Dissolved carbohydrates in streamwater determined by HPLC and pulsed amperometric detection

Dissolved total saccharides (DTS) and dissolved free monosaccharides (DFMS) in streamwater were determined by high-performance liquid chromatography with pulsed amperometric detection (HPLC-PAD). HPLC identification was verified with gas chromatography/mass spectrometry measurements. The method for DTS was improved by using a column with an anion exchange capacity of 4,500 Fq and a mobile phase of 350 mM NaOH. The detection limits for individual monosaccharides ranged from 2 to 14 nM. The average recovery for monosaccharide standards was 82% after hydrolysis, and 75% of the monosaccharides in streamwater hydrolysates were recovered following a desalting procedure. Hydrolysis of model substances showed recoveries of monosaccharides between 78 and 98%. The C.V. for a hydrolyzed stream sample was 15% for the DTS. Stream samples stored at room temperature after filtration and acidification to pH 1.1 were stable for at least 23 d. Concentrations of DTS in White Clay Creek, including sugar alcohols and amino sugars, ranged from 0.64 to 12.70 PM and accounted for 2.9-12.1% of the dissolved organic carbon pool. Neutral sugars dominated the DTS pool, and glucose and galactose were the most abundant molecules. Concentrations of DFMS ranged from 0.05 to 0.38 PM and accounted for 0.06-0.33% of the dissolved organic carbon pool. Carbohydrates are important components of the dissolved organic matter pool in aquatic environments. They provide a significant source of C and energy for heterotrophic bacteria (Mtinster and Chrost 1990). In natural waters, dissolved carbohydrates have been estimated to make up from 1 to -30% of the dissolved organic C (DOC) (Thurman 1985). Dissolved total saccharides (DTS) consist of dissolved free monosaccharides (DFMS) and dissolved combined saccharides (i.e. oligoand polysaccharides). DTS concentration and composition in natural waters have only been analyzed following preconcentration, and there are no data on minor sugars such as sugar alcohols and few data on amino sugars. The determination of the molecular composition of carbohydrates not only provides information on the composition of the DOC pool but also provides an indication of the DOC sources and their transformations (Cowie and Hedges 1984; Hedges et al. 1994). Current methods for the determination of carbohydrates in natural waters include calorimetric assays (Johnson and Sieburth 1977; Burney and Sieburth 1977), gas chromatography (Cowie and Hedges 1984; Ochiai and Nakajima 1988), and liquid chromatography (Mopper 1977; Ittekkot et al. 1982; Wicks et al. 1991). Calorimetric assays do not provide information on the composition of the saccharides and are prone to interferences. Gas chromatography (GC), in combination with a flame ionization detector (FID), allows the identification of individual molecules, and with mass spectrometry (ms), the confirmation of that identity. However, GC involves time-consuming derivatization procedures and drying of samples that increases the possibility of contamination. The combination of high-performance liquid chromatography and pulsed amperometric detection (HPLC-PAD) has been reported as a promising alternative, as it combines high sensitivity and accuracy (Wicks et al. 1991; Mopper et al. 1992; Jorgensen and Jensen 1994). In our hands, environmental samples required no preconcentration step, thus reducing the chances of contamination. With the HPLC-PAD technique, carbohydrates dissociate in a strongly alkaline solution, separate as anions on an anion-exchange resin, and are detected by oxidation at the surface of a gold electrode. Jorgensen and Jensen (1994) discussed some significant drawbacks with the method, such as long periods for baseline equilibration, changes in retention time due to the interference of carbonate, and time-consuming cleaning procedures for the analytical columns. The objective of our work was to examine the analytical method for the determination of DTS in streamwater and seek improvements in the HPLC-PAD technique. We focused especially on sensitivity without preconcentration, high sample throughput, and the measurement of combined carbohydrates. Our studies included testing two different analytical columns, investigating the influence of the mobile phase, and verifying the results with GUMS measurements. Additionally we assessed sample recovery following storage, hydrolysis, and desalting.