Diurnal variability in alkaline phosphatase activity and the potential role of zooplankton

Daily light–dark cycles drive the circadian rhythm of many ocean processes including photosynthesis, gene expression, and zooplankton diel vertical migration (DVM). In phosphate deplete surface ocean regions, microbes produce metalloenzymes, such as alkaline phosphatases (AP), to access dissolved organic phosphorus. Here, we provide novel evidence of diurnal variation in AP activity (APA) in the subtropical North Atlantic using two independent datasets, with APA being two‐ to three‐fold higher at night. We demonstrate that zooplankton are a source of AP and postulate that zooplankton DVM is a source of enhanced AP in the surface waters at night, with reduction or degradation of AP during the day. Our results challenge the current assumption that APA is linear over a 24‐h period. While future ocean scenarios predict intensification and expansion of oceanic phosphate limitation, our findings indicate a role for zooplankton in regenerating phosphate that is currently missing in conceptual and numerical models.

migrate to the surface at night and to depth during the day (Longhurst 1991). The influence of DVM on biogeochemical nutrient cycles has been studied in relation to nitrogen, carbon, and iron (Steinberg et al. 2000(Steinberg et al. , 2002Schmidt et al., 2016) but is less well-studied in relation to phosphorus (P).
Throughout, much of the low-latitude Atlantic Ocean, surface phosphate concentrations are chronically low (<10 nmol L −1 ) and limit or colimit microbial growth, nitrogen fixation, and primary productivity (Moore et al. 2013). The prevalence and intensity of phosphate stress is predicted to increase because of enhanced stratification and nitrogen fertilization (Capotondi et al. 2012;Jickells et al. 2017). Under phosphate-stressed conditions, microbes may synthesize enzymes to access dissolved organic phosphorus (DOP) to meet their P requirements (Mahaffey et al. 2014). The DOP pool is composed of esters, anhydrides, and phosphonates and can be hydrolyzed by a range of enzymes including nucleotidases, phosphonatases, adenosine-triphosphatases, and alkaline phosphatases (AP) (Duhamel et al. 2010).
AP is a metalloenzyme group with iron or zinc cofactors that can hydrolyze the labile monoester pool, which may account for as much as 75% of the DOP pool (Kolowith et al. 2001;Hoppe 2003). AP activity (APA) rapidly increases below 30 nmol L −1 phosphate and thus is often used to infer phosphate stress or limitation across transects, over times series, and in bioassay experiments (Mahaffey et al. 2014). Such studies assume no diurnal variation in APA, despite there being a light dependence and diurnal variation in phosphate uptake (Duhamel et al. 2012).
Here, we use data from two field campaigns and a previously published dataset in the subtropical North Atlantic Ocean to address the following questions: (1) is there diurnal variability in APA, and (2) do zooplankton influence APA in the surface ocean? Quantifying the diurnal variability in APA and the role of zooplankton in surface ocean P cycling will allow more accurate assessment and detection of future changes in ocean P dynamics, which are inextricably linked to the nitrogen and carbon cycles.

Near-surface APA
Seawater was collected on board the RRS James Cook during summer 2017 (JC150: 22 June to 12 August) between 30 W and 60 W and 22 N and 23 N in the subtropical North Atlantic (Fig. 1) using a trace metal clean "towed fish" at~2 m depth. This sampling method allowed us to obtain a high spatial and temporal resolution of near-surface sampling. In total, 130 samples were collected at 2-h intervals at a spatial resolution of 0.23 AE 0.07 W (26 AE 8 km; Fig. 1).
APA rates were measured in unfiltered seawater using the fluorometric substrate 4-methylumbelliferyl-phosphate (MUF-P, Sigma Aldrich; Ammerman 1993). Samples were incubated with either 500 nmol L −1 (51 W-30 W) or 2000 nmol L −1 (60 W-51 W) MUF-P at in situ conditions. Fluorescence was measured every 2 h for up to 12 h using a Turner 10AU field fluorometer (long wavelength oil lab filter kit 10-302R, 365 nm emission, 455 nm excitation). A calibration and daily single standard were produced using 4-methylumbelliferone (MUF, Sigma Aldrich, concentration range 0-2000 and 400 nmol L −1 , respectively) to convert the rate of change in fluorescence to MUF-P hydrolysis rate, considered synonymous with volumetric APA (APA vol , nmol P L −1 h −1 ). Fluorescence measurements were corrected for abiotic hydrolysis or MUF-P degradation using blanks (boiled sample plus 500 nmol L −1 MUF-P).
To interrogate diurnal patterns in the data, APA vol was normalized to local (within AE 1 ) APA vol measured at midnight (AE1 h). This gave a value of the fractional change in volumetric APA vol across a 24-h cycle, termed δAPA vol .

Zooplankton incubations and APA
To investigate the influence of zooplankton on APA (APA zoo ), replicate experiments were conducted on the RRS James Clark Ross (JR15007: June to July 2016, n = 7) and on the RRS James Cook (JC150, n = 5) ( Fig. 1). Zooplankton were collected using WP2-size net rings (57 cm diameter) fitted with 200 μm mesh size and nonfiltering cod ends. During both cruises, plankton tows were performed at night from 200 m to the surface at 0.5 m s −1 (Giering et al. 2012). This depth was chosen to ensure sufficient biomass for replicate incubations and for intercruise consistency. Samples were immediately sorted after tows with healthy, active organisms being twice washed prior to incubation by gentle sequential transfer into filtered surface seawater (0.7 μm GF/F during JR15007; 0.2 μm Durapore during JC150). Incubations consisted of taxa that were abundant at each station, namely calanoid copepods (JC150 and JR15007) and gelatinous organisms (JR15007 only,~15%).
Next, during JR15007, healthy active organisms were gently pipetted into 10 liter unfiltered surface seawater and incubated in the dark in triplicate at an incubation density of 14 AE 5 mg L −1 . Samples were sacrificed at each time point (every 30 to 60 min) during the 9 h incubations. Prescreened (63 μm mesh) subsamples were taken for determination of APA (10 μmol L −1 MUF-P; see section Near-Surface APA), and the remaining sample was filtered onto preweighed 0.7 μm GF/F and stored at −20 C for later determination of dry weight biomass.
During JC150, calanoid copepods were gently pipetted into 1 liter surface seawater and incubated in the dark in triplicate with an incubation density of 7 AE 5 animals L −1 in 1 liter polycarbonate bottles. Duplicate experiments were conducted using filtered (0.2 μm Durapore) and unfiltered surface seawater, to investigate whether APA zoo was influenced by excretion or grazing processes. Incubation bottles were spiked with MUF-P (final concentration 2 μmol L −1 ) and fluorescence measurements were taken every 30 to 60 min for 6 to 8 h. This duration was sufficiently long for animal adaptation and grazing (Nöges 1992) and sufficiently short that the change in fluorescence remained linear, consistent with the previous APA incubations.
APA was determined from the linear change in fluorescence over time. All APA zoo measurements were blank corrected (boiled seawater control to monitor abiotic changes and surface seawater only control to measure background APA).

Dissolved phosphate
During JR15007, samples for phosphate analysis were collected using 20-liter Niskin bottles on a stainless-steel rosette frame supporting a Seabird 911 conductivity, temperature, depth (CTD) instrument and fluorometer. During JC150, samples were collected using the "towed fish" at~2 m depth and the CTD. Samples were collected directly into aged, acid-washed, Milli-Q rinsed, and sample rinsed 60 mL HDPE Nalgene ® bottles. Nutrients were analyzed on board using a Bran and Luebbe segmented flow colorimetric auto-analyzer (AAIII) using techniques described in Woodward and Rees (2001). Nutrient reference materials (KANSO Technos) were analyzed daily. On JR15007 and JC150, precision was between 2% and 3%. Limits of detection were 0.02 μmol L −1 for phosphate during JR15007 but lower during JC150 (0.5 nmol L −1 ) due to the use of a 2 m liquid waveguide (Zhang and Chi 2002).

Flow cytometry
During JC150, samples for picophytoplankton and heterotrophic bacterial analysis were collected from the "towed fish" sampling system using 125 mL polycarbonate bottles (acid washed, Nalgene ® ). Immediately after sampling, 1.9 mL of sample was pipetted into 2 mL cryovials containing 20 μL of 50% glutaraldehyde solution (Fisher Scientific), inverted 10 times, and stored at 4 C for 1 h. Samples were then flash frozen with liquid nitrogen and stored at −80 C.
Samples were analyzed using a flow cytometer (FACSort™ B43) at Plymouth Marine Laboratory using techniques described by Tarran et al. (2001). Four groups were identified and enumerated: Prochlorococcus, Synechococcus, picoeukaryotes, and total heterotrophic bacteria (high and low nucleic acid groups), which collectively were used as total cell abundance.

Diurnal variation in APA
There was significant diel periodicity in δAPA vol (R 2 = 0.52, p = 0.017; Fig. 2a) and APA normalized to total cell abundance (δAPA cell , R 2 = 0.41, p = 0.011; Fig. 2b), which more than doubled between the minimum rates in early afternoon when irradiance was highest (~0.6 kW m −2 ) and maximum rates at night (Fig. 2a). Applying the same derivation to the data of Wurl et al. (2013), collected in the subtropical North Atlantic (U.S. GEOTRACES GA03, 15 October to 27 November 2010, 06 November to 11 December 2011, RV Knorr; Fig. 1), revealed a remarkably similar diurnal variation in δAPA vol and with rates doubling between day and night (R 2 = 0.55, p = 0.07; Fig. 2c).

Zooplankton and APA
APA vol was significantly higher in incubations with zooplankton than without zooplankton, such that APA zoo was positive throughout (Table 1). During JC150, APA zoo was up to 17-fold higher in filtered seawater and up to 6-fold higher in unfiltered seawater compared to controls. Once corrected for blank and background activity and normalized to the number of animals in each incubation, APA zoo was similar in filtered and unfiltered seawater experiments (9-61 nmol P animal −1 d −1 and 7-71 nmol P animal −1 d −1 , respectively). APA zoo in filtered seawater incubations showed no correlation to APA control , whereas in unfiltered seawater incubations APA zoo was highest when APA control was lowest (Table 1).
During JR15007, APA doubled over a 9-h period and biomass-normalized APA (APA wt ) increased 2.6-fold (Fig. 3). The linear increase in APA wt over incubation time (R 2 = 0.96, p < 0.001) suggests that AP release was relatively constant when normalized to zooplankton biomass. We suggest this was not a stress response because AP and other Pho regulon genes are not upregulated in zooplankton as a stress response (Lauritano et al. 2012) and there was no mortality during our short incubations (<10 h), which were designed to avoid animal mortality following Nöges (1992).
During JC150, there were too few zooplankton to accurately measure biomass in each incubation. Instead, body size to mass equations (Fernandez Araoz 1991) were used to estimate mean zooplankton mass (1.6 AE 0.3 mg), and APA vol was normalized to zooplankton biomass based on the number of animals in each incubation. Combined APA wt data from both cruises showed a significant logarithmic relationship with phosphate integrated across the surface 200 m (i.e., tow depth), with the highest APA wt at the lowest integrated phosphate values (Fig. 4).

Discussion
Using two independent datasets, we have shown that there is a significant two-fold diurnal variation in APA vol that is zonally, seasonally, and annually persistent, suggesting that it is a consistent feature in the subtropical North Atlantic. Phosphate concentrations are known to be chronically low and limit productivity in our study region, where APA is commonly used as an indicator of phosphate limitation (Mahaffey et al. 2014). Yet, the degree of phosphate stress in the microbial community did not have an apparent effect on δAPA vol or δAPA cell (Fig. 2). Such a diurnal variation in APA vol may be driven by circadian rhythms in microbial gene expression (Cohen and Golden 2015;Frischkorn et al. 2018). For example, changes in cellular enzymes can be a physiological response to the light-dark cycle, to oxidative stress, or to modulate cellular resource demands (Saito et al. 2011;Morris et al. 2016). The marine diazotroph Crocosphaera watsonii alternates its enzyme inventory across a diel cycle to economize cell quotas by transferring iron between N 2 fixation and photosynthetic processes (Saito et al. 2011). Such a process could also apply to AP, which has zinc or iron metal cofactors (Coleman 1992;Yong et al. 2014). An indirect circadian control may occur as a result of certain symbiotic unicellular cyanobacteria, such as UCYN-A and UCYN-B, inducing diurnal P stress in their photosynthetic hosts as a result of diel variation in N 2 fixation (Tripp et al. 2010;Zehr et al. 2016).
Alternatively, our study shows that the presence of zooplankton significantly enhanced APA in our incubations and so we hypothesize that zooplankton DVM can significantly contribute to the diurnal variation in APA vol in the surface ocean. In our study, δAPA vol was highest at night coinciding with zooplankton DVM to the surface waters (Fig. 2). Zooplankton phosphatase activity has been previously studied in freshwater (Wynne and Gophen 1981) and coastal regions (Jamet and Boge 1998) but to our knowledge has not been directly studied in the open ocean. There are several mechanisms by which zooplankton may enhance APA, including release from their prey during grazing, release of digestive enzymes to the particulate or dissolved pool through excretion or mortality, or release via their bacterial epibiont community (Båmstedt 1988;Davis and Mahaffey 2017;De Corte et al. 2018).
Focusing on the western basin (JC150), the diurnal variation in APA vol ranged from 31 nmol P L −1 d −1 in the day to 74 nmol P L −1 d −1 at night (data not shown). Applying values of APA zoo from Table 1 to a grazing density of 1.5 animals L −1 (Hays et al. 1996) over a 12-h grazing period (Bianchi and Mislan 2015), we estimate that APA zoo could contribute between 9 and 53 nmol P L −1 d −1 (21% to 124%) of the diurnal variation in APA vol . Meanwhile, the relationship between APA zoo and phosphate (Fig. 4) is similar to that of APA and phosphate reported by Mahaffey et al. (2014). Therefore, this reveals a potential link between APA zoo and phosphate stress of Table 1. Volumetric (nmol L -1 d −1 ) and normalized (nmol animal −1 d −1 ) rates of zooplankton associated APA (APA zoo ) during JC150 (03 July 2017 to 08 August 2017) in unfiltered and 0.2 μm filtered surface seawater and associated control activity (APA control , nmol L -1 d −1 ). SD is the standard deviation of three replicate incubations. Normalized APA zoo values have been corrected for blank associated activity.  the ambient microbial community, suggesting a contribution from the bacterial epibiont to APA zoo (De Corte et al. 2018). If zooplankton are a nocturnal source of AP in the surface ocean, then to resolve the diurnal variation in APA vol there must be a net sink of AP during the day. Factors driving AP loss in the surface could include microbial degradation of the enzyme to access protein as an N source, loss associated with sinking particles and fecal material, competitive binding and inhibition at the active site, or daytime photodegradation and oxidative inhibition of enzyme activity (Boavida and Wetzel 1998;Scully et al. 2003;Baltar et al. 2013;Davis and Mahaffey 2017). Future work into the turnover time of AP under different conditions in seawater is needed to improve our understanding of potential AP sink terms.
Phosphate-limited regions are predicted to expand and intensify due to enhanced anthropogenic and natural nitrogen addition in excess of phosphorus (Kim et al. 2014;Jickells et al. 2017). Future warming will enhance stratification at low latitudes, reducing marine productivity because of the restricted vertical nutrient supply (Bopp et al. 2013;Capotondi et al. 2012). The absolute contribution of APA zoo in the future is unclear, with changes in prey abundance, refuge depth and predation all being influential factors (Hays 2003;Cohen and Forward Jr. 2009;Chust et al. 2014). Future changes in the relative importance of APA zoo to surface APA are also unclear, as evidence of a similar relationship between phosphate concentration and APA zoo (Fig. 4) and APA vol associated with the microbial community (Mahaffey et al. 2014) suggests that both may increase under phosphate limitation. Clearly, more work is required to unravel the complex interplay among zooplankton biomass, autotrophic biomass, and phosphorus recycling that underpins the accuracy of future climate predictions.
The major implications of our findings are: (1) there is significant diurnal variation in APA vol in the surface ocean, which is an important consideration when interpreting hydrographical survey work or time series data investigating phosphorus dynamics; and (2) zooplankton provide a significant source of AP to the surface ocean during DVM, encouraging non-Redfieldian DOM remineralization and significantly contributing to ambient dissolved APA vol . Our study demonstrates that changes in DVM because of enhanced stratification, water transparency, temperature gradients, and oxygen depletion will not only influence nitrogen regeneration and carbon export but also surface ocean P regeneration via reduction of zooplankton-derived AP. Given the predicted increase in phosphate stress in subtropical oceanic regions, any changes in zooplankton biomass or behavior will have significant ramifications for phytoplankton community structure and biomass.

Conclusions
We have shown that there is a seasonally and spatially persistent doubling of APA vol between day and night in the surface subtropical North Atlantic using two independent datasets. Furthermore, we have provided evidence that APA is enhanced by the presence of zooplankton and that the degree of enhancement has a robust relationship with surface layer phosphate. Thus, we conclude that zooplankton are a potentially important source of AP in the surface ocean and that this signal combined with degradation or removal of the enzyme during the day dominates the diel variability in surface APA vol . These findings reveal a critical role of zooplankton in the surface cycling of phosphorus that may be vital to improve our understanding of surface ocean biogeochemistry in the contemporary ocean and accurate prediction of future ocean scenarios.