Microbial fluxes of free monosaccharides and total carbohydrates in freshwater determined by PAD-HPLC

FEMS Microbiology Ecology 14 (1994) 79-94 © 1994 Federation of European Microbiological Societies 0168-6496/94/$07.00 Published by Elsevier

79

FEMSEC 00525

Microbial fluxes of free and total carbohydrates by PAD-HPLC

monosaccharides in freAlwater determined

Nie l s O . G . J 0 r g e n s e n * a n d R e g i t z e E. J e n s e n

Microbiology Section, Department of Ecology and Molecular Biology, Royal Veterinary and Agricultural University, Rolighedsvej 21, DK-1958 Frederiksberg C, Denmark

(Received 27 September 1993; revision received 7 January 1994; accepted 21 January 1994)

Abstract: A new sensitive pulsed amperometric detection (PAD) method for measurements of mono- and disaccharides in nM concentrations was used in combination with high performance liquid chromatography (HPLC) to study fluxes of dissolved free and combined carbohydrates (DFCHO and DCCHO) in lake water. In a diel study concentrations of individual free saccharides typically were 5-50 nM, while total DFCHO concentrations ranged from 67 to 224 nM. No diel trends in concentration changes were obvious. At in situ light-dark conditions, dominant DFCHO were galactose, glucose, fructose and mannose/xylose. In addition to these saccharides, an increased abundance of melibiose and arabinose was measured in a parallel dark incubation. In a 118 h laboratory incubation of 1.0 /zm filtered lake water, concentrations of DFCHO decreased from 194 nM (at 12 h) to a minimum of 54 nM (at 73 h). Dominant DFCHO were glucose, fructose and cellobiose. During the incubation DCCHO varied from 1.27 to 2.20/xM. Glucose, galactose and cellobiose made up 40, 30 and 10 mol-%, respectively, of the DCCHO. Fructose was degraded during hydrolysis of the DCCHO. A decline of DCCHO at 55 h was reflected in a simultaneous increase of DFCHO, but otherwise no similarities between the two saccharide pools were found. Increased DCCHO concentrations and a high assimilation of glucose and fructose that was not reflected in a decline of their concentrations, both indicate that carbohydrates were produced during the experiment. Polysaccharides were probably excreted by the bacteria. Net assimilation of glucose and fructose sustained 14-19% (diel study) and 32% (long-term study) of the net bacterial carbon requirement.

Key words." PAD-HPLC; Monosaccharide; Carbohydrate; Fresh water; Carbon flux

Introduction

In freshwater env i ronments dissolved carbohydrates have b e e n es t imated to make up from less than 1 to about 30% of the pool of dissolved organic carbon (DOC) [1,2]. The carbohydrates consist of combined, hydrolyzable polymers and

* Corresponding author. Tel: 38282625; Fax: 38282624; E-mail [email protected]

SSDI 0168-6496(94)00008-K

free componen ts ( D C C H O and D F C H O , respectively). The concen t ra t ion of D C C H O in different freshwaters is highly variable. Thus, in eutrophic lakes D C C H O concent ra t ions of 0 . 1 4 - 5 0 / z M (in uni ts of glucose) have been repor ted [3-6]. Total pools of D F C H O appear less variable, with concentra t ions of 2 . 5 - 1 0 / ~ M (in uni ts of glucose) in different eutrophic freshwaters [2,6].

Measured concent ra t ions of D F C H O may be inf luenced by the appl ied analytical procedure . Genera l ly lower concent ra t ions of D F C H O are

80

determined in analysis of individual sugars than with assays for total concentration measurements, such as the anthrone procedure [7]. Determina- tions of glucose illustrate this. Using chromatographic or enzymatic procedures, natural concentrations of glucose of 25-150 nM typically have been measured in fresh waters [8-10]. Relative to the reported total DFCHO pools mentioned above, glucose should make up only about 2 mol-% of the DFCHO. This appears unlikely, as glucose in most studies of individual saccharides in natural waters has been found a dominant D FCHO [10-13]. Possibly colorimetric assays like the anthrone method include not only free saccharides, but also saccharides associated or combined with other substances. A similar difference in concentrations of DCCHO measured with the anthrone method and a gas chromatographic procedure has been found [14]. The authors sug- gested that manipulations, e.g. hydrolysis and derivatization, prior to chromatographic analysis may have influenced the measurements.

The methodological uncertainties encountered in analysis of saccharides seem reduced with pulsed amperometric detection (PAD) that allows a direct analysis of both reducing and non-reducing sugars, for example, see a review by Johnson and LaCourse [15]. With PAD, sugars are separated as anions in an alkaline solution by high pressure liquid chromatography (HPLC). In an amperometric on-line detector the sugar anions are oxidized at the surface of a gold electrode. The electrode is exposed to a pulsating positive and negative potential to remove oxidation products at the electrode surface. The PAD technique is very sensitive and makes detection of most sugars at a pmol level possible [16,17]. This im- plies that no derivatization procedure is required for analysis of DF C HO at natural concentrations.

The PAD-HPLC procedure was used in the present study to measure diel concentration changes of natural pools of free monosaccharides in the Danish Lake Fures0. In addition, bacterial assimilation of the two dominant sugars, glucose and fructose, and of free amino acids was measured. The assimilation rates were related to the total bacterial carbon flux. In a parallel laboratory experiment the bacterial utilization of both

DFCHO and DCCHO was followed to study total carbohydrate pools as nutrients to bacterial populations.

Materials and Methods

Sampling Water samples for studies of microbial fluxes

of dissolved free and combined saccharides and amino acids were collected in the mesotrophic Lake Fures0 (9.5 km 2, maximum depth of 38 m) in Northern Zealand, Denmark, on November 7 1990. Subsurface water samples were incubated in two 10 1 clear and two 10 1 dark glass bottles suspended at the water surface of the lake. Sub- samples from the bottles were taken at 3 h intervals from 06.00 to 24.00 h. Incubation in bottles in the lake was chosen to ensure that identical water volumes were sampled throughout the studied period. On November 9 1990 lake water was collected and brought to the laboratory, filtered through 1.0 /zm polycarbonate filters and incubated in 2 1 glass bottles in the dark at in situ temperature (6°C) in a shaking bath. Subsamples were taken at 12 to 24 h intervals from the bottles over a period of 5 days.

In both experiments water samples for analysis of dissolved free carbohydrates and amino acids were filtered through 0.2 ~ m membrane filter cartridges (Sartorius, FRG). The initial 2 ml fil- trate was discarded, after which water volumes of 2 ml (for DFAA, dissolved free amino acid, analysis) and 15 ml (for saccharide analysis) were filtered into pre-ashed (500°C for 5 h) glass vials and frozen until analysis.

Microbiological analyses Bacterial production was measured from in-

corporation of [3H]thymidine according to Fuhr- man and Azam [18] and incorporation of L- [3H]valine according to Jorgensen [19]. In both cases 5 ml triplicate water samples and a control sample with 2% final concentration of formalde- hyde were used. For thymidine a concentration of 10 nM [3H]thymidine (20 Ci mmol - j , DuPont NEN Research Products, USA) was used, but for valine the addition was 100 nM (10% [3H]valine,

45-65 Ci mmol-1 DuPont NEN Research Prod- ucts, USA, and 90% non-radioactive valine). Af- ter addition of [3H]thymidine or [3H]valine the subsamples were incubated for typically 40 min at 6°C, after which the incubation was stopped by addition of 5 ml cold 10% trichloroacetic acid (TCA). After 30 min of TCA extraction the samples were filtered thought 0.2 /~m membrane filters. Radioactivity of the filters was assayed with liquid scintillation counting (LSC).

Bacterial incorporation and respiration of glucose and fructose were measured with uniformly labelled o-[14C]glucose (250-360 mCi mmol - l ) and D-[14C]fructose (200-350 mCi mmol-1), both from DuPont NEN Research Products, USA. Water samples of 20 ml were added to 20000 DPM of [14C]glucose or [14C]fructose and incubated at 6°C for 30 to 45 rain. The incubations were stopped with buffered (pH 8.5) formalde- hyde to a 2% final concentration. To measure respiration of the two monosaccharides, the incubation bottles were closed with rubber stoppers from which plastic cups with folded paper wicks were attached. The wicks had previously been soaked with 500 /xl Carbo-Sorb CO2-absorber (Packard Instruments, the Netherlands). With a hyperdermic needle 200 /xl 85% H3PO 4 was added to the water sample in each bottle. After 1 h on a shaking table, the CO2-traps were transferred to 20 ml scintillation vials and radioassayed by LSC. Bacterial incorporation of glucose and fructose was determined by 0.2 /zm membrane filtration of the acidified water samples after removal of the CO2-traps. The filters were radioassayed by LSC. Addition of [14C]glucose and [14C]fructose to the water samples made up a maximum of 1.8 nM of each monosaccharide, corresponding to 3.7-4.0% of their ambient concentrations.

Incorporation and respiration of free amino acids was determined from addition of 20000 DPM of an equimolar mixture of L-[14C(U)]amino acids (glutamic acid, serine, glycine and alanine; 58-285 mCi mmo1-1 (Amersham, UK)). Other- wise the procedure for amino acid assimilation was identical to that used for glucose and fructose. Addition of the four 14C amino acids increased the natural concentration of these amino

81

acids by a maximum of 2 nM, or 0.9-4.1% of their ambient concentration.

Phytoplankton production was determined in the lake from 06.00 to dark at 17.00 h. Two light and one dark bottle of 110 ml were added 2/zCi H[14C]O~ -. At dark the content of the bottles was filtered through 25 mm G F / F filters (Whatman, UK) and radioassayed. The natural concentration of total CO 2 was measured by gas chromatography after acidification of 20 ml lake water in a closed serum bottle.

Bacterial abundances were determined by acri- dine orange direct counts according to Hobbie et al. [20].

Chemical analysis PAD-HPLC analysis of saccharides. Equipment

for the chromatographic analysis of mono- and disaccharides consisted of two Waters 510 pumps, a Waters 712 WISP autosampler, a Waters 464 pulsed amperometric detector with a gold electrode, two computer controlled solvent select valves and a Waters Maxima 820 or Millennium data acquisition and processing module (all from Millipore/Waters Associates, USA). Due to leakage of the viscous NaOH solvents from the autosampler injector, a manual Rheodyne 7125 injector (Rheodyne Inc., USA) was used in some of the analyses. For separation of the saccharides a CarboPac PAl 250 × 4 mm column and a Car- boPac PA 52 × 3 mm guard column (Dionex Cor- poration, USA) were used. For reduction of solvent pump pulses, both pumps had pulse dampening units installed. An additional pulse dampening LP-21 device (Scientific Systems Incorp., USA) was positioned on the solvent line between pumps and autosampler. In preliminary analyses a Waters solvent conditioning system with He sparging was used. However, degassing of the solvents introduced a large negative peak close to the peaks of glucose and fructose, so degassing was not used routinely. The negative peak may have been due to dissolved oxygen in the injected samples (J. Rich, personal communication).

Chromatographic solvents. NaOH at different concentrations were the dominant HPLC solvents. Solutions of NaOH were prepared from a 50% low carbonate NaOH stock solution (J.T.

82

Baker, the Netherlands). Carbonate inactivates the PAl column. To reduce the carbonate content of the solvents, a 50% N a O H solution, in which carbonates will precipitate, was used. Only N a O H from the upper bottle volume free of precipitate was used. In most analyses mono- and disaccharides were separated with a 14.3 mM NaOH. For conditioning of detector electrodes and the PAl column in between sample injec- tions, a 150 mM N a O H concentration was used. Weekly cleaning and conditioning of the column was carried out with 1 M NaOH. For short- term storage of the PAl column, e.g. overnight, the 300 mM N a O H solution was adequate, but for long-term storage a 1 M N a O H was used.

Chromatographic conditions. For analysis of natural water samples, settings of the pulse am- perometr ic detector were the following: (1) E 1 = 50 mV, T~ = 15 cycles and t = 0.3 s; (2) E 2 -= 800 mV, T 2 = 10 cycles and t = 0.2 s; (3) E 3 - - 4 0 0 mV, T 3 = 15 cycles and t = 0.3 s. The current of the electrode was set to 0.1 or 0 .2/xA.

Typical solvent composition for separation of mono- and disaceharides were: (A) 14.3 mM NaOH, and (B) 150 mM NaOH. The solvent composition was: 100% A (0 to 14.0 rain) ~ 100% B (linear gradient 14.0 to 15.5 min)--+ 100% B (15.5 to 17.5 rain) ~ 100% A (linear gradient 17.5 to 19.0 rain) ~ 100% A (19.0 to 70 rain).

Regenerat ion of the PAl column was necessary after analysis of 30 to 40 water samples, otherwise the separation decreased. Depending on the actual solvent (14.3 or 150 mM NaOH), the following sequence was used for column re- generation (at a 1 ml min-1 flow rate): 14.3 mM N a O H (or 150 mM NaOH) ~ water ~ 1 M HCI (for removal of monovalent cations) ~ water 100 mM oxalic acid (for removal of di- and triva- lent cations) ~ water ~ 1 M N a O H ~ 14.3 or 150 mM NaOH. Flushing with each solvent lasted about 15 rain, with a gradient period of about 5 rain between changes of solvents.

Sample pretreatment. To raise the detection level of the P A D - H P L C analysis all samples for measurements of free mono- and disaccharides were concentrated 10 to 15 times before injection. Water samples of 15 ml were freeze-dried and redissolved in 1.0 or 1.5 ml Milli-Q water

(Millipore). Precipitates in the redissolved samples were removed by filtration through 0.2 ixm filter mini-cartridges. Addition of [14C]glucose and [~4C]fructose to parallel natural water samples demonstrated a recovery of these two saccharides of at least 97.3% after sample treatment, suggesting that insignificant amounts of monosaccharides were lost during the procedure.

Individual saccharides of combined carbohydrates (DCCHO) were determined in 1.0 ml 0.2 /xm filtered water samples. The samples were transferred to glass ampoules and Suprapur HC1 (Merck, FRG) was added to a final concentration of 1.5 M HCI. The ampoules were flushed with N2, sealed and heated to 100°C for 4 h. After cooling the hydrolyzed samples were frozen and freeze-dried to remove HC1 and water. The dried samples were redissolved in Milli-Q water, soni- cated for 5 s and filtered though 0.2 /xm membrane filters.

The HC1 concentration of 1.5 M and the 100°C hydrolysis tempera ture were chosen after several preliminary experiments. PAD-HPLC analysis indicated that a stronger HC1 concentration and a higher hydrolysis tempera ture reduced the concentration of both tested carbohydrates and of D C C H O in natural water samples. Fructose and a portion (20-70%) of arabinose were found to be destroyed during the hydrolysis.

For PAD analysis of mono- and disaccharides in the 10-15 × concentrated water samples or in the unconcentrated, hydrolyzed samples, volumes of 100 /xl typically were injected into the HPLC system.

HPLC analysis of DFAA. Dissolved free amino acids were analyzed as fluorescent o-phthaldialdehyde derivatives by HPLC according to Lin- droth and Mopper [21] and J0rgensen et al. [22].

Results

PAD-HPLC Analysis of saccharide standards. The chro-

matographic equipment was routinely tested with two different standard mixtures of saccharides, containing 750 nM of each of either arabinose, galactose, glucose, mannose and fructose, or fu-

0.1

> E

0.0

-0.1

0.3

0.1

> E 0.2

0.0

0.5

1.0 -

0.9 -

0.4

0 . 3 -

0 . 2 -

0 . 8 .

0.7 --

0 . 6 2

0,5 --

i

5

® A

e

i

10 15

0 .4 .

0 . 3 -

0.2 -"

0.1

0

_o= ==

d

5 1Q 15

5

0.2

i .

o

J J

5 10 15

D

o=

r

10 15

Retention time (min)

83

cose, rhamnose, glucose and xylose (Fig. 1 A and B). As expected for an isocratric analysis, separation of the saccharides was reduced with an in- creasing retention time. The baseline noise was partly caused by pump pulses, despite two pulse dampening devices having been installed. The baseline noise was significantly reduced using higher NaOH concentrations, as larger peaks were obtained. However, at NaOH concentrations above about 20 mM, saccharides do not separate efficiently. In preliminary analyses a post-column pump with 300 mM NaOH and a flow of 0.3 ml rain-1 was connected to the post- column eluent. Larger peaks were measured, but unfortunately an increased peak width made the identification of closely eluating peaks difficult (data not shown). Therefore all analyses were performed with 14.3 mM NaOH as the only eluent.

Analysis of natural water samples. PAD chromatograms of a 13.3 X concentrated water sample and an unconcentrated, bydrolyzed water sample are shown in Fig. 1C and D. Mannose and xylose often coeluted and could not always be separated. Note that fructose was degraded during the hydrolyses. Individual saccharides in the water samples will be presented later. Free saccharides can probably be determined without a concentration step prior to analysis in most natural water samples. However, pump pulses of the present HPLC system made the concentration step necessary.

Reproducibility of the PAD-HPLC analysis. Concentrations of mono- and disaccharides were determined from duplicate or occasional tripli-

Fig. 1. PAD-HPLC chromatograms of mono- and disaccharides. Standard mixture of 750 nM of arabinose, galactose, glucose, mannose and fructose (A), and fucose, rhamnose, glucose and xylose (B). Monosaccharides in a 13.3x concentrated Lake Fures¢ water sample collected at 06.00 on November 7 1990 (C). Mono- and disaccharides in a hydrolyzed, unconcentrated sample from the long-term batch experiment at 55 h (1)). Chromatographic conditions: the solvent was 14.3 mM NaOH; the electrode current was 0.1 /~A; 100 /zl samples were injected. Note: the mV scales are relative and cannot be compared. During data conversion of the

original chromatograms these values were changed.

84

cate HPLC analyses of the water samples. For concentrations < 100 nM (before correction for the 10-15 × concentration step of the D F C H O samples), deviations from the means of the individual saccharides on the average were 16.6% (range 5.4-31.8%, n = 48). For concentrations > 100 nM, deviations of the means on the average were 7.9% (range 0.2-21.4%, n = 122). These variations were representative to both saccharide standards, natural water samples and natural samples spiked with saccharide standards (data not shown). Blank samples (Milli-Q water) showed only glucose peaks. In 4 of 10 blanks, glucose concentrations < 18 nM (unconcentrated or 10 × freeze-dried concentrated Milli-Q water) or < 32 nM (hydrolyzed MiUi-Q water) were measured. The measured concentrations were not corrected for the blanks values.

Diel changes of DFCHO and DFAA and bacteria in Lake FuresO

Glucose. In both the in situ light-dark and dark incubated samples, large concentration changes of glucose were observed during the 18 h study period (Fig. 2A). In the in situ light-dark incubations the concentration of glucose varied between 9 (24.00) and 49 nM (18.00). Concentrations of glucose in the dark incubated samples decreased from 35 nM (at 06.00 and 09.00) to 9 nM at midnight. The average diel glucose concentration was higher in the in situ light-dark (28 nM) than in the dark incubations (19 nM).

Bacterial gross assimilation (incorporation and respiration) of glucose varied from 0.4 to 3.1 nmol 1-1 h-~ (Fig. 2B). At 09.00 and 12.00 different assimilation rates occurred in the in situ light-dark and the dark samples, but otherwise a rather similar assimilation was measured in both sets of samples. Changes in concentration and assimilation in the in situ light-dark incubations were similar at a 5% level (Spearman rank corre- lation analysis). The total diel assimilation in the two incubations did not differ statistically from each other (paired t-test, P < 0.05). Respiration of glucose made up 8 -13% of the gross assimilation at 06.00, 09.00 and at 24.00 in both incubations. From 12.00 to 21.00 the respiration in-

L~

E c

L:

s

T

c

E

o

O

1

2

_.Cd- • DARK

~]rne

Fig. 2. Concentrations and gross assimilation rates of glucose (A and B respectively) and fructose (C and I)) in Lake Fures¢ on November 7 1990. The water was incubated at in situ light-dark conditions (Light) or in the dark (Dark). Natural light and dark periods are indicated on panel D. n = 4 (con-

centrations) and n = 6 (assimilation rates); + 1 S.D. shown.

creased and varied from 23-25% (in situ light- dark) and 24-33% (dark).

Fructose. The concentration of fructose ranged from 12 (21.00; in situ light-dark incubation) to 44 nM (24.00; dark incubation) (Fig. 2C). The dark incubated samples showed minor variations in concentrations, but in the in situ light-dark incubations the concentration first declined from 37 (09.00) to 12 nM (21.00) and then increased to 44 nM (24.00). The average fructose concentration was lower in the in situ light-dark (26.9 nM) than in the dark incubations (32.3 nM).

Gross assimilation of fructose in the in situ light-dark-incubated samples was below 1.0 nmol

1- x h - 1 during most of the period, but simultaneous with the higher concentration at 24.00, the assimilation doubled (Fig. 2D). In the dark incubation the assimilation varied significantly during the period, with the highest values between 09.00 and 15.00. The average assimilation in the dark- incubated samples was 46% larger than assimilation in the in situ light-dark incubation. Respira- tion of fructose varied during the studied period, with values between 16 and 29% in both incubations. No correlations between changes of concentrations and assimilation rates of fructose were found.

DFCHO. Changes of the total DFCHO concentrations in the two sets of incubations were rather similar (Fig. 3A). From 06.00 to 15.00 the concentration decreased from 119 to 67 (in situ light-dark) and 78 nM (dark). In the evening the concentrations increased to maxima of 185 (dark) and 224 nM (in situ light-dark) at 21.00. The composition of monosaccharides in the in situ light-dark incubations differed from that in the dark incubations (Fig. 3B and C), but no trends were found. In the in situ light-dark incubations galactose, glucose, mannose/xylose and fructose each constituted 18-27 mol-% of the DFCHO. Arabinose and melibiose were less important and made up 5 -7 mol-%. In the dark incubations the monosaccharide spectrum had changed; fructose was dominant, followed by mannose/xylose and glucose. The relative importance of arabinose and melibiose had doubled compared to the in situ light-dark incubations. Trace amounts ( < 2 nM) of cellobiose were measured in both sets of incubations.

DFAA. The concentration of DFAA in the in situ light-dark incubations decreased from 309 to 174 nM during the day period, but varied from 162-233 nM in the evening (Fig. 4A). In the dark incubations DFAA declined to 97 nM at midnight (Fig. 4B). Gross DFAA assimilation at in situ light-dark ranged from 6.5 (21.00) to 18.4 nmol 1-1 h-1 (24.00). In the dark the assimilation decreased during most of the period, with rates of 2.6 (21.00) to 10.0 nmol 1-1 h - I (09.00). The diel, integrated DFAA assimilation was 2.0-fold larger in the in situ light-dark than in the dark incubations. Respiration of the amino acids in the

85

2 5 0

2OO

c 1 5 o

L~ 1 o o c3

5 0

O LIGHT A • DARK r~ --~-~. (~

I J ' I m + - - ~ - m ~ - I r 6 9 12 15 18 21

T;me 2 4

~o

3 0

2 5

2 0

15

10

5

0

LIGHT ~~

ARA ~AL GLLI MAN/XYL FRU MEL

o 30 o 25 2O 15 10 5 0

DARK ~ ~ C f

ARA GAL GLU MAN/XYL FRU MEL

Fig. 3. Total concentrations of dissolved free carbohydrates (DFCHO) in the diel in situ light-dark and dark incubations (A). Natural light and dark periods are indicated. Average composition of individual monosaccharides in the in situ light-dark (B) and dark (C) incubations, n = 4; + 1 S.D. shown.

in situ light-dark incubations fluctuated between 14 to 45%, with a mean of 27%. In the dark incubations a higher respiration of 23-65% (mean 46%) was found. No correlations between concentration changes and assimilation rates were found. The most abundant DFAA in both sets of incubations were serine, glycine and alanine.

Bacteria. The number of bacteria varied from 0.93 to 1.22 x 106 m1-1 in the in situ light-dark incubations (Fig. 5A). In the dark incubations the numbers were more variable and ranged from 0.86 to 1.39 x 106 ml-x. The bacterial production was measured from incorporation of [3H]thymidine and [SH]valine. The thymidine incorporation indicated a declining production during most of the period in both sets of incubations (Fig. 5B).

86

The incorporation of valine was more variable, demonstrating 4.4- (from 06.00 to 15.00) to 7.4- fold variations (06.00 to 21.00) in the in situ light-dark and the dark incubations, respectively (Fig. 5C). The diel, integrated incorporation was 1.5- (thymidine) and 1.4- (valine) fold larger in the in situ light-dark than in the dark incubations, respectively.

Long-term changes of DFCHO, DCCHO and DFAA in batch cultures of Lake Fures¢ water

DFCHO. Concentration changes of dissolved free saccharides in the 1.0 /zm filtered water demonstrate that both saccharide-producing and saccharide-consuming processes occurred during the 118 h incubation period (Fig. 6A). Measured concentrations of DFCHO ranged from 54 nM (73 h) to 194 nM (12 h), with peaks at 12, 55 and 118 h. Dominant saccharides were glucose (18-43 nM), fructose (9-57 nM), cellobiose (3-52 nM), melibiose (8-21 nM) and mannose/xylose (5-27

. • . . •

500 • • Concentrot~on • Asstmdabon ±

g 2o° t - 150

1 O0

5° LIGHT

o ] I , 4-- 4 - - = - ~ 4 , I , ! 6 9 12 ~5 ",8 21 24

2O

L m. 1 6 ~

i

12 E c

g

4 "~

350

I 4

: \ q

s~ { B 4

6 q 12 i [. 18 ~" "

T ime 'Y')

1 2

DARK t

9 7

5

24

Fig, 4. Concentrations and gross assimilation rates of dissolved free amino acids ( D F A A ) in the diel in situ light-dark (A) and in the dark incubations (B). Natural light and dark periods are indicated. For concentration measurements pooled samples of two replicate bottles were analyzed. Assimilation

rates: n = 6; +_ 1 S.D. shown.

1 6

~4 t ~2

®o 70 t 0 8 ~

0 o O6 t

0 4

c~ o 2

DO I 6

r

P

9 12 15 18 21 24

I

7

E o~ v

& o u c

%

I m_

7

CL

m.

lO

8

6

4

2

0

1808 I 1500

1200

900

6 9 12 15 18 21 24

o :u; / ' \ [ • DARK

/ \ i

0 1 L + - - + - - ~ - - + I < , . * 1 , i , I 6 9 12 15 18 21 24

T i m e (h )

Fig, 5, Bacterial numbers (A), incorporation of 10 nM [3H]thymidine (B) and 100 nM [3H}valine (C) in the Lake

Fures¢ on November 7 1990; n = 6; +_ 1 S.D. shown.

nM). Arabinose and galactose were less important and ranged each from 1 to about 9 nM. During the initial 12 h incubation period cellobiose was reduced from 52 to 20 nM, simultaneous with the increase of glucose and fructose. Cellobiose is the disaccharide unit of cellulose and thus, an enzymatic degradation of cellobiose from 52 to 20 nM would result in 64 nM glucose. Since the glucose concentration was only 43 nM at 12 h, there probably was a significant uptake of glucose within the initial incubation period.

Gross assimilation (incorporation and respiration) of glucose and fructose varied considerably during the incubation (Fig. 6B). From an initial glucose and fructose assimilation of 1 nmol 1-1 h-~, the assimilation of glucose increased to 9.7 nmol 1-1 h -1 (118 h). The assimilation of fruc-

tose also increased and peaked at 55 h (6.9 nmol 1-1 h - l ) , but relative to glucose lower values occurred at 73 and 118 h. The respiration of glucose ranged from 9% (t = 0) to 82% (t = 12 h), with a mean of 27%. For fructose minimum and maximum respiration values also were measured at t = 0 h (5%) and t = 12 h (56%), with a mean of 23%.

DFAA. The concentration of free amino acids decreased from an initial level of 450 nM, to 68 nM at 118 h (Fig. 6C). Gross assimilation of the D F A A raised significantly during the experiment, but a decline occurred at 55 h, simultaneous with the DFCHO peak, and at 118 h. Within the initial 12 h most (92-93%) of the assimilated

20O

16O

120

©

40

A T • [ct,a~ ?rcHo 0 ~' (~elhDb:OSe

20 40 60 80 1 O0 120

10 • , "T B • G l u c o s e 0 F r u c t o s e ~

J : 8

E o 4

o

20 40 60 80 100 ~20

600 " ~ C 0 C o n c e n t r a t i o n • A s s i m i l a t i o n 40 I

,oo / ~o -

2oo- i ~o

Too - i-I r~ .~ /~ 0 I~_I F-- J I ~ , , - - 0

0 20 40 60 80 1 O0 120

T i m e ( h )

Fig. 6. Concentrations of dissolved free mona- and disaccharides (DFCHO) (A), gross assimilation rates of glucose and fructose (B) and D F A A (C) in the long-term experiment; n = 4 (for concentrations of DFCHO; for concentrations of D F A A one pooled sample was analyzed) or n = 6 (assimila-

tion rates); 5:1 S.D. shown.

3.0

87

2 5

" ~ 2.0

O 1.5 I o

1.0

0.5

0,0

1000

A

, , ~ , , , ~ . . . . . . 2 40 6 80 100 120

• Glucose • Cellobiose a Monnose /Xy lose 900 [3 Galoctose o Arobinose o MeliDiose B

o 400 o I o ~ 300

T

l ' ~ o o o o @ o

0 J J ~ i J , i ~ t ,

o 20 ,o 6o so ~oo ~2o

T i m e ( h )

Fig. 7. Concentrations of total dissolved combined carbohydrates (DCCHO) (,4,) and individual saccharides of the carbohydrates (B) in the long-term experiment; n = 2. Mean concentrations and maximum and minimum ranges are indicated.

D F A A was respired, but during the rest of the period, the respiration varied from 27 to 48%. Composition of the D F A A did not differ significantly from that found in the diel experiment.

DCCHO. The concentration of dissolved combined saccharides ranged from 1.273 to 2 .204/zM (Fig. 7A). Carbohydrates were both removed and produced during the incubation. Thus, from 0-24 h and 55-73 h the DCCHO concentration increased from 1.440 to 2.204 /.tM and 1.273 to 2.197 /zM, respectively. The concentration decline of 388 nM from 45 to 55 h coincided with the 72 nM increase of the free saccharides (Fig. 6A). Interestingly, the DCCHO concentration at t---0 h and t = 118 h were similar. From the assimilation rates above, the total bacterial con- sumption of glucose and fructose during the 118 h was estimated to 670 nM. This indicates that

88

there was a production of these two monosaccharides in the batch cultures. A likely source of the monosacchar ides was bacterial ly p roduced polysaccharides, as discussed later.

Composition of the saccharides demonstrates that the observed DCCHO decline at 55 h mainly was due to a reduction of cellobiose and mannose/xylose, as the proportion of the other saccharides either increased (arabinose) or largely remained unchanged (glucose, galactose and melibiose) (Fig. 7B). The dominant combined saccharides were glucose and galactose, each making up 40-30 mol-% of the DCCHO pool. Unfortunately fructose and to some extent arabinose were degraded during the hydrolysis. Since the fraction of arabinose lost during hydrolysis was variable (20-70%), the actual concentrations

1 6

~ 14 1 2

0 1 0

0 8 o ~v 0 6

~ 0 4 C~

0 2

~ 14

i 12 r

i

~o

E 8

6

4

% 2 ~ 0

20

j J J J ~

P

/%' f 20 40 60 80 1 O0 120

n r ~ r ] , 0

O 20 40 60 80 100

q

k 1 5

5 E

? O0 -~ v-- , n r ~ r ~ - -

0 20 40 60 80 1 O0 120 ~ ] m e ( h )

Fig. 8. Bacterial numbers (A), incorporation of 10 nM [3tt]tbymidine (B) and incorporation of 100 nM [ 3H]valine (12)

in the long-term experiment; n = 6; +_ 1 S.D. shown.

of arabinose in the lake water may have been different from that shown in Fig. 7B. In addition to the saccharides shown in Fig. 7B, fucose was found to make up 0.4-3.3 tool-% of the DCCHO.

Bacteria. The bacterial abundance varied from 0.78 to 1.43 cel ls× 106 ml -~, with minor in- creases at 45 and 55 h (Fig. 8A). Incorporation of thymidine and valine fluctuated during the experiment. Until 45 h the bacterial incorporation of the two compounds varied opposite each other, i.e. a high thymidine incorporation coincided with a low valine incorporation and vice versa• But simultaneous with the DCCHO decline and the DFCHO increase at 55 h, both methods indicated an increased production.

Discussion

Dissolved free saccharides in freshwaters have demonstrated significant short-term fluctuations [10,13, present study], but the mechanisms caus- ing such changes are largely unknown [23]. A detailed insight into these concentration changes may provide new information on production, sources and utilization of the saccharides, includ- ing exoenzymatic degradation [24].

So far such intensive analyses of individual saccharides have been impeded by laborious and time-consuming analytical procedures. With the PAD-HPLC technique a large number of mono- and disaccharide analyses now can be performed with no or minor sample work-up. However, the procedure still has some limitations. Major ad- vantages of PAD-HPLC are the following: (1) a high sensitivity that allows direct measurements of most saccharides at a 5 nM (or less) detection limit, corresponding to 0.5 pmol in a 100 /~1 injection volume; (2) no sample clean up, at least of freshwater samples (seawater samples require a desalting procedure [17]); and (3) the analysis is inexpensive and requires no derivatization.

Disadvantages of the procedure appear to be: (1) the long period of equilibration of the detector before analysis; a 1-2 h equilibration period is required when working in the 0.1-0.2 /zA range (to get a high sensitivity), when the detector has been switched off; (2) the susceptibility to

89

pump pulses to obtain maximum sensitivity. Low-pulse pumps or efficient pulse dampening devices must be used; and (3) the time-consuming cleaning of the column at regular intervals. The inconvenience of the equilibration period can be diminished by analysis of a large number of samples in a continuous series, e.g. with an autoinjec- tor; or, if using manual injection, an overnight solvent flow at a low rate with the detector turned on. Similarly, the column clean-up can be auto- mated with a computer controlled solvent deliv- ery system. In addition to saccharides, the PAD- HPLC procedure can also be used for analysis of amino sugars and uronic acids [16].

DFCHO. The DFCHO composition in Lake Fures0 water indicates that galactose, glucose, fructose and mannose/xylose were of equal importance. This composition deviates from DF- CHO spectra measured in other freshwaters, in which glucose has been found the dominant DF- CHO, and in which mannose and xylose were, less abundant [10,13,16]. The most important sources of DFCHO in freshwater appear to be enzymatic degradation of polymeric particulate and dissolved organic material, and phytoplankton extracellular release [23]. Since glucose is a major component of polysaccharides in microalgae [25], a dominance of glucose may be expected at periods with a high phytoplankton abundance. Simi- larly, in waters with vascular plants high concentrations of glucose have been related to degradation of cellulose [16]. In the present long-term experiment cellobiose initially was the major free saccharide in the water, but after 12 h the cellobiose concentration was reduced, simultaneous with an increase of glucose (and fructose). This may indicate that there was an enzymatic degradation of cellobiose to glucose and that cellulose was a source of glucose in the water. Specific sources of the measured mono- and disaccharides were not studied, but all saccharides, except melibiose, are constituents of both microalgae and vascular plants [16,25].

In Lake Fures0, the high abundance of saccharides other than glucose may illustrate both a preferential assimilation of glucose and partly fructose, relative to other saccharides, and a production of the individual monosaccharides before

and during the experiments. In the 18 h field study period, assimilation of glucose consumed 85% (in situ light-dark) and 112% (dark) of the average, ambient glucose pool. For fructose, the values were 52% (in situ light-dark) and 63% (dark). Since glucose and fructose did not demonstrate similar reductions in the concentrations, the present observations indicate that a production of glucose and fructose occurred. In the long-term experiment, the bacterial activity peak at 55 h (observed with both valine and thymidine incorporation) coincided with a reduction of DC- CHO and an increase of DFCHO. This suggests that DCCHO were degraded to DFCHO by enzymatic activity. But apparently only a portion of the produced DFCHO were taken up by the bacteria, as the concentration of DFCHO in the water increased. The DFCHO and DCCHO changes indicate that abundance of saccharides in both experiments were influenced by the bacterial activity.

Other sources of DFCHO (and DCCHO) in Lake Fures¢ water probably were bacterial extracellular production of polysaccharides. Polysac- charide production by bacteria have been observed in several species and appears among other factors to be regulated by the nutritional state of the cell, e.g. starvation [26-29]. Bacterial polysaccharide production must have caused the DC- CHO increase in the long-term experiment during the initial 24 h and after 55 h (Fig. 7), since no other organisms were present. Glucose and fructose were among the produced monosaccharides. During the 118 h incubation, 670 nM of glucose and fructose were assimilated; but since there was a net production of 31 nM of free glucose and fructose (difference between start and final concentrations), a production of the two monosaccharides of about 700 nM occurred. This production corresponds to about 46/zg C 1-1 118 h -~, or 41% of the net bacterial carbon production (valine incorporation, Table 1). This appears to be a substantial loss of cell material. It is uncertain if the polysaccharide production was caused by lack of certain nutrients, or if release of polysaccharides is commonly occurring in natural populations of aquatic bacteria. The polysaccharide production emphasizes that natural dy-

90

namics of DCCHO and D F C H O probably are more complex than generally expected.

The presently measured DFCHO concentrations align with other studies of either glucose or total DFCHO in freshwater [8-10] and seawater [11,12,17], but higher concentations have been measured in eutrophic lakes [5,13] and in water of macrophyte beds [16].

DCCHO. Major saccharides of the DCCHO were glucose and galactose (Fig. 7B). Among other saccharides that occasionally made up more than 10 mol-% of the DCCHO were cellobiose and arabinose. The dominance of galactose and glucose agrees with other studies, although mannose also have been found abundant in lake water ([3], Tranvik and Jcrgensen, in prep.). Glu- cose is expected to be abundant as it is a major component of many carbohydrates e.g. of cellulose, hemicellulose and of storage polysaccharides of phytoplankton [12,16]. Galactose is less abundant than glucose in nature, but has been found in structural carbohydrates of wood and in marine diatoms [12,16].

The composition of DCCHO in the long term-experiment was surprising stable, consider- ing that the DCCHO concentration increased from 1.44-2.20 /xM and 1.27-2.20 IxM between 0-24 h and 55-73 h, most likely due to bacterial exopolysaccharides. If these polysaccharides differed in composition from the lake water DC-

CHO, a different saccharide composition should be expected. Major saccharides of exopolysaccharides are glucose and galactose [27]. Both proportion and concentration of galactose increased between 12-24 h and 53-73 h (the concentrations went up by 18 and 21%, respectively), simultaneous with the total DCCHO increase, but glucose was reduced between 55-73 h (Fig. 7B). This may indicate that the exopolysaccharides had a relative large content of galactose. The fluctuating pools of DCCHO-arabinose (from 1.4-18.1 mol- % between 24 and 73 h) may be related both to a production of exopolysaccharide and to exoenzymatic degradation, but no literature data on microbial arabinose turnover could be found. In addition, loss during hydrolysis of the DCCHO may have influenced the abundance of arabinose, as mentioned previously.

DCCHO concentrations in Lake Fures¢ water fall within the concentration range previously measured in studies of individual saccharides of carbohydrates in lakes [3,5,14], but are lower than DCCHO concentrations determined with colorimetric assays [6]. Possibly non-standard sugars or sugar derivatives of natural D C C H O are included in the colorimetric assays, but not in the chromatographic procedures. In addition, manipula- tion of the water samples before chromatographic analysis may influence the measured concentrations [14].

Table 1 Bacterial production (incorporation of valine and thymidine), assimilation of glucose, fructose and DFAA and primary production in Lake Fures¢ a

Method D{el study Diel study Long-term study (in situ light-dark incubation) (dark incubation) /~g C 1-1 118 h -1 /~gCl -I 18h -l /xgCl -I 18h -1

Valine incorporation b 14.81 + 9.69 10.39 _+ 6.78 112.73 +_ 31.31 Thymidine incorporation c 6.86 _+ 1.28 (46.3%) 4.53 + 0.44 (43.6) 52.83 5:5.94 (46.9%) Glucose assimilation 1.44 + 0.44 (9.7%) 1.29 + 0.09 (8.9%) 23.28 5:2.27 (20.7%) Fructose assimilation 0.69 + 0.13 (4.7%) 1.00 5:0.10 (9.6%) 12.70 + 2.14 (11.3%) DFAA assimilat ion 4.88+1.01(32.0%) 2.395:0.27(23.0%) 39.365:4.3 (34.9%) Primary production 12.48 5:1.34 (84.3%) nd nd

a All rates are net values _+ 1 S.D. Numbers in parentheses indicate values relative to bacterial production based on incorporation of valine. b Assuming that valine made up 7.4% of the bacterial proteins, that 64% of the bacterial dry weight was protein and that 55% of the dry weight was carbon [32].

Assuming a conversion factor of 1.5 × 1018 cells mol- 1 thymidine [33], a cell volume of 0.I/xm -3 and 0.35 pmol C/xm -3 [34]. nd: no data.

91

Assimilation of DFCHO. In both sets of experiments a substantial uptake of glucose and fructose was measured. By comparing assimilation rates and actual pools in the diel experiment, it was estimated that 85-112% and 52-63% of the ambient glucose and fructose, respectively, were turned over during the studied period. In the experiment, assimilation of glucose appeared to influence the concentration, as statistically similar changes of occurrence and uptake of glucose were found in the in situ light-dark incubations. In the long-term study, the glucose and fructose pools were turned over 12 and 8 times, respectively, during the incubation. Except for a simultaneous concentration and assimilation peak at 55 h, no similarities between variations of pools and uptake occurred.

A minor portion (16-33%) of the assimilated glucose and fructose typically was respired in the diel experiment and at most sampling times in the long-term experiment. However, a lower glucose respiration of 8-13% was measured in the morning and at midnight in the diel study, and of fructose and glucose in the long-term study at the start of the experiment (5 and 9% respiration, respectively). These occasionally low respiration values possibly indicate periods of special metabolic events of the bacteria, e.g. intense periods of synthesis of specific compounds required for cell production. Similarly, the measured 82% respiration of glucose (at 12 h in the long-term experiment) may indicate a period of a large energy requirement for synthetic processes. The present respiration in both experiments is rela- tively low compared to previous respiration values of 27-56% (glucose) and 29-67% (fructose) measured in eutrophic Danish lakes [10], but it agrees with values determined in the German PluBsee [30].

Occurrence and assimilation of DFAA. Except for a temporary increase (between 18.00 and 21.00) in the diel in situ light-dark experiment, ambient concentrations of DFAA declined in the diel and long-term experiments. The measured concentration range (< 100 to 455 nM) and composition (serine, glycine and alanine were dominant) both agree with previous observations of Danish lakes [31].

From the assimilation rates in the diel experiment it can be estimated that 84% (in situ light- dark) and 68% (dark) of the DFAA pools were consumed during the incubation. In the dark the DFAA pool actually was reduced by 69% between 06.00 and 24.00, indicating that no production of DFAA occurred during the experiment. At in situ light-dark the concentration only was reduced by 48% between 06.00 and 24.00, although the assimilation rates indicated a 84% reduction. This difference indicates that a production of DFAA occurred, as supported by the observed concentration increase at 18.00 (Fig. 4A). The higher DFAA pool in the in situ light- dark incubations probably can be referred to release during phytoplankton primary production. The production of DFAA at in situ light-dark may also explain the two-fold higher DFAA assimilation in these incubations than in the dark incubations. A comparison of assimilation rates and DFAA pools in the long-term experiment indicates that the ambient DFAA pools were turned over about 5 times during the 118 h incubation. The source of these amino acids was probably dissolved combined amino acids that were degraded by extracellular enzymes [32].

The significant variation of the amino acid respiration (from 14% in the diel in situ light-dark incubation, to 93% at start in the long-term experiment) indicates that amino acids were in- volved in different metabolic activities. An initially high respiration rate is generally found in batch cultures of aquatic bacteria (Jorgensen, Kroer and Coffin, unpublished results), probably indicating production of energy. The higher average respiration in the dark (46%) than in the in situ light-dark condition (27%) in the diel study speculatively may be related to the absence of photosynthetic activity, e.g. the lack of phytoplankton excretion products usually used for energy production. Most of the measured respiration values fall within the 40-65% range commonly found in lakes [10,31].

Microbial carbon fluxes. Bacterial production and assimilation in the incubations were con- verted to carbon fluxes to allow a comparison of the various microbial processes. Incorporation of thymidine was calculated as bacterial carbon flux

92

using l i t e ra tu re values , as the ac tual convers ion factors were not d e t e r m i n e d . Since ca lcu la t ion of the bac te r i a l ca rbon flux f rom inco rpo ra t ion of val ine does not r equ i re f ie ld-specif ic convers ion fac tors [19,32], va l ine -based p roduc t i on ra tes will be used as m e a s u r e s of the bac te r i a l p roduc t ion . Bac te r i a l p roduc t i on m e a s u r e d f rom thymid ine i nco rpo ra t i on m a d e up 46 .3 -46 .9% of the produc t ion d e t e r m i n e d with i nco rpo ra t ion of va l ine (Table 1). T h e p h y t o p l a n k t o n p r imary p roduc t i on d id not sus ta in all of this bac te r ia l ca rbon re- qu i r emen t , as it me t only 84.3% of the ne t bac te - r ial p r o d u c t i o n (Table 1).

I n c o r p o r a t i o n (net ass imi la t ion) of glucose, f ruc tose and f ree amino acids m a d e up 47.4, 41.5 and 66.9% of the bac te r i a l ne t p roduc t ion in the in situ l ight -dark , the da rk and long- te rm incuba- t ions, respect ive ly (Tab le 1). In the da rk incuba- t ions, i nco rpo ra t i on of f ruc tose was s imilar to the i nco rpo ra t i on o f glucose, bu t in the o t h e r two expe r imen t s g lucose i nco rpo ra t ion was abou t two-fold l a rge r than inco rpo ra t ion of fructose. T h e da t a in Tab le 1 ind ica te tha t i nco rpo ra t ion of glucose and f ruc tose was i m p o r t a n t re la t ive to i nco rpo ra t i on of all D F A A . Thus, in te rms o f carbon , i nco rpo ra t i on of glucose and f ruc tose equa l ed 4 4 - 9 2 % of the amino acid incorpora t ion . In o t h e r s tudies o f f r eshwate r and mar ine bac te - ria, D F A A ass imi la t ion has b e e n found to sus ta in f rom abou t 30 to 90% of the bac te r i a l ca rbon r e q u i r e m e n t [22,31]. A p p a r e n t l y D F A A were o f re la t ively less i m p o r t a n c e as a bac te r i a l ca rbon source in the lake wa te r dur ing the exper iments .

The p r e sen t expe r imen t s show tha t changes of pools and fluxes of free sacchar ides and carbohy- d r a t e s in bo th na tu ra l lake wa te r and in con- t ro l l ed ba tch cu l tures a re complex and difficult to in t e rp re t . In fu ture s tudies on D F C H O and DC- C H O dynamics in na tu ra l waters , p roduc t i on of ex t race l lu la r po lysacchar ides by bac te r i a should be inc luded , as this p rocess a p p e a r s to be an i m p o r t a n t source of mono- and d isacchar ides .

Acknowledgements

W e wish to t hank Dr. L. Tranv ik for his con- s t ruct ive cr i t ic ism of the manuscr ip t , and Mill i-

p o r e / W a t e r s Assoc ia t e s in D e n m a r k for provid- ing c h r o m a t o g r a p h i c de ta i l s on P A D - H P L C analysis of mono- and d isacchar ides . T h e s tudy was s u p p o r t e d by the Dan i sh Na t iona l Science Re- search Counci l , g ran t no. 11-8298 (for pu rchase of pu l sed a m p e r o m e t r i c de tec tor ) .

References

10

11

12

Satoh, Y., Shoji, S., Satoh, H. and Takanashi, M. (1987) Dissolved organic matter in colored water from mountain bog pools in Japan. I. Seasonal changes in the concentration and molecular weight distribution. Arch. Hydrobiol. 110, 589-603. Miinster, U. (1991) Extracellular enzyme activity in eutrophic and polyhumic lakes. In: Microbial Enzymes in Aquatic Environments (Chrost, R.J., Ed.), pp. 96-122. Springer-Verlag, New York. Stabel, H.H. (1977) Gebundene Kohlenhydrate als stabile Komponenten in Sch6hsee und in Scenedesmus Kulturen. Arch. Hydrobiol. Suppl. 53, 159-254. Itoh, T. (1978) Seasonal fluctuations in the amount of chlorophyll-a and dissolved organic carbon in water of Lake Biwa. Japan J. Limnol. 39, 75-81. Miinster, U. (1985) Investigations about structure, distri- butions and dynamics of different organic substrates in DOM of Lake Plugsee. Arch. Hydrobiol. Suppl. 70, 429- 480. S0ndergaard, M. and Borch, N.H. (1992) Decomposition of dissolved organic carbon (DOC) in lakes. Arch. Hydro- biol. Beih. 37, 9-20. Johnson, K.M. and Sieburth, J. McN. (1977) Dissolved carbohydrates in seawater. I. A precise spectrophotomet- ric analysis for monosaecharides. Mar. Chem. 5, 1-13. Gocke, K., Dawson, R. and Liebezeit, G. (1981) Availabil- ity of dissolved free glucose to heterotrophic microorgan- isms. Mar. Biol. 62, 209-216. Riemann, B., Scndergaard, M., Schierup, H.-H., Bossel- mann, S., Christensen, G., Hansen, J. and Nielsen, B. (1982) Carbon metabolism during a spring diatom bloom in the eutrophic Lake Mosso. Int. Revue. Ges. Hydrobiol. 67, 145-185. Jcrgensen, N.O.G. (1990) Assimilation of free monosaccharides and amino acids relative to bacterial production in eutrophic lake water, Arch. Hydrobiol. Beih. 34, 99-110. Liebezeit, G., B61ter, M., Brown, I.F. and Dawson, R. (1980) Dissolved free amino acids and carbohydrates at pycnocline boundaries in the Sargasso Sea and related microbial activity. Oceanol. Acta. 3, 357-362. Ittekkot, V., Brockmann, U., Michaelis, W. and Degens, E.T. (1981) Dissolved free and combined carbohydrates during a phytoplankton bloom in the northern North Sea. Mar. Ecol. Prog. Ser. 4, 299-305.

13 Miinster, U. (1984) Distribution, dynamic and structure of free dissolved carbohydrates in the PluBsee, a north Ger- man eutrophic lake. Verh. Int. Verein. Limnol. 22, 929- 935.

14 Ochiai, M. and Ukiya, T. (1981) Seasonal variations of dissolved organic constituents in Lake Nakanuma during March and November 1979. Verh. Int. Ver. Limnol. 21, 682-687.

15 Johnson, D.C. and LaCourse, W.R. (1990) Liquid chromatography with pulsed electrochemical detection. Anal. Chem. 62, 589A-597A.

16 Wicks, R.J., Moran, M.A., Pittman, L.J. and Hodson, R.E. (1991) Carbohydrate signatures of aquatic macrophytes and their dissolved degradation products as determined by a sensitive high-performance ion chromatography method. Appl. Environ. MicrobioL 57, 3135-3143.

17 Mopper, K., Schultz, C.A., Chevolot, L., Germain, C., Revuelta, R. and Dawson, R. (1992) Determination of sugars in unconcentrated seawater and other natural waters by liquid chromatography and pulsed amperometric detection. Environ. Sci. Technol. 26, 133-138

18 Fuhrman, J.A. and Azam, F. (1980) Bacterioplankton sec- ondary production estimates for coastal waters of British Columbia, Antarctica and California. Appl. Environ. Mi- crobiol. 39, 1085-1095

19 J0rgensen, N.O.G. (1992) Incorporation of [3H]leucine and pH]valine into protein of freshwater bacteria: Uptake kinetics and intracellular isotope dilution. Appl. Environ. Microbiol. 58, 3638-3646.

20 Hobbie, J.E., Daley, R.J. and Jasper, S. (1977) Use of Nuclepore filters for counting bacteria by fluorescent mi- croscopy. Appl. Environ. Microbiol. 33, 1225-1228.

21 Lindroth, P. and Mopper, K. (1979) High performance liquid chromatographic determinations of subpicomole amounts of amino acids by precolumn fluorescence derivatization with o-phthaldialdehyde. Anal. Chem. 51, 1667- 1674.

22 Jcrgensen, N.O.G., Kroer, N. Coffin, R.B., Yang, X.-H. and Lee, C. (1993) Dissolved free amino acids, combined amino acids and DNA as sources of carbon and nitrogen to marine bacteria. Mar. Ecol. Prog. Ser. 98, 135-148.

23 Miinster, U. and Chrost, R.J. (1990) Origin, composition and microbial utilization of dissolved organic matter. In: Aquatic Microbial Ecology (Overbeck, J. and Chrost, R.J., Eds.), pp. 9-46. Springer Verlag, New York.

93

24 Middelboe, M. and Sondergaard, M. (1993) Bacterioplank- ton growth yield: Seasonal variations and coupling to sub- strate lability and /%glucosidase activity. Appl. Environ. Microbiol. 59, 3916-3921.

25 Brown, M.R. (1991) The amino-acid and sugar composition of 16 species of microalgae used in mariculture. J. Exp. Mar. Biol. Ecol. 145, 79-99.

26 Tempest, D.W. and Neilssel, O.M. (1987) Growth yield and energy distribution. In: Escherichia coli and Salmonella typhimurium (Neidhardt, F.C., Ed.), pp. 797-806. Ameri- can Society for Microbiology, Washington, D.C.

27 Whitfield, C. (1988) Bacterial extracellular polysaccharides. Can. J. Microbiol. 34, 415-420.

28 Wrangstadh, M., Conway, P.L, and Kjelleberg, S. (1989) The role of an extracellular polysaccharide produced by the marine Pseudomonas sp. $9 in cellular detachment during starvation. Can. J. Microbiol. 35, 309-312.

29 Wrangstadh, M., Szewzyk, U., Ostling, J. and Kjelleberg, S. (1990) Starvation-specific formation of a peripheral exopolysaccharide by a marine Pseudomonas sp., strain $9. Applied. Environ. Microbiol., 56, 2065-2072.

30 Gocke, K. (1976) Respiration yon gel6sten organichsen Verbindungen durch natiirliche Mikroorganismen- Populationen. Ein Vergleich zweischen verschiedenen Biotopen. Mar. Biol. 35, 375-383.

31 J0rgensen, N.O.G. (1987) Free amino acids in lakes: Con- centrations and assimilation rates in relation to phytoplankton and bacterial production. Limnol. Oceanogr. 32, 97-111.

32 Billen, G. (1991) Protein degradation in aquatic environments. In: Microbial Enzymes in Aquatic Environments (Chrost, R.J., Ed.), pp. 123-143. Springer Verlag, New York.

32 Simon, M. and Azam, F. (1989) Protein content and protein synthesis rates of planktonic marine bacteria. Mar. Ecol. prog. Ser. 51,201-213.

33 Riemann, B., Bell, R.T. and J0rgensen, N.O.G. (1990) Incorporation of thymidine, adenine and leucine into natural bacterial assemblages. Mar. Ecol. Prog. Ser. 65, 87-94.

34 Bj0rnsen, P.K. (1986) Automatic determinations of bacterioplankton biomass by means of image analysis. Appl. Environ. Microbiol. 51, 1199-1204

Documents

Microbial fluxes of free monosaccharides and total carbohydrates in freshwater determined by PAD-HPLC