Utilization and transport of L-arabinoseby non-Saccharomyces yeasts

Eric P. Knoshaug Æ Mary Ann Franden ÆBoris U. Stambuk Æ Min Zhang Æ Arjun Singh

Received: 24 October 2008 / Accepted: 24 May 2009 / Published online: 14 June 2009

� Springer Science+Business Media B.V. 2009

Abstract L-Arabinose is one of the sugars found in

hemicellulose, a major component of plant cell walls.

The ability to convert L-arabinose to ethanol would

improve the economics of biomass to ethanol

fermentations. One of the limitations for L-arabinose

fermentation in the current engineered Saccharomy-

ces cerevisiae strains is poor transport of the sugar.

To better understand L-arabinose transport and use

in yeasts and to identify a source for efficient

L-arabinose transporters, 165 non-Saccharomyces

yeast strains were studied. These yeast strains were

arranged into six groups based on the minimum time

required to utilize 20 g/L of L-arabinose. Initial

transport rates of L-arabinose were determined for

several species and a more comprehensive transport

study was done in four selected species. Detailed

transport kinetics in Arxula adeninivorans suggested

both low and high affinity components while De-

baryomyces hansenii var. fabryii, Kluyveromyces

marxianus and Pichia guilliermondii possessed a

single component, high affinity active transport

systems.

Keywords Non-conventional yeast �L-Arabinose utilization � Sugar transport �Mutagenesis

Introduction

The potential of biomass as an energy source is

widely recognized. Biomass is the only renewable

feedstock that can provide biofuels, mainly ethanol,

to replace or augment the petroleum-derived liquid

transportation fuels in significant quantities. Cur-

rently, commercial ethanol production is through

fermentation of sugars derived from agricultural

feedstocks. These fermentations rely, almost exclu-

sively, on strains of the yeast Saccharomyces.

Although primary feedstocks vary regionally, corn-

starch in the United States, sugarcane in Brazil, and

sugar beets in Europe, in all these cases only the

hexose sugars are fermented to produce ethanol. For

the economic production of ethanol from less expen-

sive and more abundant lignocellulosic biomass

such as agricultural residues, herbaceous crops, and

short rotation hardwoods, a fermentative organism is

required that would ferment the pentose sugars

D-xylose and L-arabinose present in these feedstocks

with efficiencies similar to that achieved with

glucose.

E. P. Knoshaug (&) � M. A. Franden �B. U. Stambuk � M. Zhang � A. Singh

National Bioenergy Center, National Renewable Energy

Laboratory, 1617 Cole Boulevard, Golden, CO 80401,

USA

e-mail: eric_knoshaug@nrel.gov

Present Address:B. U. Stambuk

Departamento de Bioquımica, Centro de Ciencias

Biologicas, Universidade Federal de Santa Catarina,

Florianopolis, SC 88040-900, Brazil

123

Cellulose (2009) 16:729–741

DOI 10.1007/s10570-009-9319-8

Over the last two decades, significant progress has

been made toward developing yeast that can ferment

the pentose sugars, L-arabinose and D-xylose, to

ethanol (Jefferies and Jin 2004; Lin and Tanaka 2006;

van Maris et al. 2006). Most of this research effort

has focused on xylose since it is the major pentose

sugar in biomass. However, L-arabinose is the third

most abundant sugar in most lignocellulosic biomass.

Moreover, L-arabinose is a major component of

sugars present in several specialized feedstocks that

are produced in large amounts every year. For

example, L-arabinose constitutes 32% of the sugars

in spent brewery grains (Carvalheiro et al. 2004),

28% in corn fiber (Park et al. 2001), 13.5% in wheat

bran (Palmarola-Adrados et al. 2005), and 6.6% in

rice straw (Sulbaran-de-Ferrer et al. 2003). Thus,

conversion of L-arabinose to ethanol would greatly

increase the economic viability of biomass to ethanol

fermentation.

Transport of a sugar into the cell is the first step in

the pathway of its conversion to ethanol. Several

reports have shown that D-xylose uptake is a rate-

limiting step in D-xylose fermentation (van Zyl et al.

1999; Eliasson et al. 2000; Gardonyi et al. 2003;

Pitkanen et al. 2003; Karhumaa et al. 2005) and

inefficient L-arabinose uptake was cited as a possi-

ble rate-limiting step in L-arabinose fermentation in

S. cerevisiae (Becker and Boles 2003; Richard et al.

2003; Karhumaa et al. 2006). It has long been known

that in S. cerevisiae, L-arabinose transport is mediated

by the galactose permease, Gal2p (Cirillo 1968).

However there have been no detailed reports on

L-arabinose transport in non-conventional yeasts

except for the suggestion that L-arabinose uptake in

Candida shehatae could take place via a proton

symporter (Lucas and van Uden 1986). In this study,

we have collected and analyzed a large number of yeast

species as potential sources of efficient L-arabinose

transporter(s) for the purpose of eventually isolating

and expressing suitable L-arabinose transporters in

S. cerevisiae.

Materials and methods

Yeast strain selection and growth

Our rationale for selecting yeast species and strains is

based on the simple reasoning that an organism that

grows well on L-arabinose must have an efficient

mechanism for the uptake of this sugar. In order to

facilitate future experiments, we limited our search

for species that were not reported to be pathogenic,

did not, or rarely, formed hyphae, and grew primarily

as single cells. For the selection of strains based on

these criteria, we relied on the comprehensive

publication on yeasts by Barnett et al. (2000),

previously described L-arabinose fermenting strains

(Dien et al. 1996; Kurtzman and Dien 1998), and

descriptions in the publications by various culture

collections. We obtained 165 strains from 123

different species from the American Type Culture

Collection (ATCC), the National Center for Agricul-

tural Utilization Research (NRRL), or the Centraal-

bureau voor Schimmelcultures (CBS; Table 1).

The routine growth medium for the growth and

maintenance of the yeast strains was YPD [1% Bacto-

yeast extract (Difco), 2% Bacto-peptone (Difco), 2%

dextrose, and when needed for solidification, 2%

Bacto agar (Difco)]. For the testing of growth on

L-arabinose, similar medium (YPA) was used in which

glucose was replaced with L-arabinose. Minimal

medium contained 0.67% yeast nitrogen base without

amino acids and 2% dextrose or 2% L-arabinose.

In order to determine auxotrophic requirements of

various mutants, ‘‘drop-out’’ media (Sigma Aldrich or

Clontech) were used. These media contain various

supplements, but are missing adenine, uracil, or one of

the amino acids.

Evaluation of strains for growth on L-arabinose

Cells, grown in YPD medium to early stationary

phase, were collected by centrifugation, washed with

water and suspended in 25 mL YPA medium at an

initial density of OD600 of 0.2. Cultures were

incubated at 30 �C with shaking at 220 rpm. Growth

was monitored by periodic measurement of the optical

density of the cultures. Utilization of L-arabinose was

determined by measuring the sugar remaining in the

medium after various periods of growth by the

analysis of the filtered media by high performance

liquid chromatography (HPLC) on Hewlett-Packard

(HP) 1090 instrument using a Bio-Rad HPX-

87H hydrogen ion resin column and an HP 1047A

external refractive index detector. The mobile phase,

0.001 N H2SO4, was run at 55 �C with a flow rate of

0.6 mL/min.

730 Cellulose (2009) 16:729–741

123

Table 1 Yeast species and strains screened

Species Strain Group OD600

Ambrosiozymamonospora

CBS-2554 2 18.1

Ambrosiozymamonosporaa

NRRL

Y-1484

2 25.5

Arxula adeninivorans CBS-8244 1 37.2

Bullera coprosmaensis CBS-8284 3 20.9

Bullera dendrophila ATCC-24608 4 4.2

Bullera globispora CBS-6981 4 23.5

Bullera mrakii CBS-8288 3 32.5

Bullera penniseticola CBS-8623 4 16.3

Bullera pseudoalba CBS-7227 3 26.8

Bullera sinensis CBS-7345 5 7.1

Bullera varaiabilis CBS-7354 3 21.6

Bulleromyces albus ATCC-18568 6 0.3

Bulleromyces albus CBS-6302 4 21.2

Candidaarabinofermentansa

NRRL

YB-1299

2 34.6

Candidaarabinofermentansa

NRRL

YB-1984

2 31.3

Candidaarabinofermentans

NRRL

YB-2248

3 26.2

Candida atlantica CBS-5263 3 28.8

Candida auringiensis CBS-6913 3 32.4

Candida auringiensis NRRL

Y-11848

3 34.6

Candida auringiensis NRRL

Y-11849

3 32.5

Candida bertae ATCC-58889 4 32.3

Candida bertae var.

bertaeCBS-4605 4 32.2

Candida blankii ATCC-18735 2 37.7

Candida blankii CBS-6734 6 9.7

Candida cellulolytica CBS-7920 2 32.4

Candida chilensis CBS-5719 6 0.2

Candida conglobata CBS-5808 4 24.5

Candida conglobata NRRL

Y-1504

3 26.4

Candida diddensiae CBS-6032 3 29.5

Candida diddensiaea NRRL

Y-7589

2 26.5

Candida entomaeaa NRRL

Y-7785

2 33.5

Candida insectorum ATCC-22940 2 35.3

Candida ishiwadae ATCC-22018 6 5.0

Candida ishiwadae CBS-7348 6 8.2

Candidamembranifaciens

CBS-1952 2 35.5

Table 1 continued

Species Strain Group OD600

Candidamembranifaciensa

NRRL

Y-2089

2 27.1

Candidamethanosorbosa

CBS-7029 1 35.6

Candida mogii CBS-2032 3 23.6

Candida nanaspora CBS-7200 3 20.8

Candida nitratophila CBS-2027 3 25.6

Candida ovalisa NRRL

Y-17662

2 14.8

Candida peltata CBS-5576 1 32.1

Candida pignaliae CBS-6071 3 23.4

Candida populi CBS-7351 4 26.0

Candida psychrophila CBS-5956 6 0.2

Candida rhagiib ATCC-22983 5 12.3

Candida rhagii CBS-4432 6 5.1

Candida santjacobensis CBS-8183 6 3.8

Candida sequanensis CBS-8118 6 3.9

Candida shehatae var.

shehataeCBS-2779 6 3.4

Candida shehatae var.shehatae

NRRL

Y-17029

6 4.8

Candida silvicultrix CBS-6269 5 8.0

Candida sonorensis CBS-6793 2 31.1

Candida succiphila CBS-7297 2 32.0

Candida succiphilaa NRRL

Y-11997

2 14.4

Candida succiphila NRRL

Y-11998

3 15.4

Candida tenuis CBS-2885 3 25.3

Candida tenuisb ATCC-10573 5 10.6

Candida vanderwaltii CBS-5524 1 34.8

Cryptococcus albidus ATCC-10666 5 13.1

Cryptococcus albidusvar. albidus

CBS-8395 3 32.8

Cryptococcus aerius CBS-155 3 23.5

Cryptococcuscellulolyticus

CBS-8294 3 27.2

Cryptococcus curvatus ATCC-10567 6 10.3

Cryptococcus curvatus CBS-5324 6 10.4

Cryptococcus luteolus ATCC-32044 6 2.0

Cryptococcus luteolus CBS-8014 2 33.8

Cryptococcus terreus ATCC-11799 4 16.5

Cryptococcus terreus CBS-7528 3 21.5

Cystofilobasidiumcapitatum

CBS-7420 6 0.2

Debaryomyceshanseniib

ATCC-2357 5 10.0

Cellulose (2009) 16:729–741 731

123

Table 1 continued

Species Strain Group OD600

Debaryomyces hansenii ATCC-36239 4 21.1

Debaryomyces hansenii CBS-941 3 32.2

Debaryomyces hanseniivar. fabryii

CBS-2753 1 45.2

Debaryomycesnepalensis

NRRL

Y-7108

1 25.8

Debaryomycespolymorphus

NRRL

Y-2022

6 6.4

Debaryomycesrobertsiae

CBS-2934 5 10

Debaryomyces yamadae NRRL

Y-11714

6 5.6

Fellomyces borneensis CBS-8282 5 19.7

Fellomyces chinensis CBS-8278 6 2.8

Fellomyces distylii CBS-8545 4 32.2

Fellomyces fuzhouensis CBS-6133 4 21.8

Fellomyces horovitziae CBS-7515 6 0.2

Fellomyces lichenicola CBS-8315 6 1.7

Fellomycesogasawarensis

CBS-8544 2 37.0

Fellomyces penicillatus ATCC-32128 5 12.4

Fellomyces penicillatus CBS-5491 5 20.0

Fellomyces polyborus ATCC-32821 6 11.5

Fellomyces polyborus CBS-8333 5 18.3

Fellomycessichuanensis

CBS-8277 6 6.4

Fibulobasidiuminconspicuum

CBS-6963 3 27.7

Filobasidium floriforme ATCC-22367 3 15.2

Filobasidium floriforme CBS-6241 4 14.1

Hansenula glucozyma ATCC-18938 6 5.4

Kluveromycesmarxianus

CBS-712 3 20.9

Kluveromycesmarxianus

CBS-1089 2 23

Kluveromycesmarxianus

CBS-1555 2 24.9

Kluveromycesmarxianus

CBS-1557 3 26.8

Kluveromycesmarxianus

CBS-2173 3 19.9

Kluveromycesmarxianus

NRRL

Y-8281

3 13.7

Kockovaellamachilophila

CBS-8607 5 12.7

Kockovaella sacchari CBS-8624 4 21.4

Myxozyma geophila CBS-7219 4 24.9

Myxozyma kluyveri CBS-7332 6 2.5

Table 1 continued

Species Strain Group OD600

Myxozyma lipomycoides CBS-7038 3 16.3

Myxozyma melibiosi CBS-2102 3 27.4

Myxozyma monticola CBS-7806 4 13.9

Myxozyma mucilagina CBS-7071 3 17.0

Myxozyma neglecta CBS-7058 3 20.2

Myxozyma vanderwaltii CBS-7517 3 26.4

Pachysolenstannophilus

ATCC-32691 3 20.1

Pichia angophorae CBS-5823 6 3.0

Pichia bovis NRRL

YB-4184

2 36.6

Pichia capsulate CBS-136 4 30.7

Pichia capsulate NRRL

Y-1842

3 26.2

Pichia ciferrii NRRL

Y-1031

6 3.4

Pichia guilliermondiia NRRL

Y-2075

2 40.0

Pichia haplophila NRRL

Y-7860

6 5.1

Pichia heimii CBS-6139 5 5.6

Pichia holstiib ATCC-13689 5 11.5

Pichia holstii CBS-2026 6 5.4

Pichia kodamae CBS-7081 5 21.9

Pichia kodamae NRRL

Y-17234

4 25.8

Pichia methanolica ATCC-46071 2 26.9

Pichia methanolica CBS-6515 1 27.5

Pichia mississippiensis NRRL

YB-1294

2 30.0

Pichia naganishii ATCC-32148 2 35.5

Pichia naganishii CBS-7259 2 38.6

Pichia nakazawae var.

nakazawaeNRRL

Y-7903

5 12.8

Pichia philogaea NRRL

Y-7813

3 21.7

Pichia rabaulensis CBS-6797 1 38.7

Pichia scolyti CBS-4802 2 29.1

Pichia scolyti NRRL

Y-5512

1 45.8

Pichia silvicola NRRL

Y-1679

3 20.8

Pichia stipitis ATCC-62970 6 9.0

Pichia stipitis CBS-5773 6 8.0

Pichia tannicola NRRL

Y-7499

6 7.3

Pichia trehalophila CBS-5361 2 33.1

732 Cellulose (2009) 16:729–741

123

Transport assays

Cells were grown on L-arabinose to mid-log phase and

collected by centrifugation, washed twice in water, and

then suspended in water. The yeast suspension was

adjusted to 30–60 mg dry weight/mL. Uptake rates and

inhibition of L-arabinose transport were measured as

previously described (Stambuk et al. 2003). For our

initial transport studies, we had purchased and used

several batches of L-(1-14C)arabinose from American

Radiolabeled Chemicals (ARC), Inc. (St Louis, MO,

USA). However, we discovered that the 1-14C-labeled

L-arabinose from this source was heavily contaminated

with other radioactive compounds (see Results). The

transport experiments reported here were done using

L-(1-14C)arabinose (54 mCi/mmol) that was custom-

synthesized by Moravek Biochemicals Inc. (Brea, CA,

USA). The analysis of the radioactive substrate was

performed by thin layer chromatography (TLC) as

described (Han and Robyt 1998) using Whatman Silica

Gel 60 A TLC plates (cat# 4410-221). We established

that the compound migrating to the position expected

for L-arabinose is, in fact, L-arabinose using a previ-

ously described method (Sturgeon 1984). Briefly, the

preparations were dried down and resuspended to a

final concentration of 180 lM in 50 lL of 80 mM Tris

HCl, pH 8.6 containing 500 lM nicotinamide adenine

dinucleotide (NAD). This suspension was then treated

with b-galactose dehydrogenase (Roche) for 1 h at

room temperature. Disappearance of L-arabinose,

which is oxidized to form L-arabinono-1,5-lactone,

was confirmed by analyzing the reaction mixture by the

TLC method described above.

Mutagenesis

Cells were grown in YPD to stationary phase,

washed, and then treated with 3% ethyl methanesul-

fonate (EMS) in 0.1 M potassium phosphate buffer

(pH 7.2) for 1.0 h or 1.5 h at 30 �C. At the end of the

treatment, EMS was inactivated by diluting the cells

in 5% sodium thiosulfate solution. The treated cells

were washed and plated on YPD medium. After

2 days of growth, the yeast colonies were tested for

auxotrophy by replica plating onto the synthetic

minimal medium. For testing of growth of mutants on

other sugars, the glucose from the minimal medium

was replaced with appropriate sugars.

Results

Utilization of L-arabinose

The first step in the selection of strains for transport

studies was to determine how well various strains

utilized L-arabinose for growth. The efficiency with

which the selected strains utilized the sugar was

Table 1 continued

Species Strain Group OD600

Pichia triangularis CBS-4094 3 31.4

Pseudozyma flocculosa CBS-167.88 5 6.6

Psuedozyma fusiformata CBS-6951 6 1.3

Pseudozyma rugulosa CBS-170.88 3 21.8

Rhodosporidiumsphaerocarpum

CBS-5939 3 27.7

Rhodotorula acuta ATCC-42713 3 32.4

Rhodotorula fragaria CBS-6254 4 16.0

Rhodotorula pustula ATCC-32034 6 6.2

Sirobasidiumintermedium

CBS-7805 4 15.9

Smithiozyma japonica CBS-7319 4 13.2

Sporidiobolus ruineniae CBS-5001 6 2.5

Stephanoascus smithiae CBS-5657 3 37.6

Sterigmatomyces elviae CBS-5922 3 29.3

Sympodiomycopsispaphiopedili

CBS-7429 3 27.5

Tremella aurantia CBS-6965 4 15.1

Tremella cinnabarina CBS-8234 6 2.5

Tremella enchephala CBS-6968 3 20.7

Tremella foliacea CBS-8228 6 1.0

Tremella fuciformis CBS-8225 6 3.3

Tremella indecorata CBS-6976 4 15.5

Tremella nivalis CBS-8487 3 25.3

Trichosporon laibachii ATCC-90037 3 19.7

Trichosporon laibachii CBS-2495 3 16.7

Trichosporon loubieri ATCC-56048 3 24.5

Trichosporon loubieri CBS-7719 4 20.7

Trichosporonmoniliiforme

ATCC-22164 3 22.9

Trichosporonmoniliiforme

CBS-2820 3 29.1

a These strains were not analyzed at 18 h; some of them could

be in group 1b These strains were not analyzed at 144 h; some of them

could be in group 4

Cellulose (2009) 16:729–741 733

123

determined by measuring their growth on L-arabinose

and by measuring the amount of L-arabinose that was

consumed by these strains during various periods of

growth. The amount of L-arabinose utilized was deter-

mined by HPLC analysis of the growth medium as

measured by HPLC analysis of medium before and after

growth. Based on this profile of L-arabinose utilization,

the 165 strains were classified into one of six groups

(Tables 1, 2). The group classification and final optical

density of the cultures when grown on L-arabinose are

indicated in Table 1 and a summary is provided in

Table 2. Further studies of L-arabinose transport

focused on the species and strains in groups 1–3.

Transport of L-arabinose

For our initial transport studies, we used L-[1-14C]arab-

inose from American Radiolabeled Chemicals Inc.

(Becker and Boles 2003) as this was the only supplier

available. However, our TLC analysis of the chemical

obtained from this source showed that it contained a

mixture of L-arabinose and other unknown compounds

and that L-arabinose constituted only a small fraction

of these chemicals (Fig. 1). It should be noted that the

Table 2 Summary of L-arabinose utilization

Group Criteria No. of species

and/or strains

1 Used [90% of available

L-arabinose within 18 h

9

2 Used all L-arabinose within 24 h 27

3 Used all L-arabinose within 48 h 51

4 Used all L-arabinose within 144 h 24

5 Used [10% of available

L-arabinose in 144 h

17

6 Did not use any or \10% of

available L-arabinose in 144 h

37

a b c 1 2 3 1 2 1

Fig. 1 Analysis of commercial L-(1-14C)arabinose. a 0.1 lL

of D-(1-14C)galactose (lane 1), D-(1-14C)xylose (lane 2) and

L-(1-14C)arabinose from ARC, Inc. (lane 3) were separated on

Whatman K6 TLC plates. The positions of migration of other

related compounds on this TLC system are indicated. b Sample

of L-(1-14C)arabinose from ARC, Inc. incubated without

galactose dehydrogenase (lane 1) and with galactose dehydro-

genase (lane 2). c 0.1 lL of L-(1-14C)arabinose custom-

synthesized by Moravek Biochemicals, Inc. The TLC was

performed as described in materials and methods

734 Cellulose (2009) 16:729–741

123

D-(1-14C)xylose and D-(1-14C)galactose obtained from

the same source are essentially pure (Fig. 1a). In the

chromatogram of the L-(1-14C)arabinose sample, we

have indicated the positions to which L-arabinose and

other related compounds migrate in this TLC system

(Fig. 1a). We did not attempt to identify the unexpected

compounds present in the L-arabinose preparations. We

used the method described by Sturgeon to establish that

the compound migrating to the position expected for

L-arabinose is, in fact, L-arabinose. This method uses

galactose dehydrogenase to convert L-arabinose to

L-arabino-lactone which is then converted to L-arabi-

nonic acid in aqueous solution (Sturgeon 1984). The

spot expected to be L-arabinose disappears when the

preparation is treated with galactose dehydrogenase

while the other spots remain unaffected (Fig. 1b).

Subsequently, we obtained L-(1-14C)arabinose from

custom synthesis by Moravek Biochemicals Inc. This

preparation was also analyzed by the methods described

above and was found to be free of contaminating

compounds (Fig. 1c).

In order to select the eventual strain(s) for isolation

of an L-arabinose transporter, we evaluated L-arabi-

nose transport rates in several strains deemed suitable

for further investigation (Table 3). We chose these

strains based on their L-arabinose utilization efficiency

(groups 1–3 in Tables 1, 2) and suitability for genetic

analysis and mutagenesis (see below). We also tested

whether L-arabinose transport in these strains is active

by measuring transport in the presence of 2,4-dinitro-

phenol (DNP), an inhibitor of active H?-symporters

(Stambuk et al. 2003). Transport of L-arabinose was

inhibited by DNP in all of the strains tested indicating

that these strains harbor active transport mechanisms

for L-arabinose (Table 3).

Detailed transport studies with A. adeninivorans,

D. hansenii var. fabryii, Kluyveromyces

marxianus, and P. guilliermondii

Based on the relatively high L-arabinose transport

velocities found in our initial transport studies (see

Table 3), we selected A. adeninivorans, D. hansenii

var. fabryii, K. marxianus (CBS-1089), and P. guil-

liermondii for detailed L-arabinose transport studies.

In the case of A. adeninivorans, the non-linear Eadie-

Hofstee plot suggests the existence of both low and

high affinity transport systems for L-arabinose in this

yeast (Fig. 2, panel a). The low affinity transport

system has an apparent Km of 250 mM and a Vmax of

20.0 nmol/mg min while the high affinity system has

an apparent Km of 0.3 mM and a Vmax of 6.7 nmol/

Table 3 Initial rates of L-arabinose transport (nmol mg-1 min-1)

Group Species Strain L-(1-14C)arabinose concentration (mM) Inhibition by

1.25 mM DNP (%)a

133 13.3 1.3

1 A. adeninivorans CBS-8244 32.4±0.5 4.4±0.3 5.2±0.2 96

2 C. arabinofermentans NRRL YB-1299 10.8±0.0 12.6±0.4 11.5±1.9 99b

2 C. blankii ATCC-18735 2.6±0.3 1.6±0.4 0.4±0.2 96

3 D. hansenii CBS-941 2.7±0.1 2.3±0.1 1.7±0.1 89

1 D. hansenii var. fabryii CBS-2753 14.2±2.5 15.2±0.7 13.3±1.5 98

1 D. nepalensis NRRL Y-7108 5.0±2.0 3.8±0.8 3.3±0.1 99

3 K. marxianus CBS-712 4.8±0.2 5.7±0.6 1.2±0.1 88

2 K. marxianus CBS-1089 28.5±0.1 20.5±1.3 20.2±0.6 96c

2 K. marxianus CBS-1555 4.4±0.3 4.5±0.9 1.0±0.0 94

2 P. bovis NRRL YB-4184 10.0±2.2 7.7±0.5 7.9±0.7 91

2 P. guilliermondii NRRL Y-2075 8.2±8.0 16.8±3.0 22.0±1.8 100d

1 P. methanolica CBS-6515 14.2±1.3 10.6±0.7 9.2±0.1 99b

1 P. scolyti NRRL Y-5512 4.6±0.8 4.1±0.1 2.9±0.2 94

a Transport inhibition assayed at 10.0 mM unless indicatedb Transport inhibition assayed at 1.3 mMc Transport inhibition assayed at 3.0 mMd Transport inhibition assayed at 0.3 mM

Cellulose (2009) 16:729–741 735

123

mg min. In our initial screen of transport activity and

inhibition of transport in A. adeninivorans, an L-

arabinose concentration of 10 mM was used

(Table 3). At this low level of substrate, only the

high affinity transport system would be impacted

showing that this component is an active transport

system. The nature of the low affinity transport system

remains to be determined.

The linear Eadie-Hofstee plots for D. hansenii var.

fabryii, K. marxianus, and P. guilliermondii suggest

that single component, high affinity transport systems

were responsible for L-arabinose uptake (Fig. 2,

panels b, c, and d, respectively). These transport

systems had Km values of 0.10, 0.14 and 0.07 mM

and Vmax values of 15.0, 24.0, and 22.5 nmol/mg min

for D. hansenii var. fabryii, K. marxianus, and

P. guilliermondii, respectively. These high capacity

transport systems allowed these species to effectively

metabolize 20 g/L of L-arabinose within 18 h for

A. adeninivorans and D. hansenii var. fabryii, and

within 24 h for K. marxianus and P. guilliermondii.

Mutagenesis of selected species

Our ultimate objective for these studies is not only to

characterize L-arabinose utilization and uptake by

various yeast species, but to identify the species that

contain efficient L-arabinose transporter(s) that could

be targeted for isolation of appropriate transporters

and engineering for expression in S. cerevisiae. Since

use of appropriate mutants is one of the approaches

for identifying and isolating genes of interest, we

examined the colony and cell morphology of the

species collected for these studies and chose only the

species that would be expected to be amenable to

the application of genetic and biochemical methods.

For example, we excluded strains such as Bullera

penniseticola, Pichia capsulate, Pichia kodamae,

Smithiozyma japonica, Sterigmatomyces elviae, and

the Myxozyma and Tremella strains that formed slimy

colonies and would be difficult to replica plate. We

also excluded species such as Ambrosiozyma monos-

pora, Trichosposon laibachii, and the Psuedozyma

strains that tended to display significant mycelial

form of growth from which it would be difficult to

obtain single colony mutants. Since no or very

limited genetic information is available for most of

the yeast species collected for this investigation, we

chose several species, mainly from groups 1 and 2,

for mutagenesis to determine if they would yield

ara- mutants (unable to use L-arabinose as a source

of carbon) at a reasonable frequency. We found that a

variety of mutants can be isolated from each of the

selected species with reasonable frequency (Table 4).

We therefore concluded that it would be possible to

isolate appropriate mutants from one or more of the

a

0

5

10

15

20

25

30

35

0 5 10 15 20

Velocity / [L-arabinose] (nmole mg-1 min-1 mmol-1)V

elo

city

(nm

ole

mg-1

min

-1)

b

0

5

10

15

20

25

30

35

0 2 4 6 8 10 12

Velocity/ [L-arabinose] (nmol mg-1 min-1 mmol-1)

Vel

ocit

y (n

mol

mg-1

min

-1)

0

5

10

15

20

25

30

35

0 50 100 150 200

Velocity / [L-arabinose] (nmol mg-1 min-1 mmol-1)

Vel

oci

ty (n

mo

l mg

-1 m

in-1

)

d

c

0

5

10

15

20

25

30

35

Velocity / [L-arabinose] (nmole mg-1 min-1 mmol-1)

Vel

oci

ty (n

mo

l mg

-1 m

in-1

)

0 50 100 150 200

Fig. 2 Eadie-Hofstee plots

of L-arabinose transport.

Initial rates of labeled

L-arabinose uptake (0.065–

592.2 mM) by L-arabinose

grown cells were

determined.

a A. adeninivorans.

b D. hansenii var. fabryii.c K. marxianus.

d P. guilliermondii

736 Cellulose (2009) 16:729–741

123

organisms that were determined to harbor suitable

L-arabinose transporter(s) (see Fig. 2). Screening

through approximately 155,000 colonies for A.

adeninivorans, 132,000 colonies for D. hansenii

var. fabryii, and 76,400 colonies for P. guilliermon-

dii, we isolated 15, 106, and 72 mutants, respectively

that were unable to grow on agar plates containing

2% L-arabinose (Fig. 3). Based on the L-arabinose

utilization pathway in fungi (Fig. 4), it is expected

that an L-arabinose transport defective mutant would

be in the class of mutants that can utilize L-arabitol

and D-xylose for a carbon source. We tested the ara-

mutants for growth on arabitol, xylose, and xylitol to

identify the mutants expected to be deficient at various

steps of the L-arabinose utilization pathway. From

these tests we identified four A. adeninivorans, 14

D. hansenii var. fabryii, and 14 P. guilliermondii

mutants with the phenotype expected of an L-arabi-

nose transport deficient mutant.

Discussion

We set out to screen a variety of strains for the purpose of

isolating an efficient L-arabinose transporter and rea-

soned that strains that grew well on L-arabinose would

have an efficient transport mechanism for this sugar. We

realize that some of the strains may also have efficient

L-arabinose transporter(s), but don’t grow well on the

sugar due to inefficiency of one or more other steps in

L-arabinose metabolism. An examination of Table 1

reveals that there is significant variability in the

efficiency of L-arabinose utilization among the various

strains of the same species. For example, various strains

of D. hansenii, which has been shown to be polyphyletic

and highly variable at the genomic level (Kurtzman and

Robnett 1997; Corredor et al. 2003), can be classified in

any one of the groups 1, 3, 4, or 5. On the other hand,

some genera and species showed little or no diversity.

The three strains of C. arabinofermentans, two strains of

C. diddensiae, three strains of C. succiphila, and five

strains of K. marxianus are in similar groups (group 2 or

3). Similarly, five of the six strains of the Trichosporon

spp. are in group 3 and the remaining one is in group 4.

All strains of A. monospora, C. auringiensis, C. bertae,

C. membranifaciens, C. shehetae, C. curvatus, F. pen-

icillatus, P. holstii, P. naganishii, P. stipitis, T. laibachii,

and T. moniliiforme were assigned to the same group as

the other strains of the same species. Generally a strain

from one collection had similar growth properties as the

equivalent strain from a different collection as expected.

For example, A. monospora (NRRL Y-1484 and CBS-

2554), C. membranifaciens (NRRL Y-2089 and CBS-

195), and K. marxianus (NRRL Y-8281 and CBS-712)

behaved identically from both collections. Similarly,

four species, A. monospora Y-1484, C. arabinofermen-

tans YB-2248, C. auringiensis Y-11848, and C. succi-

phila Y-11997 and Y-11998, were previously reported

to ferment L-arabinose (Dien et al. 1996; Kurtzman and

Dien 1998) and also grew well on L-arabinose in our

screening. In our study, A. monospora Y-1484 and

C. succiphila Y-11997, utilized 20 g/L of L-arabinose in

24 h (group 2) while the rest consumed all L-arabinose

in 48 h (group 3). However, there were instances where

the equivalent strains from different collections had

somewhat different growth properties as is the case for

P. kodamae (NRRL Y-17234 and CBS-7081) and

P. scolyti (NRRL Y-5512 and CBS-4802).

Initial uptake rates varied widely for strains in

groups 1–3. No clear relationship could be established

Table 4 EMS-induced mutagenesis of selected species

Species Strain No. of

colonies

tested

No. of

auxotrophs

obtained

Mutant

frequency

(%)

Types of auxotrophs

obtained

A. adeninivorans CBS-8244 14,440 5 0.03 Ade-, Arg-, Met-

C. arabinofermentans NRRL YB-1299 19,750 13 0.07 Ade-, Arg-, Lys-

D. hansenii var. fabryii CBS-2753 24,400 26 0.11 Ade-, Arg-, His-, Ile-, Leu-, Lys-, Met-,

Thr-, Trp-, Tyr-

D. nepalensis NRRL Y-7108 2,760 5 0.18 Ade-, Arg-

K. marxianus CBS-712 32,650 4 0.01 Ade-, VaL-

P. guilliermondii NRRL Y-2075 6,920 10 0.14 Ade-, Arg-, Lys-, Thr-, Tyr-

P. methanolica CBS-6515 27,200 13 0.05 Ade-, Arg-, Lys-, Met-

Cellulose (2009) 16:729–741 737

123

between the initial uptake rates and the time it took to

completely consume all L-arabinose from the culture

media. Strains in group 1 did not always have higher

transport rates than those in groups 2 or 3. The

transport of the sugar is, of course, just one (first) step

in its metabolism and the velocities of other enzymes

in the pathway, the flux rates of the intermediates of

the pathway, and various metabolic regulatory mech-

anisms contribute to the rate at which L-arabinose

would be utilized by these organisms.

We selected three species to determine if mutants

unable to utilize L-arabinose for growth could be

isolated. We were able to isolate a number of such

mutants from each of these species (Fig. 3). Except

for the mutants that can utilize xylose but do not grow

on xylitol, all the mutants can be explained by a

single mutation in one of the steps in the L-arabinose

utilization pathway in yeasts (Chiang and Knight

1960; Fig. 4). This pathway has been confirmed to

be functional by demonstrating that S. cerevisiae

a

b

c

Xylitol+4

Xylitol-0

Xylose+4

Xylitol+0

Xylitol-0

Xylose-0

Arabitol+4

Xylitol+2

Xylitol-7

Xylose+9

Xylitol+0

Xylitol-2

Xylose-2

Arabitol-11

L-Arabinose-15

Xylitol+14

Xylitol-0

Xylose+14

Xylitol+7

Xylitol-2

Xylose-9

Arabitol+23

Xylitol+43

Xylitol-9

Xylose+52

Xylitol+6

Xylitol-25

Xylose-31

Arabitol-83

L-Arabinose-106

L-Arabinose- 72

Xylitol+14

Xylitol-0

Xylose+14

Xylitol+3

Xylitol-2

Xylose-5

Arabitol+ 19

Xylitol+33

Xylitol-3

Xylose+36

Xylitol+0

Xylitol-17

Xylose-17

Arabitol-53

Fig. 3 Phenotypes of

L-arabinose-negative

mutants obtained from

a A. adeninivorans.

b D. hansenii var. fabryii.c P. guilliermondii. The

numbers of mutants in each

class are indicated in the

boxes. The grey shaded boxindicates the mutant class in

which transport is expected

to have been impacted

738 Cellulose (2009) 16:729–741

123

expressing genes for each of the enzymatic steps can

utilize L-arabinose for growth (Richard et al. 2003).

The mutants that do not grow on xylitol may have

defects in a specific xylitol transporter, as recently

described for Aspergillus niger (van de Vondervoort

et al. 2006) and, since they do not grow on

L-arabinose either, another mutation at an earlier step

in the metabolic pathway of L-arabinose may exist.

Alternatively, another as yet unknown pathway for

xylose utilization may exist.

The only known yeast transporter for L-arabinose,

galactose permease (Gal2p) of S. cerevisiae, trans-

ports L-arabinose by a facilitated diffusion mecha-

nism and functions well as long as the external

concentration of L-arabinose is higher than the

internal cellular concentration (Cirillo 1968; Kou

et al. 1970) with a transport velocity of 0.32 nmol/

mg min at 10 mM L-arabinose (Becker and Boles

2003). We have found that Gal2p has a very high

capacity for L-arabinose transport, but only at high

concentrations of L-arabinose (Knoshaug and Singh,

unpublished). The L-arabinose transport rates in the

non-conventional yeast strains were higher than that

reported for Gal2p in S. cerevisiae at low concentra-

tions of L-arabinose. In particular, the active, high

affinity transport systems of K. marxianus and P.

guilliermondii had transport velocities that were,

respectively, 64 and 53 times greater than the

reported velocities for Gal2p of S. cerevisiae. It

would be interesting to express these transporters in

S. cerevisiae and to determine if they perform in

S. cerevisiae as well as they do in their native species.

In a recent report Fonseca and co-workers ana-

lyzed L-arabinose transport kinetics by two yeast

species, C. arabinofermentans NRRL YB-2248 and

P. guilliermondii PYCC 3012 (Fonseca et al. 2007).

Although their results indicated presence of both

high- and low-affinity transport systems for L-arab-

inose in these yeasts, unfortunately these data have

to be evaluated with care as these authors used

L-[1-14C]arabinose from American Radiolabeled

Chemicals Inc., a substrate that appears to be heavily

contaminated with other labeled compounds (see

Results). Thus, the radioactive uptake transport rates

they obtained could reflect the entrance of any of the

labeled contaminants, and not just L-[1-14C]arabi-

nose. Indeed, our data shown in Table 3 and Fig. 2

indicate that both yeast species have single high

affinity L-arabinose transport system, and although

some variations might be expected in different yeast

isolates, this discrepancy could be easily explained by

the contaminated substrate these authors used.

Our data show that A. adeninivorans has, interest-

ingly, a combination of low and high affinity, high

capacity L-arabinose transport systems that allow the

complete assimilation of 20 g/L of L-arabinose within

18 h demonstrating that, through the combination of

transport systems with different affinities for L-arab-

inose, a strain can be developed that can utilize all

L-arabinose quickly. S. cerevisiae, well established as

an ethanol producer, would be the preferred organism

for engineering since non-Saccharomyces yeasts have

thus far not been shown to be capable of producing

significant quantities of ethanol from L-arabinose

(Dien et al. 1996; Kurtzman and Dien 1998).

Provided the pathways for conversion of L-arabinose

to ethanol currently established in S. cerevisiae are

capable of handling a high flux of L-arabinose, one

can envision a scenario where Gal2p, in concert with

a high affinity transporter for L-arabinose introduced

from one of the species described here, would be able

L-arabinose

L-arabinose

L-arabitol

L-xylulose

D-xylose

D-xylulose

D-xylulose-5-P

xylitol

Fig. 4 Fungal pathway for L-arabinose metabolism

Cellulose (2009) 16:729–741 739

123

to efficiently uptake all the available L-arabinose in

any given substrate and yield a highly efficient L-

arabinose fermenting strain.

Acknowledgments This work was funded by the United

States Department of Energy’s Office of the Biomass Program,

the Corn Refiners Association, and the National Corn Growers

Association. We thank C. Kurtzman for providing some of the

strains used in this study.

References

Barnett JA, Payne RW, Yarrow D (2000) Yeasts: characteris-

tics and identification. Cambridge University Press,

Cambridge

Becker J, Boles E (2003) A modified Saccharomyces cerevi-siae strain that consumes L-arabinose and produces

ethanol. Appl Environ Microbiol 69:4144–4150. doi:10.

1128/AEM.69.7.4144-4150.2003

Carvalheiro F, Duarte LC, Medeiros R et al (2004) Optimi-

zation of brewery’s spent grain dilute-acid hydrolysis for

the production of pentose-rich culture media. Appl

Biochem Biotechnol 113–116:1059–1072. doi:10.1385/

ABAB:115:1-3:1059

Chiang C, Knight SG (1960) A new pathway of pentose

metabolism. Biochem Biophys Res Commun 3:554–559.

doi:10.1016/0006-291X(60)90174-1

Cirillo VP (1968) Galactose Transport in Saccharomyces ce-revisiae I. Nonmetabolized sugars as substrates and

inducers of the galactose transport system. J Bacteriol

95:1727–1731

Corredor M, Davila AM, Casaregola S et al (2003) Chromo-

somal polymorphism in the yeast species Debaryomyceshansenii. Antonie Van Leeuwenhoek 84:81–88. doi:

10.1023/A:1025432721866

Dien BS, Kurtzman CP, Saha BC et al (1996) Screening for

L-arabinose fermenting yeasts. Appl Biochem Biotechnol

57/58:233–242. doi:10.1007/BF02941704

Eliasson A, Christensson C, Wahlbom CF et al (2000) Anaerobic

xylose fermentation by recombinant Saccharomyces cere-visiae carrying XYL1, XYL2, and XKS1 in mineral med-

ium chemostat cultures. Appl Environ Microbiol 66:3381–

3386. doi:10.1128/AEM.66.8.3381-3386.2000

Fonseca C, Romao R, de Sousa HR et al (2007) L-Arabinose

transport and catabolism in yeast. FEBS J 274:3589–3600.

doi:10.1111/j.1742-4658.2007.05892.x

Gardonyi M, Jeppsson M, Liden G et al (2003) Control of

xylose consumption by xylose transport in recombinant

Saccharomyces cerevisiae. Biotechnol Bioeng 82:818–

824. doi:10.1002/bit.10631

Han NS, Robyt JF (1998) Separation and detection of sugars

and alditols on thin layer chromatograms. Carbohydr Res

313:135–137. doi:10.1016/S0008-6215(98)00250-X

Jefferies TW, Jin YS (2004) Metabolic engineering for improved

fermentation of pentoses by yeasts. Appl Microbiol Bio-

technol 63:495–509. doi:10.1007/s00253-003-1450-0

Karhumaa K, Hahn-Hagerdal B, Gorwa-Grauslund MF (2005)

Investigation of limiting metabolic steps in the utilization

of xylose by recombinant Saccharomyces cerevisiae using

metabolic engineering. Yeast 22:359–368. doi:10.1002/

yea.1216

Karhumaa K, Wiedemann B, Hahn-Hagerdal B et al (2006)

Co-utilization of L-arabinose and D-xylose by laboratory

and industrial Saccharomyces cerevisiae strains. Microb

Cell Fact 5:18. doi:10.1186/1475-2859-5-18

Kou SC, Christensen MS, Cirillo VP (1970) Galactose trans-

port in Saccharomyces cerevisiae II. Characteristics of

galactose uptake and exchange in galactokinaseless cells.

J Bacteriol 103:671–678

Kurtzman CP, Dien BS (1998) Candida arabinofermentans, a

new L-arabinose fermenting yeast. Antonie Van Leeu-

wenhoek 74:237–243. doi:10.1023/A:1001799607871

Kurtzman CP, Robnett CJ (1997) Identification of clinically

important ascomycetous yeasts based on nucleotide

divergence in the 50 end of the large-subunit (26S) ribo-

somal DNA gene. J Clin Microbiol 35:1216–1223

Lin Y, Tanaka S (2006) Ethanol fermentation from biomass

resources: current state and prospects. Appl Microbiol

Biotechnol 69:627–642. doi:10.1007/s00253-005-0229-x

Lucas C, van Uden N (1986) Transport of hemicellulose

monomers in the xylose-fermenting yeast Candida she-hatae. Appl Microbiol Biotechnol 23:491–495. doi:10.

1007/BF02346066

Palmarola-Adrados B, Choteborska P, Galbe M et al (2005)

Ethanol production from non-starch carbohydrates of

wheat bran. Bioresour Technol 96:843–850. doi:10.1016/

j.biortech.2004.07.004

Park NH, Yoshida S, Takakashi A et al (2001) A new method

for the preparation of crystalline L-arabinose from ara-

binoxylan by enzymatic hydrolysis and selective fer-

mentation with yeast. Biotechnol Lett 23:411–416. doi:

10.1023/A:1005681032082

Pitkanen JP, Aristidou A, Salusjarvi L et al (2003) Metabolic

flux analysis of xylose metabolism in recombinant Sac-charomyces cerevisiae using continuous culture. Metab

Eng 5:16–31. doi:10.1016/S1096-7176(02)00012-5

Richard P, Verho R, Putkonen M et al (2003) Production of

ethanol from L-arabinose by Saccharomyces cerevisiaecontaining a fungal L-arabinose pathway. FEMS Yeast

Res 3:185–189. doi:10.1016/S1567-1356(02)00184-8

Stambuk BU, Franden MA, Singh A et al (2003) D-Xylose

transport by Candida succiphila and Kluyveromycesmarxianus. Appl Biochem Biotechnol 105–108:255–263.

doi:10.1385/ABAB:106:1-3:255

Sturgeon RJ (1984) Arabinose. In: Bergmeyer HU (ed)

Methods of enzymatic analysis, 3rd edn. Verlag Chemie,

Weinheim, pp 427–431

Sulbaran-de-Ferrer B, Aristiguieta M, Dale BE et al (2003)

Enzymatic hydrolysis of ammonia-treated rice straw.

Appl Biochem Biotechnol 105–108:155–164. doi:10.

1385/ABAB:105:1-3:155

van de Vondervoort PJI, de Groot MJL, Ruijter GJG et al

(2006) Selection and characterisation of a xylitol-dere-

pressed Aspergillus niger mutant that is apparently

impaired in xylitol transport. Appl Microbiol Biotechnol

73:881–886. doi:10.1007/s00253-006-0527-y

740 Cellulose (2009) 16:729–741

123

van Maris AJA, Abbott DA, Bellissimi E et al (2006) Alcoholic

fermentation of carbon source in biomass hydrolysates

by Saccharomyces cerevisiae: current status. Antonie

Van Leeuwenhoek 90:391–418. doi:10.1007/s10482-006-

9085-7

van Zyl WH, Eliasson A, Hobley T et al (1999) Xylose

utilisation by recombinant strains of Saccharomycescerevisiae on different carbon sources. Appl Microbiol

Biotechnol 52:829–833. doi:10.1007/s002530051599

Cellulose (2009) 16:729–741 741

123