Specimens were collected from Jiangxi, China, in 2022. Living specimens were initially frozen at -20 °C for 24 hours and subsequently thawed at room temperature for 2 hours to facilitate the extraction of soft parts. The soft parts were then fixed in 95% ethanol. Empty shells were cleaned, dried, and preserved at room temperature. Photographs were taken by camera and edited in Adobe Photoshop CC 2015 (Adobe, San Jose, US). Maps were made in ArcGIS Pro (Esri, Redlands, US).

Genomic DNA was extracted from muscle preserved in 95% ethanol using a TIANamp Marine Animals DNA Kit (Tiangen Biotech, China). The quality and concentration of the DNA were checked on 1% agarose gel electrophoresis and NanoDrop 2000 (Thermo Scientific, USA). The qualified genomic DNA sample was sent to Novogene (Beijing, China). Two datasets were used for phylogenetic analyses: the partial cytochrome c oxidase subunit 1 (COI) sequences and the mitochondrial genomes. The polymerase chain reaction (PCR) systems, conditions, and primer pairs of COI are followed by Chen et al. (2023). After quality controls, the library was successfully prepared and sequenced in the Illumina NovaSeq 6000 platform, yielding 12 Gb of data, with 2 × 150 bp paired-end reads. After discarding low-quality reads, clean reads were obtained and assembled de novo using CLC Genomic Workbench v. 12.0 (https://digitalinsights.qiagen.com). Contigs identified as mitogenome sequences were inspected manually for overlap at the beginning and end, resulting in a circular mitogenome. Geneious v. 11.0 (https://www.geneious.com) (Kearse et al. 2012) was used to check the complete mitogenome and analyze nucleotide composition. Strand asymmetry was calculated using the formulas: GC skew = (G−C)/(G+C) and AT skew = (A−T)/(A+T). The initial annotation of the mitogenome was carried out using the MITOS web server (http://mitos.bioinf.uni-leipzig.de/index.py) (Bernt et al. 2013). ARWEN (http://130.235.244.92/ARWEN) (Laslett and Canback 2008) was also used to identify the locations of all tRNA genes. The annotations of two rRNA genes were further refined based on the positions of neighboring genes. PCGs were initially identified using the Open Reading Frame Finder (ORF Finder) implemented on the NCBI website (http://www.ncbi.nlm.nih.gov/orffinder) and BLAST searches (http://blast.ncbi.nlm.nih.gov).

The COI dataset consists of 51 sequences of the family Unionidae and two outgroup taxa (Table 1). The mitochondrial genome dataset consists of 81 sequences (including 12 PCGs (excluding ATP8) and two rRNA genes) of the family Unionidae and nine outgroup taxa (Table 2). The COI sequences were aligned using MEGA X (Kumar et al. 2018) and checked manually. MUSCLE (Edgar 2004) was used for codon alignment of all PCGs implemented in MEGA X (Kumar et al. 2018). MAFFT (Katoh et al. 2019) with the Q-INS-i algorithm was used for the alignment of ribosomal genes (12S and 16S rRNAs). Gblocks (Castresana 2000) were used to exclude ambiguous regions from the alignment of each gene. PartitionFinder (Lanfear et al. 2017) was used to determine the best-fit partitioning schemes and substitution models under greedy search. Predefined data blocks for partitioning scheme searches were designated by gene region (rRNA gene) or codon position (PCG). Branch lengths were allowed to be unlinked, and model selection and partitioning schemes were determined using the AICc method (Table 3).

ML analyses were performed in IQ-TREE v. 1.6.12 (Minh et al. 2013) using the Ultrafast bootstrap approach (Minh et al. 2013) with 10,000 iterations. Bayesian inference (BI) analysis was conducted in MrBayes v. 3.2.6 (Ronquist et al. 2012). Four simultaneous runs with four independent Markov Chain Monte Carlo (MCMC) were implemented for 10 million generations, and trees were sampled every 10,000 generations with a burn-in of 25%. The convergence was checked with the average standard deviation of split frequencies < 0.01 and the potential scale reduction factor (PSRF) ~ 1. Trees were visualized in FigTree v.1.4.3 (http://tree.bio.ed.ac.uk/software/figtree/).

NCU_XPWU: Laboratory of Xiao-Ping Wu, Nanchang University (Nanchang, Jiangxi, China).

The alignment of the COI gene had lengths of 606 characters. Within these alignments, 260 sites were variable, and 236 sites were parsimony informative. The maximum likelihood and Bayesian analyses produced consistent phylogenies (Figs 1, 2). Although a single gene marker cannot effectively address inter tribes relationships in the subfamily Unioninae, the new species formed an independent branch that does not belong to any tribe. Its systematic position needs further verification through mitochondrial genome phylogenetic analysis.

Figure 1. 

Maximum likelihood tree inferred from COI gene sequences. Bootstrap supports are shown on the left of nodes on the tree if greater than 50%.

Figure 2. 

Bayesian inference tree inferred from COI gene sequences. Posterior probabilities are shown on the left of nodes on the tree if greater than 50%.

The length of the mitogenome is 16,240 bp. The maternal mitogenomes of the new species contain the 13 PCGs typically found in metazoan mitochondrial genomes, the type-specific F-orf described for all Unionida mitogenomes with the DUI system, 22 transfer RNA (tRNA), and two ribosomal RNA (rRNA) genes (Fig. 3A). Most genes were encoded on the light strand (L-strand), whereas 11 genes (COI, COII, COIII, ND3, ND4, ND4L, ND5, ATP6, ATP8, tRNA^Asp, and tRNA^His) were located on the heavy strand (H-strand). The location of the F-orf was between ND2 and tRNA^Glu, and the gene order of the mitogenomes was consistent with other Unioninae and Ambleminae species. The rrnS gene of it was located between tRNA^Arg and tRNA^Lys, while the rrnL gene was located between tRNA^Tyr and tRNA^Leu. The overall base composition was A (25.7%), T (38.1%), C (12.2%), and G (24.0%), with the A + T content of 63.8%. Nucleotide asymmetry of the mitochondrial strands was assessed by AT skew and GC skew that were −0.19 and 0.33, respectively.

Figure 3. 

A. Gene map of the F-type mitochondrial genome of Globunio mirificus gen. et sp. nov.; B. Maximum likelihood and Bayesian inference tree inferred from mitochondrial genome sequences. Bootstrap supports/posterior probabilities are shown on the left/right of nodes on the tree if they are greater than 50%.

The alignment of the mitogenomes had lengths of 11,547 characters. Within these alignments, 7,300 sites were variable, and 6,604 sites were parsimony informative. The maximum likelihood and Bayesian analyses produced largely consistent phylogenies (Fig. 3B). The new species formed an independent branch in the subfamily Unioninae and was sistered with the clade consisting of Unionini + Aculamprotulini with a relatively well support rate (BS/PP = 70/0.99).

Family Unionidae Rafinesque, 1820

Subfamily Unioninae Rafinesque, 1820

 Globunio mirificus Chen, Dai, Huang & Wu, sp. nov.

https://zoobank.org/C2C368AA-A510-4FDC-B58C-C037009B42CB

Figs 4, 6C, D

Holotype.

24_NCU_XPWU_GM01, ♀, Honggutan District [红谷滩区], Nanchang City [南昌市], Jiangxi Province [江西省], China, 24.68146°N, 109.69794°E, leg. Zhong-Guang Chen & Yu-Ting Dai, September 2022.

Figure 4. 

Globunio mirificus gen. et sp. nov. A. Holotype; B–K. Paratypes.

Paratypes.

n = 10, 24_NCU_XPWU_GM02, Honggutan District [红谷滩区], Nanchang City [南昌市], Jiangxi Province [江西省], China, 24.68146°N, 109.69794°E, leg. Zhong-Guang Chen, December 2021; 24_NCU_XPWU_GM03–11, other information same as holotype.

Diagnosis.

Same as the tribe.

Description.

Shell small-sized, elongated-globular, inflated, thick, solid. Anterior extremely short, inflated, and round; posterior long and flat. Anterior margin rounded, dorsal margin straightened, and slope downward at an obtuse angle, usually covered in fine upward wrinkles; ventral margin weakly curved. Umbo inflated, under dorsal margin, almost at the very front of the shell, often eroded. Central of shell with two rows of posterior dorsal spines (usually detached with only attachment marks), the anterior row longer and the posterior row shorter. Periostracum yellowish-green with olive-green rays and several thick growth lines. Hinge short and strong. Hinge developed. Left valve with two pseudocardinal teeth, anterior tooth small, triangular-shaped, posterior tooth well-developed, rectangular-shaped; right valve with a single well-developed, pyramidal pseudocardinal tooth. Both valves with two lateral teeth: left valve external weak, internal well-developed; right valve external well-developed, internal weak. Mantle attachment scars on the pallial line obvious. Anterior adductor muscle scars deep, samll; posterior adductor muscle scars shallow, orbicular-shaped. Umbo cavity open, deep. Nacre milky white.

Measurements.

Shell length 15.07–32.37 mm, width 9.83–21.54 mm, height 13.06–19.30 mm.

Etymology.

The species is named after the Latin mirificus for remarkable, referring to the remarkable shell morphology of this species.

Vernacular name.

奇异球蚌 (qí yì qiú bàng).

Distribution and ecology.

Known from three localities of the Changjiang River Basin: the Ganjiang River at Nanchang, the Dongtinghu Lake at Yueyang, and the Qingyijiang River at Wuhu (Fig. 5). Living in the slow-flowing rivers and lakes with muddy and sandy substrates alongside dozens of other freshwater mussels (Fig. 6).

Figure 5. 

Distribution of Globunio mirificus gen. et sp. nov. Solid star: type locality; hollow stars: other localities.

Figure 6. 

Habitats and live animals of Globunio mirificus gen. et sp. nov. A. Type locality, Ganjiang River at Honggutan of Nanchang in 2022; B. Type locality, Ganjiang River at Honggutan of Nanchang in 2023; C, D. Live animals.