Antonie van Leeuwenhoek

, Volume 99, Issue 1, pp 13–23

Phylogeny of the ascomycetous yeasts and the renaming of Pichia anomala to Wickerhamomyces anomalus

Authors

    • Foodborne Bacterial Pathogens and Mycology Research UnitNational Center for Agricultural Utilization Research, Agricultural Research Service, U.S. Department of Agriculture
Review Paper

DOI: 10.1007/s10482-010-9505-6

Cite this article as:
Kurtzman, C.P. Antonie van Leeuwenhoek (2011) 99: 13. doi:10.1007/s10482-010-9505-6

Abstract

In this review, the phylogeny of the ascomycetous yeasts is discussed, with emphasis on the genus Pichia and its synonym Hansenula. The genus Pichia, as defined from phenotype, had nearly 100 assigned species, but the number of species has been reduced to 20 following phylogenetic circumscription on Pichia membranifaciens, the type species of the genus. The remaining species of Pichia have been reassigned to 20 different genera, many of which are newly described, such as Wickerhamomyces. The reason for reclassification of Pichia anomala in the genus Wickerhamomyces is discussed.

Keywords

Pichia anomalaWickerhamomyces anomalusBiocontrol yeastsPhylogeny

Introduction

Gene sequence comparisons have had a profound effect on the taxonomy and systematics of yeasts. Prior to the widespread application of gene sequence comparisons, identification of species and their placement in genera was dependent on phenotype, i.e., cell morphology, ability to ferment sugars and growth on various carbon and nitrogen compounds. It had been believed that characteristics such as ascospore morphology or ability to utilize nitrate as a sole source of nitrogen predicted genetic relationships and could serve as descriptors for defining genera. The use of nuclear DNA reassociation experiments introduced the first quantitative molecular biological technique to yeast classification, and comparisons of DNA relatedness showed that glucose fermentation, nitrate assimilation and ascospore morphology can be variable among strains of a species (Price et al. 1978; Kurtzman 1984). Despite the great impact of DNA reassociation studies on yeast taxonomy, genetic resolution from this method extends no further than to closely related species.

With the introduction of rapid DNA sequencing technologies, opportunity was provided to quickly and accurately identify species and to understand broader species relationships. Initial work focused on sequencing ribosomal RNA and the genes coding for rRNA. Of particular interest were domains 1 and 2 (D1/D2) of the large subunit RNA and its gene. This region is ca. 500–600 nucleotides in length and shows sufficient substitutions to resolve most yeast species. Furthermore, the flanking regions are highly conserved and one set of primers can be used to amplify and to sequence this region for essentially all known ascomycete and basidiomycete yeasts, as well as other fungi (Kurtzman and Robnett 1998; Fell et al. 2000). Certain other genes, such as actin (Daniel and Meyer 2003) and translation elongation factor-1α (Kurtzman and Robnett 2003) also provide good resolution of yeast species, but the priming regions for these genes are often not well conserved. As a consequence of ease of use, the D1/D2 region has been widely adopted, which has given diagnostic sequences, a barcode, for essentially all known yeasts. Because of interspecific hybridization, differences in nucleotide substitution rates among various lineages and other genetic changes, reliance on a single gene for species identification can lead to occasional errors (Peterson and Kurtzman 1991; Groth et al. 1999), which is one reason for the increased application of multi-locus sequence typing (MLST).

Circumscription of yeast genera

The question of what is a genus must be nearly as old as the question of what is a species. The definition offered in the Dictionary of Fungi (Kirk et al. 2008) states that there are “no universally applicable criteria by which genera are distinguished, but in general the emphasis is now on there being several discontinuities in fundamental characters”. The discontinuities that systematists now look for are those in phylogenetic trees. The polyphyly of many yeast genera was revealed from single gene phylogenetic analyses (e.g., Liu and Kurtzman 1991; Cai et al. 1996; Kurtzman and Robnett 1998), but statistical support for the clades as determined from bootstrap analysis was often quite low for the short gene sequences initially compared, and it was uncertain whether the clades really represented genera. Inclusion of several genes in an analysis often strengthens basal branches in phylogenetic trees allowing detection of clades that appear to be genera. By analyzing the genes in various combinations, congruence of the gene trees can also be tested. From multigene analyses, it became clear that the earlier single-gene indications of widespread polyphyly among genera were correct, and that phenotypic characters such as glucose fermentation, nitrate utilization, presence of septate hyphae and ascospore morphology often do not serve as descriptors for genetically defined genera.

In this review, changes in classification of the ascomycetous yeasts resulting from gene sequence analyses will first be discussed followed by a discussion of the reclassification of Pichia anomala. An overview of relationships among ascomycetous yeast genera is presented in Fig. 1 and was determined from phylogenetic analysis of concatenated gene sequences for nuclear large subunit rRNA, nuclear small subunit rRNA and translation elongation factor-1α. Although additional genes will be needed to provide well-supported basal lineages, support for the genera, which are represented by the type species of each genus, is relatively strong. The analysis also shows that the ascomycetous yeasts are members of a single lineage (Saccharomycotina), as indicated from other multigene phylogenetic studies (e.g., Fitzpatrick et al. 2006; James et al. 2006).
https://static-content.springer.com/image/art%3A10.1007%2Fs10482-010-9505-6/MediaObjects/10482_2010_9505_Fig1_HTML.gif
Fig. 1

Phylogenetic relationships among ascomycetous yeast genera, as represented by type species, were determined from neighbor-joining analysis of a concatenated dataset of gene sequences from LSU rRNA, SSU rRNA and translation elongation factor-1α. Published family designations are listed. Where known, coenzyme Q determinations are given in brackets. Bootstrap values were determined from 1000 replicates. Y and YB prefixes are NRRL strain numbers. T = type strain, NT = neotype strain, A = authentic strain. Type species of the genera Ascobotryozyma (anamorph, Botryozyma), Coccidiascus, Endomyces, Helicogonium, Phialoascus, Saitoella, Macrorhabdus, and Schizoblastosporon (teleomorph, Nadsonia), some of which do not grow on common laboratory media, were not included in the analysis (CP Kurtzman, CJ Robnett, unpublished data)

Species assignments in the genus Pichia have been markedly affected by gene sequence analysis, which in retrospect is not too surprising. The diagnosis of the genus Pichia, based on phenotype, included the following: multilateral budding on a narrow base, presence or absence of hyphae and pseudohyphae, ascospores may be hat-shaped, hemispheroidal, or spherical with or without a ledge, sugars may be fermented and nitrate is utilized by some species as a sole source of nitrogen. Using this definition, the last monographic treatment of Pichia included nearly 100 species (Kurtzman 1998), which then represented about 20% of known ascomycetous yeasts.

The polyphyletic nature of Pichia was shown by a number of molecular comparisons, but was most obvious from analysis of a dataset of D1/D2 LSU rRNA gene sequences that included all known ascomycetous yeasts (Kurtzman and Robnett 1998). Species now accepted in Pichia are shown in Fig. 2. Among them are species of Issatchenkia, a genus that had been characterized by roughened, spherical ascospores. Closely related to I. orientalis are Pichia norvegensis, P. cactophila and P. pseudocactophila, which in contrast to I. orientalis, form hat-shaped ascospores. Species of Saturnispora had previously been assigned to Pichia, but were transferred because they have spherical ascospores with an equatorial ring. For this particular clade, the ascospore morphology is common to all species. Other genera shown in Fig. 2 are Kregervanrija, Komagataella and Phaffomyces, and all include former members of Pichia.
https://static-content.springer.com/image/art%3A10.1007%2Fs10482-010-9505-6/MediaObjects/10482_2010_9505_Fig2_HTML.gif
Fig. 2

Phylogenetic relationships among species of the genera Pichia, Saturnispora, Kregervanrija, Komagataella and Phaffomyces determined from maximum parsimony analysis of concatenated gene sequences from LSU rRNA, SSU rRNA and EF-1α. Genus names in the Pichia clade are those used prior to reclassification following multigene analysis. Bootstrap values are from 1000 replicates. Strain accession numbers are NRRL. T = type strain. From Kurtzman et al. (2008)

Species assigned to the genera Barnettozyma, Lindnera (≡Cyberlindnera, see later discussion), Starmera and Wickerhamomyces are given in Fig. 3. Pichia anomala is a member of the Wickerhamomyces clade and most closely related to P. ciferrii, P. lynferdii, P. subpelliculosa, P. sydowiorum and Candida silvicultrix. Because of their phylogenetic placement, Pichia species in this clade were reassigned to the new genus Wickerhamomyces. Noteworthy is that many other species of Pichia are now placed in the genera Barnettozyma, Lindnera and Starmera. Another major change resulting from multigene phylogenetic analysis was recognition of the polyphyly of the Saturn-spored species of Williopsis and their reassignment to four separate genera, Barnettozyma, Lindnera, Ogataea and Wickerhamomyces. Further distribution of Pichia species is illustrated in Fig. 4. Yamada et al. (1994, 1995) detected from rRNA sequence analysis that methanol assimilating yeasts assigned to Pichia were unrelated to P. membranifaciens and comprised two separate clades for which they proposed the genera Ogataea and Komagataella. Multigene analysis supports these proposals as well as the proposed genera Kuraishia and Nakazawaea.
https://static-content.springer.com/image/art%3A10.1007%2Fs10482-010-9505-6/MediaObjects/10482_2010_9505_Fig3_HTML.gif
Fig. 3

Phylogenetic relationships among species of the genera Barnettozyma, Lindnera (≡Cyberlindnera, see text), Starmera and Wickerhamomyces determined from maximum parsimony analysis of concatenated gene sequences from LSU rRNA, SSU rRNA and EF-1α. Species previously classified in the genera Pichia and Williopsis are listed with those genus names. Bootstrap values are from 1000 replicates. Strain accession numbers are NRRL. T = type strain, NT = neotype strain, A = authentic strain. From Kurtzman et al. (2008)

https://static-content.springer.com/image/art%3A10.1007%2Fs10482-010-9505-6/MediaObjects/10482_2010_9505_Fig4_HTML.gif
Fig. 4

Phylogenetic relatedness among species of the methanol assimilating genus Ogataea and neighboring genera determined from maximum parsimony analysis of concatenated gene sequences from LSU rRNA, SSU rRNA, EF-1α and mitochondrial SSU rRNA. Genus names given in the tree are those used prior to reclassification following multigene analysis. Bootstrap values are from 1000 replicates. T = type species, A = authentic strain. Strain accession numbers are NRRL. From Kurtzman and Robnett (2010)

The species just discussed are characterized by formation of the ubiquinone coenzyme Q-7, or much less frequently coenzyme Q-8. However, many Pichia species form coenzyme Q-9 and these species are phylogenetically distant from the preceding taxa (Figs. 1, 5). Based on the analysis presented in Fig. 5, these species were assigned to six new genera. In addition, the coenzyme Q-9 forming species P. ofunaensis and P. tannicola were recognized to be members of the genus Zygoascus (Kurtzman and Robnett 2007).
https://static-content.springer.com/image/art%3A10.1007%2Fs10482-010-9505-6/MediaObjects/10482_2010_9505_Fig5_HTML.gif
Fig. 5

Phylogenetic relationships among species of genera assigned to the Debaryomycetaceae determined from neighbor-joining analysis of concatenated gene sequences from D1/D2 LSU rRNA and SSU rRNA (Kurtzman and Suzuki 2010). Genus names given in the tree are those used prior to reclassification following the preceding analysis. Bootstrap values are from 1000 replicates. Strain accession numbers with Y and YB prefixes are NRRL

Taxonomy and systematics of Wickerhamomycesanomalus

The placement of Pichia anomala in the genus Wickerhamomyces by Kurtzman et al. (2008) following multigene phylogenetic analysis has raised concern whether these results argued for a new genus or whether P. anomala should have been reassigned to an earlier described genus, such as Hansenula (discussions, First International Pichia anomala Mini-Symposium, Uppsala, Sweden, 2010). As discussed in this review, phylogenetic analysis of gene sequences has caused major changes in the classification of yeasts. The goal of this work has been to develop a phylogenetic system of classification for which names of taxa convey natural relationships and are therefore predictive of the genetic properties common to related species. Consequently, with this view, most scientists would not now accept using the old name Saccharomyces anomalus Hansen (1891), the name under which W. anomalus was first described. With inclusion of the binomial Pichia anomala, W. anomalus has 45 synonyms to its credit, which includes six ascosporic genus assignments (Kurtzman 1998). This is, however, far fewer than the number of synonyms attributed to Saccharomyces cerevisiae (Vaughan-Martini and Martini 1998), and for all species, this reflects the impossible task of separating species or assigning them to genera when only phenotypic characters are used.

The following discussion traces the taxonomic history of W. anomalus and the considerations that led to its reassignment in the newly described genus Wickerhamomyces (Kurtzman et al. 2008). This account begins in 1891, when Hansen isolated and described as Saccharomyces anomalus a strain from brewer’s yeast that was fermentative, had ovoid to sausage-shaped cells, formed a pellicle in culture, gave a smell of esters and produced hat-shaped ascospores. In 1904, Hansen transferred S. anomalus, as type species, and Saccharomyces saturnus, a species that forms Saturn-shaped ascospores, to his newly described genus Willia. However, Sydow and Sydow (1919) reported prior usage of the name Willia for a genus of mosses and transferred S. anomalus and S. saturnus to the newly described genus Hansenula. Earlier, von Nägeli (1879) reported a yeast from grape must that formed a pellicle and produced an ester flavor. It is uncertain if the yeast was isolated in pure culture, but this yeast was named Saccharomyces sphaericus by von Nägeli (1879) and has been regarded as conspecific with H. anomala (Lodder and Kreger van Rij 1952, Wickerham 1970). Because S. sphaericus was described 12 years earlier than S. anomalus, it is the name of taxonomic priority for the species. However, there is no type material for either S. sphaericus or S. anomalus and it is impossible to know if they are conspecific or even if they are the same species as the Wickerhamomyces anomalus that we recognize today. The absence of type material renders each of these taxa invalid although, as discussed below, neotypes were proposed for Hansenula anomala. Fischer and Brebeck (1894) described the species Endoblastoderma pulverulentum, an apparent ascosporic species, which was based on a strain isolated by Beijerinck from lager beer produced at a brewery in Rotterdam, The Netherlands. Wickerham (1951) and Lodder and Kreger-van Rij (1952) considered E. pulverulentum as a synonym of H. anomala, and the type culture of this species is CBS 2230 (JCM 2376, NRRL Y-1785, VKM Y-118). Although the genus Endoblastoderma is of uncertain circumscription and has been dismissed in earlier taxonomic monographs (e.g., Lodder and Kreger van Rij 1952), the species name ‘pulverulentum’ is valid and has priority over ‘sphaericus’ and ‘anomalus’ because type material is available.

Hansenula, the oldest genus name among species now assigned to Wickerhamomyces, was based on species with hat-shaped ascospores, as characterized by H. anomala (Hansen 1904, Sydow and Sydow 1919). The second species assigned to Hansenula (via Willia) was H. saturnus, which is characterized by Saturn-shaped ascospores. Later, species with Saturn-shaped ascospores were transferred to the genus Williopsis, which has been shown from gene sequence analysis to be polyphyletic (Kurtzman and Robnett 1998; Kurtzman et al. 2008). The phenotypic distinction between Hansenula and Pichia has been problematic, as discussed by Lodder and Kreger van Rij (1952). Bedford (1942) provided the clearest phenotypic separation of the two genera by specifying that species of Hansenula assimilated nitrate as a sole source of nitrogen whereas species of Pichia could not utilize nitrate. Kurtzman (1984) showed from nuclear DNA reassociation experiments that Hansenula minuta and Pichia lindneri were 75% related and likely to be conspecific. From this, the phenotypic distinction between Hansenula and Pichia was erased and Hansenula species were transferred to Pichia, the genus of taxonomic priority. Furthermore, gene sequence analyses show that species which had been assigned to Hansenula are intermingled with nitrate-negative species such as those assigned to the four genera given in Fig. 3.

In order to conserve the genus Hansenula, owing to absence of a valid type species, Lodder and Kreger-van Rij (1952) selected an isolate referred to as that of H. Schnegg (probably CBS 110) to be the neotype strain of H. anomala. Unknown to Lodder and Kreger-van Rij (1952), Wickerham (1951) had selected NRRL Y-366 for neotype, a strain of H. anomala of uncertain history that earlier had been received from F.W. Fabian. Most likely with the idea of maintaining the name ‘anomala’, neither Wickerham (1951) nor Lodder and Kreger-van Rij (1952) discussed the possibility of conservation of Hansenula through neotypification of the earlier named Saccharomyces sphaericus or use of the validly described species Endoblastoderma pulverulentum, for which type material is available. Because no formal proposal was made to conserve Hansenula on a neotype of either S. sphaericus or S. anomalus or through use of E. pulverulentum, Hansenula is invalid.

From the preceding discussion, two problems have become apparent, conservation of the species name ‘anomalus’ and conservation of Hansenula. Wickerhamomyces (Hansenula, Pichia) anomalus is a widely used name and a proposal to conserve the species name anomala (-us) against ‘sphaericus’ or ‘pulverulentum’ must be made and is likely to be successful. Conservation of the genus Hansenula is a different issue. Pichia and Hansenula were phenotypically indistinguishable until Bedford (1942) separated the two genera on the test for nitrate utilization. Multigene sequence analyses have now shown that Hansenula, as defined by Bedford (1942) and many taxonomists who followed him, was highly polyphyletic. In view of this, why is there a need to conserve Hansenula and what would be the type species? The genus could be conserved with the species ‘anomalus’ when that name is conserved. In some cases, a genus can also be conserved with a new type species because of its importance to science or technology. This was done for the genus Kluyveromyces because of the widespread usage of the name K. lactis, which would have been lost without conservation of the genus Kluyveromyces on a new type species in the K. lactis clade (Kurtzman et al. 2001). However, many would consider the former species Hansenula polymorpha as far more important than Hansenula anomala because ‘H. polymorpha’ is widely used in biotechnology (e.g., Gellissen 2002). Because the two species are in widely separated clades (Fig. 1), there would be no justification for including W. anomalus in Hansenula if it were conserved on H. polymorpha. For the reasons of polyphyly and the need for retypification of a type species, it was clear to Kurtzman et al. (2008) that rather than trying to justify retention of Hansenula from retypification with an uncertain type species, a new approach was needed. This approach was classification of species in a phylogenetically circumscribed genus, which was described as Wickerhamomyces, and which is typified on the validly described species W. canadensis.

As shown in Fig. 3, many of the species formerly classified in Hansenula, Pichia and Williopsis are resolved into four major clades from multigene phylogenetic analysis. Bootstrap support for some of these clades is somewhat weak, but the genetic distances shown argue for considering the clades as four separate genera, i.e., Barnettozyma, Wickerhamomyces, Lindnera and Starmera, rather than as a single genus. Unfortunately, Minter (2009) found an earlier usage for the genus name Lindnera in the validly published plant genus Lindnera Fuss (1866), Tiliaceae, thus rendering Lindnera Kurtzman et al. (2008) a nom. illegit. The new genus name Cyberlindnera Minter (2009) was proposed and the species assigned to Lindnera Kurtzman et al. were transferred to Cyberlindnera as new combinations. Because it is unlikely that Lindnera Kurtzman et al. can be conserved against Lindnera Fuss, species assigned to Lindnera Kurtzman et al. should now be considered as members of Cyberlindnera.

Conclusions

Gene sequence analysis has provided a means for accurate identification of yeast species. An extension of this approach has been to group species on the basis of their phylogenetic relatedness and the resulting clades have been interpreted as genera. The result has been a much improved understanding of relationships among the yeasts and, it presents the possibility of predicting genetic and physiological properties from phylogenetic placement of species. This restructuring of yeast classification, however, has resulted in new genus assignments for many species, including Pichia anomala, the subject of this series of papers.

Copyright information

© U.S. Government 2010