Evaluation of laser scanning confocal microscopy as a method for characterizing reef-building coral tissue thickness and Symbiodiniaceae fluorescence

When reef-building corals are exposed to temperatures exceeding their thermal maxima in the presence of solar irradiance, they expel their obligate algal symbionts, a phenomenon known as coral bleaching that has led to mass die-offs globally and is increasing in severity and frequency (Hughes et al., 2018). Corals exhibit a wide range of physiological responses to thermal stress, both within and between species and at different life stages (Chou et al., 2016; Cunning et al., 2016; Krueger et al., 2015; Marshall and Baird, 2000). Physiological metrics including tissue thickness, Symbiodiniaceae population density and chlorophyll content are common variables in assessing coral stress response (e.g. Edmunds et al., 2012; Krediet et al., 2015; Nielsen et al., 2018; Rotmann and Thomas, 2012; Sudek et al., 2012), and are key indicators of coral health and resilience (Ainsworth et al., 2008; Edmunds et al., 2014; Loya et al., 2001). However, current methodologies often require destructive practices and are not well suited for measurement on small juvenile colonies. Here, we review current tools and introduce an application of laser scanning confocal microscopy (LSCM) for measurement of physiological condition.

Current methods for characterizing coral tissue thickness and Symbiodiniaceae population density and chlorophyll content often require destructive sampling, preventing repeated measurements on live individuals. Tissue thickness has historically been measured on bisected samples with a caliper or microscope (Ainsworth et al., 2008; Cruz-Piñón et al., 2003; Rotmann and Thomas, 2012), tissue penetration on calcified cross-sections (Edmunds et al., 2012), or by histological preparation (comparison in Loya et al., 2001). Two common metrics of Symbiodiniaceae communities are photosynthetic pigment content and population density, both of which relate to the quantity and quality of translocated photosynthates provided to the coral host (Roth, 2014). Chlorophyll content is measured by extraction from homogenized tissue and measured by spectrophotometry (Fitt et al., 2000; Jeffrey and Humphrey, 1975; Wall et al., 2019) or HPLC (Ambarsari et al., 1997; Daigo et al., 2008). Symbiodiniaceae population densities are recorded by counting cells in a homogenized sample on a hemocytometer (Ainsworth et al., 2008; Nielsen et al., 2018), as well as with flow cytometry (Krediet et al., 2015) and qPCR (Cunning and Baker, 2013). However, the precision of these methods is dependent on sample preparation (Cunning and Baker, 2014), requiring a minimum density for accuracy and destructive sampling, which complicates studies tracking coral bleaching and obtaining accurate densities in juveniles. Additional methodologies to measure these characteristics on living corals and in early life stages would increase our capacity to understand coral physiological response to stress.

LSCM technology has the capacity to address this need through non-destructive visualization of fluorescent proteins in host tissues and Symbiodiniaceae chlorophyll. The technology is widely used in the medical industry, but has only been used on corals in a small number of studies on calcification (Ohno et al., 2017; Wall and Nehrke, 2012), polyp morphology (Sivaguru et al., 2014), intracellular pH buffering (Gibbin et al., 2014, 2015), microbial communities (Garren and Azam, 2012), disease progression (Caldwell et al., 2017), bleaching in the anemone Aiptasia sp. (Matthews et al., 2016), infection of Symbiodiniaceae cells (Bay et al., 2011; Biquand et al., 2017) and tissue depth (Yost et al., 2013).

We build upon previous utilization of LSCM to measure coral physiological characteristics and apply these methods to track bleaching. On tissue samples of four species, we optimized LSCM methods to measure (1) depth of coral and symbiont fluorescence as an indicator of tissue thickness and (2) Symbiodiniaceae fluorescence as an indicator of Symbiodiniaceae cell densities, and compare these values with those obtained using other methodologies. We then test the utility of LSCM to measure these characteristics in adults and juveniles of two species, and relate Symbiodiniaceae fluorescence to chlorophyll a content and pigmentation during a thermal stress experiment. We show that LSCM measurements are significantly correlated to comparative methodologies and provide proxy measures of physiological condition on living corals.

To examine the validity of LSCM methodology to measure proxies of tissue thickness and Symbiodiniaceae populations, we collected six (n=6) fragments (<1 cm) of each of four reef-building coral species (Montipora capitata Dana 1846, Porites compressa Dana 1846, Pavona varians Verrill 1864 and Pocillopora acuta Lamarck 1816) at patch reef no. 20 in Kāneʻohe Bay, Hawaiʻi, in March 2017 under the HIMB Special Activities Permit 2017 (Division of Aquatic Resources, Hawaiʻi). Imaging was conducted on fixed tissue samples to allow for repeated measurement with multiple methodologies. Fragments were rinsed in 1 μm filtered seawater and tissue samples were collected, avoiding areas of recent growth (i.e. no branch tips or plate edges) and immediately fixed in 1:4 zinc-buffered formalin (Z-Fix Concentrate, Anatech, Ltd) diluted with 1 μm filtered seawater and refrigerated. Samples were decalcified in 10% hydrochloric acid for 2–8 h depending on the thickness of skeleton and then stored in 1× phosphate-buffered saline at 4°C in the dark until imaging.

Samples were scanned and imaged using Zeiss LSCM 710 and Zen Black processing software (2011 v14.0.16.201). Excitation wavelength of 405 nm with a Diode laser (UV) allowed for visualization of fluorescent proteins (FPs) in the host and Symbiodiniaceae fluorescence in adult samples at a magnification of 2.5×. In the method described here, we did not image or analyze specific FPs, their location or proposed function; rather, we utilized the fluorescent properties of coral tissue and Symbiodiniaceae to characterize: (1) depth of tissue fluorescence as a proxy for tissue thickness, and (2) Symbiodiniaceae fluorescence as a proxy measurement of population density and chlorophyll a content. LSCM image acquisition settings are provided in Table S1. Tissue samples were imaged in a glass bottom dish (0.17 mm) filled with 1:4 Z-Fix diluted with filtered seawater for inverted imaging.

Symbiodiniaceae fluorescence values were collected using Zen Black processing software using maximum intensity projections. Fluorescence intensity values were measured using region of interest (ROI) tracing (n=3 ROI per sample) and averaged for each coral sample. Following image acquisition of coral samples, we calibrated the Symbiodiniaceae fluorescence values using a linear equation to a red fluorescence standard with the Red InSpeck Microscope Image Intensity Calibration Kit (Molecular Probes, Sigma-Aldrich). Symbiodiniaceae fluorescence is in units of percent relative intensity to the calibration standard (% relative intensity, RI). As a comparative metric, Symbiodiniaceae population densities were measured with a hemocytometer. Tissues were homogenized in 1:4 Z-Fix solution diluted with filtered seawater and cells were counted in three replicates with technical duplicates. Population density was calculated as the total number of cells in each sample normalized to the planar surface area (cm²) of the tissue sampled.

We utilized fluorescence of FPs in coral tissue and fluorescence of Symbiodiniaceae to characterize depth of fluorescence as a proxy for tissue thickness. We describe a method to measure depth of fluorescence on three-dimensional models built from acquired z-stack images, which do not require bisection for analysis and can be applied for non-destructive measurement on coral samples. Depth of fluorescence on cross-sections was measured in triplicate and averaged for each coral sample. As a comparison to LSCM measurements, tissue thickness of fixed and bisected coral tissue samples was measured using a dissecting microscope and Fiji software v1.0 (Schindelin et al., 2012). Coral tissue samples were bisected using a razor blade and imaged with a scale bar on an Olympus dissecting microscope (Olympus Corporation of the Americas, Center Valley, PA, USA). Thickness was measured at four locations on the bisected slice and averaged for each coral sample.

Adult and 1-year-old juvenile corals of a perforate-tissue species, M. capitata (n=13 adults, n=17 juveniles), and an imperforate-tissue species, P. acuta (n=14 adults, n=14 juveniles), were exposed to an ex situ simulated thermal stress to test the utility of LSCM in measuring tissue thickness and Symbiodiniaceae fluorescence across physiological states and life stages (collected in Kāneʻohe Bay, Hawaiʻi, under the HIMB Special Activities Permit 2019 Division of Aquatic Resources, Hawaiʻi). Perforate-tissue corals (e.g. M. capitata) have porous skeletons with three-dimensional mesh-like tissue connections occurring through the skeletal framework, while imperforate-tissue species (e.g. P. acuta) are characterized by a thinner 'veneer' of tissue atop a non-porous skeleton (Yost et al., 2013). Fragments (1–3 cm in length) of adult M. capitata, adult P. acuta and 1-year-old juvenile P. acuta corals were cut from colonies (n=21 colonies, n=2 fragments per colony) using a coral saw (Gryphon Diamond Band Aquasaw) and held in outdoor flow-through raceways. One-year-old M. capitata juveniles attached to aragonite plugs were isolated by trimming aragonite plugs. Owing to the small size (∼0.05–0.15 cm²) of these juveniles, they were not cut into duplicate fragments. Adult coral fragments and fragments of juvenile P. acuta were secured using hot glue to small aragonite tiles and placed in 1 liter containers (n=2 containers per treatment) in outdoor flow-through water tables in ambient sunlight under two layers of shade cloth (peaking at ∼100 µmol photons m⁻² s⁻¹ daily) and supplied with ambient sand-filtered seawater.

One coral fragment of each colony pair was exposed to either high or ambient temperature treatments for 14 days (P. acuta) or 20 days (M. capitata) until visible bleaching was observed. Montipora capitata juvenile corals without a colony pair were randomly placed into each treatment. The high temperature treatment was controlled by heating water supply with a Neptune Apex Aquarium controller system (Neptune Systems, Morgan Hill, CA, USA) in combination with 1500 W titanium aquarium heaters. Fragments were placed in experimental tank systems on 11 August 2019 and allowed to acclimate to ambient conditions until initial LSCM imaging on 15 August. On 16 August, temperature in the high temperature treatment was increased 1°C per day for 3 days until a maximum daily temperature of 31.3°C was reached. During the treatment period, temperature was recorded every 15 min (HOBO Pendant loggers). Mean temperature in the ambient treatment was 29.3°C (∼1.0°C diel fluctuation) and 30.9°C (∼1.0°C diel fluctuation) in the high treatment.

We measured depth of fluorescence and Symbiodiniaceae fluorescence using LSCM at initial (0 days of exposure) and final (P. acuta, 14 days of exposure; M. capitata, 20 days of exposure) time points of thermal exposure on live corals (n=58 corals). LSCM acquisition settings for M. capitata and P. acuta are displayed in Table S1. Depth of fluorescence (n=3 measurements per coral) and Symbiodiniaceae fluorescence (n=3 measurements per coral) of corals were measured as described previously.

To compare LSCM measurements with comparative methodologies, we randomly selected and sampled corals (n=26) at the end of treatment exposure for tissue thickness measured on bisections on a dissecting microscope as previously described. To provide comparison to Symbiodiniaceae fluorescence measurements, we randomly selected and sampled corals (n=23) for chlorophyll a content by freezing fragments at −80°C. Tissue was then removed from the skeleton with an artists' airbrush and filtered (0.2 µm) seawater. Chlorophyll a content was measured by extraction of chlorophyll from homogenized tissue in 100% acetone and measurement in duplicate by spectrophotometry (Jeffrey and Humphrey, 1975). Chlorophyll a extractions were not successful for M. capitata juveniles (n=7) because of their small size and presence of turf algae, and are not included in analysis. Chlorophyll a content was normalized to surface area (µg cm⁻²). Surface area of coral samples were measured using the wax dipping method (Veal et al., 2010). Prior to sampling, we collected photographs of all corals (n=57) with a black and white color standard and measured bleaching pigmentation score as a ratio to a white standard (Chow et al., 2016) in Fiji software v1.0 (Schindelin et al., 2012).

All data analysis was conducted in R statistical programming (v3.6.1, https://www.r-project.org/) and RStudio v1.2.5019 (https://rstudio.com/). Differences in depth of fluorescence, Symbiodiniaceae fluorescence and population densities between species on fixed coral tissue samples were analyzed by one-way ANOVA tests. Physiological characteristics of corals exposed to a bleaching stress measured by LSCM were analyzed with a linear mixed-effects model approach in the lme4 package in R (Bates et al., 2015). Depth of fluorescence and Symbiodiniaceae fluorescence were analyzed with a separate model for each species with treatment, time point and life stage as fixed effects and coral sample nested within colony as a random intercept. Post hoc analyses were performed using estimation marginal means tests in the emmeans package in R (https://cran.r-project.org/web/packages/emmeans/index.html). Owing to small sample sizes or violations of assumptions of analysis, the effect of treatment and life stage on comparative metrics for tissue thickness, chlorophyll a content and bleaching score were analyzed by non-parametric Kruskal–Wallis rank sum tests and Dunn post hoc tests. Assumptions of residual normality for all analyses were assessed with quantile–quantile plots, and the assumption of homogeneity of variance of residuals was confirmed with Bartlett's tests in the car package (Fox and Weisberg, 2011). Pearson and Spearman correlations (where assumption of normality was violated) were used to test for relationships between LSCM and comparative methodologies.

There was a significant effect of species on depth of fluorescence (P<0.01) and tissue thickness (P<0.01; Fig. 1A). Depth of fluorescence of P. acuta was less than that of other species (post hoc P<0.01), and P. varians depth was less than that of M. capitata (P<0.01; Fig. 1A). As measured on a dissecting microscope, tissue thickness of P. acuta was less than that of other species (post hoc P<0.01) with no significant differences between other species (post hoc P>0.05). Measures of depth of fluorescence and tissue thickness were not significantly different (post hoc P>0.05). The coefficients of variation (CV) for LSCM and comparative methods were similar at 44.6% and 46.3%, respectively. LSCM revealed a significant difference in depth of fluorescence between P. varians and P. compressa, whereas comparative measurement of tissue thickness did not (Fig. 1A). There was a significant positive correlation between measurements produced by both methods (r=0.82, P<0.01; Fig. 1B).

There was a significant effect of species on Symbiodiniaceae fluorescence (% relative intensity) measured by LSCM (P<0.01; Fig. 1C). Fluorescence of P. compressa was less than that of M. capitata and P. varians (post hoc P<0.01), whereas fluorescence of P. acuta was not significantly different from that of other species (post hoc P>0.05; Fig. 1C). There was a significant effect of species on population densities (P<0.01), with values in M. capitata and P. varians greater than those in P. acuta (P<0.05; Fig. 1D). The CV for Symbiodiniaceae fluorescence was less than that for population density at 33.0% and 39.9%, respectively. There was a significant positive correlation between these two metrics (r=0.54, P<0.01; Fig. 1E).

Physiological characteristics of corals throughout the bleaching experiment are summarized in Table S2. Depth of fluorescence of M. capitata was greater in adults (post hoc P<0.05) and lower at the final time point (post hoc P<0.01; Fig. 2A, Table 1). In P. acuta, there was a significant interaction between life stage, treatment and time point (P=0.03; Fig. 2A, Table 1). Symbiodiniaceae fluorescence of M. capitata was lower in high temperature (post hoc P<0.01) and lower at the final time point (post hoc P<0.01; Fig. 2B, Table 1). In P. acuta, fluorescence was higher in juvenile corals (post hoc P=0.048) and there was an interactive effect of treatment and time point (P<0.01; Fig. 2B, Table 1). LSCM images of adult P. acuta and M. capitata in ambient and high temperature are displayed in Fig. 2C.

Comparative metrics of corals are summarized in Table S3. Tissue thickness of M. capitata was greater in adults (post hoc P=0.01), but there was no effect of treatment (P>0.05; Table 1). Pocillopora acuta tissue thickness was not different between treatments or life stages (P>0.05; Table 1). The CV of tissue thickness was higher than that of depth of fluorescence at 49.1% and 37.2%, respectively. There was no significant effect of treatment on chlorophyll a content in adult M. capitata corals (P>0.05; Table 1). In P. acuta, chlorophyll a content was lower in high temperature (post hoc P<0.01), with no effect of life stage (P>0.05; Table 1). The CV of chlorophyll a content was greater than that of Symbiodiniaceae fluorescence measured by LSCM at 77.3% and 36.4%, respectively. There was a significant effect of treatment (P<0.01), but not life stage (P>0.05), on M. capitata bleaching pigmentation score (Table 1). Bleaching score was greater in high temperature in both M. capitata and P. acuta (post hoc P<0.01; Table 1), verifying that tissue paling occurred during the experiment. The CV of bleaching score was lower than that of Symbiodiniaceae fluorescence at 32.7% and 37.2%, respectively.

There was a positive correlation between depth of fluorescence and tissue thickness (r=0.89, P<0.01; Fig. 2D). Symbiodiniaceae fluorescence was significantly correlated to chlorophyll a content (r=0.67, P<0.01; Fig. 2E) and bleaching pigmentation score (r=−0.57, P<0.01; Fig. 2F).

LSCM provides proxy measures of tissue thickness and Symbiodiniaceae population densities and chlorophyll a content demonstrated by correlation with comparative methodologies widely used in the field of coral biology. In tissue samples of adult corals, LSCM was more sensitive to differences in tissue thickness between four species as it detected a difference between P. varians and P. compressa whereas comparative measurement did not. As expected from previous reports (Loya et al., 2001; Sudek et al., 2012; Yost et al., 2013), the imperforate-tissue species P. acuta had thinner tissues than the perforate-tissue species M. capitata and P. compressa. Previously, Yost and colleagues (2013) used LSCM to measure tissue depth in corals, in which they fixed and decalcified coral samples and measured tissue thickness by visualizing fluorescence on two-dimensional images of bisected tissues. Building upon this method, we demonstrate that depth of fluorescence in tissues in three-dimensional models eliminates the need for destructive sampling. We also compared Symbiodiniaceae fluorescence with population densities in these fixed tissue samples. These metrics were positively correlated, and were highest in M. capitata and P. varians and lowest in P. acuta and P. compressa. Symbiodiniaceae population densities were within ranges previously reported (Edmunds et al., 2014; Sudek et al., 2012). Previous studies have used LSCM to monitor symbiont infection and presence (for example, in Aiptasia sp.; Biquand et al., 2017; Matthews et al., 2016). LSCM methodologies presented in the present study offer the additional advantage of quantifying Symbiodiniaceae populations beyond presence and absence.

To test the utility of LSCM in monitoring coral physiological condition, we induced bleaching by exposing adult and juvenile M. capitata and P. acuta corals to a simulated thermal stress. In this experimental context, LSCM was well suited for characterizing tissue thickness as well as Symbiodiniaceae population densities and chlorophyll content across bleaching states, especially in small juvenile colonies. This method provides an advantage of simultaneous data collection without destructive sampling, resulting in higher available sample size as compared with comparative methods. We expanded our investigation of Symbiodiniaceae fluorescence to include comparison to tissue pigmentation and chlorophyll a content. Symbiodiniaceae intracellular chlorophyll content varies both spatially and temporally (Fitt et al., 2000), and it is expected that shifts in concentration of fluorescent photosynthetic pigments will influence LSCM measurements.

In general, P. acuta corals were more sensitive to high temperature than M. capitata. Depth of fluorescence, Symbiodiniaceae fluorescence, chlorophyll a content and pigmentation were lower at high temperature whereas only Symbiodiniaceae fluorescence and pigmentation were affected in M. capitata. With thin tissues and limited energetic stores, Pocillopora sp. are considered more thermally sensitive than species with thick tissues, such as M. capitata (Putnam et al., 2017). Life stage was an important influence on physiological condition in both species. In M. capitata, juvenile corals had thinner tissues and decreased depth of fluorescence than adults. Interestingly, Symbiodiniaceae fluorescence was higher in juveniles of P. acuta and juveniles retained greater depth of fluorescence in response to high temperature. This is indicative of higher resilience to bleaching in early life stages of P. acuta in the present study. Previous research has also documented elevated resilience of juveniles, although this is often species specific (Álvarez-Noriega et al., 2018). LSCM provided an advantage in characterizing M. capitata juveniles as our attempts to extract chlorophyll a failed as a result of their small size. The increased detection of the effects of life stage and treatment on physiological condition using LSCM may be due to higher sample size, ability to image small colonies and reduced variation compared with tissue thickness and chlorophyll a content.

During this experiment, we examined the relationships between depth of tissue fluorescence and tissue thickness as well as between Symbiodiniaceae fluorescence, coral pigmentation and chlorophyll a content. Depth of fluorescence was a valid proxy measure of tissue thickness as these metrics were strongly correlated and displayed higher values in M. capitata than in P. acuta, as expected (Loya et al., 2001; Yost et al., 2013). There were significant correlations between Symbiodiniaceae fluorescence and chlorophyll a content, as well as between fluorescence and tissue pigmentation. Although we are limited from attributing differences in Symbiodiniaceae fluorescence singularly to shifts in either population density or chlorophyll content, increases in both of these metrics are associated with higher potential for performance and provision of energy to the coral host (Roth, 2014). Future work using this method should acknowledge this limitation and avoid attributing shifts in fluorescence exclusively to either population density or chlorophyll content.

The application of LSCM methodology provides a new and useful tool for repeated measurements of living corals, tracking physiological states such as bleaching, and collecting measurements on juvenile colonies. The capacity of LSCM methods to detect significant shifts in physiological condition when comparative methods did not, suggests that it may be useful in detecting early indicators of physiological stress. A primary advantage of this method is the ability to measure key characteristics simultaneously without destructive sampling, allowing researchers to track individual corals over time and throughout physiological stress. Owing to the size constraints for scanning coral samples on the confocal microscope, this method is best suited for studies in which collecting biopsies of adult colonies is feasible, or for repeated measurement of young juvenile colonies settled on appropriately sized substrate in the laboratory. In addition, it is important to monitor expansion and contraction of coral tissues, as excessive movement could interfere with depth of fluorescence measurements. As LSCM is often cost-prohibitive, it is important to consider the experimental contexts in which it would be beneficial, and the trade-offs compared with other methods. In future work, LSCM methods can be utilized to characterize shifts in coral physiological condition across environmental conditions, particularly in early developmental stages.

We dedicate this manuscript to the legacy of our coauthor, friend and mentor, Dr Ruth Gates. Thank you to Pam Omidyar for supporting research at the Confocal Microscope Facility at the Hawai'i Institute of Marine Biology. We thank J. Hancock for assistance in experimental setup. We thank C. Drury and two anonymous reviewers for comments that greatly improved the manuscript.