東北大学機関リポジトリTOUR

Morphotypes revision, population genetics, and

demographic history of Sinotaia quadrata in

East Asia

著者

YE BIN

学位授与機関

Tohoku University

学位授与番号

11301

博

士

論

文

Morphotypes revision, population genetics, and

demographic history of Sinotaia quadrata

in East Asia

（東アジアにおけるヒメタニシ

_{(Sinotaia quadrata)}

の

形態型と集団遺伝および進化史）

令和２年度

東北大学大学院生命科学研究科

生態発生適応科学専攻

Bin YE

Dissertation

for obtaining the academic degree

Doctor of Science

Morphotypes revision, population genetics, and

demographic history of

Sinotaia quadrata

in East Asia

September, 2020

Ecological Developmental Adaptability Life Sciences

Graduate School of Life Sciences

Tohoku University

Advisor: Prof. Dr. Satoshi Chiba

Bin YE

B7BD3602

Content

ABSTRACT ... 1

KEYWORDS ... 2

1 GENERAL INTRODUCTION ... 3

1.1

T

AXONOMY OF

S

INOTAIA QUADRATA

... 3

1.1.1 Genus of Sinotaia ... 3

1.1.2 Taxonomic History and Controversy ... 5

1.2

P

HYLOGENY OF

S

INOTAIA QUADRATA

... 10

1.2.1 Phylogeny of viviparids ... 10

1.2.2 Phylogeny of Sinotaia quadrata ... 11

1.3

B

IOGEOGRAPHY OF

S

INOTAIA QUADRATA IN

E

AST

A

SIA

... 12

1.3.1 Biogeographic studies in East Asia ... 12

1.3.2 Population genetics and history of Sinotaia quadrata ... 13

2 CHAPTER I REVISION FOR MORPHOTYPES OF SINOTAIA QUADRATA: EVIDENCE FROM

MOLECULAR ANALYSIS AND SHELL MORPHOLOGY ... 15

2.1

I

NTRODUCTION

... 15

2.2

M

ATERIALS AND

M

ETHODS

... 16

2.2.1 Sampling for phylogenetic analysis ... 16

2.2.2 Sampling for microsatellite and morphological analysis ... 16

2.2.3 DNA extraction and microsatellite genotyping ... 18

2.2.4 Phylogenetic analysis of Sinotaia and viviparids ... 19

2.2.5 Morphological analysis of type materials of Sinotaia ... 20

2.2.6 Landmark-based geometric morphometric analyses ... 20

2.3

R

ESULTS

... 23

2.3.1 Phylogeny and shell outline variation of Sinotaia ... 23

2.3.2 Population genetic variation ... 25

2.3.3 Shell shape variations and model-based clustering ... 25

2.3.4 Relationship between genetic differentiation and shape variation ... 28

2.3.5 Habitat effect on shell shape ... 30

2.4

D

ISCUSSION

... 31

2.4.2 Shell morphologic variation of S. quadrata ... 31

2.4.3 Inconsistency between genetic and shape variance of S. quadrata ... 32

3 CHAPTER II HUMAN−GEOGRAPHIC EFFECTS ON VARIATIONS IN THE POPULATION

GENETICS OF SINOTAIA QUADRATA THAT HISTORICALLY MIGRATED FROM CONTINENTAL

EAST ASIA TO JAPAN ... 33

3.1

I

NTRODUCTION

... 33

3.2

M

ATERIALS AND

M

ETHODS

... 35

3.2.1 Sampling, DNA extraction and genotyping ... 35

3.2.2 Population genetics ... 35

3.2.3 Population structure ... 37

3.2.4 Isolation by distance and gene flow ... 37

3.2.5 Human and bioclimatic effects on genetic divergence ... 38

3.2.6 Model checking for population history using microsatellite ... 41

3.3

R

ESULTS

... 43

3.3.1 Population genetics with microsatellite data ... 43

3.3.2 Population structure with microsatellite ... 45

3.3.3 Isolation by distance and gene flow ... 48

3.3.4 Human and environmental correlates of genetic divergence ... 50

3.3.5 Demographic history and introduction ... 51

3.4

D

ISCUSSION

... 59

3.4.1 Demographic history and ancient migration ... 59

3.4.2 Recent introduction, population admixture, and isolation ... 60

3.4.3 Long-term geographical isolation in Japan and modern human use in continental East

Asia ... 61

4 CHAPTER III PALEOGEOGRAPHIC CHANGES IN EAST ASIA SHAPE THE FLUCTUATING

DEMOGRAPHIC HISTORY OF SINOTAIA QUADRATA ... 63

4.1

I

NTRODUCTION

... 63

4.2

M

ATERIALS AND

M

ETHODS

... 67

4.2.1 Sampling for ddRAD-seq analysis ... 67

4.2.2 ddRAD-seq library ... 67

4.2.6 Ecological niche modelling ... 72

4.3

R

ESULTS

... 73

4.3.1 Phylogeny and species tree based on ddRAD ... 73

4.3.2 Genetic structure based on ddRAD ... 74

4.3.3 Extented Bayesian Skyline Plot ... 78

4.3.4 Demographic model simulation ... 79

4.3.5 Ecological niche models ... 84

4.4

D

ISCUSSION

... 87

4.4.1 Genetic variation of contemporary populations ... 87

4.4.2 Population fluctuating in the Pleistocene ... 87

4.4.3 Population outbreak during the Last Glacial Maximum ... 88

5 GENERAL DISCUSSION ... 90

5.1

T

AXONOMY AND PHYLOGENY OF

S

INOTAIA QUADRATA

... 90

5.2

B

IOGEOGRAPHY OF

S

INOTAIA QUADRATA

... 91

5.3

D

EMOGRAPHY OF

S

INOTAIA QUADRATA

... 91

ABBREVIATION ... 93

APPENDIX ... 94

REFERENCE ... 111

ACKNOWLEDGEMENT ... 129

Abstract

The systematics of species in the genus Sinotaia (Gastropod: Viviparidae: Bellamyinae) has been a complex and controversial issue since the nineteenth century and continues to the present day. The type species Sinotaia quadrata exists in many transitional forms, and the validity of some species similar to S. quadrata requires confirmation. Thus, phylogenetic relationships for species related to S. quadrata were reconstructed based on 16S rRNA and mitochondrial cytochrome oxidase subunit I (COI) genes to reveal the invalid classifications of these species. The shell outline morphology was compared based on the type materials of most species in the genus Sinotaia from the initial publications. The admixture of phylogeny and uniform shape variance suggested that some species of Sinotaia should be revised to morphotypes instead of over-classified species. Here, microsatellite markers and landmark-based geometric morphometric analysis were used to examine the genetic differentiation and shell shape variations of S. quadrata in both China and Japan. An inconsistency between genetic variations and shape variations was detected, thus suggesting that using minor shell shape variations is an invalid way to classify each described instance of S. quadrata as an independent species. The shell shape variation has no clear separations but shows plasticity under various local environments. The molecular phylogeny, genetic diversity, and shell morphology provided evidence for revising some species related to S. quadrata, especially for the eight species widely distributed in China and East Asia.

Anthropogenic factors potentially affect observed biogeographical patterns in population genetics, but the effects of ancient human activities on the original patterns created by natural processes are unknown. A widely distributed freshwater snail species in East Asia, Sinotaia quadrata, was used to investigate this issue. It is unclear if S. quadrata in Japan was introduced from China, and how different human uses and varying geographic patterns affect the contemporary population genetics between the two regions. Thus, the demography of S. quadrata and its genetic structure in Japan and continental East Asia were detected. Results showed that S. quadrata populations first naturally migrated from continental East Asia to Japan, which is associated with the ancient period in Japanese geohistory (about 70,000 years ago). They were then artificially introduced in association with human movements in two recent periods (about 8000 and 1200 years ago). Populations in different parts of Japan have their own sources. Natural migration in the ancient period and artificial introduction in the recent period suggest that the population distribution is affected by both the geohistory of East Asia and the history of human expansion. In the background of the historical migration and introduction, contemporary populations in the two regions show different genetic patterns. Population divergence levels were significantly correlated with geographical patterns in Japan, and significantly correlated with human interventions variables in continental East Asia, suggesting that long-term geographical isolation is likely the major factor that shaped the contemporary population genetics in Japan, while modern human uses are likely the major factor in continental East Asia. These preliminary results show a complex demography and unusual genetic patterns in the contemporary populations for a common freshwater snail and are of

significance to determine the historical formation and contemporary patterns of biogeography in Japan and continental East Asia.

The ancient geological patterns and climate change, such as the alternation of land and sea, the formation of lakes and rivers have important influence on the species diversification and biogeographic patterns. In East Asia, there were close paleogeographical relationships among the continent, the Korean peninsula, and the Japanese archipelago. The frequently changes of paleographical patterns since Pleistocene, reflected by the repeated shelf exposure in the regions connecting with continental East Asia and the Japanese archipelago, potentially played an important role in shaping the demographic history of this species in East Asia. To investigate the demographic history of S. quadrata influenced by the natural process of paleogeographical changes in the ancient times, the double digest restriction site associated DNA sequencing (ddRAD-seq) was used for higher precise analyses. Results showed that the population size changed associated with the sea level changes in the early Pleistocene, particularly, was highly consistent with the largest sea level changes in the Last Glacial Maximum (LGM). The non-continental populations increased during the early Pleistocene (~ 2.3−0.8 Ma) and shrank to the small sizes during the mid-Pleistocene (~ 0.8−0.13 Ma) probably because of the larger variations of climate and the possible paleogeographical changes that totally made the Japanese archipelago separated during the mid-Pleistocene. Then, during the LGM, populations encountered size outbreaks probably because they received migrants from continental East Asia through the paleo-river systems developed on the exposed ECS land bridge and through the narrow strait channel that had low-salinity surface water caused by injection of large continental paleo-rivers. The non-continental populations of S. quadrata were then replenished and increased rapidly in the warming up climate after the LGM. The divergence between lineages was also consistent with the period of sea level changes. The ancient river systems on the exposed East China Sea land bridge in the Last Glacial Maximum (LGM) would be corridors for the dispersal of S. quadrata. The migration route of S. quadrata was likely form the east of continent to the southwest of Japan (Kyushu) passing the southern part of Korean peninsula, and then arrived to the other parts of Japan. The alternation of sea and land and the paleographic changes had important influence on the evolutionary process of freshwater species in East Asia.

Keywords

Viviparidae, East Asia, taxonomy, biogeography, demographic history, human effects, geographic isolation, paleogeography change, molecular phylogeny, landmark analyses, geometric morphometrics, microsatellite, restriction site associated DNA sequencing (RAD-seq)

1 General Introduction

Sinotaia quadrata (Benson, 1842) (Gastropoda: Vivipadidae: Bellamyinae) is a native and widespread

viviparid freshwater snail species in China, and is distributed across East Asia including Japan, South Korea, and Vietnam (Hirano et al. 2015; Shu et al. 2014), as well as being present as an invasive species in regions including Europe and South America (Cianfanelli et al. 2017; Ferreira et al. 2017; Ovando & Cuezzo 2012). It feeds on small algae or large plant epidermis or other organic matter, and it can usually be collected on aquatic plants or bottom sediment in a shallow area of both natural and aquaculture water bodies. The female breeds throughout the year, and the juvenile develops to sexual maturity in one year (Liu et al. 1979). It is gonochoristic with internal fertilization, and the brood pouch contains encapsulated eggs, and it also has no dispersive larval phase (Liu et al. 1979; Van Bocxlaer & Strong 2019). S. quadrata is treated as a common food resource in China, because it has wide adaptability to various freshwater environments, a fast growth rate, and a large reproductive capacity (Wen et al. 2018). Thus, it is often used as the indicator species of water environment for studies of ecotoxicology (e.g. Zhu et al. 2014) and environment sciences (e.g. Shi et al. 2017). However, as one of the most common freshwater snails in East Asia, the taxonomy, biogeography, and phylogeny of S. quadrata still have many controversies at present. Studies focusing on the biogeography of S. quadrata have conducted in China but were questionable because they ignored the pending taxonomic issues. There are still many blanks in its evolutionary biology, particularly for its demographic and phylogeographic processes in East Asia with complex geographical history. Therefore, this study attempts to answer some of the complex and controversial issues by revising taxonomic issues of S. quadrata, revealing biogeographic patterns both on continental East Asia and in the Japanese archipelago, and estimating the demographic history of these populations.

1.1 Taxonomy of Sinotaia quadrata

1.1.1 Genus of Sinotaia

The taxonomy of species in the Sinotaia (Hass, 1939) genus is a complicated and controversial issue. Most of the species names accepted currently have undergone confusing changes in classification. Initially, Sinotaia species were contained to the main group of the genus Paludina (Benson 1842; Heude 1890; Reeve 1863) and later to a main group of the genus Vivipara (Kobelt 1909). Some species were defined as variations of Vivipara

quadrata but were not clearly classified as a unique group (Kobelt 1909). Then, Annandale (1924) defined these

Chinese species as a peculiar group called the Viviparae Angulares Group, and classified three different forms for the type species V. quadrata in the context that a large series of them were barely distinguishable. Subsequently, Prashad (1928) argued that this group had close relationship with and evolved from the Indian species, and called this group the Vivipari Dissimiles Group, which is an independent group different from the

Then, Haas (1939) defined the taxonomy of Asiatic viviparids to two large groups based on the shell size, thickness, and shell adornments, and firstly defined Sinotaia as a subgenus of Taia (“Sino-” was annotated as “China”) and Paludina quadrata as the type species of it. Almost in the same period, Yen (1939) defined that

Snotaia aeruginosa, Sinotaia turritus, Sinotaia ecarinatus, Sinotaia dispiralis, Sinotaia reevei, and Sinotaia lapillorum were the corresponding subspecies of S. quadrata (which was called Viviparus quadratus at that time). Furthermore, Yen (1943) summarized this genus and made a relatively complete list of species, and still considered many species as subspecies of S. quadrata (which was called Bellamya quadrata at that time). After that, several new species were added to this genus, and members of the Sinotaia genus gradually became richer (Brandt 1968; Datsenko 2001; Gozhik & Prysjazhnjuk 1978; Liu et al. 1982; Ng et al. 2014; Zhang et al. 1981). Recently, Qian et al. (2014) used the specimens that were best suited to the original figures to give a review of the Sinotaia genus, and provided a helpful conchological identification for these species. They also pointed out that further molecular analysis was essential to confirm the validity of these species. Thus, Sinotaia is finally accepted as the genus name for these viviparid species distributed in East Asia (MolluscaBase 2019a).

However, it is worth noting that the genus Sinotaia in East Asia has a controversial name, Bellamya (Jousseaume, 1880), which is proved to be the specific genus distributed in Africa. The first time that Bellamya was used to classify these Asiatic freshwater species started from Yen (1943), which mentioned that these Chinese species was different from the typical species of Viviparus, the other genera such as Cipangopaludina and Angulyagra, but was close to African species such as Bellamya unicolor and Bellamya dissimilis (Prashad 1928). Subsequently, Liu et al. (1979) continued to use Bellamya as the genus name to summary several important species in China such as B. quadrata, Bellamya aeruginosa, Bellamya purificata, and Bellamya

limnophila. Since then, most of Chinese researchers followed their classification to study these freshwater snail species widespread in China in various fields, such as ecotoxicology (e.g. Zhu et al. 2014) and environment sciences (e.g. Shi et al. 2017). Recently, Qian et al. (2014) argued that previously misclassified “Bellamya” for Chinese groups should be Sinotaia. In addition, the same freshwater snails distributed in Japan and South Korea was classified as S. quadrata (e.g. Hirano et al. 2015; Park 2015), as well as invasive populations in Europe and South America (Cianfanelli et al. 2017; Ferreira et al. 2017; Ovando & Cuezzo 2012). Nevertheless, Gu et al. (2019) analyzed the molecular phylogeny of Chinese “Bellamya” species based on mitochondrial cytochrome oxidase subunit I (COI) genes and obtained an independent clade that was clearly separated from African

Bellamya. But they did not recognize the taxonomic issues of these Chinese species. Instead, they argued that

the differentiation between Chinese “Bellamya” species and African Bellamya species was related to the landscape dynamics of each region. These conclusions were inaccurate because they misplaced an independent genus in East Asia to another genus in Africa and India (Sil et al. 2019; Stelbrink et al. 2020). Moreover, the original designation of Sinotaia made a good solution for arranging the status of the genus based on the shell size, thickness, and shell adornments: Bellamya, Lecythoconcha, and Idiopomus were group A and Mekongia,

with the modern molecular phylogeny of Asiatic viviparid: Bellamya in India and Africa formed highly different clades from Chinese viviparid, and Mekongia, Cipangopaludina, Heterogen, Margarya, and Torotaia are phylogenetically close to Sinotaia (Hirano et al. 2015; Hirano, Saito, Tsunamoto, Koseki, Prozorova, et al. 2019; Stelbrink et al. 2020). Thus, we have reason to believe that Haas’s taxonomy (Haas 1939) is relatively close to the real phylogeny. Sinotaia should be the valid genus name instead of “Bellamya” for this specific group of viviparids in East Asia.

1.1.2 Taxonomic History and Controversy

Sinotaia quadrata is the type species of the genus Sinotaia (Haas 1939). Its original designation is P. quadrata

(Benson, 1842), and it has had several synonyms: V. quadrata (Benson 1842), V. quadratus quadratus (Yen 1939), Taia (Sinotaia) quadrata (Haas 1939), and B. quadrata (Liu et al. 1979; Yen 1943), before being finally accepted as S. quadrata (MolluscaBase 2019b; Ovando & Cuezzo 2012; Qian et al. 2014). Many species in the genus Sinotaia distributed in East Asia especially in the type locality of China were designated mainly based on

Figure 1．Distribution of the main viviparid genera in Asia and Africa. Data are obtained from Global Biodiversity Information

minor variations in shell morphology by the 19th- and early 20th-century authors (Van Bocxlaer & Strong 2019). It has many transitional forms, which in large series can often hardly be distinguished, and the validity of species needs further investigations (Annandale 1924; Qian et al. 2014). For example, S. aeruginosa was initially described as very closely allied to P. quadrata by Reeve (1863), and S. purificata was initially described as very closely alied to P. aeruginosa by Heude (1890). Here, we list the original designations of eight common species in China and show the history of their name changes, indicating that these species are probably different forms, variations, or morphotypes of S. quadrata (Table 1). These forms and variations of S. quadrata that were defined by previous taxonomists (Annandale 1924; Kobelt 1909; Yen 1939) have gradually been ignored by recent researchers (Gu et al. 2019; Liu et al. 1979; Qian et al. 2014), which caused over-classification for species around S. quadrata. The complicated and controversial taxonomy for the common viviparid in East Asia made it difficult to share research results especially for phylogenic and population genetic analyses. In addition, although a total of 25 species of genus Sinotaia are accepted at present (MolluscaBase 2019a), many species are very similar to the type species S. quadrata (Table 1, Figure 2). Therefore, there is an urgent need to revise the systematics of these species around S. quadrata.

Table 1. The original designations of eight common species in China and the history of species name change.

Year Species Type locality Reference

Sinotaia quadrata (Benson, 1842)

1842 Paludina quadrata Benson Zhoushan, Zhejiang, China Benson (1842), p487 1905 P. quadrata var. reevei Deutzenberg et Fischer nom.nov Tonkin, Luc-Nam, Vietnam Dautzenberg and Fischer

(1905), p160-161

P. quadrata var. aeruginosa Reeve

P. quadrata var. bizonalis Mollendorff mss. Tonkin, Luc-Nam, Vietnam

P. quadrata var. heudei Deutzenberg et Fischer

1909 Vivipara quadrata Benson China; Vietnam Kobelt (1909), p120-124

Vivipara (quadrata var.) heudei Dautz. & Fischer. Guangdong, China

Vivipara (quadrata var.) reevei Dautz. & Fischer.

1924 Vivipara quadrata Benson Yunnan, China Annandale (1924), p408-409

Vivipara quadrata (form) minor Nevill Gulangyu, Xiamen, China

Vivipara quadrata (form) limnophila Mabille Erhai, Yunnan, China

Vivipara quadrata (form) dispiralis Heude Yunnan, China

1928 Viviparus quadratus Benson: Vivipari Dissimiles Group China Prashad (1928), p166-168 1937 Viviparus quadrata (Benson 1842) Hunan, China Yen (1937), p18

Viviparus quadrata heudei (Deutzenberg et Fischer 1905) Hunan, China

1939 Viviparus quadratus quadratus (Benson 1842) Zhoushan, Zhejiang, China Yen (1939), p35-37

Viviparus quadratus reevei (Deutzenberg et Fischer 1905) China

Viviparus quadratus heudei (Deutzenberg et Fischer 1905) Yangtze River, China

1939 Taia (Sinotaia) quadrata. Haas China Haas (1939), p96 1943 Bellamya quadrata (Benson 1842) Zhoushan, Zhejiang Yen (1943), p126

Syn. Bellamya purificata Heude 1890 Xiang River, Hunan

1979 Bellamya quadrata (Benson, 1842) China; Korea; Japan Liu et al. (1979), p14 Syn. Sinotaia quadrata (Benson, 1842)

1909 Vivipara (quadrata var.) aeruginosa Reeve Guangzhou, Guangdong, China Kobelt (1909), p123-124

Vivipara (quadrata var.) heudei Dautz. & Fischer. Guangzhou, Guangdong, China

1924 Vivioara quadrata (form) minor Gulangyu, Xiamen, China Annandale (1924), p408 Syn. Paludina aeruginosa var. minor Nevill

1928 Viviparus quadratus var. aeruginosus Reeve: Vivipari Dissimiles Group

China Prashad (1928), p166-168

1937 Viviparus quadrata aeruginosa Reeve Yiyang, Hunan, China Yen (1937), p19 1939 Viviparus quadratus aeruginosus (Reeve 1863) China Yen (1939), p36 1943 Bellamya quadrata orientalis (Lea 1860) China Yen (1943), p126

Syn. Bellamya aeruginosa Reeve 1863 Guangzhou, Guangdong, China Syn. Bellamya quadrata heudei (Deutzenberg et Fischer 1905) Yangtze valley, China

1979 Bellamya aeruginosa (Reeve) China Liu et al. (1979), p16 2014 Sinotaia aeruginosa (Reeve 1863) n. comb. South China Qian et al. (2014), p3-11

Sinotaia purificata (Heude, 1890)

1890 Paludina purificata Heude Xiang River, Hunan, China Heude (1890), p176 1909 Vivipara (quadrata var.) purificata Heude Kobelt (1909), p127 1928 Viviparus purificatus Heude: Vivipari Dissimiles Group Prashad (1928), p166-168 1943 Bellamya purificata Heude 1890 Xiang River, Hunan, China Yen (1943), p126

Syn. Bellamya quadrata (Benson 1842) Zhoushan, Zhejiang, China Syn. Bellamya chengtehensis Taki 1936 Chengde, Hebei, China

1979 Bellamya purificata (Heude) China Liu et al. (1979), p15 Syn. Idiopoma chengtehensis

2014 Sinotaia pyrificata (Benson 1842) n. comb. Xiang River, Hunan, China Qian et al. (2014), p3-11

Sinotaia lapillorum (Heude, 1890)

1890 Paludina lapillorum Heude Ningguo, Anhui, China Heude (1890), p177 1939 Viviparus quadratus lapillorum (Heude) Ningguo, Anhui, China Yen (1939), p37 1943 Bellamya quadrata lapillorum (Heude 1890) Ningguo, Anhui, China Yen (1943), p126 2014 Sinotaia quadrata form lapillorum (Heude 1890) n. comb. Liuzhou, Guangxi, China Qian et al. (2014), p3-11

Sinotaia dispiralis (Heude, 1890)

1890 Paludina dispiralis Heude Yunnan, China Heude (1890), p175 1909 Vivipara dispiralis Heude Kobelt (1909), p120 1924 Vivipara quadrata (form) dispiralis Yunnan, China Annandale (1924), p409 1939 Viviparus quadratus dispiralis (Heude 1890) Yunnan, China Yen (1939), p36 1943 Bellamya quadrata dispiralis (Heude 1890) Yunnan, China Yen (1943), p126 2014 Sinotaia dispiralis (Heude 1890) n. comb. Dinghai, Zhejiang, China Qian et al. (2014), p3-11

Sinotaia lapidea (Heude, 1890)

1890 Paludina lapidea Heude Kienté, Ngan-houé Heude (1890), p175 1909 Vivipara lapidea Heude Kienté, Ngan-houé Kobelt (1909), p129 1928 Viviparus lapideus Heude: Vivipari Dissimiles Group Prashad (1928), p166 2014 Sinotaia lapidea (Heude 1890) n. comb. Dinghai, Zhejiang, China Qian et al. (2014), p3-11

Sinotaia turritus (Yen, 1939)

1939 Viviparus quadratus turritus n. subsp. Yen, 1939 Guangzhou, Guangdong, China Yen (1939), p36 1943 Bellamya quadrata turria (Yen, 1939) Guangzhou, Guangdong, China Yen (1943), p126 2014 Sinotaia turritus (Yen, 1939) n. comb. Youjiang, Guangxi, China Qian et al. (2014), p3-11

Sinotaia angularis (O. F. Müller, 1774)

1877 Paludina angularis Müller

1890 Paludina angularis Müller Guangzhou, Guangdong, China Heude (1890), p120 1924 Syn. Vivipara quadrata (Benson), in part Annandale (1924), p408 2014 Sinotaia angularis (O. F. Müller, 1774) n. comb. Guilin, Guangxi, China Qian et al. (2014), p3-11

Figure 2. Type materials of 22 species in the genus Sinotaia extracted from the initial publications. Scales are 10 mm, except for V, X, Y, l, and m (scales are 5 mm). According to the references, some specimens have no scales and their sizes are estimated approximately according to their description. Some specimens are recorded in Senckenberg Forschungsinstitut und Naturmuseum Frankfurt, Germany (SMF), Indian Museum, India (Ind.Mus.), Naturhistorisches Museum Wien, Austria (NHMW), Mollusk Museum, Faculty of Tropical Medicine, Mahidol University (TMMU), and Zhejiang Museum of Natural History, China (ZMNH). Sinotaia quadrata: A. Paludina quadrata Benson, 1842. Reeve 1863, Plate IV, Species 17. B. Paludina quadrata Benson, 1842. Heude 1890, p. 175, Pl. XL. Fig. 10. C. Vivipara quadrata Benson, 1842. Annandale 1924, p. 408, Pl. 17, Fig. 2. D. Viviparus

aeruginosa Reeve, 1863. Reeve 1863, Species 41, Pl. VII, Fig. a, b. I. Paludina aeruginosa Reeve, 1863. Heude 1890, p. 175, Pl. XL. fig. 5. J. Vivipara (quadrata var.) aeruginosa Reeve, 1863. Kobelt 1909, p. 124, Taf. 21. Fig. 1-2. K. Vivipara quadrata form minor Nevill, 1885. No.2321.Ind.Mus. Annandale 1924. p. 408-409, Pl. 17, Fig. 5. Sinotaia angularis: L. Paludina

angularis Mull.(?). Heude 1890, p. 175, Pl. XL. Fig. 7. M. Vivipara angularis (in part) Nevill. Annandale 1924. p. 408, Pl. 17, Fig. 1. Sinotaia dispiralis: N. Paludina dispiralis Heude, 1890. Heude 1890, p. 175, Pl. XXXIX. Fig. 12. O. Vivipara quadrata form dispiralis Heude, 1890. Annandale 1924, p. 409. Pl. 17, Fig. 3. P. Viviparus quadrata dispiralis Heude, 1890. SMF38873. Type Material. Yen 1939, p. 36, Taf. 3, Fig. 10. Sinotaia lapidea: Q. Paludina lapidea Heude, 1890. Heude 1890, p. 175, Pl. XL, Fig. 2. R. Paludina lapidea Heude, 1890. Heude 1890, p. 175, Pl. XL, Fig. 2a. Sinotaia lapillorum: S. Paludina lapillorum Heude, 1890. Heude 1890, p. 177, Pl. XL, Fig. 11. T. Paludina lapillorum Heude, 1890. Heude 1890, p. 177, Pl. XL, Fig. 11a. U.

Viviparus quadratus lapillorum Heude, 1890. SMF38874. Yen 1939, p. 37, Taf. 3, Fig. 13. Sinotaia turritus: V. Viviparus quadratus turritus Yen, 1939. SMF40241. Holotype. Yen 1939, p. 36, Taf. 3, Fig. 8. Sinotaia angulata: W. Vivipara angulata

var. acutecarinata Kobelt, 1909. Kobelt 1909, p. 136-137, Taf. 27, Fig. 8-10. Sinotaia acutecarinatus: X. Viviparus

acutecarinatus Kobelt, 1909. SMF38832. Paratypodie. Yen 1939, p. 39, Taf. 3, Fig. 23. Sinotaia annulatus: Y. Viviparus annulatus Yen, 1939. SMF40244. Holotype. Yen 1939, p. 39, Taf. 3, Fig. 26. Z. Sinotaia barboti Sinzov, 1884. Sinzov 1884, p.

11, Taf. 9, Fig. 16-17. Sinotaia guangdungensis: a. Vivipara quadratus var. guangdungensis Kobelt, 1906. SMF38890. Holotype. Kobelt 1909, Taf. 21, Fig. 12, 14; Yen 1939, p. 37, Taf. 3, Fig. 14; Ng et al. 2014, Fig.2. b. Viviparus polyzonatus Frauenfeld, 1862. SMF38823. Yen 1939, p. 37, Taf. 3, Fig. 15. c. Vivipara polyzonata Frauenfeld, 1862. NHMW68037. Lectotype, designated by Yen (1939). Ng et al. 2014, Fig.3A. d. Sinotaia guangdungensis Kobelt, 1906. NHMW68038. Paralectotypes. Ng et al. 2014, Fig.3B. e. Vivipara polyzonata Frauenfeld, 1862. NHMW68038. Paralectotypes. Ng et al. 2014, Fig.3C. f. Sinotaia guangdungensis Kobelt, 1906. NHMW68038. Paralectotypes. Ng et al. 2014, Fig.3D. Sinotaia delavayana: g. Paludina delavayana Heude, 1890. Heude 1890. p. 175, Pl. XL. fig. 1. Sinotaia ecarinata: h. Vivipara quadrata ecarinata Kobelt, 1909. Kobelt 1909. p. 123 non 232, Taf. 21. Fig. 15, 16. i. Viviparus quadratus ecarinatus Kobelt, 1909. SMF7659. Lecotype. Yen 1939, p. 36, Taf. 3, Fig. 9. Sinotaia reevei: j. Vivipara (quadrata var.) reevei Dautzenberg & Fischer, 1905. Kobelt 1909. p. 125, Taf. 25. Fig. 10. k. Viviparus quadratus reevei (Dautzenberg & Fischer, 1905). SMF45378. Yen 1939, p. 37, Taf. 3, Fig. 11. l. Sinotaia mandahlbarthi Brandt, 1968. TMMU187. Paratypes. Ng et al. 2018, Fig.3T. m. Sinotaia atrurrolli Brandt, 1968. TMMU2020. Paratypes. Ng et al. 2018, Fig. 3S. n. Sinotaia qionghaiensis Qian, Fang & He, 2014. ZMNH-AIM671. Holotype. Qian et al. 2014. p. 9, Fig. 21. o. Sinotaia xichangensis Qian, Fang & He, 2014. ZMNH-AIM672. Holotype. Qian et al. 2014. p. 9, Fig. 22. p. Sinotaia datunensis Qian, Fang & He, 2014. ZMNH-AIM673. Holotype. Qian et al. 2014. p. 9, Fig. 23. Sinotaia margaryoides: q. Vivipara margaryoides Annandale, 1924. Paratypes. Annandale 1924, p. 410, Pl. 17, Fig. 6. r.

1.2 Phylogeny of Sinotaia quadrata

Comparing with the taxonomy based on morphological characters with minor variations, the molecular phylogeny has deeper ability to reveal the real systematic relationships and is a highly reliable and important tool for addressing biological questions for instance controversial taxonomy among species (Pace et al. 2012; Yang & Rannala 2012). The freshwater viviparids are nearly cosmopolitan except for absence in South America. The phylogeny of viviparid taxa is complicated and often involves biogeographic patterns on continent or even global scale. In recent years, the phylogenetic analysis of viviparids has made great progress and helps to understand the biogeographical patterns in the corresponding regions. For Asian regions, it has the greatest diversity of viviparids (Hirano, Saito, Tsunamoto, Koseki, Prozorova, et al. 2019; Prashad 1928; Strong et al. 2008), but the systematics of some species is still in a state of confusion that persists to the present day and requires comprehensive revision (Van Bocxlaer & Strong 2019). Sinotaia is one of the most common genera distributed in East Asia and has been studied in phylogeny but was misplaced to the African Bellamya genus (Gu et al. 2019). This phylogenetic analysis ignored the controversial taxonomy of species in Sinotaia mentioned in section 1.1 but gave a redundant classification and an invalid conclusion for this genus. It needs to study more for phylogeny of Sinotaia especially for the type species S. quadrata.

1.2.1 Phylogeny of viviparids

To discuss the phylogeny of S. quadrata, it needs to firstly summary the status of phylogenetic analysis of other viviparids. Molecular phylogeny of the two genera of Bellamya and Neothauma in the lakes of the Rift Valley area of Africa inferred that the African viviparids were the sister-group to a clade comprising Asian species (Sengupta et al. 2009). The African viviparids showed highly disjunct biogeographic patterns under the central role of the largest African river system, the Cango (Schultheissβ et al. 2014). The repeated disruptions and distributions of the viviparids during the formation of the East African Rift System caused the disjunct distributions among groups of viviparids in different water divides, for instance, southern and northern, Lake Malawi and the Middle Congo, the Victoria region and the Okavango/Upper Zambezi aera. These results suggested that the historical biogeography of viviparids in Africa is driven by consecutive vicariance events rather than by dispersal (Schultheissβ et al. 2014). In addition, because of the reflooding of the lake after severe Pleistocene low stands, the speciation processes of the modern Bellamya fauna in Lake Malawi had both a sudden recent demographic increase and a spatial expansion after a genetic bottleneck (Schultheissβ et al. 2011). For the Bellamya fauna in the Indian subcontinent, it was suggested that there were at least two dispersal events into India from Southeast Asia, and the into-India dispersal was governed by the geography and climatic oscillations during the Cenozoic driven by the effects of collision, marine transgression, and the paleoclimatic

(Hirano et al. 2015; Hirano, Saito, Tsunamoto, Koseki, Prozorova, et al. 2019). Combining with the morphological analysis, the molecular phylogeny provides essential evidences to make revisions for more reliable systematic relationships. On genus level, subgenus Tchangmargarya was elevated as an independent genus containing Tchangmargarya multilabiata and previous species Margarya yangtsunghaiensis (Du et al. 2013; Zhang et al. 2015). Anularya was defined as a new genus containing previous species Margarya mansuyi and Margarya bicostata (Zhang et al. 2015). Heterogen was added by revising a previous species

Cipangopaludina japonica as Heterogen japonica (Hirano, Saito, Tsunamoto, Koseki, Prozorova, et al. 2019;

Hirano, Saito, Tsunamoto, Koseki, Ye, et al. 2019). In the regions of Southeast Asia including parts of Indochina (including Thai-Malay Peninsula and Singapore, etc.) and East Indies (including Philippines, Java and Sulawesi, etc.), genera Mekongia, Torotaia, and Angulyagra were close to East Asian genera (Hirano, Saito, Tsunamoto, Koseki, Prozorova, et al. 2019; Stelbrink et al. 2020; Stelbrink et al. 2019), while genera Anulotaia, Idiopoma, Filopaludina, Taia, and Trochotaia were close to African and India genera (Sil et al. 2019; Stelbrink et al. 2020).

1.2.2 Phylogeny of Sinotaia quadrata

Despite a stable phylogenetic relationship within viviparid genera, the phylogenetic analysis of genus

Sinotaia in species level is still weakened and problematic. The phylogeny of Sinotaia in China has been studied

(Gu et al. 2019; Wang et al. 2017), but there are several controversial premises and conclusions in their studies, making a more confusion status for researching this common but specific Asian viviparid. Firstly, as mentioned previously in section 1.1, Gu et al. (2019) ignored the taxonomic complexity and controversy of this viviparid group but inherited historical species names that were misplaced into African Bellamya just by rare evidence of shell shape (Liu et al. 1979; Yen 1939). These historically inaccurate identifications made inaccurate premises that the Chinese Sinotaia was misused in “Bellmaya” and combined with African Bellamya despite the two groups are monophyletic in different continents. They did not attempt to accept the Chinese group as an independent genus Sinotaia, but argued that the differences between two groups of “Bellamya” were related to the landscape dynamics in different continents. Secondly, in species level, the controversies among species or forms around S. quadrata (Table 1, Figure 2) were also disregarded. The eight species of Sinotaia in China were over-classified, and the variance among these species were overstated and explained by unclear analysis of shell shape (Gu et al. 2019). Again, even though they got the results that all the eight species were phylogenetically close and mixed with each other, they failed to accept that these species probably were invalid and should be revised as forms or morphotypes of S. quadrata. Finally, they did not estimate the relationship between molecular phylogeny and shell shape that can provide evidence for addressing taxonomy issues. Their results of geometric morphometrics analysis seemed to be also overestimated. Therefore, a more accurate and complete phylogeny of genus Sinotaia should be studied. In addition, a new revision for species or forms around S. quadrata based on phylogenetic and morphological analysis is needed.

1.3 Biogeography of Sinotaia quadrata in East Asia

1.3.1 Biogeographic studies in East Asia

East Asia, especially for the Sino-Japanese Region, is a hot spot area for biogeographic studies. The geological history of East Asia, especially for the opening of Japan Sea that forms the Japanese archipelago originated from the margin of East Asia continent, has played an important role in the biodiversity and species differentiation (Tojo et al. 2017). The Japanese archipelago is originated from the accretionary prism of margin of East Asia with continental rifting of a “double-door” mode (Otofuji et al. 1985). The continental fragments originated from different part of thick continental lithosphere began to rotate in different directions since the early Miocene (during about 21−11 million years ago, Ma) and opened the Japan Sea. The northeast Japan separated from the Far East Siberia region and the southwest Japan separated from the Korean peninsula (Barnes 2003; Isozaki et al. 2011; Martin 2011; Nakama et al. 2010; Otofuji et al. 1985). In the period ~15 Ma, the two parts came to be located in their present position, but the southwest part still connected with continental East Asia (Takenaka & Tojo 2019). During about 15 − 5 Ma, the two fan-shaped spreading basins rifted closer and remained separated by a deep sea, over an area called the “Fossa Magna” (Kitada et al. 2017; Otofuji et al. 1985; Takenaka & Tojo 2019). As vigorous orogenic activity occurring in the Fossa Magna region, the Honshu island of Japan was formed during this period (Kaizuka 1984; Takenaka & Tojo 2019). During the Last Glacial Maximum (LGM, about 18,000−24,000 years ago), the coastline of Japan still connected with continental East Asia at the region of present parts of South Korea and Kyushu of Japan (Aoki et al. 2019; Kimura 2000).

Thus, the species distribution and biodiversity between continental East Asia and the Japanese archipelago have close relationship on the complex background of geological history. In addition, during the last glacial period, the East China Sea (ECS) land bridge exposed and submerged repeatedly resulting in fluctuation sea level that shaped the population structure of many temperate species, and probably acted as a dispersal corridor for species both distributed in continental East Asia and Japan (Lee et al. 2016; Liu et al. 2018; Sota & Hayashi 2007; Takenaka & Tojo 2019; Tojo et al. 2017; Zhang et al. 2016). In most cases, populations in Japan are originated from the continental populations associated with the formation geohistory of Japanese archipelago, and migrated through the dispersal corridor of ECS land bridge during the last glacial period, and finally structured after completely separated from the continent (e.g. Kinoshita et al. 2019; Shalabi et al. 2016; Wepfer et al. 2016; Xiang et al. 2018; Xiang et al. 2017; Yang et al. 2016; Zhao et al. 2019). In addition, the human-vectored gene flow also played an important role in population differentiation in East Asia (e.g. Du et al. 2018; Jang-Liaw et al. 2019; Kameyama et al. 2017; Kuwayama et al. 2017; Suzuki et al. 2015). For taxa of freshwater snails, a few biogeographic studies in East Asia region particularly for Japan, China

migration from Japanese source populations (Miura et al. 2013). For the Japanese Semisulcospira species in Lake Biwa, there were two distinct lineages and they radiated into 15 extant species rapidly and concurrently, suggesting that the recent lake expansion triggered the adaptive radiation in the ancient Lake Biwa (Miura et al. 2019). The role of ancient lakes in genetic and phenotypic diversification of viviparid snails were also proved by Hirano, Saito, Tsunamoto, Koseki, Prozorova, et al. (2019). The Semisulcospira libertina populations in north Taiwan were likely to originate from Japan and populations in south Taiwan were likely to originate from South China or South Asia (Chiu et al. 2017). Freshwater snails of Pomatiopsinae in Japan were suggest to have multiple colonization of the Eurasian Continent (Kameda & Kato 2011). Phylogeography of freshwater planorbid snails in the Japanese archipelago revealed that lineages within the islands were migrated from Eurasian continent through different routes and were diversified after the development of the Japanese archipelago (Saito et al. 2018). For the freshwater viviparid snails, Cipangopaludina chinesis was likely native to the Kyushu region in Japan, indicating its special origins during the geological ancient periods (Hirano, Saito, Tsunamoto, Koseki, Ye, et al. 2019). For the object species S. quadrata, it would be similar with the situation of

C. chinesis but need to test further.

1.3.2 Population genetics and history of Sinotaia quadrata

Previous studies showed that the S. quadrata population showed a weak divergence across large geographical distances in China, suggesting that floods, human translocations, and the large population size probably weakened the genetic differentiation (Gu, Husemann, et al. 2015; Gu, Zhang, et al. 2015; Gu, Zhou, et al. 2015). However, there is a lack of data on climate and human effects to support the conclusions in these previous studies, which also did not compare the different circumstances in the various S. quadrata distribution areas. It is suggested that S. quadrata in Japan was introduced from China, but this remains inconclusive (Hirano et al. 2015). Fossils of S. quadrata in China were found in the Late Pleistocene and Early Holocene sediments of 11,000–6000 years ago (Huang et al. 2007; Wang 1983; Wang 1961). The oldest S. quadrata fossils in Japan are from the southern Kyushu, which is also about 10,000–6000 years ago (Matsuoka unpublished), but fossil records in other parts of Japan (Honshu) are more recent about 4000–1000 years ago (Kurozumi 2013). S.

quadrata is currently the most abundant species among Japanese viviparid species, but its presence in the fossil

records that are within the past 10,000 years are very rare relative to other viviparid species (Kurozumi 2013). Thus, S. quadrata in Japan was likely introduced from continental East Asia before 10,000 years ago.

Historically, the ECS land bridge might serve as a dispersal corridor that contributed to population distribution from continental East Asia to Japan during the Pleistocene (Kameyama et al. 2017; Saito et al. 2018; Wepfer et al. 2016; Zhang et al. 2016; Zhao et al. 2019). The land bridge has been influenced by glacial-interglacial cycles during the Pleistocene (Voris 2000), resulting in its repeated exposure and submergence, and this played an important role in shaping the population structure across the land bridge (Zhang et al. 2016).

Combining with the fossil information of S. quadrata, one hypothesis that the migration of S. quadrata into Japan is likely to contain a natural process could be developed. It originally distributed in some regions of the Eurasian margin. These regions were separated from the continent and formed parts of the Japanese archipelago in geohistory, and finally, populations in these regions settled down and experienced complex evolutionary processes with a long-term geographical isolation. Later in recent periods, associating with human communication between Japan and continental East Asia that has continued for thousands of years, S. quadrata may also have been subsequently introduced several times into Japan by human activity more recently. Therefore, S. quadrata migration from continental East Asia to Japan might contain both processes of ancient migration caused by the division of Japan from Eurasian margin and recent introduction driven by human communication.

2 Chapter I Revision for morphotypes of Sinotaia

quadrata: evidence from molecular analysis and shell

morphology

2.1 Introduction

Asian regions have the greatest diversity of viviparids (Hirano, Saito, Tsunamoto, Koseki, Prozorova, et al. 2019; Prashad 1928; Strong et al. 2008), but the systematics of some species is still in a state of confusion and requires comprehensive revision (Van Bocxlaer & Strong 2019). S. quadrata (Benson, 1842) (subfamily Bellamyinae) is a widespread viviparid freshwater snail species that is distributed across East Asia, including China, Japan, South Korea, and Vietnam (Hirano et al. 2015; Shu et al. 2014). It is also present in other regions as an invasive species, including Europe and South America (Cianfanelli et al. 2017; Ferreira et al. 2017; Ovando & Cuezzo 2012). It has wide adaptability, fast growth, and large reproductive capacity, and it can usually be collected from aquatic plants or bottom humus in both natural and aquacultural water bodies (Liu et al. 1979).

Sinotaia quadrata is the type species of the genus Sinotaia (Haas 1939). Its taxonomic history and

controversy has been introduced in section 1.1.2. There is a need to revise the systematics of these species. Some morphological and karyological studies have also raised doubts about the validity of some species (Ferreira et al. 2017; Park 2015; Wu et al. 2000). In addition, a group of sequence data of additional specimens belonging to eight species was deposited in GenBank, and phylogenetic studies based on previous species classifications have shown that these species are phylogenetically mixed and were not recovered reciprocally monophyletic (Cianfanelli et al. 2017; Gu et al. 2019; Hirano et al. 2015; Wang et al. 2017). However, the validity of these species has not been considered in previous studies. Therefore, the purpose of this study is to revise the over-classified species in genus Sinotaia based on molecular phylogeny, genetic diversity, and morphological variation, especially for the eight species closely related to S. quadrata (Table 1).

Phylogenetic trees with mitochondrial 16S rRNA and COI genes of related viviparid species were reconstructed to give a basic phylogenetic relationship for Sinotaia. The type materials of Sinotaia species from original publications were extracted for morphological analysis at the species level (Figure 2). Then, the potential species of S. quadrata that was proved to be monophylitic were pooled together to analyse genetic diversity using microsatellite loci of nuclear genes and to analyse the correlations between genetic and shape variations. Shape variations among different habitat factors were also compared because the minor shape variations of S. quadrata are likely to be affected by environmental factors. Based on phylogenetic relationships and variances in morphology and genetics, it might be expected that several species that are close to S. quadrata

are invalid and should possibly be revised to corresponding morphotypes.

2.2 Materials and Methods

2.2.1 Sampling for phylogenetic analysis

Phylogeny of Sinotaia and other viviparids was reconstructed based on 16S and COI for addressing the taxonomic issues. First, the data set of 16S rRNA and COI sequences was collected from GenBank. Because there are repeated 16S and COI sequences of S. quadrata among different individuals in GenBank, it need to reduce the taxa number and get clear results in phylogeny analysis. Then, blast analyses were conducted to remove repeated sequences and select unique sequences to represent taxa from different localities and cover different ‘species’ similar to S. quadrata. According to the previous phylogenetic relationship established for the Sinotaia genus (Gu et al. 2019; Hirano et al. 2015; Hirano, Saito, Tsunamoto, Koseki, Prozorova, et al. 2019; Stelbrink et al. 2019), the following related sequences of viviparid were added to the dataset: Cipangopaludina, Heterogen, Anularya, Mekongia, Torotaia, Angulyagra, Filopaludina, and Taia in Asia, Bellamya and Neothauma in Africa and India, and Viviparus in Europe (Schultheissβ et al. 2014; Schultheissβ et al. 2011; Sengupta et al. 2009; Sil et al. 2019). A sequence of Pomacea canaliculata was used as the outgroup taxon. The dataset finally had 96 taxa of 16S and COI genes for the phylogeny reconstruction of viviparids (Appendix 1).

2.2.2 Sampling for microsatellite and morphological analysis

Fifteen sites were sampled on continent East Asia. Sampling sites that covered several representative river systems were selected and were not intensively sampled all over mainland China, because previous studies showed that S. quadrata has a weak divergence across large geographic distances in China (Gu 2013; Gu, Husemann, et al. 2015; Gu, Zhang, et al. 2015). Two sites of Yilan and Tainan on Taiwan Island were included because, although they are not on the mainland, they are much closer to the mainland compared to the larger distance that separates Japan from the continent. Two sites in Vietnam were also sampled because they have connections through the freshwater system on the same continent (East Asia). Then, 18 sites were sampled uniformly across Japan, which is separated from continental East Asia by seas. Sites in Japan were sampled because possible divergence on the Japanese archipelago must be considered. Overall, 33 sampling locations with 778 individuals in Japan (number of populations, Npop = 18; number of individual samples, n = 426), China

(Npop = 13, n = 326) and Vietnam (Npop = 2, n = 26) (Table 2) were explored. These samples were randomly

collected from lakes, ponds, or water channels that were mostly around rice fields, reservoirs, parks, or villages at each site. All sites were related to human habitats, and a few samples were even bought from markets (Table 2). Because individuals from Vietnam were relatively few and all bought from the market, they were not used

Geographic Distance Matrix Generator (GDMG) v1.2.3 (Ersts 2018), based on the geographic coordinates of the sampling location in each population (Figure 3, Table 2).

Table 2. Sinotaia quadrata samples for microsatellite analysis (N = 778, Npop = 33) and landmark-based geometric morphometric

analysis (N = 383 was choosen, Npop = 31, populations in Vietnam were removed). † Part of samples was bought from market.

The approximate coordinates of market samples were from the seller's information. Final coordinates used for geographic distance calculation were estimated according to the centroid of all sampling sites. For each sampling site, the following are indicated: GPS coordinates (in decimal degrees), total sample size (NS), average number of alleles (NA), observed (HO) and expected (HE) heterozygosities, and private allelic richness Ap(10). HWE indicated the test significance for deviation to Hardy– Weinberg equilibrium: * p < 0.05; ** p < 0.01; *** p < 0.001.

NO. Site Label Longitude Latitude n NA HO HE HWE Ap

Japan 0.4919 0.6692 1 Aomori AM 141.32549 40.71090 13 6.833 0.5399 0.7495 *** 0.07 2 Akita AK 141.32549 39.91159 25 9.417 0.5272 0.7214 *** 0.27 3 Miyagi MY 140.06500 38.20503 21 6.583 0.4991 0.5728 *** 0.11 4 Fukushima FK 140.30995 37.41986 23 9.417 0.4837 0.7287 *** 0.17 5 Nagano NGN 137.97012 36.27872 28 7.417 0.6012 0.6512 *** 0.13 6 Kanagawa KN 139.36815 35.34148 48 6.250 0.2858 0.4508 *** 0.13 7 Shizuoka SZ 137.73195 34.74079 24 6.583 0.5600 0.6444 *** 0.16 8 Nagoya NGY 136.97387 35.17167 24 8.167 0.6013 0.7664 *** 0.21 9 Shiga SG 136.28629 35.49361 23 12.667 0.7218 0.8604 *** 0.30 10 Biwako BW 136.07092 35.24254 20 10.167 0.6717 0.8050 *** 0.28 11 Nara NR 135.73082 34.67111 22 6.917 0.6588 0.7415 *** 0.10 12 Osaka OS 135.52899 34.56866 13 10.250 0.6338 0.8457 *** 0.19 13 Kagawa KGW 133.80557 34.26782 31 10.083 0.5152 0.7264 *** 0.29 14 Kochi KC 133.59092 33.52595 21 6.167 0.4905 0.5415 *** 0.07 15 Yamaguchi YM 132.21441 34.14158 33 7.833 0.4827 0.7140 *** 0.18 16 Kumamoto KT 130.65568 32.75014 22 6.333 0.4066 0.5654 *** 0.18 17 Nagasaki NGS 129.71472 33.17973 9 4.500 0.3563 0.5897 *** 0.10 18 Kagoshima KGS 130.25028 31.82426 26 4.300 0.2282 0.3709 *** 0.11 China-Vietnam 0.6149 0.8398 19 Changchun CC 125.30809 43.85745 27 11.500 0.5318 0.8148 *** 0.15 20 Anhui† _{AN 117.38800 31.71570 20 14.833 0.6418 0.8905 *** 0.60} 21 Nanjing NJ 118.78347 32.10670 48 20.583 0.6426 0.879 *** 0.36 22 Hangzhou HZ 120.11192 30.27164 42 20.917 0.6174 0.8780 *** 0.44 23 ChunAn CA 119.14888 29.49127 20 13.750 0.6830 0.8803 *** 0.67 24 Hubei† _{HB 112.63410 30.92060 16 14.000 0.5649 0.8952 *** 0.41} 25 Jiujiang JJ 115.99929 29.70966 20 13.333 0.6254 0.8615 *** 0.42 26 Sichuan SC 30.63369 104.07721 9 7.417 0.4715 0.8098 *** 0.27 27 Yangzonghai YZH 102.99134 24.89683 32 12.833 0.6516 0.8099 *** 0.30 28 Shilin SL 103.31128 24.81002 20 8.583 0.6530 0.7504 *** 0.42 29 HongKong HK 114.11083 22.51421 37 11.667 0.5027 0.7516 *** 0.23 30 Yilan YL 121.77606 24.66726 19 8.417 0.7350 0.8582 *** 0.33 31 Tainan TN 120.29436 23.19994 10 9.988 0.6732 0.8387 *** 0.11 32 Cao Bang† _{CB 106.26256 22.66817 6 4.900 0.6433 0.7103 0.0581 0.40} 33 Lang Son† _{LS 106.82302 21.73836 26 9.250 0.4716 0.7316 *** 0.29}

2.2.3 DNA extraction and microsatellite genotyping

Specimens with muscle tissues used for DNA extraction were stored in 99 ethanol. The genomic DNA of each sample was extracted using a modified phenol–chloroform method, which was described by Hirano et al. (2015). After pre-testing several microsatellite markers for S. quadrata, which were developed by Gu et al. (2012a, b), we chose 12 loci for stable amplification (Table 3). In the PCR protocol, normal forward primers and fluorochrome-labeled forward primers were used, and a 7-bp PIG-tail (5′–GTTTCTT–3′) to the 5′ end of reverse primers was added to enhance genotype specificity (Brownstein et al. 1996). Each PCR amplification reaction was performed in a total volume of 2 μL, which included the following: 1 μL of 2× PCR MasterMix (QIAGEN, Hilden, Germany), 0.1 μL of forward primer mixture that was mixed with four different fluorochrome-labeled

Figure 3. Sampling locations of S. quadrata distributed in Japan and the continental East Asia. Abbreviation

letters represent location names. Japan: AM–Aomori, AK–Akita, MY–Miyagi, FK–Fukushima, NGN– Nagano, KN–Kanagawa, SZ–Shizuoka, NGY–Nagoya, SG–Shiga, BW–Biwako, NR–Nara, OS–Osaka, KGW–Kagawa, KC–Kochi, YM–Yamaguchi, KT–Kumamoto, NGS–Nagasaki, KGS–Kagoshima. China: CC– Changchun, AN–Anhui, NJ–Nanjing, HZ–Hangzhou, CA–ChunAn, HB–Hubei, JJ–Jiujiang, SC–Sichuan, YZH–Yangzonghai, SL–Shilin, HK–HongKong, YL–Yilan, TN–Tainan. Vietnam: CB–Cao Bang, LS–Lang Son. Abbreviations are also shown in Table 2.

DNA (dried in advance, and the concentration was not quantified). Thermocycling conditions were as follows: 95°C for 5 min + 30 × (95°C for 30 s, 60°C for 90 s, 72°C for 30 s) + 60°C for 30 min (BioRad T100TM _Thermal

Cycler, Hercules, CA, USA). PCR products were genotyped using an ABI 3130 genetic analyzer (Applied Biosystems, Waltham, MA, USA), and allelic sizes were scored by GeneMapper v4.0 (Applied Biosystems) using an internal size standard (GeneScan LIZ500, Applied Biosystems). Questionable specimens and poor amplification with hard-to-read genotypes were re-extracted and re-amplified. Genotypes that have less than 20% of missing data were considered for all analyses.

Table 3. Twelve microsatellite loci primers used for Sinotaia quadrata. * indicates that PIG-tail 5’–GTTTCTT–3’ is added at the 5’

end of each reserve primer.

Locus Flurescent label Primer (5' to 3')* GenBank

ATG142 FAM F: ATGCCGATTATTCTGATTCTGG JX019326

- R: GTTTCTTACGCAAGTTTCATTCATGTATGTC TC298 VIC F: CTCCAAAGACTGTTACTGCTACGA JX019072 - R: GTTTCTTCACACAAACTAGGTAAGGGGACAT TXH79 NED F: TCCTGATTCAAGGACGTTGTGC JN555804 - R: GTTTCTTGACTGGTGCTGGCTGTAGTGC TXH12 PET F: AGGCCTCAGCTTGAATCCCTA JN555765 - R: GTTTCTTCGGCTCCCATTTTGAGCATTG

GATA328 FAM F: CCTGCGTCAATTTAAAACCATAG JX020456

- R: GTTTCTTGGGTAGGTAGGTGGGTAAGTGAG

TXH223 VIC F: GCTTGCCACCGCATTGTAGC JN555901

- R: GTTTCTTCACAACTTGGTCACAAAGCCG

CAG41 NED F: TTTGCTGCGTTTACTCGTCCTG JX019413

- R: GTTTCTTCTCGTTCTTGGGCTGGGTGTT

TXH30 PET F: CACATAGAAGGTCACACGT JN555777

- R: GTTTCTTGAATTCCAAACTCAGACAACGG

CAG178 FAM F: CATGACATAACACCCCTACCCTT JX020456

- R: GTTTCTTAATTTGGTATCTTGGAATCTGACG TXH65 VIC F: TGTGACTAAGTGTGTTTGCATG JN580068 - R: GTTTCTTCTACCAGGTGCTTGGTGTG CCT238 NED F: ACATAAGTGCTTGCGATAGTGCG JX019326 - R: GTTTCTTCTACCAGGTGCTTGGTGTG TXH113 PET F: CAAGCATGGATGCAGAACTC JN555824 - R: GTTTCTTCTCGTTGGTCCGATACAACC

2.2.4 Phylogenetic analysis of Sinotaia and viviparids

Sequences of 16sRNA and COI were aligned using the multiple sequence comparison by log-expectation (MUSCLE) method with MEGA v6.06 (Tamura et al. 2013). The best-fit partition models for the final two-gene

dataset were investigated in PartitionFinder 2 (Lanfear et al. 2016) by using a greedy algorithm with codon-

position data blocks for COI and linked branch lengths. The best-fit scheme revealed by the corrected Akaike information criterion (AICc) suggested four partitions: 16S = GTR+I+Г, COI1st _{= GTR+I+Г, COI}2nd _{= HKY+I+Г, and}

COI3rd _{= GTR+I+Г. These partition models were used for phylogenetic reconstruction with Bayesian inference}

(BI) using MrBayes 3.2 (Ronquist et al. 2012). The settings for BI were ngen = 5,000,000, samplefreq = 1000, and other default settings. In addition, phylogenetic analyses were also performed with maximum likelihood (ML) using AVX vector instructions in RAxML 8.0.0 (Stamatakis 2014), but used the only available substitution model GTR+Г for all partitions. The bootstrapping of ML was performed by convergence criteria of autoMRE (the analysis was automatically stopped after 400 bootstrap replicates). Result trees were annotated using FigTree v1.4.3 (Rambaut 2016).

2.2.5 Morphological analysis of type materials of Sinotaia

The photographes of type materials extracted from previous publications (Figure 1) were used to illustrate shell morphology differences among species in the genus Sinotaia. Some specimens of S. quadrata collected in different localities were also used to analyze shape differences from the type materials. In total, specimens from most of the distribution areas of China, Japan, South Korea, Vietnam, as well as invasion areas of Italy, and Argentina (Cianfanelli et al. 2017; Ferreira et al. 2017; Hirano et al. 2015; Ovando & Cuezzo 2012) were analyzed. Because some specimens from the original publications had no size scales, the landmark-based method mentioned below was not suitable for these specimens. Therefore, the entire shell shape outlines were evaluated using elliptical Fourier analysis using the R package Momocs v1.2.9 (Bonhomme et al. 2014) (Figure 4). The elliptical Fourier transform were analyzed using the efourier function with a harmonic setting of 99. Finally, principal component analysis (PCA) of shell shape outlines was conducted using the function PCA.

2.2.6 Landmark-based geometric morphometric analyses

Figure 4. Example of photography data for the morphological analysis using shell shape outline.

a total of 31 sampling sites and 746 samples were used (two sites in Vietnam were not used because samples were few to perform shape variation analyses). To detect the differences of genetic and shape variance, these sampling sites were divided into levels of two countries (Japan and China), six groups (Japan East: JE, Japan Middle: JM, Japan West: JW; China East: CE, China West: CW, and China South: CS, Table 4) and 31 populations (the names of the sites). Table 4. Population groups used for comparison between genetic and shape variations.

Group Region Population

JE Japan East Aomori, Akita, Miyagi, Fukushima, Nagano, Kanagawa, Shizuoka, Nagoya

JM Japan Middle Shiga, Biwako, Nara, Osaka

JW Japan West Kagawa, Kochi, Yamaguchi, Kumamoto, Nagasaki, Kagoshima CE China East Changchun, Anhui, Nanjing, Hangzhou, ChunAn, Hubei, Jiujiang

CS China South HongKong, Yilan, Tainan

CW China West Sichuan, Yangzonghai, Shilin

Firstly, the discriminant analysis of principal component (DAPC) analyses without previous information were performed at these levels using the adegenet v2.1.0 package in R (Jombart 2008; Jombart et al. 2010). This method is suitable for detecting genetic structures and it does not make assumptions such as Hardy–Weinberg equilibrium and linkage disequilibrium in population genetics. A number of clusters K ranging from 1 to 30 were chosen for DAPC. The maximum value of K was chosen to be twice the actual simulation optimal K (K = 15), which was determined based on the lowest Bayesian information criterion (BIC) (Jombart et al. 2010). The optimal number of retained principal components (PCs) was selected using the optim.a.score function (Jombart 2008; Jombart et al. 2010). Then, the DAPC analyses with the previous information were performed at levels of countries, groups, and populations. For better comparison with the results of the PCA of shapes at the population level, the significant PCs from the DAPC were detected by a broken-stick test using the PCDimension v1.1.9 package in R (Coombes & Wang 2018), and the results were averaged in each population. Then, the Pearson’s distance for significant PCs from the DAPC was estimated using the factoextra v1.0.5 package in R (Kassambara and Mundt, 2017). The Pearson’s distance calculated from significant PCs of the DAPC could imply genetic differences because it had a significant correlation with the FST estimated by the ENA method (Mantel test: r = 0.41, p < 0.001). In total, 383 adult individuals with complete shells were chosen from the 31 sampling sites for morphological analyses. The shell morphologic variation was estimated using the two-dimensional landmark-based geometric morphometric method. Each shell was positioned in a bed of sand of contrasting color with the aperture facing upward and then photographed using a Canon EOS X5 camera. The photography equipments were mounted on a high-precision stand to allow for repeatable orientation and positioning, although there were no significant effects of object placement, camera height, or digitization inconsistency on the shape variations (Vaux et al. 2017). Then, to define approximately consistent placement of landmarks, two “combs” were added to the image

of each shell using Adobe Photoshop CS6 (Figure5A). Totally, six fixed landmarks and 16 semi-landmarks around the outline of the shell and aperture were digitized, and the digitization was performed using ImageJ 1.x to obtain the x and y coordinates of the landmarks (Schneider et al. 2012). Semi-landmarks were slid using the function define.sliders in the R package geomorph v3.0.7 (Adams & Otárola-Castillo 2013) (Figure 5B). Twenty randomly sampled shells were re-photographed and re-digitized to examine the error associated with photography and digitization using the morphol.disparity function. The error was found to be negligible (< 5%) in comparison with the biologically meaningful variations among populations. The generalized Procrustes analyses with landmarks were performed using the package geomorph v3.0.7 in R (Adams & Otárola-Castillo 2013). Firstly, the Procrustes distance between pairwise populations were calculated based on the mean shape of each population. Next, a PCA of the Procrustes analyses were conducted across all individuals and all landmarks using the function plotTangentSpace.

Similar to the computation of Pearson’s distance for the PCs of DAPC (as described in the section on population genetics), the broken-stick model on the proportion of variances of PCs were used to determine the number of statistically significant PCs of shape using the package PCDimension v1.1.9 in R (Coombes & Wang 10 15 20 25 30 35 -3 0 -2 5 -2 0 -1 5 -1 0 spec[, 1] sp ec[ , 2 ] 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 5 mm 1 2 3 11 17 18 4 5 6 7 8 9 10 12 13 14 15 16 19 20 21 22 A B Figure 5. Landmarks used for geometric morphometric analysis of Sinotaia quadrata. A, Landmark positions were optimized to capture the shape variation around the aperture and body whorl with six fixed landmarks (blue) and 16 semi-landmarks (red). B, Sliding of semi-landmarks: selected semi-landmarks were filled in red, and red lines were drawn to connect the three related semi-landmarks with the flanking landmarks in blue.