Bacterioplankton drawdown of coral mass-spawned organic matter

Abstract

Coral reef ecosystems are highly sensitive to microbial activities that result from dissolved organic matter (DOM) enrichment of their surrounding seawater. However, the response to particulate organic matter (POM) enrichment is less studied. In a microcosm experiment, we tested the response of bacterioplankton to a pulse of POM from the mass-spawning of Orbicella franksi coral off the Caribbean coast of Panama. Particulate organic carbon (POC), a proxy measurement for POM, increased by 40-fold in seawater samples collected during spawning; 68% degraded within 66 h. The elevation of multiple hydrolases presumably solubilized the spawn-derived POM into DOM. A carbon budget constructed for the 275 µM of degraded POC showed negligible change to the concentration of dissolved organic carbon (DOC), indicating that the DOM was readily utilized. Fourier transform ion cyclotron resonance mass spectrometry shows that the DOM pool became enriched with heteroatom-containing molecules, a trend that suggests microbial alteration of organic matter. Our sensitivity analysis demonstrates that bacterial carbon demand could have accounted for a large proportion of the POC degradation. Further, using bromodeoxyuridine immunocapture in combination with 454 pyrosequencing of the 16S ribosomal RNA gene, we surmise that actively growing bacterial groups were the primary degraders. We conclude that coral gametes are highly labile to bacteria and that such large capacity for bacterial degradation and alteration of organic matter has implications for coral reef health and coastal marine biogeochemistry.

Fig. 1

Microcosm sampling. Spawn and Non-spawn microcosm experiments were independently conducted in triplicate 10-liter polycarbonate carboys in the dark. Samples were taken at all time points for bacterial production, microbial abundance, and enzyme activities; the 0 and 66 h time points were also sampled for 16S rRNA sequencing, FT-ICR MS analysis, organic carbon and nitrogen concentration, and inorganic nutrient concentration. Carboys were well mixed by 5-times inversion before sampling with sterile-glass serological pipettes or a peristaltic pump. *One sampling was conducted at the 0 h time point; an 8-liter aliquot from the field sampling carboy was transferred into carboy “X” and then sampled to ensure that the seawater had been exposed to the same container composition and surface area

Fig. 2

a Total organic carbon (TOC) concentration in microcosms Spawn and Non-spawn. TOC = particulate organic carbon (POC) + dissolved organic carbon (DOC) + colloidal carbon. For each fraction the error is represented by the mean ± SD of technical replicates (spawn_0h, n = 2) or by single measurements of multiple biological replicates (Spawn_66h, n = 3; Non-spawn_0h, n = 2; Non-spawn_66h, n = 6). Asterisks indicate significant difference in TOC drawdown (TOC_0h–TOC_66h) between treatments: ****p < 0.0001. b Bacterial production. Error bars represent ± SD of the mean (Spawn_0h, n = 3; Spawn_66h, n = 6; Non-spawn_0h, n = 6; Non-spawn_66h, n = 18). Asterisks indicate significant difference between treatments for each time point: *p < 0.05; **p < 0.01; ****p < 0.0001. c A range of bacterial carbon demand (BCD) (right y-axis) was estimated for microcosm Spawn on the basis of integrated BP rate measurements and a hypothetical range (5–30%) of bacterial growth efficiency. BCD = BP + bacterial respiration (BR). The model indicates that a broad range of BR and non-bacterial respiration (NBR) could have satisfied the observed TOC drawdown in microcosm Spawn (left y-axis)

Fig. 3

Bacterial abundance (a) and virus-like-particle (VLP) abundance (b) in microcosm Spawn (black) and Non-spawn (gray) samples. At least 20 fields of view were enumerated for each microscopy sample (Spawn_0h, n = 1; Spawn_66h, n = 3; Non-spawn_0h, n = 2; Non-spawn_0h, n = 6). Error bars represent the standard error of the mean. c Virus-to-bacteria ratio (VBR). Error is represented by ± SD of the mean. Asterisks indicate significant difference between treatments for each time point: ****p < 0.0001. Insufficient replication precluded statistical evaluation between treatments at the 0 h time point for data in (a–c)

Fig. 4

Hydrolysis rates of enzyme substrates for microcosm Spawn samples (black) and Non-spawn samples (gray). a Protease; b Lipase; c Alkaline phosphatase; d Chitinase; e α-glucosidase; and f β-glucosidase. Error bars represent ± SD of the mean (Spawn_0h, n = 2; Spawn_24,44,66h, n = 6; Non-spawn_0h, n = 4; Non-spawn_24,44,66h, n = 12). Asterisks indicate significant difference between treatments for each time point: *p < 0.05; **p < 0.01; ***p < 0.001; ****p < 0.0001. Insufficient replication precluded statistical evaluation between treatments for the 0 h time point

Fig. 5

Actively growing bacterial community analysis. Spawn_0h (a) and Spawn_66h (b) show the number of taxa and summed mean relative abundances of OTUs. T Total DNA, B BrdU-labeled DNA, T&B total & BrdU-labeled DNA. c Mean relative abundance of BrdU-labeled taxa in microcosm Spawn samples (Spawn_0h, n = 1; Spawn_66h, n = 3). BrdU-labeled OTUs that have ≥0.1% relative abundance are shown, and taxa from Spawn_66h samples had to be found in at least two of the three replicates. Class designations: α, Alphaproteobacteria; γ, Gammaproteobacteria

Fig. 6

FT-ICR MS analysis of dissolved organic matter (see Table S1 for sample details; Table S2 for data summary). a Cluster dendrogram based on the relative spectral peak-heights of all molecular formulas (MFs) identified in microcosm samples. Distance calculated by Bray–Curtis measure and the complete-linkage method. b Van Krevelen diagram showing identified MFs in microcosm Spawn samples (Spawn_0h, n = 1; Spawn_66h, n = 3). Black crosses (n = 82), MFs unique to microcosm Spawn_0h; black dots (n = 532), MFs unique to Spawn_66h. Colored dots (n = 3 007), MFs ubiquitously identified in both Spawn_0h and Spawn_66h samples (all replicates). Color scale represents relative change from Spawn_0h to Spawn_66h, derived by calculating relative peak-height ratios for each molecular formula (Spawn_66h/(Spawn_66h + Spawn_0h)). Ratios as “percent relative change” are shown: red, 20% relative increase; gray, no change; blue, 20% relative decrease. Black crosshair, weighted-mean element ratio of all MFs detected in microcosm Spawn_66h samples (n = 16,907); gray crosshair, weighted-mean element ratio of the unique MFs (n = 532) in the same samples. S, microcosm Spawn.