Microbial Ecology

, Volume 52, Issue 1, pp 90–103

A Multivariate Analysis of Soil Yeasts Isolated from a Latitudinal Gradient

Authors

    • Department of Microbiology and Molecular GeneticsOklahoma State University
Article

DOI: 10.1007/s00248-006-9066-4

Cite this article as:
Vishniac, H.S. Microb Ecol (2006) 52: 90. doi:10.1007/s00248-006-9066-4

Abstract

Yeast isolates from soil samples collected from a latitudinal gradient (>77°S to >64°N) were subjected to multivariate analysis to produce a statistical foundation for observed relationships between habitat characteristics and the distribution of yeast taxa (at various systematic levels) in soil microbial communities. Combinations of temperature, rainfall (highly correlated with net primary productivity), and electrical conductivity (EC) could explain up to ca. 44% of the distribution of the predominant yeast species, rainfall and pH could explain ca. 32% of the distribution of clades in the most common orders (Filobasidiales and Tremellales), whereas vegetation type (trees, forbs, and grass) played the same role for orders. Cryptococcus species with appropriate maximum temperatures for growth predominated in most soils. Cryptococcus species in the Albidus clade of the Filobasidiales predominated in desert soils; Cryptococcus species of other clades in the Filobasidiales and Tremellales predominated in wetter and more-vegetated soils, with Tremellalean species favored in soils of lower pH or higher EC. The predominance of Cryptococcus species in soils has been attributed to their polysaccharide capsules, particularly important when competing with bacteria in arid soils.

Introduction

The investigation of yeasts in soils has produced limited insight into the factors in their distribution and only quasi-anecdotal accounts of the importance of individual species in the communities of soil yeasts. How could it be otherwise, when the techniques of identification and of systematics have changed so greatly? Fonseca et al. [6] speak of “the ubiquitous C. albidus” and then describe the dissection of that “species” into four species in an “Albidus” clade, while also recognizing some species earlier synonymized with or subsumed into Cryptococcus albidus and other species that were not as members of that clade. What is the real distribution of this species? The same report describes “Aerius” clade taxa as “appear to occur predominantly in soils and are only rarely isolated from other sources,” suggesting that environmental factors are important at a supraspecific level. The present report describes an analysis of the distribution of soil yeasts in the MYSW (personal) collection. The MYSW collection was made, using consistent methods of isolation, from soil samples collected over nearly 30 years from more than 21 sites at latitudes ranging from 77°36′S through 9°47′N to 64°46′N. These isolates have now been characterized by commercial sequencing of a portion of rDNA, as well as by previously conventional methods, allowing more accurate identification at both specific and supraspecific levels. This analysis was undertaken to provide a less anecdotal (i.e., statistical) basis for observed relationships between habitat characteristics and yeast distribution at specific and supraspecific levels.

Methods and Materials

Fertile collection sites (i.e., those that yielded yeast isolates) and their characteristics are listed in Table 1. When sample size allowed, pH, conductivity [electrical conductivity (EC) in μmhos/cm], and total soluble salts (TSS in ppm) of extra-Antarctic samples were determined by the Soil, Forage, and Water Analytical Laboratory of this university. The inorganic cation content and water potential of selected Antarctic samples were reported by Klingler and Vishniac [11]. Meteorological data derive primarily from the National Oceanic and Atmospheric Administration Web site. Temperature (mean annual in °C) and precipitation (in cm year−1) records for the year of collection were used as far as possible. Unfortunately, many of these records were incomplete, and the reporting station was not always close to the collection site. When appropriate, temperatures were adjusted by the lapse rate for differences in altitude. Antarctic meteorological data were taken from Friedmann et al. [7]. Data for the Atacama were taken from McKay et al. [14]. Net primary production (NPP, g C m−2 year−1) figures are a consensus of data relating to similar biomes from Beadle et al. [3], Ito and Oikawa [8], and Roy et al. [22], modified as suggested by gross observations at these individual sites.
Table 1

Site characteristics

Site

Collection

Coordinates

Soil

Vegetation

NPP (g C m−2 year−1)

Altitude (m)

Temperature (°C, mean annual)

Rain (cm/year)

TSS (ppm)

EC (mmhos/cm)

pH

Ross Desert, Antarctica (cold desert)a

Summers 1973–1974, 1980–1984

          

 Arid highlands (11 sites)

A801-3, 8, 30; A812-23, 24; A823-1, 2; A834-51, 53, 57, 59

ca. 77–78°S, 160–161°E

Sand/gravel

None visible

3

1400–1600

−24.15

0.45

123a

187a

7.4b

 Dry melt stream, University V. (three sites)

A812-20; A834-63, 66

77°52′S, 160°40′E

Gravelly scree

Traces dry thalli

10

1400

 

1

   

Beringia (± tundra)

 Nome, Alaska

Nom94, 20July94

64°30′N, 165°25′W

Wet, peaty

Dense grass and forbs

350

5

−3.58

51.36

2713

4110

6.9

 Providenya, RFE

Pro94, 24July94

64°27′N,173°11′W

Sand/gravel

Scattered forbs, trees

230

300

−5.16

71.23

840

1272

5.1

 Yttygran, RFE

Ytty94, 25July94

64°26′N, 96°30′W

Wet, peaty

Sparse grass, forbs

200

ca. 100

−5.00

71

   

 Atacama Desert, Chile

  

Sand/gravel

None visible

50

 

6.00

0.1

2184

3310

7.9

 158 km S of Antofagasta

Ata83-9, 8May83

24°41′S, 65°44′W

   

1400

     

Ca. 30 km N of Agua Verde

Ata83-12, 9May83

    

2000

     

Gobi Desert, China (desert)

 Near Shapotou

G86, 22October86

64°46′N, 172°23′W

Sand

A rare forb

100

1200

7.30

10

   

Costa Rica (tropical forest)

 Carara

Car95,15April95

9°47′N, 84°37′W

Dark, sandy

Transitional dry forest

800

ca. 500

22.50

203.2

1426

2160

6.5

 Lomas Barbudal

LoBar95, 16April95

10°27′N, 85°15′W

Dark, silty

Dry forest

600

ca. 100

27.30

152.4

4000

5670

6.6

 Monteverde

Mntv95, 17April95

10°26′N, 83°59′W

Dark, sandy loam

Cloud forest

2000

ca. 1500

18.00

488.8c

2010

3045

4.8

 La Selva

LaSel95, 21April95

10°18′N, 84°48′W

Brown, silty

Lowland rain forest

1000

ca. 50

25.80

396.2

3742

6060

4

Iceland (± tundra)

 Westmannaeyjar

Ice99-1, 22July99

63° 30′N, 20°15′W

Tephra

Dispersed Silene

100

ca. 300

4.78

34.59

88

122

7.5

 Ring Rd., path to glacier

Ice99-3, 24July99

63°26′N, 19°15′W

Scree/sand

Sparse grass, forbs

200

ca. 100

4.37

11.34

240

353

5.1

 Gothafoss

Ice99-5, 29July99

65°37′N, 17°38′W

Sandy

Alpine flora

250

ca. 300

3.55

18.18

365

544

6.3

 Chihuahua Desert, Mexico (desert)

  

Sand/gravel

   

26.42

13.95

   

 Rincon Colorado

RC93B,D; 6March93

25°30′N, 101°20′W

 

Sparse cacti, yucca

125

1200

  

223

338

8.3

 El Baril

EB93B,D; 6March93

25°52′N, 101°9′W

 

None visible

120

1260

  

458

695

8.1

USA

 Bromide, OK ( grass)

BSR92, 13March92

34°26′N, 96°30′W

Gray aggregates

Scattered pasture grasses

350

ca. 260

14.55

260

   

 Estes Park, CO (forest)

       

108

   

 Krumholtz

Col812-K, 26October81

40°12′N, 105°?′W

Silty clay

Krumholtz

325

3580

−4.4

 

436

660

5.6

 Long's Peak

Col812-LPN, 4November81

40°15′N, 105°37′W

Gray aggregates

Krumholtz

325

3900

−8.34

 

463

411

5.6

“Horseshoe Ridge”

Col812-HSR, 3January82

40°24′N, 105°?′W

Mostly sand

Krumholtz

325

2850

2.68

 

271

660

5.5

aThe samples from the arid highlands and from the dry melt stream were each treated as coming from a single site.

bSingle representative soil sample.

cIncludes contribution as clouds.

All samples were imported and destroyed in accordance with the regulations of the United States Department of Agriculture. Soil samples were collected, using native materials (e.g., rocks as pushers), in sterile WhirlpakT bags. Samples were held at site temperatures or refrigerated during transport and were refrigerated on arrival at the laboratory. Weighed subsamples were then used for spreading on dilution plates, overnight enation [29] in shaken liquid medium, sprinkle plates, and/or simulated in situ enation (soil in test tubes with a few drops of liquid medium), depending on the estimated or determined size of the yeast population. With the exception of the earliest Antarctic samples, medium M3C [33] was used in all of these methods, as a medium capable of allowing growth from very small inocula and as the least selective (between yeasts) known [29]. Although the absence of selectivity has been demonstrated experimentally only for certain Antarctic yeasts [30], the enation system has not, in practice, discriminated against common or allochthonous soil yeasts [29]. Population size in the case of sprinkle plates and simulated in situ enation was calculated for statistical purposes on the assumption that one microcolony (i.e., one colony developing on a grain of soil) or one biotype (derived from simulated in situ enation through subsequent dilution for spread plates) was (minimally) equivalent to two viable cells. Incubation temperatures approximated habitat temperatures, using refrigerated incubators set from 10 to 25°C at 5° intervals.

The colonies randomly picked from plates were initially identified (after the customary streaking out for purification) by standard methods [35] and modifications previously employed [31]. Isolates selected as representatives of sites and biotypes were sent to MIDI Labs (Newark, DE) for sequencing rDNA encoding the D2 loop of LSU RNA, using the proprietary method of PE Biosystems. New sequences so determined have been deposited in GenBank. Representative isolates were deposited in the American Type Culture Collection, the Centraalbureau voor Schimmelcultures, the Culture Collection of Fungi from Extreme Environments of the University of Tuscia, the Collezione dei Lieviti Industriali of the Dipartimento di Biologia Vegetale e Biotecnologie Agroambientali (DBVPG) of the University of Perugia, and the Northern Regional Research Laboratory collection of the United States Department of Agriculture (not all isolates were sent to each depository).

Statistical analyses were performed using Microsoft Excel 2004 for Mac version 11.1.1 and Canoco 4.5 [27], a program for multivariate analysis.

Results and Discussion

Yeast abundance and the species (listed alphabetically) isolated at each collection site are reported in Table 2. Some data are missing—this results in large part from preservation failures. Although species assignment has become much easier than it was at the start of the MYSW collection, it was still not possible to make definite species assignments for some isolates. When the rDNA sequence was similar to more than one species, but the biotype did not precisely match that of any described species, both specific epithets have been used, with precedence given to the first-established species. More often, neither the biotype nor the sequence determined provided an exact match. In such cases, the specific epithet of the most similar species was used, qualified as a “var.” if the sequence differed in fewer than three bases, but as a “cf.” (to the most nearly related species) if the sequence differed in three or more bases. The latter are here considered undescribed new species.
Table 2

Yeast communities

Site

cfu/g soil

Taxon

Frequency

Example

GenBank

Ross Desert, Antarctica

ca. 1 microcolony

    

 Arid highlands

Cryptococcus albidosimilis

0.08

A823-2Y761=ATCC 76863=CBS 7711=NRRL Y-17463

AB032667

Cr. friedmannii

0.07

A812-24a-Y432/29=NRRL Y-17509

AB032674

Cr. vishniacii

0.52

A801-3a-Y86/19=NRRL Y-17474

 

Debaryomyces hansenii

0.03

A812-24a-Y450/41=NRRL Y-12773

 

Lalaria sp.

0.08

  

Leucosporidium scottii

0.03

A812-24a-Y440/28=NRRL Y-17464

 

Mesophilic Ascomycete

0.08

  

Sterigmatomyces sp.

0.03

  

Trichosporon cutaneum

0.03

  

 Dry melt stream

 

Cryptococcus antarcticus

0.75

A812-20bY693=ATCC76663=CBS 7687=NRRL Y-17461

AB032670

Cr. vishniacii

0.13

A812-20b-Y493/43=NRRL Y-17465

 

Ustilago sp.

0.13

  

Beringia

 Nome

8 × 102

Cryptococcus aquaticus

0.08

Nom94 Y7=CBS 9273=DBVPG 7820

 

Cr. albidus

0.02

Nom94 Y50=CBS 9276= DBVPG 7823

 

Cr. friedmannii/saitoi

0.02

Nom94 Y18=CBS 9271=DBVPG 7818

 

Cr. gastricus/gilvescens

0.02

Nom94 Y1=BS 9272=DBVPG 7819

 

Cr. liquefaciens var.

0.04

Nom94 Y46=CBS 9275=DBVPG 7822

 

Cr. terricola

0.02

Nom94 Y45=CBS 9274=DBVPG 7821

 

Cr. victoriae

0.78

Nom94 Y2=CBS 9263=DBVPG 7812

 

Cystofilobasidium sp.

0.02

Nom94 Y47=CBS 9277=DBVPG 7916

 

 Providenya

3 × 104

Bensingtonia cf. yamatoana

0.27

Pro94 Y29=CBS 9278=DBVPG 7841

AY204483

Cryptococcus aerius var.

0.05

Pro94 Y54=CBS 9287=DBVPG 7848

 

Cr. cf. aerius B

0.02

Pro94 Y57=CBS 9290=DBVPG 7851

AY189152

Cr. antarcticus var.

0.12

Pro94 Y8=CBS 9291=DBVPG 7852

 

Cr. aquaticus

0.01

Pro94 Y69=CBS 9294

 

Cr. cf. aquaticus

0.01

Pro94 Y50=CBS 9300

 

Cr. friedmannii/saitoi

0.09

Pro94 Y43=CBS 9295=DBVPG 7855

 

Cr. gastricus/gilvescens

0.03

Pro94 Y37=CBS 9297=DBVPG 7857

 

Cr. cf. victoriae

0.01

Pro94 Y46=CBS 9299=DBVPG 7946

AY204479

Cr. terricola

0.01

Pro94 Y55=CBS 9301=DBVPG 7858

 

Cr. cf. terricola

0.03

Pro94 Y64=CBS 9303=DBVPG 7860

vd. AY180910

Cr. watticus var.

0.27

Pro94 Y5=CBS 9281=DBVPG 7844

 

Cystofilobasidium sp.

0.01

Pro94 Y62=CBS 9304=DBVPG 7893

 

Leucosporidiella creatinivora

0.03

Pro94 Y14=CBS 9305=DBVPG 7861

 

Rhodosporidium cf. kratochvilovae

0.01

Pro94 Y45=CBS 9306=DBVPG 7947

AY368154 var.

 Yttygran

54 × 101

Candida norvegica

0.02

Ytty94 Y31=CBS 9313=DBVPG 7840

 

Cryptococcus friedmannii/saitoi

0.02

Ytty94 Y20=CBS 9319=DBVPG 7836

 

Cr. gastricus/gilvescens

0.56

Ytty94 Y48=CBS 9307=DBVPG 7827

 

Cr. tephrensis var.

0.02

Ytty94 Y42=CBS 8968

 

Cr. terricola

0.02

Ytty94 Y6=CBS 9318=DBVPG 7835

 

Cr. victoriae

0.10

Ytty94 Y22=CBS 9314=DBVPG 7832

 

Cr. victoriae var.

0.04

Ytty94 Y28=CBS 9316=DBVPG 7834

 

Cr. watticus

0.08

Ytty94 Y1=CBS 9317=DBVPG 7831

 

Cystofilobasidium sp.

0.02

  

G. pullulans var.

0.02

Ytty94 Y39=CBS 9320=DBVPG 7838

 

Leucosporidium cf. scottii

0.02

Ytty94 Y24=CBS 9321=DBV PG 7837

AY204480

Mrakia frigida

0.06

Ytty94 Y12=CBS 9322

 

Trichosporon sp.

0.02

Ytty94 Y44=CBS 9324=DBVPG 7839

 

Atacama Desert, Chile

0.2 microcolony

    

 158 km S of Antofagasta

 

Cryptococcus cf. albidus B

1

AT83-9 Y1=CBS 9481=DBVPG 4638

DQ004250

 ca. 30 km N of Agua Verde

Cr. cf. albidus B

1

AT83-12 Y41= CBS 9485

as DQ004250

Gobi Desert, China near Shapotou

ca. 7 microcolonies

Cryptococcus chernovii

0.14

G86 Y4=CBS 9488=DBVPG 4643

 

Cr. tephrensis

0.43

G86 Y1

 

Filobasidium floriforme

0.43

G86 Y3=CBS 9487=DBVPG 4644

 

Costa Rica

 Carara

2.67 ± 1.22 × 102

Bullera cf. surugaensis

0.04

Car95 Y20 CBS 9327 DBVPG 7864

AY204482

Candida humilis

0.04

Car95 Y5=CBS 9333=DBVPG 7869

 

Candida cf. nitratophila

0.04

Car95 Y14=CBS 9337=DBVPG 7895

 

Cryptococcus cf. heimaeyensis A

0.35

Car95 Y15=CBS 9325=DBVPG 7862

AY204484

Cr. cf. heimaeyensis B

0.04

Car95 Y10=CBS 9328=DBVPG 7865

AY250757

Cr. podzolicus var.

0.04

Car95 Y18 CBS 9326 DBVPG 7863

 

Saccharomyces exiguus var.

0.13

Car95 Y1= CBS 9330=DBVPG 7866

 

Saccharomyces dairenensis

0.04

Car95 Y3

 

Williopsis mucosa

0.22

Car95 Y8=CBS 9325=DBVPG 7862

 

Williopsis cf. saturnus

0.04

Car95 Y7=CBS 9335

as AY204886

Fermentative Ascomycete Y12

0.09

Car95 Y6, Y12

 

 Lomas Barbudal

1.68 ± 1.17 × 102

Bullera cf. coprosmaensis

0.12

LoBar95 Y3=CBS 9343=DBVPG 7876

AY230853

Candida cf. parapsilosis

0.18

LoBar95 Y4=CBS 9347=DBVPG 7879=NRRL Y-27516

AY204882

Candida sp.

0.06

LoBar95 Y8=NRRL Y-17456

 

Cryptococcus cf. laurentii B

0.06

LoBar95 Y19=CBS 9345=DBVPG 7743

AY250756

Cr. luteolus var.

0.12

LoBar95 Y10=CBS 9340=DBVPG 7874

 

Cr. cf. luteolus B

0.06

LoBar95 Y11=CBS 9341=DBVPG 7875

AY204481

Cr. cf. neoformans

0.06

LoBar95 Y1=CBS 9342=DBVPG 7737

AY204487

Rhodotorula araucariae

0.06

LoBar95 Y18=CBS 9346=DBVPG 7878

 

Sporidiobolus ruineniae

0.29

LoBar95 Y5=CBS 9338=DBVPG 7872

 

 LaSelva

16.14 ± 2.56 × 103

Candida sp. nov.

0.04

LaSel95 Y29=CBS 9363=DBVPG 7897

 

Cryptococcus podzolicus var.

0.68

LaSel95 Y1=CBS 9355=DBVPG 7886

 

Williopsis cf. saturnus

0.29

LaSel95 Y3=CBS 9362=DBVPG 7949=NRRL Y-27521

AY204886

 Monteverde

17.37 ± 0.60 × 103

Cryptococcus podzolicus

1

Mntv95 Y1=CBS 9349=DBVPG 7881

 

Iceland

 Westmannaeyar

32.74 ± 2.01 × 102

Cryptococcus heimaeyensis

0.27

Ice99-1ToM Y8=CBS 8933=ATCC MYA-1759

 

Cr. tephrensis

0.55

Ice99-1ToM Y5=CBS 8935=ATCC MYA-1765

 

Cr. terricola

0.02

Ice99-1TxM Y21=CBS 9211

 

Cr. victoriae var.

0.11

Ice9-1ToM Y11=CBS 8936=ATCC MYA-1761

 

Dioszegia cf. hungarica

0.05

Ice99-1ToM Y23=CBS 9208=DBVPG 7782

DQ004251

Leucosporidiella creatinivora

0.01

Ice99-1TxM Y5=CBS 9210=DBVPG 7744

 

Leucosporidiella fragariae

0.01

Ice99-1ToM Y58=CBS 9209

 

 RingRoad

ca. 2 × 102

Cryptococcus antarcticus var.

0.14

Ice99-3TxM Y26=CBS 9216=DBVPG 7763=ATCC MYA-1212

 

Cr. friedmannii/saitoi

0.03

Ice99-3TxM Y30=CBS 9231=DBVPG 7758

 

Cr. gastricus/gilvescens

0.27

Ice99-3TxM Y9=CBS 9217=DBVPG 7747

 

Cr. terricola

0.22

Ice99-3TxM Y24=CBS 9220=DBVPG 7749

 

Cr. watticus

0.05

Ice99-3TxM Y7=CBS 9228=DBVPG 7756

 

cf. Dothichiza pityophila

0.03

Ice99-3TxM Y31=CBS 9229=DBVPG 7757

 

G. pullulans

0.24

Ice99-3TxM Y20=CBS 9224=DBVPG 7753

 

“Black yeast B”

0.03

Ice99-3TxM Y39=CBS 9230

 

Gothafoss

8.66 ± 4.10 × 102

Cryptococcus aerius

0.06

Ice99-5ToM Y5=CBS 9245=DBVPG 7776

 

Cr. cf. aerius A

0.03

Ice99-5ToM Y10=CBS 9246=DBVPG 7777

AY180911

Cr. antarcticus var.

0.17

Ice99-5ToM Y1=CBS 9232=DBVPG 7764=ATCC MYA 1213

 

Cr. friedmannii/saitoi

0.40

Ice99-5ToM Y11=CBS 9233=DBVPG 7765

 

Cr. terreus

0.03

Ice99-5ToM Y6=CBS 9239=DBVPG 7771

 

Cr. terricola

0.03

Ice99-5TxM Y19=CBS 9248=DBVPG 7779

 

Dioszegia cf. hungarica

0.03

Ice99-5TxM Y23

vd. DQ004251

Chihuahua Desert, Mexico

 Rincon Colorado

2–4 × 102

Cryptococcus albidus

0.65

RC93-2B Y21=CBS 9377=DBVPG 7917

 

Cr. albidus cf. var. ovalis

0.03

RC93-2B Y19=CBS 9381=DBVPG 7920

AY230850

Cr. diffluens

0.09

RC93-2B Y23=CBS 9382=DBVPG 7921

 

Cr. cf. laurentii A

0.14

RC93-2B Y12=CBS 9379=DBVPG 7919

AY230852

Rhodosporidium cf. diobovatum

0.04

RC93-2B Y22=CBS 9384=DBVPG 7922

DQ004252

Unique unknown Cr. YN 3, 7

0.03

  

 El Baril

2.3 × 103

Cryptococcus albidus

0.50

EB93-4B Y15=CBS 9392DBVPG 7926

 

Cr. albidus var. ovalis

0.06

B93-4D Y42=CBS 9410=DBVPG 7942

 

Cr. cf. albidus A

0.06

EB93-4D Y19=CBS 9411=DBVPG 7932

as AY230849

Cr. chernovii

0.10

EB93-4B YN23=CBS 9405=DBVPG 7910

 

Cr. chernovii var.

0.04

EB93-4D Y58

 

Cr. diffluens

0.04

EB4B YN 19, 26, 32,47 Lost

 

Cr. diffluens var.

0.01

EB93-4B YN35=CBS 9402= DBVPG 7907

 

Cr. cf. laurentii A

0.05

EB93-4B Y5, 10; YN 17, 20, 22 Lost

 

Cr. magnus

0.07

EB93-4B Y9=CBS 9404=DBVPG 7941

 

Cr. uzbekistanensis var.

0.07

EB93-4B Y16=CBS 9398=DBVPG 7911

 

Filobasidium floriforme

0.01

EB93-4D Y4=CBS 9415= DBVPG 7913

 

USA

 Bromide, OK

ca. 2 × 102

Bullera oryzae var.

0.04

BSR Y7=CBS 9456=DBVPG 7759

 

Cr. cf. aquaticus

0.02

BSR92 Y49= NRRL Y-17510

DQ004253

Cr. friedmannii

0.02

BSR92 Y29=CBS 9460=DBVPG 4616

 

Cr. heveanensis var.

0.04

BSR92 Y22=CBS 9459=DBVPG 4590

 

?Cr. laurentii

0.02

BSR92 Y10 Lost

 

Cr. cf. luteolus A

0.10

BSR92 Y5=CBS 9455=DBVPG 4614

DQ354576

Cyst. capitatum

0.13

BSR92 Y44=CBS9463=DBVPG 4633

 

Cyst. inf.-miniatum

0.02

BSR92 Y33=CBS 9461=DBVPG 4617

 

Deb. hansenii

0.04

BSR92 Y3, 4 Lost

 

G. pullulans var.

0.02

BSR92 Y21=CBS 9458=DBVPG 4615

 

Occultifer externus var.

0.06

BSR92 Y13=CBS 9457=DBVPG 4632

 

Rh. mucilaginosa

0.43

BSR92 Y41=CBS 9462=DBVPG 4618

 

Estes Park, CO

 Krumholtz

ca. 3 × 103

Cryptococcus cf. aerius A

0.06

Col812-K168=CBS 9620=DBVPG 7955

AY367039

Cr. antarcticus var.

0.31

Col812-K133=CBS 9615=DBVPG 7950

 

Cr. diffluens var.

0.09

Col812-K191=CBS 9621=DBVPG 7956

 

Cr. friedmannii/saitoi

0.06

Col812-K130=CBS 9622=DBVPG 7957

 

Filobasidium floriforme

0.03

Col812-K189=CBS 9623=DBVPG 7958

 

G. pullulans

0.06

Col812-K162=CBS 9618=DBVPG 7953

 

Tr. cutaneum var.

0.125

Col812-K123=CBS 9617=DBVPG 7952

 

Tr. moniliiforme

0.25

Col812-K163=CBS 9616=DBVPG 7951

 

 Horseshoe Ridge

ca. 2 × 103

Cr. antarcticus var.

0.125

Col812-HSR Y144=CBS 9629=DBVPG 7964

 

Cr. diffluens var.

0.17

Col812-HSR Y148=CBS 9631=DBVPG 7966

 

Cr. friedmannii/saitoi

0.08

Col812-HSR Y153=CBS 9632=DBVPG 7967

 

G. pullulans

0.08

Col812-HSR Y156=CBS 9634=DBVPG 7969

 

Lalaria sp.

0.125

Col812-HSR Y174=CBS 9636=DBVPG 7971

 

Rhodotorula laryngis var.

0.04

Col812-HSR Y184=CBS 9638

 

Rhodotorula cf. nothogfagi

0.08

Col812-HSR Y170=CBS 9635=DBVPG 7970

AY368495

Tr. moniliiforme

0.125

Col812-HSR 150=CBS 9633=DBVPG 7968

 

 Long's Peak

ca. 1.5 × 103

Cr. aerius

0.07

Col812-LPN Y401=CBS 9626=DBVPG 7961

 

Cr. cf. aerius A

0.20

Col812-LPN 181=CBS 9627=DBVPG 7962

as AY367039

Cr. antarcticus

0.13

Col812-LPN Y138=CBS 9625=DBVPG 7960

 

Cr. antarcticus var.

0.27

Col812-LPN Y198=CBS 9624=DBVPG 7959

 

Rhodosporidium cf. kratochvilovae

0.07

Col812-LPN Y172=CBS 9628=DBVPG 7963

AY368154

Analysis of the distributions of Table 2 was restricted by the limited occurrence of many species. Over 38% (52) of the listed species were isolated from only a single site. The rest tended largely to occur only at sites in the same geographic area (18 or over 13%) or if not on the same quasicontinent, then at equivalent latitude or climate (a further 38 or nearly 30%)—occurrence at more than three sites was rare. The remaining isolates, although occurring at more than one site, sometimes seemed to have rather unlikely distributions. For example, Cr. albidus was a dominant presence in Mexican soils, with a maximum growth temperature to match, but was perhaps fortuitously present in frigid Nome. Aerial transport can account for the presence of not a few microbes in unsuitable habitats. Soil may also contain yeast species which primarily occupy aboveground niches from which they fall directly to soil. Fermentative yeasts are usually associated with fruits. The species associated with leaves at various stages (“epibionts”) form a group distinct from those dominating in soil and usually appear in soil at lower frequencies than “pedobionts” (species associated with mineral soil), although “epibionts” are more numerous and more diverse in their usual association [1, 4]. Disregarding frequencies ≤0.10 removed some, but not all, of the unlikely species/sample combinations. Unfortunately, the exclusion of exogenous species could not be guaranteed. Large populations of yeasts on decaying leaves or falling with flowers or fruits may land in soil at relatively high frequencies. Low frequencies may indicate the presence of exogenous species but might also indicate that a “pedobiont” was nearing the limit of its realized niche. Fermentative Ascomycetes, such as Candida parapsilosis (frequency 0.18), Saccharomyces exiguus var. (0.13), and Williopsis mucosa (0.22), have been considered unlikely “pedobionts” [1]. The production of ballistospores, as by Bensingtonia, Bullera, and Sporidiobolus species, suggests adaptation to an aboveground environment. Bensingtonia cf. yamatoana (0.27) was dominant at Providenya and Sporidiobolus ruineniae (0.29) at Lomas Barbudal; both were included in the analyses of dominants occurring at single sites. Bullera cf. coprosmaensis (0.12) was also included when analyzing all species with frequencies >0.10.

The numbers of species recurring at similar latitudes suggested that temperature was a major factor in yeast distribution; the correlation of adjusted (for altitude) mean annual temperature with latitude was −0.8474. When analysis of the available data for the four most widely distributed species did not show any significant correlations between temperature and species frequency, canonical correspondence analysis (CCA) using Canoco was employed for the communities of Table 2. Exploratory runs of the automatic forward selection feature of Canoco contrasted the use of all species vs dominant species, species occurring at more than a single site vs single-site species, and all occurrences vs significant (defined as frequencies >0.10) occurrences in various combinations. (For analytical purposes, the Antarctic sites referred to as arid highlands and dry melt stream were treated as single sites, although more than one site are comprehended in these descriptions.) Forward selection of environmental variables always implicated mean annual temperature and cm year−1 rainfall as significant (P<0.05 determined by Monte Carlo permutations) variables. In runs including single-site species, these were the only significant environmental variables; in runs using only species occurring at more than one site, EC was added to this list. Because rainfall and NPP were highly correlated in these data (correlation coefficient > 0.98), the choice of rainfall effectively removed NPP from consideration during forward selection; both variables were therefore used in CCA.

The result of CCA using the 10 species (in the 16 samples for which the appropriate data were available) occurring in significant frequencies at more than one site are shown in Figs. 1a, b. The canonical eigenvalues associated with the four environmental variables (temperature, rain and NPP, and EC) explained about 42.44% of the total inertia (6.598) in this analysis, in the proportions shown in Fig. 2. Variation partitioning [18] indicated that temperature was responsible for ca. 13% of the variation, the union of rainfall and NPP for about 16% (with the intersection of rain and NPP accounting for 11%), and EC for ca. 13%. Axis 1 in Fig. 1a contributed 0.980 eigenvalues (P=0.0080). It is dominated by the influence of rainfall and NPP, with Cryptococcus podzolicus isolated at the greater rainfall end of the axis and the other eight Cryptococcus species (and the single Trichosporon species) all on the negative side of this axis. This arrangement reflects the predominance of desert sites but may also reflect a morphological difference between Cr. podzolicus and the other Cryptococcus species. Cr. podzolicus produces simple pseudohyphae, whereas the other Cryptococcus species does not. Axis 2 (0.901 eigenvalues) is influenced by both temperature and EC. The upper left quadrant reflects the low maximum temperatures for growth (Tmax) of the species located there; the lower left does not do as well. When axis 1 is plotted against axis 3 (0.861 eigenvalues) in Fig. 1b, these variables are better separated; the lower left quadrant now contains the species with the highest Tmax in more appropriate order, with Cr. albidus (Tmax > 30 to > 35°C) followed by Cryptococcus chernovii (Tmax > 25°C, weak growth at 30°C). The upper left quadrant now fails to reflect the lower Tmax values appropriately.
https://static-content.springer.com/image/art%3A10.1007%2Fs00248-006-9066-4/MediaObjects/248_2006_9066_Fig1_HTML.gif
Figure 1

Canonical correspondence analysis of species of soil yeasts occurring at frequency >0.1. (a) Species occurring at more than one site, axes 1 and 2; (b) species occurring at more than one site, axes 1 and 3 [note the separation of conductivity (EC) and temperature]; (c) triplot of all species occurrences at frequency >0.1. Solid symbols (triangles) represent species and species clusters; empty symbols represent sites—squares, desert sites; circles, forest sites; diamonds, tundra-like sites. Boldface species in cluster lists indicate presence at more than one site (compare a and b). Species in clusters are as follows: cluster 1—Bullera cf. coprosmaensis, Candida cf. parapsilosis, Cryptococcus luteolus, Sporidiobolus ruineniae; cluster 2—Cryptococcus albidus, Cr. chernovii, Cr. cf. laurentii; cluster 3—Cr. cf. heimaeyensis, Saccharomyces exiguus, Williopsis mucosa; cluster 4—Cryptococcus gastricus/gilvescens, Cr. terricola, Guehomyces pullulans; cluster 5—Cryptococcus diffluens, Lalaria sp; Cr. antarcticus, Trichosporon cutaneum, T. moniliiforme; Cr. cf. aerius A.

https://static-content.springer.com/image/art%3A10.1007%2Fs00248-006-9066-4/MediaObjects/248_2006_9066_Fig2_HTML.gif
Figure 2

Variation partitioning of environmental variables significantly affecting distribution of species present at more than one site (Fig. 1a).

Figure 1c illustrates the results of adding “single-site” species (resulting in 30 species and 18 active samples) to this analysis. When Fig. 1b is compared with Fig. 1c, it becomes obvious that the added species have not changed the relative positions of the first set of species (boldfaced in the legend to Fig. 1c). The added total inertia of this analysis (11.371) was accompanied by a drop in canonical eigenvalues, resulting in the explanation of only a little over 19% of the species/site associations. The credibility of this analysis must then be assessed using other than statistical evidence.

Temperature, a variable invariably selected, is easy to justify. Van Uden [28] considered maximum growth temperature (Tmax) as important in both identification and speciation; noting that the infraspecific range of Tmax generally varied no more than around 5°C, he successfully predicted a new species on the basis of an excessive range within strains attributed to the same species. Van Uden's rule may seem to be belied by the data in the latest handbook of yeast [2] in which many species are shown with greater Tmax variation. Cr. albidus, for example, is listed as including strains with Tmax varying from 25 to 35°C. When related species were redefined on the basis of rDNA sequences [6], even Cr. albidus conformed to Van Uden's prediction. Tmax for all of the seven strains then ascribed to Cr. albidus was 30°C; the examination of 33 other phenotypically “Cr. albidus” strains (including some of those examined in Barnett et al.) sorted them out into other species with Tmax falling within Van Uden's guidelines. The maximum growth temperatures (Tmax) determined in this laboratory for isolates of the graphed species were generally concordant with their average annual site temperatures, progressing from Cryptococcus vishniacii in the Antarctic desert to Cr. albidus and Cr. chernovii in the Chihuahuan Desert. Although this does not add to our knowledge of individual species' ecology, it does add statistical verification to Van Uden's [28] views. The unanswered question is as follows: why is the maximum temperature range so narrow and so closely tuned to environmental temperatures?

Theoretically, the importance of rainfall and the resulting NPP is obvious. These yeasts are, with few exceptions, heterotrophic saprobes, and all are ultimately dependent on photosynthate. In practice, positions along axis 2 in Figs. 1a, b do not clearly reflect the site vegetation data of Table 1. To be sure, the extreme left is occupied by desert communities, dominated by Cryptococcus species. The extreme right position of Cr. podzolicus has already been the subject of comment. Only in Fig. 1c do clusters 1 and 3 occupy intermediate positions appropriate for the Costa Rican sites from which they came. The position there of the fermentative ascomycetous yeasts (Candida, Saccharomyces, and Williopsis species) suggests that this axis could possibly be viewed as one from oligotrophy (left) to copiotrophy (right). Fermentative ascomycetous yeasts such as these were characterized by Polyakova et al. [21] as “saccharolitics capable of utilizing mono- and disaccharides with the production of ethanol. Such yeasts are rarely isolated from mineral soil horizons, since their typical habitats...are rich in easily metabolizable organic substances. The presence of these yeasts...can be explained by the fact that they occur there in a state of dormancy or the survival of unfavorable conditions.” In a word, they are copiotrophs. Moreover, in excluding them from active soil communities, Polyakova et al., like Babyeva and Chernov [1], among others, were echoing long and widely held views (see [20]). At the other end of this continuum, the case for oligotrophy relies on the scarcity of potential substrates in mineral horizons, a scarcity more likely in all horizons of desert soils. Experimental evidence is lacking.

Rainfall affects not only NPP but also the composition of the soil microbiota and the degree to which soluble inorganic ions are eluviated. The predominance of Cryptococcus species in the desert sites of Fig. 1 is a reflection of their function in soil microbial communities, communities in which bacteria usually outnumber (although not necessarily outmass) yeasts. Cr. albidus, a model representing similar single-celled and encapsulated cryptococci, competes with other microbial heterotrophs in soil. It was successful in competing with the native bacteria in soil microcosms only under conditions of limited water content [32]. The xerotolerance of encapsulated microbes is attributable to their capsules; the encapsulated bacteria in the microcosms required a much higher water content for competitive growth than did Cr. albidus.

It is paradoxical that the arid conditions that favor albidus-like yeasts also tend to produce highly salty soils in which such yeasts cannot grow. Table 1 includes only productive sites, of which the highest EC occurred in two Costa Rican sites lacking any of the left side species of Fig. 1, although temperatures were roughly the same as (but rainfall greater than) those in the Chihuahuan Desert, which contained such species. Soil from the Negev, with a much higher salt content, failed to produce any yeasts (personal observation), but this could equally have resulted from the higher average annual temperatures of this region. The same cannot be said of soil samples from the arid highlands of the Ross Desert; fewer than 44% of these desert sites were productive. Lack of productivity was traced to inorganic ion content [11]. The halotolerance of multisite yeasts Cr. albidus, Cr. antarcticus, and Cr. vishniacii falls between 1 and 2 M NaCl (5.8 and 11.7%); they are considered poorly halotolerant. Yet the endpoints of growth rate vs NaCl molarity differed when all were grown at the same temperature (15°C) on a medium containing 0.5% glucose (luxury for Cr. vishniacii!)—Cr. vishniacii ca. 1.2 M, Cr. antarcticus 1.45–1.7 M, Cr. albidus ca. 1.85 M [34]. Experimental study of osmotolerance in yeasts is rare unless the yeast concerned is either an extremophile or of direct economic importance (neither of which is true of the yeasts of Table 2), but these results render it plausible that even minor variations in EC may, in combination with temperature maxima and degree of oligotrophy, define the realized niches of soil yeast species.

The polyphyletic composition (see [5, 6]) of Cryptococcus suggests caution in generalizing about the genus. Because the family Cryptococcaceae is no better constituted, the next formal rank—order—was used to group yeasts with significant occurrence. Orders have been well defined by molecular phylogeny. The phylogenetic affiliations of all species with a frequency >0.10 on at least one site are given in Table 3. Affiliations were assigned following the reports of Kurtzman and Robnett [12], Fell et al. [5], Fonseca et al. [6], Scorzetti et al. [25], Nagahama et al. [16], Guffogg et al. [9], and Middelhoven et al. [15], with reference to GenBank data for taxa not included in these reports. Orders with single representatives (Ustilaginales, Taphrinales) were omitted from subsequent analyses; soil is, in any case, more likely to be a refuge than a habitat for these plant pathogens. Automatic forward selection of variables important for the 21 samples and six orders produced probabilities <0.05 only for forbs, trees, and grass (in order of increasing P values). CCA using these environmental variables (Fig. 3) explained nearly 33% of a total inertia of 2.250. The positions of the orders in the landscape of grass, forbs, and trees shown cannot necessarily be taken as applicable to other data sets. The Sporidiales and Cystofilobasidiales are shown as highly associated with grass; in another study [26], Cystofilobasidium capitatum, Rhodotorula aurantiaca, and Rh. glutinis had significant frequencies in forest soils. That these species were identified only by their physiological profiles is immaterial, as their phenotypes would still place them in these orders. The association of the (rather few) species in the Saccharomycetales with trees can be considered similarly fortuitous. The Filobasidiales and Tremellales are much better represented in this data set; their separation along the woody axis seems reasonable, given the association of many macrofungal Tremellalean species with wood decay and with ligninolytic and cellulolytic activity, activity absent in the Filobasidiales.
Table 3

Affiliations of species occurring at frequencies >0.10

Species (n sites)

Order (n sites)

Clade (n sites)

(Hymenomycetes; 39)

Cryptococcus albidus (2)

Filobasidiales (22)

Albidus (15)

Cr. cf. albidus B (2)

  

Cr. antarcticus (7)

  

Cr. diffluens (1)

  

Cr. friedmannii/saitoi (1)

  

Cr. vishniacii (2)

  

Cr. cf. aerius A (1)

 

Aerius (2)

Cr. terricola (1)

  

Cr. chernovii (2)

 

Floriforme (3)

Filobasidium floriforme (1)

  

Cr. gastricus/gilvescens (2)

 

Cylindricus (2)

Cr. heimaeyensis (1)

Tremellales (13)

Victoria (7)

Cr. cf. heimaeyensis A (1)

  

Cr.tephrensis (2)

  

Cr. victoriae (3)

  

Bullera cf. coprosmaensis (1)

 

Luteolus (4)

Cr. luteolus var. (1)

  

Cr. podzolicus (2)

  

Cr. watticus (1)

 

Holtermannia (1)

Cr. cf. laurentii A (1)

 

Bulleromyces (1)

Guehomyces pullulans (1)

Cystofilobasidiales (2)

Mrakia (1)

Cystofilobasidium capitatum (1)

 

Cystofilobasidium (1)

Trichosporon cutaneum (1)

Trichosporonales (2)

Cutaneum (2)

Tr. monilliiforme (2)

  
 

(Urediniomycetes, 3)

 

Bensingtonia cf. yamatoana (1)

Sporidiobolales (3)

Colacogloea (1)

Rhodotorula mucilaginosa (1)

 

Curvibasidium (1)

Sporidiobolus ruineniae (1)

 

Ruineniae (1)

 

(Ustilaginomycetes, 1)

 

Ustilago sp.

Ustilaginales (1)

Unassigned

 

(Ascomycetes, 5)

 

Williopsis cf. saturnus (1)

Saccharomycetales (4)

Pichia anomala

W. mucosa

  

Saccharomyces exiguus var. (1)

 

Saccharomyces (1)

Candida cf. parapsilosis (1)

 

Debary./Lodder. (1)

Lalaria sp. (1)

Taphrinales (1)

Unassigned

https://static-content.springer.com/image/art%3A10.1007%2Fs00248-006-9066-4/MediaObjects/248_2006_9066_Fig3_HTML.gif
Figure 3

Canonical correspondence analysis of orders represented in Fig. 1c. (Orders represented by single species—Taphrinales and Ustilaginales—have been omitted.) Note the separation of Tremellales and Filobasidiales along the vector “trees.”

Unfortunately, whereas some Tremellalean yeasts are cellulolytic [17], no yeasts are able to degrade lignin [24], although basidiomycetous yeasts currently classified as Tremellales, Filobasidiales, Trichosporonales, and Cystofilobasidiales are able to assimilate various products of lignin degradation (see [23], but be aware that the classification given there is no longer current).

There is no useful formal systematic level between the genus Cryptococcus and the orders containing the Cryptococcus species of this data set, but there is an informal one; clusters of phylogenetically close species are known as clades. The clades of Table 3 were therefore analyzed. The clades of the Saccharomycetales, Cystofilobasidiales, Sporidiobolales, as well as Taphrinales and Ustilaginales were omitted, largely on the grounds of poor representation at this level. Automatic forward selection picked out average annual rainfall and pH as appropriate variables (P<0.05); the results of CCA conducted with these variables (and NPP) are shown in Fig. 4. Rainfall, NPP, and pH accounted for a little over 32% of the total inertia (4.193) of the nine clades in the 18 samples for which appropriate data were available. The disappearance of the space between the Tremellales (white symbols) and the Filobasidiales (black symbols) is striking, particularly in the close proximity of Victoria and Albidus, the two best represented clades in these orders. The far right position of Luteolus results primarily from the dominance of Cr. podzolicus in the well-watered tropical forests with the lowest pH values; the other species attributed to this somewhat poorly supported clade were never present as dominants. Although pH is one of the defining characteristics of soil types, the importance of pH for yeast distribution was something of a surprise. Many yeasts grow well over a wide pH range, exhibiting acid and alkali tolerance but not acidophily [10] or alkaliphily [13]. The importance of pH is conditional upon its combination with other factors in multivariate analysis; Cr. podzolicus did not occur in soil of equally low pH from a Slovakian coniferous forest [26]. The significance of the pH values of these soil samples may lie in the soil chemistry that produced them; they may here stand proxy for undetermined (and less easily determined) habitat characteristics.
https://static-content.springer.com/image/art%3A10.1007%2Fs00248-006-9066-4/MediaObjects/248_2006_9066_Fig4_HTML.gif
Figure 4

Canonical correspondence analysis of clades: Trichosporonales, empty circle; Tremellales, empty triangles; Filobasidiales, filled triangles. Clades of other orders shown in Fig. 3 were generally represented by fewer than two species and so have been omitted. Note the close pairing of clades from Filobasidiales and Tremellales scored with these environmental variables.

Conclusions

It has been reported many times that yeasts in mineral soil are overwhelmingly basidiomycetous and largely members of the polyphyletic genus Cryptococcus; these analyses apply a statistical imprimatur to these reports and to the importance of some environmental variables. Whereas the success of Cryptococcus species can often be attributed to their capsules, a feature not dependent on phylogenetic details, clades and orders containing “Cryptococcus” species occupied distinct ecological spaces. Desert soils were dominated by Cryptococcus species in the Filobasidiales, Albidus clade, ranging from the Antarctic Cr. vishniacii or Cr. antarcticus of slightly less arid sites in the highest latitudes to Cr. albidus and Cr. albidus-like undescribed new taxa in the Atacama and Chihuahuan Deserts. Tundra-like soils in Beringia and Iceland were dominated mainly by a Cr. gastricus/gilvescens-like species (Filobasidiales, nr. Cylindricus); Cr. gilvescens is characteristic of other tundra soils in Russia [1]. Other Beringian and Icelandic soils were less vegetated (dominant Cr. heimaeyensis, Tremellales, Victoria clade), more vegetated (dominant Cr. friedmannii/saitoi-like species, Filobasidiales, Albidus clade), or with higher EC (dominant Cr. victoriae, Tremellales, Victoria clade). The high altitude, midlatitude forest soils of Colorado were mainly dominated by a variant of Cr. antarcticus, that of a dry tropical forest by a Cr. heimaeyensis-like undescribed new taxon, whereas the tropical rain- and cloud-forest soils were dominated by Cr. podzolicus (Tremellales, nr. Luteolus). In all of these analyses, much has remained “unexplained.” The origins of “unexplained” variation may be various [19], but in these analyses, it is probable that greater representation of individual species would have added explanatory value. It is unfortunate that so much of the work of earlier investigators, some of whom did include site characteristics and species frequencies, has been rendered unusable by what have become uncertain identifications.

Acknowledgments

The patient assistance of Michael Palmer was helpful with statistical aspects of this article. Thanks are also due to Penelope Boston, E. Imre Friedmann, and Becky Johnson for expert assistance in collecting and transporting some soil samples and to Gianluigi Cardinali, Walter Hempfling, and Cletus Kurtzman for productive time spent in their laboratories.

Copyright information

© Springer Science+Business Media, Inc. 2006