INTERNATIONAL
JOURNAL
OF SYSTEMATIC BACTERIOLOGY,
OCt. 1986, p. 52A-530
Vol. 36, No. 4
0020-7713/86/040524-07$02.00/0
Copyright 0 1986, International Union of Microbiological Societies
Interspecific Discontinuity in the Genus Clavispora Rodrigues de
Miranda by Phenetic Analysis, Genomic Deoxyribonucleic Acid
Reassociation, and Restriction Mapping of Ribosomal
Deoxyribonucleic Acid
MARC-ANDRE LACHANCE,l* HERMAN J. PHAFF,2 WILLIAM T. STARMER,3 ALICE MOFFITT,2 AND
LORRAINE G. OLSON'
Department of Plant Sciences, University of Western Ontario, London, Ontario, Canada N 6 A 5B7l; Department of Food
Science and Technology, University of California, Davis, California 956162; and Department of Biology, Syracuse
University, Syracuse, New York 132103
Heterothallic strains belonging to the biologically distinct yeast species CEavispora opuntiae and Clavisporu
Zusituniae were studied by three different methods. The type cultures of the two yeast species exhibited 8%
relatedness as measured by reassociation of unique deoxyribonucleic acid. Ten strains of C . opuntiue and nine
strains of C . Zusitaniae were compared on the basis of their physiological phenotypes and the restriction maps
of their ribosomal deoxyribonucleicacids (rDNAs). Although the two species possessed many similaritiesas well
as certain amounts of intraspecific variation by both approaches, they appeared to constitute well-defined
entities. Unlike C . opuntiae, C . Zusitaniae always utilized L-rhamnose as the sole carbon source and was
resistant to 10 mg of cycloheximide per liter. Strains of C . opuntiae did not utilize L-lysine as the sole nitrogen
source or utilized it very weakly, whereas all strains of C . Zusitaniae grew rapidly on this compound. By
contrast, the hydrolysis of Tween 80 a d the utilization of lactic acid, citric acid, and hexadecane tended to be
more pronounced in C . opuntiae. The rDNA repeating Unit was 9.0 kilobases long in C . lusitaniae, as compared
with 7.6 kilobases in C. opuntiae. The conserved region identified previously in the rDNA of C . opuntiae was
found almost intact in the rDNA of C . Zusitaniae, but the variable regions differed substantially between the two
species.
The genus Clavispora was proposed by Rodrigues de
Miranda (6) to accommodate the teleomorphic state of
Candida lusitaniae and its synonym Candida obtusa.
Clavispora spp. are characterized by the formation of clavate ascospores with indistinct warts. Clavispora lusitaniae
comprises interfertile heterothallic strains isolated from various sources, including humans and other warm-blooded
animals, and materials of plant origin (7). Clavispora
opuntiae is also represented in nature by interfertile
heterothallic strains, but its habitat is restricted almost
entirely to necrotic tissue of various species of the prickly
pear cactus, Opuntia (4, 9). Mating does not take place
between the two species, and thus Clavispora is a good
example of a yeast genus in which biological species (3) are
clearly defined. It is therefore a useful model with which the
significance of different taxonomic approaches may be evaluated.
This paper presents a comparison of the two known
species of Clavispora based on the variation among physiological phenotypes normally investigated in the course of
yeast identification, the relatedness of their unique deoxyribonucleic acids (DNAs), and the patterns observed among
the restriction maps of their ribosomal DNAs (rDNAs). The
last method has been used previously to assess intraspecific
variation within C. opuntiae (2). In that study, 10 strains of
C.opuntiae (not all the same as those used here) were found
Ib share conserved regions in their rDNAs, but they also
exhibited certain differences in their variable regions. We
now show that this last approach is a valuable method for the
study of yeast species delineation and that its results are
* Corresponding author.
consistent with those of other means of establishing yeast
species boundaries.
MATERIALS AND METHODS
Microorganisms. The origins of the strains used in this
study are shown in Table 1. Strains WN9-8 and WN9-20
were obtained in the course of an industrial screening
program conducted by Weston Research, Toronto, Ontario,
Canada. The holotype (type culture) of C. lusitaniae and its
isotype were obtained from the Centraalbureau voor Schimmelcultures, Delft, The Netherlands. All other Clavispora
strains were isolated in the course of various ecological
studies conducted by some of us and including cruises CF
8205 and CF 8314 of the research vessel Cape Florida.
Phenotypic characterization. Physiological responses were
determined by replica plating by the procedures recommended by van der Walt and Yarrow (13). Resistance or
susceptibility to triterpene glycoside inhibitors found in agria
cactus (Stenocereus gummosus) was determined by observing growth on YM agar (Difco) supplemented with 2%
ground dried agria tissue after 4 days at 25°C. Mating types
were determined by mixing actively growing cultures
pairwise on 2% agar containing 1%ground Opuntia tissue
and verifying the presence or absence of zygotes or asci by
microscopy.
Determination of nucleotide composition and reassociation
of unique genome DNA. Nonrepetitive genome DNA was
extracted, purified, and radioiodinated as described by Phaff
and co-workers (4). The guanine-plus-cytosine content of
each )genome DNA was calculated from buoyant density
values in cesium chloride. Reassociation experiments were
conducted in triplicate by the methods of Price et al. (5).
524
Downloaded from www.microbiologyresearch.org
by
IP: 88.99.165.207
On: Sun, 18 Jun 2017 17:04:53
VOL. 36, 1986
VARIATION IN THE GENUS CLAVZSPORA
525
TABLE 1. Strains used in this study
Strain"
Clavispora lusitaniae
CBS 4413'
CBS 6936T
81-467C (UCD)
82-429 (UCD)
82-606B (UCD)
83-1156-2 (UWO)
83-1068-1 (UWO)
83-1080-1 (UWO)
79-257-1 (SU)
Mating
typeb
-
+
-
-
-
+
+
Isolation substrate
Source of isolation
substrate
Digestive tract
Peel juice
Cladode rot
Cladode rot
Cladode rot
Drosophila bromeliae
Fruit
Fruit
Trunk rot
fig
Citrus sp.
Opuntia phaeacantha
Opuntia stricta
Cephalocereus royenii
80-29 (SU)
WN9-8 (UWO)
+c
Leaf rot
Effluents
WN9-20 (UWO)
-
Effluents
Clavispora opunt iae
77-279T (UCD)
78-54OA' (UCD)
81-677-1 (UCD)
83-718-1 (UWO)
-
+
83-754-1 (UWO)
83-803-2 (UWO)
83-1074-1 (UWO)
79-241-3 (SU)
81-333-2 (SU)
82-106 (SU)
84-505-1 (SU)
+
+
Ipomoea sp.
Opuntia stricta
Opuntia stricta
Zdria columnaris
Agave sp.
Nielsen Chocolate
factory
Nielsen Chocolate
factory
Cladode rot
Cladode rot
Cladode rot
Decaying fruit
Opuntia stricta
Opuntia stricta
Opuntia lindheirneri
Cephalocereus royenii
Decaying fruit
Drosophila mulleri
Decaying fruit
Stem rot
Opuntia stricta
Opuntia stricta
Opuntia stricta
Myrtillocactus cochal
Cladode rot
Cladode rot
Cladode rot
Opuntia _ficus-indica
Opuntia wentiana
Opuntia _ficus-indica
Locality
Portugal
Israel
Rincon Mountains, Ariz.
Font Parisienne, Haiti
Beef Island, British Virgin Islands
Grand Cayman, Cayman Islands
Cayman Brac, Cayman Islands
Cayman Brac, Cayman Islands
El Arenoso, Baja California
North, Mexico
Tucson, Ariz.
Toronto, Ontario, Canada
Toronto, Ontario, Canada
Yarrawonga, Australia
Borallon, Australia
Ozona, Tex.
Little Conception Island,
Bahamas
Great Inagua, Bahamas
Great Inagua, Bahamas
Cayman Brac, Cayman Islands
San Telmo, Baja California North,
Mexico
Guadalajara, Mexico
Prudencio, Venezuela
Kamuela, Hawaii, Hawaii
Unknown
Debaromyces
hansenii CBS 767=
I, Isotype; ', holotype. Strain numbers are culture collection numbers used in all three institutions that we are affiliated with. Culture collections in which the
original isolates are deposited are given in parentheses: UCD, Department of Food Science and Technology, University of California, Davis; UWO, Department
of Plant Sciences, University of Western Ontario; and SU, Department of Biology, Syracuse University.
Strains with the same mating type as the holotype are designated + ; those with the same mating type as the isotype are designated - .
Mating with strain 83-1156-2 only.
Mapping of rDNA. The preparation, restriction, electrophoresis, and mapping of rDNA were performed as described by Lachance et al. (2), except that ethanol precipitation between each endonuclease digestion was eliminated.
The endonuclease BamHI was used in addition to ApaI,
EcoRI, KpnI, and XhoI, which were used previously.
Multivariate analysis. Physiological responses were coded
for each strain as 0.0 (no growth), 0.33 (weak), 0.67 (slow or
latent), or 1.0 (strong and rapid). Restriction sites were
coded for each strain as 0.0 (absent), 0.5 (variable), or 1.0
(fixed). Clustering was performed on cosine matrices of
unstandardized data by the equally-weighted-pair agglomeration method. An analysis of the principal components was
performed on centered, unranged data (covariance matrix).
FORTRAN programs LOSIDE and PCA (devised by M.-A.
Lachance), used for these purposes, followed common algorithms. See Sneath and Sokal(8) for a general discussion of
the methods of multivariate analysis and their applications in
taxonomy.
RESULTS
Physiological responses. The mean responses of 9 strains of
C. lusitaniae and 10 strains of C. opuntiae to 68 physiological tests are shown in Table 2. Inspection of these data
reveals that utilization of L-rhamnose as the sole carbon
source and growth in the presence of 10 mg of cycloheximide
per liter unequivocally distinguished C. lusitaniae from C .
opuntiae. In addition, C . opuntiae generally exhibited a
weak or negative utilization of L-lysine as the sole nitrogen
source, whereas C. lusitaniae grew efficiently on this compound.
Cluster analysis of physiological responses. The group
structure elicited by cluster analysis of the physiological
responses of Clavispora strains is shown in Fig. 1. Strains of
C. opuntiae formed a dense cluster in which strains 81-333-2
and 83-718-1acted somewhat as outliers. Examination of the
raw data (not shown) indicated that strain 83-718-1 differed
from the others by its failure to hydrolyze Tween 80 and by
its susceptibility to agria inhibitors. Other differences were
mostly in the rate of utilization of certain compounds. For
example, the assimilation of L-sorbose was very rapid in
strain 83-718-1, very weak in strain 81-333-2, and strong but
slow in the rest. Unlike most other strains, strain 81-333-2
failed to grow in the presence of 10% sodium chloride.
Strains of C. lusitaniae exhibited somewhat more nutritional
variability. Strains CBS 6936=,WN9-8, and WN9-20 formed
a separate cluster, probably owing to their ability to grow in
the presence of 50% glucose, a trait not shared by other
strains. Strain CBS 4413' appeared to be the most unusual in
Downloaded from www.microbiologyresearch.org by
IP: 88.99.165.207
On: Sun, 18 Jun 2017 17:04:53
526
INT. J. SYST.BACTERIOL.
LACHANCE ET AL.
TABLE 2. Distribution of physiological responses in 19
77-279
Clavispora strainsa
78-54011
84-505-1
Mean response' of:
Test
79-241-3
C. lusitaniae C . opuntiae
82-106
Inulin
0
0
83-754-1
Sucrose
1
1
3 83-1074-1
Hanose
0
0
83-803-2
Melibiose
0
0
8 1 -333- 2
Galactose
1
1
83-718-1
Lactose
0
0
83- 1068- 1
Trehalose
1
0.83
83-1080-1
Maltose
1
1
PI
.'p
Melezitose
1
1
83-1156-2
!3
Methyl-a-D-glucoside
0.07
0.37
80-29
Starch
0
0
E 79-257-1
Cellobiose
1
1
CBS 6936
Salicin
1
1
VN-9-8
Sorbose
0.88
0.93
VN-9-20
I=Rhamnose
1
0
CES 4413
11-Xylose
0.85
0.97
I=Arabinose
0.19
0
D- Arabinose
0
0
1 .o
0,95
0.90
D-Ribose
0.11
0.03
Methanol
0
0
Cosine
Ethanol
1
1
FIG. 1. Phenogram of Clavispora strains based on their relsopropanol
0
0
sponses to 68 physiological tests.
n-Butanol
0
0
Glycerol
0.63
0.80
Erythritol
0
0
its clustering position. It had important differences such as
Ribitol
0.81
0.43
its ability to utilize D-glucosamine efficiently and its failure to
Xylitol
0.78
0.60
utilize D-xylose, which sets it apart from all other strains,
Galactitol
0
0
including strains of C. opuntiae. It also had several minor
~Mannitol
1
1
mGlucito1
phenotypic differences from other strains.
1
1
myo-Inositol
0
0
Ordination by principal-componentanalysis of physiological
Lactic acid
0.33
0.67
responses. Clustering methods are known to impose group
Succinic acid
1
1
structure, sometimes even on data which are not intrinsically
Citric acid
0.89
1
structured. For this reason, it was of interest to analyze the
m-Malic acid
0.85
1
physiological
responses by means of a trend-seeking method
Gluconic acid
0.22
0.40
which tends to detect group structure where it actually
Glucono-A-lactone
0.26
0.30
exists. The ordination of Clavispora strains on the first two
2-Ketogluconate
1
0.97
principal components of their physiological responses is
i~-Glucosamine
0.15
0.30
"VAcetylglucosamine
0.92
shown in Fig. 2. The first component (47% of variation)
1
Tannic acid
0.63
0.73
clearly separated members of the two species, with some
,4cetone
0
0
minor intraspecific variation detected in this component
Ethyl acetate
0
0
(Fig. 2). The correlations between the physiological reHexadecane
0.19
0.67
sponses
and the first two components (Fig. 3) revealed that
Vitamin-free medium
0
0
Amino-acid-free medium
1
1
Growth at 4°C
0.52
0.63
Growth at 30°C
1
1
OCBS 4413
Y ((2%)
Growth at 37°C
1
1
Gelatin hydrolysis
0
0
Casein hydrolysis
0
0.07
Hydrolysis of Tween 80
0
0.43
,4cid production
0
0
0 m-9-8
1
Nitrate
0
0
0 m-9-20
Nitrite
OCBS 6936
0 03-154-1
0
0
0 83-118-1
0 81-333-2
Ethylamine
1
1
Lysine
1
0.23
0 71-219b-,
Cadaverine
1
1
0 03-1080-1
Y (47%)
NaCl(5%)
1
1
0 84-505-1
0 83-1068-1
NaCl (10%)
0.96
0.63
NaCl(l5%)
0.19
0.03
0 83-tl%-2
0 02-106
Glucose (50%)
0.30
0
Cycloheximide (10 mg/liter)
1
0
Cycloheximide (100 mg/liter)
0.07
0
Synthesis of starchlike compounds
0
0
0 80-29
Diazonium blue B
0
0
079-287-i
Glucose fermentation
1
1
FIG. 2. Ordination of strains of C . Zusitaniae (open circles) and
Agria resistance
1
0.90
C. opuntiae (solid circles) by principal-component analysis of their
'Coded as described in the text.
physiological responses.
Downloaded from www.microbiologyresearch.org
by
-
.s
I'
1
I
~~
IP: 88.99.165.207
On: Sun, 18 Jun 2017 17:04:53
VOL.36, 1986
VARIATION IN THE GENUS CLAVISPORA
0
0
Glucose 50%
0.5
0
~-arabinose
0 Cyclohcximide 1 0 0
~~ucosmine
0
Citrate
Lactate
.
I
i bose
0 RhNWlOle
-0.5
0 Cyclohex~nide I 0
0 Lysine
I
Ribitol
0
nrlrte
0
+
nexaoecane
N-acetyl-gIucosamine
0 nethyl-a-0-gIuCOside
NaCl 10%
0
-0.3
0
Xylitol
NSCl 15X
0
Sorbosc
~
'
~~~~
0
xy1ose
0 Tannin
0 40C
0 GIUCOn8te
0 GI ucono-A- I actone
FIG, 3. Plot of component-descriptor correlations (r[Y]) between the physiological responses of Clavispora strains and the first
two components (Fig. 2).
in addition to responses for L-rhamnose, L-lysine, and 10 mg
of cycloheximide per liter, other responses tended to discriminate between members of the two species, although not
as unequivocally. Among these were growth on hexadecane,
Clavispora opuntiae
83-71 8-1
83-803-2
82- 106
83-754-1
83-1074- 1
77-279
78-540A
84-505-1
81-333-2
79-241 -3
/
/
/Y+-X-A+/ Y A L X A L /DA+-X-I-A-E-/
/+A+-X&---E-/
/ I H - X - H r - /
/ b H - X - H - A 4 - /
/-h-X-M-Ir-/
/P-X-AC-X-A-E-/
/+A+-X-AC-A+/
/ Y H - X r l L - A - /
527
citric acid, and lactic acid and the hydrolysis of Tween 80,
which were stronger in C. opuntiae.
The second component (12% of variation) was representative of traits which varied mostly within C. Zusitaniae (Fig.
2). It resolved strains 79-257-1 and 80-29 at one end of its
spectrum and strain CBS 4413' at the other end. Growth in
the presence of 15% NaCl and growth on glucono-A-lactone
and gluconic acid were negatively correlated with the second
component (Fig. 3), although this combination of traits was
not a property of any particular strain. Among the responses
showing a positive correlation with the second component,
the utilization of L-arabinose was encountered sporadically
in some strains of C. Zusitaniue. Growth on D-glucosamine,
as mentioned above, was strong in strain CBS 4413', and can
be considered unique to that strain, although occasional
weak responses occurred in strains throughout the genus.
The significance of growth on 50% glucose with respect to
the two industrial isolates has been mentioned before.
Growth on malic acid was variable throughout the genus,
and its position near the positive end of the second component was largely due to its weaker utilization by strain 79-257-1.
Invariant traits (mean value of 0 or 1 for both species [Table 21)
were automatically excluded from this analysis.
Characteristics of nonrepetitive DNA. The nucleotide composition of unique DNAs from selected yeast strains and
their percent relative reassociation with the DNA of strain
CBS 6936= are shown in Table 3. A difference of 1.4 to 2.4%
was detected in the nucleotide composition of strains from
the two CZuvispora species, and their interspecific relatedness as estimated by DNA reassociation was only near 8%.
Restriction maps of rDNAs. Restriction maps of tandemly
repeated rDNA clusters are shown in Fig. 4. Note that the
order of the restriction sites is reversed from that used
previously for certain strains of C. opuntiae (2). The repeating units are now shown as standard transcriptional units,
based on the approximate positions of the rRNA cistrons
(M.-A. Lachance, unpublished data). Strains of C. Zusitaniae
differed from strains of C. opuntiae in the length of their
repeating units, estimated at 9.0 and 7.6 kilobases, respectively (Fig. 4). Length polymorphisms were not detected in
TABLE 3. Nucleotide composition and reassociation of unique
DNAs from selected yeast strains with the DNA of strain CBS
6936Ta
Strain or other DNA
mol% guanine +
cytosine (SD)
~~~~~d
(SD)
% Relative
binding
~~
Clavispora lusi taniae
CBS 6936
79-257-1
83-1080-1
83-1156-2
83-1068- 1
80-29
CBS 4413
WN-9-8
WN-9-20
U
Clavispora lusitaniae
CBS 6936T
82-429
CBS 4413'
82-606B
81-467C
45.2 (0.1) [316
45.1 (0.1) [3]
45.2 (0.2) 141
45.7 (0.4) [4]
45.2 (0.3) [3]
76.2 (0.3)
72.3 (1.9)
72.0 (2.8)
70.7 (1.0)
67.0 (1.4)
(100)
94.7
94.2
92.5
87.4
Clavispora opuntiae
81-677-1
77-279=
43.7 (0.3) [31
43.3 (0.2) [816
9.3 (0.7)
9.2 (0.5)
8.5
8.4
Debaromyces hansenii 38.3 (0.5) [ll]
CBS 767=
6.1 (0.4)
4.1
Calf thymus
5.2 (1.0)
2.8
Reannealing was performed between 0.2 kg of '=I-labeled DNA from
strain CBS 6936= and 200 kg of unlabeled DNA from each strain listed.
Reassociation was allowed to proceed for 25 h in 0.5 ml of 280 mh4 phosphate
FIG. 4. Restriction maps of rDNA of Clavispora strains. ApaI
buffer at 65°C. Zero-time binding was 0.64%, and self renaturation was 3.2%
(A), BamHI (B), EcoRI (E), and XhoI (X) sites are shown. Lower(5). The number of nucleotide composition determinations is shown in
case letters identify heterogeneous sites. KpnI sites were absent. kb,
brackets.
Values reported by Phaff
Kilobases.
Downloaded from www.microbiologyresearch.org
by et al. (4).
IP: 88.99.165.207
On: Sun, 18 Jun 2017 17:04:53
1 kb
528
LACHANCE ET AL.
-.%
83-1074-1
3
84-505-1
83-754-1
82-106
INT. J. SYST.BACTERIOL.
79-241-3
83-803-2
.S
.3
S
83-718-1
CBS 4413
80- 29
VN-9-8
83-1068-1
CBS 6936
1 ,o
0,8
0,9
0 4
7
0,6
Cosine
FIG. 5. Phenogram of Cluvispora strains based on the restriction
sites of their rDNAs.
C . opuntiae, except possibly in strain 83-1074-1, which may
exhibit length heterogeneity (2). C . lusitaniae may be variable from strain to strain in the length of its rDNA units.
However, the absence of a restriction endonuclease able to
produce only one cut in the repeating unit for every strain
precludes confirmation of this at this time. Five restriction
sites were conserved through all the strains studied in both
species, except for strain WN9-20, which was heterogeneous
in its first ApaI site. The constant presence in C. lusitaniae
of two XhoI sites not found in C. opuntiae and the constancy
of a BamHI site found exclusively in C . opuntiae are
noteworthy.
Cluster analysis of restriction maps. To facilitate the comparison of restriction maps with physiological data, cluster
analysis and principal-component analysis (see below) were
performed on the restriction site data. The group structure
identified by clustering (Fig. 5 ) was consistent with the
partition of strains into their respective species. In contrast
to the cluster structure elicited with physiological data (Fig.
l), intraspecific variation was minimal as compared with
-
w-9-20
83-1060-1
80-e0
83-1000-1
0 83-1156-2
QD w-9-8
00 679-257-1
00 C86 4413
Q D C 8 S 6936
T
Y2 ( 8 % )
0
84-u)5-1
a 83-1074-1
78-54OA
a 19-279
0
a 81-333-2
a 83-754-1
0
82-106
a 83-803-2
I
0
83-718-1
FIG. 6. Ordination of strains of C. lusituniae (open circles) and
C. opuntiae (solid circles) by principal-component analysis of the
restriction sites of their rDNAs.
interspecific variation; in other words, the clusters representing the species were much tighter. The strains identified
as outliers on the basis of their physiological differences
generally did not stand out as special on the basis of their
restriction maps.
Principal-component analysis of restriction maps. The two
components derived from restriction site data concentrated
more information (69 and 8% [Fig. 61) than did the two
components summarizing physiological data (47 and 12%
[Fig. 2]), perhaps a sign that more discontinuity exists in the
map data, but most certainly due in large part to the smaller
number of variables in the map datum set. Again, the first
component (Fig. 6) sharply separated representatives of the
two species, and the second component depicted intraspecific variations. Strains 83-718-1and 84-505-1constituted the
extremes of a gradient identified in C. opuntiae. This gradient revolves around ApaI and EcoRI sites found in the
variable region of the repeating unit. The extremes of the
gradient are tied to the presence of a unique ApaI site in
strain 83-718-1 and of a unique XhoI site in strain 84-505-1.
There was less intraspecific variation in C.lusitaniae, and it
was resolved by the third component (data not shown),
which accounted for only 6% of the total rDNA restriction
map variation.
DISCUSSION
Species discontinuity in the genus Clavispora. The data
presented above are consistent with the existence of a sharp
discontinuity between C. lusitaniae and C . opuntiae. We
know of no exception to the prezygotic isolation separating
these heterothallic yeast species, as evidenced by their lack
of interspecific mating. However, when considered singly,
nutritional traits such as L-rhamnose assimilation, L-lysine
utilization, or resistance to 10 mg of cycloheximide per liter
83-71 8-1
83-803-2
82-106
-
1
83-754-1
/B-AW-A4X-A-E / 84-505- 1
/B-AEX-A-Ed-E/
83- 1074- 1
/B-AEX-A-E-A-/
79-241-3
/B-AEX-A-E-A+?/
81-333-2
/B-AE X-A-E-A4 / 77-279
78-540A
/X-XAW-A-E-E-A-/
/X-XAEX-A-E-E-/
/X-XAEXXA-E-E+/
/X-XAEXxA-E-E-ab/
/X-XAEX-A-E-E-B/
/X-XAEXXA-E-E--/
/X-XAWxA-E-E-/
/X-XAEXXA-E-E-/
/X-XaEXXA-H-/
83- 1080- 1
CBS 6936
83-1068-1
83-1156-2
79-257-1
CBS 4413
80-29
WN-9-8
WN- 9- 20
.z
a
?
v)
2
3
FIG. 7. Hypothetical filiation among restriction maps of rDNAs
in Clavispora strains. The maps are abbreviated (see the legend to
Fig. 4) and are not to scale.
Downloaded from www.microbiologyresearch.org by
IP: 88.99.165.207
On: Sun, 18 Jun 2017 17:04:53
VOL.36, 1986
VARIATION IN THE GENUS CUVISPORA
normally would not be given enough taxonomic significance
to define species boundaries. Taken as a whole, these and
other nutritional traits exhibited strong correlations, a sign
that their variation is the result of divergent evolution.
The relatively small difference in nucleotide composition
between the DNAs of the two Clavispora species, reported
initially by Phaf€et al. (4), would have left some doubt about
their divergent status. The finding that C. opuntiae and C .
Zusitaniae exhibited only 8% relative DNA similarity is
indeed surprising, but it certainly confirms their species
delineation. The low value may be the result of the exclusively divergent evolution expected to follow a sharp disruption in interfertility such as that which is presumed to have
occurred in the genus Clavispora.
Phylogenetic interpretation of restriction maps. A hypothetical phylogenetic history of the restriction maps is shown
in Fig. 7. It is based on principles enunciated previously (2),
the fundamental assumption being that the recurring presence of a particular restriction site in the majority of strains,
especially in different species, represents an ancestral state
and consequently that the absence of such a site represents
a derived state. Accordingly, the map found for strains
77-279= and 78-540A’ may be viewed as ancestral for C .
opuntiae within the present context (the fact that these
strains are the holotype and an isotype of the species is
purely fortuitous). Many other strains isolated from localities as diverse as the Caribbean Sea, the continental United
States, Africa, and Australia are known to share maps
identical to these two maps (2; Lachance, unpublished data).
In C. lusitaniae, the map of strain WN9-8 may be regarded
as ancestral because it has the greatest number of fixed,
common restriction sites. Intensified sampling might reveal
that other restriction sites are more broadly distributed in
other strains of C. lusitaniae and thus confer upon other
restriction maps a more ancestral status. Likewise, several
alternate filiations among the maps regarded as derived are
equally possible, so the diagram in Fig. 7 should only be
viewed as a working hypothesis.
The assumptions set above do not necessarily match
everyone’s view on the phylogenetic interpretation of restriction maps, although Templeton (12) has shown that the
convergent appearance of a restriction site in two individuals
is much less likely than the convergent disappearance of
such a site. In the present context, these assumptions do not
permit speculation as to which of the two Clavispora species
possesses maps that may be considered ancestral, because
the sites shared by the two species are invariant. If the
“outgroup” principle (10) is followed, however, C.
lusitaniae should be viewed as ancestral, because its rDNAs
are more typical of other yeasts. Published (14) and unpublished (Lachance) data show that the rDNAs of strains of
several other yeast species do not exhibit unequivocal
homology with those found in Clavispora strains. In terms of
length, however, other yeasts are more similar to C.
lusitaniae than to C. opuntiae. Besides, the ecological specialization of C. opuntiae (4, 9) compared with the more
diverse sources of C. lusitaniae (7) also indicate that C .
opuntiae is a more derived taxon. The suggestion that a map
like that of C. lusitaniae 83-1080-1 could have given rise to
the map regarded as ancestral in C. opuntiae is highly
speculative but nonetheless interesting. It implies that the
speciation event was associated with the inversion of a small
ApaI-EcoRI fragment in the variable regions of the maps.
Other putative ancestral maps are equally possible, for
example, that of strain 79-257-1, with its fixed BamHI site
which could be homologous to that found in all strains of C.
529
opuntiae examined so far. The interest of such hypotheses is
that they can be tested by the purification and fine mapping
or sequencing of small fragments.
Tajima and Nei (11) have proposed a relationship between
the time of divergence of two species and the number of
restriction site differences between their DNAs. The time in
million years may be calculated from the formula t = 5/2X,
where ’6 = -log,S/r. Parameter A is the number of fixed
point mutations per base per million years, S is the proportion of restriction sites shared by the individuals being
compared, and r is the number of base pairs per recognition
sequence, in this case, 6. The mutation rate of yeasts is not
known because they lack identifiable fossils. Hunt et al. (1)
have determined a rate of 4 x
for Drosophila spp., a
value in the same order of magnitude as values published for
sea urchins and primates (5 x
and 1 x
respectively). This would trapslate into maximum intraspecific
divergence times of 6 million years for our strains of C.
lusitaniae and 11 million years for our strains of C. opuntiae.
The calculated time of speciation is about 15 million years,
assuming the pathway depicted in Fig. 7. There is no way of
evaluating the accuracy of these estimations, because a
calibration of evolutionary rates in yeasts is lacking. The
unicellular nature of yeasts and, in the present case, their
haploid state, may cause them to possess much higher fixed
mutation rates, and the divergence times calculated here
could be overestimated.
Geographic and ecological aspects. The strains used in this
study represent a relatively broad spectrum of habitats and
localities of isolation. The holotype of C. lusitaniae was
isolated from fruit materials in Israel, and the isotype came
from animal sources in Portugal. Other strains representing
that species came from industrial effluents in eastern Canada
and from different plant or animal hosts in semiarid areas of
Mexico, the southern United States, and the Caribbean Sea.
This diversity of origins was reflected more in the variation
among physiological responses than in the divergence at the
molecular-genetic level (compare Fig. 2 and Fig. 6). The
converse was true in C. opuntiae. The strains of C.opuntiae
were all from cacti, although they were isolated from many
different land masses in the mostly semiarid regions of the
Americas, Hawaii, and Australia. Diversity was moderate at
the molecular level and practically nonexistent at the level of
expressed physiological phenotypes (see Fig. 2 and Fig. 6).
As pointed out earlier, the most unusual strain of C.
opuntiae was 83-718-1. Its differences at the physiological
and rDNA levels might be linked to its unusual isolation
from a columnar cactus. Strain 79-241-3 is also special in that
it is not an Opuntia isolate. Its rDNA is rather unique, but it
blends well with other strains at the nutritional level.
Taxonomic aspects. A previous study of rDNA restriction
mapping in C. opuntiae (2) left open the question of the
applicability of this approach to species delineation. The
data reported here are very encoufaging in this connection.
Although the two known species of Clavispora differ consistently by only 2 or perhaps 3 of 68 physiological responses,
repeated DNA mapping agreed with unique DNA reassociation and provided an additional independent indication
of the boundary between the two species.
ACKNOWLEDGMENTS
This project was sponsored by operating grants from the Natural
Science and Engineering Research Council of Canada to M.-A.L.
and from the National Science Foundationto H.J.P. (DEB 81 08898)
and to W.T.S. (DEB 81 08679 and BSR 84 13168).
Downloaded from www.microbiologyresearch.org by
IP: 88.99.165.207
On: Sun, 18 Jun 2017 17:04:53
530
INT. J. SYST.BACTERIOL.
LACHANCE ET AL.
The technical assistance of V. Aberdeen and the collectioh efforts
of J. F. S. Barker, J. C. Fogelman, A. Fontdevila, W. B. Heed, and
W. R. Johnson are gratefully acknowledged. We are indebted to
J. E. Fein of Weston Research, Toronto, Ontario, Canada, for
authorizing the use of industrial isolates.
LITERATURE CITED
1. Hunt, J. A., T. J. Hall, and R. J. Britten. 1981. Evolutionary
distances in Hawaiian Drosophila measured by DNA reassociation. J. Mol. Evol. 17:361-367.
2. Lachance, M. A., A. Luttikhuis, L. L. Anweiler, D. L. Paylor,
and L. G. Olson. 1985. Assessment of intraspecific variation in
Cfavispora sp. by restriction mapping of tandemly repeated
deoxyribonucleic acid. Int. J. Syst. Bacteriol. 35462466.
3. Mayr, E. 1970. Populations, species, and evolution: an abridgment of animal species and evolution. Harvard University
Press, Cambridge.
4. Phaff, H. J., M. Miranda, W. T. Starmer, J. Tredickj and
J. S. F. Barker. 1986. Cfavisporaopuntiae, a new heterothallic
yeast occurring in necrotic tissue of Opuntiu species. Int. J.
Syst. Bacteriol. 36:372-379.
5. Price, C. W., G . B. Fuson, and H. J. P h a . 1978. Genome
comparison in yeast systematics: delimitation of species within
the genera Schwanniomyces, Saccharomyces, Debaryomyces,
and Pichia. Microbiol. Rev. 42:161-193.
6. Rodrigues de Miranda, L. 1979. Cfuvispora,a new yeast genus
of the Saccharomycetales. Antonie van Leeuwenhoek J. Microbiol. 45479-483.
7. Rodrigues de Miranda, L. 1984. Cfuvispora Rodrigues de
Miranda, p. 120-122. In N. J. W. Kreger-van Rij (ed.), The
yeasts: a taxonomic study, 3rd ed. Elsevier Science Publishers
B. V., Amsterdam.
8. Sneath, P. H. A., and R. R. Sokal. 1973. Numerical taxonomy.
W. H. Freeman & Co., San Francisco.
9. Starmer, W. T., and H. J. Phaff. 1983. Analysis of the community structure of yeasts associated with the decaying stems of
cactus. 11. Opuntia species. Microb. Ecol. 9:247-259.
10. Stevens, P. F. 1980. Evolutionary polarity of character states.
Annu. Rev. Ecol. Syst. 11:333-358.
11. Tajiina, F., and M. Nei. 1982. Biases of the estimates of DNA
divergence obtained by the restriction enzyme technique. J.
Mol. Evol. 18:115-120.
12. Templeton, A. R. 1983. Phylogenetic inference from restriction
endonuclease cleavage site maps with particular reference to the
evolution of humans and the apes. Evolutidn 37:221-244.
13. van der Walt, J. I'., and D. Yarrow. 1984. Methods for the
isolation, maintenance, classification and identification of
yeasts, p. 45-104. In N. J. W. Kreger-van Rij (ed.), The yeasts:
a taxonomic study, 3rd ed. Elsevier Science Publishers, B. V.,
Amsterdam.
14. Verbeet, M. P., H. van Heerikhuizen, J. Klootwijk, R. D.
Fontijn, and R. D. Plahta. 1984. Evolution of yeast ribosomal
DNA: molecular cloning of the rDNA units of Kfuyveromyces
lactis and Hunsenulu wingei and their comparison with the
rbNA units of other Succharomycetoideae. Mol. Gen. Genet.
195:11&125.
Downloaded from www.microbiologyresearch.org by
IP: 88.99.165.207
On: Sun, 18 Jun 2017 17:04:53