Methyl-Mercury Degradation
Pathways: A Comparison among
Three Mercury-Impacted Ecosystems
M A R K M A R V I N - D I P A S Q U A L E , * , ?
J E N N I F E R A G E E , ? C H A D M C G O W A N , ?
R O N A L D S . O R E M L A N D , ?
M A R T H A T H O M A S , ?
D A V I D K R A B B E N H O F T , ? A N D
C Y N T H I A C . G I L M O U R #
U.S. Geological Survey, Water Resources Division/MS 480,
345 Middlefield Road, Menlo Park, California 94025,
Department of Environmental Toxicology, University of
California, Santa Cruz, California 95064, U.S. Geological
Survey, Water Resources Division, 8505 Research Way,
Middleton, Wisconsin 53562, and The Academy of Natural
Sciences, Estuarine Research Center, 10545 Mackall Road,
St. Leonard, Maryland 20685
We examined microbial methylmercury (MeHg) degradation
in sediment of the Florida Everglades, Carson River
(NV), and San Carlos Creek (CA), three freshwater
environments that differ in the extent and type of mercury
contamination and sediment biogeochemistry. Degradation
rate constant (kdeg) values increased with total mercury (Hgt)
contamination both among and within ecosystems. The
highest kdeg's (2.8-5.8 d-1) were observed in San Carlos
Creek, at acid mine drainage impacted sites immediately
downstream of the former New Idria mercury mine, where
Hgt ranged from 4.5 to 21.3 ppm (dry wt). A reductive
degradation pathway (presumably mer-detoxification)
dominated degradation at these sites, as indicated by the
nearly exclusive production of 14CH4 from 14C-MeHg,
under both aerobic and anaerobic conditions. At the
upstream control site, and in the less contaminated
ecosystems (e.g. the Everglades), kdeg's were low (e0.2
d-1) and oxidative demethylation (OD) dominated degradation,
as evident from 14CO2 production. kdeg increased with
microbial CH4 production, organic content, and reduced
sulfur in the Carson River system and increased with
decreasing pH in San Carlos Creek. OD associated CO2
production increased with pore-water SO42- in Everglades
samples but was not attributable to anaerobic methane
oxidation, as has been previously proposed. This ecosystem
comparison indicates that severely contaminated sediments
tend to have microbial populations that actively degrade
MeHg via mer-detoxification, whereas OD occurs in heavily
contaminated sediments as well but dominates in those
less contaminated.
Introduction
Methylmercury (MeHg) is a heavy metal organo-toxin formed
primarily by sulfate reducing bacteria in anoxic sediments
(1, 2). Due to concerns regarding its bioaccumulation in
aquatic food chains, much attention has been focused on
factors controlling MeHg production under various envi-
ronmental conditions (3-8). The balance of MeHg produc-
tion and degradation ultimately controls MeHg concentra-
tion, yet comparatively little attention has been given to the
degradation process (9), which may proceed by a number of
abiotic and biotic pathways in the natural environment.
Abiotic pathways include photodegradation (10) and the
reaction with sulfide to form dimethylmercury (Me2Hg) and
HgS (11).
Biotic degradation also takes a number of forms. The most
thoroughly researched involves mercury resistance in bacteria
possessing genes of the mer-operon ((12-14) and references
therein). This capacity appears widespread in nature and
has been found for both gram negative and gram-positive
bacteria and under aerobic and anaerobic conditions (15-
17). The mer-operon can be carried on plasmids and other
transposable elements and transferred among different
bacteria species. ?Broad-spectrum? resistance refers to the
ability of bacteria to detoxify both inorganic Hg(II) and
organomercurials, including MeHg. This contrasts with
?narrow-spectrum? resistance, in which only Hg(II) detoxi-
fication occurs. The transcription of the specific detoxification
and transport proteins is regulated by an organomercurial-
responsive MerR protein in the first case and by a Hg(II)
responsive regulatory MerR in the second case. Unique to
broad-spectrum resistance is the mer-B gene that encodes
for the organomercurial-lyase enzyme, which cleaves MeHg,
forming CH4 and Hg(II) as end-products. The associated
mer-A gene, common in both resistance types, produces the
enzyme mercuric reductase, which further reduces Hg(II) to
volatile elemental Hg0 (18). In this way, broad-spectrum
mercury resistant microbes are able to detoxify MeHg by
converting it to a form that may readily evade from the
immediate environment.
An alternative anaerobic, non-mer-mediated, degradation
pathway has been demonstrated for the sulfate reducing
bacteria Desulfovibrio desulfuricans (19), where 2 mol of
MeHg react with microbially produced sulfide to form an
unstable dimethylmercury sulfide (MeHg)2S intermediate,
which decomposes to Me2Hg and HgS, as in the abiotic
pathway above. Me2Hg is then degraded to MeHg and CH4.
Thus, the production of CH4 from MeHg is common to both
of the above reductive demethylation (RD) pathways. It is
unknown if the non-mer RD pathway is induced or regulated
by ambient MeHg concentrations, as is mer-detoxification.
However, genes regulating the production of sulfide and the
degradation of MeHg were shown to be located on the same
plasmid in Clostridium cochlearium T-2 (20), and it has been
proposed that the MeHg degradation pathway in C. cochle-
arium and D. desulfuricans is one and the same (21).
Reports of CO2 as a major bacterial end-product of MeHg
degradation in anaerobic sediments led to the proposal of
an oxidative demethylation (OD) pathway, which has since
been demonstrated in freshwater, estuarine, and alkaline-
hypersaline sediments (22-24). OD is thought to represent
a cometabolism of MeHg analogous to the metabolism of
other small organic substrates (e.g. C1 compounds) by
heterotrophic bacteria and as such does not represent an
active detoxification response. Sulfate reducing, methano-
genic, and aerobic bacteria have all been implicated in this
pathway. While the production of CO2 from MeHg is what
defines OD, the production of both CO2 and CH4 via OD is
also possible and would be analogous to the production of
both end-products in the degradation of methanol or
* Corresponding author phone: (650)329-4442; fax: (650)329-4463;
e-mail: mmarvin@usgs.gov.
? U.S. Geological Survey, Menlo Park, CA.
? University of California.
? U.S. Geological Survey, Middleton, WI.
# The Academy of Natural Sciences.
Environ. Sci. Technol. 2000, 34, 4908-4916
4908 9 ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 34, NO. 23, 2000 10.1021/es0013125 CCC: $19.00 ? 2000 American Chemical Society
Published on Web 10/14/2000
monomethylamine by methanogens (23, 24). In contrast, no
CO2 is formed from MeHg by either of the two RD pathways.
Recently however, the evidence for OD has come into
question. It has been suggested that the formation of 14CO2
from 14C-MeHg degradation experiments may simply reflect
mer-detoxification followed by anaerobic 14CH4 oxidation to
14CO2 (25).
The specific environmental factors that control the relative
importance of these biotic pathways in a particular system
are largely unknown, although MeHg and/or Hg(II) con-
centration are likely important. Since mer-detoxification of
MeHg is induced by the presence of the substrate, some
threshold concentration is needed to induce transcription
(26). A similar threshold concentration might not be necessary
for OD to occur if this pathway represents a cometabolism
of MeHg. We hypothesize that OD dominates at low in-situ
MeHg concentrations, when induction of mer-degradation
is minimal. It is also unknown if Hg(II) reduction to Hg0
occurs under conditions favoring OD. If this capacity is
lacking, an extended residence time for Hg might be predicted
in systems where OD dominates. Thus, our limited under-
standing of the OD pathway and its potential importance in
the global Hg cycle has led us to further investigate this
process in various ecosystems throughout the past decade.
We have observed OD in all systems investigated to date and
focus here on the three most intensively studied. The Florida
Everglades represents a moderately contaminated system
with a nonpoint-source atmospheric Hg input, whereas San
Carlos Creek (CA) and Carson River (NV) exhibit significantly
higher Hg levels owing to ongoing point-source and historic
regional contamination, respectively. Here we attempt to
decipher the relative importance of these various microbial
pathways under different environmental conditions by
comparing MeHg degradation dynamics, in terms of CH4
and CO2 end-products, both within and among these three
very different ecosystems. We also directly test the hypothesis
that anaerobic CH4 oxidation can account for the current
and previous reports of OD.
Methods
Sites and Field Sampling. Mercury loading to the Florida
Everglades is primarily in the form of atmospheric deposition
(27), with sediment total mercury (Hgt) concentrations
typically 0.1-0.5 ppm (dry wt) (28). Surface sediment was
collected from ten Everglades wetland sites (Figure 1) during
December 1996, July 1997, January 1998, and June 1998, as
part of the South Florida Aquatic Cycling of Mercury in the
Everglades (ACME) project (29, 30). Sample depth varied from
the top 0-4 cm (July 1997 and June 1998) to the complete
unconsolidated surface floc layer (top 4-10 cm; Dec 1996
and Jan 1998). These sites represent a 130-km north-south
transect along an eutrophication gradient, stemming from
high phosphate inputs from the Everglades Agricultural Area
(EAA) (31). Sites were located in the Water Conservation Areas
(WCA), the experimental Everglades Nutrient Removal (ENR)
zone, and the more pristine Everglades National Park (ENP).
Specific site descriptions have been given elsewhere (24, 32,
33).
The Carson River U.S. EPA Superfund site, in western
Nevada, was originally contaminated during the mid- to late
1800s with elemental Hg0 used in the processing of gold and
silver ores associated with the Comstock load. The river flows
northeast and drains into a desert/wetland evaporation basin
at the terminus. Historic Hg inputs are from smaller
tributaries originating near Virginia City and from a major
amalgamation facility near Dayton. Due to the reworking
and mobilization of these sediments over the past century,
downstream locations currently have benthic Hgt concen-
trations as high as several hundred ppm (34). Sediment
samples (0-4 cm) were collected during October 1998 at 13
sites, over a 100-km stretch from Carson City to the wetlands
FIGURE 1. Maps of the San Carlos Creek (CA) [a], Carson River (NV) [b], and Everglades (FL) [c] ecosystems, with sampling sites given
as (g) and towns/cities as (b). Part [c] includes the Everglades Agricultural Area (EAA), Water Conservation Areas (WCA), and Everglades
National Park (ENP).
VOL. 34, NO. 23, 2000 / ENVIRONMENTAL SCIENCE & TECHNOLOGY 9 4909
region. Sites were categorized into five types depending on
location and major features (n given in [ ]): river [5], Lahontan
reservoir [2], agricultural drain [2], wetland [3], and playa [1].
Descriptions of the Carson River and associated wetlands
are given elsewhere (35, 36).
San Carlos Creek (SCC), located in the Diablo Mountain
range of central California, intersects the former New Idria
mercury mine, which operated for 118 years (1854-1972)
and was the second largest producer of elemental Hg0 in
North America (37). The creek is impacted both by acid mine
drainage (AMD) and mercury contamination from un-
processed cinnabar (HgS) ore and roasted-ore waste. Surface
sediment (0-4 cm) was initially sampled in October 1997 at
a non-AMD control site (C-1) located 3.2 km upstream of the
mine and at an AMD site (AMD-3) located 1.2 km downstream
of the mine. A second sampling in January 1999 included the
two previous sites plus two additional sites associated with
a short (<0.2 km) feeder stream from the New Idria mine
flowing into SCC. AMD-1 was located directly in front of the
mine, where subsurface AMD emerges as surface flow. AMD-2
was 0.1 km downstream, adjacent to a settling pond at the
base of large roasted-ore waste pile. Detailed site descriptions
are given elsewhere (37, 38).
Sediment collection for microbial assays was conducted
by hand with acid cleaned polycarbonate core tubes in all
ecosystems. Holding times prior to the initiation of 14C-MeHg
incubations varied from <0.3 to 95 days (Table 1). When
incubations were not initiated within a few hours of sample
collection, sediment was stored at 5 ?C in completely filled
acid-cleaned mason jars until further analysis.
Sediment Assays. Sediment was subsampled (3 cm3) into
13 cm3 serum vials, which were crimp sealed and flushed
with O2-free N2 gas. Radiolabeled MeHg (as 14CH3HgI) was
added (2-42 nCi*100 ?L-1) to each. The final 14C-MeHg
amendment levels (2-52 ng Hg*cm-3 wet sed or 15-2400
ppb dry wt, median ) 134 ppb) were higher than in-situ
MeHg (<10 ppb dry wt) for these systems (28, 39). Samples
were vortexed (30 s) and incubated in the dark at room
temperature (17-22 ?C) for 6 to 28 h (Table 1). Incubations
were arrested with the addition of 1 mL of NaOH (3 N). Each
site/depth sample set was replicated (n ) 2-3) and included
one autoclaved killed control. A high specific activity 14C-
MeHg stock (54 mCi*mmol-1, Amersham Corp., Arlington
Heights, IL) was used in all investigations, and 14C-end-
products were quantified by a CH4 combustion and CO2
trapping technique, followed by liquid scintillation counting
(CT-LSC) (24). The serum bottle headspace was first flushed
with commercial air (35-30 mL*min-1 for 15 min), while
vortexing, to drive off 14CH4. This end-product was combusted
to 14CO2 in an inline furnace (850 ?C, using a CuO catalyst)
and subsequently trapped in a solution of 8 mL of methanol
and 3 mL of monoethanolamine. Nearly 100% 14CH4 extrac-
tion efficiency achieved by twice amending the sample with
1 mL of pure unlabeled CH4 during the flushing period.
Samples were subsequently acidified with 1 mL of 6 M HCl
to convert base-fixed aqueous 14CO32- to gaseous 14CO2, which
was then flushed from the bottle using N2 and similarly
trapped as above in a new CO2-trap. Pure nonlabeled CO2
was also twice added (1 mL) to samples during this second
flushing step to facilitate the removal of 14CO2 from the
original sample. Scintillation cocktail (ScintiVerse II, Fisher
Scientific) was added to all 14CO2 traps, and samples were
counted by LSC.
Pseudo-first-order MeHg degradation rate constant (kdeg)
values were calculated as kdeg ) -ln(1-f)*time-1, where f
was the fraction of added 14C-MeHg degraded to 14CH4 +
14CO2 (kill corrected). The relative amount of 14CO2 produced
was expressed as the percentage of total gaseous end-
products recovered (henceforth called %14CO2) and was
calculated from kill-corrected data as %14CO2 ) [14CO2/(14-
CH4+14CO2)]*100. While most 14C-MeHg incubations con-
sisted of only one time point (Table 1), multipoint time
courses (20-120 h) were also conducted at selected sites
from each ecosystem. In these cases, kdeg was calculated from
the slope of the initial linear portion of each [%MeHg
degraded versus time] curve. All determinations of statistically
significant relationships were based on the P < 0.05 criteria
for the slope of a linear model.
Methane (14CH4) oxidation was assessed in parallel with
14C-MeHg degradation in all three ecosystems. Samples from
SCC in 1997 were amended with 14CH4 (5 nCi*250 ?L-1, sp.
act. ) 56 mCi*mmol-1, purity ) 97.5%, Amersham Corp.)
and incubated 20 h, under both aerobic and anaerobic (static)
conditions. Everglades (January 1998) sediment samples (3
cm3) were amended with 14CH4 (6 nCi*100 ?L-1, added to the
gas phase of sealed vials) and incubated statically (no shaking)
for 6-12 h, under both aerobic and anaerobic conditions.
Carson River samples were slurried (3 cm3 sediment plus 1
mL of anoxic DI water), amended with 14CH4 (15 nCi*250
?L-1), and incubated on a gyrating shaker table (150 rpm)
under anaerobic conditions (only) for 45 h. Radiotracer
14CH4 (sp. act. ) 21.1 mCi*mmol-1), used for Everglades and
Carson River experiments, was obtained from and originally
produced in the laboratory of B. Ward (UC Santa Cruz, CA)
from methanogenic cultures incubated with H14CO3- (per-
sonal communication). 14CH4 and 14CO2 was subsequently
quantified by the CT-LSC method in all cases.
Sediment Hgt was quantified using acid digestion, Sn-
reduction, gold trapping, and cold vapor atomic fluorescence
spectrophotometry (CVAFS) detection (40, 41). MeHg was
assayed by distillation (42), aqueous phase ethylation, G-C
separation, and CVAFS detection (43). Assays were conducted
at the following three institutions: Everglades sampless
Academy of Natural Sciences (St. Leonard, MD), Carson River
samplessUSGS (Madison, WI), and New Idria sampless
USGS (Menlo Park, CA). All institutions used similar equip-
ment and assay conditions.
As measures of organic content, dry sediment samples
were subject to weight loss on ignition (LOI) analysis (44) (all
three systems) and to particulate carbon (PC) analysis (Carson
River only) measured with a Carlo Erba elemental analyzer
(Model 1500). Sediment pH was measured (Carson River and
SCC only) by inserting a pH electrode (Cole-Parmer, model
59002-72) directly into homogenized sediment. Acid-volatile-
TABLE 1. Conditions Used for Methylmercury (MeHg) Degradation Measurement via 14CH3HgI Incubation and CT-LSC Detection of
Gaseous 14C End-Products
sample set date
holding time
prior to assay (d)
incubation
duration (h)
14C-MeHg added
(nCi*cc sed-1)
total MeHg added
(ng Hg*cc sed -1)
Carson R. (NV) Oct 1998 93-95 24 14 52
Everglades (FL) Dec 1996 <0.3 22-28 0.5 2
Everglades (FL) July 1997 <0.3 7-8 2-4 7-15
Everglades (FL) Jan 1998 <0.3 6-12 2-3 7-11
Everglades (FL) June 1998 <0.3 6-8 3 11
San Carlos Creek (CA) Oct 1997 14 20 3 11
San Carlos Creek (CA) Jan 1999 30 23 2-3 7-11
4910 9 ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 34, NO. 23, 2000
sulfur (AVS) in Carson River samples was determined
spectrophotometrically (45) after zinc-acetate trapping of
H2S from acidified whole sediment (46). AVS was determined
similarly (47) in Everglades samples. Pore-water from Carson
River sediment was collected under anaerobic conditions
via centrifugation and was assayed for SO4-2 via ion-
chromatography (48) and for free sulfide (45). Everglades
pore-water was collected by direct filtration of whole
sediment or by using an in-situ interstitial pore-water
sampler, with free sulfide analyzed using an ion-specific
electrode (28). Methanogenesis in Carson River samples was
measured as the net CH4 production, over 7 days, quantified
by gas chromatography with flame ionization detection.
Methanogenesis in Everglades (1997) samples was measured
as the conversion of radiolabeled H14CO3- (spec. act. ) 54.4
mCi*mmol-1; ICN Biomedicals, Irvine, CA) to 14CH4, quanti-
fied via gas chromotography with gas proportional counting
detection (22).
Results
Values of kdeg increased with increasing Hgt (20-12 700 ppb
dry wt) in the Carson River (Figure 2a). Two distinct regional
groupings were observed, with the wetland sites exhibiting
a stronger MeHg degradation response to Hgt than river,
reservoir, and agricultural drain sites (combined). All sites
exhibited >80% 14CO2, except two river sites which were 20-
30% (Figure 2b). Everglades Hgt concentrations (Figure 2c-
d) fell into a low and narrow range (70-320 ppb dry wt)
compared to the Carson River. Everglades kdeg's were
consistently low (0.03-0.23 d-1) and similar to previously
measured values (24). The %14CO2 ranged from 15 to 92%.
Neither kdeg nor %14CO2 varied as a function of log[Hgt] in the
complete Everglades data set. However, significant relation-
ships between kdeg and specific mercury fractions (e.g. bulk
sediment Hgt and MeHg, pore-water Hgt) were found when
individual sampling dates were analyzed (Figure 3a-c),
although, specific results were not consistent among all dates.
SCC Hgt levels were very high at both the control site
(4800-9600 ppb dry wt) upstream and AMD sites (4500-
21 300 ppb dry wt) downstream of the New Idria mercury
mine (Figure 2e-f). A consistent spatial trend of low kdeg
(e0.1 d-1) and high %14CO2 (37-74%) upstream of the mine,
and high kdeg (0.8-1.5 d-1) and minimal %14CO2 (1-4%) below
the mine, was observed for both sampling dates.
Time course kdeg's (Figure 4a-g) ranged from 0.017 d-1
for the modestly contaminated Everglades ENR-103 site, to
5.8 d-1 for the severely contaminated SCC AMD-3 site. This
latter value was significantly larger than 1.5 d-1 depicted in
Figure 2e for the same site and date. The lower value was
calculated using the single 20-h data point so as to be
comparable with the kdeg for January 1999 SCC, which was
based on a single-point (22 h) incubation. The nonlinear
time courses (Figure 4a,e,g) point out the potential for the
underestimation of kdeg when calculated from a single time
point, particularly from a prolonged incubation. MeHg
degradation was slow (<0.1 d-1) and increased linearly with
time for both Everglades sites and SCC site C1. In contrast,
after an initial rapid rate (0.45 d-1), MeHg degradation slowed
over time at Carson River site F1. A similar, but more
pronounced, rapid initial degradation followed by a much
slower rate was observed for SCC AMD-3 sediment, under
both oxic and anoxic incubation conditions. The %14CO2 was
high (g40%) and remained relatively constant over time for
Carson River F1, Everglades ENR-103, and SCC C1 (anaero-
bic). There was a distinct decrease in %14CO2 with time at
both Everglades 3A-15 and SCC C1 (aerobic). Very little
(<0.05%) 14CO2 was produced in both aerobic and anaerobic
SCC AMD-3 time courses.
Carson River kdeg's increased with a number of bio-
geochemical parameters associated with the transition from
FIGURE 2. Log-linear plots of MeHg degradation rate constant (kdeg)
[a,c,e] and %14CO2 end-product [b,d,f] versus Hgt concentrations in
sediment from the Carson River (NV) 1998 [a,b], Everglades (FL)
1996-1998 [c,d], and San Carlos Creek (CA) 1997/1999 [e,f]. Carson
River data is grouped by ecosystem zone, Everglades data by
sampling date, and San Carlos Creek data by region and sampling
date. Least-squares regression line and associated r 2 is given in
[a] for data grouped by either wetlands or all other sites (excluding
playa).
FIGURE 3. Significant linear regressions of MeHg degradation rate
constant (kdeg) values versus various mercury pools in the Everglades
data set. Depth intervals for kdeg data were variable (top 4 to 10 cm)
during December 1996 and 0-4 cm during both July 1997 and June
1998. The corresponding mercury pool concentrations represent
the 0-4 cm average depth interval in all cases.
VOL. 34, NO. 23, 2000 / ENVIRONMENTAL SCIENCE & TECHNOLOGY 9 4911
low-organic river to comparatively high-organic wetland
sediments, including methanogenesis rate, sediment PC and
AVS, pore-water sulfide (Figure 5a-d), and LOI (not shown,
similar to PC graph 5c). These parameters did not covary
with Hgt (data not shown). In contrast, no significant
relationships were found between kdeg and methanogenesis,
AVS, or sulfide in the Everglades data (no PC data). A weak
negative relationship with LOI was seen for July 1997 but
was heavily weighted by a single data point (not shown). No
relationship between kdeg and LOI was seen for SCC (not
shown), although degradation increased with decreasing
sediment pH, which ranged from 8.1 to 8.7 at C1 and from
to 2.6 to 7.1 at the AMD sites (Figure 5e). No significant
relationship between pH and kdeg was observed for the Carson
River data, where pH varied over a much narrower range
(6.9-8.2).
Methane oxidation was investigated to determine if this
microbial process could account for any of the 14CO2
production routinely observed for 14C-MeHg degradation
experiments. In four Everglades sediments, 48-69% 14CO2
was produced from 14C-MeHg under anaerobic conditions,
whereas only 0-4% 14CO2 was produced from 14CH4 during
contemporaneous incubations (Table 2). In contrast, 9-23%
14CH4 oxidation was observed in aerobic samples from the
two sites. A corresponding increase in the %14CO2 from 14C-
MeHg was also observed under aerobic conditions, although
the total amount of MeHg degraded decreased slightly in
both cases. In a similar set of parallel incubations (anaerobic
only), no 14CO2 was produced from 14CH4 (detection limit ca.
0.1%) at any of 13 Carson River sites (not shown), while end-
product %14CO2 from 14C-MeHg ranged from 24 to 98%
(Figure 2b). Finally, no 14CH4 oxidation was observed at either
SCC (1998) site under either aerobic or anaerobic conditions,
during a 20-h incubation (not shown). A positive relationship
between %14CO2 and pore-water SO42- was observed in three
of the four Everglades sampling dates (Figure 6), although,
a similar relationship was not evident in the Carson River
data.
FIGURE 4. Time course experiments: percent MeHg degraded
(closed square) and percent 14CO2 end-product (open square) versus
time for Carson River (NV), site F1 [a]; Everglades (FL), sites ENR-103
and 3A-15 [b,c]; and San Carlos Creek (CA), sites C1 and AMD-3
[d-g]. Degradation rate constants (kdeg's) are calculated from the
initial linear portion of each % degradation versus time plot. The
maximum time point used for each regression is noted in each
case, as are aerobic or anaerobic incubation conditions.
FIGURE 5. Significant linear regressions of biogeochemical
variables (CH4 production, acid-volatile-sulfur (AVS), particulate
carbon (PC), pore-water sulfide (H2S), and sediment pH) versus
MeHg degradation rate constants (kdeg) for 0-4 cm depth interval
sediment in the Carson River (NV) 1998 [a-d], and San Carlos Creek
(NV) 1997/1999 [e].
TABLE 2. Parallel 14C-MeHg Degradation and 14CH4 Oxidation
Experiments Conducted under Both Anaerobic and Aerobic
Conditions with Florida Everglades Whole Sediment (Jan
1998)a
site
MeHg degradation
kdeg (d-1)
MeHg end-product
%14CO2
14CH4 oxidation
to 14CO2 (%)
Anaerobic Incubations
LOX 0.06 (0.01) 53 (7) 0.8 (0.8)
TS-7 0.11 (0.01) 48 (5) 0.2 (0.1)
2Bs 0.09 (0.02) 57 (23) 4.1 (1.6)
ENR-103 0.07 (0.01) 69 (8) 0.0
Aerobic Incubations
LOX 0.04 (0.00) 75 (9) 23 (3)
TS-7 0.05 (0.01) 102 (14) 9 (1)
a Standard deviations are given in parentheses. Replication was n
) 3 and n ) 2 for 14C-MeHg and 14CH4 amended samples, respectively.
Incubation time was 6-12 h.
4912 9 ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 34, NO. 23, 2000
Discussion
The positive relationship between kdeg and log[Hgt] in the
Carson River data set (Figure 2a) reconfirms earlier findings
for this system in which the demethylation rate increased
among three sites with increasing Hg contamination (23).
The two distinct spatial groupings in the current data indicate
that MeHg degrading bacteria in the wetlands were more
responsive to Hg contamination than bacteria in other
regions. This may reflect differences in the composition,
abundance or activity of the respective microbial communi-
ties, and/or differences in the MeHg availability. The %14CO2
data suggests that OD dominated MeHg degradation even
at the most contaminated sites. This appears in contrast to
results from the earlier investigation noted above, in which
%14CO2 decreased with increasing Hg contamination and
demethylation rate. Such a trend would suggested a shift in
microbial populations, from those invoking OD at low Hg
levels to those invoking RD at higher contamination levels.
The natural selection of bacteria, able to invoke mer-
detoxification of both inorganic and organic mercury, has
been shown in other Hg contaminated sediments (18, 49).
The lack of any clear relationship between %14CO2 and Hgt
in the current study may indicate that the long sediment
holding time (>90 days), prior to 14C-MeHg incubation,
impacted the original community composition so as to
obscure this relationship. Alternatively, the apparent trend
in the earlier report may have been spurious due to the limited
number of observations (n ) 3) or because mer-detoxification
was inadvertently stimulated in contaminated sediments due
to the high 14C-MeHg amendment levels used (1800 ng
Hg*cm-3) compared to the current study (2-52 ng Hg*cm-3).
The positive relationship between kdeg and various
mercury pools evident in the Everglades data (Figure 3)
further illustrates the potential for increased degradation with
increasing contamination within an ecosystem. The incon-
sistency in the types of significant regressions observed
among sampling dates partially reflects the fact that while
sediment for both MeHg degradation assays and Hg-
speciation analysis was collected at the same site and date,
these samples were collected by two different groups of
researchers, often tens of meters apart and not always at the
same depth intervals (see Figure 3 legend). Thus, the analysis
of kdeg and Hg-speciation relationships is less than optimal
for this data set. However, since within-site variation in kdeg
and Hg-species concentrations was presumably smaller than
regional (among-site) variations (not directly tested), sig-
nificant trends were detected in some cases. Further, since
the range of Hgt concentrations is much smaller in the
Everglades compared to the Carson River, the expected
response of the microbial community to increasing Hg
contamination in the Everglades is expected to be more subtle
and significant relationships more difficult to decipher.
Finally, other geochemical factors assuredly influence kdeg
values and thus partially obscure the direct influence of
increased contamination in the Everglades.
Comparisons among systems further confirm the positive
relationship between kdeg and Hgt. In the Everglades, where
Hgt values were comparatively low, kdeg's were likewise
consistently low (Figure 2c). In SCC, where Hgt was very high
both above and below the New Idria mine, kdeg's were high
at all AMD sites and low at the control site (Figure 2e). Mercury
bioavailability likely accounted for this among-site difference
within SCC. The source of Hg upstream is primarily recal-
citrant and insoluble weathered cinnabar (HgS) abundant
throughout the local area, whereas the downstream source
includes particle-bound Hg(II) liberated by acidic conditions
within the mine and the leaching of roasted-ore waste
adjacent to the mine (37, 38). Thus, the higher levels of
bioavailable Hgt at sites downstream of the mine would be
more prone to select for bacterial populations that actively
degrade MeHg.
As with the 1998 Carson River data, the lack of any clear
relationship between %14CO2 and log[Hgt] in the Everglades
(Figure 2d) indicates that something other than Hgt alone
influences MeHg degradation pathways or stoichiometric
end-product ratios. The positive relationship between %14-
CO2 and pore-water SO42- (Figure 6) suggests that this anion
plays a role in MeHg degradation pathway, although it is
unclear if this role is biological (e.g. mediating sulfate
reduction) or abiotic (e.g. affecting MeHg-complex forma-
tion). In either case, the consistent production of 14CO2 at all
sites demonstrates that OD was active, if not dominant. We
infer that in-situ Hg was not high enough to induce a strong
RD response in the Everglades sites. The situation appears
altogether different in SCC sediments, where the lack of
significant 14CO2 production in AMD sites indicates that RD
dominated degradation, as might be predicted under severely
contaminated conditions. Abundant 14CO2 production ob-
served upstream of the mine (at C-1) supports the hypothesis
that a strong RD response is invoked when Hg is not only
very high in concentration but also bioavailable in form.
Time course experiment results (Figure 4) for Carson River
site F1 and SCC site AMD-3 indicate that the 14C-MeHg
amendment was sequestered into at least two pools; one
readily available to the resident microbial community and
one less available. This was apparent from the initial rapid
degradation rates followed by a slowing or cessation of
degradation and was in contrast to the slow linear degradation
seen in both Everglades sites and SCC site C1. Variations in
substrate availability, due to refractory MeHg-complex
formation with dissolved and/or particulate phases, may
partially account for these spatial differences. Carson River
site F1 had low organic content (4% LOI) compared to the
two Everglades sites (80-91% LOI), suggesting that very
organic-rich sediments may sequester a larger fraction of
MeHg into slowly degrading refractory pools. We conclude
that sediment organic content was not responsible for the
large spatial differences in kdeg observed for SCC because
LOI percentages were similar for SCC sites C1 and AMD-3
(11% and 17% LOI, respectively) and no significant relation-
ship between kdeg and LOI was observed for SCC data.
However, large differences in solid-phase composition were
evident, with the AMD sites primarily composed of orange
colored flocculent material, presumably iron(III)-oxy-hydroxy
sulfate precipitate, typical of acid mine drainage (50). The
distribution of 14C-MeHg between organic and inoganic solid
phases was not directly assessed. However, since 55-75% of
the 14C-MeHg amendment was readily degraded within 5 h
at AMD-3 (Figure 4e,g), we speculate that much of the
substrate was associated with the iron(III)-oxy-hydroxy
sulfate fraction and that this portion was readily available to
the MeHg degrading microbial population.
FIGURE 6. Significant linear regressions of the percent 14CO2 end-
product versus pore-water SO42- concentrations (0-4 cm depth
interval) from the Everglades data set.
VOL. 34, NO. 23, 2000 / ENVIRONMENTAL SCIENCE & TECHNOLOGY 9 4913
The nearly constant %14CO2 produced in five of seven
sites (Figure 4) indicates that one pathway dominated MeHg
degradation in most cases. Specifically, OD is implicated in
the case of Carson River F1, Everglades ENR-103, and SCC
C1 (anaerobic), and mer-detoxification is implicated in the
case of SCC AMD-3. We emphasize mer-detoxification in the
latter case and not the alternative RD pathway (via reaction
with H2S), which would have been inhibited under aerobic
conditions. Similarly for SCC C1, the change in the %14CO2
trend, from constant and high under anaerobic conditions
to decreasing with time under aerobic conditions, also
suggests mer-detoxification may have been preferentially
stimulated under aerobic conditions. The clear decrease in
%14CO2 with time at Everglades 3A-15 [anaerobic] and SCC
C1 [aerobic] (Figure 4c,f) indicate that OD and RD were
simultaneously active with RD dominating, since %14CO2
would be expected to increase if OD dominated. An alterna-
tive explanation for decreasing 14CO2 with time is that different
microbial groups are capable of OD but with different
stoichiometric end-product 14CO2/14CH4 ratios and/or at
different rates. It is important to emphasize that %14CO2 end-
product measurements alone do not indicate the relative
importance of OD verses RD, particularly with single time
point incubations, as some 14CH4 may also be an OD end-
product (23, 24). Only in cases where either 14CH4 or 14CO2
is the exclusive end-product can either RD or OD, respectively,
be solely inferred. Then, only under aerobic conditions can
the non-mer Me2Hg intermediate pathway be ruled out and
mer-detoxification surmised (e.g. SCC AMD-3).
Amendments with 14C-MeHg, above ambient MeHg levels,
may have stimulated RD to varying degrees at some sites
(e.g. Figure 4c,f), which may partially account for the wide
range of %14CO2 values observed in Everglades samples
(Figure 2d). It is noteworthy that when observed, the decrease
in %14CO2 was immediate and did not involve a lag time,
suggesting that the bacterial community was preacclimated
to MeHg (18, 51). Alternatively, the addition of 14C-MeHg to
organic-poor sediments may have stimulated heterotrophic
bacteria capable of using MeHg as an organic substrate. If
so, this would cause us to overestimate the importance of
OD, as the rate of 14CO2 produced would presumably be
higher than that of nonlabeled CO2 produced from in-situ
MeHg levels. Previous experiments with Everglades sediment
did show a significant increase in %14CO2 produced with
increasing 14C-MeHg over a large amendment range (50-
4000 ng MeHg*g dry sed-1) but no significant increase in
%14CO2 over a much smaller range (2-18 ng MeHg*g dry
sed-1) (24). Amendment concentrations in the current work
were varied over a wide range (in ng MeHg*g dry sed-1:
Everglades, 16-2600; Carson River, 68-270; SCC, 16-61), due
to large variations in sediment porosity. While the corre-
sponding amount of carbon added from 14C-MeHg was small
on a volumetric basis (0.01-0.25 nmol C*cm-3), and only a
fraction of the added radiolabel may be available for
degradation, it is uncertain if the 14C-MeHg amendment levels
used in the current experiments resulted in a significant
stimulation of heterotrophic activity. This possibility cannot
be ruled, especially for some of the low organic sediments
of the Carson River systems.
It is not surprising that a clear relationship between Hgt
and kdeg was not evident in all cases, as other environmental
factors undoubtedly also impact observed MeHg degradation
rates. The increase in kdeg with methanogenesis, sediment
organic content (PC), and reduced S suggest that within-
system regional differences in benthic microbiology and/or
geochemistry are important in the Carson River (Figure 5a-
d). It is difficult to assess the relative contribution and
mechanism of each of these covarying parameters, as a
control on microbial MeHg degradation, without conducting
controlled experiments. Taken together, however, they depict
a shift to higher kdeg's going from organic-poor (river) to
comparatively organic-rich and sulfidic (wetland and agri-
cultural drain) sites with higher rates of anaerobic metabo-
lism. These relationships were statistically independent (P
> 0.05) of increases in kdeg due to increasing Hgt. The increase
in kdeg with increasing pore-water sulfide might suggest the
non-mer RD pathway (via Me2Hg formation), although the
major end-product (>80%) was14CO2 and not 14CH4 in most
cases (Figure 3b). Thus, it would appear that it was OD, not
RD, which dominated degradation. It is unknown if OD can
also be carried out on (MeHg)2S or Me2Hg, but such reactions
could explain the spatial variation in the Carson River system.
The lack of positive relationships in the Everglades data
set, similar to those noted above for the Carson River, may
be due to the difficulty in detecting such relationships with
such a comparatively low and narrow range of kdeg's values.
Alternatively, the relative influences of individual environ-
mental controls on MeHg degradation may differ among
systems. Specifically, the large difference in sediment organic
content (as assessed by LOI) between the Everglades (33-
91%, median ) 80%, n ) 25) and the Carson River (1-12%,
median ) 2%, n ) 13) may partially account for the contrast
in kdeg values among these ecosystems, for different reasons.
The consistently low kdeg values in the organic-rich Everglades
could reflect a high degree of MeHg-organic (or MeHg-
reduced-S) complex formation, thereby decreasing MeHg
availability to bacteria. While benthic anaerobic metabolism
is presumably not carbon limited in the Everglades, organic
substrate appears to limit microbial rates in the Carson River
system, as evident from the increase in both methanogenesis
and SR along a transect from organic-poor river sites to
comparatively organic-rich wetland sites (data not shown).
This increase in microbial rates parallels the increase in MeHg
degradation for the Carson. Additional unpublished sequen-
tial extraction experiments conducted with 14C-MeHg
amended Carson River sediment indicates decreasing dis-
solved (water-extractable) and readily exchangeable (acid-
extractable) MeHg pool size and an increase in MeHg-organic
complex (base-extractable) pool size, with increasing organic
content (data not shown). Assuming that the dissolved and
readily exchangeable MeHg pools are more available for
degradation than the MeHg-organic complex pool, then the
increase in the overall activity of the MeHg degrading
community more than compensates for the decrease in
bioavailable MeHg pool size along the Carson River organic
gradient.
The apparent increase in kdeg with decreasing pH in SCC
sediments (Figure 5e) was not due to abiotic acid cleavage
of the methyl group from MeHg, since the kdeg's presented
were kill corrected and represent microbial degradation only.
Acidophilic bacteria were thus clearly involved in MeHg
degradation at the AMD sites. To our knowledge, this is the
first time that a significant MeHg degradation capacity has
been suggested for this general bacterial group.
The lack of significant anaerobic 14CH4 oxidation, in any
of the three ecosystems, demonstrates that this process could
not explain the 14CO2 produced from anaerobic 14C-MeHg
degradation, as has been recently proposed (25). The fact
that CH4 oxidation was readily observed in Everglades
samples incubated aerobically demonstrates our ability to
detect this process. The corresponding increase in %14CO2
from 14C-MeHg, under aerobic conditions, indicates that
some of this 14CO2 could have been due to aerobic oxidation
of 14CH4 produced from either RD or OD.
A clear demonstration of OD in pure culture remains
outstanding to date. The methylotrophic methanogen GS-
16 (22), subsequently named Methanolobus taylorii sp. nov.
(52), produced 14CO2 from 14C-MeHg (%14CO2 ) 29-46%)
when grown on trimethylamine, although the total amount
of 14C-MeHg degraded was low (<3%). No anaerobic 14CO2
4914 9 ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 34, NO. 23, 2000
production from 14C-MeHg was detected for two sulfate
reducing strains (Desulfovibrio desulfuricans LS and ND 132)
and one methanogen (Methanococcus maripaludis), in a
subsequent study (25). However, the high 14C-MeHg amend-
ment level used (500 ng/cm3) was far in excess of typical
environmental contamination levels and in excess of the
levels used in the current study (2-52 ng/cm3). Subsequently,
mer-detoxification may have been induced, giving rise to
detection of 14CH4 only. It was not noted whether these
bacteria were screened for the mer-operon. Further, previous
work by Baldi et al. (19) demonstrated that D. desulfuricans
degrades MeHg by the non-mer RD pathway (via reaction
with H2S), even under SO42- limited conditions. While the
above study (25) cites low SO42- conditions for the culture
media, it is also possible that low sulfide levels also existed
in the sulfate reducing cultures and that the non-mer RD
pathway was subsequently responsible for the detection of
14CH4 as the sole gaseous end-product.
The current study demonstrates strong within-system and
among-system differences in MeHg degradation rates and
pathways. Systems or regions with low Hg contamination
exhibited low kdeg's and OD dominated the degradation
pathway. A much wider range of kdeg's was observed at higher
contamination levels as other environmental factors become
important, such as overall metabolic rates, sediment reduced
S, organic matter concentrations, and substrate availability.
The sequestering of MeHg by various solid phase fractions,
as a key factor in mediating MeHg availability to bacteria,
should be more fully investigated. Only under conditions of
extreme contamination with bioavailable Hg was a strong
RD pathway clearly dominant, and only then under aerobic
conditions was mer-detoxification specifically implicated.
While OD appears widespread in natural systems, its
unambiguous demonstration in pure culture remains elusive,
and the pathway specifics remain unknown. More work is
needed to reconcile the results from the limited number of
pure culture experiments with those from whole sediment
field measurements.
Acknowledgments
This work was supported by the U.S. EPA (DW14955409-
01-0) and the U.S. Geological Survey Place-Based Studies
Program. The authors wish to thank S. Luoma (USGS, Menlo
Park, CA) and C. Alpers (USGS, Sacramento, CA) for initial
manuscript reviews; J. DeWild, S. Olund, and M. Olson (USGS,
Madison, WI) for Carson River Hgt and LOI analysis; C. Kendall
and S. Silva for PC analysis; G. Riedel for Everglades Hgt
analysis; B. Ward (UC Santa Cruz, CA) for the 14CH4; and
three anonymous reviewers for their helpful comments. We
extend special thanks to Priya Ganguli and Khalil Abu-Saba
for introducing us to San Carlos Creek.
Literature Cited
(1) Compeau, G. C.; Bartha, R. Appl. Environ. Microbiol. 1985, 50,
498.
(2) Gilmour, C. C.; Henry, E. A.; Mitchell, R. Environ. Sci. Technol.
1992, 26, 2281.
(3) Choi, S.-C.; Bartha, R. Bull. Environ. Contam. Toxicol. 1994, 53,
805.
(4) Regnell, O.; Tunlid, A.; Ewald, G.; Sangfors, O. Can. J. Fish. Aquat.
Sci. 1996, 53, 1535.
(5) Benoit, J. M.; Gilmour, C. C.; Mason, R. P.; Riedel, G. S.; Riedel,
G. F. Biogeochemistry 1998, 40, 249.
(6) Jackson, T. A. Can. J. Fish. Aquat. Sci. 1988, 45, 97.
(7) Barkay, T.; Gillman, M.; Turner, R. R. Appl. Environ. Microbiol.
1997, 63, 4267.
(8) Hadjispyrou, S. A.; Anagnostopoulos, A.; Nicholson, K.; Nim-
fopoulos, M. K.; Michailidis, K. M. Environ. Geochem. Health
1998, 20, 19.
(9) Compeau, G. C.; Bartha, R. Appl. Environ. Microbiol. 1984, 48,
1203.
(10) Sellers, P.; Kelly, C. A.; Rudd, J. W. M.; MacHutchon, A. R. Nature
1996, 380, 694.
(11) Deacon, G. B. Nature 1978, 275, 344.
(12) Robinson, J. B.; Tuovinen, O. H. Microbiol. Rev. 1984, 48, 95.
(13) Hobman, J. L.; Brown, N. L. In Metal Ions In Biological Systems:
Vol. 34 - Mercury and Its Effects on Environment and Biology;
Sigel, A., Sigel, H., Eds.; Marcel Dekker: New York, 1997; Chapter
19.
(14) Hobman, J. L.; Wilson, J. R.; Brown, N. L. In Environmental
Microbe-Metal Interactions; Lovley, D. R., Ed.; Am. Soc. Mi-
crobiol.: Washington, DC, 2000; Chapter 8.
(15) Spangler, W. J.; Spigarelli, J. L.; Rose, J. M.; Flippin, R. S.; Miller,
H. H. Appl. Micriobiol. 1973, 25, 488.
(16) Radosevich, M.; Klein, D. A. Bull. Environ. Contam. Toxicol.
1993, 51, 226.
(17) Weber, J. H.; Evans, R.; Jones, S. H.; Hines, M. E. Chemosphere
1998, 36, 1669.
(18) Barkay, T.; Turner, R. R.; VandenBrook, A.; Liebert, C. Microb.
Ecol. 1991, 21, 151.
(19) Baldi, F.; Pepi, M.; Filippelli, M. Appl. Environ. Microbiol. 1993,
59, 2479.
(20) Pan-Hou, H. S. K.; Hosono, M.; Imura, N. Appl. Environ.
Microbiol. 1980, 40, 1007.
(21) Baldi, F. In Metal Ions In Biological Systems: Vol. 34 - Mercury
and Its Effects on Environment and Biology; Sigel, A., Sigel, H.,
Eds.; Marcel Dekker: New York, 1997; Chapter 8.
(22) Oremland, R. S.; Culbertson, C. W.; Winfrey, M. R. Appl. Environ.
Microbiol. 1991, 57, 130.
(23) Oremland, R. S.; Miller, L. G.; Dowdle, P.; Connell, T.; Barkey,
T. Appl. Environ. Microbiol. 1995, 61, 2745.
(24) Marvin-DiPasquale, M. C.; Oremland, R. S. Environ. Sci. Technol.
1998, 32, 2556.
(25) Pak, K.; Bartha, R. Bull. Environ. Contam. Toxicol. 1998, 61, 690.
(26) O'Halloran, T. Science 1993, 261, 715.
(27) Dvonch, J. T.; Graney, J. R.; Keeler, G. J.; Stevens, R. K. Environ.
Sci. Technol. 1999, 33, 4522.
(28) Gilmour, C. C.; Riedel, G. S.; Ederington, M. C.; Bell, J. T.; Benoit,
J. M.; Gill, G. A.; Stordal, M. C. Biogeochemistry 1998, 40, 327.
(29) Krabbenhoft, D. P. Aquatic Cycling of Mercury in the Everglades
(ACME); U.S. Geological Survey: Accessed (23 May, 2000) at
<http://sofia.usgs.gov/projects/evergl_merc>.
(30) Krabbenhoft, D. Mercury Studies In The Florida Everglades; Fact
Sheet FS-166-96; U.S. Geological Survey: Middleton, WI, 1996.
(31) Izuno, F. T.; Sanchez, C. A.; Coale, F. J.; Bottcher, A. B.; Jones,
D. B. J. Environ. Qual. 1991, 20, 608.
(32) Hurley, J. P.; Krabbenhoft, D. P.; Cleckner, L. B.; Olson, M. L.;
Aiken, G. R.; Rawlik, P. S. Biogeochemistry 1998, 40, 293.
(33) Miles, C. J.; Fink, L. E. Arch. Environ. Contam. Toxicol. 1998, 35,
549.
(34) Wayne, D. M.; Warwick, J. J.; Lechler, P. J.; Gill, G. A.; Lyons, W.
B. Water, Air, Soil Pollut. 1996, 92, 391.
(35) Hallock, R. J.; Hallock, L. L. Detailed Study of Irrigation Drainage
In and Near Wildlife Management Areas, West-Central Nevada,
1987-90, Part B, Effect on Biota in Stillwater and Fernley Wildlife
Management Areas and Other Nearby Wetlands; Water Resources
Investigations Report 92-4042B.; U.S. Geological Survey: Carson
City, NV, 1993.
(36) Hoffman, R. J.; Hallock, R. J.; Rowe, T. G.; Lico, M. S.; Burge, H.
L.; Thompson, S. P. Reconnaissance investigation of water quality,
bottom sediment, and biota associated with Stillwater Wildlife
Management Area, Churchill county, Nevada, 1986-87; Water-
Resources Investigations Report 89-4105; U.S. Geological Sur-
vey: Carson City, NV, 1989.
(37) Ganguli, P. M. Master's Thesis, University of California, Santa
Cruz, CA, 1998.
(38) Ganguli, P. M.; Mason, R. P.; Abu-Saba, K. E.; Anderson, R. S.;
Flegal, A. R. Environ. Sci. Technol. 2000, in press.
(39) Marvin-DiPasquale, M.; Oremland, R. S. Methylmercury Forma-
tion and Degradation in Sediments of the Carson River System;
Interim Report I to U.S. EPA. Region IX; U.S. Geological Survey:
Menlo Park, CA, 1999.
(40) Gill, G. A.; Fitzgerald, W. F. Mar. Chem. 1987, 20, 227.
(41) Bloom, N. S.; Fitzgerald, W. F. Anal. Chim. Acta 1988, 208, 151.
(42) Horvat, M.; Liang, L.; Bloom, N. S. Anal. Chim. Acta 1993, 282,
153.
(43) Bloom, N. Can. J. Fish. Aquati. Sci. 1989, 46, 1131.
(44) APHA. In Standard Methods for the Examination of Water and
Wastewater, 15th ed.; Franson, M. A. H., Ed.; Am. Public Health
Association: Washington, DC, 1981; p 97.
(45) Cline, J. D. Limnol. Oceanogr. 1969, 14, 454.
(46) Ulrich, G. A.; Krumholz, L. R.; Suflita, J. M. Appl. Environ.
Microbiol. 1997, 63, 1627.
VOL. 34, NO. 23, 2000 / ENVIRONMENTAL SCIENCE & TECHNOLOGY 9 4915
(47) Fossing, H.; J?rgensen, B. Biogeochemistry 1989, 8, 205.
(48) Dionex. Installation Instructions and Troubleshooting Guide for
the IONPAC AG4A-SC Guard Column/IONPAC AS4A-SC Ana-
lytical Column; Document No. 034528; Dionex Corporation:
Sunnyvale, CA, 1992.
(49) Nakamura, K.; Sakamoto, M.; Uchiyama, H.; Yagi, O. Appl.
Environ. Microbiol. 1990, 56, 304.
(50) Webster, J. G.; Swedlund, P. J.; Webster, K. S. Environ. Sci.
Technol. 1998, 32, 1361.
(51) Spain, J. C.; Pritchard, P. H.; Bourquin, A. W. Appl. Environ.
Microbiol. 1980, 40, 726.
(52) Oremland, R. S.; Boone, D. R. Internat. J. System. Bacteriol. 1994,
44, 573.
Received for review May 29, 2000. Revised manuscript re-
ceived August 31, 2000. Accepted September 5, 2000.
ES0013125
4916 9 ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 34, NO. 23, 2000