An integrative taxonomy approach reveals Saccharomyces chiloensis sp. nov. as a newly discovered species from Coastal Patagonia

Species delineation in microorganisms is challenging due to the limited markers available for accurate species assignment. Here, we applied an integrative taxonomy approach, combining extensive sampling, whole-genome sequence-based classification, phenotypic profiling, and assessment of interspecific reproductive isolation. Our work reveals the presence of a distinct Saccharomyces lineage in Nothofagus forests of coastal Patagonia. This lineage, designated Saccharomyces chiloensis sp. nov., exhibits 7% genetic divergence from its sister species S. uvarum, as revealed by whole-genome sequencing and population analyses. The South America-C (SA-C) coastal Patagonia population forms a unique clade closely related to a previously described divergent S. uvarum population from Oceania (AUS, found in Australia and New Zealand). Our species reclassification is supported by a low Ortho Average Nucleotide Identity (OANI) of 93% in SA-C and AUS relative to S. uvarum, which falls below the suggested species delineation threshold of 95%, indicating an independent evolutionary lineage. Hybrid spore viability assessment provided compelling evidence that SA-C and AUS are reproductively isolated from S. uvarum. In addition, we found unique structural variants between S. chiloensis sp. nov. lineages, including large-scale chromosomal translocations and inversions, together with a distinct phenotypic profile, emphasizing their intraspecies genetic distinctiveness. We suggest that S. chiloensis sp. nov diverged from S. uvarum in allopatry due to glaciation, followed by post-glacial dispersal, resulting in distinct lineages on opposite sides of the Pacific Ocean. The discovery of S. chiloensis sp. nov. illustrates the uniqueness of Patagonia’s coastal biodiversity and underscores the importance of adopting an integrative taxonomic approach in species delineation to unveil cryptic microbial species. The holotype of S. chiloensis sp. nov. is CBS 18620T.


Introduction
The concept of a species is a highly debated term in biology, especially when applied to microorganisms [1].Species delineation, the process of defining and distinguishing species within a taxonomic group, faces important challenges in fungi and yeast, mostly because the species' morphological concept does not apply [2].Over time, advances in molecular biology and genomic analysis have significantly driven the identification of different species in eukaryotic microorganisms, regularly based on the phylogenetic and biological species concept [1,3,4].In this sense, yeasts with a known sexual cycle can be assessed regarding reproductive isolation, following the biological species concept [5].At the genetic level, most yeast species can be identified based on sequence divergence in the more rapidly evolving internal transcribed spacer (ITS), as well as the slowly evolving SSU (Small Subunit) and LSU (Large Subunit) rRNA genes.Hence, it is accepted that nucleotide divergence exceeding 1% in the D1/D2 domain of the LSU rRNA gene suggests distinct species [6,7].However, the use of these markers may be influenced by population genetics and demographic processes, especially in temperate and arctic regions.In these regions, glaciation processes in the past have driven some species and sub-populations into allopatric distributions during colder periods and sympatric distributions during warmer periods [8][9][10].Phases of allopatry and post-glacial range expansions can manifest as intricate genetic structures within and between species.For instance, while closely related species may share identical ITS and D1/D2 sequences they can show considerable divergence elsewhere in the genome [11,12].Thus, using a few genetic markers alone can lead to overly conservative species delineation [7].
Recent advances have enabled the use of whole-genome data for species delineation, which is essential for comparing pairwise average nucleotide identity (ANI) between species across the entire genome.ANI quantifies the DNA-level similarity between two strains, correlating with DNA-DNA hybridization.ANI values exceeding 95 ± 0.5% typically suggest conspecificity [13][14][15].Recently, Lachance et al. (2020) suggested in a large data set on haplontic, heterothallic Metschnikowia species that ANI of 95% constitutes a good threshold for species recognition, while values lower than 95% (ANI' > 5%) provided evidence that speciation has taken place [16].Recently, long-read sequencing technologies have significantly improved genome annotations and ANI estimations by only considering symmetrical nucleotide identity comparisons between ortholog pairs, known as average nucleotide identity by orthology (OrthoANI) [17].Therefore, a widely accepted consensus delineation threshold for yeast species is an OANI value of 95% [16].However, delineating yeast species still requires careful consideration of all available data, including genetic variation, phenotypic divergence, ecological niches, and population history.A general current view is to accommodate the so-called 'integrative taxonomy' approach, incorporating genomic, reproductive, and morphological characters for species delineation [18,19].The concept of integrative taxonomy acknowledges the influence of ecological and evolutionary processes on species boundaries, providing a robust approach to defining new species [19].
The Saccharomyces genus is the most prominent and well-studied group across the Saccharomycotina subphylum.This genus includes eight species, many of them with significant economic and scientific importance [5].Recent advances benefiting from bioprospecting efforts and using genomics and phylogenetics tools have revealed two additions to the Saccharomyces clade: Saccharomyces eubayanus and Saccharomyces jurei [20][21][22].S. eubayanus, first identified in the temperate forests of Patagonia, coexists sympatrically with S. uvarum [23][24][25].Despite their taxonomic closeness, in-depth genomic analyses have revealed substantial genetic diversity and evolutionary divergence between these yeast species [23,26].Comparative genomic studies have emphasized differences in chromosomal structure and gene content, highlighting their genetic distinctiveness [26,27].The genetic relationship between S. eubayanus and S. uvarum is characterized by evidence of shared ancestry and common glaciation periods in Patagonia, hinting at a shared evolutionary origin within the Patagonian region approximately 16 MYA [28,29].During these periods, vast ice sheets covered a substantial portion of the land in southern Patagonia, including some regions of the southern Pacific islands.This was succeeded by an extensive floristic recolonization and ecological succession during warmer periods [30,31].These events probably played a vital role in shaping the current genetic structure of Saccharomyces species in Patagonia [23].
S. uvarum populations in the Patagonian Andes contain significant genetic variation, more than anywhere else in the world, and they have diverged into two major sub-groups: South America SA-A and SA-B [28,32].Intriguingly, Holarctic (HOL) strains from Europe and North America clustered alongside the SA-A strain, indicating possible migration patterns in both directions between Europe and America [32].Intriguingly, a vastly divergent lineage of S. uvarum (AUS) has been identified in Tasmania [28] and New Zealand [33], forming a distinct branch within this yeast species.The AUS lineage represents an early diverging cluster and demonstrates partial reproductive isolation from other lineages [28].Although most S. uvarum populations are found in Patagonia, there have been no reports of the presence of the AUS lineage in South America, nor reports of Saccharomyces isolation efforts in Coastal Patagonia or nearby islands with similar geography.The factors contributing to the significant genetic divergence in AUS could be linked to an allopatric speciation process due to geographical separation, underscoring the complexity of the species and the potential for isolated populations to evolve distinct genetic and phenotypic patterns.
In this study, we assessed the diversity of Saccharomyces lineages in Coastal Patagonia and discovered a novel Saccharomyces lineage isolated from Nothofagus dombeyi forests on islands in Southern Chile.Phylogenomic analysis revealed a close relationship with the AUS population previously described in S. uvarum.We applied an integrative taxonomy approach that included phylogenomic assessments, postzygotic reproductive isolation evaluations, and phenotypic characterization.The new lineage displayed considerable nucleotide divergence, reproductive isolation, and a range of structural variants, including notable large translocations found to be polymorphic between Chilean and Australian populations.Our data suggests that the Chilean and Australian populations originated in allopatry.As such, we propose recognizing them as a distinct species named Saccharomyces chiloensis sp.nov.

Yeast isolation
Samples were collected and processed as previously reported [23,34] from bark samples from 106 Coigu ¨e (Nothofagus dombeyi) trees from three different localities in coastal Patagonia: Chiloe ´, Reserva Costera Valdiviana, and San Jose de la Mariquina (Table A in S1 Table ).Saccharomyces identification was performed as previously reported [23].First, the internal transcribed spacer (ITS) was amplified and sequenced using primers ITS1 and ITS4 [35] to discriminate between Saccharomyces and non-Saccharomyces colonies.Subsequently, differentiation between Saccharomyces species was performed using the polymorphic marker RIP1 by amplification, enzyme restriction, and Sanger sequencing as previously reported [23,25].Furthermore, we used 42 strains reported as S. uvarum from our yeast collection [23] to reveal the complete diversity landscape of Saccharomyces isolates from Patagonia (Table B in S1 Table ).Representative sequences for ITS and RIP1 can be found under the code PP618688-PP618690 and PP578963-PP578965.
To obtain individual consensus genomes for each strain, we used the VCF data together with the CBS7001 T genome.First, we used the vcf-subset mode of vcftools 0.1.16to extract individual SNPs from each strain.Then, we indexed each VCF with IndexFeature mode from GATK 4.2.3.0.Finally, using FastaAlternateReferenceMaker from GATK 4.2.3.0 we obtained consensus fasta genome per strain.

Phylogenomic reconstruction
To perform the phylogenomic analysis using our biallelic SNP dataset, we transformed a VCF file containing 1,504,923 SNPs into phylip format using vcf2phylip [44].This was used as input for the IQ-TREE [45] to generate a maximum likelihood phylogeny with the ultrafast bootstrap option and ascertain bias correction (-st DNA -o CL815 -m GTR+ASC -nt 8 -bb 1000) [46].The number of parsimony informative sites was 833,534.Trees were visualized on the iTOL website [47].

Population genetic analyses
First, a VCF file's thinned version was generated with vcftools 0.1.16(-thin 500) [41].Ancestry estimation was explored by running ADMIXTURE [50] using k = 2 to k = 10 five times per k using different seed numbers.Cross-validation errors of each run were used to choose the best number of populations.Results were visualized using Pong [51].Additionally, to perform a clustering analysis using SMARTPCA [52], we kept only those individuals with no evidence of admixture in the ADMIXTURE analysis.Using the same data set of phased individuals with Beagle 3.0.4[53], we performed a fineSTRUCTURE analysis as previously reported [54] considering a constant recombination rate between consecutive SNPs based on the S. cerevisiae average recombination rate (0.4 cM/kbp) [55].Pairwise nucleotide diversity (π) and Fixiation index (F st ) were estimated using the R package PopGenome [56].To predict the number of generations since the most recent common ancestor of any pair of lineages we used the mutation parameter [57], θ w = 2N θ μ, in which N θ represents the number of generations.Considering that the single-base mutation has been reported similar for multiples yeast species [58,59], we used the mutation rate μ = 1.84 x 10 −10 (mutations per generation) previously reported in S. cerevisiae for calculations [60].

Species delineation analysis
To delineate species boundaries based on genetic differences, we used the average nucleotide identity (ANI) as a main criterion [61].All ANI values were calculated against the S. uvarum reference genome, CBS7001 T , by using OrthoANI algorithm [17].As described for yeast, we used the 95-96% ANI value as cut off for species delinitation [15,16].

Spore viability analysis of interpopulation hybrids
To test spore viability between crosses from different lineages, diploid cells were sporulated on 2% potassium acetate agar plates (2% agar) for at least 7 days at 25˚C.Meiotic segregants were obtained by dissecting tetrad ascospores treated with 10 μL Zymolyase 100T (50 mg/mL) on YPD agar plates with a SporePlay micromanipulator.Spores were crossed against the S. uvarum haploid CBS7001 T strain derivate (HOL, HO::NatMx) [62].Hybrid candidates were picked from selective media [62] and confirmed by PCR amplification of the MAT locus (hybrids were Mat-a/Mat-α, HO::NatMx).The hybrids were then sporulated as previously mentioned.At least 96 ascospores per hybrid were dissected in YPD plates and incubated at 25˚C for two days.Spore viability was calculated using the formula (n˚of viable spores/n˚of dissected spores) × 100 [21,63].Additionally, spores from hybrids with low reproductive success were sporulated and dissected for a F2 collection.

Nanopore sequencing, de novo assembly and genome annotation
For nanopore sequencing EXP-NBD104 barcodes and SQK-LSK109 adapters were ligated, and library was loaded on R9.4.1 flowcells; sequencing was performed on a MiniIon device (Nanopore, Oxford, UK).Raw fast5files were converted for fastq files using Guppy, followed by Porechop for the removal of adapters and barcodes [64].Canu v2.2 [65] with default setting was used for assembly, followed by two rounds of both racon v1.4.3 [66] and pilon v1.2.4 [67] for long and short read polishing.Genome assemblies were annotated using the pipeline LRSDAY [68] using the complete genome of the CBS7001T type strain as training input for AUGUSTUS [49].Gene function was predicted using as reference the S. cerevisiae S288c genome, all the other parameters were set as default.Genome completeness was assessed by using BUSCO 5.5.0 [69].Assemblies were compared with CBS7001 T using nucmer [70], and structural variants (SVs) were called using MUM&Co [71].Calls on the ribosomal DNA were not considered.Divergence was re-estimated from whole genome alignment by using Mum-mer4 dnadiff command [70].

Introgression identification
Introgressions from S. uvarum were identified using a combination of SNP density analysis and mapping mean coverage within 1 kb non-overlapping windows [28].To achieve this, we mapped one representative strain from each S. uvarum lineage to a de novo assembly and called variants as previously described [23].Regions exhibiting zero SNP density and more than 20x mapping coverage were selected as potential introgressions.

Phenotypic diversity among isolates
The different isolates were subjected to morphological, physiological, and biochemical characterization under solid media using the replica-plating technique.This physiological and biochemical characterization was done using the standard methods described by Kurtzman et al., with a few modifications [72].The growth of the yeast strains was tested on YNB plates containing 2% of each carbon source.For cell morphology analysis, observations were made using a Zeiss AXIO Imager M2 microscope equipped with differential interference contrast optics (Axiocam 506 mono) after 3 days of growth in YPD broth, incubated at 25˚C.Ascospore production was conducted on 2% potassium acetate agar plates (2% agar) for at least 7 days at 25˚C.Pseudohyphae and true hyphae formation were detected using the Dalmau plate culture method, as described by Kurtzman et al [72].
To determine the phenotypic diversity of the different isolates, we estimate the growth and biomass production in several environmental conditions that are representative of different yeast habitats, either in nature or in industrial settings, as previously described [73].For this, we performed a high-throughput phenotyping assay in 96-well microculture plates.Briefly, cells were pre-cultivated in 200 μL of a YNB medium supplemented with 2% glucose for 48h at 25˚C.Then, strains were inoculated to an optical density (OD) of 0.03-0.1 (wavelength 620 nm) in 200 μL of media and incubated without agitation at 20˚C for 24 h (YNB control) and 48-96 h under other conditions.OD was measured every 30 min using a 620 nm filter.Each experiment was performed in triplicates.The conditions considered were Yeast Nitrogen Base (YNB) supplemented with 2% at 25˚C; 2% glycerol, 2% glucose 2% NaCl; 2% maltose; 2% Fructose; 2% Sucrose; 2% Galactose and 2% glucose + 5% ethanol.Area under the curve (AUC) was estimated with the R package growthcurver [74].

Statistical analysis
All statistical analyses were performed using biological triplicates.The difference was considered statistically significant at p-value < 0.05.The comparison of The Area Under the Curve (AUC) between strains/lineages was performed using a one-way ANOVA.The heat map was created using R package, ComplexHeatmap [75].

A highly divergent Saccharomyces lineage discovered in coastal Patagonia
To identify the Saccharomyces lineages in the Coastal Patagonian region [23,73], we conducted a survey targeting N. dombeyi forests along the South Chilean Pacific coast (Fig 1A).We collected bark samples from different sampling points in the Chiloe island and the Valdivian coastal area, isolating 106 Saccharomyces strains (Table A in S1 Table).Subsequently, we performed genotyping on these isolates by sequencing the highly polymorphic RIP1 marker, enabling the discrimination between S. uvarum and S. eubayanus species.Interestingly, our genotyping assay unveiled three distinct genotypes for the RIP1 marker: two patterns corresponding to S. uvarum and S. eubayanus, while a third, matched to a high divergent S. uvarum lineage (AUS from Almeida 2014), identified in 43 isolates (S1A Fig and  [28,76], along with one representative S. eubayanus isolate (Table B in S1 Table) [23].In total, we analyzed 100 strains: 17 strains isolated from Argentina, 56 from Chile, 5 from Australia, and 19 from the Holarctic regions (Europe, North America, and Asia) (Tables A and B in S2 Table ).The maximum-likelihood phylogenetic tree, employing S. eubayanus as an outgroup, reaffirmed earlier observations: the AUS isolates exhibited the greatest genetic divergence from the SA-A and SA-B South American populations (Fig 1B).Notably, the isolates from coastal Patagonia formed a distinct lineage, closely clustering with AUS, strongly suggesting the presence of a unique independent lineage.
We further utilized the ADMIXTURE software to delve into the population structure of Saccharomyces isolates, using from k 2 to 10 (S2  ).Patterns of pairwise nucleotide variation using Illumina reads showed that AUS and SA-C were, on average 3.69% and 4.06% divergent from the S. uvarum reference genome (CBS 7001 T , HOL), respectively (S5

Table and S5 Fig).
Interestingly, the average pairwise nucleotide divergence between the AUS and SA-C lineages was, on average, 1.03%, suggesting a genetic differentiation between these two lineages (S5 Table ).The high genetic divergence between AUS and SA-C relative to S. uvarum suggests that these groups of strains could be reproductively isolated and represent a distinct evolutionary lineage.

Low spore viability between SA-C and other S. uvarum lineages
To assess the impact of nucleotide sequence divergence on offspring viability, we crossed spores from SA-C and AUS isolates with the S. uvarum CBS7001 T reference strain, estimating spore viability.Crosses involving representative SA-C and AUS strains (CL1604 and ZP966 strains, respectively) exhibited markedly lower spore viability (6.02% and 5.09%, respectively, S6 Table and Fig 2).The resulting spores were sterile, and several did not undergo tetrad formation.On the contrary, crosses between representative strains from different S. uvarum populations with CBS7001 T had higher spore viability, ranging between 50% and 81%, consistent with their lower genetic divergence (S6 Table and Fig 2).This reproductive isolation suggests distinct barriers within the SA-C and AUS lineages compared to other S. uvarum populations.
We crossed representative SA-C and AUS strains (CL1604 and ZP966 strains, respectively) to explore potential lineage relationships between SA-C and AUS, revealing a 25.7% spore viability (S6 Table ).This indicates genetic differences beyond nucleotide divergence, likely involving additional genetic variants, such as structural variants (SVs).Interestingly, diploid SA-C and AUS strains exhibited normal spore viabilities.In addition, we identified a full HO gene version within these strain's genomes, suggesting that they are homothallic and ruling out inherent reproductive issues (S6 Table ).These results suggest reproductive isolation between SA-C, AUS, and other S. uvarum populations, possibly due to nucleotide divergence.Additionally, SA-C and AUS lineages may exhibit additional genetic variants impacting their reproductive status.Considering the nucleotide divergence and reproductive isolation, we propose classifying the SA-C and AUS lineages as a distinct species, named Saccharomyces chiloensis sp.nov.for their most frequent isolation in the Chiloe island.

Coastal Patagonian lineages exhibit large genomic rearrangements
To identify SVs that could shed light on the low spore viability between the SA-C and AUS S. chiloensis sp.nov lineages, we obtained de novo assemblies using ONT long-read sequences from a subset of 5 representative strains from both S. chiloensis sp.nov lineages (S7 Table ).Using the assemblies, we identified SVs with Mum&Co of at least 50 bp relative to the S. uvarum and S. eubayanus reference genomes (CBS7001 T and CBS12357 T , respectively).To ensure accuracy and avoid false positives stemming from assembly errors, we filtered the dataset for SVs size (> 1kb), resulting in the identification, on average, of 72.6 and 85.8 SVs at the lineage level between the SA-C S. chiloensis sp.nov.lineage and S. eubayanus and S. uvarum, respectively (S6 Fig and S8 Table).Interestingly, we identified five large rearrangements across all SA-C S. chiloensis sp.nov.strains: one reciprocal translocation (ChrVI-ChrX t , ChrX-ChrVI t ), two non-reciprocal translocation between three chromosomes (ChrXV-ChrII t and ChrII-ChrVII t ), one inversion (ChrXIII I ) and a deletion on the right arm of ChrVII relative to S. uvarum (Fig 3A).The translocated segments spanned approximately 54 (ChrVI-ChrX t ), 318 (ChrX-ChrVI t ), 433 (ChrXV-ChrII t ) and 143 (ChrII-ChrVII t ) kb, and comprised 19, 145, 212, and 64 genes, respectively.The inversion spanned approximately 487 kb, covering 225 genes (S9 Table ).A similar comparison between the AUS lineage and S. uvarum revealed the presence of the reciprocal translocation previously described between analysis using ADMIXTURE.The optimum of k = 7 groups is shown.Lineages identified are: South America (SA-A (green), SA-B1 (blue), SA-B2 (yellow), SA-B3 (orange) and SA-C (purple)), Australia (AUS, red) and Holarctic (HOL, brown).).
Since many genes are affected by SVs, we decided to explore how gene content varies between Saccharomyces species.We first annotated all the de novo assembled genomes and then constructed a phylogeny considering both S. chiloensis sp.nov lineages and the different Saccharomyces reference genomes (Fig 4A).Interestingly, we reveal the apparent greater genetic distance between SA-C and AUS relative to S. uvarum compared to other lineages within the same species (Fig 4A and S10 Table ).We used the Ortho Average Nucleotide Identity (OANI) tool to explore the mean nucleotide identity between S. chiloensis sp.nov and S. uvarum lineages.The SA-C vs. S. uvarum comparison showed 92.9% identity, while the AUS vs. S. uvarum comparison showed 93.3%.These are lower values than any of the other  ).When we performed a similar analysis using highly divergent lineages in S. paradoxus [77] and S. kudriavzevii (nucleotide divergence ~4.7%), the comparisons denoted values near the 95% species delineation limit (S10 Table ), demonstrating that the SA-C and AUS lineages are more divergent than others in the genus (Fig 4C).
In addition, we assessed the presence of nuclear genome contributions from S. uvarum in S. chiloensis sp.nov.We detected introgression signals ranging from 2-79 kb on chromosomes V (~70 kb), XI (~49 kb) and XVI (~19 kb) (S7 Fig and S11 Table).These findings conclusively establish the existence of a distinct lineage of Saccharomyces strains in coastal Patagonia and Tasmania, exhibiting significant divergence from the established S. uvarum lineages.The notable nucleotide divergence and OrthoANI values strongly suggest potential reproductive isolation, emphasizing the likelihood of these lineages representing a separate species.
Finally, we estimated the divergence time since the most recent common ancestor between lineages using the population genomic data as previously described in yeast [57,79] Our results indicate that S. chiloensis sp.nov.would represent a distinct Saccharomyces species.Additionally, the lack of many SVs between the AUS lineage and S. uvarum strongly supports that nucleotide divergence, rather than SVs, is responsible for the reduced spore viability between S. chiloensis sp.nov.and S. uvarum.In addition, the significant SV count between SA-C and AUS reinforces our observations of the lower offspring viability among these S. chiloensis sp.nov.lineages.Etymology: Saccharomyces chiloensis (chi.lo.en´sis.N.L. f. adj.chiloensis of or pertaining to the Chiloe ´island (Chile), where this yeast was found).

Phenotypic diversity across
On YM agar after 3 days at 25˚C, the cells are globose, ovoid or elongate, 2.8-6.3X 4.1-7.8μm, and occur singly or in small clusters (Fig 5A).In Dalmau plates after 2 weeks on cornmeal agar, pseudohyphae are either not formed or are rudimentary.Budding cells are transformed directly into persistent asci containing one to four globose ascospores formed after 11 days on YM agar at 25˚C (Fig 5B).Fermentation of glucose is positive.Assimilation of carbon compounds: positive for glucose, galactose, maltose (variable, weak/slow), sucrose, melibiose (weak/slow), raffinose, and D-manitol (variable, weak/slow).Negative for L-sorbose, cellobiose, trehalose, lactose, melezitose, inulin, soluble starch, D-xylose, L-arabinose, D-arabinose, D-ribose, L-rhamnose, ethanol, glycerol, erythritol, ribitol, galactitol, D-glucitol, salicin, DLlactate, succinate, citrate, myo-inositol, methanol, hexadecane, xylitol, acetone, ethylacetate, 2-propanol, D-gluconate, and N-acetyl-D-glucosamine. Assimilation of nitrogen compounds: (A) Maximum likelihood (ML) phylogenetic tree using different Saccharomyces species and lineages where long-reads are available.The tree was built considering 5,772 orthologues genes using OrthoFinder.Branch lengths represent the average number of substitutions per site (tree scale).Representative lineages from each species were selected, and annotated genomes were obtained from previous studies using publicly available data.(B) OANI values were estimated using each genome from the ONT assemblies across the different lineages against S. uvarum reference genome (CBST7001 T ).Dashed red and orange lines denote the 95% and 96% species delineation thresholds.(C) Plot depicting OANI values and offspring viabilities comparing different sister Saccharomyces species (blue dots: S. cerevisiae x S. paradoxus [55,78]; S. jurei x S. mikatae [21] and S. eubayanus x S. uvarum), Saccharomyces lineages within the same species (red dots, intraspecies crosses in S. cerevisiae [78], S. kudriavzevii [78], S. paradoxus [78]  negative for lysine, nitrate, and nitrite.Growth in amino-acid-free medium is positive.Growth at 37˚C is negative.Growth in 1% acetic acid is negative.Growth on YM agar with 10% sodium chloride is negative.Growth in 50% glucose/yeast extract (0.5%) is negative.Acid production is positive.Starch-like compounds are not produced.In 100 μg cycloheximide ml -1 growth is negative.Urease activity is negative.Diazonium Blue B reaction is negative (S12 Table ).The CL1206 strain was designated as the type strain.The holotype of S. chiloensis sp.nov., strain CBS 18620 T , is preserved in a metabolically inactive state at the CBS Yeast Collection of the Westerdijk Fungal Diversity Institute in Utrecht, the Netherlands.Additionally, two isotypes of S. chiloensis sp.nov., were deposited as strain RGM 3578, and PYCC 10014 in the Chilean Culture Collection of Microbial Genetic Resources (CChRGM) at the Agricultural Research Institute (INIA) in Chile, and the Portuguese Yeast Culture Collection in Portugal, respectively.
Next, we explored phenotypic differences between S. chiloensis sp.nov.and S. uvarum.Through measurements of microbial growth in microcultures, we calculated the area under the curve (AUC) of growth curves under eight conditions that included different carbon sources, stresses, and environmental parameters (S13 Table ).Hierarchical clustering of the AUC data revealed two clades, clearly distinguishing S. uvarum and S. eubayanus strains from S. chiloensis sp.nov (Fig 5C).Notably, specific traits were identified as characteristic to S. chiloensis sp.nov., establishing differences from its sister species.We observed a significantly lower AUC in S. chiloensis sp.nov.strains grown under maltose and ethanol 5% relative to S. uvarum (p-value < 0,05, ANOVA, S13 Table ), suggesting a lower ethanol tolerance and capacity to assimilate carbon sources other than glucose and fructose.These findings highlight a distinct phenotypic profile in S. chiloensis sp.nov.relative to its sister species, S. uvarum.
Next, we sought to determine whether S. chiloensis sp.nov.exhibited intraspecies phenotypic variation.To investigate this, we compared the AUC among strains obtained from three different locations.Chiloe ´(SA-C-C), Reserva Costera Valdiviana (SA-C-RCV) and Australia (AUS).Interestingly, the AUS strains exhibited the lowest fitness across conditions compared to the other locations (p-value < 0,05, ANOVA, (S13 Table and Fig

Discussion
Species concepts and species delineation methods are essential when new lineages exhibit enough divergence to be declared a distinct species.However, the issue has been controversial due to the diversity of organisms' lifestyles and reproductive modes [80].For the case of microorganisms, the lack of fossil records for time's calibration and their different reproductive modes, faces important challenges for new species identification, being a controversial topic for a long time [19,81].Moreover, species delineation is essential to determining community variability and diversity [82].Defining and classifying a microbial species mostly relies on utilizing a handful of markers, which may pose a significant challenge for correctly assigning phylogeny.For example, in many arbuscular mycorrhizal fungi, ITS markers were not resolutive enough relative to other nuclear markers [16,83,84].When ITS and D1/D2 sequences provide insufficient phylogenetic precision, additional identification methods should be used.obtained from the Area Under the Curve (AUC) data and normalized using Z-score per column (D) The Area Under the Curve (AUC) across all environmental conditions grouped by species, lineages and locality.Different letters reflect statistical differences between strains with a P-value < 0.05, one-way analysis of variance (ANOVA).Glc = glucose, EtOh = Ethanol.https://doi.org/10.1371/journal.pgen.1011396.g005 Generally, there is growing recognition that a single set of criteria does not adequately describe the diversity among fungal lineages [12,16,19,81].An alternative is the utilization of integrative taxonomy, i.e., the combination of various species concepts: genome sequence-based classification, phenotype, and quantifying reproductive barriers [12,81].Our study highlights the challenges associated with traditional species delineation methods in yeasts, mainly when relying on single genetic markers such as the ITS and LSU rRNA genes.Advances in genomics and the application of whole-genome data, specifically OANI have proven essential in overcoming these limitations [12,16,82].The recognition of S. chiloensis sp.nov as a distinct evolutionary lineage, supported by OANI, emphasizes the significance of employing wholegenome-based approaches for accurate species delineation.Furthermore, mating tests, demonstrating reproductive isolation between both S. chiloensis sp.nov populations and S. uvarum, further support the identification of a taxonomically independent lineage.Soon, it would be interesting to assess the genomic compositions of surviving spores from interspecies crosses and determine whether viability was possible due to recombination events and postzygotic barriers between divergent genomes.A similar approach recently deciphered in Cryptococcus the genetic separation between C. amylolentus, C. wingfieldii and C. floricola, exhibiting *6% genetic divergence, significant chromosomal rearrangements and hybrid sterility [85].
As revealed by our study, the evolutionary history of Saccharomyces species in Coastal Patagonia unveils a complex interplay of historical events and genetic differentiation (Fig 6A).The genetic relationship between S. chiloensis sp.nov and S. uvarum, hints at an allopatric speciation process, where a new species is generated due to a physical barrier.Intriguingly, the S. chiloensis sp.nov SA-C and AUS populations are found in Pacific islands located 42˚S and associated with native Nothofagus forests, suggesting a common tree host with Gondwanan origin [86].The restricted geographic distribution of the S. chiloensis sp.nov.populations on each side of the Pacific Ocean supports a secondary dispersal process driven by a vector, most likely bird.However, more evidence is needed to understand yeast dispersal across continents.
In particular, S. chiloensis sp.nov is only found in glacial refugia during Pleistocene times, suggesting that this population did not expand after the glaciation period.Indeed, our analysis suggests that S. chiloensis sp.nov with S. uvarum split into two main clades during the middle Pleistocene (ca.0.4-0.059Ma), likely due to ice barriers during the recurrent glaciation periods in Patagonia.Later, the SA-C and AUS populations split during the Late Pleistocene.Similar findings were found between Lachancea cidri populations, estimating clades divergence between Australian and Patagonian lineages during the Late Pleistocene [10].The potential impact of historical glaciation events on the genetic structure of Saccharomyces species in Patagonia, as inferred from shared ancestry and common glaciation periods, highlights the role of environmental factors and genetic drift in yeast evolutionary processes.The current sympatric existence of S. chiloensis sp.nov with S. eubayanus and S. uvarum in Patagonia suggests unique ecological niches and adaptive strategies related to these past events in the southern South Hemisphere.Future bioprospecting studies may expand our understanding of the genetic and ecological dynamics of S. chiloensis sp.nov, including its population structure and geographical distribution in the South Hemisphere.
The vast distribution of Nothofagus forests in Patagonia may promote a large genetic differentiation between Saccharomyces lineages, representing its main habitat in the South Hemisphere.In this sense, large genomic rearrangements, including chromosomal translocations and inversions, suggest that structural variations have shaped the genetic distinctiveness within the Patagonian S. chiloensis sp.nov population.The Australian population, on the other hand, does not show any major differences in structural variants from other S. uvarum populations.The presence of large structural variants is commonly observed within and between yeast species [78,79,[87][88][89][90][91].For example, S. cerevisiae strains exhibit a median of 240 SVs [91], where massive genomic rearrangements have been reported in the Malaysian S. cerevisiae population, and a rapid accumulation of inversions and translocations in the South American S. paradoxus (formerly known as Saccharomyces cariocanus) that resulted in extensively altered genome configurations [89].These large genomic rearrangements caused the partial reproductive isolation between these two lineages.However, in each case, the nucleotide divergence is below 1%; therefore, these lineages cannot be classified as distinct species.The reproductive isolation observed between S. chiloensis sp.nov and other S. uvarum lineages highlights the importance of considering both nucleotide divergence and structural variations in understanding the evolutionary trajectories of yeast species.Notably, the genetic divergence between S. chiloensis sp.nov and its closest relative, S. uvarum, exceeded that observed between other Saccharomyces populations, such as the average 4.6% sequence divergence between American and European S. paradoxus populations.This divergence and OANI values above the 95% threshold indicate that the two populations in S. paradoxus belong to the same species.On the contrary, S. chiloensis sp.nov exhibited nucleotide divergence greater than 6%, mating efficiency less than 6% and OANI values below the 95% suggested threshold relative to S. uvarum.In addition, we found that S. chiloensis sp.nov strains exhibited a distinct phenotypic divergence, marked mainly by differences in sugar consumption associated with carbon sources, such as maltose.These differences may correlate with the presence of structural variants in subtelomeric regions impacting the distribution of gene families associated with sugar consumption.These results demonstrate phenotypic differences between S. chiloensis sp.nov.and different lineages in S. uvarum.Interestingly, S. chiloensis sp.nov.strains exhibited a limited ability to grow under maltose conditions, suggesting that this species may have potential applications in producing low-alcohol beers.Additionally, investigating the metabolic pathways and regulatory mechanisms underlying this strain's maltose utilization and aroma production could provide insights into utilizing this species for specific brewing applications.Altogether, our integrative taxonomic approach demonstrates that S. chiloensis sp.nov.represents a distinct evolutionary lineage.
Our findings emphasize the need for continued exploration and isolation of yeast strains to unravel the full extent of yeast diversity.As demonstrated in this study, advanced genomic tools and integrative taxonomy are crucial to uncovering hidden lineages and understanding the evolutionary forces shaping yeast populations.Future research should focus on expanding the repertoire of S. chiloensis sp.nov.strains from various geographic regions and habitats to precisely describe the genetic and phenotypic diversity within the species.Furthermore, similar genomic studies in diverse microbial systems can facilitate the identification of novel species and elucidate their roles and potential contributions to natural ecosystems and biotechnological applications.Understanding Saccharomyces species' genetic and ecological dynamics, including the newly discovered S. chiloensis sp.nov, can have broader implications for biotechnological applications.Genetic diversity within Saccharomyces species, exemplified by distinct lineages in Coastal Patagonia, could offer valuable traits for industrial applications and explore its biotechnological potential, including its fermentation potential for low-alcohol beers.In conclusion, the discovery of Saccharomyces chiloensis sp.nov. in Coastal Patagonia expands our understanding of yeast diversity, challenges traditional species delineation methods, and highlights the importance of integrative taxonomy in unraveling the evolutionary dynamics of yeast species.The ecological and biotechnological implications of this discovery open new avenues for research in yeast biology and have broader implications for the fields of microbiology and biotechnology.
Fig).The analysis suggests 7 populations as the best-fitting model (S3 Fig), which revealed distinct lineages predominantly corresponding to the geographical origin of the isolates.This analysis sheds light on a different lineage encompassing the coastal Patagonian isolates (hereafter termed SA-C), closely clustering with the AUS population.Moreover, multiple S. uvarum lineages were identified within South America (SA-B1, SA-B2, and SA-B3) (Fig 1C).This topology was consistently supported by both F st and PCA analyses (Fig 1D and 1E and S3 Table), demonstrating substantial genetic differentiation between the AUS and SA-C lineages compared to the SA-A and SA-B Andean Patagonian lineages.FineSTRUCTURE analysis echoed these findings, highlighting an equivalent number of distinct lineages (S4 Fig).Estimating genetic diversity (π) revealed in the SA-C and AUS populations a mean π = 0.0032 and 0.0042, respectively, surpassing the genetic diversity observed in other S. uvarum populations (Fig 1D and S4 Table

Fig 1 .
Fig 1. Geographic distribution and population structure of Coastal Patagonian isolates.(A) World map illustrating the distribution of sequenced genomes of S. uvarum globally (black circles), with specific locations in Chile in Andean (purple) and Coastal (blue) Patagonia where new strains were isolated.(B) Unrooted maximum likelihood tree inferred from an alignment of 100 strains.Red lines indicate the average nucleotide distance between lineages using Illumina reads.The tree contains five main lineages: South America (SA-A, SA-B and SA-C), Australia (AUS) and Holarctic (HOL).Tree scale is substitutions per site (total SNPs across the whole genome size).(C) Non-scaled Maximum likelihood (ML) phylogenetic tree and population S. eubayanus CL815 was used as outgroup.(D) Population differentiation (F ST ) and nucleotide diversity (π) of the seven lineages.Circle size indicates π value.Dashed circles represent the F ST value between groups.(E) Distribution of genomic variation in 100 strains based on the first two components from a PCA analysis.Color codes match the clusters presented in Panel C. Each dot on the graph represents an individual strain.https://github.com/ropensci/rnaturalearth/blob/master/data/df_layers_physical.rda.https://doi.org/10.1371/journal.pgen.1011396.g001ChrVI-ChrX t (Fig 3B), indicating that this reciprocal translocation would be shared across S. chiloensis sp.nov.However, not all SVs are conserved across S. chiloensis sp.nov lineages (Fig 3C).In this sense, the two non-reciprocal translocations (ChrXV-ChrII t and ChrII-Chr-VII t ) along with the ChrXIII inversion and Chrm VII deletion are exclusively found in the SA-C lineage, likely arisen following the split of the AUS and SA-C lineages along the SA-C branch (Fig 3C and S9 Table

Fig 2 .
Fig 2. Offspring viability against S. uvarum.Plot depicting spore viability between lineages and genetic divergence (measured as nucleotide identity).Each dot compares one representative strain mated against the S. uvarum reference strain CBS7001 T .A linear regression model was applied, with an Adjusted Rsquared of 0.8891 and a p-value of 0.001.95% confidence intervals are in gray.https://doi.org/10.1371/journal.pgen.1011396.g002 . The divergence time between S. uvarum and the SA-C and AUS lineages corresponds to approximately 473-59 KYA (Fig 4D), likely dating back to the Middle Pleistocene.On the other hand, the divergence between AUS and SA-C was dated to approximately 171-21 KYA, likely during the late Pleistocene (Fig 4D).

Fig 3 .
Fig 3. Structural variants identification between Saccharomyces chiloensis sp.nov.and other Saccharomyces species.(A) Genome synteny analysis of the S. uvarum reference strain CBS7001 T , and (A) S. chiloensis sp.nov.CL1206 strain (SA-C) and (B) ZP966 strain (AUS).The dot plot representation depicts the DNA sequence identity between the genomes.Translocation and inversions are highlighted in red circles and grey rectangles indicate the chromosomes involved.Translocations were mapped in all the S. chiloensis sp.nov.strains sequenced in this study.(C) Genome synteny analysis of the S. chiloensis sp.nov.CL1206 strain (SA-C) and ZP966 strain (AUS) strains.https://doi.org/10.1371/journal.pgen.1011396.g003

Fig 4 .
Fig 4. Phylogenetic relationship across different Saccharomyces species.(A)Maximum likelihood (ML) phylogenetic tree using different Saccharomyces species and lineages where long-reads are available.The tree was built considering 5,772 orthologues genes using OrthoFinder.Branch lengths represent the average number of substitutions per site (tree scale).Representative lineages from each species were selected, and annotated genomes were obtained from previous studies using publicly available data.(B) OANI values were estimated using each genome from the ONT assemblies across the different lineages against S. uvarum reference genome (CBST7001 T ).Dashed red and orange lines denote the 95% and 96% species delineation thresholds.(C) Plot depicting OANI values and offspring viabilities comparing different sister Saccharomyces species (blue dots: S. cerevisiae x S. paradoxus[55,78]; S. jurei x S. mikatae[21] and S. eubayanus x S. uvarum), Saccharomyces lineages within the same species (red dots, intraspecies crosses in S. cerevisiae[78], S. kudriavzevii[78], S. paradoxus[78] and S. uvarum), and S. chiloensis sp.nov.x S. uvarum crosses (green dots).NRT = Non-Reciprocal Translocations and RT = Reciprocal Translocation.Spore viabilities from other crosses were obtained from Liti et al, 2006 [78].(D) Estimation of the divergence time since the last common ancestor between lineages.Time between (1) S. eubayanus and S. uvarum/S.chiloensis sp.nov.lineages; (2) S. uvarum and S. chiloensis sp.nov.; (3) the SA-C and AUS S. chiloensis sp.nov.lineages.Timeline is shown in Thousand's years (Ky).

Fig 5 .
Fig 4. Phylogenetic relationship across different Saccharomyces species.(A)Maximum likelihood (ML) phylogenetic tree using different Saccharomyces species and lineages where long-reads are available.The tree was built considering 5,772 orthologues genes using OrthoFinder.Branch lengths represent the average number of substitutions per site (tree scale).Representative lineages from each species were selected, and annotated genomes were obtained from previous studies using publicly available data.(B) OANI values were estimated using each genome from the ONT assemblies across the different lineages against S. uvarum reference genome (CBST7001 T ).Dashed red and orange lines denote the 95% and 96% species delineation thresholds.(C) Plot depicting OANI values and offspring viabilities comparing different sister Saccharomyces species (blue dots: S. cerevisiae x S. paradoxus[55,78]; S. jurei x S. mikatae[21] and S. eubayanus x S. uvarum), Saccharomyces lineages within the same species (red dots, intraspecies crosses in S. cerevisiae[78], S. kudriavzevii[78], S. paradoxus[78] and S. uvarum), and S. chiloensis sp.nov.x S. uvarum crosses (green dots).NRT = Non-Reciprocal Translocations and RT = Reciprocal Translocation.Spore viabilities from other crosses were obtained from Liti et al, 2006 [78].(D) Estimation of the divergence time since the last common ancestor between lineages.Time between (1) S. eubayanus and S. uvarum/S.chiloensis sp.nov.lineages; (2) S. uvarum and S. chiloensis sp.nov.; (3) the SA-C and AUS S. chiloensis sp.nov.lineages.Timeline is shown in Thousand's years (Ky).https://doi.org/10.1371/journal.pgen.1011396.g004 5D), particularly under galactose and higher maltose concentrations.These results indicate a separation of the isolate's phenotypic profile according to their geographic origin, either from Patagonia or Australia, demonstrating substantial intraspecies phenotypic variation (Fig5C).

Fig 6 .
Fig 6.Saccharomyces species.(A) Hypothetical speciation-dispersal processes explaining the sympatric distribution of S. chiloensis sp.nov.and S. uvarum, together with the disjunct distribution of S. chiloensis sp.nov. in Chile and Tasmania.First, an ancestral Sacharomyces population experienced allopatric speciation (1) due to an ice barrier during the Middle Pleistocene glacial maximum (~0.4 mya).Secondly S. chiloensis sp.nov.was dispersed by vectors to Tasmania (2), generating the actual distribution at present.(B) Maximum likelihood (ML) phylogenetic tree of the Saccharomyces genus using long-reads.The tree was built considering 6,615 orthologous genes using OrthoFinder.Branch lengths represent the average number of substitutions per site (tree scale).https://doi.org/10.1371/journal.pgen.1011396.g006

Table A in S1Table ) .
Further Sanger sequencing of these isolates' ITS, LSU, and SSU regions confirmed a 100% sequence identity with S. uvarum, strongly suggesting that these isolates likely belong to this species (S1B and

S1C Fig and Table A in S1 Table).
To unravel the phylogenetic relationships of the coastal Patagonia strains within the S. uvarum phylogeny, we obtained short-read whole-genome sequences from 19 isolates representing various localities in Coastal Patagonia.Additionally, we sequenced 37 S. uvarum isolates previously obtained by our group from diverse locations in Andean Patagonia (Table B in S1 Table and Fig 1A), and we included 43 S. uvarum genomes from previous reports (Table C