Cryptic diversity and species distributions
Assessing biodiversity is essential to understand evolutionary and ecological processes and to observe changes in the distribution of species as a consequence of climate change and other anthropogenic stressors. A fundamental unit of biodiversity are species, though intraspecific genetic diversity is also considered a crucial factor (Błażewicz, Jóźwiak, Menot, & Pabis, 2019; Laikre, 2010; Brix et al., 2020). Delimiting species can be challenging in cases where differences between putative species are small and intraspecific variability and interspecific variation are not clearly demarcated (Kaiser et al., 2021). In the case of Haploniscus bicuspis , several genetically divergent lineages were recovered, of which some probably constitute separate, morphologically cryptic species. Commonly observed intraspecific COI distances within deep-sea isopods were usually below 6% (e.g., Brix, Riehl, & Leese, 2011; Brix et al., 2018a, 2018b; Havermans et al., 2013; Brandt, Brix, Held, & Kihara, 2014; Bober, Riehl, Henne, & Brandt, 2018), but intraspecific distances of more than 8% have been reported (e.g., Riehl, Lins, & Brandt, 2018). Conversely, interspecific COI distances as low as 4.6% were reported (Riehl et al., 2018), possibly suggesting that some of the larger intraspecific distances may be due to cryptic diversity, but commonly exceed 10% (e.g., Havermans et al., 2013; Brandt et al., 2014; Brix et al., 2018a, 2018b). In the following, we will discuss whether the five lineages might constitute different species.
Lineages I, II and III (ICOI-IIICOI and IRAD-IIIRAD) are consistently delimited from lineages IV and V in COI and nuclear loci, and with a few exceptions also in the proteomic data. However, these three lineages are not consistently delimited from each other. There is a clear discordance between the mitochondrial and nuclear data, with individuals being assigned to different lineages (among I, II and III) based on the respective analyses. Our data also strongly suggests multiple instances of inter-lineage hybridization and gene flow most notable at station 1219 (see Figure 1) where all three lineages appear to have hybridized. The genetic distances in COI among these lineages are also smaller than among most other deep-sea isopods. We therefore suggest that these three lineages constitute a single species (referred to as H. bicuspisI-III in the following).
Lineages IV (IVCOI and IVRAD) and V (VCOI and VRAD) are consistently differentiated from each other and from H. bicuspis I-III in COI and nuclear loci. COI p -distances between lineage V and all other lineages exceed commonly observed intraspecific distances within deep-sea isopod species. For lineage IV, COI p -distances are lower in comparison to H. bicuspis I-III than commonly observed among deep-sea isopods. Furthermore, the p -distances are comparable to those observed within H. bicuspis I-III. However, in the nuclear loci, lineage IV is the most divergent. No instances of hybridization have been inferred for lineages IV or V. Distribution modelling suggests that their habitats are ecologically differentiated from those of H. bicuspis I-III and at least partially differentiated from each other. We therefore tentatively suggest that lineage IV and V constitute two morphologically cryptic species withinH. bicuspis (termed H. bicuspis IV and H. bicuspisV in the following).
Similar patterns of cryptic diversity have been observed for other Icelandic Isopoda, which were previously assumed to be widespread species (Brix et al., 2014; Jennings, Brix, Bober, Svavarsson, & Driskell, 2018). The geographic distribution of the different putativeH. bicuspis species fit well into the distribution patterns of other benthic deep-sea isopods around Iceland (Brix et al., 2018). Our discovery that the waters around Iceland alone may be inhabited by three cryptic species within H. bicuspis strongly questions the occurrence of H. bicuspis in the South Atlantic Ocean (Brökeland and Wägele, 2004; Menzies, 1962). The presence of further cryptic diversity in this presumably widespread species is highly likely. We deliberately did not perform molecular clock analyses to date the differentiation of the herein observed species, because evolutionary rates for deep-sea crustaceans are still lacking. A recent study suggests that the commonly applied rate of 1.4% per million years (Knowlton & Weigt, 1998) is too conservative and rather suggest a K2P divergence of 5-5.2% per million years for arctic marine invertebrates (Loeza-Quintana et al., 2019). If such rates are applicable to deep-sea benthic peracarids, such as Haploniscus , the divergence between the three putative species started roughly half a million to a million years ago. This is relatively young and suggests recent and putatively ongoing speciation processes.
The distribution of the three species appears to be largely governed by water masses and associated ecological parameters. In particular, the GIF Ridge represents a crucial barrier separating species. Whether the GIF Ridge is a physical barrier that hinders dispersal or just separates water masses and thereby shapes species distributions remains unanswered. The GIF Ridge has been discussed as an isolation barrier for many isopod species in the North Atlantic, for instance anthuridean isopods (Negoescu & Svavarsson, 1997), valviferan isopods (Stransky & Svavarsson, 2006), and desmosomatid and nannoniscoid isopods (Brix & Svavarsson, 2010). Haploniscus bicuspis sampled in this study were found between 316 m depth at station 1136 and 2422 m depth at station 1172, therefore they can easily cross the saddle depth of the GIF Ridge. Previously, H. bicuspis have been collected in the Arctic, even in shallow waters of 198 m depth (Brandt, 1997).
Ocean temperature is highly variable around Iceland. North of the GIF Ridge, bottom water temperature can be as low as -0.9 °C, and it only becomes warmer as it reaches the shelf (up to about 3 °C, Jochumsen et al., 2016). In contrast, south of the GIF Ridge, the North Atlantic Water can reach up to 10.5 °C (Hansen & Østerhus, 2000). Another environmental variable that may limit species distribution is the sediment structure, which was previously found to be important in peracarid crustacean distribution around Iceland (Stransky & Svavarsson, 2010). Ostmann et al. (2014) found that sediment characteristics vary in the surrounding waters of Iceland, with coarser sand found around the Reykjanes Ridge and more silt and clay found in the deep-sea east of the GIF Ridge. Oxygen concentration may also influence migration and gene flow between isopods, especially in the deep sea. Expanding oxygen minimum zones may have contributed to allopatric speciation in the past (White, 1988).
The distribution models provide interesting hypotheses that should be tested in future studies. These include the hypothesized distribution ofH. bicuspis IV (and partly V) along the Norwegian coast, the occurrence of H. bicuspis I-III south of the GIF Ridge off the coast of Greenland, and whether H. bicuspis IV, V, or both occur in the abyssal plains south of Iceland. The former would require dispersal along a narrow habitable corridor along the north-west of Great Britain, whereas the expansion of H. bicuspis I-III south of the GIF Ridge would require crossing a region with unfavorable ecological conditions. This could have been mediated by transport via the north to south overflow of arctic deep-water. Whether H. bicuspis IV, V, or both occur on the abyssal plains of the North Atlantic south of Iceland cannot be answered with the available data. The ecological conditions appear suitable for both (possibly slightly better for IV), and sympatric occurrences of these two species were already observed for the Reykjanes Ridge. By examining additional populations of H. bicuspis genetically, future studies may shed light on these questions.
A comparison of proteomic profiles revealed differences between specimens from the northern (I-III) and southern (IV and V) species, and also between the two southern species IV and V. However, differentiation between species IV and V was less distinct. Differences between species are very small and occur mainly in a few relatively low-expressed proteins, which show differences in presence and relative intensities. Similarly, high similarity in proteomic patterns and only a small number of differentially expressed proteins were observed in cryptic mosquito species (Müller et al., 2013; Dieme et al., 2014). Like H. bicuspis, peaks in the mosquito Anopheles gambiae Giles, 1902 species complex were shifted by only a few Daltons, potentially reflecting few amino acid substitutions within proteins or minor post-transcriptional modifications (Müller et al., 2013). These small shifts in masses of specific proteins may thus infer polymorphisms of these molecules in H. bicuspis . Differences between the putative species were at least partly caused by a prevalence of a certain allele in the respective putative species. Despite apparent gene flow interruptions, differences between IV and V were less pronounced. Proteomic profiles depend on physiological responses to the environment, e.g., variations in proteomic profiles varied with season and habitat in ticks (Karger, Bettin, Gethmann, & Klaus, 2019). Thus, it remains unanswered whether the higher similarity of putative species IV and V results from very similar environmental impacts and thus comparable selection pressures in the past causing a similar physiological response. We interpret the high similarity between the differentH. bicuspis species found in this study as a reflection of recent or ongoing speciation processes, as reported in fishes (Takács et al., 2014, Maasz et al., 2020). It is likely that young species are both morphologically and physiologically cryptic, revealing the limit of proteomic fingerprinting for species identification in such instances.
Haploniscus bicuspis was first described from the Norwegian Sea (Sars, 1877) north of the GIF Ridge and a few hundred kilometers east of our sampled area, but no genetic information is currently available for Norwegian representatives. Given the species’ distribution around Iceland, we suggest that H. bicuspis I-III represents the trueHaplonicus bicuspis (Sars, 1877) because the known and modelled distribution of H. bicuspis I-III extends well into the waters east of Iceland towards the Norwegian Sea and would correspond to the subspecies H. b. bicuspis described by Wolf (1962). However, Wolff differentiated the two subspecies based on the morphology of pleopod I, and both variants were found in H. bicuspis I-III (unfortunately no adult males were available for the other two putative species). In juvenile males the respective section of the male pleopod I is smooth and narrow and widens as they mature into preparatory males (see also Wolff 1962). It is plausible that the different pleopod I shapes assigned to the two subspecies are in fact successive developmental stages, with the distinct corner ascribed to H. b. bicuspis being the later stage. Even though pleopod I morphology might not be useful to distinguish Wolff’s subspecies, H. b. tepidusmight still represent a valid species (H. tepidus ). Whether this corresponds to H. bicuspis IV or H. bicuspis V cannot be answered with the data currently available, as H. b. tepidus was described from the Reykjanes Ridge.