Download KIEBER, ROBERT J., LINDA H. HYDRO, AND PAMELA J. SEATON

Limnol. Ocennogr., 42(6), 1997, 1454-1462
0 1997, by the American Society of Limnology
and Oceanography,
Inc.
Photooxidation of triglycerides and fatty acids in seawater: Implication
toward the formation of marine humic substances
Robert J. Kieber, Linda H. Hydra, and Pamela J. Seaton
Department of Chemistry and Marine Science Program, University
Wilmington, North Carolina 28403-3297
of North Carolina at Wilmington,
Abstract
Unsaturated fatty acids rapidly photodecomposed in filtered seawater to aliphatic aldehydes and an o-oxocarboxylic acid (CHO-(CH,),,-COOH) upon a 5-h exposure to ambient sunlight. The photoreactivity of fatty acids
increased as their degree of unsaturation increased. Palmitic acid displayed no photodegradation while monounsaturated oleic acid photodecomposed to nonanal and 9-oxononanoic, respectively. Photooxidation of the polyunsaturated linolcic acid produced significantly higher concentrations, primarily hexanal, and a wider variety of aldehydes.
The rate of photodegradation was more than 10 times greater for the polqtnsaturated fatty acids than for the
monounsaturated fatty acids. Photolysis of the triglyceride, trilinolein, gave En aldehyde profile similar to its component fatty acid, indicating each undergoes a similar mechanism of photodegradation. Twenty percent of the initial
triglyceride was photodegraded during the 6-h irradiation. The rapid photoreactivity of unsaturated fatty acids may
explain their low in situ levels relative to saturated fatty acids, the latter of which are found in the dissolved phase
at relatively higher levels.
Fluorescent hydrophobic dissolved organic matter, operationally defined as humic substances, increased after
irradiation of trilinolein-treated seawater over a 14-d period. The concentraticn of humic products was enhanced by
postphotolysis addition of ammonia. Dissolved ammonia concentrations declined as the dissolved organic nitrogen
of the extracted humics increased, indicating the added ammonia was becoming incorporated into the organic
structure.
Fatty acids are chemically and biologically labile compounds that comprise -3% of the total dissolved organic
matter (DOM) in seawater (Kattner et al. 1983b). They,
along with triglycerides, are involved in many biogeochemically significant processes occurring within the sea. Of particular importance is their potential role as precursors to several other important types of DOM, including aldehydes and
humic substances.
Several mechanisms have been proposed for the autooxidation of fatty acids to aldehydes in seawater. It has been
suggested that oleic acid (A9 C,,,,) is autooxidized to octanal, nonanal, and decanal while palmitoleic acid (A9 C,,,,) is
autooxidized to hexanal, heptanal, and octanal (Schauenstein
et al. 1977; Gschwend 1979; Gschwend et al. 1980, 1982).
o-Oxocarboxylic
acids have been detected in marine aerosols and the dominance of the C, o-oxoacid (CHO-(CH,),COOH) was proposed to be due to photooxidation of A9
unsaturated precursors such as oleic (A9 C,,,,), palmitoleic
(A9 C,,,,), linoleic (89,12 C,,:,), and linolenic (A9,12,15 C,,:,)
fatty acids (Kawamura and Gagosian 1987). Although mechanisms for the photooxidation of fatty acids have been proposed, they have never been verified experimentally in seawater. This paper describes photoproduction of aldehydes
and o-oxocarboxylic acids from fatty acids in seawater, under ambient sunlight conditions, as well as in situ aldehyde
concentration in four coastal marine waters.
Understanding the photolability of fatty acids and triglycerides in seawater is important for reasons beyond their role
Acknowledgments
Financial assistance was provided by the Center for Marine Science Research (No. 142) at UNCW. We also thank Andrea Li for
assistance in the nitrogen analyses.
in the production of aliphatic aldehydes and o-oxocarboxylic
acids. Fatty ac-.d and triglyceride photooxidation is significant because 0:’ the suggested role this process plays in the
formation of marine humic substances. As early as 1972,
researchers prcposed that marine humic acids form in situ
from degradation products of marine plankton, rather than
being transported from continents (Nissenbaum and Kaplan
1972). There are two prevailing views concerning the formation of marine humic substances. The polymerization
view considers that humic substances are macromolecules
formed through abiotic or enzyme-mediated polymerization
of simple organic molecules, including amino acids, fatty
acids, sugars, and phenols. However, it is very unlikely that
simple organic molecules will aggregate and polymerize to
humic substances, given the micro- or nanomolar concentrations at which they exist in ocean waters, or in large, dilute,
freshwater lakes, such as the Antarctic lakes, where extracted
humic substances were found to have significant similarities
to marine humics (M&night
et al. 199 1, 1994). A second
proposed pathway for humic substance formation is through
continued oxid ation, condensation, and transformation of existing biopolymers. Polyunsaturated lipids released into the
water column by plankton have been invoked as a potential
precursor for humic substance formation. Harvey and Boran
(1982a,b) and Harvey et al. (1983, 1984) autooxidized trilinolein in a seawater matrix under both light and dark conditions and found that the XAD-extracted product had infrared, ‘H-NMR, and 13C-NMR spectra that were very similar
to those of marine humic substances. The 613Cvalue of their
synthetic product was also similar to the S13Cvalues of natural marine fulvic acids (Harvey and Boran 1985). The authors proposed a mechanism that involves free radical autooxidative cross-linking of unsaturated fatty acid chains of
1454
Photooxidation
of triglycerides
the triglyceride, producing humic substances with both aliphatic and aromatic regions containing hydroxyl, ether, peroxide, and carbonyl functional groups. This mechanism provides an insightful starting point to our understanding of the
formation of marine humic substances. However, it does not
address the role of the photodegradation products of triglycerides, such as aldehydes, in the humification process. These
photooxidation products are potentially more reactive than
their precursors and may therefore be significant source material in the formation of hydrophobic DOM such as humics.
Marine humic substances are composed of between 3 and
5% nitrogen, although the chemical form of N in marine
humics is poorly known. Synthetic humic substances produced from trilinolein-treated seawater, however, contained
only 0.86% N (Harvey et al. 1984; Schnitzer 1985). The
possible photooxidation of triglycerides and fatty acids has
important ramifications with respect to the inclusion of N
into humics. In addition to being a direct source material for
humification, aldehyde photoproducts are much more reactive toward ammonia or amino-containing organic compounds, relative to the parent unoxidized molecules. The
question regarding the incorporation of N into marine humic
substances is particularly important because of the suggested
role humics have in the stimulation of primary productivity
(Moran and Hodson 1990, 1994; Carlsson and Graneli 1993;
Graneli and Moreira 1990; Moran et al. 1991). The primary
goals of this research are to describe, both qualitatively and
quantitatively, the degree of fatty acid and triglyceride photodegradation in seawater under ambient sunlight and to explore possible mechanisms for the incorporation of N into
humic substances resulting from the reactive photodegradation products.
Materials
and methods
Reagents and standards-All
chemicals were obtained
from Fisher Scientific or Aldrich Chemical and were reagent
grade or HPLC grade unless stated otherwise. Aldehydes
were purchased from Aldrich Chemical. Fatty acids and triglycerides were purchased from Sigma Chemical. 9-0x0nonanoic acid was synthesized through oxidative cleavage
(KMnOJNaIO,)
of oleic acid. The oxo-acid was converted
to its dinitrophenylhydrazone (DNPH) derivative, which was
purified through column chromatography for use as a standard (Seaton and Kieber unpubl.) Stock fatty acid (lo-15
mM) and triglyceride (-3 mM) standards were prepared in
acetonitrile. and chloroform, respectively. All fatty acid and
triglyceride stock solutions were stored at 4°C where they
were stable for several months. 2,4-DNPH was recrystallized
twice from a 70 : 30 (vol : vol) solution of acetonitrile and
water and a third time from 100% acetonitrile. Crystals were
dried under vacuum and stored in the dark in airtight 30-ml
Teflon vials. The derivatizing reagent was prepared by dissolving 20 mg of recrystallized 2,4-DNPH in a 15-ml solution containing concentrated HCl, water, and acetonitrile
(2:5: 1) (vol: vol : vol). Carbonyl contamination
was
removed daily with a 2-ml extraction using redistilled (inhouse) carbon tetrachloride (Ccl,) added to the reagent,
shaken on a wrist action shaker for 5 min, and centrifuged
and fatty acids
1455
at 2,000 rpm for 5 min. The organic layer was removed and
the reagent solution reextracted twice more as described
above. Stock solutions of aldehydes (-0.1 mM) were prepared in acetonitrile and kept in the dark at 4°C where they
were stable for several months. Immediately prior to HPLC
analysis serial dilutions of the stock were made in deionized
water (Milli-R06 Plus/Milli-Q Plus, Millipore) and derivatized with 2,4-DNPH.
HPLC instrumentation-The
HPLC mobile-phase gradient was generated by an E-Lab model 2020 gradient programmer connected to data acquisition software (OMS Tech)
installed in an IBM-compatible PC and to an Eldex model
B-100-S pump (Eldex Laboratories) containing an inert solenoid valve placed on its low pressure side. Derivatized,
concentrated, and eluted (from C,, cartridges) samples were
injected into a Valco six-port injector (Valco Instruments)
containing a 2,000+~,1 sample loop. Derivatized hydrazones
were separated on an 8-mm i.d. X 20-cm C,, reversed-phase
Radial-pak cartridge housed in a Radial compression module
(Millipore), detected by an Isco V4 variable wavelength absorbance detector (Isco) at 370 nm, and processed by the ELab system. A two-solvent gradient elution at 2 ml min- I
was utilized with solvent A as 10% acetonitrile in water, and
solvent B 100% acetonitrile: isocratic at 40% B for 10 min,
40% B to 75% B in 7 min, 75% B to 100% B in 11 min,
isocratic at 100% B for 8 min, then isocratic at 40% B for
9 min.
Aldehydes analysis -Aldehydes
were determined by the
method of Kieber and Mopper (1990). Filtered samples (50250 ml) were derivatized with 2,4-DNPH reagent for 1 h in
the dark at ambient temperatures. Derivatized hydrazones
were extracted and concentrated using C,, cartridges that
were precleaned and preconditioned by rinsing with 2 X 5
ml of acetonitrile followed by 2 X 5 ml of filtered Gulf
Stream seawater. To avoid contamination from plastic tubing, derivatized samples were placed in 50-ml glass-syringe
reservoirs (Fisher) that were connected in line to C,, cartridges. The derivatized aliquots were pulled through C,, cartridges via a dialysis pump (Cole-Parmer) at a flow rate of
3.5 ml min-l. After the sample was loaded onto the cartridge, salts were rinsed off with deionized water and the
remaining liquid blown out with air. Hydrazones were eluted
with 2 ml of acetonitrile into 7-ml Teflon vials, and 3 ml of
deionized water was added to mimic the HPLC mobile
phase. Diluted derivatized extracts were analyzed by HPLC,
Photochemical studies-Fatty
acids and trilinolein from
stock standards were added to 0.2~pm filter-sterilized Gulf
Stream seawater (1 FM and 0.33 ~.LM concentrations, respectively) prior to placement in six sterile 500-ml quartz,
round-bottomed flasks with ground glass quartz stoppers.
The quartz flasks containing the fatty acid- or triglyceridetreated seawater were placed directly on cork rings in direct
sunlight for a minimum of 5 h on a grassy area near the
chemistry building on the UNC-Wilmington
campus. Dark
samples were analogous in every way to the light samples
except they were wrapped in aluminum foil to restrict light
exposure.
1456
Kieber et al.
To determine the degree of thermal degradation of fatty
acids during the photolysis studies, various fatty acids were
added to 0.2~pm filter-sterilized
Gulf Stream seawater,
placed in 500-ml of round-bottomed quartz flasks wrapped
in aluminum foil, and submerged under isothermal conditions at 25 or 40°C for 5 h. The latter temperature was the
highest attained during the unthermostated photolysis studies. Samples were analyzed for aldehyde concentrations initially and after the 5-h exposure at the two different temperatures. No statistical difference was found between the
25 and 40°C samples (P < 0.01, n = 28), indicating undetectable thermal degradation of fatty acids at 40°C during
the 5-h exposure.
Photolysis of trilinolienPostphotolysis time studies of
trilinolein were also preformed, and the rate of humic substance formation in two separate experiments was determined. In the first experiment, trilinolein (0.33 ~.LM)in filtersterilized Gulf Stream seawater was irradiated (as described
above) and the water from three quartz flasks was combined
in a sterile 2-liter Erlenmeyer flask. Aliquots were taken over
a 14-d period for analysis of UV absorbance at 250 nm and
fluorescence, as described below. In the second experiment,
trilinolein was photolyzed as above, but after combining the
water from the three quartz flasks, ammonia was added to
final concentration of 1 ~.LM.Aliquots were again taken over
a 14-d period for absorbance and fluorescence. Ammonia
concentrations and dissolved organic nitrogen (DON) levels
in the C,, cartridge-extracted DOM in this latter experiment
were also measured, as described below.
Hydrophobic DOM extraction-Fluorescent,
hydrophobic
DOM, operationally defined here as humic substances, was
extracted using C,, cartridges by the method of Amador et
al. (1990). This technique was chosen over conventional
methods using XAD resins because C,, is more efficient at
removing the chromophoric humic substances from seawater. Amador et al. (1990) reported that C,, extraction of humic substances was between 22 and 84% more effective relative to XAD-2. Kuo et al. (1993) also recovered 83% of an
aquatic fulvic acid using C,, solid-phase extraction. In addition, Amador et al. (1990) found C,, extraction was better
able to retain the UV-visible and fluorescence characteristics
of the isolated humic material relative to XAD. Because
these were the characteristics of most interest in this study,
C,, extraction was the most logical choice of an extraction
procedure. A second advantage of using C,, extraction is that
methanol is the eluant rather than 0.1 M NaOH. Therefore
any degradation due to high pH is avoided. Note that some
of the unreacted triglycerides and fatty acids will also be
retained by the C,,. However, their fluorescence and absorbance are negligible relative to the isolated humic material.
C,, cartridges were preconditioned by washing with 2 X
5 ml methanol followed by 2 X 5 ml humic-free Gulf Stream
water. Humic-free Gulf Stream water was obtained by passing filtered Gulf Stream water through C,, cartridges, preconditioned by washing with 2 X 5 ml methanol, followed
by 2 X 5 ml of deionized water. Samples for humic extraction were filtered, acidified to pH 2.2-2.5 with 0.3 M HCl,
and loaded onto C,, cartridges. The cartridges were washed
with 2 X 5 n-l of deionized water to remove residual salts
and humic su3stances were eluted with 4 ml of methanol,
which was brought to a final volume of 5 ml with methanol.
The relative c,ancentration of humics in the extract was determined by fluorescence intensity of the extract in a manner
similar to Amador et al. (1990). Fluorescence was chosen
because it is a highly sensitive spectrophotometric technique
capable of detecting extremely small quantities of humic
substances prcduced. It is also highly selective because two
wavelengths are employed, allowing us to distinguish between other nonpolar entities that might be extracted by the
C,, packing. The sample was diluted 4/10 in methanol and
fluorescence rr easurements (Turner model 112) made against
a methanol blank at 360 nm excitation and 490 nm emission,
respectively. Fluorescence is reported as quinine sulfate units
(QSU) where 1 QSU is the fluorescence of a 1.3 nM solution
of quinine sull’ate in 0.1 N H,SO,. A 100 QSU solution in
0.1 N H,SO, v/as set to read -73 fluorescence units against
a 0.1 N H,SO, blank and readings converted to change in
(A) QSU liter- I relative to initial fluorescence values taken
immediately after photolysis. Relative humic concentrations
were also detemined by UV absorbance relative to a methanol blank at Z!50 nm and converted to a change in (A) absorbance liter- l relative to initial absorbance values taken
immediately after photolysis.
Nitrogen anczlyses-Total dissolved nitrogen was analyzed
using a persulfate oxidation method in which nitrogen compounds are oxidized to nitrate by potassium peroxodisulfate
under alkaline conditions (Valderrama 1981). The resulting
nitrate is quantified by Technicon Autoanalyzer. Ammonia
(NH,) concentrations were determined by the method of Parsons et al. (198 9) in which ammonia reacts with hypochlorite
under slightly alkaline conditions to form monochloramine,
which, in the presence of phenol, nitroprusside ions, and an
excess of hypcchlorite, gives indophenol blue that is measured at 630 nm.
Results
Photooxidation of fatty acids-Unsaturated
fatty acids
added to sterile seawater and irradiated for 5 h under ambient
sunlight conditions photooxidized to produce a variety of
aliphatic aldehydes (Figs. 1, 2; Table 1). There were undetectable changes in the concentration of aldehydes produced
in dark controls suggesting little or no degradation of the
parent fatty acd (Table 1). Photodegradation increased as
the degree of tinsaturation of the precursor fatty acid increased. Photodegradation of palmitic acid was negligible
(Fig. 1B) while oleic acid produced nonanal and 9-oxononanoic acid (Fig:;. lC, 2). Diunsaturated linoleic acid degraded to a greater extent relative to oleic acid and yielded higher
concentrations and a greater variety of aldehydes (Fig. 1E).
Hexanal and 9-oxononanoic acid were the principal aldehydes produced, along with smaller amounts of butanal, pentanal, and sevel-al unidentified carbonyl compounds. Triunsaturated linolenic acid also photooxidized primarily to
hexanal and 9-oxononanoic acid as well as several unidentified carbonyl compounds (Fig. 1F).
The rate of Jdehyde production, based on the primary
Photooxidation of triglycerides and fatty acids
1457
STANDARD
-..
PALMTIC
LINOLENIC$
7
Time (min)
Fig. 1. HPLC chromatograms of derivatized aldehydes after photolysis in the presence of 1 FM
fatty acid-treated filter-sterilized Gulf Stream seawater. (A) Aldehydes standard (45.5-364 nM
range). (B) Palmitic (C,,,,) acid. (C) Palmitoleic (A9 C,,,.) acid. (D) Oleic (A9 C,,,,) acid. (E)
Linoleic (A9,12 C,,,,) acid. (F) Linolenic (A9,12,15 C,,,.) acid. Cl-Cl0
= straight-chain aliphatic
aldehydes, MG = methyl glyoxal, GLY = glyoxal, 9-0X = 9-oxononanoic acid.
aldehyde produced and normalized to light intensity, is presented in Table 1. The rate was over 10 times greater for the
diunsaturated fatty acid and triglyceride relative to the monounsaturated species.The percent conversion of the initial fatty acid or triglyceride after photolysis is also presented in
Table 1. Twenty percent of the original triglyceride was photodegraded during the 6-h irradiation with lesser amounts of
degradation for the individual fatty acids. These should be
viewed as minimum conversion and production rates because several of the peaks in the chromatograms presented
in Fig. 1, especially for the diunsaturated species, were not
identifiable by our analytical scheme.
The above photooxidation of fatty acids to aldehyde results are consistent with earlier reports and proposed mechanisms (Jalliffier-Merlon et al. 1991; Gschwend 1979;
Gschwend et al. 1980,1982; Kattner et al. 1983b; Kawamura
and Gagosian 1987). Photolysis of unsaturated fatty acids
results in oxidative cleavage of the double bond, producing
aliphatic aldehydes and an o-oxocarboxylic acid. Oleic acid
(A9 C&, for example, produced nonanal and 9-oxononanoic
acid (1, 2, and 3, respectively, Fig. 2). 9-Oxononanoic acid
is also produced upon photolysis of palmitoleic (A9 C,,,,),
linoleic (A9,12 C&, and linolenic (A9,12,15 C,,,,) acids, all
of which have their first double bond at C,-,, (Fig. 1).
The photochemical lability of unsaturated fatty acids reported above may explain in part why in situ concentrations
of dissolved unsaturated fatty acids are low in natural waters
relative to their saturated counterparts (Scribe et al. 1991;
Hama 1991; Kattner et al. 1983a,b). Comparison of fatty
acid profiles of particulate and dissolved fractions from a
spring phytoplankton bloom in the North Sea (Kattner et al.
1983a,b) showed that unsaturated fatty acids were significantly less prevalent than saturated fatty acids in the dissolved fraction than in the particulate fraction. The dissolved
fraction contained -63% saturated, 32% monounsaturated,
and 6% polyunsaturated acids, whereas coexisting particles
contained -5 1% saturated, 37% monounsaturated, and 12%
polyunsaturated acids.
Photooxidation of trilinolein--The triglyceride trilinolein
containing three linoleic acid chains were also investigated
for its photochemical reactivity (Fig. 3). The primary photodegradation product was hexanal (66 &I). This concentration was similar in magnitude to the hexanal (42 nM) produced during photolysis of linoleic acid (Fig. 1). Photolysis
of trilinolein also produced relatively smaller amounts of
butanal, pentanal, and 9-oxononanoic acid, which is the
same suite of degradation products found upon irradiation
of linoleic acid. This similarity suggeststhat a similar mechanism is responsible for the photodegradation of the triglyceride and its component fatty acid.
Natural aldehydes-Natural aldehyde concentrations (Fig.
4) were determined at four coastal North Carolina sites (Fig.
5). Straight-chain, aliphatic aldehydes from butanal through
decanal and methyl glyoxal were detected at sites 1, 2, and
1458
Kieber et al.
A
0
10
30
Time (min)
0
2o
30
10
Time (rniii
1
Fig. 2. HPLC chromatogramsfrom photolysis of 1 PM oleic (A9
C,,:,). (A) Prior to photolysis. (B) After 5 h photolysis.
Fig. 3. HPLC chromatograms from photolysis of 0.33 FM trilinolein-treated filter-sterilized Gulf Stream seawater. (A) Prior to
photolysis. (B) After 5 h photolysis.
3 at relatively low concentrations (<2 nM). The more UVabsorbent waters of the Alligator River contained higher
concentrations (2-10 nM) of the same aldehydes found at
the other sites. Their presence in the water column is suggested to be due, in part, to the photooxidation of various
fatty acid precursors described above. The relatively low
concentrations reported here might result from the biological
and chemical lability of aliphatic aldehydes (Hydro 1996).
Formation qf humic substances from trilinolein photolys&-The photooxidative lability of unsaturated fatty acids
and triglycerides reported above may play an important role
in the production of marine humic substances. The generation of reactive photolysis products, such as 9-oxononanoic
acid, is significant because the reactivity of the aldehyde
functionality WI th ammonia or organic amines provides an
avenue for incorporation of nitrogen into the humic struc-
Table 1. Rate of aldehyde production, normalized to light intensity, from photolysis cBa variety of fatty acids and triglycerides. Percent
conversion is the difference between initial and final fatty acid concentrations where the final fatty acid concentration was estimated based
on the concentration of the primary aldehyde photoproduct generatedduring 5 h of irradiation. ND = not detectable.
Fatty acid
Trilinolein
Linoleic
Oleic
Palmitoleic
Initial aldehyde (nM)
4
1
ND
0.4
Aldehyde (nM)
Darkt=5h
Light t = 5 11
3
66
2
42
2.5
ND
0.6
2.6
Rate X 100
[nM(W-h)-I cm21
240
110
6
6
Percent conversion
20
4
0.3
0.3
Photooxidation
of triglycerides
1459
and fatty acids
Site 1 (Myrtle Grove)
J
s
8
E
8
8
E
$
2:
ss~s~~p
Aldehyde
Aldehyde
Site 3 (Chowan River)
8
s
8
s
s
Aldehyde
8
8
susssegg
$
Aldehyde
Fig. 4. Distribution of aldehydes in coastal North Carolina seawater.Error bars represent + 1 SD
based on y1= 3. Absorbance values at 300 nm were 0.06-o. 16 for sites 1, 2, and 3. Absorbance
value at 300 nm for site 4 was 1.7. Note scale change for site 4.
ture. In order to test this hypothesis, two separate, postphotolysis studies with trilinolein were performed and the rate
of humic formation monitored over time. In the first experiment, photolyzed trilinolein-treated seawater was the only
reactant. In the second experiment, after photolysis of the
trilinolein-treated
seawater, ammonia was added to a final
concentration of 1 FM. Aliquots, taken from each experiment, were analyzed for aldehydes, UV absorbance, and
fluorescence over a 14-d period.
In the trilinolein-only-treated
sample the concentration of
humic substances, as measured by change in fluorescence
intensities of C,, extract, decreased after irradiation through
799/v
78w
77W
76w
74%
TE4
36’N -
.Z
5
J
75ow
- 36ON
Atlantic
Ocean
35’N -
73W
- 35’N
Longitude
Fig. 5. Map of North Carolina sampling locations. W = Wilmington; E = Edenton; C = Columbia; WB = Wrightsville Beach.
the first day relative to fluorescence values taken immediately after photolysis (Fig. 6A). These early fluorescence
data most likely represent photobleaching of fluorescence
from ambient humic-type material in the seawater used as
the matrix for the photolysis study. This would be consistent
with photobleaching already known to occur during irradiation of a variety of oceanic samples by ambient sunlight
(Kouassi 1986; Kieber et al. 1990); however, it is unclear
why the process continues after irradiation ceases. After the
postphotolysis minimum, the concentration of humic substances gradually increased, reaching approximately starting
levels after 2 weeks. It is unclear from these data whether
new humic-type material was formed from the spiked trilinolein, or if there is a reversal of the photobleaching of the
ambient DOM. The key feature to these photolysis studies,
and the primary focus of this research, is the dramatic impact
of postphotolysis addition of ammonia tp the photolyzed,
trilinolein-treated
seawater. This is particularly important,
given the photolability of the triglyceride reported earlier and
the reactivity of the photofragments toward ammonia or organic amines. When ammonia was added after irradiation,
the magnitude of the increase in fluorescence and humic substances concentration was greatly enhanced (Fig. 6B). In this
latter experiment the final concentration of humic substances
dramatically exceeded initial levels and therefore cannot be
due simply to reversible photobleaching of ambient DOM
present in the water. Another significant feature of ammonia
addition is the lack of a postphotolysis fluorescence minimum. This implies that addition of ammonia enhances the
rate of the humification process. Similar conclusions can be
drawn when humic substance concentrations are monitored
by change in UV absorbance rather than fluorescence (Fig.
6). In the trilinolein-treated sample there was an initial decrease in absorbance followed by a gradual increase over
1460
Kieber et al.
::
;c-0-q
8
\/C-O-CH
I
4
II
II
I
5
IO
15
I
]A
Time (d)
hv,
02
P
-O-C&
8-0-CH
f
.Z
0.06
L
0.03
=E
{g
u)cv
a4
d
--4-i,
1 B
6
0
-0.03
-0.06
0
5
I
15
IO
it-0-CH
-4
8
I
I
HflC/C3\--b
6
Fig. 6. Concentration of humic substances as measured by fluorescence change (A QSU quinine sulfate units liter-l) and absorbance change over time of C,, cartridge extract in methanol after
photolysis of 0.33 p,M trilinolein-treated Gulf Stream seawater. (A)
Trilinolein only. (B) Tiilinolein with postphotolysis addition of ammonia (3 FM).
time. As with fluorescence, the magnitude of the increase
was enhanced by addition of ammonia, and no postphotolysis minimum was observed.
During the postphotolysis addition of ammonia experiment, aqueous ammonia concentrations and DON concentrations of C,,-extracted humic substances were also monitored over the 14-d period (Fig. 7). Aqueous ammonia
concentrations declined over time, and DON in the extracted
humic substances increased. The loss in dissolved ammonia
T
T
fc-o-c&
2O
Time (d)
0.8
NH3
I
0.3
go=6
so.4
v
go.2
02, OH, 03, Hz0
many steps
I
OH
0
7
Fig. 8. Proposed mechanisms for the incorporation of nitrogen
into marine hunic substances from photolyzed trilinolein. Structures 6 and 7 are representative of a family of related structures
containing covalently bound nitrogen that might be formed after
photolysis.
and concurrent increase in humic-bound nitrogen suggest
that ammonia is incorporated into the humic structure as it
is enhancing the rate of humic substances formation.
0
0
0
10
5
Discussion
15
Time (d)
Fig. 7. Aqueous ammonia (A) and dissolved organic nitrogen (B)
concentrations as a function of time. Error bars represent k 1 SD
based on y1 = 3.
The aldehydes generated during the photolysis of trilinolein may play an important role in humification of marine
triglycerides. Aldol-type condensations between the photogenerated aldehyde and carbonyls on one of the chains of
the partially oxidized triglyceride would result in increased
Photooxidation
of triglycerides
crosslinking in the humic substances. Additionally, the reactivity of aldehydes toward ammonia and organic amines may
explain the observation that addition of ammonia to photolyzed trilinolien in seawater resulted in increased humification and incorporation of nitrogen into the humic structure.
Aldehydes react readily with ammonia and amines to form
Schiff bases. Schiff bases have been proposed as intermediates in sedimentary and aqueous humic substance formation (Francois 1990; Thorn and Mikita 1992). One possible
mechanism for the synthesis of marine humic substances
from photolyzed triglycerides is proposed in Fig. 8 and is
very similar to the earlier mechanism proposed by Harvey
et al. (1983, 1985), except that it includes a significant addition, namely, the incorporation of nitrogen into the humic
structure. Photooxidation of trilinolein (4) first produces hexanal and the aldehyde (5). This initial photooxidation step
was verified experimentally during our photolysis studies.
Reaction of the aldehyde (5) with ammonia would produce
a Schiff base (6). Free radical-initiated crosslinking of the
fatty acid strands and subsequent oxidation, cyclization, dehydration, and ester hydrolysis, as proposed by Harvey et
al. (1983, 1985) would give a product incorporating covalently bound nitrogen, such as 7. This mechanism is also
consistent with the experimentally observed fluorescence
and absorbance increases in extracted humic substances after
addition of ammonia to photolyzed trilinolien in seawater.
Structures, such as 6 and 7 in Fig. 8, are proposed as
possible intermediates in the humification process and are
meant to represent a family of related structures that contain
covalently bound nitrogen. Although this mechanism suggests one possible pathway for the incorporation of ammonia
into humic substances, other organic amines, such as amino
acids, amino sugars, and nucleic bases (purines and pyrimidines), could also be incorporated into the humic structure
through similar processes. Our findings contribute a significant addition to our understanding of how nitrogen might
be incorporated into marine humic substances. Our continuing research will hopefully lead to a better understanding of
the nature of humic-bound nitrogen and may help unravel
the role of humic substances in the aquatic nitrogen cycle.
References
J. A., I? J. MILNE, C. A. MOORE, AND R. G. ZIKA. 1990.
Extraction of chromophoric humic substancesfrom seawater.
Mar. Chem. 29: 1-17.
CARLSSON, l?, AND E. GRANELI. 1993. Availability of humic bound
nitrogen for coastal phytoplankton. Estuarine Coastal Shelf Sci.
AMADOR,
36: 433-447.
R. 1990. Marine sedimentary humic substances:Structure, genesis, and properties. Aquat. Sci. 3: 41-80.
GRANELI, E., AND M. 0. MOREIRA. 1990. Effects of river water of
different origin on the growth of marine dinoflagellates and
diatoms in laboratory culture studies. J. Exp. Mar. Biol. Ecol.
FRANCOIS,
136: 89-106.
l? M. 1979. Volatile organic compounds in seawater.
Ph.D. Thesis, Woods Hole Oceanographic Institute.
0. C. ZAFIRIOU, AND R. B. GAGOSIAN. 1980. Volatile organic compounds in seawaterfrom the Peru upwelling region.
Limnol. Oceanogr.25(6): 1044-1053.
-,
R. E C. MANTOURA, R. P SCIIWARZENBACH, AND
R. ‘B. GAGOSIAN. 1982. Volatile organic compounds at a coast-
GSCHWEND,
-,
-
and fatty acids
1461
al site. 1. Seasonalvariations. Environ. Sci. Technol. 16: 3138.
T 1991. Production and turnover rates of fatty acids in
marinc particulate matter through phytoplankton photosynthcsis. Mar. Chem. 33: 213-227.
HARVEY, G. R., AND D. A. BORAN. 1982~. The synthesis of marine
humic substances.Trans. Am. Geophys. Union 63: 62.
-,
AND -.
1982b. Physical, chemical and NMR spectroscopic properties of dissolved humic substancesfrom the
Gulf of Mexico. Trans. Am. Geophys. Union 63: 62.
L. A. CHESAL, AND J. M. TOKAR. 1983. The
st&turc of’marine fulvic and humic acids. Mar. Chem. 12:
119-132.
S. R. PIOTROWICS, AND C. I? WEISEL. 1984. Synthesis of mirine humic substancesfrom unsaturatedlipids. Nature 309: 244-246.
;,
AND -.
1985. Geochemistry of humic substancesin
seawater,p. 233-247. Zn G. R. Aiken, D. M. McKnight, and
R. L. Wershaw [eds.], Humic substancesin soil, sediment, and
water: Geochemistry, isolation and characterization. John Wiley & Sons.
HYDRO, L. 1996. Photooxidation of triglycerides and fatty acids in
seawater: Implication toward the formation of marine humic
substances.Master’s thesis, Univ. North Carolina at Wilmington.
JALLIFFIER-MBRLON,
E., J. MARTY, V. DENANT, AND A. SALIOT.
1991. Phytoplanktonic sourcesof volatile aldehydesin the Red
River Rhone Estuary. Estuarine Coastal Shelf Sci. 32: 463HAMA,
482.
G., G. GERCKEN, AND K. EBERLEIN. 1983~. Development of lipids during a spring plankton bloom in the northern
North Sea, I. Particulate fatty acids. Mar. Chem. 14: 149-162.
AND K. HAMMER. 1983b. Development of lipids
during a spring plankton bloom in the northern North Sea, II.
Dissolved lipids and fatty acids. Mar. Chcm. 14: 163-173.
KAWAMURA, K., AND R. B. GAGOSIAN. 1987. Implications of ooxocarboxylic acids in the remote marine atmospherefor photo-oxidation of unsaturated fatty acids. Nature 325: 330-332.
KIEBER, R. J., AND K. MOPPER. 1990. Determination of picomolar
concentrations of carbonyl compounds in natural waters, including seawater, by liquid chromatography. Environ. Sci.
Technol. 24: 1477-1481.
X. ZHOU, AND K. MOPPER. 1990. Formation of carbonyl
compounds from UV-induced photodegradation of humic substances in natural waters: Fate of riverine carbon in the sea.
Limnol. Oceanogr.35: 1503-1515,
KOUASSI, A. M. 1986. Light induced alteration of the photophysical properties of dissolved organic matter in seawater.M.S.
Thesis, Univ. Miami.
Kuo, C. Y., T. C. HUNG, W. L. JENG, AND Y. S. HWAN~. 1993.
Extraction, isolation and purification of humic substances.Bull.
Inst. Chem. Acad. Sin. 40: 91-103.
MCKNI~HT, D. M., G. A. AIKEN, AND R. L. SMITH. 1991. Aquatic
fulvic acids microbially based ecosystems: Results from two
desert lakes in Antarctica. Limnol. Oceanogr.36: 998-1006.
-,
E. D. ANDREWS, S. A. SPAULDING, AND G. A. AIKEN.
1994. Aquatic fulvic acids in algal-rich Antarctic ponds. Limnol. Oceanogr.39: 1972-1979.
MORAN, M. A., AND R. E. HODSON. 1990. Bacterial production on
humic and nonhumic components of dissolved organic carbon.
Limnol. Oceanogr.35(S): 1744-1756.
-,
AND -.
1994. Support of bacterioplankton production by dissolved humic substancesfrom three marine environments. Mar. Ecol. Prog. Ser. 110: 241-247.
R. J. WICKS, AND R. E. HODSON. 1991. Export of dissolved organic matter from a mangrove swamn ecosvstem:EvKATTNER,
1462
Kieber et al.
idence from natural fluorescence, dissolved lignin phenols and
bacterial secondary production. Mar. Ecol. Prog. Ser. 76: 175184.
NISSENBAUM, A., AND I. R. KAPLAN. 1972. Chemical and isotopic
evidence for the in situ origin of marine humic substances.
Limnol. Oceanogr. 17: 570-582.
PARSONS, T., Y. MAITA, AND C. LALLI. 1989. A manual of chemical
and biological methods for seawateranalysis. PergamonPress.
SCHAUENSTEIN, E., H. ESTERBAUER, AND H. ZOI,LN~R. 1977. Aldehydes in biological systems. Their natural occurrence and
biological activities. Pion.
SCHNITZER, M. 1985. Nature of nitrogen in humic substances,
p. 303-325. Zn G. R. Aiken, D. M. McKnight, and R. L.
Wershaw reds.], Humic substances in soil, sediment, and
water: Geochemistry, isolation and characterization. John
Wiley 82 Sons.
SCRIBE, l?, J. FILI,AUX, J. LAURIELLARD, V. DENANT, ANL) A. SALIOT. 1991. Fatty acids as biomarkers of planktonic inputs in
the stratified estuary of the Krka River, Adriatic Sea:Relationship with pigment. Mar. Chem. 32: 299-312.
THORN, K. A. AND M. A. MIKITA.
1992. Ammonia fixation by
humic substances; a nitrogen- 15 and carbon-13 NMR study.
Sci. Total Environ. 113: 67-87.
VALDERRAMA, J..C. 1981. The simultaneous analysis of total nitrogen and total phosphorus in natural waters. Mar. Chem. 10:
109-122.
Received: 1 July 1996
Accepted: 20 December 1996