A preliminary survey of zoantharian endosymbionts shows high genetic variation over small geographic scales on Okinawa-jima Island, Japan

Introduction

Symbiodinium is an important genus of dinoflagellate photosymbionts found in tropical and subtropical marine ecosystems. These “zooxanthellae” transfer energy to their invertebrate hosts in nutrient-poor environments, enhancing the growth of reef-building corals and other reef organisms such as zoantharians (Muscatine & Cernichiari, 1969; Baker, 2003). They also serve a key role in establishing the thermal tolerance of coral colonies and shaping the adaptive response of reef organisms to climate change (Sampayo et al., 2008; Thornhill et al., 2014). As climate change intensifies and diversity patterns are expected to alter, it is increasingly important to map marine species distributions on scales both large (McClanahan et al., 2014; Reimer et al., 2017) and small (Mieszkowska & Lundquist, 2011).

The biogeography of Symbiodinium is determined by many factors, including host species distributions and environmental parameters (LaJeunesse et al., 2004). To determine Symbiodinium biogeography, it is critical to use DNA markers with the resolving power to delineate between and among species (LaJeunesse & Thornhill, 2011). Our ability to discern different lineages of Symbiodinium has improved in tandem with the resolution of the molecular markers used to identify them. Originally, highly divergent lineages or “clades” designated A–I were confirmed via the sequencing of 18S and 28S ribosomal DNA (Rowan & Powers, 1991; Pochon & Gates, 2010), followed by the delineation of numerous subclades via phylogenetic analyses of the internal transcribed spacer regions (e.g., internal transcribed spacer 2 (ITS2); LaJeunesse, 2001). More recently, sequences from domain V of the chloroplast large subunit (cp23S) ribosomal DNA (Santos, Gutierrez-Rodriguez & Coffroth, 2003; Kirk et al., 2009; LaJeunesse & Thornhill, 2011; LaJeunesse, Parkinson & Reimer, 2012) and the non-coding region of the plastid minicircle (psbA^ncr; Takishita et al., 2003; LaJeunesse & Thornhill, 2011) have provided even greater resolution, often corresponding to the species level (LaJeunesse, Parkinson & Reimer, 2012). Microsatellite markers have also been developed to investigate fine-scale diversity in Symbiodinium (Pettay & LaJeunesse, 2007; Grupstra et al., 2017).

As a result of the development and implementation of finer resolution molecular markers, it is apparent that the extent of Symbiodinium diversity may be much greater than previously recognized based on studies using only internal transcribed spacer regions. For example, in examinations of Symbiodinium within the common reef zoantharian Palythoa tuberculosa (Esper, 1805) across a latitudinal gradient of 800 km in the Northern Red Sea, psbA^ncr sequences revealed up to four unique lineages, despite ITS2 results showing only one single “subclade” lineage (Reimer et al., 2017). The distribution of these individual psbA^ncr lineages strongly correlated with sea surface temperature (SST) differences of approximately 1 °C. Such results demonstrate a need to re-examine previously reported Symbiodinium diversity with high-resolution markers, as fine scale, ecologically important differences may have been missed.

In biogeography, isolation by distance (IBD) generally refers to how genetic differences between individuals or populations increase with increasing geographical distance due to limitations on dispersal. In Symbiodinium species, other factors in addition to IBD that may act as drivers of evolution include host species associations (LaJeunesse et al., 2004), UV light levels (correlates of turbidity or chlorophyll-a (chl-a) concentrations; LaJeunesse et al., 2010; Tonk et al., 2014), and ocean temperatures (LaJeunesse et al., 2010; Tonk et al., 2014). Correspondingly, differences in Symbiodinium associations have been noted over large oceanic or latitudinal gradients (LaJeunesse & Trench, 2000; LaJeunesse et al., 2004; Reimer et al., 2017), within different host species (Frade et al., 2008a; Thornhill et al., 2014), at different light levels (Frade et al., 2008b; Sampayo et al., 2008; Kamezaki et al., 2013), and in areas of extremely high (Hume et al., 2013) or low ocean temperatures (Chen et al., 2003).

Okinawa-jima Island is ∼1,200 km² in area, just over 100 km in length, and 3–30 km in width. Although it is not “large” when compared to oceanic or regional scales, the surrounding marine environment encompasses a variety of ecosystems including large bays, muddy tidal flats, patch reefs within lagoons, and fringing reefs (Fujita et al., 2015). Additionally, recent population genetic studies on scleractinian corals (Shinzato et al., 2015), sea cucumbers (Soliman, Fernandez-Silva & Reimer, 2016), and amphipods (White, Reimer & Lorion, 2016) have described unexpected genetic structure among locations around Okinawa-jima Island, particularly between the Kuroshio-influenced west coast and the more isolated east coast.

Based on ITS2 sequence analyses of Symbiodinium within P. tuberculosa in Southern Japan (including Okinawa), we previously reported that subclade C1 or closely related types are dominant (Reimer, Takishita & Maruyama, 2006). Here we use psbA^ncr sequences to re-examine how much variation exists within Symbiodinium C1 associating with P. tuberculosa, specifically focusing on Okinawa-jima Island’s shallow coral reef environments. We additionally explore potential associations between observed diversity and the environment. This study is intended to complement our recent work in the Red Sea (Reimer et al., 2017) by focusing on the same host species and describing the extent of symbiont diversity on a much smaller geographical and environmental scale.

Materials and Methods

Results

Phylogenetic analyses

Two different Symbiodinium ITS2 types were detected. The first type (n = 17) matched 100% with previously reported Symbiodinium subclade C1 from P. tuberculosa in Southern Japan (DQ889743; DQ889741) (Fig. 1A). The other type (n = 5) differed by one base pair and was also 100% identical to previously reported Symbiodinium from P. tuberculosa in Southern Japan (DQ480639). This second type formed a subclade within C1 (ML = 99%, NJ = 74%, MP = 61%, BI = 0.90) (Fig. 1A). These two types formed a large, moderately supported clade separate from subclade C3 (ML = 69%, NJ = 83%, MP = 64%, BI = 0.70), so we considered all of our sequences to be subclade C1 or “C1-related” (Fig. 1A; Table S1).

Sixty-three unique Symbiodinium psbA^ncr forward sequences were recovered. Two sequences were too short to be included in the final alignment but were long enough to identify to lineage (described below, Table S1). The resulting psbA^ncr ML tree showed 33 specimens within a large, well-supported clade (ML = 100%, NJ = 100%, MP = 100%, BI = 1.00) that we designated as “lineage 1” (Fig. 1B). Specimens belonging to lineage 1 (n = 33) were recovered from all nine locations. Another 19 specimens from Oku, Teniya, and Mizugama formed a separate monophyly (“lineage 2”; ML = 76%, NJ = 93%, MP = 100%, BI = 0.98). Additionally, seven specimens from Wase, Nerome, Bise, Mizugama, and Odo formed a monophyly (“lineage 3”; ML = 75%, NJ = 100%, MP = 100%, BI = 0.77). Finally, two specimens from Bise and Nerome formed another monophyly (“lineage 4”; ML = 100%, NJ = 100%, MP = 100%, BI = 1.00). The between-lineage molecular distances for psbA^ncr ranged from 0.105 to 0.256, while the within-lineage distances were much smaller, ranging from 0.003 to 0.021 (Table 2).

Table 2:

Pairwise genetic distances among Symbiodinium psbA^ncr lineages isolated from P. tuberculosa in Southern Japan.

	Lineage 1	Lineage 2	Lineage 3	Lineage 4*
Lineage 1	0.003
Lineage 2	0.137	0.005
Lineage 3	0.184	0.128	0.021
Lineage 4*	0.256	0.198	0.105	0.000*

DOI: 10.7717/peerj.3740/table-2

Notes:

Shaded diagonal values represent within-lineage distances.

* Lineage 4 was represented by two identical sequences.

Lineage distributions

We next examined the distribution of Symbiodinium lineages across locations (Fig. 2). Wase, Amami Oshima Island was dominated by lineage 1 (4/5 P. tuberculosa colonies), as was Bise (9/11), Uken (3/3), Kyan (4/4), and Odo (6/7). Lineage 2 was dominant at Oku (8/10), Teniya (5/7), and Mizugama (6/8), while lineage 3 was dominant at Nerome (5/8) and also appeared at Bise (2/7). Lineage 4 members only appeared in one colony each at Wase, Mizugama, and Odo.

Finally, we examined the range of environments in which each lineage could be found (Fig. 3). Symbiodinium lineage 1 appeared at all sites and thus all environments in this study (SST^avg = 24.6–25.35 °C; SST^max = 29.43 °C; SST^min = 20.35 °C; chl-a 0.08–0.50 mg/m³). Lineage 2 was only observed at Oku, Teniya, and Mizugama, where SST^avg ranged between 24.95 and 25.15 °C, with SST^min of 20.69 °C (Oku) and SST^max of 29.16 °C (Teniya, Mizugama), and chl-a ranged between 0.08 and 0.25 mg/m³. Lineage 3 was only found at Bise and Nerome, with SST^avg of 24.85–24.95 °C, SST^min of 20.47 °C, SST^max of 29.33 °C (both Nerome), and chl-a of 0.15–0.30 mg/m³. Lineage 4 members were only observed once each at three locations, but these stretched across the geographic range of this study (Wase, Mizugama, Odo).

Discussion

Using the high-resolution psbA^ncr marker, we identify four Symbiodinium lineages associated with P. tuberculosa on Okinawa-jima Island. The lineages feature surprisingly unique distributions over a small geographic scale, and would be considered at most two entities based on lower-resolution ITS2 data (Fig. 1). Because the between-lineage molecular distances for psbA^ncr (0.105–0.256; Table 2) are greater than those reported between the psbA^ncr sequences of two divergent Symbiodinium ITS2 types (∼0.045 for C26a vs. C31; LaJeunesse & Thornhill, 2011), these lineages likely represent reproductively isolated species rather than populations within a species.

Lineage 1 appears to be a widely distributed generalist, at least over the range of this study in the Central Ryukyus (Fig. 2). It is found at all sites and often occupies the majority of colonies at a given site. Lineage 3 is observed only at the two sites on the northwestern coast of Okinawa-jima Island, and thus appears to have narrow geographical and environmental components to its distribution (SST^avg = 24.85–24.95 °C; chl-a = 0.15–0.30 mg/m³) (Fig. 3). Lineage 4 is present in very low numbers (n = 3) across the latitudinal range of the study, making it difficult to infer its environmental niche.

Lineage 2 has a somewhat restricted range, found at only three locations (where it also occupied the majority of colonies): Oku, Teniya, and Mizugama. These three sites are located near river outflows, suggesting this Symbiodinium lineage may be able to tolerate changes in salinity more effectively than the others. Unfortunately, fine-scale salinity data are not yet available to investigate this trend further. Chl-a levels at these sites are generally low (<0.25 mg/m³), while the SST^avg range is intermediate (24.95–25.15 °C). Lineage 2 is absent at other locations within this SST^avg range; for example, at Bise and Nerome, where only lineages 1 and 3 are present, and at Uken, where only lineage 1 is detected. Although Uken’s SST^avg is firmly in the middle of the SST range investigated in this study, Uken’s SST^min and SST^max are generally more extreme than those at other sites (Table 1) due to Uken’s position within shallow Southern Kin Bay, which is isolated from the stabilizing temperature effects of the open ocean (Montani, 1996). Additionally, chl-a levels at Uken are higher than those of all other locations (0.50 mg/m³).

By focusing on a very small area of the Northwest Pacific, and by using a rapidly evolving molecular marker, we could resolve a much finer scale of Symbiodinium biogeography than has previously been recognized in the region. Some psbA^ncr lineages appear to be partitioned on the basis of temperature (SST^max) differences on the scale of 0.1–0.3 °C, much lower than the 0.7–1.0 °C observed in previous studies (Baums, Devlin-Durante & LaJeunesse, 2014; Reimer et al., 2017). It also appears SST stability (e.g., differences between SST^max and SST^min), as well as fine scale salinity differences (not measured here) may play an important role in the distribution of Symbiodinium diversity. Chl-a levels are generally low at all locations, although it should be noted that only generalist lineage 1 is found in Uken (although n = 3), the site with the highest chl-a levels. Overall, the contribution of chl-a (a proxy for turbidity) is not clear, as all P. tuberculosa specimens were collected from very shallow waters (0–2 m).

It is surprising that such biogeographic patterns could be uncovered given the low numbers of host P. tuberculosa at some locations. However, these low numbers also caution that the symbiont species’ distributions are unlikely to have been completely resolved. Another issue is that the fine-scale environmental variation among sites falls below the margin of error in the satellite datasets; future research of this nature must utilize more precise instrumentation. We did not examine host P. tuberculosa population genetics, and it remains to be seen if fine-scale host structure may also play a role in the observed patterns. Previous studies indicate cnidarian host and symbiont genetic structure can be mismatched (Baums, Devlin-Durante & LaJeunesse, 2014; Leydet & Hellberg, 2016), although clear cases of matching genetic structure have also been observed (Bongaerts et al., 2010; Prada et al., 2014).

As our ability to discern between different lineages of Symbiodinium has increased, so too has our understanding of the complexity and nuances of Symbiodinium diversity and distribution. The present work supports previous studies that show Symbiodinium evolution is driven largely by specialization to different environmental niches (Frade et al., 2008b; LaJeunesse et al., 2010; Kamezaki et al., 2013; Reimer et al., 2017), and that specialization may occur on much finer micro-environmental scales than usually addressed. Further characterization of the Symbiodinium–P. tuberculosa symbiosis within the Ryukyus or comparable island chains should be carried out to confirm similar symbiont structuring based on fine-scale environmental heterogeneity.

The Symbiodinium diversity patterns on Okinawa-jima Island highlight three major considerations for future investigations of this kind. First, it is advantageous for researchers to obtain fine-scale environmental data of the study area if available, so as to better characterize niches that might otherwise be deemed homogenous. Second, researchers should refrain from considering pooled specimens or results from different nearby locations as representative of a larger area without thoroughly checking for fine scale environmental patterns. Third, as has been suggested elsewhere (LaJeunesse & Thornhill, 2011; Reimer et al., 2017), researchers should incorporate both ITS2 data (to tie to past work) as well as rapidly evolving markers (to uncover hidden diversity) when investigating Symbiodinium biogeography, as a failure to address fine-scale niche adaptation could lead to a misinterpretation of results. These suggestions should improve the design of Symbiodinium studies on all geographic scales, from local to regional to global.

Supplemental Information