Technical Note: Towards resolving in situ, centimeter-scale location and timing of biomineralization in calcareous meiobenthos – the calcein–osmotic pump method

Insights into oceanographic environmental conditions such as paleoproductivity, deep-water temperatures, salinity, ice volumes, and nutrient cycling have all been obtained from geochemical analyses of biomineralized carbonate of marine organisms. However, we cannot fully understand geochemical proxy incorporation and the fidelity of such in species until we better understand fundamental aspects of their ecology such as where and when these (micro)organisms calcify. Here, we present an innovative method using osmotic pumps and the fluorescent marker calcein to help identify where and when calcareous meiofauna calcify in situ. Method development initially involved juvenile quahogs (Mercenaria mercenaria); subsequent method refinement involved a neritic benthic foraminiferal community. Future applications of this method will allow determining the in situ growth rate in calcareous organisms and provide insights about microhabitats where paleoceanographically relevant benthic foraminifera actually calcify.

stand geochemical proxy incorporation and the fidelity of such in species until we better understand fundamental aspects of their ecology such as where and when these (micro)organisms calcify. Here, we present an innovative method using osmotic pumps and the fluorescent marker calcein to help identify where and when calcareous meiofauna calcify in situ. Method development initially involved juvenile quahogs (Mercenaria mercenaria); subsequent method refinement involved a neritic benthic foraminiferal community. Future applications of this method will allow determinations of in situ growth rate in calcareous organisms and provide insights about microhabitats where paleoceanographically relevant benthic foraminifera actually calcify. 15 Biomineralized carbonate of marine organisms such as foraminifera, coccolithophores, and ostracods has provided an abundance of geochemical data critical to our understanding of modern-day oceanographic conditions and processes as well as critical to reconstructions of paleoceanographic conditions and processes. While geochemical proxies of planktic and benthic foraminiferal tests (shells) have yielded copious 20 insights to past sea-surface temperatures, salinity, ice volumes, deep-water temperatures, oceanic circulation patterns, nutrient cycling and paleoproductivity (e.g., Katz et al., 2010;Allen and Honisch, 2012), in the vast majority of cases, initial proxy calibration was developed from core-top sampling and field calibrations. Culturing studies have also contributed greatly to our understanding of the mechanisms controlling these Introduction Discussion Paper | Discussion Paper | Discussion Paper | Discussion Paper | edge on these topics (reviewed by Katz et al., 2010), there remain some significant issues regarding fundamental and emerging proxies. In brief, a variety of factors complicate proxy interpretations; the most common ones in this context include "microhabitat preferences", "vital effects", and rapid changes in carbonate chemistry occurring in the uppermost sediment column. Microhabitats refer to the micron-or millimeter-scale 5 distribution of foraminifera with respect to the sediment-water interface, some other physical structure (e.g., worm tube), or chemocline. Vital effects, which can include ontogenetic differences (McCorkle et al., 2008;Filipsson et al., 2010), are physiological processes that impact test geochemistry, although some researchers include environmental processes in the definition of "vital effects" (de Nooijer et al., 2014). 10 In particular, changes in environmental parameters occurring in the uppermost part of the sediment might affect proxy reconstruction and it is crucial to obtain an increased understanding of where in the sediment biomineralization occurs. For example, one of the most often used temperature proxies, foraminiferal Mg / Ca, relies on temperature, but is also influenced by carbonate ion concentration (Elderfield et al., 2006;Rosen-15 thal et al., 2006;Healey et al., 2008;Raitzsch et al., 2008) and pH (e.g., Russell et al., 2004), which vary significantly with sediment depth. Stable isotopes of oxygen and carbon (δ 18 O and δ 13 C) also are impacted by the carbonate ion effect (Spero et al., 1997;Bijma et al., 1999;Lea et al., 1999;McCorkle et al., 2008). With increasing sediment depth, pore water becomes increasingly depleted in δ 13 C due to a steep gradient in 20 pore-water dissolved inorganic carbon (DIC) because of organic carbon remineralization. This same process is simultaneously lowering the carbonate ion concentration. Thus, for benthic foraminifera, determining where within (or on) the sediments they calcify is very important for determining the fidelity of their test chemistry and the resulting proxy relationships, as well for improving the precision of proxy reconstructions. Introduction by McCorkle et al. (1990), abundance peaks of Rose Bengal stained foraminifera are typically several centimeters thick yet these authors showed species' δ 13 C ranges of 1-2 ‰, which may suggest calcification in a narrower depth horizon. While it is possible that foraminiferal species calcify at the sediment-water interface or a particular sediment horizon, thereby incorporating only the DIC from bottom waters or the hori-5 zon's pore waters, subsequent migration into more oxygen-depleted zones characterized by extremely low δ 13 C values hypothetically results in an apparent disequilibrium between ambient conditions and foraminiferal calcitic tests (McCorkle et al., 1990;Stott et al., 2002). Such activity would explain at least in part the disequilibrium observed in down core studies (e.g., McCorkle et al., 1990). In reality, while benthic foraminiferal calcification horizons are inferred from distribution studies (e.g., Stott et al., 2002), the actual depth of calcification or the related geochemistry is not known, especially for paleoceanographically relevant species, such as, for example, Cibicides spp., Uvigerina spp., and Oridorsalis umbonatus. Furthermore, distribution patterns may not be reliable given that the classically employed method to distinguish live from dead foraminifera, 15 Rose Bengal stain, has been shown to undoubtedly also stain foraminiferal carcasses (i.e., dead foraminifera, Bernhard et al., 2010). The requirements for monospecific (single species') analyses as well analyzing specimens within a well-defined size range to avoid biases caused by vital effects or microhabitat effects can minimize geochemical proxy uncertainty (Ravelo and Hillaire-Marcel, 2007;Katz et al., 2010), but at this time 20 it is not established that all conspecifics calcify in the same microhabitats and/or depth horizons or that vital and ontogenetic effects are consistent among an entire population of a given species To resolve some of these unknowns, we developed a method that will assist in documenting the timing and location of calcification in sediments for calcareous benthic 25 meiofauna. The method employs commercially available osmotic pumps to deliver calcein, which is a fluorescent compound that binds to calcium in biomineralized structures as it is precipitated (e.g., Monaghan, 1993;Medeiros-Bergen and Ebert, 1995;Collin and Voltzow, 1998;Hernaman et al., 2000;Moran, 2000). Using full immersion incu- bations, calcein has been used to mark bivalves (e.g., Kaehler and McQuaid, 1999;Moran and Marko, 2005;van der Geest et al., 2011) and also has been used in laboratory studies regarding foraminiferal calcification Dissard et al., 2009;Filipsson et al., 2010;Denoyelle et al., 2012;Nardelli et al., 2014;Kurtarkar et al., 2015). In this contribution, we describe a novel point-source calcein dispensa-5 tion method and show proof-of-concept for quahog (hard clam) bivalves and benthic foraminifera.

Osmotic pumps and calcein
The means used to dispense the calcein are ALZET ™ osmotic pumps ( Fig. 1a; 10 DURECT Corporation, Cupertino, CA, USA). Osmotic pumps are devices designed to deliver pharmaceuticals to animals; as originally intended, they are installed under the skin of an animal. Different osmotic pump models allow different delivery rates and durations. We used model 2ML2 or 2ML4, each with a reservoir of 2 mL. The 2ML2 was designed to dispense (in mammals) at a rate of 5 µL h −1 for 14 days; the 15 2ML4 at a rate of 2.5 µL h −1 for 28 days. Dispensation rate depends on the model, as noted, but also on osmolality and temperature of the environment. A calculator to determine flow rate under specific relevant conditions conveniently exists on the ALZET web page: (http://www.alzet.com/products/guide_to_use/pump_selection.html). In our quahog incubations, we expected each 2ML2 to dispense for about 2 months and, in 20 the foraminifera incubations, for about 4 months. Because incubations were performed in seawater, to avoid corrosion, we replaced the stainless steel tubing that is standard in the ALZET osmotic pumps with PEEK (PolyEtherEtherKetone) tubing. The osmotic pumps were filled with a concentrated solution of calcein (100 mg L −1 ; Discussion Paper | Discussion Paper | Discussion Paper | Discussion Paper | heretofore referred to as the "port", could be placed facing downward or upward, depending on the research objective. To avoid clogging of the port, for downward pointing osmotic pumps, a thin plastic film (conventional kitchen plastic wrap) was loosely wrapped over the dispensation end during pump emplacement into sediments. After the pump was located within sediments as desired, the thin film was removed by gently 5 pulling one edge vertically so as to minimize disturbance to the sediments. Upward facing osmotic pumps did not require such protection during emplacement.

Bivalve incubations
Our initial incubations employed juvenile bivalves (quahogs and surf clams; initially ∼ 5 mm and 1 cm in length, respectively), starting in December 2012. Intertidal sedi-10 ments that were collected from a local salt pond were divided into four containers so that each had a sediment layer of ∼ 10 cm of sediment. Bivalves were seeded into the sediments at a density of about 1 bivalve per square centimeter. Both species (Mercenaria mercenaria; Spisula solidissima similus) are surface dwelling or shallow-infaunal taxa. One calcein-filled osmotic pump was placed into each container so that the port 15 was located in the container center, just below the sediment-water interface so that the calcein would emanate near the sediment-water interface. These containers were initially maintained at 7 • C. During that time, containers were installed in a recirculating seawater system containing ∼ 10 L. Two containers were installed in each circuit. Salinity was monitored weekly with a refractometer and adjusted 20 to 35 as needed. After the first ∼ 3 weeks, in order to increase bivalve calcification rate, the containers were thereafter maintained at room temperature (21 • C). Due to logistic reasons, circulating the containers at 21 • C was not possible so each container was aerated with an aquarium bubbler. During this time, salinity was monitored and new seawater was added approximately every 2 weeks.
Throughout the incubations, living algal food (Dunaliella tertiolecta; Isochrysis galbana) was provided to each container every week  concentrated via gentle centrifugation and ∼ 40-50 mL was introduced into each container, so as not to disturb the sediment-water interface, each week. Every 2-3 weeks whole specimens (live) were removed from containers noting their location with respect to the osmotic pump. Specimens were typically burrowed into the top cm. Each bivalve was examined with epifluorescence microscopy (see below) to 5 determine if they had incorporated the fluorescent marker calcein. After examination, each individual was placed back in the sediment near its original location. After ∼ 2 months (i.e., the approximate end of calcein dispensation), the osmotic pumps were removed from the sediments and the quahogs allowed to grow for another ∼ 2 weeks. Then, the quahogs were removed from the sediments and preserved in 70 % ethanol.  (Bothner et al., 1981). The site is well known as being a 15 sediment-focus area, causing sediments to be muddy. Sediments were collected with an Ocean Instruments MC800 multicorer. The top 14-18 cm of designated multicores was extruded into 40 cm long core liners of identical internal diameter (i.e., 10 cm). Bottom waters from the collection site, which were collected in Niskin bottles attached to a CTD-rosette sampler, were carefully introduced into each core liner to produce a 20 seawater header of ∼ 20 cm. These introductions did not visibly disturb the sedimentwater interface of any core. Cores were maintained near in situ temperature (8-10 • C) and brought back to our shore-based cold room (7 • C).
As with the bivalve incubations, one calcein-filled osmotic pump was placed into each core so that the dispensation port was located in the core center. In most cases, the 25 port was placed just below the sediment-water interface. In other cases, the port was placed deep (4 cm) below the sediment-water interface. The cores were maintained at 7 • C. Due to logistic reasons, circulating the cores with seawater was not possible so each core was aerated with an aquarium bubbler. Salinity was monitored with a refractometer weekly and adjusted to 35 as required. Throughout the foraminiferan incubations, living algal food (Dunaliella tertiolecta; Isochrysis galbana) was provided to each core every week, as noted above for bivalves. After ∼ 4 months, each osmotic pump was gently removed and each core was sub-5 sampled as follows. All overlying water was carefully removed. A plastic ring identical in diameter to the core barrel was placed atop the barrel and the core barrel was gently lowered 1 cm so that 1 cm of the core extended into the ring. A thin stainless steel plate was then passed between the core barrel and ring to isolate the surface cm. Our goal was to obtain samples at horizontally and vertically discreet distances from the osmotic 10 pump port. Because surface sediments had high water content, the first core sectioned was subsampled by taking 8-mm-diameter syringe cores along four radii in the surface centimeter (Fig. 1b). The remaining 0-1 cm sediments were retained separately. All subsequent subsampling of 1 cm intervals was configured in concentric rings (Fig. 1b, c). Thus, the next 1 cm interval was extruded into the large diameter ring, 15 three thin-walled plastic rings were concentrically placed into the core ∼ 1 cm, the thin stainless steel plate was used to slice the core horizontally while the concentric rings were held in place, and the sediments delimited between concentric rings were placed into plastic bottles, and properly labeled (as center, inner, outer or rim) along with depth interval below the sediment-water interface. The 1-2, 2-3, 3-4, and 4-5 cm intervals 20 of each core were subsampled using this concentric ring approach. All sediment samples were preserved in 70 % ethanol. Each sediment sample was sieved with artificial seawater over a 63 µm screen and > 63 µm fraction microscopically examined.

Microscopy
Epifluorescence microscopy (480 nm excitation; long pass 518 nm emission) was used 25 to assess calcein incorporation. Preserved materials were examined with a Leica FLIII stereomicroscope equipped with epifluorescence capabilities and an Olympus DMP70 Discussion Paper | Discussion Paper | Discussion Paper | Discussion Paper | digital camera. Whole quahogs were examined. Whole foraminifera, obtained from the > 63 µm fraction of sieved sediment aliquots, were also examined. Once imaged at low magnification, select quahog shells were cut with an Isomet slow-speed rock saw (0.4-mm-thick blade) to obtain valve cross sections. These valve cross sections had to be polished with fine grit wet-dry sandpaper to obtain a 5 smooth surface. To remove organics, shells were exposed to 3 % sodium hypochlorite for 20 min. After rinsing and drying, these were examined and imaged with the epifluorescence-equipped Leica FLIII stereo-dissecting microscope and DMP digital camera and/or with an Olympus Fluoview Confocal Laser Scanning Microscope (CLSM).

Proof of concept
The surf clams only survived 1-3 weeks in the cores, but the quahogs remained active and grew, evidenced by the observations that many (∼ 25 %) of the quahogs had portions of their shell that were fluorescent (Fig. 2). The observation that the umbo 15 and other oldest parts of each quahog valve were not fluorescent is consistent with the fact that we seeded the cores with juveniles that were ∼ 0.5 cm in length. Keeping in mind that bivalves add material to their valves in increasingly larger concentric annuli, somewhat akin to tree rings, it was noted that the fluorescence appeared as bands, where some portions fluoresced and others did not. Incorporation of calcein into newly 20 precipitated carbonate was confirmed by CLSM of bleached valves (Fig. 2g).
Because the system used during bivalve incubations was recirculating or lacking flow, it is important to consider the maximum concentration of calcein possible if all contents of the osmotic pump were dispensed into the seawater. We calculate that, at most, the recirculating seawater would have had 0.02 mg L −1 calcein. For mollusc Introduction  (e.g., Kaehler and McQuaid, 1999;Klunzinger et al., 2014). Thus, the maximum possible calcein concentration in recirculating seawater during the first 3 weeks of bivalve incubations was far below the level required to fluorescently label calcite. Of course if the bivalves were near the concentrated source of calcein (i.e., pump port, 100 mg L −1 calcein), then we predict the new calcite would be fluorescent, 5 as observed.
In the static (but aerated) setup, the maximum concentration of calcein possible if all contents of the osmotic pump were dispensed into the seawater was 0.2 mg L −1 , which again is far below the minimum mollusc threshold of 80 mg L −1 . There are two more lines of evidence that further support the inference that the lack of constant water 10 replacement did not cause artifactual calcein labeling in the bivalves: (1) the seawater overlying the cores did not have strong fluorescence when viewed with the appropriate optics using the stereo dissecting microscope (i.e., there was no significant background signal).
(2) Not all bivalves were labeled with calcein. Some quahogs did not fluoresce in any part of their valves, indicating they were sufficiently removed from the point 15 source throughout the incubation. Some of the calcareous benthic foraminifera in the cores exhibited bright fluorescence while others did not (Fig. 3a, b). As established for the calcein labeling method, the non-fluorescent calcareous specimens either did not calcify during the incubation  or were too far from the osmotic pump port to incorporate 20 calcein. Some of the calcareous foraminiferal tests fully fluoresced (Fig. 3a-d), while others had only one or two brightly fluorescent chambers (Fig. 3e, f). It is possible that the fully fluorescent specimens were the result of reproduction during the experiment (Filipsson et al., 2010;Hintz et al., 2004). In contrast to brightly fluorescent rotalids, entire tests of miliolid (porcelaneous) calcareous foraminifera fluoresced dimly; no ag-25 glutinated foraminifera fluoresced (not shown). It is known that at least some miliolids fluoresce after incubation in calcein, even without calcification . It is also established that when rotalid (hyaline) calcareous foraminifera add new chambers, a thin veneer of calcite is precipitated over existing chambers (Erez, 2003;Nehrke , 2013). Such a growth habit explains the differential fluorescence patterns in some foraminiferal specimens, where 1-2 chambers are brightly fluorescent and remainder of the test has less intense fluorescence (Fig. 3f).
As noted for the non-recirculating bivalve incubations, we do not believe the calcein concentration of the overlying seawater would exceed the minimum labeling threshold 5 in the foraminiferal incubations even if the entire osmotic pump contents were released. For foraminiferal calcite labeling, a calcein concentration of 10 mg L −1 has been typically used previously Denoyelle et al., 2012;Nardelli et al., 2014;Filipsson et al., 2010) although concentrations as low as 5 mg L −1 reportedly also labeled foraminiferal calcite (Dissard et al., 2009;Kurtarkar et al., 2015). The maximum 10 possible calcein concentration in overlying waters of our foraminiferal incubations was less than 1 mg L −1 . As for bivalves, if growing calcareous foraminifera were near the concentrated source of calcein (i.e., pump port, 100 mg L −1 calcein), then we predict new calcite would be fluorescent, as observed. Unfortunately, the calcein-labeled foraminiferal densities were insufficient to deter-15 mine the vertical and horizontal extent of calcein diffusion into our muddy sediments. Specifically, calcein labeled foraminifera were absent from all small-volume radial subsamples of the 0-1 cm interval of one core. Calcein-labeled foraminifera were found, however, in the remaining bulk 0-1 cm interval of the first multicore. Time and resource limitations prohibited full processing of additional multicores; spot checks in those sam-20 ples did not yield convincing fluorescent foraminiferal calcite.

Notes and caveats
In the course of our method development, a number of lessons were learned. To assist future users of the method, these topics are discussed. Orienting the osmotic pump port downwards is problematic in fine-grained and/or low water-content sediments be-25 cause conditions likely impede calcein dispensation or the diffusion is too localized to expose many specimens to calcein. It is expected that sediments with high water content (e.g., sediments with "fluff" layers) would not impede dispensation as much as 9454 Introduction

Tables Figures
Back Close

Full Screen / Esc
Printer-friendly Version

Interactive Discussion
Discussion Paper | Discussion Paper | Discussion Paper | Discussion Paper | more compacted or consolidated sediments. Attempts to test the osmotic pump port at 4 cm depth did not result in any fluorescent specimens, but we do not know if that was due to lack of calcification or spatially limited calcein diffusion.
To document specifics regarding infaunal calcification horizons, it will be critical to determine the extent of calcein diffusion into sediments. Diffusion will vary with sediment grain size, water content, compaction, hydrodynamics, and community composition (e.g., presence or absence of bioturbators). Initial verification tests should be considered prior to initiating a lengthy or complicated experiment.
Calculations based on expected dispensation rate, temperature, and salinity can provide estimated duration of calcein efflux. Osmotic pumps are single use; they will not dispense if refilled. Per manufacturer's instructions, osmotic pumps will not perform well if handled without clean gloves.
The cytoplasm of at least one benthic foraminiferal species autofluoresces using excitation and emission wavelengths similar to those for calcein (Apotheloz-Perret-Gentil et al., 2013). The foraminiferal species known to autofluoresce lacks a carbonate test, 15 so it cannot be confused with our calcein-labeling approach. If there are calcareous foraminifera with similarly autofluorescent cytoplasm, distinguishing between cytoplasmic fluorescence (from viability indicators reliant on similar excitation and emission wavelengths) and carbonate fluorescence is not difficult if one considers the patterns and shapes of the signal (Nardelli et al., 2014).

Potential applications
The calcein-osmotic pump method can be used without modification to assess growth rates and calcification locations of juvenile and meiofaunal metazoans with calcareous hard parts (e.g., gastropods, echinoids, ostracods). These units can be deployed in shallow marine waters near shellfish fisheries and in reef areas with sediment pockets. Introduction As noted, our method will help to better understand foraminiferal microhabitats. Such knowledge will help to minimize uncertainty and increase precision in records of paleoceanographic proxies preserved in foraminiferal tests. For instance, recently the difference in the δ 13 C for epifaunal Cibicides wuellerstorfi and for deep infaunal Globobulimina spp. was used to reconstruct bottom-water oxygen concentration (Hoogakker et 5 al., 2015), an approach that could be further improved by this method. Foraminiferalbased ecology studies under shifting environmental conditions, such as varying water oxygen concentration (Nardelli et al., 2014) or pH conditions would also benefit from our calcein-osmotic pump method. Of course, our method should be considered a first step given that most paleoceanographically-relevant foraminifera presumably are able 10 to migrate within the sediment column. It is unknown, however, if benthic foraminifera actually migrate vertically in situ and, if they do, where they calcify. Regardless, additional refinements to the approach will be required to prevent or minimize foraminiferal migration during and/or after calcification. Our osmotic pump method can be further modified to deploy these units in deep- 15 sea sediments using a Remotely Operated Vehicle (e.g., Jason) or a Human Occupied Vehicle (e.g., Alvin). Further, calcein-filled osmotic pumps can be installed in habitats that are spatially restricted, such as hydrocarbon seeps or brine pools, where we have little growth data for any sediment-dwelling species. 20 While calcein has been used in growth studies for a variety of organisms, to our knowledge, calcein has not been used as a point source to determine calcification in the environment. Most studies using calcein to determine growth rates immerse entire specimens in the laboratory and then release them into nature for later recapture. Our new calcein-osmotic pump approach can help pinpoint where and when meiofaunal organ-Introduction Discussion Paper | Discussion Paper | Discussion Paper | Discussion Paper | de Nooijer, L. J., Hathorne, E. C., Reichart, G. J., Langer, G., and Bijma, J.: Variability in calcitic Mg / Ca and Sr / Ca ratios in clones of the benthic foraminifer Ammonia tepida, Mar. Micropaleontol., 107, 32-43, 2014. Denoyelle, M., Geslin, E., Jorissen, F. J., Cazes, L., and Galgani, F.: Innovative use of foraminifera in ecotoxicology: A marine chronic bioassay for testing potential toxicity of drilling 5 muds, Ecol. Indic., 12, 17-25, 2012. Dissard, D., Nehrke, G., Reichart, G. J., Nouet, J., and Bijma, J.: Effect of the fluorescent indicator calcein on Mg and Sr incorporation into foraminiferal calcite, Geochem. Geophys. Geosyst., 10, Q11001, doi:10.1029/2009gc002417, 2009. Elderfield, H., Yu, J., Anand, P., Kiefer, T., and Nyland, B.: Calibrations for benthic foraminiferal