at 200 rpm in Cyst formation was determined by observing the formation liquid Na+-free Burk medium made with deionized water in of rounded, phase-bright cells which, when stained by the acid-washed glassware (24). Optimal growth of the Na+- method of Vela and Wyss (38), exhibited reddish brown dependent strains typically required the addition of 1 mM exines and green central bodies. Flagellar distribution was NaCl to Burk medium (24). observed by using negatively stained (1% [wthol] uranyl Media for determining the optimum growth pH contained acetate in distilled water) preparations that were viewed with 100 mM Na+ and buffers of Good et al. (11). The buff- a Philips model 410 transmission electron microscope. ers used were piperazine-N,N'-bis(2-ethanesulfonic acid) (PIPES) (pH 6.0 to 7.0), N-2-hydroxyethy1piperazine-Nf-2- RESULTS AND DISCUSSION ethanesulfonic acid (HEPES) (pH 7.0 to KO), and tris(hy- droxymethy1)methylaminopropanesulfonic acid (TAPS) (pH Basic characterization of strains. There is no selective 8.0 to 9.0). medium for the isolation of Na+-dependent azotobacters. Oxidative-fermentative (0-F) tests were performed by All of the strains have been identified after isolation from using the modified marine broth described by Lemos et al. soil-sprinkled plates (this study) or after prior enrichment in (18). This medium contained marine broth (Difco) supple- static tubes containing liquid nitrogen-free medium (24). mented with 0.001% phenol red, 0.3% agar, and 0.05% Tris, Aerobic nitrogen-fixing colonies appeared on soil-sprinkled was adjusted to pH 7.6, and contained 1% glucose as plates after 1 week of incubation, and the colonies with a required. 0-F tests were also performed in tubes of Burk brown to black pigment were picked for pure culture isola- medium containing phenol red and agar plus 1% glucose and tion by restreaking. In strain 184T, the pigment in the 0.1% NaNO, as a nitrogen source. The production of gas colonies has been shown to be a catechol melanin, and its bubbles during growth was considered positive N, evolution production was stimulated by copper ions in the medium during growth, and nitrate utilization was followed colori- (34). These heavily pigmented azotobacters accounted for metrically (35). 110% of the aerobic nitrogen-fixing colonies on most plates, The G+C contents were determined by the thermal melt- and of those colonies picked, only 40 to 50% were found to ing method described by Johnson (13), using DNAs prepared be Na+ dependent. from Escherichia coli b and Azotobacter vinelandii UW (= A comparison of the new isolates with the strains exam- OP = ATCC 13705) as standards. Determinations were made ined previously showed that growth on melibiose was a trait with a Philips model PU 8740 UV-Vis scanning spectropho- found in almost all Na+-dependent strains (15 of 17 strains) tometer fitted with a model PU 8764 temperature program- (Table 1). The same number of strains did not grow in the mer and a Magnavox model RGB monitor. When this presence of 50 mM Rb+, which is an antagonist of Na+ (24). method and instrumentation were used, the thermal melting In contrast, the Na+-independent strains in our collection point of E. coli b DNA was 8935°C (previously published rarely used melibiose as a growth substrate (two of eight value, 90.5"C [13]), and the thermal melting point of Azoto- strains were positive), and all eight strains usually tolerated bacter vinelandii DNA was 97.96 to 98°C (calculated G+C Rb+ . content, 67.73 to 67.8 mol%; previously published value, New Na+-dependent isolates were obtained from Cana- 64.9 to 66.5 mol% [36]). dian soils from northern British Columbia, the Yukon, and Catalase and peroxidase detection. Catalase activity was Saskatchewan, as well as locations on the University of detected by observing bubbles of 0, evolved after a drop of Alberta campus. One strain, which was tentatively identified 5% H,O, was placed on an young colony or mixed with as Azotobacter chroococcum, was isolated from a saline soil active cells on a microscope slide (35). Negative reactions in Egypt (this isolate was obtained from Y. Z. Ishac, Ains were not improved by increasing the peroxide concentration Shams University, Shobra, Cairo, Egypt). Na+-dependent to 20%. strains also have been isolated recently from soils in Loui- Catalase activity was also detected in cell extracts after siana (37a). electrophoresis and staining for activity (33). Cell extracts Catalase and peroxidase electromorphs. The Na+-depen- were prepared as described by Page and von Tigerstrom dent strains were catalase negative in the H20, spot test (28). Approximately 100 to 300 pg of cell extract protein was (Table 1) (27). Encapsulated strains (e.g., strain D [Table 11) subjected to electrophoresis in an 8% nondenaturing poly- sometimes trapped very fine bubbles of 0, in this test and acrylamide gel (3,rather than the starch gel recommended appeared to be weakly positive when they were observed VOL.41, 1991 AZOTOBACTER SALINESTRIS SP. NOV. 371

TABLE 1. Basic characteristics of strains

NaC- Growth in the presence Catalase Strain(s) dependent or: Origin activity' growth Melibiose Rbf ATCC 9043T' - + ATCC~ ATCC 7493 + + ATCC 34, 88, 52, 86 - - Alberta' 109, 184T, 188, 221f + + Alberta' 46, 82, 186, 253 + + Alberta' PC24 + - Pelly Crossing, Yukon ML95 + + Muncho Lake, British Columbia WL92 + + Watson Lake, Yukon DSI + + University of Alberta DSIA + + University of Alberta DSII - - University of Alberta A - - Saskatoon, Saskatchewan B + + Rein, Saskatchewan D + - Foam Lake, Saskatchewan F - + Storoway , Saskatchewan B5 + + Egyptg

a Growth was measured as positive or negative in liquid Burk medium with or without 1 mM NaCl (to determine the Na+ requirement), on solid Burk medium containing 1% melibiose as a sole carbon source, and on solid Burk medium containing 50 mM RbCl (to determine Rb+ tolerance). Growth was scored after at least 3 days of incubation at 30°C. Catalase activity was assayed by using the standard H,O, spot test. Azotobacter chroococcum type strain. ATCC, American Type Culture Collection. See reference 24. Strain 184 is the Azotobacter salinestris type strain (this study). See reference 21. with a magnifying lens (27). Na+-independent strains always active catalase electromorphs in each cell extract (Table 2). gave clearly positive catalase reactions. Although the level One of these electromorphs had a mobility similar to the of catalase activity of the Na+-dependent strains was very mobility of the catalase electromorph found in the Na+- low, the enzyme could be demonstrated electrophoretically dependent strains. The presence of multiple electromorphs in cell extracts after staining for catalase activity. The within one strain is considered to be unusual and suggests mobility of each electromorph (a clear band against a blue that either posttranslational modification or gene duplication background) was measured before the staining pattern faded has occurred in the Na+-independent strains (33). However, (33). Analysis of the data showed that all of the Na+- multiple forms of catalase also are formed by E. coli, dependent strains examined contained one weakly active Bacillus subtilis, Pseudomonas putida, and Pseudomonas catalase electromorph and that the electromorphs of these syringae pv. phaseolicola (17, 19, 20). The fact that this strains were apparently identical (Table 2). In contrast, the enzyme exists in multiple forms attests to its importance in Na+-independent strains contained two to five strongly promoting cell survival in aerobic environments. The peroxidase activity ,found in cell extracts also was present as multiple electromorphs in strains belonging to TABLE 2. Peroxidase and catalase electromorphs both Na+ nutritional types (Table 2). In this case, two of the three electromorphs had very similar mobilities and were Mobility relative to the dye front found in strains of both nutritional types, while a third Strain(s) or standard" Peroxidase Catalase (faster) electromorph was characteristic of each nutritional electromorphs electromorphs type. ATCC 9043Tb 0.31, 0.49, 0.61 0.091, 0.13,' 0.50 Nitrogen fixation. Optimal nitrogen fixation occurred in the ATCC 7493 0.25, 0.50 0.033, 0.13,' 0.50 presence of 12.5 pg of sodium molybdate per ml, but very 34 0.24, 0.49, 0.61 0.041, 0.13' low levels of nitrogen fixation and growth were detected in 52 No bands 0.13,' 0.50 molybdate-starved cultures (Fig. 1). Nitrogen fixation in the A 0.23, 0.49, 0.61 0.060, 0.13' absence of molybdate has been observed previously in F 0.22, 0.49, 0.61 0.14' Azotobacter vinelandii and Azotobacter chroococcum and DSII 0.23, 0.49, 0.62 0.050, 0.13' was caused by an alternative molybdate-independent nitro- 184Td 0.29, 0.48, 0.72 0.074 genase (9, 12). The growth of molybdenum-starved Na+- 221 0.29, 0.48, 0.72 0.070 dependent strain 184T was enhanced by the addition of 0.025 188, 109 0.29, 0.49 0.070 DSI, DSIA 0.29, 0.48, 0.72 0.070 to 25 pg of sodium vanadate per ml (Fig. 1). One of the B5 No bands 0.070 alternative nitrogenases found in Azotobacter vinelandii and Horseradish peroxidase 0.13, 0.19 No bands Azotobacter chroococcum has been shown to be a vanadium Beef liver catalase No bands 0.070 nitrogenase (9, 12). Sodium tungstate, an antagonist of molybdate or vanadate function in nitrogenase activity, a Strains were selected from those shown in Table 1. 184T 'Azotobacter chroococcum type strain. inhibited strain growth in nitrogen-free medium (Fig. 1). ' Two electromorphs are present. Strain 184T grew optimally in nitrogen-free medium incu- Azotobacter salinestris type strain (this study). bated at 35°C and grew poorly in the same medium incubated 372 PAGE AND SHIVPRASAD INT. J. SYST.BACTERIOL.

200 TABLE 3. Differential use of carbon sources by Na+-dependent and Na+-independent strains

% of strains positive for utilizationb Carbon sourcen Na+-dependent Na+-independent strains strains -h 1 E \ D-Galactonic acid lactone 66 0 0) =L Cellobiose 66 0 Y Glucuronic acid 100 66 $ 100 Melibiose 100 66 I- a-Methyl-galactoside 66 33 Man nose 33 66 a Psicose 33 100 Quinic acid 17 100 Gentiobiose 17 100 a-C yclodextrin 0 50

~~~~~ ~ Carbon sources were tested in Biolog ES and GS multiwell plates (see Materials and Methods). 0 In this comparison we used six Na+-independent strains and six Na+- .1 1 10 100 1000 10000 dependent strains.

CONCENTRATION (ng/ml) FIG. 1. Nitrogen-fixing growth of Azotobacter salinestris sp. strains belonging to each nutritional type were screened for nov. Molybdate-starved strain 184T was inoculated into liquid Burk their ability to oxidize more than 100 carbon sources by nitrogen-free medium containing 1 mM NaCl and 1%glucose but no added molybdate, to which we added sodium molybdate (O), using Biolog multiwell plates. We found no carbon source sodium vanadate (a),or sodium tungstate (0).The cultures were that was used exclusively by all isolates of one Na+ nutri- incubated for 24 h at 30°C with shaking at 200 rpm, and the amount tional type. The carbon sources which were used disparately of cell protein formed was used as an estimate of growth. (233% difference) are shown in Table 3. However, some caution must be exercised in evaluating the results obtained with these multitest plates. Although at 28°C (Fig. 2). Azotobacter chroococcum ATCC 9043T most Na+-independent azotobacters were not able to use grew well at temperatures ranging from 28 to 37°C (Fig. 2). In melibiose for growth (Table l), they were able to oxidize the presence of ammonium acetate, strain 184= growth was melibiose according to the Biolog system test (Table 3). Similarly, none of the strains used cellobiose or mannose for constant (345 k 22 kg of protein per ml) through the temperature range from 28 to 35"C, a finding similar to that growth (data not shown), but the strains did oxidize these observed with strain ATCC 9043T (data not shown). compounds in Biolog plates (Table 3). This problem may Carbon source utilization. Differentiation of Na+-depen- have arisen because of other undisclosed basal medium dent and Na'kdependent strains was not possible on the components present in the multitest system, which perme- basis of carbon source utilization patterns. Representative abilized the cells and allowed substrate penetration and oxidation. For example, low concentrations of yeast extract have been shown to permeabilize Azotobacter vinelandii cells to antibiotics (2), and Tris buffers promote envelope disruption (4). However, cells incubated in Burk minimal salts medium appear to be unable to obtain some of these substrates for growth. The use of 0.5% melibiose in nitrogen-free medium to select for Na+-dependent strains from soil samples has not been successful, but has resulted in the selection of melibiose-utilizing Na+-independent Azotobacter chroococcurn strains. As noted by Thompson and Skerman (37), all 19 Azotobacter chroococcum strains which they examined used melibiose for growth, although they did not examine the Na+ dependence of their strains. In E. coli, melibiose transport is Na+ dependent (39), which may explain the coincidence of melibiose utilization and the Na+-dependent nutritional type. In any event, Na+ is absolutely required for the growth of the Na+-dependent strains with all of the carbon and 26 28 30 32 34 36 38 nitrogen sources tested to date (24; this study). pH range for growth. In media buffered with the buffers of TEMPERATURE ("C) Good et al. at concentrations of at least 100 mM, all of the Azotobacter strains grew optimally at pH 7.2 to 7.5, and FIG. 2. Optimum temperatures for nitrogen-fixing growth of Azotobacter salinestris sp. nov. and Azotobacter chroococcum growth was limited at pH 6.2 and 8.2 (with N, or NO,- as the ATCC 9043=. Strain 184T (0)and Azotobacter chroococcum (W) nitrogen source) or pH 9.0 (with NH,+ as the nitrogen were incubated in Burk nitrogen-free medium containing 1 mM source) (24; data not shown). NaCl and 1% sucrose. Growth conditions and analysis were as Na+-dependent acid production. Azotobacter spp. are sup- described in the legend to Fig. 1. posed to have strictly respiratory metabolism, and the for- VOL.41, 1991 AZOTOBACTER SALINESTRIS SP. NOV. 373

9 8 400 I 6oo I 500 300 7 8 c' 400 0 =L & v 200 z 300 w 63i I- 7 0 100 :200 5 100 01 - ' . I I " - I 6 0 10 20 30 40 50 4 0 1 2 3 4 NaCl (pgiml) FIG. 3. Sodium-dependent growth and acid production by Azo- PERCENT NaCl tobacter salinestris sp. nov. Strain 184= was inoculated into sodium- free Burk media containing 1% glucose, 15 mM ammonium acetate, FIG. 4. Acid production by nitrogen-fixing Azotobacter salines- and different concentrations of NaCl. The cultures were incubated tris sp. nov. at different NaCl concentrations. Strain 184= was (see the legend to Fig. l), and the amount of culture protein (0)and inoculated in sodium-free Burk nitrogen-free media containing 1% culture fluid final pH (W) were determined. glucose and different NaCl concentrations and incubated (see the legend to Fig. l), and the amount of culture protein (0)and culture fluid final pH (0)were determined. mation of acid during growth is unusual. Thompson and Skerman (37) noted that Azotobacter chroococcum strains often produced acid during incubation with carbon sub- solid media contained 20 g of CaCO, per liter and the 1 mM strates that were not used for growth. Similarly, the Na+- NaCl required for Na+-dependent strain growth, there was dependent strains produced acid in nitrogen-free medium no solubilization of the CaCO,. However, there was a thin containing the non-growth-promoting substrates mannose, (1-mm) zone of CaCO, solubilization when 0.5% NaCl was xylose, cellobiose, arabinose, and lactose at concentrations added to the growth medium, and this zone was very clear of 1% (wthol). However, the Na+-dependent strains also and larger (2 to 3 mm) when the medium contained 1to 1.5% produced acid from the growth-promoting substrates melibi- NaC1. Acid production during growth on growth-promoting ose, galactose, mannitol, sucrose, glucose, and fructose carbon sources and CaCO, solubilization are characteristics under the same conditions. The acids excreted by strain 184T that are not shared by strains of Azotobacter chroococcum have been shown to be glutamic acid, succinic acid, and (1). citric acid (25). 0-F test. When an 0-F test was performed in marine broth The production of acid by the Na+-dependent strains was inoculated with the Naf-dependent strains (Table l), marine stimulated by Na+ ions (Fig. 3). Azotobacter chroococcum broth alone became alkaline, and acid was formed in the ATCC 9043T did not acidify its growth medium in response oxidative and fermentative (marine broth plus glucose) tubes to increased NaCl availability (data not shown). Acid pro- within 7 days. Growth in all of the tubes was at the surface duction by strain 184T appeared to be the factor that limited or at the oil-water interface (in the fermentative tubes). The the growth of this strain in the presence of higher concen- acid, which extended down through 25% of the tube, was trations of NaCl. For example, N,-fixing cells grew in media reabsorbed in the oxidative test during another 7 days of containing 51.25% NaC1, and growth did not occur in the incubation. Acid continued to be formed in the fermentative presence of 1.5% NaCl, because the medium had been tubes during incubation and within 9 days filled the tubes. acidified to pH 4.50 (Fig. 4). Acid continued to be produced In contrast, the Naf -independent Azotobacter chroococ- in media containing up to 2.75% NaCl, presumably because cum strains (Table 1) produced an alkaline reaction in marine the metabolism of the cells continued in the absence of cell broth alone and failed to form significant acid after 7 days in growth (Fig. 4). Similarly, strain grew and acidified the oxidative or fermentative tubes. media containing NH,+ and 12.0% NaCl (data not shown). The reaction of the Na+-dependent strains in this 0-F test Growth ceased at an NaCl concentration of 2.25% NaCl with marine media was most similar to the reactions of the because the medium had been acidified to pH 6.03. Acid marine Vibrio spp. and Aeromonas hydrophila, although production continued, in the absence of cell growth, in these bacteria did not give an alkaline reaction in marine media containing 2.5 to 3% NaCl. broth alone (18). The upper limit of NaCl concentrations tolerated by the The 0-F test was repeated by using Burk defined medium Na+-dependent strains increased to 2.5% when testing was containing glucose and NaNO, as a nitrogen source. There done on media solidified with 1.8% agar (24). The reason for was no visible growth of the Na+-independent strains in the this is unknown but may be related to the Na+-binding fermentative tubes, whereas the Na+-dependent strains pro- capacity of agar (24). However, nitrogen-fixing cells did not duced a clearly visible band of growth below the oil and an grow well on solid medium containing this concentration of acid reaction in the medium. All strains gave an alkaline NaCl, and growth appeared to flourish only at NaCl concen- reaction in the oxidative test. The amount of NO,- left in trations that were 12%. these tubes was decreased, but NO2- and N2 were not Na+-dependent CaCO, solubilization. When nitrogen-free produced. 374 PAGE AND SHIVPRASAD INT. J. SYST.BACTERIOL.

These data suggested that the Naf-dependent strains were Azotobacter salinestris, and it is possible that this organism not respiring anaerobically , but were capable of assimilatory will be isolated from this fragile environment. NO3- reduction. The growth of the Na+-dependent strains Molecular biology. In 1987 the nearly complete sequences in, the fermentative tubes was probably a reflection of their of the 5s rRNAs from Azotobacter chroococcum 52 and lower 0, requirement and the growth-promoting effect of Azotobacter salinestris were determined by H. Van den oxygen exclusion (27). Eynde and R. De Wachter (Department of Biochemistry, Habitat. All of the Azotobacter salinestris strains isolated Universiteit Antwerpen, Antwerp, Belgium). These unpub- to date were isolated from soils. These soils had neutral to lished sequences confirmed that Azotobacter salinestris is a slightly alkaline pH values. Alberta soils tend to be saline member of the genus Azotobacter. In this study, we deter- because this geographic region, as well as the land through mined that the thermal melting point of Azotobacter salines- the center of North America from the Arctic Ocean to the tris DNA is 96.68 to 97.08"C and that the G+C content is Gulf of Mexico, was a sea bottom some 64 X lo6 years ago 65.19 to 65.98 mol%, characteristic of Azotobacter spp. (36). (32). However, the Na+ requirement of the Na+-dependent In addition, the detailed 16s rRNA sequencing and catalog- strains is only 1 mM NaCl, which means that these soils do ing conducted by Moore (Department of Biology, University nlot have to be very saline (24). Azotobacter chroococcum of Houston, Houston, Texas) have confirmed that Azotobac- strains have been isolated from saline soils in Egypt (21), and ter salinestris is a separate species of the genus Azotobacter one of the isolates obtained from the Egyptian collection was (22a). On the basis of these physiological, ecological, and Azotobacter salines- identified as Azotobacter salinestris. In the Egyptian study, molecular data, we propose the name Azotobacter chroococcum tris sp. nov. for the Na+-dependent azotobacters. the greatest numbers of strains Description of Azotobacter salinestris sp. nov. Azotobacter were isolated from soils containing 0.7 to 3.5% total soluble salinestris (sal. in. es' tris. N. L. adj. salinus, saline; L. suff. salts when an isolation medium containing 0.2 to 0.7% NaCl -estris, belonging to, living in; N. L. adj. salinestris, living in was used (21). Nitrogen fixation by strains isolated from saline environments). The data below are summarized from saline soils was most efficient when 0.05 to 0.1% NaCl was this study and previous studies (24, 25, 27, 34) which have present. Mahmoud et al. concluded that there are specific been conducted with type strain 184. The type strain does strains of Azotobacter chroococcum in saline soils (21). not form appreciable capsule polysaccharide, whereas other Dicker and Smith isolated Azotobacter spp. from salt marsh isolates may form considerable amounts of capsule. sediments in Delaware (6). These organisms were present in Gram negative. Young cells (18 h old) are elongated and significant numbers (lo7 cells per g of dry mud) in the top 1 oval with pointed ends (2 by 3 to 4 pm) and occur singly, in cm of marsh mud along with Desulfovibrio and Clostridium pairs, or chains of six to eight cells in actively growing spp. Nitrogen fixation rates were not affected by oxygen, cultures. Motile by means of peritrichous flagella (Fig. 5). indicating that there were stable aerobic (or microaerobic) Each flagellar filament has a wavelength of 2.7 to 2.9 pm and and anaerobic microenvironments within the mud samples. an amplitude of 0.36 to 0.55 pm and is 9 to 10 Fm long. Higher cell densities of Azotobacter isolates were obtained Poly-P-hydroxybutyrate is formed. Older cells are rounder, when the cultures were not shaken (7), which is character- larger (diameter, 3 to 5 pm), pleomorphic, and nonmotile istic of the oxygen-sensitive, microaerophilic nature of Azo- and form cysts. Encystment is induced in young cultures by tobacter salinestris (26). One halophilic Azotobacter isolate incubation with 0.2% P-hydroxybutyrate as the sole carbon from this marsh exhibited Na+-dependent nitrogen fixation source. and respiration (7). This strain was tentatively identified as Growth in all media is Na+ dependent. Optimal growth is an Azotobacter chroococcum strain which rapidly produced promoted by rl mM NaC1. Cells grow well in the presence acid during growth with mannitol in nitrogen-free medium of 0.25 mM Na+ and 25 mM Li+ or 25 mM Mg2+; 15 to 25 (7), which is characteristic of Azotobacter salinestris. mM K+ or 15 to 25 mM Rb+ is antagonistic to growth. It is possible that Azotobacter salinestris also will be Optimal growth requires >5 pM soluble iron. found in marine environments. The Na+-dependent growth Colonies develop on solid media after 2 to 3 days of phenotype is characteristic of marine bacteria (30). Marine incubation at 30°C. Cells are killed by 100 pM H20,. alzotobacters were described in a review of older literature Colonies are not formed in numbers proportional to the by ZoBell (40). These organisms were Azotobacter chroo- dilution of a culture because of the presence of toxic oxygen roccum-like strains that were associated with surface algae products in normal media. Capsules are absent in the type and required NaCl for their development, but their isolation strain, but may be formed by other isolates. Colonies are was sporadic. Pshenin (29) found Azotobacter strains in the brown to brownish black because of the production of a Hack Sea throughout the water column in stratified layers water-insoluble, cell-associated catechol melanin. Water- from the surface to the sediments. The number of azotobac- soluble catechol formation precedes melanization. Melaniza- ters in the open sea (>lo3 to >lo4 cells per liter) increased tion is greatest in aerated, nitrogen-free medium and is with summer blooms of plankton, which may have provided promoted by trace amounts (0.25 pg/ml) of Cu' or Cu2+. nutrients for Azotobacter growth. Significantly, it is cur- Nitrogen fixation is optimal with molybdate, but nitrogen rently believed that nutrients for marine microbial growth is fixed by molybdate-starved cells in the presence of vana- are not evenly distributed throughout the water column but date. Nitrogen fixation is optimal at 35°C and minimal at ire concentrated at freshwater-seawater interfaces in estu- 28°C. Nitrogen-fixing growth in liquid medium is initially aries (the halocline) and at stable thermocline regions of the microaerophilic, but the cells can adapt to growth in vigor- ocean (10,22,41). These zones may have increased available ously aerated medium. Cells inoculated at a low density into carbon derived from plankton decay, decreased salinity, and nitrogen-free medium or aerated vigorously before aeroad- microaerobic conditions (10, 41). Halocline layers are aptation do not grow. NO,- is assimilated without NO,- or formed where river flow into areas of low tidal amplitude is N, production. Urease is produced, large (41), a situation that Pshenin probably encountered at Catalase negative in the H20, spot test. A single, weakly the Black Sea sample sites. The growth conditions created at active catalase electromorph is present, which migrates on haloclines seem to be ideal for the nitrogen-fixing growth of nondenaturing polyacrylamide gels with the same relative VOL.41, 1991 AZOTOBACTER SALINESTRIS SP. NOV. 375

FIG. 5. Negatively stained preparation of Azotobacter salinestris sp. nov. Bar = 1 pm. mobility as beef liver catalase. Multiple peroxidase electro- Strains have been isolated from slightly saline, neutral to morphs are present, and a single superoxide dismutase is slightly alkaline soils from the Canadian provinces Alberta present. All of the enzyme activities are dependent on iron. (type strain 184), Saskatchewan, British Columbia, and the Optimal growth occurs at pH 7.2 to 7.5, and growth is Yukon and from a saline soil from Egypt. limited at pH 6.2 and 8.2 (nitrogen fixing) or pH 9.0 (with The G+C content is 65.19 to 65.98 mol%. ammonium acetate). Acid is produced during aerobic or Strain 184 is the type strain; this strain has been deposited microaerophilic growth in a sodium-dependent manner. in the American Type Culture Collection, Rockville, Md., as Higher available NaCl concentrations promote greater acid- strain ATCC 49674. ification of the medium until a growth-inhibiting pH is attained. The acid released by nitrogen-fixing cultures on ACKNOWLEDGMENTS agar-solidified medium containing 0.5 to 1.5% (wthol) NaCl is sufficient to solubilize CaCO,. The study and discovery of Azotobacter salinestris would never have occurred without the fine collection of aerobic nitrogen-fixing An alkaline reaction is produced in marine broth alone, bacteria assembled by F. D. Cook (Professor Emeritus, Department and an acid reaction is produced in marine broth containing of Soil Science, University of Alberta) and given to W.J.P. Excel- glucose under oxidative or fermentative conditions. lent technical assistance was provided by Linda Jackson, Dana Strains are very susceptible to streptomycin and kanamy- Smith, and Dena Adachi. We also thank Richard Sherburne (De- cin (inhibited by 11 pg/ml) and resistant to chloramphenicol partment of Medical Microbiology and Infectious Diseases, Univer- (40 pg/ml). sity of Alberta) for electron microscopy and Julia Foght for help The carbon substrates used for growth are the same as with the G+C content determinations. Soil samples also were those described previously for Azotobacter chroococcum, collected by Ted Cook and Phil Fedorak. including 0.25% sodium benzoate, 0.5 to 21.0% fructose, Operating funds for this study were obtained from the Natural Sciences and Engineering Council of Canada. galactose, glucose, mannitol, melibiose, starch, and sucrose. The substrates not used for growth include 0.5 to 21.0% REFERENCES arabinose, cellobiose, glutamate, glycerol, lactose, man- Becking, J.-H. 1981. The family Azotobacteraceae, p. 795-817. nose, rhamnose, and xylose. Acid is produced during incu- In M. P. Stan-, H. Stolp, H. G. Triiper, A. Balows, and H. G. bation with the growth-promoting substrates fructose, galac- Schlegel (ed.), The prokayotes. A handbook on habitats, isola- tose, glucose, mannitol, melibiose, and sucrose and with the tion and identification of bacteria, vol. 1. Springer-Verlag, non-growth-promoting substrates arabinose, cellobiose, lac- Berlin. tose, mannose, rhamnose, and xylose. Bingle, W. H. 1988. Transformation of Azotobacter vinelandii 376 PAGE AND SHIVPRASAD INT. J. SYST.BACTERIOL.

OP with a broad host range plasmid containing a cloned chro- Annu. Meet. Am. Soc. Microbiol. 1990, H251, p. 196. mosomal nif-DNA marker. Plasmid 19:242-250. 23. Page, W. J. 1983. Formation of cystlike structures by iron- 3. Butler, M. J., and M.-A. Lachance. 1987. The use of limited Azotobacter vinefandii strain UW during prolonged N,N,N' ,"-tetramethyl phenylenediamine to detect peroxidase storage. Can. J. Microbiol. 29:111&1118. activity on polyacrylamide electrophoresis gels. Anal. Biochem. 24. Page, W. J. 1986. Sodium-dependent growth of Azotobacter 162:443445. chroococcum. Appl. Environ. Microbiol. 51510-514. 4. Cho, K. Y., L. Pope, and 0. Wyss. 1974. Formation of proto- 25. Page, W. J. 1987. Iron-dependent production of hydroxamate by plasts in Azotobacter vinefandii. Arch. Microbiol. 101:337-342. sodium-dependent Azotobacter chroococcum. Appl. Environ. 5. Davis, B. J. 1964. Disc electrophoresis. 11. Method and applica- Microbiol. 53: 1418-1424. tion to human serum proteins. Ann. N.Y. Acad. Sci. 121:404- 26. Page, W. J., and S. K. Collinson. 1982. Molybdenum enhance- 427. ment of nitrogen fixation in a Mo-starved Azotobacter vinefandii 6. Dicker, H. J., and D. W. Smith. 1980. Enumeration and relative Nif- mutant. Can. J. Microbiol. 28:1173-1180. importance of acetylene-reducing (nitrogen-fixing) bacteria in a 27. Page, W. J., L. Jackson, and S. Shivprasad. 1988. Sodium- Delaware salt marsh. Appl. Environ. Microbiol. 39:1019-1025. dependent Azotobacter chroococcum strains are aeroadaptive, 7. Dicker, H. J., and D. W. Smith. 1981. Effects of salinity on microaerophilic, nitrogen-fixing bacteria. Appl. Environ. Micro- acetylene reduction (nitrogen fixation) and respiration in a biol. 54:2123-2128. marine Azotobacter. Appl. Environ. Microbiol. 42:740-744. 28. Page, W. J., and M. von Tigerstrom. 1982. Iron- and molybde- 8. Duff, J. T., and 0. Wyss. 1961. Isolation and classification of a num-repressible outer membrane proteins in competent Azoto- new series of Azotobacter bacteriophages. J. Gen. Microbiol. bacter vinefandii. J. Bacteriol. 151:237-242. 24:273-289. 29. Pshenin, L. N. 1963. Distribution and ecology of Azotobacter in 9. Eady, R. R. 1989. The vanadium nitrogenase of Azotobacter. the Black Sea, p. 383-391. In C. H. Oppenheimer (ed.), Sym- Polyhedron 8:1695-1700. posium on marine microbiology. Charles C Thomas, Publisher, 10. Fauzi, R., and C. Mantoura. 1987. Organic films at the halocline. Springfield, Ill. Nature (London) 328579-580. 30. Reichelt, J. L., and P. Baumann. 1974. Effect of sodium chloride 11. Good, N. E., G. D. Winget, W. Winter, T. N. Connolly, S. Izaea, on growth of heterotrophic marine bacteria. Arch. Microbiol. and R. M. M. Singh. 1966. Hydrogen ion buffers for biological 97: 329-345. research. Biochemistry 5:467-477. 31. Robson, R. L., and J. R. Postgate. 1980. Oxygen and hydrogen 12. Joerger, R. D., and P. E. Bishop. 1988. Bacterial alternative in biological nitrogen fixation. Annu. Rev. Microbiol. 34:183- nitrogen fixation systems. Crit. Rev. Microbiol. 16:l-14. 207. 13. Johnson, J. L. 1981. Genetic characterization, p. 457458. In P. 32. Russell, D. A. 1989. An odyssey in time. The dinosaurs of North Gerhardt, R. G. E. Murray, R. N. Costilow, E. W. Nester, A. America, p. 120-141. University of Toronto Press, Toronto, W. Wood, N. R. Krieg, and G. B. Phillips (ed.), Manual of Ontario, Canada. methods for general bacteriology. American Society for Micro- 33. Selander, R. D. A. Caugant, H. Ochman, J. M. Musser, biology, Washington, D.C. K., M. N. Gilmour, and T. S. Whittam. 1986. Methods of multilocus 14. Jurtshuk, P., Jr., J.-K. Liu, and E. R. B. Moore. 1984. Com- enzyme electrophoresis for bacterial population genetics and parative cytochrome oxidase and superoxide dismutase analy- systematics. Appl. Environ. Microbiol. 51:873-884. ses on strains of Azotobacter vinefandii and other related Shivprasad, and W. J. Page. free-living nitrogen-fixing bacteria. Appl. Environ. Microbiol. 34. S., Catechol formation and mela- Azotobacter chroococcum: 47~1185-1187. nization by Na+-dependent a pro- 15. Jurtshuk, P., Jr., T. J. Mueller, D. N. McQuitty, and W. H. tective mechanism for aeroadaptation? Appl. Environ. Micro- Riley. 1978. The cytochrome oxidase reaction in Azotobacter biol. 55: 1811-1817. vinelandii and other bacteria, p. 99-121. In H. Degn, D. Lloyd, 35. Smibert, R. M., and N. R. Krieg. 1981. General characteriza- and G. C. Hill (ed.), Functions of alternative terminal oxidases. tion, p. 413419. In P. Gerhardt, R. G. E. Murray, R. N. Proceedings of the 1lth Federated European Biochemical Soci- Costilow, E. W. Nester, W. A. Wood, N. R. Krieg, and G. B. eties, Copenhagen. Pergamon Press, Oxford. Phillips (ed.), Manual of methods for general bacteriology. 16. Jurtshuk, P., Jr., T. J. Mueller, and T.-Y. Wong. 1981. Isolation American Society for Microbiology, Washington, D.C. and purification of the cytochrome oxidase of Azotobacter 36. Tchan, Y.-T., and P. B. New. 1984. Genus I. Azotobacter vinefundii. Biochim. Biophys. Acta 637:374382. Beijerinck 1901, 567AL, p. 220-229. In N. R. Krieg and J. G. 17. Katsuwon, J., and J. Anderson. 1989. Response of plant-colo- Holt (ed.), Bergey's manual of systematic bacteriology, vol. 1. nizing pseudomonads to hydrogen peroxide. Appl. Environ. The Williams & Wilkins Co., Baltimore. Microbiol. 552985-2989. 37. Thompson, J. P., and V. B. D. Skerman. 1979. Azotobacter- 18. Lemos, M. L., A. E. Toranzo, and J. L. Barja. 1985. Modified aceae: the taxonomy and ecology of the aerobic nitrogen-fixing medium for the oxidation-fermentation test in the identification bacteria. Academic Press, Inc., New York. of marine bacteria. Appl. Environ. Microbiol. 49:1541-1543. 37a.Tsong, H., and P. F. Mathmeier. 1989. Abstr. Annu. Meet. Am. 19. Loewen, P. C., and J. Switala. 1987. Multiple catalases in SOC.Microbiol. 1989, 1-91, p. 232. Bacillus subtilis. J. Bacteriol. 169:3601-3607. 38. Vela, G. R., and 0. Wyss. 1964. Improved stain for visualization 20. Loewen, P. C., and B. L. Triggs. 1984. Genetic mapping of katF, of Azotobacter encystment. J. Bacteriol. 87:476-477. a locus that with katE affects the synthesis of a second catalase 39. Wilson, T. H., and D. M. Wilson. 1983. Sugar-cation cotransport species in Escherichia cofi.J. Bacteriol. 160:668475. systems in bacteria, p. 1-39. In E. Elson, W. Frazier, and L. 21. Mahmoud, S. A. Z., M. El-Sawy, Y. Z. Ishac, and M. M. Glaser (ed.), Cell membranes. Methods and reviews, vol. 1. El-Safty. 1978. The effects of salinity and alkalinity on the Plenum Press, New York. distribution and capacity of N,-fixation by Azotobacter in 40. ZoBell, C. E. 1946. Marine microbiology. A monograph on Egyptian soils. Ecol. Bull. (Stockholm) 26:99-109. hydrobiology, p. 156-157. Chronica Botanica Co., Publishers, 22. Mitchell, J. G., A. Okubo, and J. A. Fuhrman. 1985. Microzones Waltham, Mass. surrounding phytoplankton form the basis for a stratified marine 41. Zutic, V., and T. Legovic. 1987. A film of organic matter at the microbial ecosystem. Nature (London) 316:58-59. fresh-waterhea-water interface of an estuary. Nature (London) 22a.Moore, E. R. B., G. E. Fox, and P. Jurtshuk, Jr. 1990. Abstr. 328: 6 12-614.