Evaluation of cryptic invasion in Japanese Undaria populations based on mitochondrial haplotypic analysis

1 Introduction

Numerous examples of intraspecific genetic diversity have been reported, and their significance in conservation and utilization have been well-documented (DeWoody et al. 2021; Swarup et al. 2021). Species without long dispersal ability tend to show clear genetic structure among regions; numerous examples have been reported in various lineages, including seaweeds (Hu and Fraser 2016). Human activities, however, sometimes promote long distance dispersal beyond the actual dispersal capacity of a given species. Indeed, examples of such anthropogenic introductions beyond the native species range have been observed for various lineages (Bonanno and Orlando-Bonaca 2019). Colonization of a non-native species from other regions, countries, or continents is usually detected as a new record of the species. On the other hand, cryptic invasion, which is an invasion by a species morphologically similar to a native one or invasion by a conspecific lineage, is often undetectable by morphology (Canavan et al. 2020; Morais and Reichard 2018). Cryptic invasion, especially that by a conspecific lineage genetically distinct from a native one but within a species range, generally occurs due to shared niche and geographical proximity between the introduced and recipient populations. The red algal species Eucheuma (Tano et al. 2015) is one of the well-studied examples in coastal environments (Morais and Reichard 2018). Considering the vast amounts of data on the genetic structure in seaweed species (Hu and Fraser 2016, and references therein) and the increasing demands on seaweed mariculture as a measure for carbon fixation, additional studies of the risk of cryptic invasion through seaweed mariculture are required.

Undaria pinnatifida (Harvey) Suringar, a brown algal species of the order Laminariales, has a heteromorphic life cycle with macroscopic sporophytes of 1–3 m in height (Lee and Yoon 1998). This species is well known as an edible seaweed within its native range (Japan, Korea and China, Yamanaka and Akiyama 1993) but, in some countries outside its native range, it is also known as an invasive species (e.g., Bonanno and Orlando-Bonaca 2019; Castric-Fay et al. 1993; Dellatorre et al. 2014; Hay and Luckens 1987; Sanderson 1990). Undaria populations, especially those on the Japanese coasts, show clear geographic structure both in mitochondrial (Uwai et al. 2006) and nuclear markers (Uwai et al. 2023). Mitochondrial haplotypes based on DNA sequences of mitochondrial cox3 and tRNA genes (a region including tTrp, tIle, and tGln and spacers between these genes, previously named “tatC-tLeu”) are clearly different by regions; several mitochondrial lineages (haplogroups) have been recognized in association with their geographic distributions, such as the northern Japan haplogroup, central Japan haplogroup and Sea of Japan haplogroup (Uwai et al. 2006). Chinese populations also have a unique mitochondrial haplogroup (the continental Asia haplogroup), although Chinese populations also include the northern Japan haplotypes (Shan et al. 2022).

Since the 1960s, Undaria mariculture has been conducted in the native range (e.g., Niwa et al. 2017; Shan et al. 2018); principally, mariculture is done by growing young sporophytes (seedlings) on cultivation rafts in the open sea (e.g., Shan et al. 2018). Seedlings are formed by filamentous gametophytes attached to a thin thread (e.g., Sato et al. 2021). Grown sporophytes are usually harvested before sporophyll formation. Recently, Undaria mariculture in Japan has been conducted intensively in northern Pacific Honshu (Miyagi and Iwate Prefectures) and in the Setonaikai Sea (Seto Inland Sea) located in the western region of Honshu (Tokushima and Hyogo Prefectures, e.g., Dan et al. 2015; Ministry of Agriculture, Forestry and Fisheries, Japan, https://www.maff.go.jp/j/tokei/kouhyou/kaimen_gyosei/); mariculture seedlings originally established in northern Pacific Honshu have been used in both regions. Introduction of materials from northern Japan into the Seto Inland Sea was reported to have started in the 1970s (Dan et al. 2015). An earlier study also suggested that invasions occurred from cultivated materials (Uwai et al. 2006), but the numbers of samples and sampling points in that investigation were quite limited.

The primary purpose of this study was to evaluate the degree of cryptic invasion through Undaria mariculture in the Seto Inland Sea. The combination of genetic analysis and phylogeographic structure analysis is a powerful tool for detecting cryptic invasion (Morais and Reichard 2018). We therefore analyzed the mitochondrial haplotypic diversity of wild populations in the Seto Inland Sea and compared it with that of cultivated materials in the area. The results of this study should help in future efforts to clarify the risk of domestic introduction in seaweed cultivation.

2 Materials and methods

Wild samples were collected from 19 sites along the Seto Inland Sea over the years 2021 and 2022 (Figure 1; Table 1). A small piece of blade was cut from each individual in the field and preserved as a silica-dried sample until DNA extraction. Materials of cultivated individuals were provided as young sporophytes from three fishery cooperatives (FC1–3), and each sporophyte was preserved separately in a small plastic bag with silica gel. Cultivation sites of each fishery cooperative are located along the shore close to each fishery cooperative.

Figure 1:

Sampling sites and frequency of haplotypes in each site. Numerals represent site-numbers shown in Table 1. Number of samples (n) is also shown. (A) Entire Seto Inland Sea; (B) Eastern region of Seto Inland Sea; the closed triangle (FC1), square (FC2) and pentagon (FC3) show three fishery cooperatives that provided cultivated samples. AI, Awaji Island; KS, Kiisuido Strait; HS, Harimanada Sea; OB, Osaka Bay. (C) Haplotypes found in cultivated samples. Colors show haplogroups, and patterns indicate haplotypes shared between cultivated materials and wild samples. Colors and patterns of haplotypes correspond to those in Figure 2. nJ, northern Japan; SJ, Sea of Japan; Ki, Kiisuido Strait; CJ, Central Japan; CA, Continental Asia.

DNA extractions were conducted using a DNeasy plant mini kit (Qiagen) following the manufacturer’s instructions. DNA solutions were purified using a GeneClean II kit (Bio 101) before PCR amplifications. Mitochondrial cox3 and a region including tTrp, tIle, and tGln and spacers between these genes (previously named “tatC-tLeu”) were amplified as described in Uwai et al. (2006), and sequences were determined using a commercial service (FASMAC, Kanagawa, Japan).

The mitochondrial haplotype of each individual was defined based on concatenated sequences of the cox3 and tRNA regions, resulting in 908 bp in alignment. Phylogenetic relationships among the haplotypes were inferred by the statistical parsimony networking method using TCS 1.21 (Clement et al. 2000) under the 95 % confidence limit. Metrics of genetic diversities, i.e., haplotype diversity (h), average number of differences among haplotypes (K) and nucleotide diversity (π) for each population (one corresponding to each site), were estimated using DNAsp ver. 6.12.03 (Rozas et al. 2017). To evaluate the significance of genetic heterogeneity within/among areas, analyses of hierarchical genetic structure (Analysis of Molecular Variance; AMOVA) were also conducted using Arlequin ver. 3.5.2.2 (Excoffier and Lischer 2010). Estimations of pairwise F _ST between populations and statistical significance (1,000 permutations) were performed by Arlequin ver. 3.5.2.2. Genetic relationships between populations were inferred based on Slatikin’s linearized F _ST values using Neighbor Net analysis implemented in SplitsTree v. 4.17.1 (Huson and Bryant 2006).

3 Results

In this study, 23 mitochondrial haplotypes were detected from 219 individuals along the Seto Inland Sea (Figure 2; Table S1). We observed five groups of haplotypes (haplogroups), which were previously defined by sequence similarity and geographical proximity in Uwai et al. (2006, 2023; one of the haplogroups, previously named Seto-Inland-Sea (Uwai et al. 2023), is hereinafter called the Kiisuido haplogroup in consideration of its distribution being limited to the Kiisuido Strait rather than whole Seto Inland Sea. Among 188 wild samples analyzed, the Sea-of-Japan haplogroup was the most dominant (44.7 %), followed in order by the northern Japan haplogroup (28.7 %) and the Kiisuido haplogroup (17.6 %). Haplotypes belonging to Continental Asia (haplotypes 23 and 30) or Central Japan (haplotype 2) were also found, but their distributions were limited (Figure 1). In mariculture materials, 4 haplotypes were identified among 31 samples analyzed; all of these belonged to the northern Japan haplogroup (Figures 1 and 2; Table 1). One of these haplotypes (haplotype 56) was identical to that previously reported from a cultivar used in this area (Niwa et al. 2017). Three of the haplotypes found in mariculture materials were also observed in wild individuals, of which haplotype 31 was the most widely distributed across the wild samples in the studied area.

Figure 2:

Statistical parsimony network among haplotypes. Haplogroups shown by colors are recognized following Uwai et al. (2006, 2023 and Shan et al. (2022); larger circles with bold perimeters and numbers indicate haplotypes observed in this study; smaller colored circles were those previously reported in Japan, China and Korea. Haplotypes are numbered following Uwai et al. (2023). nJ, northern Japan; SJ, Sea of Japan; Ki, Kiisuido Strait; CJ, Central Japan; CA, Continental Asia.

Despite the limited geographic range analyzed in this study, the haplotype frequencies observed were different among areas (Table 1). In the eastern Seto Inland Sea (sites 1–12), a larger number of haplotypes (21 haplotypes from 136 samples) were observed compared to the western area (3 haplotypes from 52 samples). Eight northern Japan haplotypes were observed in this area, and the northern Japan haplogroup prevailed, especially in populations 7–9 and 11. On the other hand, only one of the northern Japan haplotypes (haplotype 31) was observed in the central to western Seto Inland Sea (sites 13–19), and the frequency of the Sea-of-Japan haplogroup increased westwards. Along the Kiisuido Strait (sites 5–9), 7 endemic haplotypes were found in addition to 5 northern Japan haplotypes; the haplotype diversity (h), average number of differences between haplotypes (K) and nucleotide diversity (π) were all largest in the Kiisuido Strait (Table 1). In AMOVA, however, genetic differences among groups were not statistically significant for either the island-based groups or area-based groups (Table 2).

Most of the pairwise F _ST values were significant (p < 0.01), even between cultivated materials (Table S2); in the Neighbor-net analysis based on linearized F _ST (Figure 3), however, two groups of populations (one corresponding to each site), were observed, and one of these included 10 wild populations with zero genetic distance. Populations 9 and 11 were grouped with the cultivated materials (FC1–3), with two other populations (populations 4 and 16) at the basal position of the branch. In contrast, populations 2, 5 and 6 were separated from these two groups and from each other by long branches with two populations (populations 7 and 8) in between.

Figure 3:

Neighbor-net analysis between populations. Numerals represent site numbers in Table 1; genetic distances were calculated among populations of wild samples at each site and cultivated materials provided by Fishery Cooperatives.

4 Discussion

The high genetic diversity in Undaria populations in the Seto Inland Sea observed in this study was at least partly attributable to anthropogenic introductions from other regions. Because strong genetic structures shown by both mitochondrial and nuclear markers have been reported in Japanese Undaria (Uwai et al. 2023), our finding of multiple coexisting haplogroups in the native range of Undaria – especially in the eastern Seto Inland Sea – is extremely exceptional. Furthermore, combinations of haplogroups were different by populations. Groups of populations in the Neighbor net analysis were hardly explainable by geographic area (such as Harimanada Sea, Kiisuido Strait and Osaka Bay), and genetic differentiations among geographic groups of populations were not significant in AMOVA, both of which were caused by distribution of the northern Japan haplogroup in multiple areas (Harimanada Sea, Kiisuido Strait and Osaka Bay). The continental Asia and the central Japan haplogroups also made the geographic structure ambiguous. Generally, from a phylogeographic point of view, unexpected distribution of a lineage has been considered as a sign of cryptic invasion (Morais and Reichard 2018); the distributions of the northern Japan, Central Japan and Continental Asia groups in the studied area could be the result of human transport, either intentionally or not. In particular, the sharing of northern Japan haplotypes between the cultivated and wild samples indicates that Undaria maricultures have caused cryptic invasion(s) in this area.

The large number and high frequency of the northern Japan haplotypes observed in the wild samples also support a hypothesis that Undaria invasions have occurred through fishery activities. Usually, transplantation for fishery purposes would result in large genetic diversity in recipient regions, because vast quantities of materials can be transported (Roman and Darling 2007). It is generally thought that the frequencies of non-indigenous genotypes are relevant to the geographic distance from the area of introduction (AOI) (e.g., Eggertsen et al. 2021). And the genetic diversity of introduced genotypes is relevant to propagule pressure (Lockwood et al. 2005; Roman and Darling 2007), which is determined by the number of introduced individuals (propagule size) and repeat number of introduction (propagule number). The high frequency and large genetic diversity of the northern Japan haplotypes observed at sites 9 and 11, where intensive Undaria mariculture has been conducted, suggested that the shores around these sites could be candidates for AOIs. The cultivated materials used in this study showed high genetic diversity, and the small number of introductory events could explain the observed diversity of northern Japan haplotypes in wild populations. Alternatively, the large number of haplotypes at sites 9 and 11 could have resulted from multiple introduction events around this area, since the diversities observed at these sites (4–5 haplotypes) were larger than those reported at each site in northern Japan (1–2 haplotypes; Uwai et al. 2006, 2023). The cultivars currently used in this area are known to have a northern Japan haplotype (Niwa et al. 2017), and mariculture seedlings from northern Japan have been introduced in this area since the 1970s (Dan et al. 2015); multiple introductions are likely to have occurred during these decades.

It is not clear why the northern Japan haplotypes occupied the populations at sites 7–9 and 11, in contrast to other sites where they were observed as one of the minor haplotypes. As a possible simple explanation, geographic proximity from the epicenter of mariculture should be considered. Shan et al. (2018) and Li et al. (2020) analyzed gene flow between cultivated Undaria populations and wild populations growing on or near cultivation facilities based on SSR markers. Wild and cultivated populations were successfully distinguished in these studies, and percurrent gene flows between wild and cultivated populations 400–500 m apart from each other (Li et al. 2020) were not detected, although genetic mixture between them was suggested in the spontaneous populations on a cultivation raft (Li et al. 2020; Shan et al. 2018). In the present study, wild populations at rocky shores (sites 9 and 11) near cultivation sites (ca. 500–2000 m or more) were strongly affected by cultivated lineages.

Moreover, northern Japan haplotypes were detected at site 7, although it is geographically 11 km apart from the area of extensive Undaria cultivation and no Undaria cultivation has been conducted on the coast, at least for several years. In the cultivated red alga Eucheuma, replacement by non-indigenous Asian lineages has been observed, even in regions where Asian lineages have never been farmed (Tano et al. 2015). The genetic impact from cultivated lineages in the eastern Seto Inland Sea, either hybridization or complete replacement, might be widespread and more serious than detected in this study, considering the maternally inherited manner of mitochondrial markers and oogamous nature of this species.

Ecologically driven mechanisms have been considered as a cause of replacement by non-indigenous lineages (Morais and Reichard 2018, and references therein), although differences in adaptation, such as differences in growth rate, between indigenous and non-indigenous lineages were not clearly evident in the present case. However, it is noteworthy that shrinking of the wild Undaria population in this area was reported in the 1960s and 1970s, when Undaria mariculture was actively performed; it was reported that no Undaria population was detected around the eastern region of the Harimanada Sea and Awaji Island, due to the extraordinarily low SST (Enomoto et al. 1983). Similarly, in the 1970s, the biomass of Undaria populations was smaller than that in the 1980s along the eastern coast of Shikoku (around sites 8 and 9) (Dan et al. 2015). Enomoto et al. (1983) attributed the recovery of Undaria populations to cultivation performed on nearby shores, which is concordant with the consideration by Morais and Reichard (2018) that abrupt expansion of a population could be a sign of cryptic invasion.

The origins of the haplogroups observed in the Seto Inland Sea, other than the northern Japan haplogroup, have not been clear. Considering the dominance in most of the areas westward from Harimanada Sea and the geographic continuity of the distribution range, the Sea of Japan haplogroup might be one of the native haplotypes; however, genetic connection between the Seto Inland Sea and the Sea of Japan is not common in other marine species (seaweed: Akita et al. 2020; Hu et al. 2017, fish: Kato et al. 2021). The phylogenetic position of the Kiisuido haplogroup separated from the other groups by 5–6 substitutions and its endemic distribution suggested that this haplogroup may also be native in this area. The other two minor lineages, one haplotype of the Central Japan group and two of the Continental Asia group, are likely to have been introduced through human activity in view of the completely identical DNA sequences reported in central Pacific Honshu (Uwai et al. 2006) and China (Shan et al. 2022), respectively. The haplotypes of the Continental Asia group (haplotypes 23 and 30) were recently reported from wild populations in Qingdao and Gouqi Island in China (as H10/25), but were not observed in cultivated individuals near the sampling site (Shan et al. 2022). Limited distributions and limited diversities suggest that introduction events of each lineage have been limited in number and scale for these lineages; identification of the geographic origins and introduction vectors for these lineages will require detailed genetic comparisons with samples from purported geographical origins.

In this study, cryptic invasion(s) was detected in the wild Undaria populations of the Seto Inland Sea based on mitochondrial haplotypic analysis. Comparisons with cultivated materials showed that fishery activities have driven the invasion; the exceptionally high genetic diversity of Undaria populations observed in the studied area can be at least partly attributed to mariculture. The present results clearly show a risk of cryptic invasion through seaweed mariculture regardless of whether the mariculture consists of fisheries or blue carbon measures to combat global warming. The impact on the recipient ecosystem by the cryptic invasion, on the other hand, was difficult to evaluate due to the similar morphology between the indigenous and non-indigenous individuals in this case. An increment of non-indigenous lineages could prevent the recovery of indigenous lineages, which was previously reported as a concern in the case of Eucheuma (Tano et al. 2015). On the other hand, there is a case that admixed populations were reported to thrive well compared to native populations declining in size by inbreeding depression (O’Donnell et al. 2017). In the present case, based on the frequency of non-indigenous haplotypes, the degree of genetic impact from the cryptic invasion would appear to have differed by site; comparisons between the local community – e.g., in faunal and floral biodiversity as well as the morphologies and phenological dynamics of Undaria populations between the sites – would assist in evaluating the impact of this cryptic invasion.