NITROGEN
FIXATION
BY BACTERIA
IN LAKE MIZE,
FLORIDA,
AND IN SOME LACUSTRINE
SEDIMENTS’
Michael A. Keirn and Patrick L. Brexonik
Department
of Environmental
Engineering,
University
of Florida,
Gainesville
32GOl
ABSTRACT
Nitrogen fixation was measured in the water column of Lake Mize, Florida,
a highly
colored, small, deep lake, and in a variety
of lacustrine
sediments, by the acetylene
A consistent seasonal and vertical pattern of fixation occurs in Lake
reduction technique.
Mize with positive rates noted only for a relatively
short period during summer stratificaMaximum rates (up to an cquivalcnt
of 3.26 ,ug
tion and only in the anoxic hypolimnion.
N/liter-hr)
were at depths of %lO m (D,,, = 25 m) and three bacterial cultures capable
of fixing nitrogen were isolated from the waters:
a heterotroph
characteristic
of the genus
Clostridium
and the purple sulfur bacteria Thiospirillum
and Chromutium.
Available
evidence suggests that fixation is primarily
heterotrophic.
Acetylene reduction
was detected
in sediments of 7 out of 25 Florida lakes and in sediments from 3 Guatemala lakes. Rates
decreased with depth in 30-50-cm cores. High concentrations
of sucrose stimulated acetyand pyruvate
did not.
lene reduction
in lake sediment, but glucose, acetate, butyrate,
These results indicate that bacterial fixation in aquatic environments
is more widespread
and significant
than previously
thought.
INTRODUClXON
Recently, Brczonik and Harper (1969)
and Stewart (1969) presented evidence
that nitrogen fixation occurs in anoxic watcrs of lakes and fjords. Brooks ct al.
( 1971) found low but measurable rates of
nitrogen fixation in the upper layers of
cstuarinc sediments, and Howard et al.
(1970) found low rates in Lake Erie
sediments, thus extending the known distribution
od nitrogen fixation in aquatic
habitats.
Fixation rates measured in anoxic environmcnts have thus far been low compared to rates of cyanophyccan fixation in
euphotic waters. Using the acctylenc rcduction method of Stewart et al. ( 1967),
Brczonik and Harper (1969) rcposrted maximum nitrogen fixation rates as fractions
of a pg N/liter-hr
in two anoxic lake waters compared to maximum algal fixation
rates of several pg N/liter-hr
in a variety
of lakes (Dugdalc
and Dugdale 1962;
Goering and Ncess 1964; Billaud 1968).
1 This investigation
was supported
in part by
Federal Water Quality
Administration
Research
Grant DGK 16010 to P. L. Brczonik
and by a
University
of Florida Biomedical
Sciences support
grant.
LIMNOLOGY
AND
OCEANOGRAPHY
Low fixation rates in anoxic waters may
supply only a small portion of the nitrolgen
requirements of the biota at a given time.
However, the long-term significance may
be greater; extrapolation
of the fixation
rate in Lake Mary, Wisconsin (Brezonik
and Harper 1969), indicated an annual
contribution
of 4-5 kg of nitrogen to the
lake.
Reported fixation rates in sediments are
also low. Brooks et al. (1971) report a mean
of 3.7 ng N fixed/g sediment-hr in the
surface scdimcnts of the Waccasassa cstuary, Florida. This amounts to an annual
input of 3 x 10” kg N to the sediments of
the 7-km” estuary, again illustrating
that
seemingly low rates of fixation may, in
fact, be geochemically significant. Conversion of ethylene production rates reported
by Howard et al. (1970) to equivalent nitrogen fixed indicates a rate in Lake Eric
sediments of about 2.7 ng N/g scdimcnthr, which is in the range of the Waccasassa
estuary rates. However, s-day incubations
in the former study vs. 1 hr in the latter
limit comparability
elf the data.
A variety of organisms may fix nitrogen
in anoxic natural waters and sediments.
Stewart (1969) lists 15 genera in the orders
Eubactcriales
and Pscudomonadalcs
for
720
SEPTEMBER
1971,
V. 16(S)
NITROGEN
FIXATION
which there is good evidence that some
strains fix nitrogen.
Biochemical s’tudics
of bacterial nitrogen fixation have used
cultures of Axotobacter and CZostricEiz~m
extensively,
but much remains to be
lcarncd about the ecological significance
of these groups. Even less is known about
the distribution
and significance of other
bacteria. Evidence for fixation by yeasts,
actinomycetes and fungi is highly controvcrsial. The prescncc of acrolbic, anaernitro’gen-fixing
obic, and photosynthetic
bacteria in aquatic environments is well
established ( e.g., Waksman et al. 1933;
Pshcnin 1959, 1963; Triipcr and Gcnovese
1968). However, the presence of bacteria
capable of nitrogen fixation dots not necessarily mean that fixation is occurring in
the environment.
The occurrence of nitrogen fixation in
Lakes Mary (Wisconsin) and Mizc (Florida) (Brezonik and Harper 1969) under
anoxic and cxtrcmcly low light conditions
suggests heterotrophic fixation or perhaps
fixation by photosynthetic
bacteria.
No
blue-green algae were found in water samples from these lakes. Stewart (1969) correlated significant
uptake oS 15N2 at a
depth of 7 m in a Norwegian fjolrd with
the presence of the photosynthetic
bactcBrooks ct al. ( 1971)
rium Pelodictyon.
isolated Clostridium-like
bacteria capable
of nitrogen fixation from nitrogen-fixing
sediments of the Waccasassa estuary; fixation is certainly heterotrophic in such scdiments since it occurs below the surface
layer.
This paper describes the annual time
course of nitrogen fixation in the anoxic
depths of Lake Mizc, a small dystrophic
lake in northern Florida, and presents evidence on the nature of the agents of fixation, Secondly, nitrogen fixation is shown
to be a fairly common phcnolmenon in scdimcnts from lakes in Florida and Guatemala. WC acknowledge the assistance of
C. IIarpcr in the initial phases of the
Lake Mizc study. A bathymetric map of
Lake Mize was supplied by F. Nordlic; R.
Yorton pcrformcd the chemical analyses;
and P. H. Smith (Department of Bactcriol-
IN
LAKE
721
MIZE
ogy) supplied advice concerning
of photosynthetic bat tcria.
MATERIALS
AND
isolation
METHODS
The acetylene reduction technique (Stcwart et al. 1967) was used to measure rates
of nitrogon fixation. The basic technique
used on water samples was described by
Brezonik and Harper (1969) and on scdiments by Brooks et al. ( 1971). Sediment
and anoxic water samples were purged
with helium; water samples with dissolved
oxygen were purged with a gas mixture
( 20% 02, 0.03% COZ, balance Ar) . Samplcs were incubated at 22C in a waterbath-shaker, and ‘ethylene was analyzed
with a Varian-Aerolgraph
600D gas chromatograph with a hydrogen flame ioaization detector and a ‘J/ inch X 6 ft (0.32
cm x 1.8 m ) Porolpak R column. Sedimcnts and anoxic water samples were incubated in the dark; aerobic samples were
incubated under daylight-type
fluorcsccnt
lighting.
Controls run with each set of
samples were carried through the identical
procedure except that 1 ml of 50% TCA
was added before acetylene addition.
Water samples for routine chemistry
were taken by Van Dorn sampler; bacteriological samples from the depths were
taken aseptically in BOD bosttlcs attached
to a special apparatus which opens the
bottle from the surface at the pull of a
string. Scdimcnt cores 30-50 cm long were
collected in Plexiglas tubes.
Acidity,
alkalinity,
and pH were run
potentiometrically
in the laboratory within
a few hours after collection of samples
in BOD bottles to avoid coatact with
air. Dissolved oxygen was determined by
the Winkler-azide
method (Amer. Public
Health Ass. 1965). Ammonia, nitrate, and
orthophosphate were mcasurcd on a Tcchnicon AutoAnalyzer,
using modifications
of the phenol-hypochlorite
method for ammonia (Tcchnicon Corp. 1969), the brucinc
method for nitrate (Kahn and Brezenski
1967)) and the Murphy and Riley ( 1962)
method for ortholphosphatc.
Total phosphatc was also measured by the Murphy
722
TABLE
MICHAEL
1.
Chemical
A.
KEXRN
of Lnke Mixe,
cha.racte&tics
Florida”
5.03
29.4
1.7
52.9
434
1.4
10.29
5.0
6.7
0.32
0.94
3.4
2.4
19
0.02
3.19
68.6
77
2
PH
Acidity
( mg CaCO&iter
)
Alkalinity
(mg CaCO,/liter)
Specific conductance
(,umho cm-l)
Color (mg Pt/liter
at pH 8.3)
Turbidity
( JTU )
Chloride
( mg/liter )
Sulfate (mg/liter)
Na (mg/liter)
K (mg/liter)
Mg (mg/liter)
Ca ( mg/liter )
Fe ( mg/liter )
Mn ( yg/liter )
F (mg/liter)
SiOa (mg/liter)
COD (mg/liter)
To.tal solids (mg/liter)
Suspended solids (mg/liter)
* Average of data collected
1970 (Shannon 1970).
from
AND
June
1968
to June
and Riley method after autoclaving at 1.02
atm for 1 hr with persulfate and sulfuric
acid. Analyses for total organic nitrogen
and ammonia in sediments were made on
slurries of fresh sediment samples using
micro-Kjcldahl
and distillation
methods.
Sedimentary total phosphate was run on
simikar samples using the persulfatc oxidation and Murphy and Riley procedures.
Volatile solids were measured according
to Standard methods (Amer. Public Health
Ass. 1965).
Bacteriological
samples were collected
aseptically at the same depths assayed for
nitrogen fixation on several occasions, and
samples from those depths where fixation
occurred were subjected to enrichment
culture for groups of nitrogen-fixing
microorganisms. Three enrichment schcmcs
were used to cover the spectrum of likdy
microbial N-fixing
agents : anaerobic or
facultative
heterotrophic
bacteria, photosynthetic bacteria, and yeasts and fungi.
Heterotrophic
nitrogen-fixing
bacteria,
both aerobic and anaerobic, were enriched
initially
using a nitrogen-free,
modified
Winogradsky’s medium (Grau and Wilson
1962) with sucrose as the carbon source.
Aliquots from each depth were mixed with
PATRICK
L.
BREZONIK
equal amounts of double strength medium
and in.cubated at 2OC both aerobically and
anoxically under 1 atm of Na, Control aliquots were carried through the same procedurc except that 1.0 g/liter NH&l
was
present in this medium. After growth was
noted by increased turbidity,
transfers
were made to fresh nitrogen-free media.
After three transfers, streak plates were
made on to the same medium plus 1.5%
agar.
Enrichment for photosynthetic
bacteria
was done with three media. Athiorhodaceac (nonsulfur purple bacteria) were enriched by placing 100 ml of lake water
from- each depth into a BOD bottle and
filling with a medium consisting of sodium
acetate (2 g/liter),
NH&l
( 1 g/liter),
&HP04
(0.5 g/liter), MgC12 (0.1 g/liter),
yeast extract (0.05 g/liter),
and adjusted
to pH 7.0. The bottles were stoppered and
incubated at 25C under 2,690 lux of continuous fluorescent light. Enrichment for
Thiorhodaceae
(green and purple sulfur
bacteria) was accomplished by inoculating
200 ml of lake water into each of two BOD
bottles containing 100 ml of a sterile slurry
of cellulose (3 g), CaS04 (3 g), NH&l
(0.139 g), KHzP04 (0.33 g), and Na2S.
7H20 (0.07 g), one adjusted to pH 7.3
(for green sulfur bacteria) and the other
adjusted to pH 8.5 (for purple sulfur
bacteria).
The bottles were capped and
incubated as described abo’ve for the Athiorhodaccae. Controls, consisting of sediment samples from a lake known to contain
these groups (Lake Alice, Gainesville,
Florida: Lackey et al. 1965), were carried
through the procedure to assure that the
schcmc was valid.
Enrichment for nitrogen-fixing
yeasts or
fungi was made by filtering aliquots of
water through 0.45-pm mcmbranc filters
which were incubated at 25C and 35C on
Sabouraud dextrose agar. All morphololgical types (about 30% of the total colonies
or 65 isolates) were picked after 4 days
growth, Gram stained, and inoculated into
a nitrogen-deficient
medium modified following Metcalfc and Brown ( 1957); this
NITROGEN
FIXATION
IN
LAKE
723
MIZE
6
3
6
18
FIG.
1.
Temperature
profiles
contained 0.3% sodium benzoatc; glucose,
sucrose, and mannitol glucose as carbon
sources at 5.0 g/liter; phosphate, and tract
metals and was buffered to pH 7.2. The
isolates were also inoculated into a coatrol
medium (i.e., the above medium plus nitrogen as NH&l).
All cultures wcrc incubated in the dark at 25C.
RESULTS
General description of Lake Mixe
Lake Mize is a small (0.86 ha) coneshaped lake in a pint forest about 24 km
northeast of Gainesville, Florida. Its maximum depth is abolut 25 m, and its morphometry coupled with the wind protcction of the surrounding forest allows the
lake to stratify thermally from February
or March to October or early November.
Table 1 summarizes the general. chemical
characteristics of the lake. The lake has a
in Lake Mize,
1969-1970.
high but variable color, evidently leached
from pine needle litter in the drainage
basin, and has typical characteristics
elf
dystrophy:
low values of pH, calcium,
and primary productivity.
Figures 1 and 2 present temperature and
dissolved oxygen profiles of the lake folr
1969 and 1970. In 1969 the lake was only
slightly thermally
stratified in February
but a marked depletion of disso’lved oxygcn was found in the hypolimnion.
By
April anoxic conditions had dcvelopcd bclow 7 m and by July there was no oxygen
below 3 m. Stratification began somewhat
later in 1970, but dissolved oxygen was
partially
dcplctcd in the bottom waters
during February.
Anoxic conditions occurred below 5 m by early April and below
3 m by early June; dissolved oxygen values
were low even in surface waters during
summer 1970.
724
MICHAEL
FIG.
2.
Dissolved
A.
KEIRN
oxygen
AND
profiles
Conditions in Lake Mize are now showing the cffccts of an increased nutrient
loading that began in fall 1968. An enclosure housing about 50 captured mallard
ducks has been located since then at the
north shore of the lake. In response to
enrichment, a lush growth of emergent
grass has occurred along the previously
nearly bare shorclinc and the rate of primary production has increased markedly.
Rates ranging from 1.9-10.4 mg Cjm3-hr
were found before 1970. Samples taken
in April and June 1970 yielded rates of
42.3 and 36.2.
Concentrations
of nitrogen and phosphorus have also increased substantially
over the past year (Table 2) presumably
Conccntrareflecting
duck enrichment.
tions of ammonia wcrc low throughout the
PATRICK
L.
in Lake
BREZONIK
Mizc,
1969-1970.
water column in spring 1969 and an inverse clinogradc distribution
occurred in
summer. By Scptembcr 1969 @limnetic
ammonia concentrations
surpassed those
in the bottom water. Concentrations wcrc
higher throughout
1970 and generally
showed more complex profiles. Total organic nitrogen was rather uniform throughout the water column during the 2 years,
and nitrate levels remained lo,w ( typically
<0.05 mg N/liter)
but somewhat variable
during the study.
Nitrogen
fixation
rates in Lake Mixe
Nitrogen fixation as mcasurcd by the
acetylene reduction
method has been
summer stratification
in
found during
1968, 1969, and 1970. Figures 3 and 4
NITROGEN
TABLE
2.
Nutrient
FIXATION
chemistry
of Lake
1969”
Parameter
Top 3, m
Total org-N (mg N/liter)
N&-N
(mg N/liter)
Org-PO& (mg P/liter)
Total-Pod
(mg P/liter)
0.82
0.03
0.018
0.087
* Collected 11 June 1969.
t Average of 3 samples taken during
IN
LAKE
Mixe
725
MIZE
during
June 1969 and 1970
(composite)
197OT (composite)
Bottom
15 m
Top 3 m
0.66
0.07
0.030
0.059
Bottom
15 m
1.02
0.28
0.10
0.19
1.35
0.50
0.0158
0.15
June 1970.
illustrate the relationship oE fixation rate
to depth during the fixation periods in
1969 and 1970. Fixation occurred only in
the hypolimnion during both years. Rates
were highest during summer 1969, and the
depth at which peak fixation occurred
bccamc greater over the course of that
summer. This tendency was not found
during 1970, when fixation started later
than in 1969 and ended sooner. Why a
shorter period of fixation occurred in 1970
is not completely clear; however the nutrient cnrichmcnt may have been a factor.
Nitrogen fixation in Lake Mizc is apparently restricted to midsummer stratification. No fixation was detected in samples
assayed in April, Scptcmbcr, October, and
December 1969 nor in February, April,
and May 1970. The annual cycle of nitrogcn fixation for 1969 and 1970 is shown
in Fig. 5. It is apparent that fixation was
much greater during 1969 both in duration
and rate. The values for total nitrogen fixation per hour were cstimatcd by summing
the product of the volumetric fixation rate
at each depth times the calculated lake
volume for each depth.
Bacterial
isolations
Enrichment tcchniqucs indicate that at
least two nitrogen-fixing
groups of bacte0
3
6
11 Jun 69 0
12
Jun 70
l
18
Jun 70
Sun 70
0
A
9 Jul 70
A
30
15
15 Jul 69 .
30 Jul 69 l
18
I
1
0
1.00
2.00
NITROGEN FIXATION
(ugN/liter-hr)
I?IG.
Lake
1
3
3.00
4.00
I
I
I
0.50
1.00
NITROGEN
RATE
3. Depth profiles of nitrogen
Mize, summer 1969.
I
0
FIXATION
J
1.50
RATE
(up N/liter-hr)
fixation
in
FIG.
Lake
4. Depth profiles of nitrogen
Mizc, summer 1970.
fixation
in
726
MICHAEL
A.
KEIRN
AND
PATRICK
TABLE
L.
3.
BREZONIK
Lakes surveyed for sediment
fixation* t
nitrogen
Hypereutrophic
Apopka ( B )
Unnamed 20 ( D )
Bivins Arm (A)
Griffin
(B )
Alice (A)
Eustis ( B )
Kanapaha (A)
Eutrophic
Hawthorne
(A)
Clear (B)
Wauberg (A)
Newnans (A)
FIG.
fixation
1970.
5. Time course of total hourly nitro,gen
at all depths in Lake Mize, 1969 and
ria-one
heterotrophic,
the other autotrophic-exist
in the depths of Lake Mize.
Six isolates elf Gram-positive
spore-forming rods that grew anaerobically but not
aerobically on nitrogen-free
media were
isolated frolm water samples taken at 3-, 5,
and 7-m depths on 18 June and 5 and 7 m
on 9 July. Seven transfers of the anoxic
isolates were made, each time to nitrogenfree media, and tests for acetylene rcduction were positive at each step. Cultures
of the six isolates grown in liquid media
Ear 24 hr gave acetylene reduction rates in
the range 1.5-10 nmolc C&l2 reduced/mg
organism N-hr.
Colonies of purple sulfur bacteria appcarcd in a water sample from 7 m after
4 weeks incubation.
Growth began at the
glass cellulose interface, hampering obscrvation of morphological
types. Transfers
of pigmented colo,nies were made into
liquid media containing
only NazS and
other inorganic salts at “pH 8.5 and incubated in the light. Phase contrast microscopy showed two morphological
forms, a
long motile, spiral-shaped organism which
we identified as a species of Thiospirillum
and a small motile rod, apparently a species of Chromatium.
The latter form prcdominated in mixed cultures and grew
rapidly under the culture conditions.
No growth on nitrogen-free media oc-
Mesotrophic
Cooter Pond (B )
Lochloosa (B)
Calf Pond (A)
Orange (A)
Oligotrophic
Watermelon
Pond
Unnamed 10 (A)
Jwwd
(B>
Moss Lee ( B )
Altho ( B )
(13 )
Ultraoligotrophic
Anderson-Cue
(A)
Gallilee ( B )
Cowpen ( B )
Sand Hill (C)
Swan (C)
* Classification
into trophic class after Shannon and
Brexonik (in press),
+ Sediment types found:
A-Brown
flocculant
unconsolidated material
with or without
undecomposed
plant
remains. B-Black
jellylike sediment.
C-Sandy
bottom.
D-This
sediment was a mixture of sand and brown, foulsmelling sludge.
curred within 7 weeks in any of the 65
yeast and fungal isolates picked from the
18 June and 9 July samples. However,
growth was observed within 2 weeks for
all isolates incubated in the test medium
plus NII&l.
Nitrogen fixation in sediments of
Florida and Guatemala lakes
Sediment cores were obtained from 16
lakes in north central Florida in September 1969 and from thcsc and 9 additional
lakes in June and July 1970. These lakes
are all sampled periodically
in conjunc-
NITROGEN
FIXATION
TABLE
4. Nitrogen
ixation rates (1 in n N/ghr; 2 in ng N/mg N- c,r) in Florida lake se i7iments
Nitrogen
Lake
Bivins Arm
Kanapaha
Orange
Moss Lee
Apopka8
Alice
Unnamed 20
Sediment
type
A
A
A
B
B
A
D
fixation
193?
28
36
9.5
1.8
17
1.20
0
IN
LAKE
727
MIZE
Stratification
of nitrogen
5.
in ng N/g-hr;
2 in ng N/mg N-hr)
lake sediments
fixation
(1
in Florida
TABLE
rate*
1
197OT
22
59
28
0
T
16
2
2.0
1.4
1.3
0
0
23
* Calculatied
assumed a C,H, : NH, molar production
ratio of 3: 2.
t Maximum rate found in profile (Table 5).
$ One dredge sample 3 November 1969.
Q Average of 5 cores.
tion with other studies (Shannon and
Brezonik, in press), and their water and
sediment characteristics are thus known.
Table 3 lists these lakes and their trophic
pcrtinen t sediment
states along with
characteristics.
Short cores (30-50 cm) from the lakes either were blended in their entirety and an
aliquot taken for measurement of nitrogen
fixation or were segmented into several
zones which were then tested for fixation
individually.
Apparent nitrogen fixation
was shown in 7 ,of the 25 lake scdimcnts in
one or both surveys (Table 4). Of these
7 sediments, 4 were peatlike (type A),
and except for Lake Apopka, these also
gave the highest rates of ethylene production. Sediment from lake No. 20 appeared to be a mixture of sand and sludge.
This shallolw lake (max depth, 2.5 m) receives some domestic waste effluent and
is at times anoxic bclolw 1 m. Sediments
from Lakes Apopka, Bivins Arm, Kanapaha, and Orange emitted a musty or
moldy oldor when blcndcd; the other scdimcnts were either odorless or smelled of
hydrogen sulfide. Lake Alice was the only
sediment with a pronounced H2S odor
that showed significant fixation. This lake
receives a large proportion of its inflow
as treated sewage effluent; five cores taken
in 1969 showed fixation, but no fixation
was found in a core taken during the 1970
survey.
Nitrogen fixation
rate
Stratum
(cm)
1
Orange Lake-l
O-37 (Shallow-water
core)
37-50
O-36 (Deep-water
core)
36-50
Orange
Lake-l
969
1.09
0.33
0.98
0.29
Biuins
1.3
0.78
59
10
3
0
1.4
1.4
0.37
0
14
22
18
5
0
1.4
2.0
1.9
0.8
0
Arm-1970
o-2
2-5
5-10
lo-15
20-25
Unnamed
o-2
24
4-6
6-8
8-10
lo-12
12-15
15-20
20-25
28
1.1
Kunapaha-1970
O-5
%15
15-25
25-38
Lake
-
970
O-33
3342
Lake
2
Lake
20-l
970
3
12
16
3
4
6
0
0
0
0.51
2.1
2.3
0.71
0.25
0.24
0
0
0
To identify the layers of sediment most
active in nitrogen fixation, we measured
profiles of acetylene reduction rates on
sediment cores from Orange Lake in the
1969 survey and on all lakes sampled in
1970. The loose, unconsolidated nature of
most of the sediments exhibiting nitrogen
fixation precluded detailed segmentation
of the cores, and only the core from the
unnamed lake, No. 20, was easily divisible
into segments. Table 5 presents the 1969
profiles for Orange Lake and profiles from
the 4 lakes exhibiting fixation in the 1970
survey. Highest rates were found in the
upper layers; this trend was most marked
when the rates are considered per gram
728
MICHAEL
A.
KEXRN
AND
TABLE 6. Effect of organic substraces on rates of
acetylene reduction
by Lake Kanapaha sediment
Relative
Substrate
rate of acetylene reduc!ion*
10-I M
10-z M
Sucrose
Pyruvate
Acetate
1.2
1.9
1.1
0.8
Butyrate
0.8
1.1
1.7
1.0
1.1
1.0
Glucose
10-3
M
lo-4
1.1
1.2
1.0
1.2
1.1
* Relative to acetylene reduction
rate of 0.50
C,H, reduced/mg
N-hr (average of 4 controls).
value is the mean of duplicate samples.
M
1.0
1.0
1.0
1.0
1.1
nmoles
Each
dry weight of sediment rather th‘an per
milligram of scdimen t nitrogen. Scdimcnt
from lake No, 20, although not as active
on a gram dry weight basis, was as active
as the other 3 when compared on a scdimcnt nitrogen basis.
Sediment samples from lakes in the
Pcten region of Guatemala surveyed by
one of us (P.L.B. ) during summer 1969
and 1970 also showed nitrogen fixation by
the acctylenc reduction method. Locations
and descriptions of these lakes arc given
by Fox ct al. (1970). In 1969, Laguna
Pctenxil sediment fixed at a rate of 3.1 ng
N/mg sediment N-hr, and green flocculant surface scdimcnt from Laguna Eckixil
fixed at a rate of 5 ng N/mg sediment-hr.
S’edimcnts from two Pet&-r lakes (Eckixil
and Sal Pctbn) wcrc sampled in detailed
profiles in 1970, and high rates of fixation
(up to 2.4 ngN/mg sediment N-hr) wcrc
found in both. Sal Pet&r scdimcnt exhibits three distinct bands ob fixation.
The
near surface sediments of these lakes arc
unusual in being composed of flocculent,
pigmented organic particles.
The green
to pink color of the sediment apparently
derives from undecomposcd algae and pcrhaps partially from photosynthetic
bacteria. The nature of thcsc sediments is under
investigation.
Relation of acetylene reduction
measurements to nitrogen fixation
in lake sediments
Acctylcnc reduction activity
be directly related to nitrogen
appears to
fixation in
PATRlCK
L.
RREZONIK
lake sediments. First, addition of 1 ml of
50% trichloroacetic
acid or 3 ml of saturated mercuric chloride immcdiatcly
and
completely
inhibits
cthylcnc production,
Second, if acctylenc reduction is related to
nitrogen-fixing
activity it would bc expectcd that molecular nitrogen would inhibit the rate of acetylene reduction.
Brooks et al. ( 1971) found that NB
inhibition
of acetylene reduction fits a
competitive enzyme inhibition pattern, corroborating the earlier finding of Schollhorn and Burris ( 1967) that acctylcnc is
a competitive inhibitor of nitrogen fixation.
A similar experiment was performed with
sediment from Lake Kanapaha, a shallow
eutrophic lake near Gainesville.
Varying
amounts of acctylcne were added to scrum bottles containing 25 ml of sediment
slurry, which had tither been purged with
the gas mixture to climinatc N2 or left
unpurged, and the samples were incubated
for 1.5 hr at 22C in the dark. A Lincweaver-Burk plot of the resulting rates vs.
substrate (acetylene)
concentrations also
fits a pattcm of competitive inhibition
of
ethylene production by N2,
IIcterotrophic
nitrogen fixation is knoswn
to depend on the availability
of organic
carbon ( Stewart 1966). Table 6 shows the
results of an experiment designed to show
the cffcct of various organic substrates on
acctylenc reduction in lake sediment. Samples of scdimcnt slurry (25 ml) wcrc
purged and injected with acetylene in the
usual manner cxccpt that before injection
of acetylene the carbon source was added
to the concentrations
shown.
Sucrose
definitely
stimulated acctylenc reduction
while the 0.10 M concentrations of sodium
acctatc and bu tyratc caused inhibition.
DISCUSSION
The agents responsible for fixation in
the two environments under discussion arc
probably hctcro trophic bacteria. In Lake
Mize fixation occurs only in a region of
low light and no dissolved oxygen. The
fact that fixation is maximum at intcrmcdiatc depths suggests that photosynthetic
NITROGEN
FIXATION
bacteria may bc the agents and they arc
in fact prcscnt in the subsurface layers
of this lake. However, the high rates of
acetylene reduction in Lake Mize samples
incubated in the dark would tend to rule
out photoautotrophic
forms as dominant
agents of fixation since photosynthesis apparently provides the source of energy for
fixation by these f o,rrns. Rhodospirillum
rubrum, for example, has been shown to
fix nitrogen anaerobically in the light, but
intact cells of this bacterium cease fixing
immediately
when placed in the dark
(Pratt and Frcnkcl 1959). Further, photosynthetic bacteria were isolated from only
one of the sampled depths at which fixation occurred (7 m). The long lag bcfo’re significant growths of these organisms
were noted in enrichment m!edia and the
fact that microscopic examination of raw
lake water and seston retained on Millipore filters failed to reveal the presence
of photosynthetic bacteria imply very low
populations in the lake.
On the other hand, heterotrophic growth
on a nitrogen-free medium was rapid for
samples from the various depths at which
fixation occurred.
Further culture work
isolated Gram-positive obligate anaerobic,
spore-forming
rods capable of using N2
as their sole nitrogen source and also capable of acetylene reduction in pure culture;
the characteristics
of thcsc isolates arc
those of a Clostridium
species. Whcthcr
this is the only hctcrotrophic
bacterium
fixing nitrolgcn in Lake Mize cannot bc
answered by this study. Other enrichment
procedures may isolate different
forms;
Arthrobacter
has been suggcstcd as another likely agent (J. Sieburth, personal
communication ) . No aerobic or facultative bacteria and no fungi or yeasts were
found that could grow on nitrogen-deficient media. Cyanophyceans are not likely
to bc agents of fixation in Lake Mize since
algal nitrogen fixation is rclatcd to photosynthesis and conditions at the depths of
fixation indicate no oxygen production.
No blue-green algae wcrc found in samples examined microscopically.
The &-
IN
LAKE
MIZE
729
served stratification
of fixation
and its
annual periodicity
may be influcnccd by
such factors as a requirement for low light
and narrow tolcrancc to high or low sulfide concentrations or Eh values, which
form gradients in the anoxic zone of this
lake.
Sediment nitrogen fixation is most likely
mcdiatcd by heterotrophic
bacteria and
fixation is stimulated by added sucrose. It
is obviously unlikely that photosynthetic
organisms are rcsponsiblc for the observed
rates since fixation extends into fairly deep
layers of sediment.
Stewart ( 1969) suggested that the lack
of oxidizable carbohydrate may limit the
significance elf hetero,trophic nitrogen fixation in both soils and natural waters.
Various reports in thle literature indicate
that heterotrophic nitrogen fixation under
labolratory conditions is incfficicnt in terms
of the amolunt ‘of carbon oxidized per unit
of nitrogen fixed. Rates reported in this
study polint either to more efficient utilization of substrate by nitrogen fixers in the
natural environment or to high concentrntions of usable energy sources at intcrmcdiatc depths in Lake Mize.
In Lake Mizc maximum nitrogen fixation
rates ranged from 0.08-3.26 pg N/liter-hr;
the upper value is within the range of
rates reported for blue-green algae in lakes
(Dugdalc and Dugdalc 1962; Goering and
Necss 1964; Billaud 1968). The annual input of nitrogen to Lake Mize by fixation
may be extrapolated from the data in Fig.
5. Calculation of yearly rates from a few
I-hr incubations is undoubtedly risky, but
even an approximate figure can be instructive. Assuming that during the period of
fixation the reaction occurred for 24 hr a
day at the rates shown in Figs. 3 and 4,
the total nitrogen contribution
in 1969
was 39.2 kg and in 1970 was 9.6 kg or
1.14 and 0.28 g/m3 of lake water per
year. Shannon ( 1970) computed a nitrogcn budget for Lake Mizc (excluding fixation) of 2.05 g/ms-yr based on land-use
patterns in the drainage basin and cstimates frolm the literature for the nutrients
730
MICI-IAEL
A.
KEIRN
AND
contributed
by runoff from the various
sources. Thus nitro,gen fixation represents
about 56 and 14% of the total nitrogen
income from other sources in 1969 and
1970, respectively, and at least for this
admittedly
unusual lake, bacterial nitrogcn fixation is a highly significant nutrient
source.
Although nitrogen fixation does not occur universally in lake sediments, its occurrence at least sometimes in 7 of the 25
Florida lake sediments suggests that the
phenomenon is more than just an environmental curiosity. The occurrence of fixation in the 3 Guatcrnala sediments tested
and in Lake Erie and Waccasassa estuary
sediments indicates a fairly widespread
distribution
of low nitrogen fixation rates
in the sediments underlying
natural waters. The range of fixatioln rates measured
in Florida sediments was 0.33-59 ng N/g
sediment-hr in surface layers and 0.02-1.1
in the bottom strata of 30-SO-cm coIrcs.
Comparable rates were found in Guatcmala lake sediments, and these rates are
in the same range as those reported for
the Waccasassa estuary (Brooks et al.
1971) and Lake Erie (Howard ct al. 1970).
It would be of great interest to determine why fixation occurs in some sediments but not in others. Examination elf
sediment characteristics
of the Florida
lakes (Shannon 1970) and of parameters
measured on the cores in this study shows
no obvious correlation between the occurrence of sediment fixation and sediment
ammonia, total organic nitrogen, phosphorus, or percent volatile solids, carbon, or
nitrogen.
Fixation
activities
should ble
most directly related to ammonia conccntration. Howcvcr, as was pointed out by
Brooks et al. ( 1971) cvcn though sediments may be apparently
nitrogen-rich,
much of the ammonia may be sorbed onto
clays and other particles and may be
unavailable to microorganisms.
Increased activity in the upper layers of
scdimcnt probably reflects the higher concentrations of oxidizable substrate in these
layers. Depth profiles of fixation were not
PATRICK
L.
BREZONIK
nearly so narrow as those found by Brooks
et al. in Waccasassa estuary scdimcnts
where fixation was confined largely to the
2s-cm stratum. The broader distribution
is undoubtedly
related to wind-induced
mixing of relatively unconsolidated
sediments in the shallow lakes. This hypothesis is supported by the results for lake No.
20, which had a sharper stratification
and
which also has a fairly compact sediment.
Stratification
of fixation
is much more
pronounced when the rates are expressed
per gram dry weight sediment than when
expressed per unit of organic nitrogen in
the sediment.
Nitrogen fixed in the sediments will bc
only partially
released to the overlying
water, depending on the degree of mixing
effected by wind action. In the loose
unconsolidated sediments found in Lakes
Kanapaha, Orange, and Bivins Arm, this
may bc considerable; in compacted sediments release to the overlying water may
be controlled by much slower diffusion
of the sediment
processes. Extrapolation
fixation rates in Tables 5 and 6 to annual
amounts indicates that the process may
contribute substantial quantities of nitrogen to the lake basin as a whole and may
in this sense be geochemically significant.
REFERENCES
AMERICAN PUBLIC HEALTH ASSOCIATION. 1965.
Standard
methods
for the examination
of
water
and wastewater,
12th ‘ed.
APHA.
769 p.
BILLAUD, V. A. 1968. Nitrogen fixation and the
utilization
of other inorganic
nitrogen
resources in a subarctic lake. J. Fish. Res. Bd.
Can. 25: 2101-2110.
BREZONIK, P. L., AND C. L. HAIWER.
1969. Nitrogen
fixation
in some anoxic lacustrine
environments.
Science 164: 1277-1279.
BROOKS, R., P. L. BI~~ZONIK, H. D. P~TNA;M,
1971. Nitrogen fixation
AND M. A. KEIRN.
in an estuarine environment:
the Waccasassa
on the Florida Gulf Coast.
Limnol. Oceanogr. 16: 701-710.
DUGDALE, V. A., AND R. C. DUGDALE.
1962.
Nitrogen
metabolism
in lakes. 2. Limnol.
Oceanogr. 7 : 170-177.
Fox, J, L., P. L. BREZONIK, I-1. D. PUTNAM, AND
1970.
A study of the
S. C. SNEDAKER
aquatic
and terrestrial
ecology
of selected
NITROGEN
FIXATION
karst lakes and basins near Flores, Guatemala. Rep. Univ. Florida. B,iomed. Sci. Support Grant Comm. 47 p. (Mim#eographed.)
GOERING, J. J., AND S. C. NEESS. 1964. Nitrogen
fixation
in two Wisconsin
lakes.
Limnol.
Oceanogr. 9 : 535-539.
GRAU, F. H., AND P. W. WILSON.
1962. Physiology of nitrogen fixation
by Buci2Zu.s poltjmyxa.
J. Bacterid.
83: 490.
HOWARD, D. L., J. I. FRED, R. M. BEEISTEW, AND
1970.
Biological
nitrogen
P. R. DUGAN.
fixation in Lake Erie.
Science 169: 61-62.
1967.
DeKAI-IN, L., AND F. T. BREZEXSKI.
termination
of nitrate
in estuarine
waters.
Automatic
determination
using a brucine
method.
Environ. Sci. Technol. 1: 492494.
LACKEY, J. B., E. W. LACKEY, AND G. B. MORGAN.
1965. Taxonomy
and ecology of the sulfur
bacteria.
Bull. Ser. 119, Eng. Ind. Exp. Sta.,
Univ. Florida, Gainesville.
METCALFE, G., AND M. BROWN, 1957. Nitrogen
fixation
by a new species of Nocu&z.
J.
Gen. Microbial.
17: 567-572.
MURPHY, J., AND J. P. RILEY.
1962. A modified single solution
method for the determination
of phosphate
in natural
waters.
Anal. Chim. Acta 27: 31-36.
PRATT, D. C., AND A. W. FREXKEL. 1959. Studies on nitrogen
fixation
and photosynthesis
of Rhodospirillum
rubrum.
Plant. Physid.
340: 333-337.
PSI-IENIN, L. N. 1959. Quantitative
distribution
of nitrogen fixing bacteria and their ecology
in the region of the Zernov Phyllophora
field
in the Black Sea. Mikrobiologiya
28: 92,7932.
IN
LAKE
MIZE
731
1963. Distribution
and ecology of hotobmter
in the Black Sea, p. 383-391.
In
C. H. Oppenheimer
[ea.],
Symposium
on
marine microbiology.
Thomas.
SCH~LLHORN, R., AND R. I-1. BURRIS. 1967. Acetylene as a competitive
inhibitor
of nitrogen
fixation.
Proc. Nat. Acad. Sci. U.S. 58:
213-216.
1970.
Eutrophication-trophic
SHANNON,
E. E.
state relationships
in north-central
Florida
lakes.
Ph.D. thesis, Univ. Florida,
Gainesville. 257 p.
-,
AND P. L. BIUZZONIK. IN PRESS. Limnological
characteristics
of north and central
Florida Lakes.
Limnol. Oceanogr. 17 ( 1).
STEWART,
W. D. P. 1966. Nitrogen fixation in
plants.
Athlone.
155 p.
-.
1969. Biological
and ecological aspects
of nitrogen fixation
by free living microorganisms.
Proc. Roy. Sot. London
Ser. B
-.
172: 367-388.
-
G. P. FITZGERALD, AND R. H. Bumus.
1967.
In situ studies on nitrogen
Fixation
using the acetylene
reduction
technique.
Proc. Nat. Acad. Sci. U.S. 58: 2071-2078.
TRITER,
H. G., AND S. GENOVESE. 1968. Charactcrization
of photosynthetic
sulfur bacteria
causing red water in Lake Faro.
Limnol.
Oceanogr. 13 : 225-232.
TECIINICON CORPORATION,
1969.
Ammonia
in
water
and waste water.
Technicon
Ind.
M&hod. Bull. 19-69w. Tarrytown,
N.Y.
WAKSMAN, S. A., M. HOTCHKTSS, AND L. C.
CAREY.
1933.
Bacteria
conccrncd
in the
cycle of nitrogen in the sea. Biol. Bull. 65:
137-167.