Extraction and quantification of microphytobenthic Chl a within calcareous reef sands

Calcareous reef sands are characterized by high concentrations of photosynthetic pigments that extend well below the sediment surface, as well as by high variability in concentrations between study sites. An important contributor to the observed variability may be differences in extraction protocols, further complicated by variability in calcareous sand characteristics. We tested the effects of freeze-drying, grinding, sonication, extraction temperature, and extraction time on quantification of Chl a content within calcareous sands. The resulting optimized extraction protocol consists of freeze-drying, grinding with a mortar and pestle for 30 s, and extracting with 100% acetone at –20°C for at least 20 h, yielding a 39% increase in Chl a content over frozen, unground samples. Using this protocol, we measured and compared ten sedimentary Chl a profiles taken in close proximity to test for relationships between surface and sub-surface concentrations. Sedimentary Chl a content at a backreef location on the south shore of O‘ahu varied between 4.33-14.25 μg g–1 dw, with distinctly higher values occurring in a relatively enriched surface layer (0-1 cm). Surface Chl a concentrations varied between 86-307 mg m–2, depending on the depth of integration (0.5-2 cm), with 73% of the full-core (0-8 cm) Chl a concentration occurring below 2 cm. The concentrations of surface and subsurface layers were significantly correlated between cores, allowing for the use of plug sampling when profile generation is not feasible or necessary to determine the magnitude of the subsurface microphytobenthic biomass and its variability over scales of meters. *Corresponding author: E-mail: [email protected] Acknowledgments This research was supported by the U.S. National Science Foundation (Grant OCE-1031947). We thank the Associate Editor, Dr. Gordon Taylor, and two anonymous reviewers for their constructive comments. We also thank Ms. Stephanie Christensen for conducting the HPLC analysis. This is SOEST contribution 9079. DOI 10.4319/lom.2014.12.126 Limnol. Oceanogr.: Methods 12, 2014, 126–138 © 2014, by the American Society of Limnology and Oceanography, Inc. LIMNOLOGY and OCEANOGRAPHY: METHODS macroalgae in a typical reef setting (Roelfsema et al. 2002), and consequently MPB in sands could account for the same amount of productivity as the latter do over the scale of the whole coral reef ecosystem (Clavier and Garrigue 1999; Werner et al. 2008). The potential magnitude of MPB contribution to ecosystem-wide benthic primary productivity, especially in view of the anticipated decline in dissolved O2 due to climate change (Keeling et al. 2010), necessitates further investigation into their role as important components of calcareous reef sand communities and habitats. Pigment distribution in sediments A key parameter in any study of benthic primary productivity is MPB biomass. However, the enumeration and identification of MPB is a time-consuming process, and consequently, photosynthetic pigments, and especially Chl a, are commonly used as proxies of MPB biomass (MacIntyre et al. 1996). Typically, a sedimentary sample of known dry weight is collected using a core of known cross-sectional area, is analyzed, and the amount of recovered pigment is normalized to (a) the dry weight of sediment to calculate the content, M, defined as μg of pigment per g of dry sediment, or μg g–1 dw, and (b) the cross-sectional area of the core to calculate the concentration, C, which is defined as mg of pigment per unit area of seafloor, or mg m–2 (see “Materials and procedures” for calculation details). Infrequently, M is converted to C using sedimentary weight-to-volume conversion factors, even though such conversions must be avoided when based on assumptions and not measurements of these factors (e.g., Tolhurst et al. 2005). Two major sampling approaches for microphytobenthic pigment determination are commonly reported in the literature: profiling and plug sampling. Profiling consists of sediment coring, sectioning of the core (typically every 0.5-1 cm), and analysis of each section for M, followed by integration over a layer ranging from the sediment-water interface (SWI) down to a chosen depth (hereby defined as the depth of integration), and then dividing by the core cross-section area to obtain a C value (e.g., Werner et al. 2008). Selection of the depth of integration is a crucial step. Integrating over a depth of several mm to 1 cm below the SWI is commonly considered reasonable (Table 1), since availability of scalar irradiance (both visible and infrared, 400-880 nm) that can be used for photosynthesis in sands is restricted to the top few millimeters of the sediment column (Lassen et al. 1992; Kühl et al. 1994), and most of the actively photosynthesizing cells are expected to aggregate there. Whereas depthrelevant information provided by profiles can be useful, the number of samples needed, especially after factoring in the need for replicates, can make this a time-consuming and costly approach. Plug sampling may greatly minimize the number of samples needed to effectively determine C at a given location. It consists of obtaining a single section that extends from the SWI down to a predetermined depth, essentially replacing the depth of integration with a sampling depth (e.g., Roelfsema et al. 2002; Grinham et al. 2007). Grinham et al. (2007) demonstrated that 8 plug samples down to 2 cm are sufficient to generate a reliable measure of C on a scale of meters, but also pointed out that M can be underestimated if high-value surface layers (e.g., 0-0.5 cm) are diluted by deeper poorer layers (e.g., 1.5-2 cm). Moreover, C can be underestimated if the sampling depth excludes deeper layers where pigments are present. Plug sampling may be best suited for cases where pigments are detectable only a few mm to 1 cm below the SWI, as is commonly the case in quiescent locations such as estuaries and in finer-grained sediments (e.g., Cartaxana et al. 2006). In high-energy settings that are characterized by the presence of permeable, coarse-sand sediments, viable pigment-containing MPB can be distributed for centimeters to decimeters below the SWI (MacIntyre et al. 1996; Roelfsema et al. 2002; Werner et al. 2008). Accordingly, available profiles of pigment content in calcareous reef sediments feature significant amounts of pigment (e.g., >3 μg g–1 dw, Fig. 1) down to 10 cm below the SWI, well below scalar irradiance penetration depths. Another common feature of the limited available data are high variability (×2-5) both between replicate samples at the same depth as well as with depth in the same profile, with com127 Hannides et al. Microphytobenthic Chlorophyll a in reef sands Table 1. Concentration of sedimentary Chl a for different sediment types. The depth from the sediment-water interface (SWI) indicates the thickness of the sedimentary layer from the sediment-water interface over which C is calculated, and corresponds to the depth of integration and sampling depth for profiling and plug sampling, respectively. Location (sediment type) Chl a concentration (mg m–2) Depth from SWI (cm) Reference Southwest lagoon, New Caledonia (calcareous sand) 59 ± 4–62 ± 5 1 Clavier and Garrigue (1999) Heron Reef (calcareous sand) 36–1153 1.5–2 Roelfsema et al. (2002) Heron Reef (calcareous sand) 92–995 1 Heil et al. (2004) Heron Reef (calcareous sand) 31 ± 3–84 ± 10 1 Werner et al. (2008) Cooloolo Passage (sand) 47 ± 2 2 Grinham et al. (2007) Oxley Creek (silt) 33 ± 3 2 Grinham et al. (2007) Georgia shelf (siliceous sand) 6–37.7 0.5 Nelson et al. (1999) Florida shelf (siliceous sand) 10.6–40.9 0.5 Nelson et al. (1999) Tagus estuary (siliceous sand) 59 ± 18 0.2 Cartaxana et al. (2006) Tagus estuary (mud) 61 ± 20 0.2 Cartaxana et al. (2006) monly occurring sub-surface maxima (Fig. 1). These patterns in calcareous reef sands may reflect dynamic processes shaping pigment distributions in the vertical dimension, e.g., sediment mixing, cell filtration, and cell motility (MacIntyre et al. 1996; Rao et al. 2012). When coupled with high irradiance and low turbidity, which commonly characterize reef habitats (Clavier and Garrigue 1999), these processes result in high content variability and deep penetration below the SWI. Because subsurface MPB (2-6 cm below SWI) has been shown to be viable (i.e., if given light, it has the potential to photosynthesize; Werner et al. 2008), sampling only surface sediment or integrating only over surficial depths to estimate C may exclude the majority of viable MPB biomass from the calculation. Consequently, given the concentration of subsurface pigments in calcareous reef sediments, subsurface sediment sampling and analysis is necessary to obtain a more comprehensive understanding of MPB-driven processes such as benthic primary productivity and photosynthesis/respiration coupling. Also, given the variability between replicate profiles (Fig. 1), the number of replicates necessary to characterize a station must be chosen carefully, and the logistical burden of generating more than 2-3 profiles per station must be considered. Moreover, the existence of a relationship between surface and sub-surface content has not been explicitly assessed and is far from certain, particularly given the aforementioned variability and the dynamic nature of calcareous reef sediment settings. These considerations dictate a need to compare profiling with plug sampling, while at the same time counter-balancing statistical rigor with logistical constraints. Sample processing and analytical considerations One potential reason for the observed variability between studies of pigments in calcareous reef sediments may be the extraction efficiency achieved with different analytical protocols (Heil et al. 2004). Whereas methods of pigment quantification in the water column are well-established (e.g., Jeffrey et al. 1997; Roy et al. 2011), similar methods for sediments are under debate (Reuss and Conley 2005; Hagerthey et al. 2006). Moreover, the nature of permeable calcareous reef sediments complicates the issue. Compared with siliciclastic grains, biogenic calcareous grains are irregularly shaped and marked by various microtopographic features such as pits, crevices, and crevasses, resulting in poorer sorting, a higher grain surfacearea-to-volume ratio, and consequently greater microbial abundances (Wild et al. 2005; Schöttner et al. 2011). The more complex nature of the substrate, therefore, results in the necessity for multiple methodological steps to detect more accurately various microbial characteristics, such as cell numbers (Wild et al. 2006; Rusch et al. 2009). The extraction efficiency of pigments from complex substrates, such as sediment, is commonly enhanced by freezedrying, grinding, and sonication, used in isolation or sequentially (Pinckney et al. 2011). Freeze-drying a sample before extraction significantly improves the extraction efficiency of pigments compared with “wet-frozen” extraction (BuffanDubau and Carman 2000), and also improves peak resolution during HPLC, irrespective of the solvent used for the extraction (Hagerthey et al. 2006; van Leeuwe et al. 2006). Moreover, water contained in sandy sediment samples can induce dilution effects that differ from sample to sample, due to high but variable permeability and poor sorting (Buffan-Dubau and Carman 2000). Changes in the solvent solution concentration due to water in the sample may also significantly shift the wavelength at which pigment absorption maxima occur (Porra 2011). Therefore, removal of water by freeze-drying may improve analytical performance and precision. Mechanical disruption via sonication is considered indispensable for calcareous reef sediments because the complex grain surface allows for the development of complex biofilms and the persistent attachment of cells and detrital materials (Wild et al. 2006; Rusch et al. 2009). Sonication has been shown to improve extraction irrespective of the solvent used, and allows shorter extraction times (Cartaxana and Brotas 2003). Sonication duration may vary from 8-10 short bursts of 3-6 s (Hagerthey et al. 2006) to several (7-15) minutes of continuous treatment (Grinham et al. 2007; Werner et al. 2008). In the case of long sonication times, the concomitant temperature increase may cause pigment degradation (Metaxatos and Ignatiades 2002), and consequently, it has been recommended to sonicate in an ice bath using cold solvents and a succession of short bursts (Reuss and Conley 2005; Hagerthey et al. 2006; van Leeuwe et al. 2006). Grinding is typically put forth as an alternative to freezedrying and sonication. Whereas no effects have been observed if samples are previously freeze-dried (Hagerthey et al. 2006), grinding can significantly improve extraction efficiency in the case of frozen samples (Heil et al. 2004). Calcareous reef sedi128 Hannides et al. Microphytobenthic Chlorophyll a in reef sands Fig. 1. Chlorophyll a (Chl a) content (dw = dry weight) profiles in calcareous reef sediments of Heron Reef, Great Barrier Reef, Australia (data replotted from the literature): (A) means and standard deviations of triplicates at each one of two stations: coarse sand on the reef flat (mean grain size, dmean = 897 μm), and fine sand on the reef edge (dmean = 590 μm) (Rao et al. 2012); (B) duplicate profiles from three stations: NB2 on the reef flat (dmean = 591 μm), RB on the reef edge (dmean = 426 μm), and Ch in the channel (dmean = 227 μm) beyond the reef edge (Werner et al. 2008). ment grain surfaces are characterized by rough microtopography that may protect pigment-containing cells from lysis during extraction. Thus, abrasion during grinding may release the cell-rich layer of the grain surfaces and dramatically increase exposure to the solvent. In addition to procedures involving freeze-drying, sonication, and grinding, various extraction durations and temperatures have been reported. Long extraction times of 24 h or more are not uncommon (Clavier and Garrigue 1999; van Leeuwe et al. 2006; Werner et al. 2008), however they are thought to result in pigment degradation, isomerization, and allomer formation (Cartaxana and Brotas 2003). On the other hand, very short extraction times of less than an hour may not be sufficient. For example, cells with shells, such as testate diatoms, may take longer to digest than cyanobacteria (Jeffrey et al. 1997). For siliceous sandy sediments, studies have indicated that extraction times greater than 3 h do not result in significantly improved efficiencies (Buffan-Dubau and Carman 2000; Cartaxana and Brotas 2003), especially when the samples have been freeze-dried and mechanically disrupted (usually by sonication). The optimal duration of extraction also depends on the extraction temperature, which must be low enough to slow down pigment degradation but high enough to induce fast and efficient dissolution (van Leeuwe et al. 2006). Sedimentary and/or calcareous matrix samples have either been extracted at 2-4°C (Heil et al. 2004; Hagerthey et al. 2006; Chevalier et al. 2010) or at –20°C (Grinham et al. 2007; Werner et al. 2008), and a study on microalgal cultures demonstrated no observed difference between the two temperatures (van Leeuwe et al. 2006). Objectives of this study The preceding overview leads to two outstanding issues in the study of MPB in calcareous reef sediments: the isolated and synergistic effects of various methodological steps on pigment extraction efficiency, and the quantification of subsurface C and its relation to surface-layer C. In this study, we developed an optimized pigment extraction protocol for calcareous reef sands by investigating the effects of freeze-drying, sonication, grinding, and extraction duration and temperature on the final M measured. Using the optimized protocol, we analyzed and compared multiple sedimentary pigment profiles taken in close proximity to test for relationships between surface and sub-surface C values when investigating distributions on the lateral scale of meters. Based on our results, we propose a sampling strategy that captures the full-core pigment inventory (in mg m–2) and its lateral variability without the time-consuming and resource-intensive requirements of fine-scale sediment profiling, where the latter is not needed. Materials and procedures Sediment sampling and sub-sampling We obtained samples for this study from the shallow (0.5 m depth) near-shore back-reef of Waialae Beach Park (21° 16.1′ N, 157° 46.5′ W) on the south shore of Oahu, Hawaii. Surface sediment (down to 1 cm below the SWI) for the optimization of the extraction protocol was collected using several 50-mL Blue Max modified-polystyrene conical (“Falcon”) tubes (BD). Cores for profiling were taken using 60-mL cut-off syringes (BD) with an inner diameter of 2.6 cm. Overlying water samples were collected in amber 250-mL Nalgene HDPE bottles (Thermo Fisher Scientific). All samples were placed on ice and kept in the dark (Buffan-Dubau and Carman 2000; Grinham et al. 2007) until arrival at the laboratory, which typically occurred within 30 min of sampling. Upon arrival to the laboratory, the syringe cores were immediately stored upright in an ultra-cold (–80°C) freezer (Cartaxana and Brotas 2003; Reuss and Conley 2005), and the surface sediment from all Falcon tubes was combined and gently homogenized with a spatula in an ice-cooled beaker. Subsamples (1-2 g, 0.5-1 mL) of the pooled, homogenized sediment were transferred to pre-weighed 13 × 100 mm borosilicate tubes (VWR), which were covered with Parafilm (Pechiney, Plastic Packaging), and stored in an ultra-cold (–80°C) freezer until analysis. The syringe cores were sectioned as soon as they were frozen (typically within 12-24 h after storage) at 0.5-cm depth intervals using a clean metal blade, and the sections were stored at –80°C in pre-weighed borosilicate tubes covered with Parafilm. Water samples were filtered onto 25-mm Whatman GF/F filters (GE Healthcare Life Sciences), which were folded in aluminum foil and stored at –80°C until extraction. The period between sampling and analysis did not exceed 2 weeks, thus minimizing storage effects. Sample processing and subsequent analysis were conducted under dim light (Jeffrey et al. 1997). Sedimentary characteristics Sedimentary characteristics at the study site (Table 2) were determined on sediment collected down to 10 cm below the SWI using a 5-cm diameter core. Grain-size distribution was determined by wet-sieving, and mean grain size, sorting, and skewness were calculated using the equations in McManus (1988). Porosity was determined gravimetrically (Breitzke 2000), whereas permeability was determined by the constanthead method (Klute and Dirksen 1986). 129 Hannides et al. Microphytobenthic Chlorophyll a in reef sands Table 2. Characteristics of the back-reef sediment (0-10 cm below the sediment-water interface) of South Oahu, Hawaii, used in this study (n=3). Results indicate the average and one standard deviation. The dimensionless measures for sorting (σ1) and skewness (SK1) are determined using the phi system (McManus 1988). Mean grain size (phi, μm) 0.98 ± 0.11, 509 ± 39 Median grain size (phi, μm) 1.08 ± 0.27, 480 ± 90 Sorting, σ1 1.78 ± 0.02 Classification (Wentworth scale) Poorly sorted Skewness, SK1 –0.23 ± 0.12 Classification (McManus) Negatively skewed Porosity 0.45 ± 0.02 Permeability (×10–11 m2) 1.86 ± 0.7 Freeze-drying All samples, except those used as frozen controls, were freeze-dried in the dark using a Virtis Benchtop K freeze drier (SP Scientific) at –80°C and pressure < 3 Pa (Hagerthey et al. 2006; van Leeuwe et al. 2006). Sediment samples weighing 12 g were typically dry within 5-10 h. Freeze-dried samples were immediately processed to avoid pigment degradation during storage (Reuss and Conley 2005). Grinding Samples were ground with a mortar and pestle for 30 s (Heil et al. 2004). The mortar and pestle were cleaned between samples using a brush and laboratory tissue paper, rinsed with 100% acetone, and allowed to dry before reuse. Extraction solvent We extracted pigments using 100% acetone, as it has been shown to recover all pigments (especially pheopigments) from sedimentary matrices more efficiently than other solvents, and results in better-resolved peaks during HPLC (BuffanDubau and Carman 2000; Werner et al. 2008). A known volume (3 mL) of cold (2-4°C) 100% reagent grade acetone (Thermo Fisher Scientific) was added to the 1.5-2 g freezedried, ground sediment subsamples, to yield a sediment-tosolvent volume ratio between 1:2 and 1:3 (Grinham et al. 2007). Such a ratio has been determined as optimal because it maximizes the amount of extracted pigment without diluting the signal significantly. Sonication After solvent addition, the samples were briefly mixed with a vortexer, and then sonicated in a bath-style Branson 200 Ultrasonic cleaner (Branson Ultrasonics) filled with chipped ice (Hagerthey et al. 2006) for a 30-s period (van Leeuwe et al. 2006; Chevalier et al. 2010). When testing required longer sonication times, these were accomplished by sonicating for 30-s periods interspersed with 15-s pauses. Extraction duration and temperature Following sonication, the samples were placed in the dark in a refrigerator at 2-4°C (Heil et al. 2004; Hagerthey et al. 2006; Chevalier et al. 2010) or in a freezer at –20°C (Grinham et al. 2007; Werner et al. 2008; Ruivo et al. 2011) for periods of time between 2 and 60 h, depending on the treatment. During extraction periods longer than 12 h, the samples were mixed by vortexing at least once every 12 h (e.g., Arar and Collins 1997; van Leeuwe et al. 2006). Chlorophyll a analysis Upon conclusion of the extraction period, the samples were vortexed, followed by centrifugation at 1750×g for 5 min (Grinham et al. 2007). A measured volume of the supernatant was removed, diluted if necessary, and Chl a and pheopigment content determined by fluorometry before and after acidification (Yentsch and Menzel 1963; Holm-Hansen et al. 1965) using a Turner TD700 fluorometer (Turner Scientific). Fluorometry was selected for its high analytical sensitivity and speed, as well as low cost, and is considered sufficient in detecting relative differences in Chl a (Pinckney et al. 1994). Selected samples were also analyzed by HPLC to determine the relationship between the two methods for our samples. Mixtures of 1-mL extract (spiked with canthaxanthin as an internal standard) plus 0.3-mL HPLC grade water were prepared in opaque autosampler vials, and 200 μL were injected into a Varian 9012 HPLC system (Varian/Agilent) equipped with a Varian 9300 autosampler (Varian/Agilent), a Timberline column heater (26°C; Timberline Instruments), and a Waters Spherisorb® 5-μm ODS-2 analytical (4.6 × 250 mm) column and guard cartridge (7.5 × 4.6 mm). Pigments were detected with a ThermoSeparation Products UV2000 detector (λ1 = 436, λ2 = 450; Thermo Fisher Scientific). A ternary solvent system was used for pigment analysis: Eluent A (methanol:0.5 M ammonium acetate, 80:20, v/v), Eluent B (acetonitrile:water, 87.5:12.5, v/v), and Eluent C (100% ethyl acetate). Solvents A and B contained an additional 0.01% 2,6-di-ter-butyl-p-cresol (0.01% BHT, w/v; Sigma-Aldrich) to prevent the conversion of Chl a into Chl a allomers. The linear gradient used for pigment separation was a modified version of the Wright et al. (1991) method: 0.0 min (90% A, 10% B), 1.00 min (100% B), 11.00 min (78% B, 22% C), 27.50 min (10% B, 90% C), 29.00 min (100% B), 30.00 min (100% B), 31.00 min (95% A, 5% B), 37.00 min (95% A, 5% B), and 38.00 min (90% A, 10% B) (Bidigare et al. 2005). Eluent flow rate was held constant at 1.0 mL min–1. Pigment peaks were identified by comparison of retention times with those of pure standards and extracts prepared from algal cultures of known pigment composition. A dichromatic equation was used to resolve mixtures of monovinyl and divinyl Chl a spectrally (Bidigare and Trees 2000). After removing the supernatant, the remaining sediment samples were dried in a drying oven at 60°C for 72 h and weighed, to calculate sedimentary Chl a content MChla (μg g –1 dw):