Dodds 1*3, Bonnie K. Ellis2 & John C. Priscu' ... Biological Station, University of Montana, Poison, MT 59860 USA; 3 current address: Division of Biology,.
Hydrobiologia 211: 253-259, 1991. 0 199 1 Kluwer Academic Publishers. Printed in Belgium.
253
Zooplankton induced decrease in inorganic phosphorus uptake by plankton in an oligotrophic lake Walter IS. Dodds 1*3, Bonnie K. Ellis2 & John C. Priscu’ ‘Department of Biology, Montana State University, Bozeman, MT, 59717 USA; 2Flathead Lake Biological Station, University of Montana, Poison, MT 59860 USA; 3 current address: Division of Biology, Kansas State University, Manhattan, KS 66506, USA Received 27 June 1989; in revised form 20 March
Key words: phosphate,
Daphnia,
1990; accepted 20 May 1990
phytoplankton,
nutrient uptake
Abstract
Experiments conducted on samples collected from a large oligotrophic lake revealed the following: (1) excretion rates of PO:- by single Daphnia thorata were below detection (5 pmol animal- ’ min - ‘) in 20 ml of oligotrophic lake water over a period of 10 min, (2) experimental addition of D. thorata to 20 ml aliquots of lake water decreased community-wide microbial uptake of PO: - on two occasions (as measured by 32POj- incorporation), and (3) the presence of D. thoratu increased uptake by organisms smaller than 1 pm, and decreased uptake by large phytoplankton. The specific mechanism for these responses remains unclear, but the results imply that when phytoplankton larger than 1 pm encounter cm scale patches of water recently occupied by Daphnia they may experience decreased PO: - availability rather than elevated concentrations of PO:- caused by excretion. We show that 32P uptake experiments using natural plankton assemblages can be influenced by the presence or absence of large zooplankton, and that neither grazing, turbulence, nor PO: - excretion can account for this influence. Introduction
Phosphorus is often important in controlling primary production in freshwaters (Schindler, 1977; Hecky & Kilham, 1988). The low levels of dissolved inorganic nutrients observed in oligotrophic systems may not sustain uptake rates necessary to support observed levels of primary production without continuous nutrient regeneration by pelagic heterotrophic organisms (McCarthy & Goldman, 1979; Priscu & Priscu, 1984). Laboratory experiments have shown that phosphorus regenerated by zooplankton can be utilized by P deficient algae (Sterner, 1976). Zooplankton excretion in natural systems has been hypothesized to increase nutrient availability
to nearby phytoplankton by forming ephemeral patches of regenerated nutrient which may sustain primary production at higher levels than would be expected given the ambient bulk nutrient concentration (McCarthy & Goldman, 1979; Lehman & Scavia, 1982a, 1982b). The importance of such patches in nature has been disputed on theoretical grounds because the patches may disperse before phytoplankton can exploit them (Currie, 1984; Jackson, 1980). Laboratory experiments using quiescent uni-algal cultures have demonstrated that Daphnia can cause ephemeral patches of PO:- which stimulate uptake, and presumably growth of algae that encounter these patches, prompting speculation that regenerated PO: - patches are important to phytoplankton in waters with low PO: - concentrations (Lehman
254 & Scavia, 1982a, 1982b). Furthermore, experimentswith natural plankton assemblagesin semicontinuous culture have shown that short-term nutrient patchescan alter algal community structure (Scavia et al., 1984). However, it has not beenconclusivelyand directly demonstratedthat PO:- patchescausedby zooplankton excretion are important to the dynamics of natural algal communities. We measured short-term 32POz- uptake by plankton in the presence and absence of D. thovutuin 20 ml aliquots of phosphorus-deficient lake water to determine if 31POi- regeneratedby D. thoratu provided short-termbenefits to the plankton. To investigatethe responseof the microbial community to the presence of D. thorutu, chlorophyll, thymidine uptake, and bacterial counts were conducted on lake water before and after incubation with D. thoratu. The effect of organic C and NH,' on 32PO:- uptake was measuredto examine potential mechanisms for the relationship between32POz- uptake and D. thorutu excretion, and 32PO:- uptake in the presenceof stirring was used to mimic the effect of turbulence causedby swimming. Materials and methods Experiments were conducted on water samples collected at 5 m with a displacement sampler @odds & Priscu, 1988)from the deepestportion (cu. 2 km from shore) of Flathead Lake, on 6 August 1987,18 September 1987and 9 November 1987.Flathead Lake is a large(460km2) deep (X = 53 m), oligotrophic (0.1-1.0 pg l- * epilimnetic chl a) lakein westernMontana, USA. Carrier free 32[P]-H,P0, (333 Bq ml- i total activity) was added to 20 ml aliquots of unscreenedlake water in all 32Puptake experiments.The reaction was terminated after 4 min by filtration of a subsample from the aliquots which did not contain any D. thoruta onto Gelman GN-6 filters (effective retention0.45 pm). Blank filters placedunder the sample filter were used to correct for 32P absorption to the filters; formalin kills showed insignificant levels of abiotically labeled particu-
late matter. Filters and total 32Pin solution were counted separately by liquid scintillation spectrometry. Maximum possiblelevelsof biologically available PO:- (BAP) were measuredusing Rigler’s bioassay (Rigler, 1966).This bioassaymeasures the responseof algal 32PO:- uptake to known concentrations of 31PO:-. If it is assumedthat net PO:- uptake increases with increasing 31POz- concentrations,it is possibleto determine maximum possible ambient 3’POj- concentration using Rigler’stechnique.Unlabeled KH,PO, was used as a source of biologically available “PO: -. The effectof added 31POi- on 32POiuptake was determined on each of 4 replicates with “PO:- added to a nominal concentration of 0, 0.32, 1.0 and 2.0 nM immediately before addition of 32P03 4 - ’ To determine effectsof D. thorutu on 32POiuptake, a wide aperturepipettewas usedto gently isolate 3-20 ml aliquots of lake water, each containingoneD. thorutu, from lake water 30 min before 32PO:- was added in the August experiments, and 5 min before32PO:- was addedin the Septemberand November experiments.Aliquots with and without D. thorutu wereboth taken from the sampleof lake water. The shorterpre-incubation in the laterexperimentswas usedto minimize excretionby D. thorutu beforethe initiation of the experiment. 32PO:- uptake was measured as described above and rates were compared to 32PO:- uptake in three,20 ml control samplesof untreatedlake water without D. thorutu. Control aliquots were taken from the same lake water sample in the same manner as the D. thorutu aliquots,exceptthat the wide aperturepipette was used to sample water without D. thoratu. The effect of the addition of D. thorata on relative rates of 32PO:- uptake by the < 1 pm size fraction was determined after 4 min 32POjincubations. Three, 20 ml aliquots containing D. thorata and threewithout were incubatedwith 32POz- as above. Following incubation, 15ml (with no D. thorutu) were filtered first through a 1 pm pore size Nuclepore membrane filter, and then on a Gelman GN-6 (0.45pm) filter. All filters were countedby liquid scintillation spectrometry;
255 blank filters placed under the samplefilters were used to correct for label absorbedby filters. The relative size distribution of chlorophyllcontaining cells was determined by size fractionation The total amount of chl a retainedupon Whatman GF/F filters (effective retention 0.7 pm) from un-filtered lake water was comparedto the amountof chl a retainedon Whatman GF/F filters from lake water which was first passed through a 1 pm Nuclepore filter. Phaeophytin corrected chl a from the various fractions was measuredfluorometrically (Strickland & Parsons, 1972). Bacteria and algal picoplankton cell counts were made on 20 ml aliquots of lake water incubated for 45 min with D. thorutu and untreated samples of lake water. Samples were stained with 4’ 6-diamidino-2-phenylindole (DAPI) (Porter & Feig, 1980),concentratedon 0.2 pm Nuclepore filters, and counted. Bacterial counts represent total DAPI fluorescing cells minus algal cells. Changes in bacterial activity associatedwith D. thoruta addition were estimatedby measuring thymidine incorporation (Fuhrman & Azam, 1982).3[H]-thymidine was added to a final concentration of 20 PM thymidine (at a final activity of 0.27 kBq ml- ‘) to six, 10ml aliquots of lake water, 3 with D. thorutu. Incubation was stopped after 45 min by the addition of 10ml aqueous ice-cold trichloroacetic acid (TCA, 10% w/v). Samples were then filtered onto 0.45pm filters (Millipore HA), rinsed with 20 ml of ice cold 5% aqueousTCA acid and countedby liquid scintillation spectrometry. The potential effect of NH,’ regenerationon 32POi- uptake was determined by enriching samples with NH,f before P addition. On 9 November 1987,three, 50 ml aliquots were preincubatedfor 12 h with a nominal concentration of 10PM NH,Cl (ambient NH: cont. < 0.5 PM). This concentration of NH,’ saturated uptake(Dodds, Priscu & Ellis, unpublisheddata), and was used to estimatethe maximum possible effectof NH: excretionby D. thorutu. After 12h, 333 Bq ml- ’ of 32[PI-H,PO, and 2 PM KH,PO, were added to the pre-incubated samples and
threeother 50 ml aliquots of lake water. Samples were filtered (0.45pm Gelman GN-6) after 40 min and filters were counted as above. To determine if dead D. thoratu produced compounds which affected32POi- uptake, four D. thorutu were killed by entrapment in the surface film of lake water for 2 h and then addedto four, 20 ml aliquots of lake water. 32PO:- uptake was determined immediately on the aliquots as describedabove. 32PO:- uptake was measured in four, 20 ml aliquots, in the presenceof 0.1 ,uM glucose and in four stirred aliquots to measure effectsof dissolved organic C and turbulenceon 32POi- uptake. Means were comparedwith t tests. This test is appropriate for comparing two means when 12< 20. When means of percentageswere compared,arc-sinetransformationwas usedto obtain normal distribution (Steele& Torrie, 1960).When variances differed more than 2 fold, nonparametric (ranked) t tests were performed. One way analysisof variance (ANOVA) was usedto compare multiple means. Results and discussion Maximum possiblelevelsof biologically available P (BAP) in Flathead Lake were determinedto be 0.97nM, 0.65 nM, and 0.64 nM on 6 August, 18 Septemberand 9 November 1987,respectively. These values are within the ranges reported by Rigler (1966) for several Canadian lakes. The consistently low levels of BAP imply that the P requirements of the plankton community in Flathead Lake are met to a largeextent by internally regeneratedP. The relationship between31POj- uptake and 32POi- uptake was calculated as follows: 3’POj- uptake utime- ’ = 32POi- uptake+ *time-‘* [“Poj-]/[=Po:-] (1) Higher [“‘PO:-] can result in higher calculated ratesof 31POj- uptake perunit time, eventhough there is a decreasein 32POi- uptake per unit time. From equation (1) if [ 32POi- ] remains
256
0.0
0.5
total
1.0
1.5
2.0
PO4 3-(rlr?lol
2.5
c 3.0
1-l)
I L c-
I-
0.0
0.5
total
1.0
1.5
P043-(r7rnol
2.0
2.5
3.0
1-l )
Fig. 1. Uptake of ‘*PO:- (A) and total PO:- uptake (B) in Flathead Lake water versus total 3’POjconcentration (ambient + added) on 9 November 1987. Data were similar for the August and September sampling dates. Error bars = 1 standard deviation (n = 3). Biologically available P concentration was 0.64 nM, points at the lowest PO: - concentration represent uptake measured at ambient concentration (0.64 nM PO:- ).
per unit time (Fig. 1B). At PO:- concentrations well below saturation, 31POi- uptake has also been shown to increase as substrate concentration increases in other lakes (e.g. Tarapchak & Herche, 1986; Button, 1985; Rigler, 1966). BAP concentrations did not increase during the course of a 4 min incubation, even with D. thoruta present. The time-course of 32POiuptake in whole water showed no decrease in uptake rate in the first 4 min (Fig. 2). Any change in 31POi - concentration during the incubation would alter the ratio of [ 31POi - ] : [ 32POj - ] and cause a change in the measured uptake rate (Fig. 1A). If there was a net increase in 31POi concentration over the course of the incubation (i.e. high rates of 31POi- excretion by D. thorata), the rate of 32POi - incorporation would have decreased owing to isotopic dilution. Excretion of 31POi- by D. thorata was not measurable in the November experiment; there was not a difference between the time-course of 32POj - incorporation measured in control lake water and that in the presence of D. thorata (Fig. 2). A decrease in 32POi - incorporation resulting from isotope dilution by regenerated 3’POi - was expected in the presence of D. thorata. Such a decrease would be the outcome of lowered specific activity, resulting in an increase in net uptake of 3’POiby phytoplankton (see Figs. lA, B and equation 1). However, in August
iii
constant, 3’POi - uptake per unit time will increase if relative increase in [ 3’POj- ] is greater than the decrease in 32POi- uptake per unit time. This was the situation we found in Flathead Lake. Experimental addition of 31PO: ~ > 1 nM over ambient BAP caused a significant decrease in incorporation of 32POz- (P < 0.05, pooled t test, n = 6, Fig. 1A). This also occurred during August and September (data not shown). Decreased 32POz - uptake at higher 31P02 ~ concentration from a higher ratio of resulted [ 3 ‘PO: - ] : [ 32POi - 1. The decrease in 32POi uptake with increased 31POz- (Fig. 1A) occurred concurrently with an increase of 31POi - uptake
2OJ
-l-J ,5-. E : oz
0
o control l Daphnia
0
10-w
5--
x
0 01 0
8
0
. 3
6
Time
9
12
(min)
Fig. 2. Time-course of 32POjm uptake in Flathead Lake water on 9 November 1987 in the presence and absence of D. thorata.
257 and September, treatments containing D. thorutu showed greater incorporation of 32POi- indicating lower total 31POi- uptake (Fig. 3). These results show that 3’POj - regeneration by D. thorutu was not significant in August and September. Furthermore, the results imply that [ 3’POi- ] was lowered by D. thorutu, or that total 32POz - uptake per unit time was stimulated. Experimental addition of D. thorutu on 9 November, caused more 32POi - uptake in the < 1 pm size fraction than in the > 1 pm fraction. The c 1 pm fraction accounted for 42.2% of the total inorganic P uptake in whole lake water without D. thorutu (n = 3, standard deviation = 2.5%). After 5 min incubation with D. thorutu, 61.4% (n = 3, standard deviation = 10.7%) of the uptake was accounted for by the < 1 pm fraction, significantly more than without D. thorutu addition (P < 0.05, t test, 4 df, t = 2.44). D. thorutu caused an overall decrease in the proportion of 31POi - uptake by phytoplankton > 1 pm, because total community uptake in November was not changed by D. thorutu (Figs. 2, 3) and uptake by the < 1 nm fraction increased at the expense of uptake by the > 1 pm fraction. Most of the uptake in the
I .-c
:
7
61
T L Q e .+c 0 a
5
39
O
0
control
T
4 3 2 1
AUG
SEPT
NOV
Fig. 3. “PO:incorporation in Flathead lake water with and without D. thoruta for three sampling dates. Values are means from 3 replicates at 4 min, error bars = one standard deviation. ANOVA showed that D. thoruta had a significant effect on 32PO:- incorporation (F = 10.71, P < 0.01). There was a significant difference between untreated lake water (control) and water with D. thoratu added on 6 August and 18 September (P < 0.05, t test); the difference was not signiticant (P > 0.35) on 9 November.
< 1 pm fraction was presumably heterotrophic since more than 95 y0 of the total chlorophyll was retained on a 1 pm filter and only 5-10% of the living cells which passed through a 1 pm filter were algal. We performed a series of experiments to establish a mechanism behind the effect of D. thorutu on 32POi- uptake. Because an increase in bacterial biomass in the D. thorutu treatment (caused by excretion of bacteria) might have accounted for the increased 32POi- uptake by the < 1 pm fraction, bacterial counts and activity (thymidine uptake) were measured. Furthermore, the effect of dissolved organic C and NH: on PO:- incorporation was investigated because D. thorutu may excrete these compounds with subsequent stimulation of 32PO:- uptake. Finally, we explored for effects of D. thorutu on 32POi - uptake mediated by potential grazing, and turbulence caused by swimming. D. thorutu did not preferentially increase 32POi- uptake in the < 1 pm fraction by increasing bacterial biomass or thymadine uptake, because ‘[ HI-thymidine incorporation and bacterial counts were similar in untreated lake water (137 kBq 1-l hrrl, n = 3, std. dev. = 18.7; 7.50 x 10’ bacteria ml- ‘) and samples exposed toD. thorutufor45min(132kBql-‘hr-‘,n = 3, std. dev. = 28.0; 7.45 x lo5 bacteria ml- ‘). Thymidine uptake rates do not have to be coupled to increased PO:- uptake rates over short time periods. Therefore, although D. thorutu did not stimulate thymadine uptake, this does not mean that bacterial 32POi- uptake was not stimulated. Bacteria are capable of luxury consumption, and PO:- uptake does not necessarily translate to growth in short-term experiments. Glucose decreased 32POi- uptake as did dead D. thorutu (which may produce 31POi- or dissolved organic material) (Table 1). These results infer a stimulation of community uptake of total PO:- (see above discussion). These effects are the opposite of those seen with live D. thorutu addition. Samples exposed for 12 h to NH,‘, utilized PO: - at a rate of 67.6 nM PO: - h - ’ (std. dev. = 8.5, n = 4); the rate with no NH,’ was
258 Table 1. Rate of 32PO:- uptake on 9 November 1987 in water collected from Flathead Lake. In all treatments n = 4, except lake water where n = 3. Probability column represents the results oft tests comparing individual treatments to the whole lake water treatment. Treatment
KBq 1-l min-’
s.d.
Probability
whole lake water dead D. thorata 0.1 PM glucose stirred
0.541 0.312 0.356 0.373
0.139 0.023 0.042 0.023
p < 0.025 p < 0.025 p > 0.15
102 nM PO:h-l (std. dev. = 45, n = 4). Because these rates were not significantly different (P > 0.05, non-parametric two-tailed t test based on ranks) we conclude that NH: excretion by D. thorata did not affect uptake of PO: -. Turbulence and grazing caused by D. thorata were not responsible for the apparent decrease in total 32POj- uptake by phytoplankton. Because 32POi- uptake of stirred treatments (at 300 rpm) was not significantly different from unstirred lake water (Table l), turbulence caused by D. thorata can be disregarded. Grazing on bacteria or phytoplankton was not important in August, September or November because there was no decrease in the rate of 32POi- incorporation into particulate matter over 4 min with D, thorata present (Fig. 3). If D. thorata removed a significant portion of the microplankton during the incubation period, the rate of 32POi - uptake would have decreased. Furthermore, if the maximum clearance rate of each D. thoruta is assumed to be 7.0 ml d- ’ (Wetzel, 1986), only 0.73% of the volume would be filtered during a 30 min incubation in 20 ml. Therefore, we concluded that grazing of microplankton in our short-term experiments in August, September and November was negligible. Our data show that D. thorata can cause a short-term decrease in 31POi- availability to phytoplankton. The most likely cause of this decrease is that D. thorata increased 31PO: - and 32PO: - uptake by heterotrophic bacteria, lowering 31PO:availability to phytoplankton. Berman (1988) documented that a variety of dissolved organic phosphorus compounds can affect 32POi - uptake and that these effects are size-
fraction-specific. Organic phosphorus excretion by living D. thorata could also account for the effects we observed. Therefore in some situations, phytoplankton may experience patchy levels of BAP caused by D. thorata not as plumes of increased BAP near D. thorata, but as decreased BAP in regions recently inhabited by D. thoruta. The short-term incubations (4 min total) may have missed episodic 31POjexcretion by D. thorata. However, our data suggest that most of the time phytoplankton experience zooplankton effects which are the opposite of those expected to be caused by 31POi- excretion by zooplankton. Zooplankton-phytoplankton interactions appear to be more complex than previously reported, and effects such as we have reported should be considered when 32P uptake measurements are made on natural phytoplankton assemblages containing large zooplankton. Acknowledgements We would like to thank J. Grover, J. Lehman, G. Redfield, D. Scavia, R. Sterner and an anonymous reviewer for their helpful criticism. This research was supported by the Soap and Detergent Association. References Berman, T., 1988. Differential uptake of orthophosphate and organic phosphorus substrates by bacteria and algae in Lake Kinneret. J. Plankton. Res. 10: 1239-1249. Button, D. K., 1985. Kinetics of nutrient limited transport and microbial growth. Microb. Rev. 49: 270-297. Currie, D. J., 1984. Microscale nutrient patches: Do they matter to the phytoplankton? Limnol. Oceanogr. 29: 211-214. Dodds, W. K. & J. C. Priscu, 1988. An inexpensive device for sampling large volumes oflake water from discrete depths. Freshwat. Biol. 20: 113-l 15. Fuhrman, J. A. & F. Azam, 1982. Thymidine incorporation as a measure of heterotrophic bacterioplankton production in marine surface waters: evaluation and field results. Mar. Biol. (Berlin) 66: 109-120. Hecky, R. E. & P. Kilham, 1988. Nutrient limitations of phytoplankton in freshwater and marine environments: a review of recent evidence on the effects of enrichment. Limnol. Oceanogr. 33: 796-822.
259 Jackson, G. A., 1980. Phytoplankton growth and zooplankton grazing in oligotrophic oceans. Nature 284: 439-440. Lehman, J. T. & D. Scavia, 1982a. Microscale nutrient patches produced by zooplankton. Proc. Nat. Acad. Sci. USA 79: 5001-5005. Lehman, J. T. & D. Scavia, 198213.Microscale patchiness of nutrients in plankton communities. Science 216: 729-730. McCarthy, J. J. & J. C. Goldman, 1979. Nitrogenous nutrition of marine phytoplankton in nutrient-depleted waters. Science 203: 670-672. Porter, K. G. & Y. S. Feig, 1980. The use of DAPI for identifying and counting aquatic microflora. Limnol. Oceanogr. 25: 943-948. Priscu, J. C. & L. R. Priscu, 1984. Inorganic nitrogen uptake in oligotrophic Lake Taupo, New Zealand. Can. J. Fish. aquat. Sci. 41: 1436-1445. Rigler, F. H., 1966. Radiobiological analysis of inorganic phosphorus in lake water. Verh. Int. Ver. Limnol. 16: 465-470.
Scavia, D., G. L. Fahnenstiel, J. A. Davis & R. G. Kreis Jr., 1984. Small-scale nutrient patchiness: Some consequences and a new encounter mechanism. Limnol. Oceanogr. 29: 785-793. Schindler, D. W., 1977. Evolution ofphosphorus limitation in lakes. Science 21: 260-262. Steel, R. G. D. & H. Torrie, 1960. Principles and Procedures of Statistics with Special Reference to the Biological Sciences. McGraw-Hill Book Co. Inc., New York. Sterner, R. W., 1986. Herbivores’ direct and indirect effects on algal populations. Science 23 1: 605-607. Strickland & Parsons, 1972. A Practical Handbook of Seawater Analysis. Fish. Res. Bd Can. 167. Tarapchak, S. J. & L. R. Herche, 1986. Phosphate uptake by microorganisms in lake water: deviations from simple Michaelis-Menten kinetics. Can. J. Fish aquat. Sci. 43: 319-328. Wetzel, R. G., 1986. Limnology. Saunders, Philadelphia.
