Spatio-temporal variability in the GDH activity to ammonium excretion ratio in epipelagic marine zooplankton

Spatio-temporal variability in the GDH activity to ammonium excretion1 ratio in epipelagic marine zooplankton2 l. Fern´ andez-Urruzola∗,N. Osma,T.T. Packard,F. Maldonado,M. G´ omez3 Marine Ecophysiology Group (EOMAR), Universidad de Las Palmas de Gran Canaria, 35017, Spain.4 Abstract5 Glutamate dehydrogenase (GDH) activities have been widely used in oceanographic research as an index of in situ6 NH+ 4excretion rates (RNH+ 4) in zooplankton. Here we study the variability in the relationship between the enzymatic7 rates and the actual rates measured in epipelagic marine zooplankton between several marine ecosystems. Although8 both measures were significantly correlated across zooplankton assemblages, the regression models yielded different9 GDH/RNH+ 4ratios across ecosystems. Accordingly, the error of a general equation increased up to ±42.5 % when10 regressing all our data together. Aside from possible interspecific differences, some of the variability was explained11 by the unequal allometric relation that each rate maintained with protein. Scaling exponents were 1.38 for GDH12 activities and 0.87 for RNH+ 4, which would induce uncertainties in the GDH/RNH+ 4ratios when organisms with different13 sizes were considered. Nevertheless, the main factor causing divergence between GDH activities and RNH+ 4was14 the potential prey availability. We compared the excretory metabolism of the zooplankton community at different15 productivity periods in waters offGran Canaria, and observed an important decrease in the RNH+ 4during stratification.16 A similar decrease was found in the internal pool of glutamate, which may be critical in the regulation of in vivo rates.17 Strengthening our knowledge of the relationship between GDH activities and the RNH+ 4will lead to more meaningful18 predictions of phytoplankton regeneration and community nitrogen fluxes across large spatial scales.19 Keywords: Zooplankton, Glutamate dehydrogenase (GDH), Ammonium excretion, Allometry, Intracellular20 glutamate21 1. Introduction22 Nitrogen is one of the most limiting nutrient elements controlling phytoplankton growth throughout the world’s23 oceans. The dissolved inorganic nitrogen availability may come either from remineralization processes in the sunlit24 layer or from introduction of new nutrients via upwelling, dinitrogen fixation and terrestrial run-off. Among all the25 inorganic nitrogen species, the recycling of the reduced form of ammonium (NH+ 4) satisfies a global mean of about26 ∗Corresponding author. Tel.: +34 928 45 44 73; fax: +34 928 45 29 22 Email address: [email protected] (l. Fern´ andez-Urruzola∗) 80 % of the primary production requirements (Harrison, 1992). It is therefore an outstanding source of nitrogen to be27 considered when assessing nutrient fluxes in any aquatic ecosystem.28 The regeneration of NH+ 4is mainly the result of both bacterial remineralization of dissolved organic matter and ex-29 cretion processes in zooplankton (Bronk and Steinberg, 2008). Here we focus on this latter component of the nitrogen30 cycle. The importance of NH+ 4excretion by zooplankton is closely related to the trophic character of the ecosystem31 and it is, in general terms, more important in oligotrophic than in eutrophic waters. Accordingly, mesozooplankton32 NH+ 4excretion has been found to be responsible from about 90 % of the primary production in oligotrophic gyres33 (Isla et al., 2004) to a low of 5 % in upwelling environments (Bode et al., 2004; Fern´ andez-Urruzola et al., 2014).34 However, the NH+ 4excretion rates (RNH+ 4) are also affected by the temperature, taxa, body size and nutritional level35 (Steinberg and Saba, 2008), so their potential contribution to the marine biogeochemical cycles varies widely in both36 time and space, highlighting the need for monitoring the zooplankton physiology in order to understand this vari-37 ability. Unfortunately, measuring RNH+ 4on live zooplankton is, not only burdened by unavoidable uncertainties, but38 so time consuming that rarely enough incubations can be made to obtain a high-resolution spatial coverage of RNH+ 4.39 This becomes even more complicated if different size fractions of zooplankton are to be studied. Aside from the effort40 investment, in vitro measurements of zooplankton metabolism are subjected to several sources of error. Factors such41 as crowding, stress caused during manipulation, and starvation in the ongoing experiments would promote a rapid fall42 in the RNH+ 4(Bidigare, 1983; Ikeda et al., 2000). Conversely, organisms injured either during collection or handling are43 prone to release more nutrients than do the healthy specimens (Ikeda et al., 1982). All these effects result in excretion44 rates that, to some extent, might be different from normal RNH+ 4in seawater.45 As part of the biochemical machinery, enzymes catalyze the synthesis of many metabolic end-products and there-46 fore, they have been extensively used in oceanography to infer rates of particular physiological processes such as, for47 example, respiration (Packard et al., 1971), nitrate uptake (Eppley, 1978) or NH+ 4excretion (Bidigare and King, 1981).48 Enzymatic assays constitute a relatively straightforward way to study the plankton metabolism that circumvents all the49 methodological constraints associated with bottle incubations. Moreover, enzyme activities can be measured quickly,50 either on-board or at a later time, as long as the biological samples were properly stored. This confers on the enzy-51 matic assays an advantage over the more direct incubation techniques. Prompted by these arguments, Bidigare and52 King (1981) introduced the analysis of the glutamate dehydrogenase (GDH) activity as a proxy for RNH+ 4in zooplank-53 ton. Since then, the GDH assay has been used to obtain a detailed RNH+ 4distribution, both depth resolved and across54 ocean regions, at sampling rates that otherwise would not have been attainable (e.g., Bidigare et al., 1982; King et al.,55 1987; Fern´ andez-Urruzola et al., 2014). More recently, Fern´ andez-Urruzola et al. (unpublished data) modeled down-56 ward nitrogen fluxes from GDH measurements through the water column in the northern Benguela. But enzymatic57 assays, such as the one for GDH, are not exempt from methodological biases. They are measurements that reflect58 the maximum rate at which the reaction may occur, not the actual one, so they have to be converted into in situ rates59 through an empirical factor. However, the relationship between enzymatic and in vivo rates is not universal, but may60 be affected by the ambiental conditions. In fact, Bamstedt (1980) demonstrated that enzymes respond to the environ-61 2 mental changes with a certain delay as compared to the physiological response. This was subsequently corroborated62 for different metabolic pathways when varying food availability in cultures of marine mysids (Herrera et al., 2011;63 Fern´ andez-Urruzola et al., 2011).64 During a year (2011), we conducted on-board incubations of mixed epipelagic zooplankton throughout different65 marine systems: North Atlantic, Benguela Upwelling and Indian Ocean. Here we present both the RNH+ 4and GDH66 activities measured at each province, and provide the most complete GDH/RNH+ 4data set published to date for ma-67 rine zooplankton. We compare our ratios with those found in the literature either for natural mixed zooplankton or68 cultured organisms, and discuss the use of a generalized GDH/RNH+ 4ratio for routinely assessing in vivo RNH+ 4from69 GDH measurements. Furthermore, oceanic mixing events may induce trophic shifts that would locally impact the70 relationship between GDH activity and RNH+ 4in the resident zooplankton community. For this reason, we chose a71 station offGran Canaria (28◦N) to evaluate the magnitude of the seasonal changes in the biomass-specific rates from72 the same location in comparison to variability found between different oceanic systems. In this light, we measured73 the intracellular levels of the main substrate for the GDH reaction, i.e., glutamate, to explore the correlation between74 physiological rates and substrate concentration. If the latter controls the former, then both should follow the same75 trend in response to the environmental changes.76 2. Material and methods77 2.1. Study regions78 The spatial variability of both RNH+ 4and GDH activities in zooplankton were analyzed from five cruises that were79 carried out during 2011. These cruises surveyed tropical and temperate waters of the Indian Ocean (IO), North Atlantic80 (NA), Canary islands (CI), and Benguela upwelling (BU) system (Fig. 1). The temporal variability in the zooplankton81 metabolism was also assessed by sampling the same station offTaliarte, Gran Canaria Island (28◦00’03” N, 15◦19’30”82 W) during the so-called “late winter bloom” (CI-LWB, characterized by the nutrient-enrichment of surface waters83 through mixing processes), and during the period of maximum stratification (CI-ST, with higher temperatures and a84 lower nutrient load in the sunlit layer that is expected to limit the phytoplankton growth). In all the cases we followed85 the same experimental procedure in order to minimize any bias associated with the methodology. Zooplankton were86 collected by vertical tows with a UNESCO WP-2 net (60 cm diameter ring, fitted with either 100 µm or 200 µm87 mesh sizes depending on the cruise) from 200 m to the surface (i.e., the epipelagic zone). Additionally, a Hansen-88 Egg plankton net with a mouth opening of 20 cm, and fitted with a 50 µm mesh size, was used during the CI-LWB89 cruise to extend our study into the 50 - 100 µm size fraction. The hauling speed was always about 0.2 - 0.3 m s−1, as90 recommended for physiological studies of live zooplankton (Sameoto et al., 2000). Once on deck, organisms were91 carefully fractionated into 50 - 100 µm, 100 - 200 µm, 200 - 500 µm, 500 - 1000 µm and >1000 µm size categories.92 This fractionation varied according to the mesh size of the sampling-net, and how much zooplankton were gathered93 in the net. Each size fraction was then transferred by siphoning into 2-L bottles filled with GF/F filtered seawater,94 3 Figure 1: Stations sampled during the CAMVALEX (F), SUCCESSION (N), and MALASPINA-2010 –legs 3/4 () and leg 7 ( )– research cruises. All samples were taken from Feb-2011 to Oct-2011. To compare seasonal differences in the zooplankton NH+ 4excretory metabolism, the Camvalex cruise was conducted twice offTaliarte (Canary islands): during the so-called late winter bloom (Apr-2011), and during the period of maximum stratification (Oct-2011). and maintained at in situ temperature. During the acclimation period, the bottles were gently aerated with an air95 pump, taking care not to damage the organisms with the bubbles. The zooplankton were thus acclimated for about an96 hour before being used in the NH+ 4excretion experiments in order to reduce the stress incurred during the course of97 sampling.98 2.2. Chlorophyll-a determinations99 Chlorophyll-a (Chl-a) was measured for each oceanic system as an estimator of phytoplankton biomass. Seawater100 was filtered through GF/F and, in some cases, stored at −20 ◦C for subsequent analyses. Pigments were extracted101 in acetone, and their concentration was determined according to two different methods depending on the cruise. The102 Chl-a samples from the NA and the IO were measured onboard using their fluorescence properties as described by103 Yentsch and Menzel (1963), while the Chl-a concentration in the CI-LWB, CI-ST and BU was spectrophotometrically104 analyzed in the land-based laboratory following the protocol of Parsons et al. (1984).105 2.3. Bottle incubations106 After acclimation, the most healthy and actively swimming zooplankters were washed in GF/F filtered seawater107 and siphoned into 60 mL gas-tight glass bottles. Each experimental batch included, at least, one control flask without108 organisms. We found little effect of container size on the oxygen consumption rates when varying the experimental109 bottles from 30 mL to 160 mL (Ikeda et al., 2000), so we chose a volume in the lower range in order to reduce110 the incubation time. Thus, we obtained a significant signal of NH+ 4release in less than 1.5 h. This achieved a111 4 compromise between those effects that density and starvation may induce on the physiological rates, and which are112 fairly constant over 1.5 h of incubation. Shortly before the incubation began, we took three replicates (10 mL each113 one) of filtered seawater to determine the dissolved NH+ 4concentrations (µM) at the starting point. Then, we incubated114 the organisms at in situ temperature (after averaging the temperatures for the upper 200 m) and in the dark for 1 - 1.5115 h, depending on the density of the experimental population. Darkness was meant to prevent any autotrophic activity116 that could utilize the available dissolved NH+ 4. Afterwards, 10 mL of seawater were siphoned offfrom each bottle117 for NH+ 4determinations. Dissolved NH+ 4was spectrofluorometrically measured according to the Holmes et al. (1999)118 method, except in the “SUCCESSION” cruise where it was determined through the phenol-hypochlorite method119 (Solorzano, 1969) due to the inability to measure fluorescence on board. We used a standard curve from 0.04 to120 10.24 µM to calibrate both the fluorescence and absorbance measurements. For the calculations of NH+ 4excretion121 rates, we subtracted the NH+ 4concentration quantified in the control flasks from those concentrations measured in the122 experimental flasks.123 2.4. Enzymatic measurements124 Once the seawater was sampled for NH+ 4excretion analyses, the zooplankters were immediately frozen in liquid125 nitrogen (−196 ◦C) and stored at −80 ◦C until enzyme analyses in the land-based laboratory. Organisms were then126 thawed, and sonicated for 45 s in 0.1 M Tris-buffer medium, made up to pH 8.6 with acetic acid. The resulting ho-127 mogenate was centrifuged for 8 min at 4000 rpm. The whole process prior to the enzymatic assay never exceeded128 20 min, with the samples being kept at 0 ◦C at all times. The supernatant was then assayed for glutamate dehydro-129 genase (GDH) activity following the method published in Bidigare and King (1981), slightly modified by applying130 the principles of fluorometry as explained in Fern´ andez-Urruzola et al. (2011) to detect the NADH production rate131 in the reaction. To ensure that the maximum velocity (Vmax) of the reaction was reached, i.e., the potential enzy-132 matic rate, we saturated the enzyme with 50 mM glutamate and 1.2 mM nicotinamide adenine dinucleotide (NAD+).133 Furthermore, 2 mM adenosine-5’-diphosphate (ADP) was added to favor the glutamate deamination that could be134 inhibited to some degree by guanosine-5’-triphosphate (GTP) molecules present in the homogenate. In addition, for135 those samples collected offthe Canary Islands, an aliquot of the supernatant fluid was simultaneously assayed for136 electron transport system (ETS) activity according to the Owens and King’s (1975) protocol. This allowed us to eval-137 uate seasonal changes in the RO2/RNH+ 4ratios. The two enzyme reactions were kinetically measured for 4 min at the138 same temperature used in the incubation experiments, so no temperature correction (Arrhenius equation) was needed.139 2.5. Intracellular concentration of glutamate140 We further studied the intracellular levels of the main substrate of the reaction (glutamate) in order to relate any141 temporal variation in the NH+ 4excretion rates at the CI station with biochemical adjustments of the GDH. Among other142 factors, the concentration of available glutamate will be critical to determine the rate at which the GDH can operate.143 Accordingly, we analyzed the intracellular concentration of free glutamate by applying the method of Beutler and144 5 Michal (1974), which uses diaphorase, tetrazolium salts and pure GDH from bovine liver (EC 1.4.1.3) to determine145 the glutamate concentration in the sample. This method overcomes the equilibrium of the GDH reaction by the146 continuous reoxidation of the NADH formed from the glutamate deamination (Eq. 1), through coupling with a second147 enzyme reaction catalyzed by diaphorase (Eq. 2):148 Glutamate +NAD++H2OGDH −−−* )−−− α−Ketoglutarate +NADH +NH+ 4(1) 149 NADH +INT +H+Diaphorase −−−−−−−−* )−−−−−−−− NAD++Formazan (2) Since the reaction proceeds stoichiometrically, we derive the intracellular glutamate concentration from quantify-150 ing the total formazan production, whose extinction coefficient is measured at 492 nm.151 2.6. Biomass determination152 Biomass was estimated as protein content using the Lowry method (Lowry et al., 1951) modified by Rutter (1967).153 Bovine serum albumine (BSA) was used as a standard.154 2.7. Statistics155 Statistical analyses were performed using SPSS for Macintosh (v 22, Inc., Chicago, USA). The normal distribution156 of data and the variance homogeneity were confirmed by the Shapiro-Wilk and the Levene’s tests, respectively. An157 ANCOVA test was applied to check for significant differences between the RNH+ 4-protein and the GDH activities-158 protein slopes. Differences in the RNH+ 4and GDH activities between locations and size categories were determined159 by one-way ANOVA tests. When necessary, Box-Cox analyses were applied to find the best transformations of the160 protein-specific data in order to achieve normality and homoscedasticity. All the regression equations and confidence161 intervals (CIs) were calculated using Sigmaplot (v 12.5, Systat Software Inc., California, USA).162 3. Results163 3.1. Characteristics of the study sites164 Sampling dates and oceanographic properties of the different provinces studied during 2011 are presented in Table165 1, along with the number of experiments conducted in each region. Mean sea surface temperature (SST) ranged166 from a low of 14.6 ◦C in the BU to a high of 25.1 ◦C in the NA. The opposite trend was observed in the averaged167 chlorophyll-a values, with their maximum in the BU (3.18 mg m−3) and the minimum in the NA (0.08 mg m−3). Both168 variables reflected the features typical of upwelling and oligotrophic environments, respectively. There were fewer,169 but still noticeable, seasonal differences in the hydrographical properties in the Canary Islands waters; during the late170 winter bloom period (CI-LWB) the waters were colder and had more phytoplankton biomass than during October171 (Table 1).172 6 Table 1: Cruise name and regions sampled during 2011 for NH+ 4excretion and GDH analyses in zooplankton. SST and SSS stand for the sea surface temperature and salinity, respectively. The range min - max (mean) is given for each physical or biological variable. The last column (exp. number) indicates the number of incubations performed at each cruise. Cruise Region Study season SST SSS Chl-a Exp. (in 2011) (◦C) (PSU) (mg m−3) number Malaspina 2010 Indian Ocean Feb. - Mar. 16.5 - 25.9 (21.5) 34.8 - 36.0 (35.5) 0.04 - 0.52 (0.18) 23 North Atlantic Jun. - Jul. 21.1 - 28.8 (25.1) 34.5 - 35.4 (34.9) 0.04 - 0.27 (0.08) 57 Camvalex Canary Islands Apr. 18.2 - 20.8 (19.3) 36.6 - 38.8 (36.7) 0.33 - 0.36 (0.34) 83 Oct. 20.6 - 23.3 (22.1) 36.8 - 36.9 (36.9) 0.22 - 0.26 (0.24) 52 Succession Northern Benguela Aug. - Sep. 12.8 - 16.2 (14.6) 34.4 - 35.8 (34.8) 0.75 - 14.34 (3.18) 32 3.2. NH+ 4excretory metabolism of zooplankton173 Fig. 2 shows the relationship between protein content in the sample and both RNH+ 4and GDH activities from174 the different marine systems surveyed, disregarding the potential effect of in situ temperature. Both variables were175 significantly correlated with the biomass (p<0.0001), even though the variance in the GDH activities that was176 explained by the protein content (62 %) was twice that for RNH+ 4(29 %). On the other hand, the slopes of the regression177 analyses were significantly different from each other (ANCOVA test, F1,243 =16.39, p<0.01), which would cause178 variability in the GDH/RNH+ 4ratio with biomass.179 RNH+ 4and GDH activities were then standardized by protein for comparison between areas and size fractions (Fig.180 3). It is noteworthy that no large zooplankton (>1000 µm) were captured in the net during CI-LWB. This was not the181 case in BU, where only the zooplankton between 500 - 1000 µm were considered due to methodological problems in182 the other size categories (since they were either contaminated with diatom chains in the case of the 100 - 500 µm size183 fraction, or below the detection limit of the method in the case of the >1000 µm size fraction). Biomass specific-RNH+ 4 184 indicated some allometry as they were, in general, higher in the smaller size fraction (Table 2). Considering the study185 area, the most significant differences in RNH+ 4were found at CI-ST and BU, where the NH+ 4release per unit of protein186 showed the lowest rates (Table 2). This variability was attributed mainly to the smaller size fractions, since the RNH+ 4 187 in the largest zooplankton (>1000 µm) was relatively invariant between the different regions (ANOVA test, F2,28 =188 1.55, p=0.231). As expected for potential measurements, the protein specific-GDH activities were always higher189 than their correspondent RNH+ 4(Fig. 3b). GDH activities depicted, however, a different pattern than those observed190 for RNH+ 4. In fact, the differences with size fraction followed the opposite trend, with the GDH activities higher in the191 largest organisms (Table 2). The variability in the GDH between regions was not so marked although, paradoxically,192 CI-LWB presented the lowest GDH activities. Nevertheless, considering the zooplankton between 100 - 1000 µm, the193 GDH activities between CI-LWB and CI-ST were comparable (Student t-test, p>0.05).194 The relationships between GDH activities and RNH+ 4at each location and size fraction, expressed as µmol NH+ 4 195 mg protein−1h−1, are presented in Table 3. Both variables were linearly related in all cases, so no transformations196 7 Figure 2: Log-scale scatterplot showing the relationship between protein content and NH+ 4excretion rates (a), and between protein content and GDH activities (b). Each data point represents different size fractions of mixed zooplankton, incubated at in situ temperature (ranging from 12.8 to 28.8◦C). The least-square linear regressions were: log RNH+ 4=0.87 log protein −0.82 (r2=0.29, n=243, p<0.0001) for NH+ 4excretion rates, and log GDH =1.38 log protein +0.31 (r2=0.62, n=247, p<0.0001) for GDH activities. Dashed lines stand for the 95 % CIs. were applied to the data. Furthermore, each data set was normally distributed (Shapiro-Wilk test, p>0.05), and their197 variance was constant across observations (Levene’s test, p>0.05). This allowed us to extract meaningful regression198 statistics and compare our slopes, which define the GDH to RNH+ 4ratio, with other published GDH/RNH+ 4means. These199 slopes were similar between the NA, IO and CI-LWB, ranging from 1.7 (NA) to 2.3 (IO) for the whole community.200 Furthermore, the ratio measured at the CI-LWB compared well with those reported in the literature for the same201 area and season (Fern´ andez-Urruzola et al., 2011; Hern´ andez-Le´ on and Torres, 1997). However, the GDH to RNH+ 4 202 relationship increased dramatically up to 6-fold during the stratification period, at the CI-ST (=13.27, p<0.0001).203 Zooplankters from other ecosystems were characterized by a higher GDH/RNH+ 4ratio, with a maximum of 43.8 in204 the marine mysid Praunus flexuosus (Bidigare and King, 1981). In general, the error of estimates (SEE) was lower205 in the monospecific experiments than in those samples of mixed zooplankton. Seeking a common relationship for206 all the study areas, we pooled all our experimental data in Fig. 4. In this case, both rates (in µmol NH+ 4sample−1 207 8 h−1units) were logarithmically transformed to reduce heteroscedasticity of the residuals, and we found the following208 relationship:209 log GDH =0.64 log RNH+ 4+0.36 (r2=0.37,n=235,p<0.0001,S EE =±42.6 %) (3) The values from the five cruises were distributed uniformly along the regression line, but all together generated a210 higher dispersion as compared to the one observed for each individual cruise. Accordingly, the standard error of the211 estimate in Eq. 3 was twice the errors found when regressing each cruise separately. Still, the linear model for the212 whole data set was significant at p<0.0001. Considering a multivariate regression in the form of213 log GDH =−2.25 +0.72 log RNH+ 4+0.12 T+0.42 Chl-a (r2=0.59,n=235,p<0.0001,S EE =±34.5 %) (4) which includes other factors such as in situ temperature (T) and chlorophyll-a (Chl-a), we improved the prediction214 of GDH activities to 59 %. Similarly, the error associated with Eq. 4 decreased by 8.1 % with respect to the simple215 regression model.216 Figure 3: Boxplot showing the biomass-specific NH+ 4excretion rates (a), and the biomass-specific GDH activities (b) in three size categories of zooplankton throughout different marine ecosystems. The lower and upper boundaries of the boxes represent the first and third quartiles of the data distribution, respectively, with the middle line indicating the median. Error bars indicate the 95 % CIs. 9 the Canary Island region, the erosion of the thermocline allows the entrainment of nutrients into the euphotic zone,330 leading to increased primary productivity (De Le´ on and Braun, 1973). This was reflected in the higher chlorophyll-a331 concentration during CI-LWB as compared to the stratification period, CI-ST (Table 1). At that time, zooplankton332 released an average of 0.39 µmol NH+ 4mg protein−1h−1, twice the protein-specific RNH+ 4found during CI-ST (Fig. 5).333 Hern´ andez-Le´ on and Torres (1997) monitored the mesozooplankton RNH+ 4from November to May offGran Canaria334 island, and also found great variability in the rates according to the trophic fluctuations (ranging between 0.02 - 0.71335 µmol NH+ 4mg protein−1h−1). Similar to our findings, their GDH activities did not follow the RNH+ 4pattern, which336 led to different GDH/RNH+ 4ratios during their study period. In fact, several studies have found that the specific GDH337 activities did not peak in the chlorophyll-a maximum, but rather it was attenuated (Fern´ andez-Urruzola et al., 2014;338 Hern´ andez-Le´ on et al., 2001; Park et al., 1986). On the one hand, it seems reasonable to presume that the plankton339 community was growing during the CI-LWB and therefore, it was in an earlier developmental stage (sensu Vinogradov340 and Shushkina, 1978) as compared to the community from the CI-ST. Nitrogen may thus limit biosynthesis, so it341 would not be energetically efficient to produce an excess of enzyme. Another plausible biochemical explanation342 was given by Park et al. (1986), who suggested a strong inhibition by the high GTP concentration generated via343 the tricarboxilic acid cycle in those organisms that were actively growing under favourable trophic conditions. In344 such a situation, GDH activities may be underestimated by the standard assay, since it would require higher levels345 of ADP to counteract the GTP effect. This could explain the lower GDH/RNH+ 4ratios during conditions of high prey346 abundance, as well as differences in the values of the y-intercepts observed in Table 3. Similar behavior in the ratio347 was observed when zooplankters were exposed to starvation in laboratory experiments (Fern´ andez-Urruzola et al.,348 2011; Park, 1986a). In addition to the allosterism associated with GTP, GDH is known to be controlled by the internal349 pool of glutamate. As in all enzymes, the substrate concentrations determine the actual rate at which the reaction350 can operate (Bisswanger, 2008); however, few attempts have been made to measure them directly. Park et al. (1986)351 calculated the effective glutamate concentration from kinetic parameters in macrozooplankton, and showed an increase352 of the glutamate pool linked with those periods of food availability. Similar findings have been made regarding the353 respiratory metabolism, for exampe Osma et al. (2016) measured a decrease in the levels of pyridine nucleotides in the354 marine dinoflagellate Oxyrrhis marina as organisms starved. In our study, internal glutamate decreased dramatically355 from April to October, which supports the hypothesis of substrate levels as a key mechanism in the regulation of RNH+ 4 356 (Hern´ andez-Le´ on and Torres, 1997). Accordingly, the higher glutamate availability may lead the organisms to excrete357 more NH+ 4per unit protein during the late winter bloom. The significant correlation between these two variables (Fig.358 6) reinforces the utility of kinetic-based models in the study of zooplankton metabolism (Packard and G´ omez, 2008).359 The measurement of biochemical parameters such as the Michaelis constant (Km), jointly with the intracellular levels360 of both the substrates and allosteric regulators, would therefore open new avenues in the approximation of the in vivo361 RNH+ 4from GDH activities. Furthermore, we studied the relationship between the respiration rates (RO2) and the RNH+ 4 362 because it serves as an index of catabolism (Mayzaud and Conover, 1988). Although it was slightly higher during363 CI-ST, the low values reflected a protein-based catabolism during the two sampling periods. This is not surprising364 16 since the small microheterotrophs, poor in fatty acids, constitute 35 - 80 % of the diet of mesozooplankton in these365 waters (Hern´ andez-Le´ on et al., 2004). So rather than a shift in the diet, changes in the availability of prey seem to be366 responsible for the variability measured in the zooplankton excretory metabolism.367 5. Conclusions368 GDH is an essential tool for mapping zooplankton RNH+ 4throughout the oceans. Unfortunately, the statistical re-369 lationship as measured by GDH/RNH+ 4, the ratio between enzymatic and physiological rates, is not universal. In this370 research we found that temporal variability in the GDH/RNH+ 4ratios from the same ecosystem could be higher than371 those between regions. Both GDH activities and RNH+ 4maintained differently allometric relationships with biomass,372 which has to be considered when comparing communities with different sized animals. Still, this effect should be373 studied on specific taxa and controlled culture conditions in order to avoid any interference from other sources of374 variability. On the other hand, abundance of prey is known to be a key factor in modulating the metabolic rates of375 zooplankton. Here we observed fluctuations in the internal glutamate pool according to the productivity regime, in376 parallel to the RNH+ 4trends. How this variation affects the actual enzymatic rates needs to be further investigated.377 Given the variability in the GDH activity to RNH+ 4relationship, we encourage a field calibration of this ratio for each378 specific community being studied.379 380 Acknowledgements. We wish to thank the crews of the BIO Hesp´ erides, BIO Atlantic Explorer, and RV Maria S. Merian for their expertise and381 enthusiastic support. We thank L. Postel and C. M. Duarte for their invitation to participate in the “SUCCESSION” (I. F.-U.) and “MALASPINA382 2010” (F. M., N. O., I. F.-U.) cruises. We are also grateful to M. Estrada and P. Mozetic for providing the chlorophyll data from the MALASPINA383 2010 cruise. Two anonymous reviewers notably improved the manuscript with their valuables suggestions. Funding was provided, in part, by384 the German Research Foundation (DFG), and by the MALASPINA 2010 (CSD2008-00077) and the BIOMBA (CTM2012-32729/MAR) projects385 granted to C. M. Duarte and M. G., respectively. I.F.-U. and N.O. were supported by postgraduate grants from the Formation and Perfection of the386 Researcher Personel Program from the Basque Government. T.T. P. was largely supported by TIAA-CREF and Social Security (USA).387 388 References389 Ar´ ıstegui, J., Montero, M.F., 1995. The relationship between community respiration and ETS activity in the ocean. J. Plankton Res. 17 (7),390 1563–1571.391 Bamstedt, U., 1980. ETS activity as an estimator of respiratory rate of zooplankton populations. The significance of variations in environmental392 factors. J. Exp. Mar. Biol. Ecol. 42 (3), 267–283.393 Berges, J.A., Roff, J.C., Ballantyne, J.S., 1993. Enzymatic indices of respiration and ammonia excretion: relationships to body size and food levels.394 J. Plankton Res. 15 (2), 239–254.395 Bidigare, R.R., 1983. Nitrogen excretion by marine zooplankton. In: E.J. Carpenter, D.G. Capone (Eds.), Nitrogen in the marine environment.396 Academic Press, Inc., New York, pp. 385–409.397 Bidigare, R.R., King, F.D., 1981. The measurement of glutamate dehydrogenase activity in Praunus flexuosus and its role in the regulation of398 ammonium excretion. Comp. Biochem. Physiol. 70 (B), 409–413.399 17 Bidigare, R.R., King, F.D., Biggs, D.C., 1982. Glutamate dehydrogenase (GDH) and respiratory electron-transport-system (ETS) activities in Gulf400 of M´ exico zooplankton. J. Plankton Res. 4 (4), 895–911.401 Bisswanger, H., 2008. Enzyme kinetics: Principles and methods. 2nd ed. Wiley-VCH Verlag GmbH & Co., Weinheim, 301 pp.402 Bode, A., Barquero, S., Gonz´ alez, N., ´ Alvarez-Ossorio, M.T., Varela, M., 2004. Contribution of heterotrophic plankton to nitrogen regeneration in403 the upwelling ecosystem of A Coru˜ na (NW Spain). J. Plankton Res. 26 (1), 11–28.404 Bronk, D.A., Steinberg, D.K., 2008. Nitrogen regeneration. In: E.J. Carpenter, D.G. Capone (Eds.), Nitrogen in the marine environment. Academic405 Press, London, pp. 385–467.406 Campbell, R.W., Boutillier, P., Dower, J.F., 2004. Ecophysiology of overwintering in the copepod Neocalanus plumchrus: changes in lipid and407 protein contents over a seasonal cycle. Mar. Ecol. Prog. Ser. 280, 211–226.408 De Le´ on, A.R., Braun, J.G., 1973. Annual cycle of primary production and its relation to nutrients in the Canary Islands waters. Bol. Inst. Esp.409 Oceanogr. 167, 1–24.410 Eppley, R.W., 1978. Nitrate reductase in marine phytoplankton. In: J.A. Hellebust, J.S. Craige (Eds.), Handbook of phycological methods.411 Physiological and biochemical methods. Cambridge University Press, Cambridge, pp. 217–223.412 Fern´ andez-Urruzola, I., Osma, N., G´ omez, M., Montesdeoca-Esponda, S., Packard, T.T., 2016. Building a model of ammonium excretion in two413 species of marine zooplankton based on glutamate dehydrogenase kinetics. Mar. Ecol. Prog. Ser. 550, 83–99.414 Fern´ andez-Urruzola, I., Osma, N., Packard, T.T., G´ omez, M., Postel, L., 2014. Distribution of zooplankton biomass and potential metabolic415 activities across the northern Benguela upwelling system. J. Mar. Syst. 140 (B), 138–149.416 Fern´ andez-Urruzola, I., Packard, T.T., G´ omez, M., 2011. GDH activity and ammonium excretion in the marine mysid, Leptomysis lingvura: Effects417 of age and starvation. J. Exp. Mar. Biol. Ecol. 409 (1-2), 21–29.418 Hans-Otto, B., Michal, G., 1974. L-glutamate: Determination with glutamate dehydrogenase, diaphorase, and Tetrazolium salts. In: H.U.419 Bergmeyer (Ed.), Methods of enzymatic analysis. Verlag Chemie GmbH, pp. 1708–1713.420 Harrison, W.G., 1992. Regeneration of nutrients. In: P.G. Falkowski, A.D. Woodhead (Eds.), Primary productivity and biogeochemical cycles in421 the sea. Plenum Press, New York, pp. 385–409.422 Hern´ andez-Le´ on, S., Almeida, C., B´ ecogn´ ee, P., Yebra, L., Ar´ ıstegui, J., 2004. Zooplankton biomass and indices of grazing and metabolism during423 a late winter bloom in subtropical waters. Mar. Biol. 145 (6), 1191–1200.424 Hern´ andez-Le´ on, S., Almeida, C., G´ omez, M., Torres, S., Montero, I., Portillo-Hahnefeld, A., 2001. Zooplankton biomass and indices of feeding425 and metabolism in island-generated eddies around Gran Canaria. J. Mar. Syst. 30 (1-2), 51–66.426 Hern´ andez-Le´ on, S., Fraga, C., Ikeda, T., 2008. A global estimation of mesozooplankton ammonium excretion in the open ocean. J. Plankton Res.427 30 (5), 577–585.428 Hern´ andez-Le´ on, S., Postel, L., Ar´ ıstegui, J., G´ omez, M., Montero, M.F., Torres, S., Almeida, C., K¨ uhner, E., Brenning, U., Hagen, E., 1999.429 Large-scale and mesoscale distribution of plankton biomass and metabolic activity in the Northeastern Central Atlantic. J. Oceanogr. 55,430 471–482.431 Hern´ andez-Le´ on, S., Torres, S., 1997. The relationship between ammonia excretion and GDH activity in marine zooplankton. J. Plankton Res.432 19 (5), 587–601.433 Herrera, A., Packard, T., Santana, A., G´ omez, M., 2011. Effect of starvation and feeding on respiratory metabolism in Leptomysis lingvura (G.O.434 Sars, 1866). J. Exp. Mar. Biol. Ecol. 409, 154–159.435 Holmes, R.M., Aminot, A., K´ erouel, R., Hooker, B.A., Peterson, B.J., 1999. A simple and precise method for measuring ammonium in marine and436 freshwater ecosystems. Can. J. Fish. Aquat. Sci. 56, 1801–1808.437 Ikeda, T., 1985. Metabolic rates of epipelagic marine zooplankton as a function of body mass and temperature. Mar. Biol. 85, 1–11.438 Ikeda, T., Fay, E.H., Hutchinson, S.A., Boto, G.M., 1982. Ammonia and inorganic phosphate excretion by zooplankton from inshore waters of the439 Great Barrier Reef, Queensland. I. Relationship between excretion rates and body size. Mar. Freshwater Res. 33 (1), 55–70.440 Ikeda, T., Kanno, Y., Ozaki, K., Shinada, A., 2001. Metabolic rates of epipelagic marine copepods as a function of body mass and temperature.441 Mar. Biol. 139, 587–596.442 18 Ikeda, T., Torres, J.J., Hern´ andez-Le´ on, S., Geiger, S.P., 2000. Metabolism. In: R.P. Harris, P.H. Wiebe, J. Lenz, H.R. Skjoldal, M. Huntley (Eds.),443 ICES Zooplankton methodology manual. Academic Press, San Diego, pp. 455–532.444 Isla, J.A., Llope, M., Anad´ on, R., 2004. Size-fractionated mesozooplankton biomass, metabolism and grazing along a 50oN-30oS transect of the445 Atlantic Ocean. J. Plankton Res. 26 (11), 1301–1313.446 King, F.D., Cucci, T.L., Townsend, D.W., 1987. Microzooplankton and macrozooplancton glutamate dehydrogenase as indices of the relative447 contribution of these fractions to ammonium regeneration in the Gulf of Maine. J. Plankton Res. 9 (2), 277–289.448 King, F.D., Packard, T.T., 1975. Respiration and the activity of the respiratory electron transport system in marine zooplankton. Limnol. Oceanogr.449 20 (5), 849–854.450 Lowry, O.H., Rosebrough, N.J., Farr, A.L., Randall, R.J., 1951. Protein measurement with the Folin phenol reagent. J. Biol. Chem. 193, 265–275.451 Mayzaud, P., Conover, R.J., 1988. 0:N atomic ratio as a tool to describe zooplankton metabolism. Mar. Ecol. Prog. Ser. 45, 289–302.452 Osma, N., Aristizabal, M., Fern´ andez-Urruzola, I., Packard, T.T., G´ omez, M., 2016. Influence of starvation on respiratory metabolism and pyridine453 nucleotide levels in the marine dinoflagellate Oxyrrhis marina. Protist 167 (2), 136–147.454 Owens, T., King, F.D., 1975. The measurement of respiratory electron transport system activity in marine zooplankton. Mar. Biol. 30 (1), 27–36.455 Packard, T.T., Devol, A., King, F.D., 1975. The effect of temperature on the respiratory electron transport system in marine plankton. Deep-Sea456 Res 22 (4), 237–249.457 Packard, T.T., G´ omez, M., 2008. Exploring a first-principles-based model for zooplankton respiration. ICES J. Mar. Sci. 65 (3), 371–378.458 Packard, T.T., Healy, M.L., Richards, F.A., 1971. Vertical distribution of the activity of the respiratory electron transport system in marine plankton.459 Limnol. Oceanogr. 16 (1), 60–70.460 Packard, T.T., Williams, P.J.L., 1981. Rates of respiratory oxygen consumption and electron transport in surface seawater from the northwest461 Atlantic. Oceanol. Acta 4 (3), 351–358.462 Park, Y.C., 1986a. Impact of starvation and feeding experiments on ammonium excretion and glutamate dehydrogenase activity of zooplankton.463 Kor. Biochem. J. 19 (1), 251–256.464 Park, Y.C., 1986b. Nitrogen regeneration and glutamate dehydrogenase activity of macrozooplankton in the Southeastern Sea of Korea. J. Oceanol.465 Soc. Korea 21 (2), 110–117.466 Park, Y.C., Carpenter, E.J., Falkowski, P.G., 1986. Ammonium excretion and glutamate dehydrogenase activity of zooplankton in Great South Bay,467 New York. J. Plankton Res. 8 (3), 489–503.468 Parsons, T.R., Maita, Y., Lalli, C.M., 1984. Manual of chemical and biological methods for seawater analysis. Pergamon Press, New York, 173 pp.469 Postel, L., Fock, H., Hagen, W., 2000. Biomass and abundance. In: R.P. Harris, P.H. Wiebe, J. Lenz, H.R. Skjoldal, M. Huntley (Eds.), ICES470 Zooplankton methodology manual. Academic Press, San Diego, pp. 83–192.471 Regnault, M., 1987. Nitrogen excretion in marine and fresh-water crustacea. Biol. Rev. 62, 1–24.472 Richardson, K., 1991. Comparison of 14C primary production determinations made by different laboratories. Mar. Ecol. Prog. Ser. 72, 189–201.473 Rutter, W.J., 1967. Protein determination in embryos. In: F.H. Wilt, N.V. Wessels (Eds.), Methods in Developmental Biology. Academic Press,474 London, pp. 671–684.475 Sameoto, D., Wiebe, P., Runge, J., Postel, L., Dunn, J., Miller, C., Coombs, S., 2000. Collecting zooplankton. In: ICES Zooplankton methodology476 manual. Academic Press, San Diego, pp. 55–81.477 Schneider, G., 1990. A comparison of carbon based ammonia excretion rates between gelatinous and non-gelatinous zooplankton: Implications478 and consequences. Mar. Biol. 106 (2), 219–225.479 Solorzano, L., 1969. Determination of ammonia in natural waters by the phenolhypochlorite method. Limnol. Oceanogr. 14 (5), 799–801.480 Sprugel, D.G., 1983. Correcting for bias in log-transformed allometic equations. Ecology 64 (1), 209–210.481 Steinberg, D.K., Saba, G.K., 2008. Nitrogen consumption and metabolism in marine zooplankton . In: E.J. Carpenter, D.G. Capone (Eds.),482 Nitrogen in the marine environment. Academic Press, London, pp. 1135–1196.483 Yentsch, C.S., Menzel, D.W., 1963. A method for the determination of phytoplankton chlorophyll and phaeophytin by fluorescence. In: Deep Sea484 Research and Oceanographic Abstracts. Elsevier, pp. 221–231.485 19