Sundareshwar, P. V., J. T. Morris, P. J. Pellechia, H. J. Cohen ...
1570
Notes
DRUFFELL, E. R. M., AND P. M. WILLIAMS. 1991. Radiocarbon in
west coast, p. 6–16. In M. Yoklavich [ed.], Marine harvest
seawater and organisms from the Pacific coast of Baja Cali-
refugia for West Coast rockfish: A Workshop. NOAA Techni-
fornia. Radiocarbon 33: 291–296.
cal Memorandum NOAA-TM-NFMS-SWFSC-255.
FERRON, A., AND W. C. LEGGETT. 1994. An appraisal of condition
, AND D. F. HOWARD. 1995. On the development of year-
measures for marine fish larvae. Adv. Mar. Biol. 30: 217–303.
class strength and cohort variability in two northern California
FRANTZ, B. R., M. KASHGARIAN, K. H. COALE, AND M. S. FOSTER.
rockfishes. Fish. Bull. 93: 710–720.
2000. Growth rate and potential climate record from a rhodol-
RAU, G. H., J.-L. TEYSSIE, F. RASSOULZADEGAN, AND S. FOWLER.
ith using 14C accelerator mass spectrometry. Limnol. Oceanogr.
1990. 13C/12C and 15N/14N variations among size-fractionated
45: 1773–1777.
marine particles: Implications for their origin and trophic re-
GEARING, J. N. 1991. The study of diet and trophic relationships
lationships. Mar. Ecol. Prog. Ser. 59: 33–38.
through 13C natural abundance, p. 201–218. In D. C. Coleman
, D. G. AINLEY, J. L. BENGTSON, J. J. TORRES, AND T. L.
and B. Fry [eds.], Carbon isotope techniques. Academic.
HOPKINS. 1992. 15N/14N and 13C/12C in Weddell Sea birds, seals,
HERMANN, A. J., B. M. HICKEY, M. R. LANDRY, AND D. F. WINTER.
and fish: Implications for diet and trophic structure. Mar. Ecol.
1989. Coastal upwelling dynamics, p. 211–253. In M. R. Lan-
Prog. Ser. 84: 1–8.
dry and B. M. Hickey [eds.], Coastal oceanography of Wash-
, C. LOW, J. T. PENNINGTON, K. R. BUCK, AND F. P.CHAVEZ.
ington and Oregon. Elsevier Science.
1998. Suspended particulate 15N versus nitrate utilization: Ob-
LARSON, R. J., W. H. LENARZ, AND S. RALSTON. 1994. The distri-
servations in Monterey Bay, California. Deep-Sea Res. 45:
bution of pelagic juvenile rockfish of the genus Sebastes in the
1603–1616.
upwelling region off central California. Calif. Coop. Ocean.
, F. P. CHAVEZ, AND G. E. FRIEDERICH. 2001. Plankton 13C/
Fish. Invest. Data Rep. 35: 175–221.
12C variations in Monterey Bay, CA: Evidence of non-diffusive
LASKER, R. 1978. The relation between oceanographic conditions
inorganic carbon uptake by phytoplankton in an upwelling en-
and larval anchovy food in the California Current: Identifica-
vironment. Deep-Sea Res. 48: 79–94.
tion of factors contributing to recruitment failure. Rapp. P.-v.
ROBINSON, S. W. 1981. Natural and man-made radiocarbon as a
Reun. Cons. Int. Explor. Mer 173: 212–230.
tracer for coastal upwelling process, p. 298–302. In F. A. Rich-
LEAMAN, B. M., AND R. J. BEAMISH. 1984. Ecological and man-
ards [ed.], Coastal upwelling. American Geophysical Union.
agement implications of longevity in some northeast Pacific
SMITH, R. L. 1968. Upwelling. Oceangr. Mar. Biol. Ann. Rev. 6:
groundfishes. Inter. N. Pac. Fish. Comm. Bull. 42: 85–97.
11–46.
LENARZ, W. H. 1980. Shortbelly rockfish, Sebastes jordani: A large
SOUTHON, J. R., D. E. NELSON, AND J. S. VOGEL. 1990. A record
unfished resource in waters off California. Mar. Fish. Rev. 42:
of past ocean–atmosphere radiocarbon differences from the
34–40.
northeast Pacific. Paleoceanography 5: 197–206.
NORTON, E. C., R. B. MACFARLANE, AND M. S. MOHR. 2001. Lipid
STUIVER, M., AND T. F. BRAZIUNAS. 1993. Modeling atmospheric
class dynamics during development in early life stages of
14C influences and 14C ages of marine samples to 10,000 BC.
shortbelly rockfish and their application to condition assess-
Radiocarbon 35: 137–189.
ment. J. Fish Biol. 58: 1010–1024.
SUTHERS, I. M., A. FRASER, AND K. T. FRANK. 1992. Comparison
PEARCY, W. G., AND M. STUIVER. 1983. Vertical transport of car-
of lipid, otolith, and morphometric condition indices of pelagic
bon-14 into deep-sea food webs. Deep-Sea Res. 30: 427–440.
juvenile cod Gadus morhua from the Canadian Atlantic. Mar.
PENG, T.-H., R. M. KEY, AND H. G. OSTLUND. 1998. Temporal
Ecol. Prog. Ser. 84: 31–40.
variations of bomb radiocarbon inventory in the Pacific Ocean.
WILLIAMS, P. M., E. R. M. DRUFFELL, AND K. L. SMITH. 1987.
Mar. Chem. 60: 3–13.
Dietary carbon sources for deep-sea organisms as inferred from
RALSTON, S. 1995. The influence of oceanographic variables on
their organic radiocarbon activities. Deep-Sea Res. 34: 253–
time series of otolith growth in pelagic young-of-the-year Se-
266.
bastes spp., p. 97–118. In D. H. Secor, J. M. Dean, and S. E.
Campana [eds.], Recent developments in fish otolith research.
Received: 5 February 2001
University of South Carolina Press.
Accepted: 1 June 2001
. 1998. The status of federally managed rockfish on the U.S.
Amended: 14 June 2001
Limnol. Oceanogr., 46(6), 2001, 1570–1577
2001, by the American Society of Limnology and Oceanography, Inc.
Occurrence and ecological implications of pyrophosphate in estuaries
Abstract—Loading of bioavailable phosphorus, traditionally
demonstrate that Ppi is readily utilized by microbes in coastal
measured as soluble reactive phosphorus (SRP), contributes to
wetland sediments in the presence of nitrogen and carbon and
the eutrophication of aquatic ecosystems. However, polyphos-
can serve as a reservoir of orthophosphate. Thus, Ppi accu-
phates are also bioavailable but escape detection by the stan-
mulation in estuaries will subsidize the in situ biogeochemical
dard method used for measuring SRP. 31P nuclear magnetic
phosphorus cycle. This has important ecological implications
resonance spectrometric analysis of sediment extracts and en-
for trophic responses and estuarine productivity.
zymatic assay of surface waters reveal heretofore unreported
presence of pyrophosphate (Ppi) in coastal wetlands. We show
that the accumulation of Ppi (the smallest chemical form of
Phosphorus (P) plays a vital role in controlling biotic pro-
polyphosphate) in coastal wetlands is related to human impact
duction in a wide range of ecosystems, ranging from fresh-
and can occur in quantities that exceed that of SRP. We further
water lakes (Hecky and Kilham 1988) to open oceans (Clark
Notes
1571
Table 1. Relative concentrations of SRP, sulfide, and Ppi and the degree of high-impact urban (HIU) areas at sampling sites. SRP and
sulfide are concentrations in marsh pore water (geometric mean of monthly measurements for 2 yr integrated over 1 m depth). HIU is high-
intensity urban area within a 5-km radius around the sampling site. Ppi is pyrophosphate abundance in NaOH
EDTA extract from surface
sediments expressed as (% of NMR-visible and as mg kg 1 dry sediment for selected sites). Sediment total P (TP), total P in NaOH
EDTA extract, and extraction efficiency (% of sediment TP) are reported. Key: FM, BM, and SM are intertidal freshwater, brackish, and
salt marsh, respectively. FSM and FOR are salt marsh fertilized with (NH ) HPO and maritime forest, respectively. SM1 and SM2 are
4 2
4
impacted and unimpacted salt marsh, respectively. ND, no data.
Ppi (% of
TP
TP
Extract
SRP
S2
HIU
total NMR
sediment
extract
efficiency
Ppi
Estuary
Site
( M)
( M)
( 106 m2)
visible P)
(mg kg 1)
(mg kg 1)
(%)
(mg kg 1)
Cooper River
FM
3
5
0.29
4.2
1752
840
48
35.3
BW
11
15
3.9
12.5
921
396
43
49.5
SW
35
1,034
14.4
56.9
572
282
49
160.5
North Inlet
SM
4.4
127
0.09
0
64.5
39
61
0
FOR
ND
ND
0.09
0
ND
ND
—
—
FSM
35
151
0.09
0
ND
ND
—
—
Edisto River
SM1
46
166
1.16
5.5
933
383
41
21
SM2
69
451
0.21
0
626
300.5
48
0
Plum Island
BM
1.7
2.3
1.64
3.9
ND
ND
—
—
SM
7.1
508.9
0.66
0
ND
ND
—
—
et al. 1998). Phosphorus occurs naturally in various organic
available P pool in coastal wetlands and that its accumula-
and inorganic forms. Because estimation of these P pools is
tion in coastal wetlands is related to the degree of
based on analytical procedures that target specific classes of
anthropogenic impact.
P compounds, significant analytical overlap complicates the
We related the relative abundance and distribution of py-
interpretation of results. For instance, the soluble reactive
rophosphate (Ppi) in marsh sediments from 10 locations in
phosphorus (SRP) pool and the true ‘‘orthophosphate’’ pool
four different estuaries within South Carolina and Massa-
are both considered as bioavailable P. However, research
chusetts, U.S.A., to the degree of urban impact around these
from lakes has shown that analytical procedures employed
sites (Table 1). These estuaries differ in their nutrient status
to measure SRP overestimate the concentration of ortho-
and the degree of anthropogenic impact. Subsequently, we
phosphate because the acidic conditions of the reaction also
explored the implications of the presence of Ppi in coastal
hydrolyze some organic P compounds (e.g., Tarapchak et al.
wetland sediments in the context of bioavailable P pools.
1982). Nevertheless, novel techniques have been developed
that circumvent these problems (Karl and Tien 1992; Thom-
Study sites—For the purpose of this study, we categorized
son-Bulldis and Karl 1998). Although, studies of eutrophi-
our sampling locations in the following categories based on
cation have traditionally focused on SRP, other forms of P,
the urban impact estimates around each site (see below). Out
such as organic P compounds and polyphosphates, are also
of a total of 11 sampling locations in four different estuaries
important. However, some of these forms cannot be detected
in South Carolina and Massachusetts, five were classified as
by the standard method used for measuring SRP (Strickland
impacted, three as minimally impacted reference sites, one
and Parsons 1972). For this reason, considerable effort is
as maritime forest, and two as fertilized sites located in a
being expended to estimate the relative proportion of various
minimally impacted estuary (Table 1). The impacted sites
P pools and their ecological significance in a wide range of
included three intertidal marshes located along the estuarine
ecosystems. In fact in certain samples, it is often observed
salinity gradient on the urbanized Cooper River estuary in
that the sum of the concentration of P in various analytically
South Carolina, a salt marsh (SM1) on the Edisto River
defined P pools is less than the total P in the sample. This
(ACE Basin) in South Carolina, and a brackish marsh (BM)
discrepancy was attributed to a missing P pool, the occur-
in the Plum Island estuary, Massachusetts. Three other salt
rence of which suggests (although indirectly) the presence
marshes located in the minimally impacted North Inlet es-
of inorganic polyphosphates in the sample (Thomson-Bulldis
tuary (in South Carolina), the Edisto River (ACE Basin)
and Karl 1998). Central to these investigations is the quest
(SM2), and the Plum Island estuary served as reference sites.
to understand the sources and ecological relevance of these
Furthermore, two fertilized plots, adjacent to the reference
operationally defined P pools. Although it has been shown
site in the North Inlet estuary, were sampled to test the effect
that certain organic P pools are enzymatically remineralized
of duration of fertilization and the type of fertilizer used on
and that such remineralization is indeed ecologically signif-
the occurrence of Ppi. One of the fertilized plots has been
icant (e.g., Clark et al. 1998), bioavailability studies and
fertilized with a commercial P fertilizer, Triple Superphos-
characterization of missing and other nonreactive P pools is
phate, for 2 yr with an annual dosage of 15 mol P m 2. The
far from complete. We show here that pyrophosphate
second experimental plot was fertilized for 15 yr with re-
(P O (variably charged 1 to 4 )), the smallest chemical form of in-
agent-grade ammonium phosphate to achieve a final annual
2
7
organic polyphosphate, can also contribute to the total bio-
dosage of 30 mol N m 2 and 15 mol P m 2. Additionally,
1572
Notes
one maritime forest site in the North Inlet estuary was also
plex data points, which were processed with 10 Hz line
sampled.
broadening. Signal averaging varied with each sample as
shown (Fig. 1). Chemical shifts are relative to 85% phos-
Estimation of urban impact—The location of each sam-
phoric acid. External standards were run as a mixture of 1
pling site in South Carolina and Massachusetts was deter-
mM solution of inorganic orthophosphate (KH PO ), repre-
2
4
mined using global positioning system (GPS), and a buffer
sentative phosphomonoesters (orthophospho-L-serine and
radius of 5-km was generated around each site using geo-
DL-glycerophosphate), and pyrophosphate (tetrasodium Ppi,
graphic information system (GIS). The buffer around each
Na P O ), in a matrix of 0.5 M NaOH
0.1 M EDTA. Peak
4
2
7
sampling site was overlaid with the respective land use and
assignments based on external standards (Fig. 1A) and lit-
land cover data to determine the type (e.g., urban, open wa-
erature (e.g., Cade-Menun and Preston 1996) are as follows:
ter) and quantity of land use and land cover within it. For
orthophosphate at 5.2 ppm, phosphomonoesters (between 3.4
the South Carolina study sites on the Cooper River estuary,
and 4.8 ppm), and a diester peak at around 0 ppm and a Ppi
the North Inlet estuary, and the Edisto River (ACE Basin),
peak at around
5.6 ppm. Higher order polyphosphates
the source data for the land use and land cover data layer
(e.g., sodium tripolyphosphate, adenosine triphosphate
were SPOT 20-m satellite images (classified according to the
[ATP], and ammonium tetrapolyphosphate) typically have
Anderson level I/II scheme) acquired during leaf-off condi-
multiple peaks at chemical shifts
6 ppm (Robitaille et al.
tions in 1989 and 1990 and rectified using a statewide da-
1991). To estimate the Ppi concentration in NaOH
EDTA
tabase of GPS points as ground control points. For the Mas-
extract of wetland sediments, we relied on the total P in the
sachusetts study sites on the Plum Island estuary, the
extract and the percent contribution of Ppi to this total P as
statewide 1 : 25,000 land use data layer was used to deter-
calculated from the 31P NMR spectrum for corresponding
mine the land use characteristics around each site, developed
sites. These values were converted to Ppi (mg kg 1 dry sed-
originally from 1 : 25,000 aerial photography acquired in
iment) for select sites based on corresponding sediment total
1971. Updates to the original land use data layers were com-
P and the extraction efficiency.
pleted using 1 : 40,000 9 by 9 color infrared photographs
acquired during 1984 and 1985. Further updates to the coast-
Ppi in surface and pore water samples—Ppi in surface
al areas of Massachusetts, including our study sites, took
and pore water was estimated using an enzymatic assay (Sig-
place in 1991 when the classification system was also re-
ma product P-7275) after optimizing it for natural samples.
vised to include additional classes (http://www.magnet.
This method was originally developed to monitor the pro-
state.ma.us/mgis/lu-doc.htm). We expressed urban impact as
duction of Ppi in biochemical studies (e.g., O’Brien 1976),
the high-impact urban (HIU) area within the 5-km radius
and consists of a cascade of enzymatic reactions, of which
around our sampling sites, estimated using ArcView
and
the first enzyme is Ppi dependent. In this assay, 2 mol of the
the land use classification as defined under the National Land
reduced form of nicotinamide adenine dinucleotide (NADH)
Cover Data (NLCD). The HIU category includes the two
is oxidized to NAD for every mole of Ppi consumed. This
NLCD class definitions that target high-intensity residential
reaction is monitored spectrophotometrically at 340 nm. Be-
areas (including apartment complexes and row houses but
cause organic acids also absorb strongly at this wavelength,
excluding single-family housing units which typically dom-
potential problems can arise when attempting to use this
inate suburban development) and urban areas under com-
technique to directly quantify the concentration of Ppi in
mercial/industrial/transportation use.
NaOH
EDTA extract of wetland sediments. This is be-
cause large amounts of organic matter is also extracted (as
31P nuclear magnetic resonance (NMR) analysis of sedi-
highly colored material) from the sediments in the NaOH
ments—Duplicate surface sediment (0–10 cm) cores were
EDTA extraction scheme, which could potentially interfere
pooled and extracted overnight at room temperature (using
with the absorbance at the wavelength of interest. Prior to
a soil : solution ratio of approximately 10 g dry sediment :
the use of the enzymatic assay (to measure Ppi in surface
100 ml 0.5 M NaOH
0.1 M ethylenediaminetetraacetic
and pore water samples), we performed a laboratory test to
acid [EDTA]) (Cade-Menun and Preston 1996). Extraction
evaluate the effect of changes in ionic strength of the water
efficiencies were calculated in a parallel experiment. Sedi-
sample on the performance of the kit. We found that the kit
ment total P was measured using x-ray fluorescence and wet
performed very well in a simulated salinity gradient created
combustion techniques. Total P in NaOH
EDTA extract
using NaCl. Additionally, a comparison of actual and esti-
were measured by acid persulfate digestion. Extraction ef-
mated Ppi concentration in a series of standards (5
M to
ficiencies are reported in Table 1. Comparable extraction ef-
200
M) showed that this enzymatic kit predicts the Ppi
ficiencies suggest that the relative differences in distribution
concentration in water samples very well (slope
1.09, r2
and abundance of P species at and among sites were not due
0.96, n
10). However, for Ppi concentrations that are
to extraction artifacts. For 31P NMR analysis, sediment ex-
below 50
M, it is helpful to maximize the ratio of sample
tracts were centrifuged at 3,200 rpm for 10 min, and 3 ml
volume : reagent mixture volume while keeping the total vol-
of the supernatant was used for NMR analysis after adding
ume of the reaction mixture constant. To the best of our
0.3 ml D O. The pH of sediment extracts was between 11.5
knowledge, this is the first use of this method to detect Ppi
2
0.4 pH units. All spectra (202.46 MHz) were collected on
in natural water samples.
a Varian Inova 500 spectrometer. Data were collected with
a 200 ppm window centered at 0 ppm. A 45 pulse width
Pore water nutrients—SRP (Strickland and Parsons 1972)
with 2.1 s interpulse delay was used to collect 24,320 com-
and sulfides (Otte and Morris 1994) in marsh pore water
Notes
1573
Fig. 1.
Solution 31P NMR spectra of sediment extracts from four representative salt marsh sites
from two watersheds. A peak corresponding to Ppi is found at
5.6 ppm in the NMR scans. (A)
External standards run in the NaOH
EDTA matrix. (B) Salt marsh located on the heavily indus-
trialized Cooper River. Total number of scans
112,000. (C) A salt marsh located in the pristine
North Inlet estuary. (D) A fertilized plot in the pristine North Inlet estuary, fertilized with ammo-
nium phosphate (a laboratory chemical that does not contain Ppi). (E) An experimentally fertilized
plot in North Inlet estuary (fertilized with commercial Triple Superphosphate). (F) Saturated solution
of commercial fertilizer Triple Superphosphate. Total number of scans
102,000 for (C) and (F).
Total number of scans
40,000 for (C) and (D). Note the absence of a peak corresponding to Ppi
( 5.6 ppm) in (C) and (D) and its presence in (A), (B), (E), and (F). Additional details on peak
assignment are provided in the text titled 31P NMR analysis of sediments.
were measured using diffusion samplers equilibrated for 1
(500
M) as a source of nitrogen and carbon. SRP was mea-
month (Sundareshwar and Morris 1999). Pore water at each
sured (Strickland and Parsons 1972) in an aliquot after a 48-
site was sampled in triplicates at 10, 25, 50, 75, and 100 cm
h incubation. Appropriate controls were carried out with or
depths. Depth-integrated geometric means of at least 12 sam-
without the addition of either Ppi or the microbial inocula.
pling dates at each site are reported here. These pore water
In addition, we also included a killed control, where toluene
constituents were measured to evaluate the occurrence of Ppi
was added as a poison. All treatments were in triplicate. The
in coastal wetlands in relation to the biologically available
treatment codes are detailed in the figure legend. The second
SRP and redox conditions.
experiment (Fig. 3B) was a parallel experiment designed as
above. However, here we used sediments from only the salt
Bioavailability of Ppi—The bioavailability of Ppi was
marsh on the Cooper River estuary. This experiment was
tested in two different experiments (Fig. 3A,B). In the first
run in duplicate, and the appropriate treatment codes are de-
experiment (Fig. 3A), surface sediment from salt marshes on
tailed in the figure legend. Duplicates of a treatment with
the heavily urbanized Cooper River estuary and the rela-
added Ppi with sediment slurry inoculum were run in three
tively pristine North Inlet estuary, South Carolina, were slur-
batches sacrificed after 24, 48, and 72 h of incubation in the
ried and used as a source of microbial inocula. These were
dark at room temperature. The two controls—Ppi addition
used to inoculate an aqueous medium supplied with Ppi (500
without inoculation and microbial inoculation of the aqueous
M) as the sole source of added P and ammonium acetate
medium without the addition of Ppi—were sacrificed after 3
1574
Notes
Fig. 2.
Relationship between Ppi occurrence and high-impact urban (HIU) area. Two relation-
ships are shown, where (1) Ppi is expressed as mg kg 1 sediment dry wt; Y
11.65X
5.95; r2
0.963, P
0.0001, n
8, excluding the two Plum Island estuary (PIE) sites, and (2) Ppi is
expressed as % NMR-visible P; Y
3.94X
0.58; r2
0.989, P
0.0001, n
10.
d of incubation. The net production of SRP and consumption
tems that are affected by anthropogenic activities may ex-
of added Ppi were monitored in aliquots from each replicated
hibit a divergence from this pattern. Such a shift in relative
treatment during the course of the experiment.
contribution of various P pools to the total P obviously will
depend on the forms in which P loading occurs. For instance,
Ppi occurrence and urban impact—We detected the pres-
in contrast to the high concentration of Ppi found in Cooper
ence of Ppi in sediments from the coastal wetlands that were
River salt marsh sediments (160.5 mg kg 1), sediments from
classified as impacted (Table 1). For instance, sediments
a salt marsh receiving inconsequential terrigenous inputs
from all three locations along the salinity gradient on the
from a small, undeveloped watershed did not contain de-
urbanized Cooper River showed the presence of Ppi. Simi-
tectable levels of Ppi (Fig. 1C). Similarly, sediments from
larly, we detected Ppi in sediments from the impacted salt
the reference site in the ACE Basin (SM2, in South Carolina)
marsh (SM1) in the Edisto River (ACE Basin), which is
did not show the presence of Ppi, even though this site sup-
influenced by a local marina and a nearby housing devel-
ported similar pore water concentrations of phosphate and
opment. The impacted brackish marsh in the Plum Island
sulfides to the impacted site (SM1) in the ACE Basin (Table
estuary that occupies a more developed, suburban landscape
1). The third reference site (SM), located in the Plum Island
(when compared to the reference site in the Plum Island
estuary, Massachusetts, also did not show the presence of
estuary) also showed the presence of Ppi in marsh sediments.
Ppi (Table 1). Furthermore, in contrast to previous reports
The contribution of Ppi to the total NMR-visible P at these
(Gressel et al. 1996; Preston and Trofymow 1998), we did
sites varied and was related to the degree of urban impact
not detect Ppi from a forest site (FOR) located in the un-
estimates for the corresponding sites. For instance, the great-
developed watershed of the North Inlet estuary. Collectively,
est concentrations in sediment, pore water, and surface wa-
our data show a trend of increasing Ppi concentration with
ters among the sites examined were in the salt marsh near
increasing degree of urbanization (Fig. 2).
the city of Charleston, South Carolina (Fig. 1B). This site
Of course, the type of urban impact is probably a stronger
(SM) is at the mouth of the highly urbanized Cooper River
determinant of Ppi loading than is the total urban area
estuary. Ppi in sediments at this site accounted for approx-
around a site. For instance, total urban area (expressed as a
imately 57% of the total extractable P that was NMR-visible.
percentage of total land cover) around the Cooper River
We also measured a concentration of 33.8
7.3
M of Ppi
brackish marsh is 1.02%, but the HIU area here is 3.9 million
in surface waters, where the corresponding concentration of
m2, whereas the total urban area (expressed as a percentage
SRP was only about 2
M. Although in minimally impacted
of total land cover) around the Plum Island estuary brackish
or oligotrophic ecosystems the contribution of SRP to the
marsh is 13.17%, but the HIU area here is only 1.64 million
total P pool is higher than that of any other P pool, ecosys-
m2. This reflects the relative differences in the type of urban
Notes
1575
impact around these sites. The urban area around the brack-
ish marsh site on the Cooper River is dominated by indus-
tries, whereas the type of urban area around the brackish
marsh site on the Plum Island estuary is mainly suburban
development. Ppi and other polyphosphates have wide in-
dustrial and domestic applications (e.g., Cordon et al. 1997),
and correspondingly, the Cooper River brackish marsh site
supports relatively higher concentration of Ppi than the Plum
Island estuary brackish marsh site (Table 1). Thus, the HIU
area within a 5-km radius around sampling sites is a good
indirect index of Ppi loading (Fig. 2).
The sources of Ppi in these coastal wetlands are exoge-
nous. In addition to the link between Ppi occurrence and
urban impact demonstrated above, the pattern of occurrence
of Ppi in the experimentally fertilized plots located in the
North Inlet estuary also supports this hypothesis. The ex-
perimental plot that has been fertilized annually with 30 mol
N m 2 and 15 mol P m 2 for 15 yr with reagent-grade
(NH ) HPO did not accumulate detectable levels of Ppi (Fig.
4 2
4
1D). In contrast, sediments from the second fertilized plot
within the North Inlet estuary that was fertilized annually
(15 mol m 2 of P) with a commercial P fertilizer (Triple
Superphosphate) for 2 yr showed the presence of Ppi (Fig.
1E). In this case however, 31P NMR analysis of this com-
mercial P fertilizer confirmed that it contains Ppi (Fig. 1F).
Thus, we detected Ppi in only those plots that were fertilized
with commercial fertilizer containing Ppi.
This is the first report of the existence and sources of Ppi,
the smallest chemical form of polyphosphate, in marine en-
vironments. Although it has been found in forest soils (Gres-
sel et al. 1996; Preston and Trofymow 1998) and lakes (Hup-
fer and Gachter 1995; Carman et al. 2000), its origin was
unclear or was suggested to be biogenic in nature. By con-
trast, in estuaries, we show a clear link between Ppi accu-
mulation and human impact (Fig. 2).
Ppi and other polyphosphates have wide industrial appli-
cations (Monsanto 1996). Ppi has also been used as a P
fertilizer in agriculture (Louge 1961), and from 1955–1960
in the United States alone, annual use of Ppi and sodium
tripolyphosphate was at least 5.44 and 45.35
106 kg of P
equivalent, respectively (Louge 1961). More recently, orga-
no-P fertilizers, such as chitosan-polyphosphate complex,
Fig. 3.
(A) Utilization of Ppi by microbial inocula from North
have also been synthesized (Frossard et al. 1994). Urban
Inlet (NI-SM) and Cooper River (CR-SM) salt marsh sediments in
wastewater also contains Ppi and polyphosphates (Florentz
an aqueous media supplied with Ppi as the sole source of P and
et al. 1983). In addition, the more unstable long-chain po-
ammonium acetate as a source of nitrogen and carbon. Shown is
lyphosphates may breakdown to the relatively more stable
the mean ( 1 SE, n
3 per treatment) production of orthophos-
Ppi form, with an abiological half-life as great as 230 yr
phate by treatment following a 48-h dark incubation at room tem-
(Monsanto 1996). Consequently, the occurrence of Ppi in
perature, where
and
indicate the presence or absence, respec-
marsh sediments can be related to human impacts.
tively. Toluene (Tol.) was added in the killed controls to inhibit
There are also biogenic Ppi sources, but they alone cannot
microbial activity. Significant differences (Tukey at
0.05) in
explain its distribution among our sample sites. For instance,
orthophosphate release between Ppi
sed. and Ppi
sed. treat-
bacteria may accumulate polyphosphates in the presence of
ments are shown as *. (B) Ppi breakdown in aqueous medium in-
oculated with Cooper River–SM sediment slurry. Shown are the
excess orthophosphate (e.g., Kromkamp 1987), but Ppi and
production of orthophosphate and the hydrolysis of Ppi following
pore water SRP concentrations were not correlated in our
1, 2, and 3 d of incubation. The controls are shown as
Ppi
sed.
study sites (P
0.74, n
9) (Table 1). In addition, a fer-
and Ppi
sed. Note that the net release of SRP and the hydrolysis
tilized plot (FSM) in the pristine North Inlet estuary that
of Ppi from these controls are over a 72-h period. A significant
was fertilized with (NH ) HPO did not accumulate Ppi to
4 2
4
increase (
0.05) in orthophosphate concentration during 1,2, and
detectable levels, even though the pore water SRP concen-
3 d of incubation over Ppi
sed. is shown as *. Error bars represent
tration at this site was similar to that at the SM site on the
1 SD, n
2.
urbanized Cooper River estuary (Table 1). This indicates that
1576
Notes
the presence of Ppi in Charleston Harbor is not a conse-
hydrolysis (under identical conditions) (Fig. 3B). Based on
quence of excessive inorganic orthophosphate loading per
the consumption rate of Ppi in this experiment, we calculated
se. However, Ppi is involved in many cellular biochemical
the rate of biotic hydrolysis of Ppi to be approximately 7%
reactions (e.g., sulfate reduction) and has been proposed as
d 1, a rate that is consistent with a previous report (Al-Kan-
an evolutionary precursor of ATP (Lipmann 1984). In biotic
ani and MacKenzie 1990). Consistent with earlier studies
reactions, Ppi is hydrolyzed by specific enzymes such as
where Ppi has been shown to support the growth of various
pyrophosphatase, and some sulfate reduction reactions are
groups of bacteria from salt and freshwater marshes (Liu et
‘‘pulled’’ to completion by the hydrolysis of the generated
al. 1982), we found that in the above experiment the slope
Ppi (Ware and Postgate 1970). Furthermore, the presence or
of Ppi hydrolysis and net production of SRP was less than
absence of Ppi and other polyphosphates have been specu-
2 (the theoretical slope of Ppi hydrolysis and the gross yield
lated to be controlled by differences in redox conditions
of SRP), suggesting that a fraction of the Ppi and the regen-
(Carman et al. 2000). In their study, they found Ppi in oxic
erated SRP may indeed have been used by bacteria for
sediments from the lakes tested but not from the Baltic Sea
growth. In laboratory experiments, Ppi amendments stimu-
sediment samples. However, in our study the pore water sul-
lated bacterial production in sediments from the reference
fide concentrations for the reference sites (where we did not
salt marsh in the North Inlet estuary (data not shown), al-
detect Ppi) ranged from 127 to 508.9
M (Table 1), whereas
though the hydrolysis of Ppi was not monitored in this case.
the pore water sulfide concentrations in our impacted sites
These data demonstrate that the loaded Ppi is actively in-
(where we detected Ppi to varying degree) ranged from 2.3
corporated into the P biogeochemical cycle.
to 1,034
M (Table 1). In our study, in contrast to the find-
The physicochemical properties of polyphosphates, which
ings of Carman et al (2000), Ppi in sediments accounted for
make them suitable for treatment of hard water, also imply
about 57% of the total NMR visible P at the SM site on the
that the loaded polyphosphates will accumulate in the coastal
Cooper River estuary, which also supported the highest pore
zone where concentrations of complexing cations are the
water sulfide concentration ( 1 mmol L 1). Interestingly, the
highest. This is consistent with the increase in precipitation
SM site also had the highest degree of urban impact among
of Ppi that we observed with increasing salinity (data not
our sites (Table 1). Furthermore, the two salt marsh sites in
shown) and suggests that the marsh surface of coastal wet-
the ACE basin supported similar pore water sulfide and SRP
lands acts as a sink for Ppi. Another important factor that
concentrations, but we detected Ppi only in sediments from
can regulate the sink strength of the coastal wetlands for P
the impacted location. Thus, we detected Ppi in marsh sed-
is the P sorption capacity of wetland sediments. For instance,
iments from sites that were classified as impacted and not
the P sorption capacity of Cooper River wetland sediments
from our reference sites (irrespective of the prevalent redox
declines along the salinity gradient (Sundareshwar and Mor-
conditions and bioavailable SRP concentrations). Further-
ris 1999), whereas the mean depth-integrated (up to 1 m)
more, the relationship between sulfide versus percent Ppi (r2
pore water SRP concentration increases (Table 1). Interest-
0.56, P
0.02, n
9) is weaker than the relationship
ingly, the pore water Ppi concentrations also increase along
between HIU and percent Ppi (r2
0.989, P
0.0001, n
the salinity gradient on the urbanized Cooper River. For in-
10). This further suggests that the sources of Ppi are ex-
stance, pore water Ppi concentrations (at 10 cm deep) in-
ogenous. The degree of accumulation of Ppi in coastal wet-
creased from 55.4
53
M at the freshwater site to 74.7
land sediments appears to reflect a legacy of historical trends
18.6
M at the brackish marsh site, with the salt marsh
in land use and urbanization within coastal watersheds. Col-
site supporting the highest concentration of 157
55
M.
lectively, our data suggest that the loading of Ppi in coastal
Thus, Ppi represents a greater source of available P than SRP,
wetland sediments is anthropogenic in origin, although its
particularly at the salt marsh site on the Cooper River. Be-
utilization is biological.
cause the hydrolysis of 1 mol of Ppi yields 2 mol of ortho-
We have found that the biological hydrolysis of Ppi is a
phosphate, Ppi trapped in the sediment matrix at this site can
function of the nutrient status of sediment microbes. When
act as an ecologically significant reservoir for P, depending
we provided Ppi as the sole source of P in the presence of
on the activity of pyrophosphatase (an enzyme that hydro-
excess nitrogen and carbon, microbial inocula from urban-
lyzes Ppi).
ized Cooper River and pristine North Inlet salt marsh sedi-
The bioavailability of Ppi and its ability to support the
ment slurries produced 120 and 160
M SRP, respectively
growth of heterotrophic bacteria from different environments
(Fig. 3A). Note that the killed controls do show some release
(Liu et al. 1982) suggests that even in the presence of high
of SRP. This is most likely due to residual enzyme activity
ambient SRP concentrations, accumulation of Ppi will sub-
(of pyrophosphatase) released from lysed bacterial cells be-
sidize the estuarine P cycle and contribute to the mainte-
cause of the addition of toluene. However, the amount of
nance of high in situ SRP concentrations. Importantly, this
SRP liberated is still significantly lower than the treatment
accumulated Ppi is not detectable by the standard molybdate
that includes both Ppi and live microbes. This clearly shows
method of SRP determination. Given the influence of P on
that utilization of Ppi is an active process, which occurs in
important trophic compartments in estuarine and other eco-
both impacted and pristine sites. Algae and bacteria have
systems (e.g., Thingstad et al. 1998; Ulanowicz and Baird
been shown to scavenge polyphosphate in the presence of
1999), it will be necessary to examine other previously un-
excess nitrate (Solorzano and Strickland 1968).
detected forms of bioavailable P, such as Ppi, to fully un-
Additionally, in a parallel experiment where we monitored
derstand the effects of nutrient loadings. Our results suggest
the net release of SRP and the consumption of Ppi, we found
that the full extent of bioavailable P accumulation in estu-
that the amount of net SRP released was a function of Ppi
aries is unknown because of the presence of Ppi.
Notes
1577
P. V. Sundareshwar
R. F. POWERS. 1996. Linkages between phosphorus and carbon
transformation in a forest soil. Biogeochemistry 33: 97–123.
Duke University Wetland Center
HECKY, R. E., AND P. KILHAM. 1988. Nutrient limitation of phyto-
Nicholas School of Environment and Earth Sciences
plankton in freshwater and marine environments: A review of
Duke University
recent evidence on the effects of enrichment. Limnol. Ocean-
Durham, North Carolina 27708
ogr. 33: 776–795.
HUPFER, M., AND R. GACHTER. 1995. Polyphosphate in lake sedi-
J. T. Morris
ments: 31P NMR spectroscopy as a tool for its identification.
Limnol. Oceanogr. 40: 610–617.
Department of Biological Sciences
KARL, D. M., AND G. TIEN. 1992. MAGIC: A sensitive and precise
University of South Carolina
method for measuring dissolved phosphorus in aquatic envi-
Columbia, South Carolina 29208
ronments. Limnol. Oceanogr. 37: 105–116.
KROMKAMP, J. 1987. Formation and functional significance of stor-
P. J. Pellechia
age products in cyanobacteria. N.Z. J. Mar. Freshwat. Res. 21:
H. J. Cohen
457–465.
L
Department of Chemistry and Biochemistry
IPMANN, F. 1984. Pyrophosphate as a possible precursor for ATP,
p. 133–135. In K. Matsuno, K. Dose, K. Harada, and D. L.
University of South Carolina
Rohlfing [eds.], Molecular evolution and protobiololgy. Plenum.
Columbia, South Carolina 29208
LIU, C. L., N. HART, AND H. D. PECK, JR. 1982. Inorganic pyro-
phosphate: Energy source for sulfate-reducing bacteria of the
D. E. Porter
genus Desulfotomaculum. Science 217: 363–364.
B. C. Jones
LOGUE, P. 1961. Utilization and economics of phosphorus and its
Belle W. Baruch Institute for Marine Biology
compounds, p. 987–1025. In J. R. Van Wazer [ed.], Phosphorus
and its compounds. V. 2. Technology, biological functions and
and Coastal Research
applications. Interscience.
University of South Carolina
MONSANTO. 1996. Pyrophosphates: Properties and I and I applica-
Columbia, South Carolina 29208
tions. A technical update. Monsanto.
O’BRIEN, E. W. 1976. A continuous spectrophotometric assay for
References
argininosuccinate synthetase based on pyrophosphate forma-
tion. Anal. Biochem. 76: 423–430.
AL-KANANI, T., AND A. F. MACKENZIE. 1990. Pyrophosphate hydro-
OTTE, M. L., AND J. T. MORRIS. 1994. Dimethylsulphoniopropionate
lysis in mineral fractions of soils, goethite, kaolinite, and mont-
(DMSP) in Spartina alterniflora Loisel. Aquat. Bot. 48: 239–259.
morillonite. Soil Science 149: 239–247.
PRESTON, C., AND J. A. TROFYMOW. 1998. Chemical and 31P NMR
CADE-MENUN, B. J., AND C. M. PRESTON. 1996. A comparison of
characterization of soil P in coastal forest chronosequences of
soil extraction procedures for 31P NMR spectroscopy. Soil Sci.
southern Vancouver Island: Effects of climate and disturbance.
161: 770–785.
Northwest Sci. 72: 46–47.
CARMAN, R., G. EDLUND, AND C. DAMBERG. 2000. Distribution of
ROBITAILLE, P.-M. L., P. A. ROBITAILLE, G. G. BROWN, JR., AND G.
organic and inorganic phosphorus compounds in marine and
G. BROWN. 1991. An analysis of the pH-dependent chemical-
lacustrine sediments: A 31P NMR study. Chem. Geol. 163:
shift behavior of phosphorus-containing metabolites. J. Mag-
101–114.
net. Resonance 92: 73–84.
C
S
LARK, L. L., E. D. INGALL, AND R. BENNER. 1998. Marine phos-
OLORZANO, L., AND J. D. H. STRICKLAND. 1968. Polyphosphate in
phorus is selectively mineralized. Nature 393: 426.
seawater. Limnol. Oceanogr. 13: 515–518.
C
S
ORDON, L. M., E. FROSSARD, H. NEWMAN, P. TEKELY, AND J.-
TRICKLAND, T. D., AND T. R. PARSONS. 1972. A practical handbook
L.M
of seawater analysis. J. Fish. Res. Board Can. 167: 1–311.
OREL. 1997. Use of 31P NMR in the study of soils and the
environment, p. 247–271. In M. A. Nanny, R. A. Minear, and
SUNDARESHWAR, P. V., AND J. T. MORRIS. 1999. Phosphorus sorp-
J. A. Leenheer [eds.], Nuclear magnetic resonance spectros-
tion characteristics of intertidal marsh sediments along an es-
copy in environmental chemistry. Oxford University Press.
tuarine salinity gradient. Limnol. Oceanogr. 44: 1693–1701.
F
TARAPCHAK, S. J., S. M. BIGELOW, AND C. RUBITSCHUN. 1982.
LORENTZ, M., P. GRANGER, AND P. HARTEMAN. 1983. Use of 31P
Overestimation of orthophosphorus concentrations in surface
nuclear magnetic resonance spectroscopy and electron micros-
waters of southern Lake Michigan: Efffects of acid and am-
copy to study phosphorus metabolism of microorganisms from
monium molybdate. Can. J. Fish. Aquat. Sci. 39: 296–304.
wastewaters. Appl. Environ. Microbiol. 47: 519–525.
THINGSTAD, T. F., U. L. ZWEIFEL, AND F. RASSOULZADEGAN. 1998.
FROSSARD, E., P. TEKLEY, AND J. L. MOREL. 1994. Chemical char-
P limitation of heterotrophic bacteria and phytoplankton in the
acterization and agronomic effectiveness of phosphorus applied
northwest Mediterranean. Limnol. Oceanogr. 43: 88–94.
as a polyphosphate–chitosan complex. Fertil. Res. 37: 151–158.
THOMSON-BULLDIS, A., AND D. KARL. 1998. Application of a novel
GRESSEL, N., J. G. MCCOLL, C. M. PRESTON, R. H. NEWMAN, AND
method for phosphorus determination in the oligotrophic North
Pacific Ocean. Limnol. Oceanogr. 43: 1565–1577.
Acknowledgments
ULANOWICZ, R., AND D. BAIRD. 1999. Nutrient controls on ecosys-
This work was supported in part by NSF, the South Carolina Sea
tem dynamics: The Chesapeake mesohaline community. J. Mar.
Grant Consortium, and a NERRs Graduate Fellowship to P.V.S
Syst. 19:(1–3) 159–172.
awarded by NOAA’s OCRM division. We thank Ronald Benner, C.
WARE, D., AND J. R. POSTGATE. 1970. Reductant-activation of in-
J. Richardson, and D. Vasudevan for their helpful comments. We
organic pyrophosphatase: An ATP-conserving mechanism in
also thank N. Flanagan for his assistance with GIS analysis and T.
anaerobic bacteria. Nature 226: 1250–1251.
Rogers, W. Willis, and P. Hiene for their assistance in the lab. Two
Received: 13 July 2000
anonymous reviewers provided helpful reviews, and we are thankful
Accepted: 23 April 2001
for their contributions.
Amended: 24 May 2001
