Fig 1.
The evolutionary history, rate, and timeline of Hanseniaspora diversification.
(A) Phylogenomic and relaxed molecular clock analysis of 1,034 single-copy OGs from a near-complete set of Hanseniaspora species revealed two well-supported lineages termed the FEL and SEL, which began diversifying around 87.2 and 53.6 mya after diverging 95.3 mya. (B) Among single-gene phylogenies in which the FEL and SEL were monophyletic (n = 946), the FEL stem branch was consistently and significantly longer (0.62 ± 0.38 base substitutions/site) than the SEL stem branch (0.17 ± 0.11 base substitutions/site) (p < 0.001; paired Wilcoxon rank–sum test). (C) Examination of the difference between FEL and SEL: stem branch lengths per single-gene tree revealed that 932 single-gene phylogenies had a longer FEL stem branch (depicted in orange with values greater than 0), while only 14 single-gene phylogenies had a longer SEL stem branch (depicted in blue with values less than 0). Across all single-gene phylogenies, the average difference in stem branch length between the two lineages was 0.45. figshare: https://doi.org/10.6084/m9.figshare.7670756.v2. AWRI, Australian Wine Research Institute; CBS, Centraalbureau voor Schimmelcultures; DSM2768, Dutch State Mines 2768; Eo., Eocene; FEL, faster-evolving lineage; Mio., Miocene; mya, million years ago; NRRL, Northern Regional Research Laboratory; OG, orthologous gene; Oligo., Oligocene; Paleo., Paleocene; Pleisto., Pleistocene; Plio., Pliocene; Quat., Quaternary; SEL, slower-evolving lineage; UTAD222, University of Trás-os-Montes and Alto Douro 222.
Fig 2.
Gene presence and absence analyses reflect phenotype and reveal disrupted pathways.
(A) Examination of gene presence and absence (see Methods) revealed numerous genes that were lost across Hanseniaspora. Specifically, 1,409 were lost in the FEL, and 771 genes were lost in the SEL. A Euler diagram represents the overlap of these gene sets. Both lineages have lost 748 genes, the FEL has lost an additional 661, and the SEL has lost an additional 23. (B) The IMA gene family (IMA1–5) encoding α-glucosidases, MAL (MALx1–3) loci, and SUC2 are associated with growth on maltose, sucrose, raffinose, and melezitose. The IMA and MAL loci are largely absent among Hanseniaspora with the exception of homologs MALx1, which encode diverse transporters of the major facilitator superfamily whose functions are difficult to predict from sequence; as expected, Hanseniaspora spp. cannot grow on maltose, raffinose, and melezitose, with the sole exception of H. jakobsenii, which has delayed/weak growth on maltose and is the only Hanseniaspora species with MALx3, which encodes a homolog of the MAL-activator protein. (C) The genes involved with galactose degradation are largely absent among Hanseniaspora species, which correlates with their inability to grow on galactose. Genes that are present are depicted in white, and genes that are absent are depicted in black. The ability to grow, the ability to weakly grow/exhibit delayed growth on a given substrate, or the inability to grow is specified using white, gray, and black circles, respectively; dashes indicate no data. (D) Most genes involved in the thiamine biosynthesis pathway are absent among all Hanseniaspora. (E) Many genes involved in the methionine salvage pathway are absent among all Hanseniaspora. Absent genes are depicted in purple. figshare: https://doi.org/10.6084/m9.figshare.7670756.v2. ADI, Acireductone Dioxygenase; ARO, AROmatic amino-acid requiring; AWRI, Australian Wine Research Institute; BAT, Branched-chain Amino-acid Transaminase; CBS, Centraalbureau voor Schimmelcultures; DSM2768, Dutch State Mines 2768; FEL, faster-evolving lineage; GAL, GALactose metabolism; IMA, IsoMAltase; MAL, MALtose fermentation; MDE, Methylthioribulose-1-phosphate DEhydratase; MEU, Multicopy Enhancer of Upstream activation site; MRI, MethylthioRibose-1-phosphate Isomerase; NRRL, Northern Regional Research Laboratory; SAM, S-AdenosylMethionine requiring; SEL, slower-evolving lineage; SUC2, SUCrose; THI, THIamine regulon; UTAD222, University of Trás-os-Montes and Alto Douro 222; UTR, Unidentified Transcript.
Fig 3.
Gene presence and absence in the budding yeast cell cycle.
Examination of cell-cycle genes revealed numerous genes that are absent in Hanseniaspora genomes. The genes not present in Hanseniaspora participate in diverse functions and include key regulators such as WHI5, components of spindle checkpoint processes and segregation such as MAD1 and MAD2, and components of DNA-damage–checkpoint processes such as MEC3, RAD9, and RFX1. Genes absent in both lineages, the FEL, or the SEL are colored purple, orange, or blue, respectively. The “e” in the PHO cascade represents expression of Pho4:Pho2. Dotted lines with arrows indicate indirect links or unknown reactions. Lines with arrows indicate molecular interactions or relations. Circles indicate chemical compounds such as DNA. figshare: https://doi.org/10.6084/m9.figshare.7670756.v2. Ama, Activator of meiotic anaphase-promoting complex; APC (or APC/C), Anaphase-Promoting Complex; Bfa, Byr-four-alike; Bm1, Biomimetic moiety glutathionesulfonic acid; Bub, Budding uninhibited by benzimidazole; Cak1, Cyclin-dependent kinase-activating kinase; cAMP, cyclic AdenosineMonoPhosphate; Cdc, Cell division cycle; Cdh, CDC20 homolog; Cdr, Candida drug resistance; Chk, Checkpoint kinase; Cks, Cdc28 kinase subunit; Clb, Cyclin B; Cln, Cyclin; Cyc, Cytochrome C; Dam, Duo1 and Mps1 interacting; Dbf, Dumbbell former; Ddc, DNA Damage Checkpoint; Doc, Destruction of Cyclin B; Dun, DNA-damage UNinducible; Esp1, Extra spindle pole bodies 1; Far1, Factor ARrest; FEL, faster-evolving lineage; Fob, Fork Blocking less; Fus3, cell fusion 3; Gin4, Growth inhibitory 4; Grr, Glucose repression-resistant; Hsl, Histone synthetic lethal; Irr, Irregular cell behavior; Kcc, K+-Cl− cotransporters; Lte, Low temperature essential; MAD, Mitotic Arrest-Deficient; MAPK, Mitogen-Activated Protein Kinase; Mbp, Mlul-box–binding protein; Mcd, Mitotic chromosome determinant; MCM, Mini-Chromosome Maintenance; MEC3, Mitosis Entry Checkpoint 3; Met30, Methionine requiring 30; Mih1, Mitotic inducer homolog; Mnd, Meiotic nuclear divisions; Mob, Mps one binder; Mps, Monopolar spindle; Mrc, Mediator of the Replication Checkpoint; Net, Nucleolar silencing establishing factor and telophase regulator; ORC, Origin Recognition Complex; Pds, Precocious Dissociation of Sisters; PHO, PHOsphate; Pom, Polarity misplaced; PP2A, Protein Phosphatase 2A; RAD9, RADiation sensitive; RFX1,; SCB, Swi4,6-dependent cell ycle box; Scc, Sister Chromatid Cohesion; SCF, S-phase kinase-associated protein, Cullin, F-box containing complex; SEL, slower-evolving lineage; Sic, Sucrose NonFermenting; Slk, Synthetic lethal karyogamy; Smc, Stability of minichromosomes; Spo, Sporulation; Swe, Saccharomyces Wee1; Swi, Switching deficient; Swm, Spore Wall Maturation; Tah11, Topo-A Hypersensitive; Tem, Termination of M phase; Tup, deoxythymidine monophosphate-uptake; WHI5, WHIskey 5; Ycg, Yeast cap G; Ycs, Yeast condensing subunit; Yhp1, Yeast Homeo-Protein 1; Yox1, Yeast homeobox 1.
Fig 4.
A panoply of genome-maintenance and DNA repair genes are absent among Hanseniaspora, especially in the FEL.
Genes annotated as DNA repair genes according to GO (GO:0006281) and child terms were examined for presence and absence in at least two-thirds of each lineage, respectively (268 total genes). 47 genes are absent among the FEL species, and 14 genes are absent among the SEL. Presence and absence of genes was clustered using hierarchical clustering (cladogram on the left) where each gene’s ontology is provided as well. Genes with multiple gene annotations are denoted as such using the “multiple” term. figshare: https://doi.org/10.6084/m9.figshare.7670756.v2. ABF1, Autonomously replicating sequence-Binding Factor 1; AWRI, Australian Wine Research Institute; CBS, Centraalbureau voor Schimmelcultures; CDC13, Cell Division Cycle 13; CSM2, Chromosome Segregation in Meiosis 2; DEF1, RNA polymerase II Degradation Factor 1; DSM2768, Dutch State Mines 2768; EAF6, Essential something about silencing 2-related acetyltransferase 1-Associated Factor 6; ECO1, Establishment of Cohesion 1; FEL, faster-evolving lineage; FYV6, Function required for Yeast Viability 6; GO, gene ontology; HPR1, HyPerRecombination 1; KRE29, Killer toxin Resistant 29; LIF1, Ligase Interacting Factor 1; LRS4, Loss of RDNA Silencing 4; MAG1, 3-MethylAdenine DNA Glycosylase 1; MCM21, Mini-Chromosome Maintenance 21; MGT1, O-6-MethylGuanine-DNA methylTransferase 1; MMS22, Methyl MethaneSulfonate sensitivity 22; MRC1, Mediator of the Replication Checkpoint 1; NEJ1, Nonhomologous End-Joining defective 1; NRRL, Northern Regional Research Laboratory; NSE1, NonStructural maintenance of chromosomes Element 1; NUP120, NUclear Pore 120; PCD1, Peroxisomal Coenzyme A Diphosphatase 1; PDS1, Precocious Dissociation of Sisters 1; PHR1, PHotoreactivation Repair deficient 1; POL32, POLymerase 32; PSY3, Platinum SensitivitY 3; P/A, presence or absence; RAD9, RADiation sensitive 9; RFA3, Replication Factor A 3; RIF1, Repressor/activator site binding protein-Interacting Factor 1; SAE3, Sporulation in the Absence of sporulation Eleven; SEL, slower-evolving lineage; SEN15, Splicing ENdonuclease 15; SIR4, Silent Information Regulator 4; SLD2, Synthetically Lethal with DNA polymerase B (II)-1 2; SLX4, Synthetical Lethal of unknown (X) function 4; SNF6, Sucrose NonFermenting 6; TAH11, Topo-A Hypersensitive 11; TDP1, Tyrosyl-DNA Phosphodiesterase 1;UTAD222, University of Trás-os-Montes and Alto Douro 222; XRS2, X-Ray Sensitive 2.
Table 1.
Rate of sequence evolution hypotheses and results.
Fig 5.
dN/dS (ω) analyses support a historical burst of accelerated evolution in the FEL.
(A) The null hypothesis (HO) that all branches in the phylogeny have the same ω value. Alternative hypotheses (B–E) evaluate ω along three sets of branches. (Bi) The alternative hypothesis (HFEL–SEL branch) examined ω values along the FEL and SEL stem branches. (Bii) 311 (31.45%) genes supported HO, and 678 (68.55%) genes supported HFEL–SEL branch. (Biii) Among the genes that supported HFEL–SEL branch, we examined the distribution of the difference between ω1 and ω2 as specified in part Bi. Here, a range of ω1–ω2 of −3.5 to 3.5 is shown in the histogram. Additionally, we report the median ω1 and ω2 values, which are 0.57 and 0.29, respectively. (Biv) 384 (38.83%) genes significantly rejected HO and were faster in the FEL than the SEL, while 237 (23.96%) significantly rejected HO and were faster in the SEL than the FEL. (Ci) The alternative hypothesis (HFEL) examined ω values along the FEL stem branch (ω1) and crown branches (ω2). (Cii) 246 (24.87%) genes supported HO, and 743 (75.13%) genes supported HFEL. (Ciii) Among the genes that supported HFEL, we examined the distribution of the difference between ω1 and ω2 as specified in part Ci. The median ω1 and ω2 values were 0.71 and 0.06, respectively. (Civ) 725 (73.31%) genes significantly rejected HO and had higher ω1 values than ω2 values, while 18 (1.82%) genes significantly rejected HO and had higher ω2 than ω1 values. (Di) The alternative hypothesis (HSEL) examined ω values along the SEL stem branch (ω1) and crown branches (ω2). (Dii) 455 (46.29%) genes supported HO, and 528 (53.71%) genes supported HSEL. (Diii) Among the genes that supported HSEL, we examined the distribution of the difference between ω1 and ω2 as specified in part Di. The median ω1 and ω2 values were 0.27 and 0.07, respectively. (Div) 481 (48.93%) genes significantly rejected HO and had higher ω1 than ω2 values, while 47 (4.78%) genes significantly rejected HO and had higher ω2 than ω1 values. (Ei) The alternative hypothesis (HFEL–SEL crown) examined ω values in the FEL crown branches (ω1) and SEL crown branches (ω2). (Eii) 272 (27.50%) genes supported HO, and 717 (72.50%) genes supported HFEL–SEL crown. (Eiii) Among the genes that supported HFEL–SEL crown, we examined the distribution of the difference between ω1 and ω2 as specified in part Di. The median ω1 and ω2 values were 0.06 and 0.07, respectively. (Eiv) 481 (21.54%) genes significantly rejected HO and had higher ω1 than ω2 values, while 504 (50.96%) genes had higher ω2 than ω1 values. figshare: https://doi.org/10.6084/m9.figshare.7670756.v2. dN, rate of nonsynonymous substitutions; dS, rate of synonymous subsitutions; FEL, faster-evolving lineage; SEL, slower-evolving lineage.
Fig 6.
Analyses of base substitutions and indels reveal a higher mutational load in the FEL compared to the SEL.
(A) Analyses of substitution patterns among codon-based alignments of 1,034 OGs revealed a higher number of base substitutions in the FEL compared to the SEL (F(1) = 196.88, p < 0.001; multifactor ANOVA) and an asymmetric distribution of base substitutions at codon sites (F(2) = 1,691.60, p < 0.001; multifactor ANOVA). A Tukey honest significance differences post hoc test revealed a higher proportion of substitutions in the FEL compared to the SEL at the first (n = 240,565; p < 0.001), second (n = 318,987; p < 0.001), and third (n = 58,151; p = 0.02) codon positions. (B) Analyses of the direction of base substitutions (i.e., G|C → A|T or A|T → G|C) revealed significant differences between the FEL and SEL (F(1) = 447.1, p < 0.001; multifactor ANOVA) as well as differences in the directionality of base substitutions (F(1) = 914.5, p < 0.001; multifactor ANOVA). A Tukey honest significance differences post hoc test revealed a significantly higher proportion of substitutions were G|C → A|T compared to A|T → G|C among sites that are G|C (n = 232,546) and A|T (n = 385,157) (p < 0.001), suggesting a general AT bias of base substitutions. Additionally, there was a significantly higher proportion of sites with base substitutions in the FEL compared to the SEL (p < 0.001). Specifically, a higher number of base substitutions was observed in the FEL compared to the SEL for both G|C → A|T (p < 0.001) and A|T → G|C mutations (p < 0.001), but the bias toward AT was greater in the FEL. (C) Examinations of transition/transversion ratios revealed a lower transition/transversion ratio in the FEL compared to the SEL (p < 0.001; Wilcoxon rank–sum test). (D) Comparisons of insertions and deletions revealed a significantly greater number of insertions (p < 0.001; Wilcoxon rank–sum test) and deletions (p < 0.001; Wilcoxon rank–sum test) in the FEL (;
) compared to the SEL (
;
). (E and F) When adding the factor of size per insertion or deletion, significant differences were still observed between the lineages (F(1) = 2,102.87, p < 0.001; multifactor ANOVA). A Tukey honest significance differences post hoc test revealed that most differences were caused by significantly more small insertions and deletions in the FEL compared to the SEL. More specifically, there were significantly more insertions in the FEL compared to the SEL for sizes 3–18 (p < 0.001 for all comparisons between each lineage for each insertion size), and there were significantly more deletions in the FEL compared to the SEL for sizes 3–21 (p < 0.001 for all comparisons between each lineage for each deletion size). Black lines at the top of each bar show the 95% confidence interval for the number of insertions or deletions for a given size. (G) Evolutionarily conserved homopolymers of sequence length 2 (n = 17,391), 3 (n = 1,062), 4 (n = 104), and 5 (n = 5) were examined for substitutions and indels. Statistically significant differences of the proportion mutated bases (i.e., [base substitutions + deleted bases + inserted bases]/total homopolymer bases) were observed between the FEL and SEL (F(1) = 27.68, p < 0.001; multifactor ANOVA). Although the FEL had more mutations than the SEL for all homopolymers, a Tukey honest significance differences post hoc test revealed differences were statistically significant for homopolymers of two (p = 0.02) and three (p = 0.003). Analyses of homopolymers using additional factors of mutation type (i.e., base substitution, insertion, deletion) and homopolymer sequence type (i.e., A|T and C|G homopolymers) can be seen in S10 Fig. (H) G → T or C → A mutations are associated with the common and abundant oxidatively damaged base, 8-oxo-dG. When examining all substituted G positions for each species and their substitution direction, we found significant differences between different substitution directions (F(2) = 5,682, p < 0.001; multifactor ANOVA). More importantly, a Tukey honest significance differences post hoc test revealed an over-representation of G → T or C → A in the FEL compared to the SEL (p < 0.001). (I) Signatures of UV-damage–associated single and double substitutions (i.e., C → T at CC sites and CC → TT) double substitutions are greater in the FEL compared to the SEL (p < 0.001 for both tests; Wilcoxon rank–sum test). figshare: https://doi.org/10.6084/m9.figshare.7670756.v2. FEL, faster-evolving lineage; OG, orthologous gene; Pro., Proportion; SEL, slower-evolving lineage.