[A-OS13] Marine ecosystems and biogeochemical cycles: theory, observation and modeling

[AOS13-P01] Marine fish distribution in the Kuroshio Extension area observed by OceanDNA

*YU ZESHU¹、伊藤 進一²、Marty Wong²、吉澤 晋²、井上 潤²、伊藤 幸彦²、兵藤 晋² (1. 東京大学大学院農学生命科学研究科、2. 東京大学大気海洋研究所)

キーワード：環境DNA、魚類分布、黒潮続流、小型浮魚類

Due to huge volume of the ocean and the difficulty of observing in stormy weather, distribution patterns of marine fish species are still underdetermined. In recent years, a new survey technique using environmental DNA (eDNA) in water has been developed and actively conducted. However, the lower density of fish in the open ocean leaves questions about the effectiveness of ‘eDNA survey in the open ocean’ (hereafter OceanDNA). We have three purposes in this research: 1. comparing the fish distribution found by OceanDNA with the net-sampling result, 2. analyzing the difference between OceanDNA data from MiFish and from qPCR method, 3. comparing the OceanDNA data from Niskin bottle and from ship-bottom-intake seawater. During the KS-18-5 cruise by Shinsei-Maru, seawater samples were collected at several stations in the Kuroshio Extension area from different depth using two methods. One is water sampling by Niskin bottles from 5 to 300 m combined with Conductivity Temperature Depth (CTD) instrument, which measured water mass distribution, and by a bucket for 0 m (130 samples). The other is water sampling from intake seawater from the bottom of the ship (15 samples). Water samples were filtered by Sterivex DNA filter (0.45μm) and the filters were stored at -20 °C, then DNA was extracted from filter after the cruise. To analyze the OceanDNA samples, we used the DNA metabarcoding method with universal MiFish primers targeting a hyper-variable region of fish mitochondrial 12S rRNA gene. The PCR results were compared to DNA library to identify the species. 60 marine fish species were identified and in terms of number of species, Myctophidae is the most dominating family that contains 15 species of all 60. The results showed spatial segregating distribution of 15 Myctophidae species, in which 4 groups were segregated by depth and 2 groups were segregated by sampling stations. On the other hand, two potential predator species of Myctophidae were found out from all 60 species, and both showed depth-segregating distribution with Myctophidae. The fish distribution pattern found here, especially the depth-segregating distribution in Myctophidae and in two prey/predator groups were consist with previous studies. We also used qPCR method to analyze the OceanDNA samples to detect 6 small pelagic fish species (Sardinops melanostictus, Engraulis japonicus, Scomber japonicus, Scomber australasicus, Trachurus japonicus, and Cololabis saira). We first compared the absence/presence of the 6 small pelagic species between MiFish and qPCR methods. When DNA of one species was found by either method, we admitted its existence. Two Scomber species and Sardinops melanostictus got high existence rate (higher than 40%) while Engraulis japonicus and Cololabis saira were found in less than 15% samples (Trachurus japonicus was found in only one sample). From the Niskin water samples, the positive significant correlations were found between data of qPCR and MiFish analyzing for the existence of two Scomber species and Sardinops melanostictus, but qPCR got more than three times discover rate compared with MiFish method. Same correlations were not found in data from ship-bottom-intake seawater. However, if pooling data of six species together, the positive significant correlation was found between data of qPCR and MiFish analyzing although qPCR got more discover rate than MiFish method in both two kinds of water. The result suggests that qPCR method we developed is a proper method to observe distribution of the small pelagic species. On the other hand, when looking at qPCR method data, discover rate from Niskin-bottles seawater and from ship-bottom-intake seawater were always similar in highly existed species. But for sparse distributed species Engraulis japonicus, it was discovered more from ship-bottom-intake seawater rather than from Niskin-bottles seawater. When using MiFish data, the discover rates were often higher from ship-bottom-intake seawater than from Niskin-bottles seawater. The data was limited and the reason why the discover rates differ between the water sampling methods is still unclear. Further comparison is essential to clarify the mechanism in future.