aldoses in various size fractions of marine organic matter: implications for carbon cycling

Aldoses in Various Size Fractions of Marine Organic Matter: Implications for Carbon CyclingAuthor(s): Annelie Skoog and Ronald BennerSource: Limnology and Oceanography, Vol. 42, No. 8 (Dec., 1997), pp. 1803-1813Published by: American Society of Limnology and OceanographyStable URL: http://www.jstor.org/stable/2838926 .

Accessed: 16/06/2014 19:24

Your use of the JSTOR archive indicates your acceptance of the Terms & Conditions of Use, available at .http://www.jstor.org/page/info/about/policies/terms.jsp

.JSTOR is a not-for-profit service that helps scholars, researchers, and students discover, use, and build upon a wide range ofcontent in a trusted digital archive. We use information technology and tools to increase productivity and facilitate new formsof scholarship. For more information about JSTOR, please contact [email protected].

.

American Society of Limnology and Oceanography is collaborating with JSTOR to digitize, preserve andextend access to Limnology and Oceanography.

http://www.jstor.org

This content downloaded from 62.122.76.60 on Mon, 16 Jun 2014 19:24:39 PMAll use subject to JSTOR Terms and Conditions

http://www.jstor.org/action/showPublisher?publisherCode=limnoc

http://www.jstor.org/stable/2838926?origin=JSTOR-pdf

http://www.jstor.org/page/info/about/policies/terms.jsp


Limnol Oceanogr, 42(8), 1997, 1803-1813 C) 1997, by the American Society of Limnology and Oceanography, Inc

Aldoses in various size fractions of marine organic matter: Implications for carbon cycling

Annelie Skoog and Ronald Benner1 Marine Science Institute, University of Texas at Austin, 750 Channel View Dr., Port Aransas, Texas, 78373

Abstract Carbohydrates are major components of marine organic matter, but few molecular-level carbohydrate analyses in

seawater have been undertaken owing to the low concentrations of individual compounds. This paper presents novel data on aldose compositions and concentrations in various size fractions of particulate and dissolved organic matter from the equatorial Pacific. Samples of high-molecular-weight (HMW, >1 kDa) dissolved organic matter (DOM) and suspended particulate organic matter (POM, >0.1 ,um) were collected by tangential-flow ultrafiltration. Aldose content of low-molecular-weight (LMW) DOM was calculated as the difference between aldose content in unfiltered samples and HMW DOM + POM. Size-fractionated and unfiltered samples were hydrolyzed with sulfuric acid, and aldoses were separated and detected by high-performance anion-exchange chromatography and pulsed amperometric detection. A detailed description of the method is provided. The concentrations of fucose, rhamnose, arabinose, glucose, galactose, xylose, and mannose were determined. In general, the predominant sugars were glucose and galactose. The HMW DOM fraction was slightly enriched in deoxysugars and galactose. The average depth- integrated size distribution of aldoses indicated that 68% of aldoses were LMW DOM, 28% were HMW DOM, and 4% were POM. Concentrations and yields (normalized to organic carbon) of aldoses in POM and HMW DOM decreased with depth, indicating selective degradation of aldoses. The molecular compositions of the degraded components of POM and HMW DOM were similar to surface compositions but the mole % glucose increased with depth, implying that glucose was preferentially preserved in a refractory structure. The fraction with the highest aldose yield was POM, followed by HMW DOM and LMW DOM. In surface waters from both stations, the largest fractions of dissolved aldoses were found in HMW DOM, whereas in subsurface waters most dissolved aldoses were found in LMW DOM. Yields and concentrations of aldoses indicated a size-related diagenetic sequence where POM represents the most reactive material and LMW DOM the most refractory material. Aldoses represented 1.7- 3.6% of DOC and 10-20% of total carbohydrates, indicating that a large fraction of the carbohydrate pool is still uncharacterized at the molecular level. Aldoses could be identified in all, presumably refractory, deep-water samples, and most of the aldoses in deep water were LMW, indicating that a factor other than molecular composition was important in determining biological availability of the aldoses.

Carbohydrates are the largest identified fraction of organic matter in the ocean, accounting for 20-30% of organic matter in marine surface waters (Pakulski and Benner 1994; Benner et al. 1992). Carbohydrates comprise 10-70% of the organic matter in the plankton cell (Romankevich 1984) and are known to be released directly to the water column by algae (Hellebust 1965; Burney et al. 1981; Ittekot et al. 1981; Mopper et al. 1995). Additionally, other plankton as well as grazing activities release carbohydrates to the water column (Cowie and Hedges 1994, 1996; Strom et al. 1997). The major classes of carbohydrates have been identified in marine waters, including uronic acids (e.g. Mopper et al. 1995), amino sugars (e.g. Kerherve et al. 1995), and neutral

' Corresponding author.

Acknowledgments This research was supported by grants from the Department of

Energy (DE-FG03-94ER61907) and the National Science Founda- tion (OCE 94-13843). Annelie Skoog is grateful for a scholarship from the Swedish Institute. We thank N. Borch and D. Kirchman for helpful discussions about the aldose analysis and M. McCarthy for assistance with sample collection. The manuscript improved from discussions with the Biogeochemistry Group at the University of Texas Marine Science Institute and the MOG group at University of Washington. This is contribution 1005 from the University of Texas Marine Science Institute.

sugars (e.g. Mopper et al. 1995; Kerherve et al. 1995; Borch and Kirchman 1997).

There have been several investigations of dissolved carbohydrates in marine waters (e.g. Handa 1970; Liebezeit et al. 1980; Ittekot et al. 1981), but all of these studies relied on con entration procedures that fractionate the sample and leave an unknown fraction of carbohydrates uncharacterized. Determination of bulk carbohydrate content is possible using the MBTH method (Pakulski and Benner 1992, 1994), but bulk measurements do not provide information on molecular composition that can provide clues about the origin and diagenetic state of the material. Recent application to marine samples of an efficient chromatographic separation of aldoses and a sensitive detection method (Mopper et al. 1992, 1995) has made direct determinations in seawater possible without preconcentration.

The distribution and relatively high concentrations of combined carbohydrates in the upper ocean are indicative of a highly reactive substrate that supports heterotrophic me- tabolism in the surface ocean (Benner et al. 1992; Pakulski and Benner 1994). High-molecular-weight (HMW) carbohydrates seem to be rapidly consumed in surface waters whereas low-molecular-weight (LMW) carbohydrates persist in the deep ocean. These observations and recent direct com- parisons of the bioreactivity of HMW and LMW components of dissolved organic matter (DOM) indicate that naturally occurring HMW DOM is of more recent origin and

1803



1804 Skoog and Benner

Table 1. Results from testing the effect of pretreatment with 12 M H2SO4 on neutral sugar yields. The Gulf of Mexico sample was unfiltered water, whereas the equatorial Pacific sample was high-molecular-weight dissolved organic matter isolated by ultrafiltration from 100 m at 2?S. Samples were analyzed in triplicate. (Diff, difference achieved with pretreatment; Student's t-test was used to assess whether differences were statistically significant; P, probability that the samples come from the same population with the same mean; n.s., no statistical difference (i.e. P > 0.1).

Gulf of Mexico Equatorial Pacific

Pretreated Pretreated Concn concn Diff. Concn concn Diff.

Aldose (nM) (nM) (%) P (nM) (nM) (%) P

Fucose 78 73 -6 n.s. 615 643 4.6 n.s. Rhamnose 40 36 -10 n.s. 532 531 -0.2 n.s. Arabinose 71 70 -1.5 n.s. 341 348 2 n.s. Galactose 96 87 -9 n.s. 830 834 0.5 n.s. Glucose 100 187 87 <0.005 422 490 16 <0.01 Mannose 67 65 -3 n.s. 455 456 0.2 n.s. Xylose 72 77 7 n.s. 477 488 2.3 n.s. Ribose 33 32 -3 n.s. 101 105 4 n.s.

is more bioreactive than LMW DOM (Amon and Benner 1994, 1996). Detailed chemical characterizations of various size classes of organic matter in the ocean would provide additional information about the diagenetic state of organic matter and its relationship to size. Aldoses have been demonstrated to be particularly sensitive biomarkers of the diagenetic state of organic matter (Cowie and Hedges 1994). In the present study, the molecular-level analyses of aldoses are used to further investigate the relationship between the size and diagenetic state of organic matter in the ocean.

Concentrations of aldoses in unfiltered water samples, particulate organic matter (POM), and HMW DOM were determined using high-performance anion-exchange chromatography (HPAE) with pulsed amperometric detection (PAD) in depth profiles from two stations in the equatorial Pacific. The concentrations and composition of aldoses in LMW DOM were calculated by mass balance as the difference between unfiltered samples, which contain POM and DOM, and the POM + HMW DOM size fractions. The stations in the equatorial Pacific span regions of high and low productivity in the open ocean (Chavez and Barber 1987) and are therefore well suited for investigations of marine organic matter cycling (e.g. Hernes et al. 1996; Peltzer and Hayward 1996; Druffel et al. 1996). A detailed evaluation of the HPAE-PAD method and protocol for the analyses of aldoses in seawater are also provided.

Materials and methods

Sampling-Samples were collected from the equatorial Pacific Ocean at 2?S, 140?W and 12?S, 135?W. Details of the sample collection and ultrafiltration procedure are described elsewhere (Benner et al. 1997). Niskin-type samplers equipped with Teflon-coated springs were used. Unfiltered samples were frozen immediately after sampling in acid- rinsed polycarbonate bottles. Two size fractions were isolated by ultrafiltration, suspended POM (0.1-60 ,tm) and HMW DOM (1-100 nm).

Sample work-up procedure-The protocol included a combination of steps used by other laboratories (Mopper et

al. 1980, 1992; Pakulski and Benner 1992; Borch and Kirch- man 1997; Dionex, Sunnyvale, California, technical notes 20 and 21). All glassware (except ampoules) was acid-rinsed (2 M HCI), rinsed three times with Milli-UV + water (Milli- pore) and combusted at 500?C for 3 h. Ampoules were combusted without acid-rinsing. Plasticware (pipette tips and funnel valves) was soaked in acid for at least 12 h, rinsed three times with Milli-UV+ water, and dried.

Unfiltered seawater samples (9 ml) were pipetted into 15- ml sample tubes and dried in a Savant SpeedVac. One mil- liliter of 12 M H2SO4 was added to the samples, which were then placed in an ultrasonic bath for 15 min. Concentrated acid pretreatment has been shown to increase the carbohydrate yield (Pakulski and Benner 1992), and we tested to see whether yields of individual aldoses would increase signifi- cantly with pretreatment. Samples from the equatorial Pacific and the Gulf of Mexico were hydrolyzed with and without 12 M H2SO4 pretreatment. Pretreatment gave a statistically significant increase in glucose yield (Table 1). It is important to have 12 M H2SO4, to avoid overshooting the pH when neutralizing samples. Concentration of H2SO4 was determined by density. The ultrasonic bath aided in dissolving salt pellets. After the ultrasonic bath, samples were stirred to a slurry with a glass rod and left to sit for 1 h 45 min. Nine milliliters of Milli-UV+ water was added (1.2 M H2SO4, final concentration of acid) and samples were stirred until salts dissolved. Samples were then hydrolyzed in a 100?C water bath for 3 h in glass ampoules. The hydrolysis was terminated by placing the ampoules in an icebath for 5 min.

Samples were neutralized by adding the samples in 1-ml aliquots (to reduce effervescence) to 1.44 g precombusted CaCO3 in 20-ml glass sample tubes with Teflon-lined caps. Before adding sample to the CaCO3, deoxyribose was added as an internal standard to a final concentration of 200 nM. Samples were stirred and then placed in an ultrasonic bath for 15 min to complete the reaction. A sample pH of -6 was achieved this way. Ba(OH)2 has been used for neutralization by other authors (Mopper et al. 1992), but we found that CaCO3 gave higher recoveries and lower blanks. Or-



Diagenetic patterns of aldoses 1805

Table 2. Detector setting found to give the least baseline noise and the greatest sensitivity using a 24 mM NaOH mobile phase. Several different detector settings recommended by Dionex (Dionex technical note 21) were tried. Potentials are relative to Ag/AgCl. The integrate column shows where in the program integration starts and ends.

Time (s) Potential (V) Integrate

0.00 +0.05 0.20 +0.05 Begin 0.40 +0.05 End 0.41 +0.75 0.60 +0.75 0.61 -0.15 1.00 -0.15

ganic contaminants in Ba(OH)2 cannot be removed by com- bustion because the melting point of Ba(OH)2 is too low. Additionally, it was easier to achieve pH 6 using CaCO3, since carbonate buffers pH. Overshooting pH 6 resulted in losses of sugars, since they undergo extensive rearrangement at pH values >7 (Pigman 1957). After neutralization, the samples were centrifuged for 10 min in a table centrifuge and the supernatants were pipetted into scintillation vials and frozen until time of analysis.

Immediately before analysis, samples were run through a mixed bed of anion (AG 2-X8, 20-50 mesh, Bio-Rad) and cation (AG 50W-X8, 100-200 mesh, Bio-Rad) exchange resins. Using different bead sizes for cation and anion exchange resins makes separation and reuse of the resins possible. Before use, the resins were Soxhlet extracted with water for 2 h and transferred to H+ form (cation-exchange resin) and HCO3- form (anion-exchange resin). Equal resin volumes were mixed and 3 ml was added to a glass extrac- tion funnel with glass wool packed in the Teflon valve. The funnel was connected to vacuum via a bell jar. The resin bed was rinsed three times with Milli-UV+ water and then with -1 ml of sample, which was discarded. A sample volume barely covering the resin was then added for deionization. The rinse was necessary to minimize sorbtive losses of sugars. The sample-resin mixture was stirred until CO2 stopped evolving (-5 min), and the sample was allowed to drain by applying vacuum. The samples were degassed with He for 1 min before injection in the chromatography system. Dis- solved oxygen in the sample gave a negative peak that in- terfered with carbohydrate quantification. Degassing for too long resulted in a positive He peak in the chromatogram, also interfering with aldose quantification.

Seawater and HMW DOM samples were analyzed in triplicate, whereas POM samples were analyzed in duplicate. HMW samples were redissolved in Milli-UV+ water before analyses. Procedural blanks were run using Milli-UV+ water, and the only detectable sugar was glucose, which was found in concentrations of 0-10 nM. The glucose concentration found in the blank was deducted from the samples. Relative standard deviations of individual sugar concentrations in seawater samples were in the range of 5-30% (avg 18%), while relative standard deviations of individual sugar concentrations in HMW DOM and POM samples were in the range 2-10% (avg 7%). The lower standard deviations

Table 3. Effect of postcolumn base addition before and after cleaning the electrode surface. Sensitivity is the slope (area/nM) of the standard curve, calculated from a four-point standard curve. The cleaning procedure consisted of polishing the electrode surface [pc- base, postcolumn base addition (600 mM NaOH at 0.5 ml min- ); diff., difference in sensitivity when postcolumn base addition was omitted].

Sensitivity before cleaning Sensitivity after cleaning

With Without Diff. With Without Diff. Sugar pc-base pc-base (%) pc-base pc-base (%)

Fucose 0.3528 0.6593 87 0.3239 0.6879 112 Deoxyribose 0.1917 0.2529 32 0.1828 0.2607 43 Rhamnose 0.2441 0.4787 96 0.5234 0.5074 100 Arabinose 0.3615 0.6296 74 0.3442 0.6744 96 Galactose 0.5100 0.7968 56 0.4954 0.8452 71 Glucose 0.5736 0.8293 45 0.5960 0.9132 53 Mannose 0.4139 0.6627 60 0.4054 0.7018 73 Xylose 0.3941 0.8432 114 0.3704 0.8626 133 Fructose 0.3192 0.4508 41 0.2922 0.5611 92 Ribose 0.3464 0.5234 51 0.3072 0.5998 95

for HMW DOM and POM samples probably resulted from the higher concentrations of carbohydrates in the analyzed samples.

Separation and detection of aldoses-Aldoses were separated with an isocratic 24 mM NaOH elution using a PA- 1 column mounted in a Dionex 500 ion chromatography system with PAD (Rocklin and Pohl 1983; Dionex technical note 20; Johnson and LaCourse 1990) using a gold working electrode and an Ag/AgCl reference electrode. Several detector settings were tried (varying time and voltage), and the detector setting (Dionex technical note 21) in Table 2 was found to give the lowest noise and highest response using 24 mM NaOH mobile phase.

Milli-UV+ water was used for the mobile phase and was sparged with He for 15 min before adding liquid, low-carbonate NaOH (Fisher). Sparging was continued for -2 min and thereafter the mobile phase was kept in He-pressurized bottles. The same mobile phase could be used for 2 d if bottles were kept pressurized.

Several NaOH concentrations for the mobile phase have been suggested (Borch and Kirchman 1997; Kerherve et al. 1995). We tried 12-30 mM NaOH and found that 24 mM NaOH gave the best resolution of the aldoses analyzed. The 24 mM mobile phase still caused slight overlap between glucose, galactose, mannose, and xylose (Fig. 1). The resolution factors (R,) were - 1, i.e. >95% of the sugar was accounted for by integrating between peak valleys in the chromatogram (Snyder and Kirkland 1979). Considering that the standard deviation is 5-30% when using the work-up procedure described in this paper, we found a 5% deviation in the integration of the peaks acceptable.

Postcolumn base addition was not necessary using 24 mM NaOH at these detector settings. On the contrary, removal of the postcolumn base addition increased the overall sensitivity by omitting the dilution caused by adding NaOH to the sample (Table 3). Additionally, omission of the postcol-




umn base addition greatly decreased baseline noise, which further decreased the detection limit.

Recoveries-Aldose recoveries were in the range of 70- 90%, with the deionization procedure being the step causing the largest losses. However, some sugars could not be detected and separated using this method. Pretreatment and hydrolysis caused large (40-60%) losses in ribose. Ribose re- acts under acidic conditions to form a furfural, a reaction that proceeds so well that it has been used to approximate concentrations of pentoses (Pigman 1957). Fructose, a common ketose, is also susceptible to this reaction and forms a calcium levulate with low solubility (Pigman 1957). Fur- thermore, fructose coeluted with an unknown compound found in all sample chromatograms and is therefore not reported here. Deoxyribose was lost from the sample during hydrolysis. This makes it possible to use deoxyribose as an internal standard, since no deoxyribose originally contained in the sample will remain after hydrolysis.

Concentrations of fucose, rhamnose, arabinose, galactose, glucose, mannose, and xylose are reported. We also found low concentrations of ribose in HMW DOM and POM samples, but the standard deviations on the determinations were high and the data are not reported. We have chosen not to correct our original data beyond compensation for losses in the steps after hydrolysis, as reflected by the deoxyribose internal standard. Because the HMW DOM and POM samples were isolated from seawater as a dry powder, the water- column concentrations were calculated by multiplying the measured yield by the dry weight recovered and dividing by the volume of water filtered.

DOC and CHN analysis-DOC measurements were made using a Shimadzu TOC-5000 analyzer as described previ- ously (Benner and Strom 1993). Atomic C: N ratios were determined with a precision of -2% with a Carlo Erba 1108 CHN analyzer, after vapor phase HCI treatment (Hedges and Stern 1984). The DOC concentrations measured at both stations were typical for open-ocean samples.

Results and discussion

Concentrations and distribution of aldoses-Unfiltered samples and LMW DOM contain both free and combined aldoses, whereas HMW DOM and POM contain only combined aldoses. In this study, we made no attempt to measure free aldoses separately.

Aldose concentrations in unfiltered samples decreased about threefold with depth at both stations (Fig. 2), which agreed with the trend in total carbohydrate concentrations measured colorimetrically by Pakulski and Benner (1994) in the same water samples from 20S. The sharp decrease in aldose concentration with depth indicated that dissolved aldoses were reactive components of marine organic matter. The concentrations of individual sugars in unfiltered samples were in the range of 5-120 nM and the fraction of DOC accounted for as aldoses ranged from 1.7 to 3.6% (Tables 4, 5). The depth-integrated aldose concentrations indicated that 68% of aldoses was LMW, 28% was HMW, and 4% was POM. Aldose concentrations in surface water were higher at

Signal

A 0 10 15 Time(min)

Signal (mV)

a b c d e fgh

B 0 5 10 15 Time(min)

Fig. 1. A. Chromatogram of a hydrolyzed unfiltered surface sample from 2?S (see Table 5 for concentrations of aldoses). B. Chromatogram from a standard with 20 nM concentrations of individual sugars. The scales of the y-axes are not the same: a, fucose; b, deoxyribose; c, rhamnose; d, arabinose; e, galactose; f, glucose; g, mannose; h, xylose.

120S than at 20S (Fig. 2), which was consistent with DOC concentrations (Table 4). Nutrient-rich water upwells at the equator in the Pacific and is transported in surface water from the equator to higher latitudes (Murray et al. 1994). Biological processes occur in the transported water that cause an accumulation of DOC with increasing distance from the equator (Peltzer and Hayward 1996).

In general, the highest aldose concentrations in the different size fractions were found in surface samples from 2-m depth (Table 5, Fig. 3). The large concentration gradient between the surface and 100 m could be the result of rapid consumption of aldoses at 100 m, but it has also been suggested that there may be more release (both relatively and absolutely) of cell exudates at specific depths, e.g. in the upper, nutrient-limited region of the euphotic zone (Long- hurst and Harrison 1989). Note the similarity in aldose con-




Aldoses (nM)

0 200 400 600 800 0- _

500

1000

? 1500- = 2000 - 2

CL ~~~~~~~~~a-20S 0 2500

3000

3500

4000 Fig. 2. Aldose concentrations at the two stations vs. depth.

centrations among the respective size-fractions below the euphotic zone at both stations (Fig. 3), indicating that the aldose distribution among size-fractions in Pacific deep water could be fairly uniform.

High concentrations of aldoses in LMW DOM were found close to the chlorophyll maximum at 120S, indicating a possible source of LMW aldoses in the chlorophyll maximum. The large concentration gradients around 100 m indicate that aldoses in LMW DOM are rapidly consumed. However, at 20S the LMW DOM concentration maximum did not coin- cide with the chlorophyll maximum. Differences in the distribution and concentrations of aldoses in the chlorophyll maximum of these two stations could result from differences in the extent of water column mixing. Upwelling at 20S results in greater water column mixing than at 120S. At 120S, the chlorophyll maximum at 100 m is trapped just below the surface mixed layer. The stratification of the two stations could therefore explain the difference in aldose distribution and allow for a source of rapidly cycled LMW DOM in the chlorophyll maximum. Other studies have found that high concentrations of monosaccharides correlate with primary production in the equatorial Pacific (Rich et al. 1996).

Molecular composition-The aldose composition of POM, expressed as the mole percentages of the sum of aldoses (Table 5), was dominated by galactose and glucose. Glucose was the most abundant aldose, which has also been shown in other studies (Tanoue and Handa 1987; Hernes et al. 1996). The POM composition, particularly of material from the chlorophyll maximum, resembled that of material collected in net plankton tows from the equatorial Pacific (-45 mole % glucose, 10-15 mole % mannose and galac-

Table 4. Sampling locations and sample organic carbon con- tents. Samples for dissolved organic carbon (DOC) analysis were filtered through 0.1-,um pore-size filters (POC, particulate organic carbon; HMW, high molecular weight).

HMW POC HMW DOC

Depth DOC POC (wt% DOC (wt% Sampling locations (m) (,AM) (,AM) C) (,AM) C)

1?30'S, 140?00'W 2 72 1.02 4.70 21.7 19.7 2o00'S, 140?00'W 100 68 0.46 3.06 14.4 23.6 2o00'S, 140?00'W 400* 51 0.71 6.10 11.7 15.1 1?57'S, 140?03'W 4,000 44 0.35 2.36 8.02 18.8

12?12'S, 134?40'W 2 82 0.93 5.16 22.2 29.0 11?58'S, 135?07'W 100 85 0.82 6.58 22.2 26.8 12?06'S, 134?55'W 200 57 0.32 3.02 12.6 20.1 12?19'S, 134?26'W 375* 53 0.43 4.56 10.6 18.4 12o00'S, 135?00'W 4,000 45 0.32 2.72 8.10 14.6 * Depth of the oxygen minimum layer.

tose, 8 mole % fucose, and 5 mole % xylose, rhamnose, and arabinose; Hernes et al. 1996). The abundance of glucose and galactose in particulate material is probably a result of the major roles of these compounds in phytoplankton biol- ogy. Glucose polymers (glucans) are major storage compounds in phytoplankton, and galactose polymers (galactans) are common structural components of phytoplankton cell walls (Romankevich 1984). The relative abundance of aldoses other than glucose decreased with depth in POM, re- sulting in an increase in mole % glucose with depth. This has also been observed in the North Pacific (Tanoue and Handa 1987). Increasing mole % glucose with depth indicates selective degradation or removal of aldoses other than glucose. Fucose, rhamnose, arabinose, and galactose showed decreasing relative abundances with depth.

HMW DOM had a rather constant aldose composition with depth, which was also found by McCarthy et al. (1996). Relative to other size fractions at all depths, HMW DOM was enriched in galactose and deoxysugars (fucose and rhamnose). It has been shown that phytoplankton exudates are rich in galactose and deoxysugars, and that these sugars are surface active (Mopper et al. 1995).

The most abundant aldose in LMW DOM was glucose, and its relative abundance increased with depth (Table 5). At 20S, mannose was the second most abundant aldose, and a very low mole % galactose was found below surface waters. At 120S, the second most abundant aldose varied. There were relatively high abundances of deoxysugars, and at 100 m (chlorophyll maximum) galactose was the second most abundant aldose.

Implications of aldose yields for reactivity of organic matter-Aldose yields were calculated as a percentage of organic carbon by dividing aldose carbon by the total organic carbon in the size fraction. Aldose yields for all size fractions ranged from 2 to 14%. Aldose yields in surface HMW DOM in our study were 6-10% (Fig. 4) and are similar to yields (10-20%) reported by McCarthy et al. (1996) for HMW DOM in surface waters of the Sargasso Sea, North Pacific, and Gulf of Mexico. The same study also reported




Table 5. Mole percentages of aldoses in different size-fractions, and the sum of individual aldose concentrations. Concentrations in the low-molecular-weight fraction were calculated by difference (neg., negative value was calculated; POM, particulate organic matter; DOM, dissolved organic matter; HMW, high molecular weight; LMW, low molecular weight).

Depth Fucose Rhamnose Arabinose Galactose Glucose Mannose Xylose Sum (m) Sample (%) (%) (%) (%) (nM)

20S 2 Unfiltered 13 2 11 19 25 17 13 351

100 Unfiltered 6 5 4 11 42 19 13 212 375 Unfiltered 5 3 4 8 57 12 11 166

4,000 Unfiltered 4 4 4 7 60 13 8 128 2 POM 8 7 9 19 38 9 12 19.2

100 POM 5 6 3 14 52 10 10 3.5 375 POM 6 7 2 10 57 9 8 10.4

4,000 POM 3 4 2 6 69 7 10 2.2 2 HMW DOM 14 13 9 20 20 12 12 218

100 HMW DOM 16 12 8 16 28 10 10 109 375 HMW DOM 16 13 6 18 28 11 9 55

4,000 HMW DOM 18 16 4 19 20 12 11 20 2 LMW DOM 10 Neg. 12 14 26 24 13 114

100 LMW DOM Neg. Neg. 1 5 53 26 15 100 375 LMW DOM Neg. Neg. 4 3 70 12 12 101

4,000 LMW DOM 1 2 4 4 68 14 8 106 120S

2 Unfiltered 11 13 11 17 24 12 13 507 100 Unfiltered 12 12 10 18 21 14 12 804 200 Unfiltered 13 11 10 8 41 9 9 252 400 Unfiltered 11 9 11 10 37 11 10 160

4,000 Unfiltered 7 4 6 8 52 14 10 167 2 POM 8 6 10 18 34 8 16 17.8

100 POM 8 6 8 16 41 10 11 10.1 200 POM 8 9 3 12 44 13 10 2.4 400 POM 8 8 3 11 48 11 11 5.1

4,000 POM 5 5 1 8 64 9 7 2.3 2 HMW DOM 14 12 9 20 18 14 14 360

100 HMW DOM 19 15 8 19 13 13 13 133 200 HMW DOM 15 16 5 14 22 13 15 73 400 HMW DOM 18 17 7 19 16 12 11 66

4,000 HMW DOM 13 11 2 19 18 14 13 28 2 LMW DOM 2 19 16 7 39 7 10 129

100 LMW DOM 10 12 11 18 23 14 12 661 200 LMW DOM 12 9 12 6 48 7 7 176 400 LMW DOM 7 4 14 4 52 4 9 89

4,000 LMW DOM 5 2 7 6 57 14 9 137

aldose yields in the range of 2-3.5% for HMW DOM from deep waters (750-4,000 m), which can be compared to our values of 1.6 and 2.1% for HMW DOM from 4,000 m.

Decreasing concentrations of organic carbon with depth in the upper ocean are commonly interpreted as being indicative of reactive material. The relative reactivity of individual components of organic matter can be estimated by determining the yield, or fraction of carbon as specific molecules, and the changes in yield with depth. Yields of relatively reactive materials decrease with depth and increasing decomposition. It has been demonstrated that the carbohydrate yield of natural organic matter is a robust indicator of diagenetic state (Cowie and Hedges 1994), i.e. low yields of bioreactive components are indicative of more highly degraded organic material. In the present study, aldose yields of HMW DOM and POM decreased with depth in the water column (Fig. 4), indicating that aldoses are relatively reac-

tive components that are selectively removed from organic matter as it ages in the ocean. Organic materials in the deep ocean have the lowest aldose yields and represent the most diagenetically altered organic matter in the ocean.

Aldose yields in LMW DOM were low (1.2-1.6%) and relatively constant with depth at the 20S station (Fig. 4). Similar but more variable yields were observed in LMW DOM at 120S (Fig. 4). A relatively high aldose yield was observed in LMW DOM from 100 m at the 120S station and, as discussed earlier, the difference between the stations could result from differences in the stability of euphotic zone waters. The low and relatively invariant yields with depth in LMW DOM are indicative of a fairly unreactive material.

Yields from the different size fractions in our study showed that material with a higher molecular weight had a higher content of aldoses. With only one exception (100 m at 20S), the POM fraction always had the highest aldose




20S 1 20S Aldoses (nM) Aldoses (nM)

0 100 200 300 400 0 100 200 300 400 0 0-

500 - 500 6

1000 1000

e olde POM1 500 org

mattr; H W DO , hih-mWlclrwih isle rai DOMer ILMW DOMW Dowmolclr

i2000 2000 id-x_ LMW DOM CL ~~~-x- LMW DOM C

(D 2500

(D 2500-

3000 -3000

3500 -3500

4000 ~~~~~~~~~4000 Fig. 3. Aldose concentrations in different size-fractions vs. depth. Value in parenthesis is the

concentration of low-molecular-weight aldoses at 100-rn depth, 12'S. POM, particulate organic matter; HMW DOM, high-molecular-weight dissolved organic matter; LMW DOM, low-molecular- weight dissolved organic matter.

yield at each depth, followed by HMW DOM and LMW DOM (Fig. 4). Aldose yields indicate the following sequence of diagenetic alteration, from least to most: POM -e HMW DOM -> LMW DOM. These data also indicate the following order of reactivity from most reactive to least reactive: POM > HMW DOM > LMW DOM.

It is important, however, to note that LMW DOM is not completely unreactive in surface waters. The composition of aldoses in LMW DOM changed substantially between surface water and the oxygen minimum layer (Table 5). The mole % glucose increased from an average of 32% in surface

water to an average of 61% at the oxygen minimum layer. This change in aldose composition with depth indicated an exchange of >50% of the aldoses in LMW DOM in the upper ocean.

The origin of suspended POM-The diagenetic sequence, suggested by aldose compositions of materials in the size range of suspended POM to LMW DOM, raises the question of a possible diagenetic connection with size classes larger than suspended POM. Aldose yields from suspended POM (Fig. 4) at 100 m were higher than yields from net plankton

20S 120S Aldose yields Aldose yields

(% of organic carbon) (% of organic carbon) 0.00 5.00 10.00 15.00 0.00 5.00 10.00 15.00

_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ - I

2 2

h HMW POM }

2 100

100 l _ I I~~ HMW DOM 0 POM i

400 E 0 LLMW DOM E M HMW DOM

a f A s ~~~~~~~~~~200 L?LMW DOM

*-400

H ~~~~~~~~ ~ ~~375 # C

4000 4000

Fig. 4. Aldose yields in different size-fractions as a percentage of total carbon in the fraction. Abbreviations as in Fig. 3.




and sinking POM (Hernes et al. 1996). At 20S, the yield from sinking POM was 2%, whereas the yield from suspended POM was 12%. At 120S, the yields were 6 and 14% for sinking POM and suspended POM, respectively. The aldose yield from suspended POM (Fig. 4) was at least twice the yield from net plankton (Hernes et al. 1996). Hence, compositional data from surface waters indicated no direct diagenetic coupling between larger size classes of suspended POM (net plankton and sinking POM) and suspended POM. Plankton tows collected material in the 26-850-,um size range, while suspended POM was in the 0.1-60-,um size range. Plankton species vary with size, and we might therefore be comparing the yield from different species of plankton when comparing yields from organic matter in plankton tows with suspended POM. This might also be true for the comparison between organic matter from sediment traps (105 m) and suspended POM from 100 m. Sinking POM caught in sediment traps is largely comprised of fecal pellets from zooplankton that have been feeding on larger size classes of phytoplankton (J. Hedges pers. comm.).

When looking at deep-water data, we find that aldose yields from sinking POM are close to yields for suspended POM. Yields from suspended POM at 4,000 m were 4.0% (20S) and 5.8% (120S), whereas values from Hernes et al. (1996) from sediment traps at 4,000 m ranged between 4.4 and 4.7%. The close agreement between yields from sinking POM and suspended POM in deep waters suggests that there is no clear diagenetic sequence from sinking POM to suspended POM in deep waters. The similarity in aldose yields suggests the possibility of exchange between sinking and suspended POM at depth.

Biological reactivity of aldoses-Based on aldose composition of POM from net plankton tows, sediment traps, and sediments, it has been suggested that glucose is preferentially degraded from organic matter and that the mole % glucose could be a diagenetic indicator for organic material in the equatorial Pacific region (Hernes et al. 1996). How- ever, our data show no such relationship for dissolved and suspended material. When we calculated the composition of the fraction of carbohydrates removed between surface and deep water, we found that the composition of the removed fraction was almost identical to the composition of surface HMW DOM and suspended POM (Fig. 5). The same observation was also made by McCarthy et al. (1996) for HMW DOM. This implies that there is no biological pref- erence for specific sugars in dissolved and suspended material. Carbohydrates are removed in similar molar ratios as they occur in reactive organic material. The mole % glucose might therefore only be valid as a diagenetic indicator for sinking particulate material.

Furthermore, these findings suggest that the relative im- portance of glucose as a substrate for microorganisms is mainly due to its greater abundance. This finding has implications for interpreting aldose fluxes in the marine environ- ment, because it suggests that using only glucose for estimating carbohydrate turnover could lead to an underestimate of the fraction of bacterial production supported by carbohydrates. Galactose, mannose, and xylose are also abundant

in seawater, and the biological utilization of these aldoses should be investigated.

A large fraction of carbohydrates are not aldoses-What are they ?-Measurements of total carbohydrates by the MBTH method (Pakulski and Benner 1992) showed carbohydrate-C concentrations in the range of 5-19 ,uM at 20S (Fig. 6). The concentration of aldose-C found at 20S was in the range of 0.8-2.2 ,uM, i.e. only -7-20% of the carbohydrate pool was identified as aldoses. Similar hydrolysis procedures were used in the two studies, so the higher concentrations found by Pakulski and Benner (1992) should be indicative of the presence of carbohydrates other than aldoses. Nuclear magnetic resonance (NMR) analysis has also been used to estimate carbohydrate concentrations (Benner et al. 1992; McCarthy et al. 1993; Pakulski and Benner 1992) and indicates carbohydrate concentrations similar to or higher than MBTH determinations.

The large amount of molecularly uncharacterized carbohydrate could be explained in a number of ways. Both NMR and MBTH carbohydrate concentration estimates are based on estimating the amount of functional groups characteristic of carbohydrates. By NMR it is possible to estimate the amount of C-O and O-C-O by integrating the peaks repre- senting resonances by carbon present in these groups. NMR studies have been undertaken on unhydrolyzed samples of UDOM. The MBTH method estimates the amount of terminal alcohol groups and terminal groups that can be con- verted to alcohol groups by reduction with KBH4. In unhydrolyzed samples, the analysis will include both monomeric sugars and free ends of sugars that are attached to other molecules. The sample can be hydrolyzed before MBTH determination to estimate the sum of combined and monomeric sugars. Aldoses that have been modified by biological and chemical reactions might still contain the functional groups that are recognized by NMR and MBTH analysis (i.e. carbohydrate-like compounds) but might not be recognized as aldoses by the HPAE-PAD separation/detection procedure.

Carbohydrates occurring in seawater are a diverse group of compounds and the terminology can be confusing. Car- bohydrates are defined as polyhydroxy aldehydes or ketones and their derivatives. Strictly speaking, the term "aldoses" includes all carbohydrates with an aldehyde end, i.e. amino sugars and neutral aldoses. In this paper, aldoses refer to neutral aldoses. A number of naturally occurring classes of carbohydrates cannot be quantified with the method used in this study (e.g. sugar alcohols, amino sugars, and uronic acids). Sugar alcohols were preserved during the hydrolysis and ion-exchange procedures used in the present study, and peaks eluting at the general times of sugar alcohols were found in the chromatograms of the samples analyzed (Fig. 1, RT < 5 min). These peaks were not resolved with the elution conditions used (retention times overlap) and could therefore not be identified. However, the response of the PAD detector is fairly uniform for varying sugar molecules, and an estimate from peak area indicated that sugar alcohol concentrations might be as high as aldose concentrations in samples analyzed. Amino sugars and uronic acids are two classes of charged carbohydrates that would be lost in the deionization procedure. Amino sugars are the basic building




2QS _ 20S POM3 2m

40 - POM, rem.

35 i EHMW DOM, 2m E I S:30 1 : HMW DOM, rem.

.2 25

20 0

o 10

0- -1) O O D (L1) W O

0 0 0 0 0 0

0Ec 0 -

VL E .

_10 POMS r-- 2m~-~~~~---S

40 S POM, 2rem E 35 - B HMW DOM, 2m C 30 - _EL oHMW DOM, rem< 0 . , 25

20 0 15 E

O 10 X ' X t

o 0 JE ?1) ~ ~ I)

tl() ()C) _ c 0 0 0 0 0 0

> O f U O~~ ~~~~ C -,

it E 0 _

Fig. 5. Aldose composition of surface (2 m) organic material relative to the component that is removed between surface and deep water (rem.). The composition of the removed material was calculated by subtracting deep-water aldose concentrations from surface aldose concentrations.

blocks of chitin and are common components of marine algae, bacteria, and zooplankton. Montgomery et al. (1990) reported suspended (>0. 1-,m size-fraction) chitin concentrations of 4-21 ,ug liter-' in the Subarctic Pacific and Del- aware Bay and suggested that these structural polysaccha- rides were degraded rapidly in surface waters. The concentrations of uronic acids have been suggested to com-

prise a minor component of marine carbohydrates in some studies (Mopper et al. 1995), whereas other studies have reported that uronic acids comprise 11-43% of polysaccha- rides isolated from the North Pacific and Bering Sea (Saku- gawa and Handa 1985). Therefore, amino sugars and uronic acids could be a substantial fraction of the marine carbohydrate pool.




Carbohydrate (,uM C) 0 10 20 30

500 X

1000

E 1500

= 2000 --Aldose,2?S

2500 --o-Aldose, 120S

3000 -x-TCHO, 20S

3500

4000 Fig. 6. Carbon concentration in aldose and total carbohydrate

fractions in the Pacific Ocean. TCHO, total carbohydrate as measured by the MBTH method (data from Pakulski and Benner 1994).

Aldoses in deep water-Refractory aldoses?-Carbohy- drates, especially free glucose, have been shown to be biologically reactive molecules. Our data on aldose yield and concentration changes with depth support the observation that aldoses are biologically reactive molecules. However, all size-fractions of organic matter had measurable quantities of aldoses in deep water, indicating that aldoses comprise part of the refractory organic matter in the ocean. Addition- ally, the relative abundance of glucose, the aldose that has been thought to be the most reactive, increased with depth. Glucose was the major aldose in deep-water POM and DOM, accounting for 60% of the total aldoses. About 80% of aldoses in deep water are LMW DOM, indicating they have a molecular weight lower than 1,000 Da. We do not know what fraction of the aldoses in LMW DOM are in combined form, but analyses from other areas indicate that free aldoses are in low concentrations in deep water (<10 nM; Skoog et al. in prep.), suggesting that most deep-water LMW aldoses are in combined form.

The existence of LMW aldoses in deep water is an ap- parent paradox. Why are they resistant to decomposition? Our data indicate that LMW DOM is the product of the diagenesis of macromolecular material. Perhaps the residual LMW-combined aldoses in deep water are formed through selective preservation of resistant components of biomole- cules. Clearly, we need to know more about the molecular architecture of refractory organic matter to better understand the mechanisms controlling the preservation of organic matter in the ocean.

Conclusions

The concentration and yield distributions of aldoses in all size fractions indicated degradation between surface and

deep water. In general, the most abundant sugars were glucose and galactose in all size fractions. HMW DOM was slightly enriched in deoxysugars and galactose when compared with the other size-fractions. When integrated over depth, most (68%) of the aldoses were in the LMW fraction, 28% was in the HMW fraction, and the smallest fraction (4%) of the aldoses was in POM. Our data also indicate that a large fraction of aldose is LMW material, i.e. a large part of aldoses is combined but has low molecular weight. When comparing aldose concentrations to total carbohydrate concentrations estimated by the MBTH method, only 7-20% of the carbohydrates were identified as aldoses. A large part of the carbohydrate pool is therefore still unidentified at the molecular level.

Degradation of aldoses between surface and deep water seemed to be nonspecific. The molecular composition of the component degraded between surface and deep water resembled that of surface organic material. No indications were found of selective degradation of glucose. In fact, the mole % glucose increased with depth in the LMW DOM and suspended POM fraction.

Yields of aldoses (aldose normalized to carbon) from the dissolved and suspended fractions indicated that the weight- fraction richest in aldose was POM, followed by HMW DOM and LMW DOM. Aldose yield data, together with concentration and distribution data, indicated a diagenetic sequence from least to most degraded: POM -4 HMW DOM -4 LMW DOM. Remember that turnover times of the various molecular groups within a size fraction probably vary.

Aldose yields from suspended POM, and POM from sediment traps and plankton tows (Hernes et al. 1996), did not indicate a diagenetic sequence between these two classes of particulate material. However, care has to be taken when comparing large POM from sediment traps and plankton tows with suspended POM collected by ultrafiltration since it is possible that these different size classes include different components of plankton.

Samples from the deep sea revealed the presence of LMW (<1,000 Da) glucose-containing compounds. Traditionally, compounds with those characteristics are considered biologically reactive. The presence of compounds with these characteristics indicates that even in LMW compounds, glucose can be complexed in a form that is unavailable to microorganisms.

References

AMON, R. M. W., AND R. BENNER. 1994. Rapid cycling of high- molecular-weight organic matter in the ocean. Nature 369: 549-552.

, AND . 1996. Bacterial utilization of different size classes of dissolved organic matter. Limnol. Oceanogr. 41: 41- 51.

BENNER, R., B. BIDDANDA, B. BLACK, AND M. MCCARTHY. 1997. Abundance, distribution, and stable carbon and isotope compositions of marine particulate and dissolved organic matter isolated by tangential-flow ultrafiltration. Mar. Chem. 57: 243- 263.

, J. D. PAKULSKI, M. MCCARTHY, J. I. HEDGES, AND P. G. HATCHER. 1992. Bulk chemical characteristics of dissolved organic matter in the ocean. Science 255: 1561-1564.




, AND M. STROM. 1993. A critical evaluation of the analyt- ical blank associated with DOC measurements by high-tem- perature catalytic oxidation. Mar. Chem. 41: 153-160.

BORCH, N. H., AND D. L. KIRCHMAN. 1997. Concentration and composition of dissolved combined neutral sugars (polysac- charides) in seawater determined by HPLC-PAD. Mar. Chem. 57: 85-95.

BURNEY, C. M., K. M. JOHNSON, AND J. M. SIEBURTH. 1981. Diel flux of dissolved carbohydrate in a salt marsh and a simulated estuarine ecosystem. Mar. Biol. 63: 175-187.

CHAVEZ, F P., AND R. T BARBER. 1987. An estimate of new production in the equatorial Pacific. Deep-Sea Res. 34: 1229- 1243.

COWIE, G. L., AND J. I. HEDGES. 1994. Biochemical indicators of diagenetic alteration in natural organic matter mixtures. Nature 369: 304-307.

AND . 1996. Assimilation and alteration of the major biochemicals on ingestion of diatoms by an herbivorous zooplankton. Limnol. Oceanogr. 41: 581-594.

DRUFFEL, E. R. M., J. E. BAUER, P. M. WILLIAMS, S. GRIFFIN, AND D. WOLGAST. 1996. Seasonal variability of particulate organic radiocarbon in the northeast Pacific. J. Geophys. Res. 101(C9): 20,543-20,552.

HANDA, N. 1970. Dissolved and particulate carbohydrates, p. 129. In D. W. Hood [ed.], Organic matter in natural waters. Inst. Mar. Sci. Univ. Alaska.

HEDGES, J. I., AND J. H. STERN. 1984. Carbon and nitrogen determinations of carbonate containing solids. Limnol. Oceanogr. 29: 657-663.

HELLEBUST, J. A. 1965. Excretion of some organic compounds by marine phytoplankton. Limnol. Oceanogr. 10: 192-206.

HERNES, P. J., J. I. HEDGES, M. L. PETERSON, S. G. WAKEHAM, AND C. LEE. 1996. Neutral carbohydrate geochemistry of particulate material in the central equatorial Pacific. Deep-Sea Res. 43: 1181-1204.

ITTEKOT, V., U. BROCKMAN, W. MICHAELIS, AND E. T DEGENS. 1981. Dissolved free and combined carbohydrates during a phytoplankton bloom in the northern North Sea. Mar. Ecol. Prog. Ser. 4: 299-305.

JOHNSON, D. C., AND W. R. LACOURSE. 1990. Liquid chromatography with pulsed electrochemical detection. Anal. Chem. 62: 589A-597A.

KERHERVE, P., B. CHARRIERE, AND F GADEL. 1995. Determination of marine monosaccharides by high-pH anion-exchange chromatography with pulsed amperometric detection. J. Chroma- togr. 718: 283-289.

LIEBEZEIT, G., M. BOLTER, I. F BROWN, AND R. DAWSON. 1980. Dissolved free amino acids and carbohydrates at pycnocline boundaries in the Sargasso Sea and related microbial activity. Oceanol. Acta 3: 357-362.

LONGHURST, A. R., AND W. G. HARRISON. 1989. The biological pump: Profiles of plankton production and consumption in the upper ocean. Prog. Oceanogr. 22: 47-123.

MCCARTHY, M. D., J. I. HEDGES, AND R. BENNER. 1993. The

chemical composition of dissolved organic matter in seawater. Chem. Geol. 107: 503-507.

, , AND . 1996. Major biochemical composition of dissolved high molecular weight organic matter in seawater. Mar. Chem. 55: 281-297.

MONTGOMERY, M. T, N. A. WELSCHMEYER, AND D. L. KIRCHMAN. 1990. A simple assay for chitin: Application to sediment trap samples from the subarctic Pacific. Mar. Ecol. Prog. Ser. 64: 301-308.

MOPPER, K., R. DAWSON, G. LIEBEZEIT, AND V. ITTEKOT. 1980. The monosaccharide spectra of natural waters. Mar. Chem. 10: 55-66.

, C. A. SCHULTZ, L. CHEVOLOT, C. GERMAIN, R. REVUELTA, AND R. DAWSON. 1992. Determination of sugars in unconcen- trated seawater and other natural waters by liquid chromatography and pulsed amperometric detection. Environ. Sci. Tech- nol. 26: 133-138.

, J. ZHOU, K. S. RAMANA, U. PASSOW, H. G. DAM, AND D. T. DRAPEAU. 1995. The role of surface-active carbohydrates in the flocculation of a diatom bloom in a mesocosm. Deep- Sea Res. 42: 47-73.

MURRAY, J. W., R. T BARBER, M. R. ROMAN, M. P. BACON, AND R. A. FEELY. 1994. Physical and biological controls on carbon cycling in the equatorial Pacific. Science 266: 58-65.

PAKULSKI, J. D., AND R. BENNER. 1992. An improved method for the hydrolysis and MBTH analysis of dissolved and particulate carbohydrates in seawater. Mar. Chem. 40: 143-160.

, AND . 1994. Abundance and distribution of carbohydrates in the ocean. Limnol. Oceanogr. 39: 930-940.

PELTZER, E. T, AND N. A. HAYWARD. 1996. Spatial and temporal variability of total organic carbon along 140?W in the equatorial Pacific Ocean in 1992. Deep-Sea Res. 43: 1155-1180.

PIGMAN W. [ED.] 1957. The carbohydrates: Chemistry, biochem- istry, physiology, Academic.

RICH, J., H. W. DUCKLOW, AND D. L. KIRCHMAN. 1996. Concen- trations and uptake of neutral monosaccharides along 140 W in the equatorial Pacific: Contribution of glucose to heterotrophic bacterial activity and the DOM flux. Limnol. Oceanogr. 41: 595-604.

ROCKLIN, R. D., AND C. A. POHL. 1983. Determination of carbohydrates by anion exchange chromatography with pulsed amperometric detection. J. Liquid Chromatogr. 6: 1577-1590.

ROMANKEVICH, E. A. 1984. Geochemistry of organic matter in the ocean. Springer-Verlag.

SAKUGAWA, H., AND N. HANDA. 1985. Chemical studies on dissolved carbohydrates in water samples collected from the North Pacific and Bering Sea. Oceanol. Acta 8: 185-196.

SNYDER, L. R., AND J. J. KIRKLAND. 1979. Introduction to modem liquid chromatography. John Wiley and Sons.

STROM, S., R. BENNER, AND M. DAGG. Sources of marine dissolved organic carbon: Laboratory investigations of protozoa, cope- pods, and phytoplankton. Limnol. Oceanogr.

TANOUE, E., AND N. HANDA. 1987. Monosaccharide composition of marine particles and sediments from the Bering Sea and northern North Pacific. Oceanol. Acta 10: 91-100.

Received: 29 August 1996 Accepted: 20 May 1997



aldoses in various size fractions of marine organic matter: implications for carbon cycling

Documents