Long distance dispersal and vertical gene flow in the Caribbean brooding coral Porites astreoides

Abstract

To date, most assessments of coral connectivity have emphasized long-distance horizontal dispersal of propagules from one shallow reef to another. The extent of vertical connectivity, however, remains largely understudied. Here, we used newly-developed and existing DNA microsatellite loci for the brooding coral Porites astreoides to assess patterns of horizontal and vertical connectivity in 590 colonies collected from three depth zones (≤10 m, 15-20 m and ≥25 m) at sites in Florida, Bermuda and the U.S. Virgin Islands (USVI). We also tested whether maternal transmission of algal symbionts (Symbiodinium spp.) might limit effective vertical connectivity. Overall, shallow P. astreoides exhibited high gene flow between Florida and USVI, but limited gene flow between these locations and Bermuda. In contrast, there was significant genetic differentiation by depth in Florida (Upper Keys, Lower Keys and Dry Tortugas), but not in Bermuda or USVI, despite strong patterns of depth zonation in algal symbionts at two of these locations. Together, these findings suggest that P. astreoides is effective at dispersing both horizontally and vertically despite its brooding reproductive mode and maternal transmission of algal symbionts. In addition, these findings might help explain the ecological success reported for P. astreoides in the Caribbean in recent decades.

Figure 1. Porites astreoides population structure across regions [Upper Keys, Lower Keys and Dry Tortugas (within Florida), Bermuda and the U.S. Virgin Islands] and depths [shallow (≤10 m), mid (15–20 m) and deep (≥25 m)].

Bar graphs show the average probability of membership (y-axis) of individuals (N = 590, x-axis) in K = 2 to K = 4 clusters (shown in ascending order) as identified by STRUCTURE. Samples were arranged in order of increasing depth within region.

Figure 2

Mean log-likelihood of K (a) and Delta K (b) values for STRUCTURE analysis of Porites astreoides samples. Values of K = 1–20 were tested by running 3 replicate simulations for each K (error bars in upper figure indicate variance among replicates).

Figure 3. Isolation-by-distance patterns in Porites astreoides.

Geographic distance explained 17% of the variation in genetic distance (F_ST) across all sampling sites (R² = 0.17, P < 0.01, Fig. a), 39% of the variation in genetic distance when the U.S. Virgin Islands sites were excluded (R² = 0.39, P < 0.01, Fig. b), and none of the variation when Bermuda sites were excluded (R² = 0.01, P > 0.05, Fig. c).

Figure 4. Principal Component Analysis (PCA) of allele frequency covariance in Porites astreoides populations.

14 of 79 axes were retained, explaining 100% of the cumulative variance. Plotted are the first and second axes explaining 38.59% (P < 0.01) and 21.28% (P < 0.05) of the variance, respectively. Axes cross at 0. The different shapes denote each of the 3 geographic locations sampled in this study (Florida, Bermuda and U.S. Virgin Islands), whereas different colors denote each of the 3 depths under comparison [shallow (≤10 m), mid (15–20 m) and deep (≥25 m)]. UK = Upper Keys, LK = Lower Keys, DT = Dry Tortugas, Bermuda = Bermuda and USVI = U.S. Virgin Islands.

Figure 5. Porites astreoides population structure by color morph (yellow/green or brown).

Bar graphs show the average probability of membership (y-axis) of individuals (N = 200, x-axis) in K = 3 clusters as identified by STRUCTURE.

Figure 6

Symbiodinium types detected in a subset of Porites astreoides corals from shallow (≤10 m), intermediate (15–20 m) or deep (≥25 m) depths, using denaturing gradient gel electrophoresis (a) versus high-sensitivity quantitative PCR (b). In (b), potential immigrants or individuals with immigrant ancestors were identified in STRUCTURE as having a probability of membership >0.90 to the deep cluster [denoted as ‘shallow (deep origin)’], or as having a probability of membership >0.90 to the shallow cluster [denoted as ‘deep (shallow origin)’]. No shallow individuals of deep-water origin were found in the U.S. Virgin Islands (denoted by nd) and no samples from Bermuda where included due to absence of patterns of depth zonation at this location. Numbers in bars indicate number of colonies assessed. USVI = U.S. Virgin Islands.