Proc. Natl. Acad. Sci. USAVol. 93, pp. 3575-3580, April 1996Biochemistry

N2 fixation in marine heterotrophic bacteria: Dynamics ofenvironmental and molecular regulation

(dinitrogenase reductase/nifHDK operon/nitrogenase/in situ hybridization/ Vibrio natriegens)

JAMES A. COYER*t, ALEJANDRO CABELLO-PASINIt, HEWSON SWIFT§, AND RANDALL S. ALBERTEI*Department of Biology, University of California, Los Angeles, CA 90024; tMarine Science Institute, State University of New York, Stony Brook, NY 11794;§Department of Molecular Genetics and Cell Biology, University of Chicago, Chicago, IL 60637; and ¶Biological Science and Technology Program, Office ofNaval Research, Arlington, VA 22217

Contributed by Hewson Swift, December 6, 1995

ABSTRACT Molecular and immunological techniqueswere used to examine N2 fixation in a ubiquitous heterotrophicmarine bacterium, the facultative anaerobic Vibrio natriegens.When batch cultures were shifted from aerobic N-replete toanaerobic N-deplete conditions, transcriptional and post-translational regulation of N2 fixation was observed. Levels ofnifHDK mRNA encoding the nitrogenase enzyme were highestat 140 min postshift and undetectable between 6 and 9 h later.Immunologically determined levels of nitrogenase enzyme (Feprotein) were highest between 6 and 15 h postshift, andnitrogenase activity peaked between 6 and 9 h postshift,declining by a factor of2 after 12-15 h. Unlike their regulationin cyanobacteria, Fe protein and nitrogenase activity werepresent when nifHDK mRNA was absent in V. natriegens,indicating that nitrogenase is stored and stable under anaer-obic conditions. Both nifHDK mRNA and Fe protein disap-peared within 40 min after cultures were shifted from N2-fixing conditions (anaerobic, N-deplete) to non-N2-fixing con-ditions (aerobic, N-enriched) but reappeared when shifted toconditions favoring N2 fixation. Thus, unlike other N2-fixingheterotrophic bacteria, nitrogenase must be resynthesizedafter aerobic exposure in V. natriegens. Immunological detec-tion based on immunoblot (Western) analysis and immuno-gold labeling correlated positively with nitrogenase activity;no localization of nitrogenase was observed. Because V. natrie-gens continues to fix N2 for many hours after anaerobicinduction, this species may play an important role in provid-ing "new" nitrogen in marine ecosystems.

Biological N2 fixation is the reduction of atmospheric N2 toammonium. The reaction is catalyzed by the enzyme nitroge-nase and requires an input of energy via photosynthesis or themetabolism of organic carbon compounds. Nitrogenase andthe ability to fix N2 are present in a wide variety of eubacteriaand some methanogenic archaebacteria (1, 2). Several generaof N2-fixing heterotrophic eubacteria have been isolated frommarine sediments and have been suggested to be importantsources of "new" nitrogen in some systems (3). PlanktonicN2-fixing heterotrophic species also have been described, buttheir importance as a source of "new" nitrogen is unclearbecause of difficulties in enumeration and/or quantifying nitro-genase activity (4-6). Thus, a reliable estimate of the abundanceof heterotrophic N2-fixing bacterioplankton in the oceans, theirpotential for N2 fixation, and the environmental factors thatregulate this potential are either unavailable or obscure.

Nitrogenase is rapidly and irreversibly inactivated by 02,which particularly targets the Fe protein (dinitrogenase re-ductase) (7). Consequently, even though a diverse assemblageof heterotrophic marine bacteria with the potential to fix N2may be present, the potential can be realized only when a suitable

The publication costs of this article were defrayed in part by page chargepayment. This article must therefore be hereby marked "advertisement" inaccordance with 18 U.S.C. §1734 solely to indicate this fact.

carbon source is available and when localized 02-depletion zonesare present (4). The capacity for N2 fixation in natural populationsof marine eubacteria is beginning to be evaluated using molecularand immunological techniques (5, 6, 8).The eubacterial heterotrophic genus Vibrio is ubiquitous in

marine and estuarine environments. Most species are free-living with some forming a major portion of the surfaceplankton; others are symbiotic with teleost fishes and squid orare pathogenic in fish and shellfish (9-13). Some free-livingand symbiotic species are bioluminescent; others are capableof N2 fixation (9, 14). Little is known, however, about N2fixation or nitrogenase activity in any of the four species ofN2-fixing Vibrio, and virtually nothing is known about theirabundance (5, 14).The facultatively anaerobic and N2-fixing Vibrio natriegens is

found in marine and estuarine environments throughout theworld and has one of the shortest doubling times (9.8 min at37°C) of any bacterium (12, 14-17). As a first attempt tocharacterize and understand the regulation of N2 fixation inthis genus, we manipulated cultures of V. natriegens to evaluatethe effects of 02 and combined nitrogen availability on (i) theinduction time and stability of nifHDK mRNA and Fe proteinabundance, (ii) nitrogenase activity, and (iii) cellular abun-dance and distribution of dinitrogenase. We anticipated thatunderstanding temporal and environmental regulation of nitro-genase gene expression and enzyme activity would expand our

knowledge of the role of this genus in the ocean nitrogen cycle.

MATERIALS AND METHODSCulture of Bacteria. Batch cultures of V. natriegens (ATCC

14048) were grown at 27°C in (i) FL medium, 1 g each ofpeptone and yeast extract per liter (5) (N-replete); (ii) UR-FEmedium without peptone and yeast extract (5, 14) (N-deplete);(iii) UR-FE medium, 100 mM NH4Cl, 0.05 g of yeast extractper liter (N-enriched); or (iv) UR-FE medium, 0.01 or 0.05 gof yeast extract per liter (N-deficient).

Cells were grown in aerobic N-replete or N-enriched media,then shifted to an anaerobic environment (N2/CO2, 99:1) withN-deficient or N-deplete media. Initial cultures (100 ml) weregrown overnight in aerobic media, diluted into 1 liter of freshmedium, and grown aerobically to OD550 = 0.4-0.5. The 0 timepoint sample was collected, and the remaining cells werewashed twice in either a N-deficient or N-deplete medium,added anaerobically to a N-deficient or N-deplete medium,and grown anaerobically. Samples were collected anaerobicallyto determine cell density (OD55o), steady-state levels of nif-HDK mRNA and Fe protein, nitrogenase activity, and immu-nological reactivity.RNA Transcript Abundance. Sodium azide (final conc., 0.02

M) was added to cell samples, and RNA was extracted (18, 19).

tTo whom reprint requests should be sent at the present address:Hopkins Marine Station, Pacific Grove, CA 93950.

3575

Dow

nloa

ded

by g

uest

on

June

1, 2

020

Proc. Natl. Acad. Sci. USA 93 (1996)

Equal amounts of total RNA (20 jug) were resuspended,electrophoresed, and transferred (19-22). Prehybridizationand hybridization were at 42°C; filters were washed twice in 5 xstandard saline phosphate/EDTA (SSPE; 1 x = 180 mMNaCl/10 mM NaH2PO4, pH 7.4/1 mM EDTA)/0.2% SDS for15 min at 50°C, 1 x SSPE/0.75% SDS for 30 min at 42°C, and0.1X SSPE/1% SDS for 10 min at 42°C (19).Probe Preparation. The probe for nifHDK mRNA was a

1.9-kb restriction fragment containing nifH and a 5' portion ofnifD from Klebsiella pneumoniae double-digested from pSA30(23) with EcoRI and BamHI at 37°C for 3 h (New EnglandBiolabs). After electrophoresis on a 0.6% agarose gel, the1.9-kb restriction fragment was excised by centrifugation (24),and random-primed labeled (Prime-a-Gene, Promega) with[ca-32P]dCTP (Amersham).Western Blot Analysis. Lysates were sonicated, cleared by

centrifugation, electrophoresed in SDS/15% polyacrylamidegels, and transferred to nitrocellulose membranes (19). Poly-clonal Fe protein antiserum (raised against the protein fromAzotobacter) and goat anti-rabbit immunoglobulin conjugatedto alkaline phosphatase were the primary and secondaryantibodies, respectively. Colorimetric detection was by a stan-dard alkaline phosphatase reaction (Promega).

Nitrogenase Activity. Nitrogenase activity was assayed bythe acetylene reduction method (25, 26). A sample (5 ml) wasdrawn anaerobically after the shift from aerobic to anaerobicconditions and injected into crimp-sealed 20-ml serum bottlespreviously flushed with N2. Acetylene was generated from

A

0© calcium carbide (27) and injected (1 ml) into the sample

bottles. Duplicate experiments and anaerobic negative con-trols (no acetylene added) were conducted. Gas phase samples(50 ,Al) were withdrawn with a gas-tight syringe (Hamilton)within 1 h of incubation. Ethylene content was determinedusing a Shimadzu flame ionization gas chromatograph (modelGC-14A) equipped with a GS-Q capillary column (J & WScientific, Folsom, CA; 30 m x 0.546 mm); ethylene produc-tion rates were determined by use of a linear regression.

In Situ Detection of Fe Protein. Cells of V. natriegens werecollected in aerobic and anaerobic conditions; negative controlcells included Escherichia coli (JM109) and bakers' yeast. Allcells were fixed overnight in 4% formalin/phosphate-bufferedsaline-M (PBS-M; 10 mM Na2HPO4/10 mM NaH2PO4/200mM NaCl, pH 7.2); subsequent steps are modified fromOrellana and Perry (28). Cells were permeabilized in 100 ,ul oflysozyme at 1 mg-ml-1 in 10 mM Tris-HCl/1 mM EDTA (pH8.0) for 2 min on ice, washed, and blocked for 15 min at roomtemperature with 1% blotto [50mM Tris-HCl, pH 7.4/100 mMNaCl/5% (wt/vol) Carnation nonfat dry milk] in PBS-M.Polyclonal Fe protein (1:100) and goat anti-rabbit conjugatedto CY-2 were the primary and secondary antibodies, respec-tively. After a washing in dim light, cells were smeared ontocoated slides (29), air dried, and viewed with epifluorescentmicroscopy [Olympus, (New Hyde Park, NY) model BH2;fluorescein isothiocyanate filter set].Immunogold Localization of Fe Protein. Cells were fixed in

4% glutaraldehyde/sorbitol buffer (0.6 M sorbitol/5 mMpotassium phosphate, pH 7.8/30 mM Hepes-KOH, pH 7.8) for2 h at 4°C and resuspended in sorbitol buffer. Loosely packedcell pellets were dehydrated in ethanol and embedded inLowicryl K4M (Chemische Werke Lowi, Waldkraiburg, Ger-many). Silver sections were mounted in presoaked (0.5%bovine serum albumin/0.1 M Tris-HCl, pH 7.4) carbon-coated

A

»/.

0)

24

Time, h

B O+, N+ 0-, N- O+, N+ O-, N-

T1 60 80 100 120 140 160"40 60 80' '80 120 140 160 '

0.8 -

* UR-Fc: O.05g/1 yeast. N+. air-0.7 0 LJR-Fe: 0).()/l yeast. N-. air-

T UR-Fc: 0.01 /1 veasl. N-. air+0.6

0.5

0.4

0.3

0.1 --0

() 6 12 18 24Time. h

C 0+, N+ O-, N-

0 60 80 100 120 140 16046 kDa-

O+,N+ -,N-

0+, N+ 0-, N-

f40 60 80 80 120 140 160

B O+,N+ O-, N- 0+, N-

0 20 40 60 80 100 120 140 160 1 2dr10 20'

30 kDa -

FIG. 1. Transcription and translation of the nif operon in V.natriegens. (A) An innoculum from the final time point for each growthcurve, either aerobic N-replete (solid symbols) or anaerobic N-deficient (open symbols), was transferred to the subsequent culture.Circles (solid or open) correspond to sampling times in B and C,whereas squares indicate samples for growth determination only. (B)Northern blot analysis and detection of nifHDK mRNA. (C) Westernblot analysis and detection of polypeptides cross-reacting with Feprotein antiserum. In B and C, the time course was set to 0 upon shiftto an anaerobic N-deplete condition.

FIG. 2. Influence of aerobic and anaerobic conditions on tran-scription of the nifoperon in V natriegens. (A) An innoculum from theaerobic N-enriched culture (N+, air-) (solid circle) was transferred toan anaerobic N-deficient culture (N-, air-) (open symbols); after 3.3h, the culture was exposed to aerobic (N-, air+) (solid triangles)conditions. Circles (solid or open) and triangles in A correspond tosampling times in B; squares (solid or open) indicate samples forgrowth determination only. (B) Northern blot analysis and detectionof nifHDK The time course was set to 0 upon shift to anaerobicN-deplete conditions.

3576 Bohmsr:Cyre l

Dow

nloa

ded

by g

uest

on

June

1, 2

020

Proc. Natl. Acad. Sci. USA 93 (1996) 3577

nickel grids and floated on 20-,ul drops of polyclonal Feprotein antisera (1:10) for 1 h. Goat anti-rabbit IgG taggedwith 20-nm colloidal gold was the secondary antibody. Sectionswere washed and stained for 15 min in 1% uranyl acetate.By use of enlarged micrographs and a computer-based

imaging system (JAVA, Jandel, San Rafael, CA), cell area in theelectron micrograph of 10 cells for each treatment was deter-mined, and the number of particles within each cell wascounted by eye. Elongated cells were selected to ensureconsistent cell orientation.

RESULTSThe plasmid pSA30 contains the nifHDK operon from Kpneumoniae (23) that hybridize to nifHDK sequences fromvirtually all N2-fixing bacteria examined to date (30, 31).Southern blot analysis confirmed that a 1.9-kb restrictionfragment containing nifH and partial nifD sequences frompSA30 hybridized to a 6.3-kb EcoRI insert from pSA30 thatcarries the nifHDK operon and the EcoRI insert from pRmR2,which contains the nifHD genes from Rhizobium meliloti (21)(data not shown). Northern blot analysis demonstrated thatthe 1.9-kb nifHD insert from pSA30 bound to RNA -5 kb insize from V. natriegens actively fixing nitrogen, but not to asimilarly sized RNA from the non-N2-fixer Vibrio harveyi (datanot shown). As the nifHDKoperon inK pneumoniae is 4 kb (7),the 1.9-kb nifHD probe undoubtedly hybridized to a nifHDKmRNA in V. natriegens. The gene nifH expresses the Fe proteinsubunit of nitrogenase (detected by Western blot analysis inthe present study), whereas nifD expresses the a subunit of thenitrogenase Mo-Fe protein (7).The doubling time of V. natriegens grown in N-replete FL

medium was -1 h at 27°C, decreasing to 4-5 h when shifted

A 0.9 ,- -

0

C46 kDa -

30 kDa -

0

0.8 * UR-Fe: N+, 0.05g/l yeasto UR-Fe: N-, no yeast

0.7

0.6

0.5 /0.4 e /0.3 _

0.2 _ * o

0.1' 0

0.0 -'0 6 12

Time, h18 24

B 0 4060 801001201401806h9h12h15h

0 40 60 80 100 120 140 160 160 6h 9h 12h 15h

FIG. 3. Transcription and translation of the nif operon in V.natriegens. (A) An innoculum from the aerobic N-enriched culture(solid symbols) was transferred to the anaerobic N-deplete culture(open symbols). Circles (solid or open) correspond to sampling timesin B and C. (B) Northern blot analysis and detection of nifHDK (C)Western blot analysis and detection of polypeptides cross-reacting withFe protein antiserum. In B and C, the time course was set to 0 uponshift to the anaerobic N-deplete condition.

to an N-deficient UR-FE medium (Fig. 1A). Both growth rateswere essentially identical to a previous study of N2 fixation inV. natriegens that used the same media and temperatures (14).The nifHDK mRNA appeared between 120 and 140 min

after the shift from aerobic N-replete FL medium to anaerobicN-deficient UR-FE medium, as indicated by binding of thenifHD probe (Fig. 1B). After the cultures were shifted back toan aerobic N-enriched medium, the nifHDK mRNA disap-peared within 40 min, but reappeared between 120 and 140min after the culture was reshifted to an anaerobic N-deficientcondition (Fig. 1B).

Cultures of V. natriegens grown in aerobic N-enrichedUR-FE medium displayed a doubling time of -1 h, similar togrowth in aerobic FL medium (Figs. 2A, 3A, and 4A). Whencultures were shifted from N-enriched UR-FE medium toeither anaerobic N-deficient or anaerobic N-deplete UR-FEmedium, nifHDK mRNA appeared between 60 and 80 min(Fig. 2B) and between 100 and 120 min (Fig. 3B) after the shift.

A

0

B

IZ._'5:

cnctC)c.)0

z

._

1 .t[

0.9 L4

0.8

0.7

0.6

0.5

0.4

0.3

(.2

0.00

UR-Fe: N+. ).05g/1 yeastLUR-Fe: N-. no yeast

6 12 18

Time, h

24

0 20 100 180 6h 9h 12h 15h

r_

0^

-6

600

500

400

300

200

100

0

90

00

100

80 v

60 O40 i e

o.20 v

0 v20 20 100 180 6h 9h 12h 15h

Time

FIG. 4. Translation of the nif operon and corresponding nitroge-nase activity in V. natriegens. (A) Cultures shifted as described in Fig.3A. (B) Western blot analysis and detection of polypeptides cross-reacting with Fe protein antiserum. (C) Nitrogenase activity measuredby the acetylene reduction assay. Densitometer values (% of maxi-mum) are shown for the 35-kDa (open circles) and 37- to 38-kDa (solidcircles) polypeptides. At each time point in A represented by circles(solid or open), 5 ml of culture was withdrawn and incubated inanaerobic N-deplete conditions and 1 ml of acetylene (see text); timepoints represented by squares (open or solid) indicate samples forgrowth determination only. Error bars are one standard error. In B andC, the time course was set to 0 upon shift to the anaerobic N-depletecondition.

Biochemistry: Coyer et al.

I ,\

!

1

I0

Dow

nloa

ded

by g

uest

on

June

1, 2

020

Proc. Natl. Acad. Sci. USA 93 (1996)

The steady-state level of the message declined between 6 and9 h after the shift (Fig. 3B). Additionally, nifHDK mRNAdeclined to undetectable levels within 10 min after culturesgrowing anaerobically in a N-deficient medium were exposedto air (Fig. 2B).

Nitrogenase Fe protein antiserum resolved three polypep-tides; all three were in the range of molecular masses (35-40kDa) reported by others using nitrogenase antisera (5, 26), andall varied in time of appearance during the experiments. Onepolypeptide of -35 kDa was absent until between 100 and 120min after shifting to anaerobic conditions (Figs. 1C, 3C, and4B). Levels of the 35-kDa polypeptide increased from 120 to160 min and remained high for the period 6-15 h. Thepolypeptide was undetectable within 40 min of shifting back toan aerobic N-enriched medium, then reappeared 120 min afterreshifting to an anaerobic N-deficient medium (Fig. 1C). Thus,the 35-kDa polypeptide was present only after induction ofnifHDK mRNA, as would be expected because nifH encodesthe Fe protein subunit of nitrogenase (35 kDa; see ref. 5).Two additional polypeptides of -37 and -38 kDa cross-

reacted with Fe protein antiserum between 100 min and 15 hafter shifting from aerobic N-enriched to anaerobic N-depletemedia (Figs. 1C and 3C; merged to form one band in Fig. 4B).

During early postshift periods (0-100 min), however, the 37-kDapolypeptide was not detected, and the 38-kDa polypeptide eitherwas not detected or was present at low levels. Both, however, werepresent at lower levels after N2-fixing cells were shifted fromanaerobic to aerobic N-enriched conditions (Fig. 1C).

Nitrogenase activity appeared 100 min after the aerobic/anaerobic shift, corresponding to the first whole-cell immu-nological detection of Fe protein, as well as to detection of theprotein (35-kDa polypeptide; Fig. 4B, see also Figs. 1C and3C). Although equivalent and high levels of Fe protein weredetected from 180 min to 15 h postshift, nitrogenase activityincreased linearly to peak levels at 6-9 h, then declined twofoldfrom 12 to 15 h (Fig. 4B). No nitrogenase activity was detectedin any of the anaerobic controls (data not shown).Immunofluorescence demonstrated presence of the Fe pro-

tein antiserum in 4 and 8 h V. natriegens cells, but not in 0 hV. natriegens, E. coli, or yeast cells (Fig. 5). Immunogolddetection revealed significantly more gold particles within the4-h [mean/pxm2 (SEM) = 5.90 (0.58)] and 8-h [6.50 (0.51)]cells than within the 0-h cells [0.38 (0.06)] (one-way ANOVA,n = 10 for each cell type, df = 2, F = 56.411, P < 0.001; Tukeytest, 0 h # 4 h = 8 h; P < 0.001) (Fig. 6). Localization of goldparticles within specific areas of the cytoplasm was not appar-

FIG. 5. Immunofluorescence of formalin-fixed cells with anti-Fe protein CY-2. (A and C) Phase contrast microscopy. (B and D) Epifluorescencemicroscopy, 6-sec exposure. (A and B) Anaerobic N-deplete V. natriegens (4 h). (C and D) E. coli (JM109), bakers' yeast, and V. natriegens (anaerobicN-deplete 4 h). No signals were present in the controls (aerobic N-replete V. natriegens, 0 h) (data not shown).

3578 Bohmsr:Cyre l

Dow

nloa

ded

by g

uest

on

June

1, 2

020

Biochemistry: Coyer et al.

ent. Particles also were associated with the cell membrane ofboth N2-fixing and non-N2-fixing cells, perhaps because ofnonspecific binding of the polyclonal Fe protein antiserum tomembrane-associated proteins and/or a contaminating anti-gen. N2-fixing cells accumulated intracellular refractive gran-ules (low electron density areas 0.6-0.7 tLm in diameter) andwere smaller than non-N2-fixing cells [mean length and diam-eter and volume (/mm3, SEM): 0 h = 3.8 x 1.6 ,m, (37.9, 0.10),4 h = 2.8 x 1.2 ,tM (16.9, 0.9), 8 h = 2.6 X 1.2 ,um (15.2, 1.0);n = 10 for each cell type; one-way ANOVA for volumes, df =2, F = 110.74,P < 0.001; Tukey test, 0 h- 4 h = 8 h;P < 0.001].

DISCUSSIONThe use of both molecular probes to detect the relative levelsof nif transcripts (mRNA) and antibodies to detect the pres-ence of nitrogenase can provide a more detailed understand-ing of how N2-fixing bacteria respond to environmentalchanges and can resolve spatial and temporal dynamics of N2fixation in the sea. Such studies can also provide insights intothe coupling and interactions of N2 fixation with the biogeo-chemical cycles in the ocean.N2 fixation in heterotrophic bacteria can be regulated at

several molecular levels. For example, transcriptional regula-tion may be modulated by specific environmental factors suchas 02, combined nitrogen, and/or the availability of a carbonsource to support energetic requirements (7, 32-34). Post-translational regulation may involve the inactivation of pro-teins by degradation or by modifications that influence theabundance and/or activity of nitrogenase (for review, see refs.7 and 35). We demonstrate that both types of regulation occurin the heterotrophic, marine N2-fixing bacterium V. natriegens.Whether this is a general feature of marine heterotrophicbacteria remains to be determined.

Transcriptional control of N2-fixation potential was sug-gested by the failure to detect nifRDK mRNA after 10 min ofshifting cells from anaerobic to aerobic conditions (see Fig.2B). However, as the relative abundance of the message alsodecreased between 6 and 9 h later under anaerobic conditionsduring mid-growth phase, factors other than anoxia mayinfluence transcription.

Posttranslational control by degradation also was suggested,as a 35-kDa polypeptide cross-reacting with Fe protein anti-serum was abundant in anaerobic conditions but undetectablewithin 40 min of shifting cells to aerobic conditions. Thedisappearance of Fe protein was independent of nifHDK

Proc. Natl. Acad. Sci. USA 93 (1996) 3579

mRNA levels (and transcriptional control) because prior to theshift, Fe protein increased when nifHDK mRNA levels de-creased. Degradation or inactivation undoubtedly was due tothe presence of 02, which inactivates nitrogenase (7). Addi-tionally, the response of the 35-kDa polypeptide to the anaero-bic-aerobic-anaerobic shift indicates that it may be crucial tonitrogenase activity and serve as a more important marker forN2-fixation capacity than the 37- and 38-kDa polypeptides,both of which vary in a less predictable fashion. The 37- and38-kDa polypeptides may correspond to some of the multiplebands reported by others for the Fe protein (5, 36).The disappearance of the 35-kDa polypeptide within 40 min

of exposure to aerobic conditions and its simultaneous reap-pearance with nifHDK mRNA between 120 and 140 min aftershifting to anaerobic conditions suggests an inability to protectnitrogenase from 02 inactivation and subsequent degradation.Consequently, new nitrogenase must be synthesized after each02 exposure. In contrast, other N2-fixing heterotrophic bac-teria recover nitrogenase activity slowly (Desulfovibrio gigas)or rapidly (Azotobacter) after 02 exposure and without proteinsynthesis (37, 38). Whether an inability to protect nitrogenasefrom 02 degradation implies a primitive or recent develop-ment is not clear (39).

Cyclic synthesis of nitrogenase in a freshwater cyanobacterium(Synechococcus) required the continued presence of nitrogenasemRNA; when the mRNA was absent or transcription was inhib-ited, no nitrogenase activity was detected (40). In the presentstudy, however, three polypeptides cross-reacting with Fe proteinantiserum continued to be detected several hours after nifHDKmRNA was absent or below detectable levels. Furthermore,maximum nitrogenase activity was measured during the periodwhen nifHDK mRNA declined to undetectable levels and sig-nificant nitrogenase activity (albeit twofold lower) remained forseveral hours after nifHDK mRNA was undetectable. Thus, thestrong correlation between levels of nitrogenase mRNA andnitrogenase activity reported for cyanobacteria (40), was notobserved in V natriegens.The inability to detect nif mRNA after 9 h does not

necessarily mean that transcription of nif genes ceased. Thedata do suggest, however, that V. natriegens produced a largeamount of nitrogenase shortly after exposure to an anaerobicN-deplete environment, possibly by overexpressing nifHDKmRNA (which became diluted on a per cell basis as the cellsachieved balanced growth); synthesizing many polypeptidesper message, resulting in detectable amounts of protein beingpresent before detectable amounts of message are detected

a^y^j *~~~*

. ....

"'K..

R ',

· ,.

*.. . o

*;*..·~~~~~~~~... ..

AFIG. 6. Immunogold localization of Fe protein in V. natriegens. Cells from an aerobic N-replete culture (0 h) (A) and 4 h (B) and 8 h (C) after

shifting to anaerobic N-deplete conditions. R, refractive granules. (X10,400.)

.. c

Dow

nloa

ded

by g

uest

on

June

1, 2

020

Proc. Natl. Acad. Sci. USA 93 (1996)

(Fe protein was detected slightly before nifHDK mRNA);and/or producing nitrogenase that was very stable underanaerobic conditions.The presence of Fe protein and nitrogenase activity when

nifHDK mRNA was undetectable suggests that nitrogenase is arelatively stable protein and can be "stored" under anaerobicconditions. Consequently, V natriegens may retain N2-fixingcapacity over a broad temporal scale, and this retention mayconvey an ecological advantage when levels ofcombined nitrogenoscillate under anaerobic conditions. The presence of a relativelystable nitrogenase throughout anaerobic periods when levels ofcombined nitrogen are high may provide cells with the capacityto rapidly fix N2 if combined nitrogen levels suddenly decline.Indeed, nitrogenase is not inhibited by ammonium in vitro, butammonium can result in a rapid, but reversible, inhibition ofnitrogenase activity in some diazotrophs (7, 41).Immunogold detection revealed nitrogenase throughout the

cytoplasm of V. natriegens cells actively fixing N2, rather thansequestered in specific locations. N2-fixing cells also containedintracellular refractive granules. Similar granules have beennoted in cells from aerobic cultures of V. natriegens that werecarbon-replete, but nitrogen and/or phosphorous-starved, andare believed to contain the storage product poly-p3-hydroxy-butyrate (12, 42, 43). Carbon reserves may be formed underhigh C/P ratios (43); similarly, high C/N ratios may triggerformation of carbon reserves in N2-fixing cells.The present study has demonstrated the utility of molecular

probes to detect the presence and temporal dynamics ofN2-fixation capacity in heterotrophic marine bacteria. Refine-ment and adaptation of these approaches and techniques tonatural bacterial assemblages should advance our understand-ing of the ecological role of heterotrophic N2 fixation in marineecosystems."

We thank Dr. G. J. Smith, D. L. Robertson, and M. Moody forvalued technical assistance and constructive comments on earlierversions of the manuscript. Drs. S. Nierzwicki-Bauer (RensselaerPolytechnic Institute, Troy, NY) and D. R. Dean (Virginia Polytech-nic Institute, Blacksburg, VA) kindly provided separate gifts of Feprotein antiserum. This research was supported by the NationalScience Foundation (Grant OCE 8820304).

1. Postgate, J. R. (1982) The Fundamentals of Nitrogen Fixation(Cambridge Univ. Press, Cambridge, U.K.).

2. Sibold, L. & Souillard, N. (1988) in Nitrogen Fixation: HundredYears After, eds. Bothe, H., de Bruijn, F. J. & Newton, W. E.(Fischer, Stuttgart), pp. 705-710.

3. Capone, D. E. (1988) in Nitrogen Cycling in Coastal MarineEnvironments, eds. Blackburn, T. H. & Sorensen, J. (Wiley, NewYork), Vol. 33, pp. 86-123.

4. Paerl, H., Bebout, B. M. & Prufert, L. E. (1988) in NitrogenFixation: Hundred Years After, eds. Bothe, H., de Bruijn, F. J. &Newton, W. E. (Fischer, Stuttgart), pp. 711-716.

5. Currin, C. A., Paerl, H. W., Suba, G. K. & Alberte, R. S. (1990)Limnol. Oceanogr. 35, 59-71.

6. Kirshtein, J. D., Zehr, J. P. & Paerl, H. W. (1993) Mar. Ecol. Prog.Ser. 95, 305-309.

7. Smith, R. J. & Gallon, J. R. (1993) in Plant Biochemistry andMolecular Biology, eds. Lea, P. J. & Leegood, R. C. (Wiley, NewYork), pp. 129-153.

8. Zehr, J. D. & McReynolds, L. (1989) Appl. Environ. Microbiol.55, 2522-2526.

9. Hasting, J. W. (1978) Methods Enzymol. 57, 125-135.10. Ruby, E. G. & Nealson, K. H. (1978) Limnol. Oceanogr. 23,

530-533.

11. Ruby, E. G., Greenberg, E. P. & Hastings, J. W. (1980) Appl.Environ. Microbiol. 39, 302-306.

12. Bauman, P., Furniss, A. L. & Lee, J. V. (1984) in Bergey's Manualof Systematic Bacteriology, eds. Krieg, N. R. & Holt, J. G. (Wil-liams & Wilkins, Baltimore), Vol. 1, pp. 518-538.

13. Ortigosa, M., Esteve, C. & Pujalte, M. J. (1989) Syst. Appl.Microbiol. 12, 316-325.

14. Urdaci, M. C., Stal, L. J. & Marchand, M. (1988)Arch. Microbiol.150, 224-229.

15. Payne, W. J. (1958) J. Bacteriol. 76, 301-307.16. Payne, W. J., Eagon, R. G. & Williams, A. K. (1961)Antonie van

Leeuwenhoek J. Microbiol. Serol. 27, 121-128.17. Eagon, R. G. (1961) J. Bacteriol. 83, 736-737.18. Chomczymski, P. & Sacchi, N. (1987) Anal. Biochem. 162,

156-159.19. Smith, G. J., Zimmerman, R. C. & Alberte, R. S. (1992) Limnol.

Oceanogr. 37, 989-1007.20. Rosen, M. K. & Villa-Komaroff, L. (1990) Focus 12, 23-24.21. Tsang, S. S., Yin, X., Guzzo-Arkuran, C., Jones, V. S. & Davison,

A. J. (1993) BioTechniques 14, 380-381.22. Ruvkum, G. B., Sundaresan, V. & Ausubel, F. M. (1982) Cell 29,

551-559.23. Cannon, F. C., Riedel, G. E. & Ausubel, F. M. (1979) Mol. Gen.

Genet. 174, 59-66.24. Heery, D. M., Gannon, F. & Powell, R. (1990) Trends Genet. 6,

173.25. Hardy, R. W. F., Holsten, R. D., Jackson, E. K. & Burns, R. C.

(1968) Plant Physiol. 43, 1185-1207.26. Capone, D. G. (1993) in Handbook of Methods in Aquatic Mi-

crobial Ecology, eds. Kemp, P. F., Sherr, B. F., Sherr, E. B. &Cole, J. J. (Lewis, Chelsea, MI), pp. 621-630.

27. Postgate, J. R. (1972) in Methods in Microbiology, eds. Morris,J. R. & Ribbons, D. W. (Academic, New York), Vol. 6B, pp.342-356.

28. Orellana, M. V. & Perry, M. J. (1992) Limnol. Oceanogr. 37,478-490.

29. DeLong, E. F., Wickham, G. S. & Pace, N. R. (1989) Science 243,1360-1363.

30. Ruvkum, G. B. & Ausubel, F. M. (1980) Proc. Natl. Acad. Sci.USA 77, 191-195.

31. Scolnik, P. A. & Haselkorn, R. (1984) Nature (London) 307,289-292.

32. Bishop, P. E. & Joerger, R. D. (1990) Annu. Rev. Plant Physiol.Plant Mol. Biol. 41, 109-125.

33. Gussin, G. N., Ronson, C. W. & Ausubel, F. M. (1986) Annu.Rev. Genet. 20, 567-591.

34. Kennedy, C. & Toukarian, A. (1987) Annu. Rev. Microbiol. 41,227-258.

35. Zehr, J. (1992) in Marine Pelagic Cyanobacteria: Trichodesmiumand Other Diazotrophy, eds. Carpenter, E. J., Capone, D. G. &Rueter, J. G. (Kluwer, Dordrecht, The Netherlands), pp. 249-264.

36. Zehr, J. P., Wyman, M., Miller, V., Duguay, L. & Capone, D. G.(1993) Appl. Environ. Microbiol. 59, 669-676.

37. Drozd, J. W. & Postgate, J. R. (1970)J. Gen. Microbiol.63, 63-73.38. Postgate, J. R. & Kent, H. M. (1985) J. Gen. Microbiol. 131,

2119-2122.39. Postgate, J. R. & Eady, R. R. (1988) in Nitrogen Fixation: Hun-

dred Years After, eds. Bothe, H., de Bruijn, F. J. & Newton, W. E.(Fischer, Stuttgart), pp. 31-40.

40. Huang, T.-C., Chow, T.-J. & Hwang, I.-S. (1988) FEMS Micro-biol. Lett. 50, 127-130.

41. Vignais, P. M., Willison, J. C., Jouanneau, Y., Colbeau, A.,Magnin, J. P., Leclerc, M. & Schatt, E. (1988) in NitrogenFixation: Hundred Years After, eds. Bothe, H., de Bruijn, F. J. &Newton, W. E. (Fischer, Stuttgart), pp. 147-156.

42. Nissen, H., Heldal, M. & Norland, S. (1987) Can. J. Microbiol. 33,583-588.

43. Holmquist, L. & Kjelleberg, S. (1993) FEMS Microbiol. Ecol. 12,215-224.

3580 Bohmsr:Cyre l

Dow

nloa

ded

by g

uest

on

June

1, 2

020