Metschnikowia okinawaensis sp. nov. and Metschnikowia ryukyuensis sp. nov., two novel yeast species isolated from flowers and young fruits in Okinawa, Japan

Abstract

Two novel yeast species belonging to the genus Metschnikowia were isolated from flowers and young fruits in Okinawa, Japan. Five strains (JCM 35539, JCM 36740^T, JCM 36743, JCM 36744, and JCM 36747) and four strains (JCM 35542, JCM 35544^T, JCM 36745, and JCM 38279) were repeatedly isolated from samples collected over 2-3 y. Because BLAST analysis of D1/D2 large subunit ribosomal DNA sequences did not clearly resolve their phylogenetic placement and intragenomic polymorphisms were detected in the internal transcribed spacer region, their relationships were further assessed by whole-genome phylogenetic analysis using available genome data from Metschnikowia and related taxa. This analysis showed that the two novel species formed a distinct clade with Candida danieliae, the closest among described species in the dataset. Analysis of MAT-locus gene collinearity revealed that each novel species possessed strains of opposite mating types, and mating assays between compatible strains produced elongated asci containing needle-shaped ascospores, which is a characteristic feature of the genus Metschnikowia. By integrating phylogenomic, mating, morphological, and physiological evidence, we conclude that these strains represent two novel yeast species, for which the names Metschnikowia okinawaensis sp. nov. and Metschnikowia ryukyuensis sp. nov. are proposed.

1. Introduction

Metschnikowia, a member of the ascomycetous family Metschnikowiaceae, is a well-documented yeast genus widely distributed and frequently associated with the winemaking environment (Vicente et al., 2020). The genus was first described more than 100 y ago (Lachance, 2011) and currently comprises over 80 recognized species, including asexual taxa historically assigned to the genus Candida (Lachance, 2016). Although the number of described species has steadily increased, many undescribed taxa are expected to occur in geographically and ecologically underexplored regions. Most Metschnikowia species reproduce by multilateral budding, and many produce distinctive needle-shaped ascospores, one of the characteristic morphological hallmarks of the genus (Lachance, 2011).

Clavispora lusitaniae is a closely related member of the family Metschnikowiaceae, and has a mating-type (MAT) locus structure similar to that of Metschnikowia (Reedy et al., 2009). Similar to haploid cultures of Clavispora lusitaniae, the haplontic species of Metschnikowia possess a heterothallic reproductive system and a unique MAT region that lacks the ability to switch between the a and alpha configurations (Lee et al., 2018; Reedy et al., 2009). Therefore, comparative analyses of closely related genera are important for understanding the diversity of reproductive systems and their underlying genetic mechanisms.

The taxonomy of Metschnikowia has been progressively clarified through DNA sequence analysis, particularly ribosomal DNA (rDNA), enabling the delimitation of species previously recognized solely based on phenotype. For example, six former Candida species (C. chrysomelidarum, C. gelsemii, C. kofuensis, C. picachoensis, C. pimensis, and C. rancensis) have been reassigned to Metschnikowia based on sequence analysis (Kurtzman et al., 2018), and DNA sequencing has facilitated the discovery of several additional species (Chai et al., 2019). However, intragenomic copy polymorphisms are frequently observed in rDNA, particularly in the internal transcribed spacer (ITS) region, which complicates species identification using a single marker. Therefore, genome-based phylogenetic analysis provides a more reliable framework for species delineation and for resolving the relationships among novel and related taxa (Liu et al., 2024; Sipiczki et al., 2018).

Wild flowers represent a major reservoir of yeast biodiversity (Lachance, 2006), and numerous yeast species have been isolated from flowers and flower-visiting insects, including flies, beetles, and bees (de Oliveira Santos et al., 2015; de Vega et al., 2012, 2014; Lachance et al., 2001). Most Metschnikowia species have been isolated from fruits, flowers, floral nectar, and floricolous insects, and are regarded as one of the major yeast lineages associated with angiosperm flowers and their insect pollinators (Brysch-Herzberg, 2004; de Vega et al., 2018; Guzmán et al., 2013; Lachance, 2016). Some species have also been detected in insect guts (Tiwari et al., 2023; Yoshihashi & Degawa, 2023). However, a few species, including M. bicuspidata and M. pulcherrima, have been recorded in other sources, such as shrimp and human skin (Butinar et al., 2005; Kuan et al., 2016). Excluding M. reukaufii, most Metschnikowia species show narrow ecological ranges and high specialization (Guzmán et al., 2013). Certain taxa also exhibit highly restricted geographic distributions: M. cerradonensis occurs only in a small region of Brazil (Rosa et al., 2007); M. drakensbergensis in the Drakensberg Mountains of South Africa (de Vega et al., 2014); and M. hamakuensis, M. kamakouana, and M. mauinuiana appear to be confined to the Hawaii Islands, where related haplontic taxa have been associated with nitidulid beetles (Lachance et al., 2005). However, some species related to Metschnikowia have been reported in geographically distant localities, suggesting that the ecology and dispersal of this lineage remain incompletely understood (Lachance & Fedor, 2014; Rosa et al., 2010).

Despite these studies, vast geographic regions remain mycologically unexplored and numerous undescribed species are expected to exist. In this study, we describe the strains isolated from flowers and young fruits on Okinawa Island, Japan, and examine their physiological, reproductive, and phylogenomic characteristics. Based on these analyses, we propose that these strains represent two novel species of the genus Metschnikowia.

2. Materials and methods

2.1 Yeast isolation and identification

Flowers of blue morning glory Ipomoea indica, shell ginger Alpinia zerumbet, Bengal clock vine Thunbergia grandiflora, and large-flowered pink sorrel Oxalis debilis were used for yeast isolation. Only samples with no visible insects were selected to focus on yeasts present on flower and young-fruit surfaces, and avoid direct isolation from insects present at the time of sampling. Samples were collected at the Nishihara campus of the University of the Ryukyus (26°14′51.3″N, 127°45′54.5″E) in Nishihara, Okinawa Prefecture, Japan, located on the main island of the Ryukyu Archipelago. Sampling was conducted between May and Aug 2020-2022 (Table 1). The region has a subtropical climate. Ipomoea indica and A. zerumbet are native to Okinawa, whereas T. grandiflora and O. debilis have been introduced and naturalized. After collection, the samples were sealed in polyethylene zipper bags and transported to the laboratory.

Table 1 - Strains used in this study.

Strain information			Isolation information		Sequence accession No.		Mating characteristics	Genome assembly
Lab ID	JCM ID	Mating type	Host plant	Collection date	ITS	D1/D2	Mating combinations yielding asci	Accession No.	Size (bp)	G+C content (%)
Metschnikowia okinawaensis sp. nov.
TN125	JCM 35539	α	Ipomoea indica	Aug, 2020			TN466, TN924, TN938	BAAKFQ010000001-BAAKFQ010004025	13,981,155	48.1
TN321	JCM 36740T	α	Alpinia zerumbet	Aug, 2020	LC805679	LC805669	TN466, TN924, TN938	BAAKFT010000001-BAAKFT010004286	14,051,056	48.1
TN466	JCM 36743	a	Alpinia zerumbet	May, 2021			TN125, TN321	BAAKFU010000001-BAAKFU010004585	14,153,940	48.1
TN924	JCM 36744	a	Alpinia zerumbet	Jun, 2022			TN125, TN321
TN938	JCM 36747	a	Thunbergia grandiflora	Jun, 2022			TN125, TN321
Metschnikowia ryukyuensis sp. nov.
TN318	JCM 35542	a	Alpinia zerumbet	May, 2021			TN926	BAAKFR010000001-BAAKFR010002626	12,772,582	49.6
TN488	JCM 35544T	a	Oxalis debilis	May, 2021	LC710264	LC710256	TN926	BAAKFS010000001-BAAKFS010003145	14,872,828	49.7
TN925	JCM 36745	a	Alpinia zerumbet	Jun, 2022			TN926
TN926	JCM 38279	α	Alpinia zerumbet	Jun, 2022			TN318, TN488, TN925	BAAKFV010000001-BAAKFV010002540	12,689,900	49.7

In the laboratory, plant material was placed in sterile 50-mL conical tubes, 10 mL of sterilized water was added, and the samples were vigorously vortexed to wash the floral surfaces. The resulting suspensions were serially diluted and spread on yeast extract agar (YEA; 0.5% yeast extract, 3% d-glucose, and 1.5% agar), yeast extract-peptone-dextrose (YPD; 1% yeast extract, 2% peptone, 2% d-glucose, and 2% agar), or potato dextrose agar (PDA). All media were supplemented with ampicillin (100 µg/mL) and chloramphenicol (100 µg/mL) to inhibit bacterial growth. Plates were incubated at 30 °C for 2-3 d. Colonies were initially classified by color and surface morphology, and cell morphology was examined using light microscopy. Purified yeast strains were suspended in 15% (v/v) glycerol and stored at −80 °C.

2.2 DNA sequencing and phylogenetic analysis

Genomic DNA was extracted from overnight YPD cultures using a Wizard Genomic DNA Purification Kit (Promega, Madison, WI, USA) following the manufacturer's protocol. The D1/D2 domain of the 26S LSU rDNA gene and the ITS region were PCR-amplified using KOD FX Neo DNA polymerase (TOYOBO, Tokyo, Japan). Primers NL1 (5′-GCATATCAATAAGCGGAGGAAAAG-3′) and NL4 (5′-GGTCCGTGTTTCAAGACGG-3′) were used for D1/D2 amplification (Kurtzman & Robnett, 1998), and ITS1 (5′-TCCGTAGGTGAACCTGCGG-3′) and ITS4 (5′-TCCTCCGCTTATTGATATGC-3′) for ITS amplification (White et al., 1990). PCR conditions were 94 °C for 10 min; 30 cycles of 98 °C for 10 s, 52 °C for 30 s, and 68 °C for 30 s; followed by 68 °C for 5 min. PCR products were purified and subjected to Sanger sequencing (Eurofins Genomics or Thermo Fisher Scientific) using the same primers.

For the D1/D2 LSU rDNA phylogeny, the sequences of Metschnikowia species, Candida danieliae, closely related undescribed Metschnikowia lineages, and outgroup taxa were retrieved from the NCBI database. The dataset included 58 Metschnikowia species for which D1/D2 sequences were available, with Australozyma species used as outgroups. The sequences were aligned using MAFFT (Katoh & Standley, 2013), and a maximum-likelihood tree was constructed using IQ-TREE 3 (Wong et al., 2025).

Because all the strains produced mixed chromatograms for the ITS region, consistent with intragenomic rDNA polymorphisms, ITS amplicons were cloned before sequencing. PCR products amplified with Cloning_ITS1_Fw (5′-GCCGCCATCCAGTTTTCCGTAGGTGAACCTGCGG-3′) and Cloning_ITS4_Rv (5′-GAGACCGGCAGATCCTCCTCCGCTTATTGATATGC-3′) were assembled into pFA6a-kanMX6 linearized by inverse PCR using Cloning_pFA6a_inv_Fw (5′-GGATCTGCCGGTCTCCCTATAGTG-3′) and Cloning_pFA6a_inv_Rv (5′-AAACTGGATGGCGGCGTTAG-3′), via Gibson assembly (Gibson Assembly Master Mix; New England Biolabs). Constructs were transformed into Escherichia coli DH5α (TaKaRa). Single colonies were selected, plasmid DNA was purified (LaboPass^TM Plasmid DNA Purification Kit Mini; Hokkaido System Science Co., Ltd.), and inserts were sequenced using Cloning_pFA6a_seq (5′-GAAAGTAATATCATGCGTCAATCGTATGTG-3′). For each strain, ≥3 distinct ITS haplotypes were recovered; the most frequent haplotype was used as the representative for downstream analyses (JCM 36740^T, LC805679; JCM 35544^T, LC710264). All alternative haplotypes are listed in Table S1.

Draft genome sequences were generated using the Illumina MiSeq System (Illumina, San Diego, CA, USA). Paired-end libraries (2 × 300 bp) were constructed according to the manufacturer's protocol and sequenced to approximately 200× coverage. Adapter-trimmed reads were assembled de novo using SPAdes v. 4.2.0 with default parameters.

For the phylogenetic analysis, one or more representative species were selected from each Metschnikowia clade recognized by Liu et al. (2024), and the corresponding genome data were retrieved from the NCBI database. Including the datasets generated in this study, twenty-three genome datasets were analyzed. Single-copy BUSCO genes (566 genes) were identified in each genome using BUSCO ver. 6.0.0 (Manni et al., 2021). Genes shared by all 23 genomes were extracted, individually aligned using MAFFT (Katoh & Standley, 2013), and concatenated into a supermatrix using custom Python scripts. A maximum-likelihood phylogenetic tree was constructed using IQ-TREE 3 (Wong et al., 2025) with 1,000 bootstrap replicates. The resulting tree was visualized using iTOL (Letunic & Bork, 2024).

2.3 Determination of mating locus and type

Yeasts of the genus Metschnikowia are heterothallic. In the many haplontic species, each strain carries either the MAT-a or MAT-α mating-type (Lee et al., 2018). Moreover, mating-type can be inferred from the characteristic gene order, or collinearity, within the MAT locus, which remains conserved for each mating-type (Lee et al., 2018). In this study, we developed an analytical procedure based on Lee et al. (2018).

First, homolog searches were performed against draft genome assemblies of JCM 35539, JCM 36740^T, JCM 36743, JCM 35542, JCM 35544^T, and JCM 38279 using MAT locus-associated genes (PAP1, OBP1, PIK1, and RCY1) and the a1, a2, and α-specific amino acid sequences of Clavispora lusitaniae (GenBank: FJ524850.1; RefSeq: XP_002614908.1) as queries. Searches were performed using BLAST (BLAST+). For each query, the best-hit locus (lowest E-value) was retained, and when multiple candidates were detected, loci colocalizing on the same contig were prioritized (Reedy et al., 2009). Sequences of hit loci were extracted using Seqkit, and their arrangement was visualized using SnapGene. Based on the inferred MAT-locus gene collinearity, MAT-specific regions were identified, and primers were designed for PCR confirmation. Strains were assigned to a mating-type when PCR yielded a product of the expected size exclusively in either the MAT-a or MAT-α reaction system.

JCM 36740^T-group MAT-a primers

F: 5′-AAATGGGCCACAGACTAGCA-3′

R: 5′-ATCGTTTTAGACCTTTGGCT-3′

JCM 36740^T-group MAT-α primers

F: 5′-GAGAGTCGTCCAGTTAAGGT-3′

R: 5′-TGAGACAGGAAAGGAGTTCA-3′

JCM 35544^T-group MAT-a primers

F: 5′-TTCATTATGATTTCCCAAGT-3′

R: 5′-TTATAATCGGTACAAAAGCA-3′

JCM 35544^T-group MAT-α primers

F: 5′-CGCGATTCAATACAATGCTA-3′

R: 5′-GCTAAAACAGAGGAAAGTTC-3′

2.4 Physiological and morphological characterization

Nine yeast strains (JCM 35539, JCM 35542, JCM 35544^T, JCM 36740^T, JCM 36743, JCM 36744, JCM 36745, JCM 36747, and JCM 38279) were characterized physiologically and morphologically following Kurtzman et al. (2011) and recent studies on Metschnikowia (Chai et al., 2019; de Vega et al., 2018). Formation of pseudo-hyphae, true hyphae, and asci was assessed on the following media at 25 °C for up to 4 wk: 5% malt extract agar (MEA; 5% malt extract, 0.5% peptone, and 1.5% agar), corn meal agar (CMA), PDA, yeast morphology agar (YMoA), yeast carbon base-ammonium sulfate (YCBAS) agar (1.1% yeast carbon base, 0.01% ammonium sulfate, and 1.8% agar), V8 juice agar (V8A; 20% V8 juice, 0.2% calcium carbonate, and 1.5% agar), and diluted V8 juice agar (1:10 dilution).

Carbon and nitrogen assimilation tests were performed on agar plates and scored within 7-21 d at 25 °C. Carbohydrate fermentation was assessed in 5 mL of liquid medium containing 2% (w/v) sugar in Durham fermentation tubes and incubated at 25 °C for 14 d. Cycloheximide tolerance was evaluated in YPD supplemented with 0.01% and 0.1% (w/v) cycloheximide. The diazonium blue B (DBB) reaction was performed in a solid medium. Growth at various temperatures (19, 25, 30, 35, 37, and 40 °C) was evaluated on YPD after 7 d.

The cells were examined directly from the agar plates without washing or fixation. Images were acquired using cellSens software (Olympus, Tokyo, Japan) on a BX53 differential interference contrast (DIC) microscope (Olympus) equipped with UPlanSApo 60×/1.35 NA and 100×/1.45 NA oil-immersion objectives and an ORCA-Spark camera (Hamamatsu Photonics, Hamamatsu, Japan).

3. Results and Discussion

Identification and delineation of novel species

From 2020 to 2022, 20 yeast strains were isolated from flowers and young fruits collected on Okinawa Island, Japan. These included JCM 35539 (=TN125), Anthracocystis heteropogonicola, Moesziomyces aphidis, and Rhodotorula paludigena from one Ipomoea indica flower; JCM 35542 (=TN318), JCM 36743 (=TN466), JCM 36744 (=TN924), JCM 36745 (=TN925), JCM 38279 (=TN926), Hanseniaspora thailandica, Hanseniaspora uvarum, Kodamaea nitidulidarum, and Kodamaea ohmeri from nine Alpinia zerumbet flowers; JCM 36740^T (=TN321) and Saturnispora silvae from four young fruits of A. zerumbet; JCM 36747 (=TN938) and Nakaseomyces nivariensis from seven Thunbergia grandiflora flowers; and JCM 35544^T (=TN488), Candida wangnamkhiaoensis, Metschnikowia koreensis, Moesziomyces aphidis, and Rhodotorula glutinis from one Oxalis debilis flower (Table 1). Metschnikowia strains recovered repeatedly over multiple years, suggesting their recurrence in the plant samples from Okinawa Island. Notably, M. hawaiiensis, the first known haplontic Metschnikowia, has also been isolated from Ipomoea acuminata, a synonym of I. indica (Lachance et al., 1990). Metschnikowia colocasiae and M. similis were isolated from I. indica in Costa Rica (Lachance & Bowles, 2004). No insects were observed on the sampled flowers at the time of collection and only flowers without visible insects were selected for yeast isolation. Previous studies have suggested that nitidulid beetles may play an important, although still poorly understood, role in the ecology of many haplontic Metschnikowia species, including endemic Hawaiian taxa (Lachance et al., 2005; Lachance & Fedor, 2014). Taken together, these observations suggest that novel Metschnikowia strains are persistent components of the local plant-associated yeast community in Okinawa.

Five strains (JCM 35539, JCM 36740^T, JCM 36743, JCM 36744, and JCM 36747) and four other strains (JCM 35542, JCM 35544^T, JCM 36745, and JCM 38279) shared identical D1/D2 LSU sequences. Although three ITS sequence variants were recovered from each strain, all ITS haplotypes were clustered within the respective strain groups, and the most frequent haplotype was adopted as the representative ITS sequence (Supplementary Table S1).

BLASTn searches against GenBank (Altschul et al., 1997) revealed that JCM 36740^T (=TN321) shared only 87.06% identity with its closest hit, Metschnikowia sp. (MG835363; D1/D2 sequence from insect guts), differing by 66 substitutions and 19 gaps. In contrast, JCM 35544^T (=TN488) showed 90.21% identity with Metschnikowia sp. GCMCC2.2748 (PQ309482) with 38 substitutions and 12 gaps.

In the D1/D2 phylogenetic tree that included available Metschnikowia species, Candida danieliae, closely related undescribed lineages, and Australozyma species as outgroups, the two novel species formed a distinct lineage within Metschnikowia (Fig. S1). This result was consistent with the BLASTn searches, which did not clearly identify closely related species based on the D1/D2 sequences alone. Although the D1/D2 tree provided broad taxon sampling across Metschnikowia, deeper relationships within the genus were not fully resolved, which is consistent with the limited resolution of this marker for some lineages. These results indicated that additional phylogenetic resolution was required.

Comparison of the representative ITS sequences revealed that JCM 36740^T (=TN321) differed from Clavispora sp. UFMG-CM-Y2963 (MG694261) with 87.13% identity (query coverage 79%), whereas JCM 35544^T (=TN488) differed from Danielia reshetovae CBS 11556 (KY102569) with 87.36% identity (query coverage 72%). A match to an uncultured fungal genomic DNA sequence (consensus03331, OU942036.1; 87.36% identity, 33 substitutions, 6 deletions) was excluded because it was not derived from a cultured strain and could not be taxonomically assigned.

In alignment with the findings observed in the D1/D2 region, we noted that higher similarity values were present among strains categorized within the same species. However, these observations were not pursued for further analysis. In addition, JCM 36740^T and JCM 35544^T differed by 38 substitutions and 15 gaps in the D1/D2 domain and 33 substitutions and 17 gaps in the ITS region, indicating substantial divergence consistent with species-level separation (Chai et al., 2019; de Vega et al., 2018). Notably, ITS-based comparisons showed relatively low query coverage and, together with the discordance between the D1/D2 and ITS results, complicated the identification of the closest named relatives. Such disparity likely reflects intragenomic rDNA polymorphisms or heterogeneity, which are common in Metschnikowia (Lachance et al., 2020).

To further clarify the relationships among these isolates, we performed a phylogenetic analysis based on single-copy BUSCO genes shared among the analyzed genomes. The resulting tree placed JCM 36740^T and JCM 35544^T in a distinct clade, together with Candida danieliae within the Metschnikowiaceae (Fig. 1). This clade comprises only the three taxa in the present dataset, indicating that C. danieliae is the closest described relative of the two novel species among the taxa analyzed. However, JCM 36740^T and JCM 35544^T were clearly separated from C. danieliae and each other, supporting their recognition as independent, novel species. This result also indicates that phylogenomic analysis was essential for resolving the placement of these isolates because comparisons based solely on D1/D2 and ITS sequences did not consistently identify their nearest described relatives.

Fig. 1. Phylogenomic relationships of the two novel species within Metschnikowiaceae.

A phylogenetic tree was constructed from concatenated single-copy BUSCO genes (566 genes) shared among all analyzed genomes using IQ-TREE 3. Bootstrap values (>50%) based on 1,000 replicates are shown at the nodes. Six strains representing two novel species (JCM 35539, JCM 35542, JCM 35544^T, JCM 36740^T, JCM 36743, and JCM 38279) formed a distinct clade with Candida danieliae. The strain names and genome accession numbers are shown in the figure labels. Bar, 0.1 substitutions per site.

The key physiological characteristics of species assigned to Metschnikowia include assimilation of d-sucrose, d-maltose, d-trehalose, d-xylose, d-gluconate, 2-keto-d-gluconate, N-acetyl-d-glucosamine, d-cellobiose, d-salicin, l-sorbose, d-mannitol, d-glucitol, carbohydrate fermentation, and growth on medium containing 0.01% cycloheximide, as well as growth at 30 and 37 °C (Lachance, 2011). In this study, the two novel isolates were distinguished from the most closely related Metschnikowia species based on their phenotypic profiles. JCM 36740^T (=TN321) assimilated raffinose (weak), methyl-α-d-glucoside (slow), d-ribose (slow), l-lysine, 0.01% cycloheximide (slow), whereas Candida danieliae assimilated few to none of these substrates (Table 2). Conversely, JCM 36740^T did not assimilate d-xylose, citrate, 2-keto-d-gluconate, d-glucuronate, whereas C. danieliae assimilated these carbon sources (Table 2). JCM 36740^T fermented trehalose within 14 d.

Table 2 - Diagnostic physiological characteristics of Metschnikowia okinawaensis sp. nov., Metschnikowia ryukyuensis sp. nov., and the related species Candida danieliae.

		JCM 36740T	JCM 35544T	Candida danieliae
Fermentation	Glucose	+	+	+
	Galactose	−	−	−,+
	Trehalose	s/+	s/+	n
Carbon assimilation	Raffinose	w	w	−
	Galactose	s/+	s/+	d
	Methyl-α-d-glucoside	s/+	+	−
	d-Xylose	−	s/+	+
	d-Ribose	s/+	s/+	−
	Ethanol	s/+	s/w	+
	Glycerol	+	+	d
	Ribitol	s/+	s/+	+
	dl-Lactate	−	s/+	−
	Citrate	−	s/+	+
	d-Gluconate	+	+	d
	d-Glucosamine	s/+	−	+
	2-keto-d-gluconate	−	−	+
	Xylitol	s/+	+	+
	50% Glucose	−	s/w	−
	d-Glucuronate	−	−	+
	l-Lysine	+	s/+	−
	10% NaCl / 5% Glucose	s/+	+	n
	0.01% Cycloheximide	s/+	s/+	d

Species/strain numbers: 1, Metschnikowia okinawaensis sp. nov. (type strain JCM 36740^T); 2, Metschnikowia ryukyuensis sp. nov. (type strain JCM 35544^T); 3, Candida danieliae (data from The Yeasts Database: https://theyeasts.org).

Growth reactions: +, positive; −, negative; w, weak; s, slow; v, variable; d, delayed; n, no data.

JCM 35544^T (=TN488) assimilated raffinose (weak), methyl-α-d-glucoside, d-ribose (slow), dl-lactate (slow), and l-lysine (slow), whereas C. danieliae failed to assimilate most of these substrates (Table 2). Conversely, JCM 35544^T did not assimilate d-glucosamine, 2-keto-d-gluconate, d-glucuronate, whereas C. danieliae assimilated these carbon sources (Table 2). JCM 35544^T fermented trehalose within 14 d. The physiological characteristics of JCM 35544^T were broadly similar to those of JCM 36740^T; however, JCM 35544^T assimilated d-xylose (slow), dl-lactate (slow), citrate (slow), and grew at 37 °C, but failed to assimilate d-glucosamine. Taken together, these phenotypic differences provide practical diagnostic characteristics that distinguish the two novel species from C. danieliae and from each other.

Table 2 summarizes the diagnostic physiological characteristics differentiating the two novel species from related Metschnikowia taxa. Based on these molecular and phenotypic lines of evidence, we propose the novel species Metschnikowia okinawaensis sp. nov. (JCM 36740^T = TN321) and Metschnikowia ryukyuensis sp. nov. (JCM 35544^T = TN488) to accommodate these yeast strains.

Determining MAT-locus gene collinearity using draft genome assemblies

Following the procedure described in the Materials and Methods section 2.3, the gene arrangement (collinearity) of the MAT locus was inferred from the contig data of JCM 35539, JCM 36740^T, JCM 36743, JCM 35542, JCM 35544^T, and JCM 38279, and the corresponding regions were extracted and visualized using SnapGene. In JCM 35539, JCM 36740^T, and JCM 38279, the gene order RCY1 (→), PAP1 (←), ALPHA (→), PIK1 (→), and OBP1 (→) was obtained, consistent with MAT-α (Fig. 2A, B).

Fig. 2. Organization and mating-type determination of the MAT locus in the two novel species.

A: Gene organization of the MAT locus in the JCM 36740^T group, B: Gene organization of the MAT locus in the JCM 35544^T group. The bottom panel shows an inverted duplication of the MAT-a locus observed in strain JCM 35544^T. C: PCR-based determination of mating-type in the JCM 36740^T group using a- and α-specific primers. Strain identities and mating types are indicated below the lanes. Metschnikowia torresii (CBS 5152) and Metschnikowia drosophilae (CBS 8809) were included as negative controls. D: PCR-based determination of mating-type in the JCM 35544^T group using a- and α-specific primers. Strain identities and mating types are indicated below the lanes. JCM 35539 from the other novel species group, as well as Metschnikowia torresii (CBS 5152) and Metschnikowia drosophilae (CBS 8809), were included as negative controls. For both panels (C, D), a 100 bp DNA ladder was used as a size marker, and arrowheads indicate the expected PCR products.

In contrast, JCM 36743, JCM 35542, and JCM 35544^T showed the order PAP1 (←), OBP1 (→), PIK1 (→), A2-1 (→), A2-2 (→), and RCY1 (→), consistent with MAT-a (Fig. 2A, B). The reference Clavispora lusitaniae MAT-a locus encodes A1 and A2. However, A1 was not detected in any of the contigs of the new Metschnikowia species. Instead, two sequences corresponding to A2 were identified and designated as A2-1 and A2-2.

For each strain, the entire MAT locus and individual loci were visualized, revealing that strains of the same species and mating-type (JCM 35539 vs. JCM 36740^T; JCM 35542 vs. JCM 35544^T) showed identical MAT locus sequences across their respective regions. In contrast, strains of the same species but different mating types (JCM 36743 [a] vs. JCM 35539 [α]; JCM 38279 [α] vs. JCM 35542/JCM 35544^T [a]) showed sequence differences even in the shared MAT-locus-associated genes. Overall, MAT-locus gene collinearity between the two novel species was conserved, however, nucleotide sequences of the corresponding MAT-locus-associated genes differed.

Although most strains harbored a single MAT locus in their draft assemblies, JCM 35544^T contained two MAT-a loci arranged in an inverted orientation and separated by a short 248-bp linker sequence (Fig. 2B). In the assembly, the two PAP1 genes were located approximately 20 kb apart, with the 248-bp linker sequence positioned between them. To verify this arrangement, PCR was attempted using primers designed to flank or anneal the short linker region between the two PAP1 genes. However, due to the limited length of the linker sequence, it is difficult to design primers with sufficient specificity for unambiguous validation. As a result, multiple amplification products of varying lengths were obtained, preventing the confirmation of the inferred gene arrangement. Therefore, this pattern may reflect either an inverted duplication in vivo or an assembly-derived artifact, the origin of which remains unresolved. Although the unusual configuration observed in JCM 35544^T could not be validated experimentally, it suggests a possible structural variation at the MAT locus in this lineage and warrants further investigation using long-read sequencing or targeted locus validation.

Determination of mating-type using PCR and mating assays

PCR was performed using the mating-type primers designed in this study. PCR assays of JCM 35539, JCM 36740^T, JCM 36743, JCM 36744, and JCM 36747 with JCM 36740^T-group MAT-a and MAT-α primer sets identified JCM 35539 and JCM 36740^T as MAT-α, and JCM 36743, JCM 36744, and JCM 36747 as MAT-a (Fig. 2C). MAT-α strains (JCM 35539 and JCM 36740^T) were then crossed with MAT-a strains (JCM 36743, JCM 36744, and JCM 36747) and co-cultured on seven media (YCBAS, CMA, V8A, diluted V8A [1:10], YMoA, 5% MEA, and PDA) at 25 °C for >4 wk. Elongated asci containing two needle-shaped ascospores were observed exclusively on the YCBAS plates and at a low frequency after two days (Fig. 3C, D).

Fig. 3. Morphological characteristics of Metschnikowia okinawaensis sp. nov.

A: Source of isolation. B: Budding cells after 3 d on YM agar at 25 °C. C, D: Elongated asci (arrowhead) containing needle-shaped ascospores formed after mixing opposite mating types on YCBAS agar and incubating for 2 d and 2 wk, respectively, at 25 °C. Bars: 10 µm.

Similarly, PCR assays of JCM 35542, JCM 35544^T, JCM 36745, and JCM 38279 with the JCM 35544^T-group primers identified JCM 35542, JCM 35544^T, and JCM 36745 as MAT-a and JCM 38279 as MAT-α (Fig. 2D). The MAT-α strain JCM 38279 was crossed with each of the three MAT-a strains and co-cultured under identical conditions on the same seven media. Elongated asci usually containing a single needle-shaped ascospore were observed only on YCBAS at a low frequency after two weeks (Fig. 4C, D).

Fig. 4. Morphological characteristics of Metschnikowia ryukyuensis sp. nov.

A: Source of isolation. B: Budding cells after 3 d on YM agar at 25 °C. C, D: Elongated asci (arrowhead) containing a single needle-shaped ascospore formed after mixing opposite mating types on YCBAS agar and incubating for 2 wk at 25 °C. Bars: 10 µm.

Interspecific crosses between complementary mating types of the two species did not produce asci or ascospores under the tested conditions.

Because the mating results were consistent with the mating-type determined by PCR using primers designed from contig-based sequence information, we considered the PCR-based mating-type assignments to be reliable. Furthermore, the formation of elongated asci and needle-shaped ascospores, which are characteristic morphological features of Metschnikowia, supports the conclusion that these new yeasts belong to the genus Metschnikowia.

3.1. Taxonomy

Metschnikowia okinawaensis sp. nov. Yamamoto, Takekata & Seike　(Fig. 3)

MycoBank no.: MB 853777

Etymology: Metschnikowia okinawaensis (o.ki.na.wa.en′sis. N.L. adj. okinawaensis, referring to the Okinawa Prefecture, Japan, where the type strain was first isolated) (Fig. 3A).

After 3 d at 25 °C on YM agar, cells are ovoid to ellipsoidal (1.7-5.9 × 2.4-7.2 µm) with multilateral budding (Fig. 3B). The colonies are white to cream-colored, convex, smooth, and have complete margins. Pseudohyphae are formed on YCBAS agar at 25 °C after 4 wk. When strains of opposite mating types are co-cultured on YCBAS at 25 °C, elongated asci containing two needle-shaped ascospores are observed after two days. Fermentation is positive for d-glucose (3 d) and d-trehalose (14 d), but negative for d-galactose, d-sucrose, d-maltose, d-lactose, and d-raffinose.

Carbon assimilation is positive for d-glucose, d-sucrose, d-raffinose (weak), d-galactose (slow), d-trehalose, d-maltose, d-melezitose, methyl-α-d-glucoside (slow), d-cellobiose, d-salicin, l-sorbose, d-ribose (slow), ethanol (slow), glycerol, ribitol (slow), d-mannitol, d-glucitol, succinate, d-gluconate, d-glucosamine (slow), N-acetyl-d-glucosamine, hexadecane (slow/weak), xylitol (slow), and d-glucono-1,5-lactone, and negative for inulin, d-melibiose, lactose, soluble starch, l-rhamnose, d-xylose, l-arabinose, d-arabinose, galactitol, myo-inositol, dl-lactate, citrate, 2-keto-d-gluconate, methanol, erythritol, and d-glucuronate. Nitrogen assimilation is positive for cadaverine, l-lysine, and ethylamine but negative for nitrate and nitrite. Growth is positive in amino acid-free medium and negative in vitamin-free medium. Growth is negative in 50% (w/v) glucose, and positive (slow) in 10% (w/v) NaCl and 5% (w/v) glucose. No growth is observed in the presence of 0.1% cycloheximide, whereas slow growth occurred at 0.01%. Growth occurs at 19, 25, 30, and 35 °C (weak at 35 °C for JCM 36740^T), but not at 37 and 40 °C. The DBB test results are negative.

Holotype. The strain JCM 36740^T, isolated from Alpinia zerumbet, was collected from the Nishihara campus of the University of the Ryukyus, Okinawa Main Island, Japan. The strain is preserved in a metabolically inactive state at the Japan Collection of Microorganisms (JCM) as JCM 36740^T. Ex-type cultures in a metabolically inactive state were deposited at the National Institute of Technology and Evaluation (NITE), NBRC 116651, and the Westerdijk Institute, Netherlands, as CBS 18606.

Sequences. GenBank/EMBL/DDBJ accession numbers: LSU D1/D2, LC805669; ITS, LC805679; genome accession numbers, BAAKFT010000001-BAAKFT010004286.

Paratypes. JCM 35539 (=NBRC 115685 = CBS 18007; genome accession numbers, BAAKFQ010000001-BAAKFQ010004025), JCM 36743 (=NBRC 116654 = CBS 18607; genome accession numbers, BAAKFU010000001-BAAKFU010004585), JCM 36744 (=NBRC 116655 = CBS 18608), and JCM 36747 (=NBRC 116658 = CBS 18609).

Metschnikowia ryukyuensis sp. nov. Yamamoto, Takekata & Seike　(Fig. 4)

MycoBank no.: MB 853820

Etymology: Metschnikowia ryukyuensis (ryu.kyu.en′sis. N. L. adj. ryukyuensis, referring to the University of the Ryukyus in Okinawa Prefecture, Japan, where the type strain was first isolated (Fig. 4A).

After 3 d at 25 °C on YM agar, cells are ovoid to ellipsoidal (2.0-3.6 × 2.8-5.2 µm) with multilateral budding (Fig. 4B). The colonies are white to cream-colored, convex, smooth, and have complete margins. Pseudohyphae are formed on CMA at 25 °C after four wk. When strains of opposite mating types are co-cultured on YCBAS medium at 25 °C, elongated asci usually containing a single needle-shaped ascospore are observed after two weeks. Fermentation is positive for d-glucose (3 d) and d-trehalose (14 d), but negative for d-galactose, d-sucrose, d-maltose, d-lactose, and d-raffinose.

Carbon assimilation is positive for d-glucose, d-sucrose, d-raffinose (weak), d-galactose (slow), d-trehalose, d-maltose, d-melezitose, methyl-α-d-glucoside, d-cellobiose, d-salicin, l-sorbose, d-xylose (slow), d-ribose (slow), ethanol (slow/weak), glycerol, ribitol (slow), d-mannitol, d-glucitol, dl-lactate (slow), succinate, citrate (slow), d-gluconate, N-acetyl-d-glucosamine, hexadecane (slow/weak), xylitol, and d-glucono-1,5-lactone but negative for Inulin, d-melibiose, lactose, soluble starch, l-rhamnose, l-arabinose, d-arabinose, galactitol, myo-inositol, d-glucosamine, 2-keto-d-gluconate, methanol, erythritol, and d-glucuronate. Nitrogen assimilation is positive for cadaverine, l-lysine (slow), and ethylamine but negative for nitrate and nitrite. Growth is positive in amino acid-free medium and negative in vitamin-free medium. Growth is positive in 50% (w/v) glucose (slow/weak), and 10% (w/v) NaCl and 5% (w/v) glucose. No growth is observed in the presence of 0.1% cycloheximide, whereas slow growth occurred at 0.01%. Growth occurs at 19, 25, 30, 35, and 37 °C (slow), but not at 40 °C. The DBB test results are negative.

Holotype. The strain JCM 35544^T, isolated from O. debilis was collected at the Nishihara campus of the University of the Ryukyus, Okinawa Main Island, Japan. The strain is preserved in a metabolically inactive state at JCM as JCM 35544^T. Ex-type cultures in a metabolically inactive state were deposited at NITE (NBRC 115690) and Westerdijk Institute (CBS 18011).

Sequences. GenBank/EMBL/DDBJ accession numbers: LSU D1/D2, LC710256; ITS, LC710264; genome accession numbers, BAAKFS010000001-BAAKFS010003145.

Paratypes. JCM 35542 (=NBRC 115688 = CBS 18010; genome accession numbers, BAAKFR010000001-BAAKFR010002626), JCM 36745 (=NBRC 116656 = CBS 18610), and JCM 38279 (=NBRC 117439 = CBS 19552; genome accession numbers, BAAKFV010000001-BAAKFV010002540).

Disclosure

The authors declare that there are no conflicts of interest.