APPLIED AND ENVIRONMENTAL MICROBIOLOGY, Oct. 2001, p. 4694–4700 Vol. 67, No.

10
0099-2240/01/$04.00⫹0 DOI: 10.1128/AEM.67.10.4694–4700.2001
Copyright © 2001, American Society for Microbiology. All Rights Reserved.

Response of the Endophytic Diazotroph Gluconacetobacter

diazotrophicus on Solid Media to Changes in Atmospheric
Partial O2 Pressure
BO PAN AND J. KEVIN VESSEY*
Department of Plant Science, University of Manitoba, Winnipeg, Manitoba R3T 2N2, Canada
Received 5 June 2001/Accepted 31 July 2001

Gluconacetobacter diazotrophicus is an N2-fixing endophyte isolated from sugarcane. G. diazotrophicus was

grown on solid medium at atmospheric partial O2 pressures (pO2) of 10, 20, and 30 kPa for 5 to 6 days. Using
a flowthrough gas exchange system, nitrogenase activity and respiration rate were then measured at a range
of atmospheric pO2 (5 to 60 kPa). Nitrogenase activity was measured by H2 evolution in N2-O2 and in Ar-O2,
and respiration rate was measured by CO2 evolution in N2-O2. To validate the use of H2 production as an assay
for nitrogenase activity, a non-N2-fixing (Nifⴚ) mutant of G. diazotrophicus was tested and found to have a low
rate of uptake hydrogenase (Hupⴙ) activity (0.016 ⴞ 0.009 ␮mol of H2 1010 cellsⴚ1 hⴚ1) when incubated in an
atmosphere enriched in H2. However, Hupⴙ activity was not detectable under the normal assay conditions used
in our experiments. G. diazotrophicus fixed nitrogen at all atmospheric pO2 tested. However, when the assay
atmospheric pO2 was below the level at which the colonies had been grown, nitrogenase activity was decreased.
Optimal atmospheric pO2 for nitrogenase activity was 0 to 20 kPa above the pO2 at which the bacteria had been
grown. As atmospheric pO2 was increased in 10-kPa steps to the highest levels (40 to 60 kPa), nitrogenase
activity decreased in a stepwise manner. Despite the decrease in nitrogenase activity as atmospheric pO2 was
increased, respiration rate increased marginally. A large single-step increase in atmospheric pO2 from 20 to
60 kPa caused a rapid 84% decrease in nitrogenase activity. However, upon returning to 20 kPa of O2, 80% of
nitrogenase activity was recovered within 10 min, indicating a “switch-off/switch-on” O2 protection mechanism
of nitrogenase activity. Our study demonstrates that colonies of G. diazotrophicus can fix N2 at a wide range of
atmospheric pO2 and can adapt to maintain nitrogenase activity in response to both long-term and short-term
changes in atmospheric pO2.

Gluconacetobacter diazotrophicus (47) (previously known as
Acetobacter diazotrophicus [15]) is a strict aerobe and an N2-
fixing endophyte originally isolated from sugarcane roots and
stems (6). It has been estimated that G. diazotrophicus can fix
up to 150 kg of N ha⫺1 year⫺1 in sugarcane (2). Such high
levels of N2 fixation have not been reported in any other system
outside legume-rhizobium symbioses. The bacterium has sub-
sequently been isolated from sweet potato (38), coffee (23),
pineapple (44), sorghum (22), finger millet (31), and several
other tropical grass species (24).
Aerobic endophytic diazotrophs require a high flux of O2 to
their respiratory systems to enable an adequate supply of re-
ductant and ATP to support N2 fixation (e.g., see reference
13), yet paradoxically, an excessive flux of O2 to the bacterium
can result in an inhibition of nitrogenase activity (14, 21, 26).
The inhibition of nitrogenase activity by O2 in aerobic dia-
zotrophs can be reversible or irreversible, depending on the
organism and the nature (i.e., duration and severity) of the
increase in O2 flux (33, 37, 39). Reversible inhibition of nitro-
genase activity (i.e., a temporary “switch-off” of the nitroge-
nase activity while O2 flux is excessive) can be due to a con-
formational change in nitrogenase, as seen in Azotobacter (11,
32), to an ADP-ribosylation of dinitrogenase reductase, as seen
* Corresponding author. Mailing address: Department of Plant Sci-
ence, University of Manitoba, Winnipeg, MB, R3T 2N2, Canada.
Phone: (204) 474-8251. Fax: (204) 474-7528. E-mail: k_vessey@umani
toba.ca.
in the purple nonsulfur bacteria (46) and Azospirillum (49), or
to a diversion of electrons from nitrogenase to other reduction
pathways, as proposed for Azotobacter (16, 29).
G. diazotrophicus has the ability to fix N2 at ambient atmo-
spheric partial O2 pressures (pO2) (i.e., approximately 20 kPa
of O2) in semisolid medium (6) and as colonies on solid me-
dium (10). The ability to fix N2 in colonies on solid medium is
especially interesting, as there is evidence that G. diazotrophi-
cus exists in situ in the intercellular spaces of sugarcane vas-
cular tissue as mucoid microcolonies (9). Dong (8) also re-
ported that colony morphology on solid medium and the
relative distribution of the bacteria within these highly muci-
laginous colony changed with changes in the partial pressure of
O2 surrounding the colonies.
Reis and Döbereiner (40) measured nitrogenase activity in
liquid cultures of G. diazotrophicus by acetylene reduction in
closed batch assays and found that activity was maximal when
the culture was at equilibrium with 0.2 kPa of O2 in the gas
phase. However, nitrogenase activity of G. diazotrophicus
grown in colonies on solid medium in response to changes in
atmospheric pO2 has not yet been well characterized. Given
that G. diazotrophicus exists in situ as microcolonies adhering
to plant cell walls (9), characterization of the response of the
bacterium on solid medium to changes in atmospheric pO2 is
particularly relevant.
The objective of our study was to characterize the effect of
atmospheric pO2 on nitrogenase activity of G. diazotrophicus
grown on solid medium using flowthrough gas exchange mea-

4694
VOL. 67, 2001 G. DIAZOTROPHICUS RESPONSE TO ATMOSPHERIC pO2 CHANGES
surements. Treatments included long-term growth of the bac-
terium on a range of atmospheric pO2 (10 to 30 kPa) and
subsequent rapid changes in atmospheric pO2 in small (5- to
10-kPa) and large (40-kPa) steps. We found that nitrogenase
activity by G. diazotrophicus is adaptive to both short-term and
long-term changes in atmospheric pO2 and that the bacterium
has a switch-off/switch-on mechanism for protection of nitro-
genase from rapid changes in atmospheric pO2.
MATERIALS AND METHODS
Organism and culture. G. diazotrophicus PAL-5 (ATCC 49037; obtained from
the American Type Culture Collection, Manassas, Va.) was cultured for 2 days
at 30°C, shaken at 150 rpm in LGI-P liquid medium (M. McCulley [Carleton
University], personal communication), a modified version of LGI medium (6).
LGI-P medium differs from the original LGI medium in containing 0.02 g of
Na2MoO4 䡠 2H2O liter⫺1, 0.1 mg of biotin liter⫺1, 0.2 mg of pyridaxol HCl
liter⫺1, and 5 ml of sugarcane juice (pressed from fresh sugarcane stem) liter⫺1,
and the final pH was adjusted to 5.5 using 1% acetic acid. Diluted cells were
spread on solid LGI-P agar medium (15 g of agar liter⫺1 plus 50 mg of yeast
extract liter⫺1). Before serial dilution, 5 ml of culture was vortexed with glass
beads to prevent clumping of the colonies and to obtain an even distribution of
individual separate colonies on the petri plates.
G. diazotrophicus was grown on solid LGI-P medium in petri plates for 5 or 6
days at 30°C prior to gas exchange measurements. Cultures were grown under
ambient atmospheric pO2 (approximately 20 kPa) or in a gas exchange chamber
(see below) with 10 or 30 kPa of O2 flowing through the chamber at approxi-
mately 65 ml per min. For cultures grown in the chamber, input air was bubbled
through a flask of water prior to being fed into the chamber to avoid desiccation
of the cultures.
Cell enumeration. The numbers of viable cells per colony of all petri plate
cultures used in gas exchange measurements were determined by plate counting.
Five colonies per plate were cut out of the agar, and as much agar subtending the
colonies as possible was removed without disturbing the integrity of the colonies.
These five colonies were then vortexed together in a small test tube containing 5
ml of 0.85% NaCl solution and glass beads. This suspension was then serially
diluted in 0.85% NaCl solution and plated. Cell number per plate was calculated
by multiplying the colony number per plate by the cell number per colony. On
average, colonies contained 107 to 108 viable cells each and 8.5-cm-diameter petri
plates contained 100 to 150 colonies each. No differences were noted in these
growth parameters for cultures grown at 10, 20, or 30 kPa of O2.
Gas exchange measurements of respiration rate and nitrogenase and hydro-
genase activities. A flowthrough gas exchange system (45) was used to measure
the effects of changing atmospheric pO2 on respiration rate and nitrogenase
activity of G. diazotrophicus colonies. The system includes computer-controlled
mass flow controllers (MKS Instruments Inc., Nepean, Canada) for gas mixing
and delivery, an infrared gas CO2 analyzer (ADC-225MKS; Analytical Develop-
ment Co. Ltd., Hoddesdon, United Kingdom) for measurement of respiration
rate, and an H2 analyzer (27) for measurement of nitrogenase activity. A cham-
ber with inner dimensions of 50 by 20 by 5 cm with four shelves for holding up
to 40 petri plates was constructed from 0.9-cm-thick acrylic sheeting. The void
volume when the chamber was fully loaded with cultures was 1.86 liters. Gas was
introduced at one end of the chamber, flowed horizontally across the petri plates,
and exited at the opposite end of the chamber.
Measurements of nitrogenase activity and respiration rate by G. diazotrophicus
colonies were taken at various atmospheric pO2. Gas mixtures fed into the assay
chamber were composed of various partial pressures of O2 in N2 (N2-O2) or in
Ar (Ar-O2). Respiration rate was measured as the rate of CO2 evolved from the
colonies with N2-O2 as the input gas. Nitrogenase activity was measured as H2
evolution in N2-O2 and in Ar-O2 (21, 26).
Gas exchange measurements were made in the following manner. Twenty to
40 petri plates containing 5- or 6-day-old cultures of G. diazotrophicus on solid
LGI-P medium were sealed in the chamber and then connected to the gas
exchange system. Gas mixtures were passed through the chamber at a rate of 500
ml min⫺1. For testing the response of G. diazotrophicus to small (5- or 10-kPa)
changes in atmospheric pO2, gas exchange measurements were initiated at the
pO2 under which the cultures had been grown (i.e., 10, 20, or 30 kPa of O2).
Initially, H2 and CO2 evolution was quantified at this pO2 in a gas mixture of
N2-O2. Once these readings had been made, the input gas was switched to Ar-O2
at the same pO2. Normally, 30 min to 1 h was required before the H2 evolution
came to steady state, and the H2 evolution rate was always measured 1 h after the
4695
switch from N2-O2 to Ar-O2. After the H2 evolution rate had been measured in
Ar-O2, the input stream was change back to N2-O2, but at a new atmospheric
pO2 (⫾5 or 10 kPa from the previous reading). Upon returning to an atmosphere
of N2-O2, 10 to 30 min was required for H2 and CO2 evolution rates to come to
steady state. This cycle was then repeated until nitrogenase activity and respira-
tion rate had been measured at all atmospheric pO2 (stepping down from initial
level of atmospheric pO2 to the lowest levels tested and then stepping up to the
highest levels tested). To observe the temporal response of G. diazotrophicus to
large (40-kPa), single-step increases and decreases in atmospheric pO2, gas
exchange measurements were initiated at the atmospheric pO2 under which the
bacteria had been grown (approximately 20 kPa), and then atmospheric pO2 was
increased in a single step to 60 kPa, where it remained for approximately 15 min
before being returned to 20 kPa in a single step. All gas exchange measurements
were made at room temperature (22 ⫾ 1°C). Preliminary experiments showed
that once steady-state rates of H2 and CO2 evolution were reached, they re-
mained steady for many hours (i.e., up to 12 h).
Production of H2 is an obligate reaction of the nitrogenase enzyme complex
during the fixation of N2 (4). The rate of H2 evolution in N2-O2 is a measure of
partial or “apparent” nitrogenase activity (i.e., proton reduction to H2 by nitro-
genase in the presence of N2 fixation) (21, 26). The rate of H2 evolution in Ar-O2
is a measure of total nitrogenase activity (i.e., in the absence of N2 as a substrate,
total electron flow through nitrogenase is used to reduce protons to H2) (20, 21,
26). In N2-O2, the proportion of total electron flow through nitrogenase being
directed to N2 fixation is known as the electron allocation coefficient (EAC) (12)
and is calculated as 1 ⫺ (H2 evolution in N2-O2 ⫼ H2 evolution in Ar-O2). EAC
can be viewed as a measure of an aspect of nitrogenase “efficiency” (i.e., the
higher the EAC, the greater the proportion of electrons going to fix N2 and the
lower the proportion of electrons going to the “wasteful” process of proton
reduction).
The accuracy of measuring nitrogenase activity by H2 evolution is dependent
upon a lack of hydrogenase activity leading to either H2 production or consump-
tion by the test organism under the assay conditions. Experiments with a non-
N2-fixing (Nif⫺) mutant of G. diazotrophicus (strain MAD3A) (42) were per-
formed to determine if H2 evolution from G. diazotrophicus was associated only
with nitrogenase activity and if the bacterium had hydrogenase uptake (Hup⫹)
activity. The Nif⫺ mutant was designed by insertional mutagenesis of the nifD
gene of wild-type G. diazotrophicus. The resulting mutant strain (MAD3A) was
generously donated by C. Kennedy, University of Arizona. G. diazotrophicus
MAD3A was grown and handled as described above for wild-type G. diazotro-
phicus PAL5 except that 200 ␮g of kanamycin ml⫺1 was added to the medium.
The growth rate of the Nif⫺ mutant was not significantly different from that of
wild-type G. diazotrophicus for the first 5 days of culture.
Three replicates of 40 plates each of G. diazotrophicus MAD3A were tested for
H2 production in air and Ar-O2 (80:20) in preliminary experiments in our gas
exchange system. MAD3A did not produce H2 production under any conditions
(data not shown).
Hydrogenase uptake activity by G. diazotrophicus was assessed in flowthrough
and closed-assay systems. For the flowthrough assay, MAD3A was grown for 4
days at ambient pO2 (approximately 20 kPa) as described above. Three replicates
of 20 petri plate cultures were then placed in the gas exchange chamber and
flushed continuously with air containing 2 ppm (vol/vol) of H2 at a flow rate of
20 ml min⫺1 for approximately 24 h. This concentration of H2 was used because
it is the typical level of H2 evolution from wild-type G. diazotrophicus in air under
our normal assay conditions. After exposure to 2 ppm of H2 for 24 h, the gas flow
rate was increased to 500 ml min⫺1 (the normal flow rate for our assays) and the
concentration of H2 exiting the chamber was measured. For the closed assay,
wild-type and Nif⫺ mutant strains of G. diazotrophicus were grown on solid
medium for 4 days. On the fifth day, 10 petri plate cultures were placed in the gas
exchange chamber and the chamber was flushed with 50 ppm of H2 in air at flow
rate of 20 ml min⫺1. After 24 h at this flow rate, the chamber was sealed, and
evolution (wild-type strain) and consumption (Nif⫺ strain) were monitored im-
mediately after sealing and then every 30 to 60 min for the next 6 to 8 h. Gas
samples (1 ml) were taken from the chamber and injected into an air stream
entering the H2 analyzer at a flow rate of 300 ml min⫺1 for analysis as described
by Layzell et al. (27). Hydrogen consumption and evolution rates were calculated
by linear regression. The tests were replicated four times each for the wild-type
and Nif⫺ strains of G. diazotrophicus.
The aerobic, facultative chemoautotroph Ralstonia eutropha (ATCC 17699)
was used as a positive control in the assessment of Hup⫹ activity. Early-log-phase
cells grown in Difco 0003 liquid medium (Becton Dickinson, Franklin Lakes,
N.J.) were plated onto Difco 0001 solid medium and grown for 4 days before
being assayed. H2 consumption by these colonies was assay as described above
for G. diazotrophicus MAD3A (i.e., closed assays at 50 ppm of H2 for 8 h).
4696 PAN AND VESSEY APPL. ENVIRON. MICROBIOL.

FIG. 2. Effect of atmospheric pO2 on EACs of nitrogenase of G.

FIG. 1. Effect of atmospheric pO2 on total nitrogenase activity (H2 diazotrophicus grown at 20 kPa of O2. Data are means plus standard
evolution in Ar-O2) of G. diazotrophicus colonies grown at 20 kPa of errors (n ⫽ 4). Data with different letters are significantly different at
O2. Data are means plus standard errors (n ⫽ 4). Results with different a P value of ⱕ0.05.
letters are significantly different at a P value of ⱕ0.05.

Four experiments were conducted to test the response of G. diazotrophicus to
changes in atmospheric pO2. The experiments consisted of (i) testing responses
of G. diazotrophicus grown at 20 kPa of O2 to small (5- or 10-kPa) stepped
changes in atmospheric pO2; (ii) testing responses of G. diazotrophicus grown at
20 kPa of O2 to a large (40-kPa) stepped change in atmospheric pO2; (iii) testing
responses of G. diazotrophicus grown at 10 kPa of O2 to small (5- or 10-kPa)
stepped changes in atmospheric pO2; and (iv) testing responses of G. diazotro-
phicus grown at 30 kPa of O2 to small (10-kPa) stepped changes in atmospheric
pO2. Gas exchange measurements in a single chamber containing 20 to 40 petri
plates of G. diazotrophicus cultures was considered a single replicate. For each
experiment, gas exchange measurements were replicated four times. All data
were normalized by calculating gas evolution per cell (cell number was deter-
mined for each replication of each experiment; see enumeration method above).
Data were analyzed using the general linear model of the SAS statistical package
(SAS Institute, Cary, N.C.), assuming a completely randomized design, and mean
separation was tested using the least-significant-difference procedure (P ⫽ 0.95).
RESULTS
severe decreases in respiration rates of 39 and 51%, respec-
tively.
Effects of a large (40-kPa) stepped change in atmospheric
pO2 on G. diazotrophicus grown at 20 kPa of O2. When G.
diazotrophicus colonies grown at atmospheric pO2 of 20 kPa
were exposed to a 40-kPa single-step increase in atmospheric
pO2, nitrogenase activity decreased rapidly and severely (Fig.
4). After this decrease, nitrogenase activity at 60 kPa of O2 was
steady at approximately 26% of the activity at 20 kPa of O2.
After 15 min at 60 kPa of O2, oxygen concentration was then
switched back to 20 kPa, and nitrogenase activity increased
almost immediately. Within 10 min of returning to 20 kPa of
O2, nitrogenase activity by G. diazotrophicus had recovered to
approximately 80% of the original activity. Changes in nitro-
genase activity (Fig. 4) and respiration rate (data not shown) in

Effects of small stepped changes in atmospheric pO2 on G.

diazotrophicus grown at 20 kPa of O2. For G. diazotrophicus
grown at 20 kPa of O2, 10-kPa stepped increases in atmo-
spheric pO2 above 30 kPa of O2 resulted in a decrease in total
nitrogenase activity (H2 evolution in Ar-O2) (Fig. 1). However,
nitrogenase was still active even at atmospheric pO2 of 60 kPa
(29% of the rate at 20 kPa of O2). Stepped decreases in
atmospheric pO2 from 20 to 10 to 5 kPa also resulted in
decreases in total nitrogenase activity. The optimal atmo-
spheric pO2 for G. diazotrophicus grown at 20 kPa of O2 were
20 and 30 kPa of O2.
Stepped increases of 10 kPa of O2 above 20 kPa had no
significant effect on the EAC of nitrogenase activity (Fig. 2).
However, as atmospheric pO2 was lowered from 20 kPa to10
and 5 kPa of O2, the EAC decreased.
As atmospheric pO2 was increased from 20 to 60 kPa of O2
in 10-kPa steps, the respiration rate of G. diazotrophicus cells
increased marginally (Fig. 3). For example, the threefold in-
FIG. 3. Effect of atmospheric pO2 on respiration rate (CO2 evolu-
crease in atmospheric pO2 from 20 to 60 kPa resulted in an tion in N2-O2) of G. diazotrophicus colonies grown at 20 kPa of O2.
11% increase in CO2 evolution per cell. In contrast, decreasing Data are means plus standard errors (n ⫽ 4). Data with different
atmospheric pO2 from 20 to 10 kPa and then 5 kPa resulted in letters are significantly different at a P value of ⱕ0.05.
VOL. 67, 2001 G. DIAZOTROPHICUS RESPONSE TO ATMOSPHERIC pO2 CHANGES
FIG. 4. Time course of the response of nitrogenase activity (H2
evolution in Ar-O2) of G. diazotrophicus colonies to large, sudden
changes in O2 concentration. The gas composition (top axis) was
changed from 20 to 60 kPa of O2, maintained for 15 min, then changed
back to 20 kPa of O2. The timeline of nitrogenase activity is typical of
the response to the changes in pO2. The vertical bars represent the
standard errors for actual rates of nitrogenase activity at steady state
(n ⫽ 4).
response to the single-step change from 20 to 60 kPa of O2
were similar in magnitude (i.e., a 74% decrease in nitrogenase
activity and an approximate 10% increase for respiration) to
those observed when the increase in from 20 and 60 kPa of O2
took place in several 10-kPa steps (Fig. 1 and 3).
Effects of small stepped changes in atmospheric pO2 on G.
diazotrophicus grown at 10 or 30 kPa of O2. G. diazotrophicus
colonies were grown under 10 or 30 kPa of atmospheric O2 for
5 to 6 days and then assayed for total nitrogenase activity at a
range of atmospheric pO2 (5 to 60 kPa) (Fig. 5 and 6). In both
FIG. 5. Effect of atmospheric pO2 on nitrogenase activity (H2 evo-
lution in Ar-O2) of G. diazotrophicus colonies grown at 10 kPa of O2.
Data are means plus standard errors (n ⫽ 4). Data with different
letters are significantly different at a P value of ⱕ0.05.
4697
FIG. 6. Effect of atmospheric pO2 on nitrogenase activity (H2 evo-
lution in Ar-O2) of G. diazotrophicus colonies grown at 30 kPa of O2.
Data are means plus standard errors (n ⫽ 4). Data with different
letters are significantly different at a P value of ⱕ0.05.
cases, maximal nitrogenase activity occurred at 10 to 20 kPa of
O2 above the atmospheric pO2 at which the colonies had been
grown. For colonies grown under 10 kPa of O2, nitrogenase
activity was maximized at 20 and 30 kPa of O2 (Fig. 5). For
colonies grown under 30 kPa of O2, nitrogenase activity was
maximized at 40 kPa of O2 (Fig. 6).
Hydrogenase uptake activity of G. diazotrophicus. There was
no detectable H2 consumption by the Nif⫺ mutant of G. dia-
zotrophicus (MAD3A) under typical conditions experienced by
the wild-type strain when nitrogenase activity was assayed (2
ppm of H2 in air in a flowthrough system at ambient atmo-
spheric pO2). However, when the strain was supplied with 50
ppm of H2 in air in a closed-assay system, we detected a H2
consumption rates of 0.016 ⫾ 0.009 ␮mol of H2 1010 cells⫺1
h⫺1. This rate is low, as the H2 evolution rate by wild-type G.
diazotrophicus assayed under the same conditions was 0.362 ⫾
0.027 ␮mol of H2 1010 cells⫺1 h⫺1 and H2 consumption rate by
the Hup⫹ aerobe R. eutropha was 0.080 ⫾ 0.003 ␮mol of H2
1010 cells⫺1 h⫺1.
DISCUSSION
The concentration of O2 at the sites of nitrogenase activity in
aerobic and microaerophilic diazotrophs is the result of the
interplay among (i) the concentration of O2 in the surrounding
atmosphere, (ii) the diffusion rate of O2 from the surrounding
atmosphere to the sites of nitrogenase activity, (iii) the con-
sumption rate of O2 in the vicinity of nitrogenase (predomi-
nantly via oxidative phosphorylation), and (iv) the role of car-
riers of O2 which facilitate diffusion in some systems (such as
leghemoglobin in legume nodules) (21). The present study
investigated responses in nitrogenase activity to changes in
atmospheric pO2 around colonies in a flowthrough system;
previous studies (5, 40) injected enough pure O2 into a previ-
ously anaerobic, closed liquid system to achieve target pO2. All
these studies enable observation of changes in nitrogenase
activity in response to relative changes in O2 flux to the bac-
4698 PAN AND VESSEY
teria; however, neither the actual flux of O2 to the diazotrophs
or the actual concentration of O2 at the sites of nitrogenase
activity was determined.
In our study, nitrogenase activity by G. diazotrophicus was
tolerant of atmospheric pO2 as high as 60 kPa (Fig. 1). These
findings are not in conflict with previous studies (5, 40) that
found that nitrogenase activity by G. diazotrophicus in liquid
culture was totally inhibited when the culture was at equilib-
rium with 6 kPa of O2 in the gas phase. These findings simply
reflect that in our study, atmospheric pO2 surrounding the
colonies was changed, and in the previous studies, the partial
pressure of dissolved oxygen in liquid cultures was changed.
However, comparison of these studies indicates that G. dia-
zotrophicus can use the milieu of a colony as an effective re-
sistance to O2 diffusion, resulting in an O2 concentration and
O2 flux within the colony which enable the bacteria to fix N2 in
a broad range of atmospheric pO2 surrounding the colony.
Using H2 production as a measure of nitrogenase activity in
closed-assay systems, Dong et al. (8, 10) showed that G. dia-
zotrophicus could fix N2 in colonies with 2 and 20 kPa of O2 in
the surrounding atmosphere and suggested that colony struc-
ture and location of bacteria within the colony played a role in
the protection of nitrogenase from excessive O2 flux. Bacterial
mucilage is known to decrease the rate of oxygen diffusion to
cells (3). The presence of extracellular polysaccharide sur-
rounding Beijerinckia derxii cells is necessary to maintain ni-
trogenase in this organism (1). Derxia gummosa forms small
nonfixing colonies if grown at 20 kPa of O2; however, if grown
at 5 kPa of O2, the bacterium forms large, highly mucilaginous
colonies which fix N2 (17, 18). The motile diazotroph Azospi-
rillum brasilense (50) is known to display aerotaxis within sus-
pensions to achieve the appropriate O2 environment for N2
fixation.
For colonies grown at 20 kPa of O2 and assayed at the same
atmospheric pO2, total nitrogenase activity was approximately
0.5 ␮mol of H2 1010 cells⫺1 h⫺1 (Fig. 1). Is this a relatively low
or high rate of nitrogenase activity? We have compared the
level of nitrogenase activity in G. diazotrophicus to that of
Bradyrhizobium japonicum in a typical soybean (Glycine max
[L.] Merr.) nodule. Based on measurements of nodules on
5-week-old soybean plants, Lin et al. (28) found that nodules
contained approximately 109 B. japonicum bacteroids each and
that total nitrogenase activity was in the range of 2.0 to 4.0
␮mol of H2 1010 cells⫺1 h⫺1. Nitrogenase activity for G. dia-
zotrophicus colonies at ambient atmospheric pO2 in our study
was approximately 12 to 25% of the rates calculated for B.
japonicum in soybean nodules. We consider such a level of
nitrogenase activity by G. diazotrophicus in colonies to be re-
markably high considering that a soybean nodule is a highly
sophisticated organ designed to provide a highly conducive
milieu (in terms of O2 flux, carbon supply, assimilation of fixed
N, etc.) for bacteroids to fix N2.
Our study is the first measure of EACs for G. diazotrophicus.
We found that the EAC of G. diazotrophicus at 20 kPa of O2
was approximately 0.6 (Fig. 2), meaning that in air, approxi-
mately 60% of electron flow through nitrogenase would be
allocated to reduction of dinitrogen and 40% would be allo-
cated to proton reduction. Again, the EAC of G. diazotrophi-
cus can be put into context by comparing it to that of rhizobia
in legume nodules. The theoretical maximum for EAC is 0.75
APPL. ENVIRON. MICROBIOL.
(i.e., at least one H2 produced for every N2 fixed by nitroge-
nase) (43). The EACs of legume symbioses are commonly
between 0.59 and 0.70 (21). The reason for the variability in the
EAC is not clearly understood (19, 26). Our measurements of
the EAC for nitrogenase in G. diazotrophicus grown on solid
medium at ambient atmospheric pO2 is in the same range as
EACs commonly observed in legume nodules.
The accuracy of measurements of nitrogenase activity by H2
evolution is dependent upon the lack of hydrogenase uptake
activity (21, 26). Although the Nif⫺ mutant of G. diazotrophi-
cus was seen to have a low level of Hup⫹ activity when supplied
with relatively high levels of H2 (50 ppm) in a closed-assay
system, under the standard conditions in our flowthrough assay
system (i.e., 2 ppm of H2), Hup⫹ activity was not detectable.
Small (5- to 10-kPa) decreases in atmospheric pO2 resulted
in declines in nitrogenase activity and respiration rate in G.
diazotrophicus grown at 20 kPa of O2 (Fig. 1). This is clearly
representative of an O2 limitation of cellular metabolism and
has been seen in other aerobically functional N2-fixing systems,
such as Azotobacter (48) and soybean nodules (25). However,
small stepwise increases in atmospheric pO2 above 20 kPa also
resulted in declines in nitrogenase activity (Fig. 1). The de-
clines in nitrogenase activity with small (10-kPa) increases in
atmospheric pO2 could occur for one of two reasons: (i) an
irreversible O2-induced denaturation of the nitrogenase en-
zyme or (ii) a reversible controlled down-regulation of nitro-
genase activity (14). The time course of nitrogenase activity in
response to single-step, 40-kPa changes in atmospheric pO2
(Fig. 4) indicates that the latter and not the former mechanism
is at work in G. diazotrophicus. The rapid decrease in nitroge-
nase activity when atmospheric pO2 was switched from 20 to 60
kPa, and the subsequent rapid recovery when the bacteria
returned to 20 kPa, indicate that G. diazotrophicus has a
switch-off/switch-on protection mechanism in response to
changes in atmospheric pO2.
Reversible inhibition of nitrogenase activity has been seen in
a number of diazotrophs in response to increases in pO2 and to
the addition of ammonium. Three underlying physiological
mechanisms have been associated with switch-off/switch-on ki-
netics of nitrogenase in diazotrophs. The switch-off/switch-on
mechanism can be the result of an O2-induced conformational
change in nitrogenase as seen in the Mo-dependent nitroge-
nase of Azotobacter (11, 32, 35, 36, 41). Switch-off/switch-on
mechanisms can also be facilitated by an ADP-ribosylation of
dinitrogenase reductase which halts nitrogenase activity. This
mechanism is coded for by the draT and draG genes and has
been observed in a number of diazotrophs, including
Rhodobacter capsulata (46), Rhodospirillum rubrum, Azospiril-
lum brasilense, and Azospirillum lipoferum (34, 49). Finally, the
nitrogenase switch-off/switch-on mechanism in a number of
diazotrophs may involve diversion of electrons from nitroge-
nase to other (unidentified) electron acceptors (16) and/or an
ATP limitation of nitrogenase activity, possibly due to a switch
to uncoupled respiratory chain as proposed for Azotobacter
vinelandii (29).
Which if any of the above identified switch-off/switch-on
mechanisms are at work in G. diazotrophicus was not investi-
gated in our study. However, it is highly unlikely that the
mechanism involves the ADP-ribosylation of dinitrogenase re-
ductase. Although Burris et al. (5) found that G. diazotrophicus
VOL. 67, 2001 G. DIAZOTROPHICUS RESPONSE TO ATMOSPHERIC pO2 CHANGES
had “a rather sluggish” response to ammonium addition and
required 10 ␮M NH4⫹ to switch off nitrogenase, they found no
evidence of ADP-ribosylation of dinitrogenase reductase or of
the draT-draG gene complex in G. diazotrophicus. Recently, S.
Norlund (personal communication) also found evidence of a
switch-off/switch-on phenomenon in G. diazotrophicus in re-
sponse to changes in pO2, possibly involving a conformational
change in nitrogenase medium by a Shethna-like protein (32).
In this study, long-term adaptation of G. diazotrophicus to
different atmospheric pO2 was tested by growing the bacterium
for 5 or 6 days at 10, 20, or 30 kPa of O2 before nitrogenase
activity was measured. Although culture conditions were not
exactly the same for all the cultures (see Material and Meth-
ods), some trends are consistent in all three treatments. During
assays of nitrogenase activity, when atmospheric pO2 was de-
creased below the concentration at which the bacteria were
cultured, nitrogenase activity was always lower (Fig. 1, 5, and
6). This appears to be due to a generalized O2 limitation of
cellular metabolism (Fig. 3) (25, 48). G. diazotrophicus cultures
which were grown under different atmospheric pO2 also
showed different optimal atmospheric pO2 for nitrogenase ac-
tivity. The optimal atmospheric pO2 for cultures grown at 10
and 20 kPa of O2 was 20 to 30 kPa of O2 (Fig. 1 and 5); the
optimal atmospheric pO2 for nitrogenase activity for cultures
grown at 30 kPa of O2 was 40 kPa of O2 (Fig. 6) The fact that
the cultures grown at the highest atmospheric pO2 showed a
higher optimal pO2 for nitrogenase activity indicates a long-
term adaptation of G. diazotrophicus colonies to different pO2.
Other aerobically functional N2-fixing systems such as A. vine-
landii (30) and the B. japonicum-soybean symbiosis (7) are
known to make long-term adaptations of nitrogenase activity
to nonambient pO2. Dong (8) noted differences in colony mor-
phology of G. diazotrophicus between cultures grown long-
term on 2 and 20 kPa of atmospheric pO2. We are currently
investigating whether morphologic and structural characteris-
tics of the colonies contribute to these long-term adaptations.
ACKNOWLEDGMENTS
We thank B. Luit and J. Foidart for their technical assistance, C.
Kennedy for donating the Nif⫺ strain G. diazotrophicus MAD3A, and
P. Hallenbeck, Université de Montréal, for useful discussions on this
paper.
This study was funded by grants from Cargill Inc. and the AAFC/
NSERC (Canada) Research Partnership Program.
REFERENCES
1. Barbosa, H. R., and F. Alterthum. 1992. The role of extracellular polysac-
charide in cell viability and nitrogenase activity on Beijerinckia derxii. Can. J.
Microbiol. 388:986–988.
2. Boddey, R. M., S. Urquiaga, V. Reis, and J. Döbereiner. 1991. Biological
nitrogen fixation associated with sugar cane. Dev. Plant Soil Sci. 48:105–111.
3. Brown, D. E. 1970. Aeration in the submerged culture of micro-organisms.
Methods Microbiol. 2:125–174.
4. Burris, R. H. 1991. Nitrogenases. J. Biol. Chem. 266:9339–9342.
5. Burris, R. H., A. Hartmann, Y. Zhang, and H. Fu. 1991. Control of nitro-
genase in Azospirillum sp. Dev. Plant Soil Sci. 48:89–96.
6. Cavalcante, V. A., and J. Döbereiner. 1988. A new acid-tolerant nitrogen-
fixing bacterium associated with sugarcane. Plant Soil 108:23–31.
7. Dakora, F. D., and C. A. Atkins. 1991. Adaptation of nodulated soybean
(Glycine max L. Merr.) to growth in rhizospheres containing nonambient
pO2. Plant Physiol. 9:728–736.
8. Dong, Z. 1995. The N2-fixing bacterium from the apoplast of sugarcane:
localization, isolation and characterization. Ph.D. thesis. Carleton Univer-
sity, Ottawa, Canada.
9. Dong, Z., M. J. Canny, M. E. McCully, M. R. Roboredo, C. F. Cabadilla, E.
Ortega, and R. Rodés. 1994. A nitrogen-fixing endophyte of sugarcane stems.
Plant Physiol. 105:1139–1147.
4699
10. Dong, Z., M. Heydrich, K. Bernard, and M. E. McCully. 1995. Further
evidence the N2-fixing endophytic bacterium from the intercellular spaces of
sugarcane stems is Acetobacter diazotrophicus. Appl. Environ. Microbiol.
61:1843–1846.
11. Drozd, J., and J. R. Postgate. 1970. Effects of oxygen on acetylene reduction,
cytochrome content and respiratory activity of Azotobacter chroococcum.
J. Gen. Microbiol. 63:63–73.
12. Edie, S. A., and D. A. Phillips. 1983. Effect of the host legume on acetylene
reduction and hydrogen evolution by Rhizobium nitrogenase. Plant Physiol.
72:156–160.
13. Flores-Encarnacion, M., M. Contreras-Zentella, L. Soto-Urzua, G. R. Agui-
lar, B. E. Baca, and J. E. Escamilla. 1999. The respiratory system and
diazotrophic activity of Acetobacter diazotrophicus PAL5. J. Bacteriol. 181:
6987–6995.
14. Gallon, J. R. 1992. Tansley Review no. 44. Reconciling the incompatible: N2
fixation and O2. New Phytol. 122:571–609.
15. Gillis, M., K. Kersters, B. Hoste, D. Janssens, R. M. Kroppenstedt, M. P.
Stephan, K. R. S. Teixeira, J. Döbereiner, and J. De Ley. 1989. Acetobacter
diazotrophicus sp. nov., a nitrogen-fixing acetic acid bacterium associated
with sugarcane. Int. J. Syst. Bacteriol. 39:361–364.
16. Goldberg, I., V. Nadler, and G. Hochman. 1987. Mechanism of nitrogenase
switch-off by oxygen. J. Bacteriol. 169:874–879.
17. Hill, S. 1971. Influence of oxygen concentration on the colony type of Derxia
gummosa grown on nitrogen free medium. J. Gen. Microbiol. 67:77–83.
18. Hill, S., J. W. Drozd, and J. R. Postgate. 1972. Environmental effects on the
growth of nitrogen-fixing bacteria. [Azotobacter chroococcum]. J. Appl.
Chem. 22:541–558.
19. Hunt, S., S. T. Gaito, and D. B. Layzell. 1988. Model of gas exchange and
diffusion in legume nodules. II. Characterization of the diffusion barrier and
estimation of the concentration of CO2, H2, and N2 in the infected cells.
Planta 173:128–141.
20. Hunt, S., B. J. King, and D. B. Layzell. 1989. Effects of gradual increases in
O2 concentration on nodule activity in soybean. Plant Physiol. 91:315–321.
21. Hunt, S., and D. B. Layzell. 1993. Gas exchange of legume nodules and the
regulation of nitrogenase activity. Annu. Rev. Plant Physiol. Plant Mol. Biol.
44:483–511.
22. Isopi, R., P. Fabbri, M. Del-Gallo, and G. Puppi. 1995. Dual inoculation of
Sorghum bicolor (L.) Moench ssp. bicolor with vesicular arbuscular mycor-
rhizas and Acetobacter diazotrophicus. Symbiosis 18:43–55.
23. Jimenez-Salgado, T., L. E. Fuentes-Ramirez, A. Tapia-Hernandez, M. A.
Mascarua-Esparza, E. Martinez-Romero, and J. Caballero-Mellado. 1997.
Coffea arabica L., a new host plant for Acetobacter diazotrophicus, and iso-
lation of other nitrogen-fixing acetobacteria. Appl. Environ. Microbiol. 63:
3676–3683.
24. Kirchhof, G., V. M. Reis, J. I. Baldani, B. Eckert, J. Döbereiner, and A.
Hartmann. 1997. Occurrence, physiological and molecular analysis of endo-
phytic diazotrophic bacteria in gramineous energy plants. Plant Soil 194:45–
55.
25. Kuzma, M. M., H. Winter, P. Storer, I. Oresnik, C. A. Atkins, and D. B.
Layzell. 1999. The site of oxygen limitation in soybean nodules. Plant
Physiol. 119:399–407.
26. Layzell, D. B. 1990. N2 fixation, NO3⫺ reduction and NH4⫹ assimilation, p.
389–406. In D. T. Dennis and D. H Turpin (ed.), Plant physiology, biochem-
istry and molecular biology. Longman Scientific and Technical, Harlow,
Essex, United Kingdom
27. Layzell, D. B., G. E. Weagle, and D. T. Canvin. 1984. A highly sensitive, flow
through H2 gas analyzer for use in nitrogen fixation studies. Plant Physiol.
75:582–585.
28. Lin, J., K. B. Walsh, D. T. Canvin, and D. B. Layzell. 1988. Structural and
physiological bases for effectivity of soybean nodules formed by fast-growing
and slow-growing bacteria. Can. J. Bot. 66:526–534.
29. Linkerhägner, K., and J. Oelze. 1995. Cellular ATP levels and nitrogenase
switchoff upon oxygen stress in chemostat cultures of Azotobacter vinelandii.
J. Bacteriol. 177:5289–5293.
30. Linkerhägner, K., and J. Oelze. 1997. Nitrogenase activity and regeneration
of the cellular ATP pool in Azotobacter vinelandii adapted to different oxygen
concentrations. J. Bacteriol. 179:1362–1367.
31. Loganathan, P., R. Sunita, A. K. Parida, and S. Nair. 1999. Isolation and
characterization of two genetically distant groups of Acetobacter diazotrophi-
cus from a new host-plant Eleusine coracana L. J. Appl. Microbiol. 87:167–
172.
32. Maier, R. J., and F. Moshiri. 2000. Role of the Azotobacter vinelandii nitro-
genase-protective Shethna protein in preventing oxygen-mediated cell death.
J. Bacteriol. 182:3854–3857.
33. Marchal, K., and J. Vanderleyden. 2000. The “oxygen paradox” of dinitro-
gen-fixing bacteria. Biol. Fertil. Soils 30:363–373.
34. Masepohl, B., R. Krey, and W. Klipp. 1993. The draTG gene region of
Rhodobacter capsulatus is required for the post-translational regulation of
both the molybdenum and the alternative nitrogenase. J. Gen. Microbiol.
139:2667–2675.
35. Moshiri, F., J. W. Kim,. C. Fu, and R. J. Maier. 1994. The FeSII protein of
Azotobacter vinelandii is not essential for aerobic nitrogen fixation, but con-
4700 PAN AND VESSEY
fers significant protection to oxygen-mediated inactivation of nitrogenase in
vitro and in vivo. Mol. Microbiol. 14:101–114.
36. Moshiri, F., B. R. Crouse, M. K. Johnson, and R. J. Maier. 1995. The
“nitrogenase-protective” FeSII protein of Azotobacter vinelandii: overexpres-
sion, characterization, and crystallization. Biochemistry 34:12973–12982.
37. Oelze, J., and G. Klein. 1996. Control of nitrogen fixation by oxygen in
purple nonsulfur bacteria. Arch. Microbiol. 165:219–225.
38. Paula, M. A., S. Urquiaga, J. O. Siqueira, and J. Döbereiner. 1992. Syner-
gistic effects of vesicular-arbuscular mycorrhizal fungi and diazotrophic bac-
teria on nutrition an growth of sweet potato (Ipomoea batatas). Biol. Fertil.
Soils 14:61–66.
39. Poole, R. K., and S. Hill. 1997. Respiratory protection of nitrogenase activity
in Azotobacter vinelandii—roles of the terminal oxidases. Biosci. Rep. 17:
303–317.
40. Reis, V. M., and J. Döbereiner. 1998. Effect of high sugar concentration on
nitrogenase activity of Acetobacter diazotrophicus. Arch. Microbiol. 171:13–
18.
41. Scherings, G., H. Haaker, H. Wassink, and C. Veeger. 1983. On the forma-
tion of an oxygen-tolerant three component nitrogenase complex from Azo-
tobacter vinelandii. Eur. J. Biochem. 135:591–599.
42. Sevilla, M., D. Meletzus, K. Teixeira, S. Lee, A. Nutakki, I. Baldani, and C.
Kennedy. 1997. Analysis of nif and regulatory genes in Acetobacter diazotro-
phicus. Soil Biol. Biochem. 29:871–874.
APPL. ENVIRON. MICROBIOL.
43. Simpson, F. B., and R. H. Burris. 1984. A nitrogen pressure of 50 atmo-
spheres does not prevent evolution of hydrogen by nitrogenase. Science
224:1095–1097.
44. Tapia-Hernandez, A., M. R. Bustillos-Cristales, T. Jimenez-Salgado, J. Ca-
ballero-Mellado, and L. E. Fuentes-Ramirez. 2000. Natural endophytic oc-
currence of Acetobacter diazotrophicus in pineapple plants. Microb. Ecol.
39:49–55.
45. Vessey, J. K. 1992. Cultivar differences in assimilate partitioning and capacity
to maintain N2 fixation rate in pea during pod-filling. Plant Soil 139:185–194.
46. Yakunin, A. F., and P. C. Hallenbeck. 1998. Short-term regulation of nitro-
genase activity by NH4⫹ in Rhodobacter capsulatus: multiple in vivo nitro-
genase responses to NH4⫹ addition. J. Bacteriol. 180:6392–6395.
47. Yamada, Y., K. Hoshino, and T. Ishikawa. 1997. The phylogeny of acetic acid
bacteria based on the partial sequences of 16S ribosomal RNA: the elevation
of the subgenus Gluconoacetobacter [sic] to generic level. Biosci. Biotechnol.
Biochem. 61:1244–1251.
48. Yates, M. G., E. M. deSouza, and J. H. Kahindi. 1997. Oxygen, hydrogen and
nitrogen fixation Azotobacter. Soil Biol. Biochem. 29:863–869.
49. Zhang, Y., R. H. Burris, P. W. Ludden, and G. P. Roberts. 1997. Regulation
of nitrogen fixation in Azospirillum brasilense. FEMS Microbiol. Lett. 152:
195–204.
50. Zhulin, I. B., V. A. Bespalov, M. S. Johnson, and B. L. Taylor. 1996. Oxygen
taxis and proton force in Azospirillum brasilense. J. Bacteriol. 178:5199–5204.