Genetic evidence for the invasion of Cymbella janischii (A. Schmidt) De Toni in Japan

Cymbella janischii (A. Schmidt) De Toni, an endemic diatom of the Paci�c Northwest, was found in 2006 in Japan, and since then, its distribution has been expanding. It was identi�ed on the basis of morphology in Japan, however, molecular analyses have not yet been performed, with no sequences known of genes derived from the Japanese isolate. Here, we analyzed rbcL, psaB, psbA, 18S rRNA, and 28S rRNA gene sequences (6526 bp in total) of the C. janischii specimens from several locations in Japan and explored their genetic relatedness with C. janischii from its country of origin (the United States) and its closely related species. We showed that all Japanese specimens had the same sequences, regardless of geographical distance, and formed a clade with the US C. janischii. The identities and the pairwise distance between the sequences of the Japanese and the US diatoms were 99.937% and 0.0003, respectively, indicating that these diatoms are extremely similar. These results provide potential genetic evidence of the recent invasion and rapid spread of C. janischii from the US in Japan.


Introduction
Invasion of alien species and their impact on local ecosystems has become a major issue in recent years (Simberloff et al. 2013).Among diatoms, invasive species are also a serious problem, as exempli ed by the well-known invasion of New Zealand by Didymosphenia geminata (Lyngbye) M. Schmidt (Kilroy et al. 2008;Blanco and Ector 2009).
Cymbella janischii (A.Schmidt) De Toni has an ability to produce copious extracellular polymeric substances that form the stalks and cause nuisance blooms as 'rock snot', similar to that of D. geminata (Khan-bureau et al. 2016).Cymbella janischii is considered an endemic diatom in the Paci c Northwest (Krammer 2002;Bahls 2007); however, it has recently been observed outside of the distribution (e.g., Arizona, Colorado, New York, Oklahoma, and Connecticut in the US and in Japan (Suzawa et al. 2011;Khan-bureau et al. 2014Khan-bureau et al. , 2016)).
Genetic analysis is useful in estimating the region of origin for invasive species, their evolution, and the pathway through which they were introduced (Cann et al. 1987;Gaut et al. 1992;Kato-Unoki et al. 2020).For example, the invasion of D. geminata in New Zealand was evidenced by the genetic distance based on a comparison of a 2230 bp sequence in the chloroplast genomes of strains from New Zealand and other countries (Kilroy and Novis 2018).For C. janischii, the genetic data from Idaho and Connecticut in the US are submitted in GenBank (Nakov et al. 2014;Khan-bureau et al. 2016).These aligned sequences (about 300 bp in the small subunit rDNA (18S)) are consistent and show a close relationship (Khanbureau et al. 2016).
In Japan, C. janischii was rst reported from the Chikugo River in 2006 (Suzawa et al. 2011) and has expanded its distribution thereafter (Suzawa and Suzawa 2021).In the previous study, C. janischii was identi ed morphologically; however, genetic analyses were not carried out.Here, we determined the genetic sequences of this species in Japanese samples, con rmed species identi cation at the molecular level, and clari ed their phylogenetic relationship with C. janischii from the country of origin (the US) and other closely related species.

Materials And Methods
Cymbella janischii specimens were collected from six different geographical sites in Japan, in 2018-2019, as shown in Fig. 1a.Each single cell was washed with sterile water by pipetting under a stereo microscope (Olympus SZX10) and stored in a sample tube at -80°C until further use.Genomic DNA was ampli ed from a single cell using REPLI-g Mini kit (Qiagen).The samples were incubated for 10 h at 30°C.Ampli ed products were diluted 20 times with TE and then used as PCR template.
To annotate the genetic classi cation, BLASTN homology search was conducted on the nucleotide collection (nr/nt) database.To con rm the phylogenetic relationship with C. janischii from the US (Idaho) and its closely related species as previously reported by Nakov et al. (2014), each gene sequence was aligned with the corresponding genes from the following most closely related taxa using MAFFT ver.7 (Katoh and Standley 2013): C. tumida, C. proxima, C. baicalensis, C. stuxbergii, C. mexicana, C. janischii, D. dentata, D. geminata, D. siberica, and Encyonema triangulum (used for the outgroup).Aligned sequences were concatenated after removing ambiguous aligning regions (total 5863 bp), and processed via the maximum likelihood (ML) analysis with partitioning by genes and codons using RAxML ver.8.2.11 (Stamatakis 2014), with the GTR gamma model and 1,000 bootstraps.The pairwise distance (pdistance) between the sequences of specimens from Japan and the US was calculated with MEGA ver.7 (Kumar et al. 2016) using concatenated and aligned sequences with gaps removed.

Results
The sequences of samples from geographically distinct sites, morphologically identi ed as C. janischii in Japan, were determined with the rbcL (1473 bp), psaB (1958 bp), psbA (895 bp), 18S (1667 bp), and 28S (533 bp) genes (total 6526 bp).The sequences of all samples were identical.The BLASTN search for each gene sequence showed the highest homology to the sequences of C. janischii registered in Genbank with the following identity rates: rbcL 100%, psaB 99.95%, psbA 99.88%, 18S 99.88%, and 28S 100%.The degenerate nucleotides of existing sequence, "R" in psaB (KJ011751) and "Y" in 18S (KJ011622), were single nucleotides "A" and "C" in samples from Japan, respectively.The phylogenetic tree resulting from BLASTN search showed that each gene sequence in the Japanese specimens belonged to the same branch as C. janischii (Fig. S1).The ML tree based on the multigene analysis also showed the Japanese sequence to be in the same branch as C. janischii with 100% bootstrap support (Fig. 1b).Hence, our genetic analysis con rmed that all diatoms in question from Japan were C. janischii.The identities and the p-distance between the aligned sequences of specimens from Japan and the US were 99.952% and 0.0002 in the chloroplast genome (total 4200 bp of rbcL, psaB, and psbA), 99.907% and 0.0004 in the nuclear genome (total 2146 bp of 18S and 28S), and 99.937% and 0.0003 in the ve gene sequences (total 6346 bp), respectively.

Discussion
This study provides molecular genetic con rmation that the stalked diatoms recently found in Japanese streams are C. janischii, and enables an estimation for the invasion from the country of origin and the recent dispersal in Japan.Geographical distance and/or isolation along with environmental differentiation can lead to genetic diversity (Sexton et al. 2014;Fernández et al. 2017).In case of D. geminata, genetic divergence (p-distance up to 0.005 in a total 2230 bp region of chloroplast genome atpF-atpH, rbcS-rbcL, and secA-rbcR)) was observed in the indigenous areas in the Northern Hemisphere, whereas genetic homogeneity was found within New Zealand, which was compatible with specimens from several indigenous areas (Kilroy and Novis 2018).The variation in US (Colorado and Montana) specimens of D. geminata in that region corresponded to a p-distance of 0.001 (Kilroy and Novis 2018).In this study, all Japanese C. janischii specimens showed the same sequence, regardless of geographical distance, and were extremely similar in sequence to isolates from the country of origin, the US.Between the Japanese and the US specimens, the p-distance is 0.0002 in the chloroplast genome regions (rbcL, psaB, and psbA) and 0.0003 in the total region (6346 bp) including the nuclear genomes regions (18S and 28S).Since the genomic regions analyzed in this study and in D. geminata are not the same, it is not possible to make a strict comparison; however, the Japanese C. janischii specimens have higher similarity than that observed in D. geminata.These results indicate that C. janischii from the country of origin invaded and spread widely in Japan in recent times.Most likely, this occurred via some object or biotic material (e.g., shing gears, the transport of sh, their eggs, and water), such as has been discussed in the case of D. geminata (Kilroy and Novis 2018).Future studies will be needed to con rm if the observed genetic differences are due to intraspeci c polymorphisms, since the genetic information comes from a limited number of specimens from the country of origin and from a single cell in each Japanese specimen.
In rivers invaded by C. janischii, the landscape becomes degraded, as observed with D. geminata.Damage to shing for Plecoglossus altivelis, a prominent Japanese sh which feeds on algae, has been reported (Ashizawa and Kaji 2019).Thus, to prevent the further spread of this diatom, C. janischii needs to be detected and monitored at an early stage using highly sensitive methods such as environmental DNA analysis.The genetic data obtained here will contribute to the design of an early detection system for this species. Figures