R E S E A R C H A R T I C L E

Molecular diversityand intragenomic variability in theyeast genusXanthophyllomyces: the origin ofPha⁄a rhodozyma?Jack W. Fell1, Gloria Scorzetti1, Adele Statzell-Tallman1 & Kyria Boundy-Mills2

1Rosenstiel School of Marine and Atmospheric Science, University of Miami, Key Biscayne, FL, USA; and 2Phaff Yeast Culture Collection, Food Science

and Technology, University of California Davis, Davis, CA, USA

Correspondence: Jack W. Fell, Rosenstiel

School of Marine and Atmospheric Science,

University of Miami, 4600 Rickenbacker

Causeway, Key Biscayne, FL, USA.

Tel.: 11 305 421 4603; fax: 1305 421 4600;

e-mail: jfell@rsmas.miami.edu.

Received 28 September 2006; revised 15 June

2007; accepted 17 June 2007.

First published online 6 September 2007.

DOI:10.1111/j.1567-1364.2007.00297.x

Editor: Teun Boekhout

Keywords

basidiomycetous yeast; Xanthophyllomyces ;

Phaffia ; molecular diversity; rRNA gene

sequence analysis; intragenomic sequence

heterogeneity.

Abstract

The teleomorphic basidiomycetous yeast Xanthophyllomyces dendrorhous is im-

portant as a commercial source of astaxanthin, which is a component of feeds for

mariculture. Phaffia rhodozyma is the anamorphic state of Xanthophyllomyces;

however, there are conflicting reports in the literature concerning the presence of a

sexual cycle in P. rhodozyma. The current study attempted to explain this enigma.

Strains were obtained from the Phaff Yeast Culture Collection (University of

California, Davis) and other sources in the northern hemisphere. Molecular

sequences of three nuclear rDNA regions were examined: the internal transcribed

spacers (ITS), intergenic spacer (IGS1) and the D1D2 region at the 50 end of the

26S gene. Different levels of genetic variability were observed in the three regions.

The D1D2 differentiated major groups of strains, while an increased variability in

the ITS suggested that the ITS region could be employed as an ecological marker.

The greatest variability was in the IGS1 region, where strains can be defined by the

presence and location of indels. Intragenomic sequence heterogeneity in the ITS

and IGS1 regions led to the hypothesis that the type strain of P. rhodozyma (CBS

5905T, UCD 67-210T) was derived as a mating-deficient basidiospore from the

parent teleomorphic strain CBS 9090.

Introduction

Phaffia rhodozyma is an economically important basidiomy-

cetous yeast, which is used in commercial feed to produce

pink-colored flesh in salmon and shrimp (Johnson, 2003).

The species was first isolated by Phaff et al. (1972) from

Japan and Alaska and described by Miller et al. (1976) as an

anamorphic species with type strain UCD 67-210T, which

was submitted to the Centraalbureau voor Schimmelcul-

tures as CBS 5905T. Golubev (1995) discovered the tele-

omorphic state based on parent–bud mating in strains from

Japan, Finland and Russia. The genus and species were

described as Xanthophyllomyces dendrorhous with the type

strain designated as CBS 7918T.

Golubev (1995) reported that a subculture of the type

strain of P. rhodozyma (VKM Y-2274T) produced a Xantho-

phyllomyces-type sexual cycle. In contrast, Kucsera et al.

(1998) and research in our laboratory failed to observe a

sexual cycle with strain CBS 5905T. In addition, we obtained

strains, from several laboratories that were labeled UCD 67-

210T or reported to be derived from UCD 67-210T. We

found that some of these strains produced sexual cycles,

whereas other strains lacked that ability. These results

suggested that some strains might be erroneously labeled.

To resolve the confusion, we examined a series of lyophilized

stocks from the years 1968, 1985 and 1986, which are

maintained in the Phaff Yeast Culture Collection at the

University of California at Davis (UCD). The series revealed

two strains that differed on their rRNA gene sequences. One

strain, represented by the 1968 lyophilized stock, was

identical to CBS 5905T. The other strain, represented by the

1985 and 1986 stocks, was unlike previously studied strains

(Fell & Blatt, 1999) and was subsequently labeled UCD 67-

210.2 (CBS 9090, Scorzetti et al., 2002). For comparative

purposes, we included all of the Phaffia/Xanthophyllomyces

strains from the Phaff Collection as well as unstudied strains

in the USDA Collection (NRRL, Peoria, IL; C.P. Kurtzman),

which had been isolated from trees in Illinois and Wiscon-

sin. The studies included observations of sexual cycles,

sequence analyses of nuclear rDNA regions (internal

FEMS Yeast Res 7 (2007) 1399–1408 c� 2007 Federation of European Microbiological SocietiesPublished by Blackwell Publishing Ltd. All rights reserved

transcribed spacers: ITS1 and ITS2, D1D2 region of the large

subunit and the intergenic spacer region 1: IGS1) and

cloning and sequence analysis of ITS and IGS1 regions to

establish the presence of intragenomic sequence heteroge-

neity.

Materials and methods

The strains studied, and the GenBank accession numbers of

their ribosomal sequences, are listed in Table 1. The three

regions examined in the nuclear rDNA were the D1D2

domain at the 50 end of the large (26S) subunit (LSU), the

ITS region including ITS1, 5.8S and ITS2 and the IGS1

region, which extends from the LSU to the 5S. Amplification

of the D1D2 and ITS regions was performed with primers

NS7, 50-GAGGCAATAACAGGTCTGTGATGC-30, and LR6,

50-CGCCAGTTCTGCTTACC-30. IGS primers were LR11,

50-TTACCACAGGGATAACTGGC-30, and 5SR, 50-GATC

GGACGGGCAGGGTGC-30. Eppendorf HotMaster Mix

(Eppendorf North America) and DyNAzyme Polymerase

(Finnzymes) were alternatively employed. The optimized

amplification program consisted of one denaturation cycle

at 94 1C for 5 min followed by 35 cycles of 1 min denatura-

tion at 94 1C, 1 min annealing at 55 1C and 3 min extension

at 72 1C. The final extension was at 72 1C for 7 min. Samples

of the resulting amplicons were analysed by gel electrophor-

esis on a 2% agarose minigel. The PCR products were

purified with the QIAquick PCR Purification Kit (Qiagene

Inc.)

The D1D2 domain was sequenced with primers F63,

50-GCATATCAATAAGCGGGAGGAAAAG-30, and (R635)

LR3, 50-GGTCCGTGTTTCAAGACGG-30. The ITS region

was sequenced with primers ITS1, 50-TCCGTAGGTGAACC

TGCGG-30, and ITS4, 50-TCCTCCGCTTATTGATATGC-30.

The IGS region employed primers LR12, 50-CTGAACGCC

TCTAAGTCAGAA-30, and 5SR, 50-GATCGGACGGGC

AGGGTGC-30. A LiCor (Lincoln, NB) automated sequencer

with IRD800 conjugate primers was employed for

Table 1. Origin of strains studied

Strain Other numbers Origin

GenBank accession numbers

D1D2 ITS IGS 1

Xanthophyllomyces dendrorhous

ATCC 24228 UCD 68-653.3 Betula papyrifera Alaska DQ870193 AF139631 AF139636

ATCC 24230 UCD 67-385 Betula tauschii Japan AF139630 AF139635

CBS 6938 VKM Y-2793 Betula sp. Finland AF444793 AF139632 AF13937

CBS 9090 UCD 67-210.2 AF444721 AF444488 DQ870210

CBS 9090 C3 DQ904247

CBS 9090 C4 DQ904243

CBS 9090 C6 DQ904244

CBS 9090 C7 DQ870211

CBS 9090 C10 DQ904245

CBS 7918T T VKM Y-2786 Betula verrucosa Russia AF075496 AF139628

NRRL Y-17430 Betula populifolia Wisconsin DQ870185 DQ870197

NRRL Y-17430 C4

NRRL Y-17431 B. populifolia Wisconsin DQ870186 DQ870198

NRRL Y-17432 B. populifolia Wisconsin DQ870187 DQ870199 DQ870215

NRRL Y-17433 B. populifolia Illinois DQ870188 DQ870200 DQ870216

NRRL Y-17434 B. populifolia Illinois DQ870189 DQ870201 DQ870217

NRRL Y-17434 C1 DQ870218

NRRL Y-17438 B. populifolia Illinois DQ870190 DQ870202 DQ870219

NRRL Y-17438 C2 DQ870220

NRRL Y-17441 B. populifolia Illinois DQ870191 DQ870203 DQ870221

NRRL Y-17441 C9 DQ870222

NRRL Y-17442 B. populifolia Illinois DQ870192 DQ870204 DQ870223

NRRL Y-17443 B. populifolia Illinois DQ913745 DQ913746 DQ913747

NRRL Y-27348 B. populifolia Illinois DQ870205 DQ870224

NRRL Y-27348 C2 DQ870225

UCD 67-202 ATCC 24229 Cornus brachypoda Japan DQ870194 DQ870206

UCD 67-203 ATCC 24201 C. brachypoda Japan DQ870195 DQ870207

Phaffia rhodozyma

CBS 5905T T UCD 67-210T Fagus crenata Japan AF189871 AF139629 AF139634

CBS 5905T C2 DQ904246

CBS 5905T C10 DQ870209

FEMS Yeast Res 7 (2007) 1399–1408c� 2007 Federation of European Microbiological SocietiesPublished by Blackwell Publishing Ltd. All rights reserved

1400 J.W. Fell et al.

sequencing. Sequences were analysed with Align IR for Mac

OS 9. The sequences with double pattern images were

resequenced on an ABI 3730 DNA Analyzer (Applied

Biosystems) and analysed with Seqman 5.51 for Mac OS X.

Following amplification and purification, amplicons re-

presenting the IGS1 and ITS regions from selected strains

(Table 1) were ligated and transformed in competent cells

(TA Cloning Kit with INVaF chemically competent Escher-

ichia coli, Invitrogen Corp.) using the manufacturer’s proto-

cols. Sixteen clones from each sample were purified (Wizard

SV 96 Plasmid DNA Purification System, Promega Corp.)

and sequenced with an ABI 3730 DNA Analyzer (Applied

Biosystems) using primers M13F (M13/pUC universal

primer code 1, Sambrook & Russell, 2001) (50-GTAAAA

CGACGGCCAGT-30) and M13R (50–GGAAACAGCTATG

ACCATG–30), which is a 3-nt 50 extension of M13/pUC

universal primer code 4 (Sambrook & Russell, 2001).

Sequences were analysed with Seqman 5.51 for Mac OS X.

Phylogenetic analysis of the D1D2 region employed max-

imum-likelihood (ML) with the HKY85 substitution model

(Hasegawa et al., 1985) and a heuristic search. The ITS and

IGS analyses compared ML with maximum-parsimony

(MP: gaps treated as missing data) and neighbor-joining

(NJ) analyses. Heuristic search consisted of random step-

wise addition and tree-bisection-reconnection branch swap-

ping (TBR). Bootstrap analyses consisted of 1000 replicates

in a full heuristic search. All analyses employed PAUP 4.0b10.

The MP, ML and NJ analyses agreed in terms of strain

relationships, and therefore the trees illustrated are those

that were considered to be the most appropriate for data

presentation and discussion. The IGS1 region was cloned for

strains CBS 5905T, 7918T, 9090, NRRL Y-17430, Y-17431, Y-

17432, Y17433, Y-17434, Y-17438, Y-17441 and Y-27348.

Strains cloned for the ITS region were CBS 5905T, 7918T,

9090, Y-17433 and Y-27348. The selection included strains

with and without apparent multiple copies.

Basidial formation was studied with CBS 9090, CBS

5905T and all of the NRRL strains. The strains were grown

for 1 month on 0.5% glucitol yeast nitrogen base (YNB)

agar, then transferred to 0.5% glucitol YNB agar at 18 1C

for 5 days.

Results

D1D2 sequence analysis

Xanthophyllomyces and Phaffia are members of the Cystofi-

lobasidiales (Fig. 1, Fell et al., 1999; Scorzetti et al., 2002),

which includes the teleomorphic genera Mrakia and Cysto-

filobasidium and anamorphic species of the genera Udenio-

myces, Cryptococcus, Itersonilia and Guehomyces. There are

three groups of strains within the Xanthophyllomyces cluster:

(1) all of the X. dendrorhous strains listed in Table 1 have an

identical D1D2 sequence with the exception of CBS 9090,

UCD 67-202 and UCD 67-203; (2) CBS 9090 and P.

rhodozyma (CBS 5905T) have identical sequences, which

differ from CBS 7918T (X. dendrorhous) by the absence of a

T at position 558 (Table 2) and a C instead of a T at position

562; (3) two UCD strains (67-202 and 67-203), as repre-

sented by 67-202, differ from CBS 5905T and CBS 7918T by

the presence of a C instead of a T at position 569 (Table 2).

ITS sequence analysis

Analysis of the ITS region (Fig. 2), which includes c. 650 bp,

suggests that alignment differences can be related to the

geographic origins of the strains. The two clusters of strains

collected from birch trees (Betula spp.) appear to be closely

related as they differ by a single nucleotide. Greater distances

are demonstrated by the type strain of P. rhodozyma (CBS

5905T), which was obtained from a beech tree (Fagus) and

the UCD strains from dogwood (Cornus brachypoda). The

distinction between CBS 5905T and CBS 9090 is of particular

interest in this study. As illustrated in Table 3, the majority

of the differences are in the ITS2 region.

IGS1 sequence analysis

There were two major difficulties associated with IGS1

analyses. (1) Some strains demonstrated intragenomic het-

erogeneity between copies of the rDNA operons. This

heterogeneity was indicated by the presence of multiple

patterns in the sequencing images. In some cases, the

secondary patterns were weak and the primary pattern

could be read. In other cases, the strength of the secondary

sequence precluded sequence interpretation. The double

pattern, which initiated in the region of position 390 (Table

4), varied in length from 30–40 bp to nearly half of the

sequence. This double pattern corresponded to the most

variable portion of the sequence. As a consequence, not all

strains were included in the IGS1 analysis (Fig. 3, Table 4).

(2) Multiple repeats and indels were present in the se-

quences, which required visual and manual alignments.

Individual strains may have indels at different positions in

the sequence, and therefore alignments varied with the

strains included in the analysis (Table 4).

The IGS1 data consisted of 687 bp. The initial 392 bases

consist of a relatively conserved region, which we examined

with NJ, MP and ML analyses. The first 60 bases from the

primer at the 50 end were trimmed due to uneven data

acquisition between strains. The results, as illustrated by an

MP tree (Fig. 3), are similar to the ITS data with the

exception that ATCC24230 and ATCC24228 are on a sepa-

rate branch and CBS 5905T and CBS 9090 have identical

sequences. The greatest sequence diversity is in the 390–687

region due to the presence of multiple indels (Table 4). The

number and location of these indels provide markers for

FEMS Yeast Res 7 (2007) 1399–1408 c� 2007 Federation of European Microbiological SocietiesPublished by Blackwell Publishing Ltd. All rights reserved

1401Origin of Phaffia rhodozyma

strain identification. Of particular importance, CBS 9090

and the P. rhodozyma CBS 5905T had identical sequence

alignments with the sole difference based on the presence of

an insertion or deletion at position 482–523. Indels, which

were present in all of the strains, were prevalent in a GT-rich

area from 392 to 560 and in a TAC-rich region from 560 to

687. As examples of strain specificity, CBS 7918T is distin-

guished by insertions at 500–520 and 560–575. Similarly,

Fig. 1. Phylogenetic tree representing nuclear

rDNA D1D2 large subunit sequence analysis of

Cystofilobasidiales prepared by maximum like-

lihood with a heuristic search (PAUP 4.0b10):

4045 rearrangements, tree score 2790. Num-

bers on the branches represent bootstrap

percentages (4 50%). Cryptococcus

fuscescens, Cryptococcus aerius and

Cryptococcus terricola formed the outgroup.

Table 2. D1D2 Sequence differences between strains of Xanthophyllomyces and Phaffia

Nucleotide position

550 560 570X. dendrorhous C A G C G C G C C T C T T T A C G G G G T C UCD 67-202 C A G C G C G C C - C T T C A C G G G G C C CBS 9090 C A G C G C G C C - C T T C A C G G G G T C P. rhodozyma C A G C G C G C C - C T T C A C G G G G T C

FEMS Yeast Res 7 (2007) 1399–1408c� 2007 Federation of European Microbiological SocietiesPublished by Blackwell Publishing Ltd. All rights reserved

1402 J.W. Fell et al.

NRRL Y17434 and Y 17442 can be separated by a 24-bp

indel in the 603–624 TAC region (Table 4).

Cloning experiments

The ITS and IGS regions were cloned to ascertain the

presence of intragenomic sequence heterogeneity, which

was indicated by the double banding pattern in the sequence

images. Strains, with and without a double banding pattern,

were sequenced. Throughout this presentation of results and

discussion, the dominant sequence (as reported to GenBank

and used for tree construction) is signified by the strain

number, i.e. CBS 9090. Clone numbers, such as CBS 9090

C4, designate the sequences obtained from clones.

The ITS cloning experiments did not reveal the presence

of heterogeneity in CBS 7918T, NRRL Y-17433 and Y-27348.

In contrast, clones from CBS 9090 (Fig. 2) had sequences

represented by (1) CBS 9090, (2) CBS 9090 C4, with 2-bp

Fig. 2. Cladogram representing the nuclear

rDNA ITS region of Xanthophyllomyces strains

and clones of CBS9090 (clones C4 and C6) as

prepared by maximum-parsimony analysis with

a heuristic search (PAUP 4.0b10). Midpoint root-

ing (MINF optimization): 2794 rearrangements.

Numbers on branches represent branch lengths.

Numbers in parentheses represent bootstrap

percentages (4 50%).

Table 3. Partial ITS sequence alignment of CBS 9090 and CBS 5905T: positions 481–660. Lower-

case letters indicate nucleotide sequence differences

045184CBS9090 GAAGCGCGGGCGGTGCCTTGACATGATAAGAAATTGTCGTCGAGtgtcGCTGTCTGTGTGCBS5905 GAAGCGCGGGCGGTGCCTTGACATGATAAGAAATTGTCGTCGAGC--TGCTGTCTGTGTG

006145CBS9090 aGTGTGTGggTTTCCTCGGGAAcaCgcaGACTAGCtaGTaCaACTAAaCGACGCATGCGACBS5905 TGTGTGTGCTTTTCCTCGGGAAGGCATGGACTAGCCGGTGCGACTAA-CGACGCATGCGA

066106CBS9090 ACTGCTTCTAACGATACTGTGTGtGtgGtC-TTtGCGGaCtgCaTGCGCACACTTCTGATCBS5905 ACTGCTTCTAACGATACTGTGTGGGGAGCCCTTCGCGGGCCCCTTGCGCACACTTCTGAT

FEMS Yeast Res 7 (2007) 1399–1408 c� 2007 Federation of European Microbiological SocietiesPublished by Blackwell Publishing Ltd. All rights reserved

1403Origin of Phaffia rhodozyma

differences from ATCC 24230 and related species and (3)

CBS 9090 C6 (and 9090 C10, Table 1), which was identical to

CBS 5905T. The clones from CBS 5905T were all identical

with one exception. CBS 5905T C2 had the appearance of the

5905x9090 hybrid; however, we conservatively treated the

sequence as a chimera. For reference, the sequence was

submitted to GenBank (Table 1).

Sequencing of the IGS1 clones demonstrated a lack of

intragenomic variability in the IGS1 regions of strains CBS

7918T, NRRL Y17431, Y17432 and Y17433. In contrast, CBS

5905T, Y17430, Y17434, Y17438, Y17441 and Y27348 con-

tained at least two different IGS1 sequences in their

genomes. Analysis of the 60–392-bp 50 region of the IGS1

(Fig. 3) demonstrated the presence of a limited number of

scattered single nucleotide substitutions. For example (Fig.

3), CBS 5905T C10 and CBS 9090 C7 had two single

nucleotide sequence differences. The other clones were not

included in Fig. 3 due to the apparent lack of significant

substitution differences.

The major differences between clones of a single strain

were in the 392–687 region, which contains indels that

represent the gain or loss of repeat units. Table 5 lists the

position of the indels, which can be located in Table 4. For

example, Y-17430 C4, as compared with the dominant

sequence, had an insert (398–416) in the GT-rich region

and a deletion (650–670) in the TAC region of multiple

repeats (Tables 4 and 5). Y-17434 C1 had two deletions: one

at 432–446, the other at 605–629. Although the clone study

was limited in numbers of strains analysed, none of the

clones from individual strains was identical to sequences

obtained from other strains. Exceptions were clones ob-

tained from CBS 9090 and CBS 5905T (Table 5, Fig. 3). CBS

5905T C10 had an insert at 482–523; as a result, CBS 5905T

C10 was identical to CBS 9090 (with the exception of the

single nucleotide substitution). Clones from CBS 9090 had

three distinct sequences: (1) the dominant CBS 9090 se-

quence; (2) CBS 9090 C7, which had a deletion at 482–523,

was identical to CBS 5905T; and (3) CBS 9090 C3, which is

Table 4. Location of indels in the 361-687-bp region of the rDNA IGS1 for strains of Xanthophyllomyces dendrorhous and Phaffia rhodozyma

FEMS Yeast Res 7 (2007) 1399–1408c� 2007 Federation of European Microbiological SocietiesPublished by Blackwell Publishing Ltd. All rights reserved

1404 J.W. Fell et al.

not included in Table 5, does not represent a deletion/

insertion event. CBS 9090 C3 (Fig. 3) has the CBS 7918T

type of sequence rather than the 5905T/9090 sequence. The

two types of sequences are distinctly different (Table 4). CBS

9090 C3 aligns with ATCC 24230 with the exception of

differences at four separate base positions (two of the

differences are in the 361–687 region, which is not included

in Fig. 3).

Teleomorphic state

The NRRL strains (Table 1) and CBS 9090 formed basidia as

illustrated by Golubev (1998). Previous studies demon-

strated that CBS 6938, ATCC 24230 and UCD 67-202

(ATCC 24229) (Kucsera et al., 1998) also produced sexual

states. CBS 5905T did not produce a sexual stage, either in

our study or the Kucsera et al. (1998) study. Based on

current information, the only known anamorphic strain is

CBS 5905T.

Discussion

The present study was initiated by the conflicting reports

concerning the ability of UCD 67-210T to produce a sexual

cycle. We initially addressed the possibility that a transfer

strain had been mislabeled during the routine maintenance

of the culture collection at UCD. This seemed unlikely

because multiple studies over the years in the UCD labora-

tory had not indicated the presence of additional

Fig. 3. Sequence analysis of nuclear rDNA IGS1

bases 60-392. The cladogram represents rela-

tionships between strains in the 50 region pre-

ceding the indels (see Table 3). Clones from CBS

9090 and CBS 5905T are included in the analysis.

The tree was prepared by maximum-parsimony

analysis (heuristic search) as an unrooted tree

(PAUP 4.0b10). One of two equally parsimonious

trees. Xanthophyllomyces dendrorhousT, type

strain; Phaffia rhodozymaT, type strain. Bases

1–60 were trimmed due to the lack of complete

sequence data for some strains. Numbers on

branches represent branch lengths. Numbers in

parentheses represent bootstrap percentages

(450%).

FEMS Yeast Res 7 (2007) 1399–1408 c� 2007 Federation of European Microbiological SocietiesPublished by Blackwell Publishing Ltd. All rights reserved

1405Origin of Phaffia rhodozyma

anamorphic strains in the Phaff Collection. In addition, a

lack of published reports on the isolation of other ana-

morphic strains suggests that the type strain (67-210T) of P.

rhodozyma represents a rare occurrence in nature. To clarify

this point, we sequenced D1D2 and ITS regions of all of the

Phaffia/Xanthophyllomyces strains (16) in the Phaff Collec-

tion (data not shown) and did not find another strain that

resembled 67-210T.

We also investigated the history of the maintenance

transfers of UCD 67-210T. The Phaff Collection has main-

tained a series of lyophilized stocks of UCD 67-210T since

1967. We sequenced subcultures from several of these

transfers and found that the April 1968 stock strain had a

nucleotide sequence identical to that of CBS 5905T. In

addition, the April 1968 stock did not produce a sexual

cycle. Of note, a subculture of 67-210T was accessioned in

May 1968 at CBS as CBS 5905T, the designated type strain of

P. rhodozyma. The stocks dated June 1985, July 1985, March

1986 and April 1986 produced a typical Xanthophyllomyces

sexual structure and they had identical IGS and ITS

sequences, which were distinct from CBS 5905T. These

subcultures were identical to strains that had been sent to

us from other laboratories, which we had presumed to be

strains mislabeled as 67-210T. To distinguish the different

strains, the June 1985 transfer stock was designated 67-210.2

and submitted to CBS, where it was accessioned as CBS

9090.

The presence of a teleomorphic strain in the transfer

stocks of 67-210T is possibly the source of the reported

sexual cycle in P. rhodozyma. In particular, Golubev (1995)

reported a sexual cycle in a subculture of the type strain,

which he accessioned as VKM Y-2274T. Golubev (personal

communication) obtained the culture from H.J. Phaff

(UCD) in 1975. Golubev’s report of a teleomorphic state

suggests that the VKM strain originated from the same line

as CBS 9090 (UCD 67-210.2) and the apparent strain shift

took place prior to 1975.

This information did not explain the origin of the

teleomorphic strain. As a possible clue to solve this dilem-

ma, we turned our attention to our observations of sequence

heterogeneity. We found that CBS 9090 and CBS 5905T are

heterologous in the ITS and IGS regions. CBS 9090 carries

the CBS 5905T ITS and IGS sequences and CBS 5905T carries

the CBS 9090 IGS sequence. Other strains of Xanthophyllo-

myces (Table 5) have multiple copies with indel differences,

yet none of these strains had copies that are identical to

other genotypes. Hypothetically, the mating-deficient strain

CBS 5905T (Kucsera et al., 1998) could have originated as a

basidiospore from the parent CBS 9090 strain in the UCD

67-210T culture tube. Conceptually, the detection of a

mating-deficient strain, followed by a teleomorph, does not

correlate with this hypothesis. The anticipation would be to

find the teleomorph, followed by the anamorph. However,

culture purification methods are generally employed in the

course of transfer and submission of subcultures to other

collections. This method entails a streak culture and transfer

of a single colony. By the repeated use of this technique over

time, the anamorphic strain could have been sent to CBS

and subsequently lost in the Phaff Collection to the more

prevalent diploid form.

In conclusion, we hypothesize that the origin of CBS 9090

in the culture collection tube labeled UCD 67-210T was

not the result of a mislabeled strain; rather, strain UCD

67-210T may have been derived from a basidiospore from

the parent teleomorph represented by strain CBS 9090. This

hypothesis could be examined by micromanipulation and

analysis of multiple basidiospores from CBS 9090. In the

meantime, researchers should be aware of the heterologous

nature of these strains and that some strains, marked as

UCD 67-210T, represent CBS 9090. The two strains can be

differentiated by ITS sequences (Table 3), either by sequence

analysis or through the use of strain-specific PCR primers or

probes.

In addition to our analysis of the source of these strains,

we gained insight on the use of sequence analysis to

distinguish specific strains (Fell & Blatt, 1999). The genus

Xanthophyllomyces appears to be associated with exudates

from trees, such as birch, beech and dogwood, in temperate

regions throughout the world. The strains isolated from the

different geographic regions demonstrate considerable gen-

otypic variability in the D1D2, ITS and IGS regions.

Although our present comparison is geographically limited,

a greater diversity can be anticipated as the search for new

habitats is expanded. The D1D2 alignments distinguish

three groups (Fig. 1), which are separated by differences at

three positions (Table 1). Additional diversity was recorded

in the ITS region (Fig. 2) with the greatest separation

between the strains isolated from Betula and the strains

Table 5. Location of IGS1 insertions between copies within individual

strains (– sequence repeat is absent, see Table 4)

CBS5905 -------

CBS5905 C10 482–523

CBS9090 482–523

CBS9090 C7 -------

NRRLY17430 -------, 650–670

NRRLY17430 C4 398–416, -------

NRRLY17434 432–446, 605–629

NRRLY17434 C1 -------, -------

NRRLY17438 650–670

NRRLY17438 C2 -------

NRRLY17441 604–649

NRRLY17441 C9 -------

NRRLY27348 -------

NRRLY27348 C2 420–435

FEMS Yeast Res 7 (2007) 1399–1408c� 2007 Federation of European Microbiological SocietiesPublished by Blackwell Publishing Ltd. All rights reserved

1406 J.W. Fell et al.

from Fagus and Cornus, which suggests that the ITS region

could be an indicator of host specificity.

Our study viewed ITS variability within the context of

differentiating strains, although many systematic studies

separate species at this level of sequence diversity. For

example, Rhodotorula glutinis and Rhodotorula graminis are

considered to be separate species based on 30% nuclear

DNA relatedness, yet they are identical in the ITS region and

differ at one base position in the D1D2. In contrast,

Sporobolomyces holsaticus is a designated synonym of Spor-

idiobolous johnsonii based on 93% relatedness; they are

identical in the D1D2 region and differ by 5 bp in the ITS.

The extensive ITS variability between CBS 5905T and CBS

9090 (30 bp) may not be unusual as mating strains of

Sporidiobolus salmonicolor differ by 26 bp (Scorzetti et al.,

2002). The use of differences in the ITS region to separate

species must be considered on a case-by-case basis: one rule

does not fit all. There are considerable ITS and IGS sequence

differences between strains of Xanthophyllomyces, but at

present, there is no additional biological data (such as

mating genetics) to suggest the presence of more than one

species. The apparent geographic strain isolation provides

an opportunity to explore the state and rate of speciation

within Xanthophyllomyces.

The diversity in the IGS1 region is due to the presence

of indels (Table 4) and differences at specific nucleotide

positions (Fig. 3). The sequence specificity suggests that

these markers could be used in ecology to track the source

and distribution of strains. The method is complicated

by the presence, in some strains, of intragenomic sequence

heterogeneity, which requires cloning to obtain readable

data. Intragenomic variation in the rDNA spacer regions

is not uncommon among fungi and reports include

ITS variation in Fusarium (O’Donnell & Cigelnik, 1997)

and IGS variability in hybrids of Cryptococcus neoformans

(Bovers et al., 2006). The sexual state in Xanthophyllomyces

is generally considered to initiate with parent–bud mating.

However, Golubev (1995) and Kucsera et al. (1998) reported

mating between independent cells. Consequently, there is

the potential for hybridization between strains. CBS 9090

carries both the 5905T/9090 and 7918T types of sequences in

the ITS and IGS regions. CBS 9090 C4 (ITS) is nearly

identical (3-bp difference) to CBS 7918T (Fig. 2) and CBS

9090 C3 (IGS) aligns with ATCC 24230 (difference of 2 bp,

Fig. 3). However, the distinct tree hosts and geographic

distributions complicate the hypothesis of hybridization.

The potential of sequence heterogeneity caused by poly-

morphisms in the rDNA repeat units could be a more

plausible explanation.

Libkind et al. (2007) reported an extended range of

Xanthophyllomyces to Argentina. The authors isolated

strains from fruiting bodies of the ascomycetes Cyttaria

hariotti, which is a parasite of a southern beech tree,

Nothofagus sp. Based on their ITS analysis, the yeast strains

from Nothofagus represent a separate clade, which is distinct

from clusters representing other tree hosts. Their IGS

analysis indicates that the Nothofagus yeasts are related to

CBS 9090. The Libkind et al. (2007) article appeared while

our manuscript was in revision following journal review.

Consequently, their data were not included in the present

study. This extended geographic range and molecular diver-

sity is important for future phylogenetic studies of strains

with consideration of haplotype network analyses (Posada &

Crandall, 2001; Templeton et al., 2005) and of the structures

and evolutionary rates of indels (Mes et al., 2000; Yamane

et al., 2006).

Strains of Phaffia/Xanthophyllomyces are extensively stu-

died and utilized throughout the world. Major corporations

produce astaxanthin for aquaculture, whereas academic

research examines various aspects including the carotenoid

synthesis pathway, culture conditions for improved expres-

sion, and new applications such as colorants for egg yolks

(Johnson, 2003). Clarification of the taxonomic designation

and molecular diversity of the strains can be important for

patent and safety concerns. The present work should assist

researchers in this regard.

Acknowledgements

This study was supported by NSF Grant DEB 0206521. We

gratefully acknowledge C.P. Kurtzman, USDA, Peoria, IL,

for supply of NRRL strains and the laboratory assistance

with mating studies provided by Nicholas Pinel with NSF

REU support. We acknowledge anonymous journal re-

viewers, whose comments and questions were instrumental

in the revision of the manuscript.

References

Bovers M, Hagen F, Kuramae EE, Diaz MR, Spanjaard L, Dromer

F, Hoogveld HL & Boekhout T (2006) Unique hybrids between

fungal pathogens Cryptococcus neoformans and Cryptococcus

gattii. FEMS Yeast Res 6: 599–607.

Fell JW & Blatt G (1999) Separation of strains of the yeasts

Xanthophyllomyces dendrorohous and Phaffia rhodozyma based

on rDNA IGS and ITS sequence analysis. J Ind Microbiol

Biotech 21: 677–681.

Fell JW, Roeijmans H & Boekhout T (1999) Cystofilobasidiales, a

new order of basidiomycetous yeasts. Int J Syst Bacteriol 49:

907–913.

Golubev WI (1995) Perfect state of Rhodomyces dendrorhous

(Phaffia rhodozyma). Yeast 11: 101–110.

Golubev WI (1998) Xanthophyllomyces Golubev. The Yeasts, A

Taxonomic Study, 4th edn (Kurtzman CP & Fell JW, eds)

pp. 718–719. Elsevier, Amsterdam, The Netherlands.

FEMS Yeast Res 7 (2007) 1399–1408 c� 2007 Federation of European Microbiological SocietiesPublished by Blackwell Publishing Ltd. All rights reserved

1407Origin of Phaffia rhodozyma

Hasegawa M, Kishino H & Yano T (1985) Dating the human-ape

split by a molecular clock of mitochondrial DNA. J Mol Evol

22: 160–174.

Johnson EA (2003) Phaffia rhodozyma: a colorful odyssey. Ind

Microbiol 6: 169–174.

Kucsera J, Pfeiffer I & Ferenczy J (1998) Homothallic life cycle in

the diploid red yeast Xanthophyllomyces dendrorohous. Antonie

van Leeuwenhoek 73: 163–168.

Libkind D, Ruffin A, van Broock M, Alves L & Sampaio JP (2007)

Biogeography, host specificity and molecular phylogeny of the

Basidiomycetous yeast Phaffia rhodozyma and its sexual form,

Xanthophyllomyces dendrorhous. Appl Environ Microbiol 73:

1120–1125.

Mes THM, Kuperus P, Kirschner J, Stepanek J, Oosterveld P,

Storchova H & den Nijs JCM (2000) Hairpins involving both

inverted an direct repeats are associated with homoplasious

indels in non-coding chloroplast DNA of Taraxacum

(Lactuceae: Asteraceae). Genome 43: 634–641.

Miller MW, Yoneyama M & Soneda M (1976) Phaffia, a new yeast

genus in the Deuteromycotina (Blastomycetes). Int J Syst

Bacteriol 26: 286–291.

O’Donnell K & Cigelnik E (1997) Two divergent intragenomic

rDNA ITS2 types within a monophyletic lineage of the

fungus Fusarium are nonorthologous. Mol Phylogenet Evol 1:

103–116.

Phaff HJ, Miller MW, Yoneyama M & Soneda M (1972) A

comparative study of the yeast florae associated with trees on the

Japanese islands and on the west coast of North America. Proc IV

IGS: Fermentation Technology Today (Terui G, ed.), pp. 759–774:

Kyoto Society of Fermentation Technology, Osaka, Japan.

Posada D & Crandall KA (2001) Intraspecific gene genealogies:

trees grafting into networks. Trends Ecol Evol 16: 37–45.

Sambrook J & Russell DW (2001) Molecular cloning: a laboratory

manual. Vol. 2, 3rd edn. Cold Springs Harbor Laboratory

Press, New York, NY.

Scorzetti G, Fell JW, Fonseca A & Statzell-Tallman A (2002)

Systematics of basidiomycetous yeasts: a comparison of large

sub-unit D1D2 and internal transcribed spacer rDNA regions.

FEMS Yeast Res 2: 495–517.

Templeton AR, Maxwell T, Posada D, Stengard JH, Boerwinkle E

& Sing CF (2005) Tree scanning. A method for using

haplotype trees in phenotype/genotype association studies.

Genetics 169: 441–453.

Yamane K, Yano K & Kawagara T (2006) Pattern and rate of

indel evolution inferred from whole chloroplast intergenic

regions in sugarcane, maize and rice. DNA Res 13: 197–204.

FEMS Yeast Res 7 (2007) 1399–1408c� 2007 Federation of European Microbiological SocietiesPublished by Blackwell Publishing Ltd. All rights reserved

1408 J.W. Fell et al.