

Yeast 14, 1267–1283 (1998)

Comparison of Expression Systems in the Yeasts

Saccharomyces cerevisiae, Hansenula polymorpha,

Klyveromyces lactis, Schizosaccharomyces pombe and

Yarrowia lipolytica. Cloning of Two Novel Promoters

from Yarrowia lipolytica

SVEN MU

} LLER

2

, THOMAS SANDAL

1

*, PETER KAMP-HANSEN

1

AND HENRIK DALBØGE

1

1

Microbial Discovery I, Novo Nordisk A/S, Novo Alle´, Building 1.B1.20, DK-2880, Bagsværd, Denmark

2

Department of Life Science and Chemistry, Roskilde University, Building 17.2, PO Box 260, DK-4000 Roskilde,

Denmark

We have compared expression systems based on autonomously replicating vectors in the yeasts Saccharomyces

cerevisiae, Schizosaccharomyces pombe, Kluyveromyces lactis, Hansenula polymorpha and Yarrowia lipolytica in order

to identify a more suitable host organism for use in the expression cloning method (Dalbøge and Heldt-Hansen,

1994) in which S. cerevisiae has traditionally been used. The capacity of the expression systems to secrete active

forms of six fungal genes encoding the enzymes galactanase, lipase, polygalacturonase, xylanase and two cellulases

was examined, as well as glycosylation pattern, plasmid stability and transformation frequency. All of the examined

alternative hosts were able to secrete more active enzyme than S. cerevisiae but the relative expression capacity of the

individual hosts varied significantly in a gene-dependent manner. One of the most attractive of the alternative host

organisms, Y. lipolytica, yielded an increase which ranged from 4·5 times to more than two orders of magnitude. As

the initially employed Y. lipolytica XPR2 promoter is unfit in the context of expression cloning, two novel promoter

sequences for highly expressed genes present in only one copy on the genome were isolated. Based on sequence

homology, the genes were identified as TEF, encoding translation elongation factor-1á and RPS7, encoding

ribosomal protein S7. Using the heterologous cellulase II (celII) and xylanase I (xylI) as reporter genes, the e

ffect of

the new promoters was measured in qualitative and quantitative assays. Based on the present tests of the new

promoters, Y. lipolytica appears as a highly attractive alternative to S. cerevisiae as a host organism for expression

cloning. GenBank Accession Numbers: TEF gene promoter sequence: AF054508; RPS7 gene promoter sequence:

AF054509.

1998 John Wiley & Sons, Ltd.

  — non-Saccharomyces yeasts; heterologous gene expression; autonomously replicating expression

vectors; selective promoter identification

INTRODUCTION
Isolation of enzyme genes from filamentous fungi

by use of the expression cloning method (Dalbøge

and Heldt-Hansen, 1994) has proved to be a very

strong alternative to traditional cloning methods

based on amino acid sequence information. So far,

approximately 200 genes have been cloned in our

laboratory, of which several have been character-

ized (Christgau et al., 1994, 1995, 1996; Kofod

et al., 1994; Draborg et al., 1995, 1996; Kauppinen

et al., 1995). Successful use of the method

demands: (i) e

ffective methods for generation of

full-length cDNAs; (ii) sensitive and reliable

enzyme assays; and (iii) a host organism which

e

fficiently secretes active forms of the cloned gene

products. The use of yeast as a host organism is

convenient because of its close evolutionary rela-

tionship with fungi. In addition, as a unicellular

microorganism, yeast o

ffers the advantages of

bacterial systems with regard to ease of manipula-

tion and growth conditions. So far, Saccharomyces

cerevisiae has been used as the host organism for

isolation of enzyme genes by expression cloning

but S. cerevisiae has been found to have certain

*Correspondence to: T. Sandal.

CCC 0749–503X/98/141267–17 $17.50

1998 John Wiley & Sons, Ltd.

Received 24 March 1998

Accepted 28 June 1998

limitations as a host for heterologous gene expres-

sion. In general, product yields are very low, with a

maximum of 1–5% of total protein, and plasmid

instability, hyperglycosylation and retention of

the proteins within the periplasmic space have

been observed (for references, see Buckholz

and Gleeson, 1991). Even though the use of S.

cerevisiae as a host organism has resulted in the

cloning of a large number of genes, the outlined

limitations of this host organism indicate that the

detection of some of the cloned heterologous

genes fails during the plate screening procedures.

Furthermore, the low expression level constitutes a

bottleneck for the expression cloning method, as it

necessitates reintroduction of the isolated genes

into a more e

ffective expression system prior to the

initial characterization of the products. Clearly,

it would be an improvement of the expression

cloning method if the host organism more e

ffi-

ciently secreted an active form of the heterologous

gene product. S. cerevisiae has been developed as a

production system for many di

fferent proteins (for

references, see Gellissen and Hollenberg, 1997) but

the increased product yield and improved quality

of the recombinant products have been achieved

by the introduction of tailor-made mutant strains

in which relevant steps of protein synthesis or

secretion have been altered. This product-specific

boosting of gene expression is not an option for

improvement of the expression cloning method in

which cDNAs from various donors encoding a

broad range of enzymes are introduced into the

host organism. Rather, it would be an improve-

ment of the expression cloning method to intro-

duce a more tolerant host organism which secretes

active forms of a broad range of heterologous gene

products more e

fficiently.

A growing number of non-Saccharomyces yeasts

have become accessible as expression systems

for heterologous gene products. The performance

of these alternative host organisms, including

Hansenula polymorpha, Klyveromyces lactis, Pichia

pastoris, Schizosaccharomyces pombe, Schwannio-

myces occidentalis and Yarrowia lipolytica, has

been reviewed in relation to S. cerevisiae (see

e.g. Buckholz and Gleeson, 1991; Gellissen and

Hollenberg, 1997) but the relative tolerance of

these organisms as hosts for expression of a broad

range of identical heterologous gene products has

not been tested and compared. In order to improve

the expression cloning method, we examined the

potential of the yeasts H. polymorpha, K. lactis, S.

pombe and Y. lipolytica, in which expression of

heterologous gene products is possible by use

of autonomously replicating expression vectors

(Gleeson et al., 1986; Bro¨ker and Ba¨uml, 1989;

Fleer et al., 1991; Fournier et al., 1991), with focus

on their ability to secrete active forms of six fungal

test gene products.

In this study we report that Yarrowia lipolytica

emerged as one of the most attractive alternative

host organisms. However, the homologue XPR2

promoter employed in the initial Y. lipolytica ex-

pression system is unsuitable in the context of

expression cloning, mainly because full induction

requires high levels of peptone in the culture

medium (Ogrydziak and Scharf, 1982). A search

for more attractive Y. lipolytica promoters was

therefore initiated. In the context of the expression

cloning method, an ideal promoter is characterized

by strength, activity in a medium suitable for

product recovery and inducibility. A strong pro-

moter is a necessary prerequisite for a high expres-

sion level and, with a very low copy number of the

centromere-linked ars18-based expression vectors

(Fournier et al., 1991; Vernis et al., 1997), this

character is even more important when Y. lipo-

lytica is used as the host organism. A tightly

regulated promoter makes it possible to separate

the growth stage from the expression stage,

thereby enabling expression of products which are

known to inhibit cell growth.

We report the cloning and the sequences of the

translation elongation factor-1á (TEF) and the

ribosomal protein S7 (RPS7) gene promoters from

Y. lipolytica. Based on tests of the novel promoters

with two heterologous reporter genes, Y. lipolytica

appears as a highly attractive alternative host

organism for use in expression cloning.

MATERIALS AND METHODS

Strains, media and growth conditions

Bacterial and yeast strains are described in

Table 1

. The various yeast transformants were

precultured in selective synthetic complete (SC)

media (Sherman, 1991) containing 2% glucose and

lacking either uracil or leucine. Of the non-

selective inducing media, 100 ml were inoculated

with precultures to an optical density at 600 nm

(OD

600

) of 0·1. Inducing media for the various

transformants were as follows: S. cerevisiae was

grown in 2

YP with 4% galactose; S. pombe and

K. lactis were grown in 2

YP with 4% glucose; H.

polymorpha was grown in a YP-derived medium

1268

. ¨

 

.

1998 John Wiley & Sons, Ltd.

Yeast 14, 1267–1283 (1998)

Table 1. Bacterial strains, yeast strains and expression vectors used in the study.

Bacterial strains

Source

MC1061

Meissner et al., 1987

DH10B

Gibco BRL

SJ2

A C600 derivate (Raleigh et al., 1988) modified at Novo Nordisk A/S

Yeast strains

Features

Source

Saccharomyces cerevisiae W 3124

Mat a ura 3-52 leu 2–3,112 his 3-Ä200

pep4-1137 Äprc1::HIS3 prb1::LEU 2 cir

+

J. Winther, Carlsberg Research Laboratory,

Department of Yeast Genetics, Denmark

Hansemula polymorpha A16

A derivate of the wild-type strain CBS4732

P. E. Sudbury, University of She

ffield, UK

Kluyveromyces lactis MW98-8C

Mat á uraA arg lys K

+

pKD1

0

M. Wesolowski-Louvel, Institut Curie,

Centre Universitaire, Orsay, France

Schizosaccharomyces pombe 972

h ura4-294

R. Egel, Department of Genetics,

University of Copenhagen, Denmark

Yarrowia lipolytica P

0

1d

A derivative of W29 Mat A: ura 3-302

leu 2-270 xpr 2-322

C. Gaillardin, Centre de Biotechnologies

Agro-Industrielles, Thiverval-Grignon, France

Expression

vectors

Host

Size

(kb)

Homologue

promoter

Replicative

element

Selection

marker

in yeast

Original source/

modifications

pYES 2·0

Saccharomyces cerevisiae

5·9 GAL1-inducible

2 ì

URA3 Invitrogen U.S.A.

pC4

Saccharomyces cerevisiae

5·9 GAL1-inducible

2 ì

URA3 Invitrogen U.S.A.

Introduction of an SfiI restriction site

pK2

Kluyveromyces lactis

7·0 LAC4-constitutive Structurally related to 2 ì URA3 pKD1 (Fleer et al., 1991)

pY1

Yarrowia lipolytica

7·8 XPR2-inducible

ars18

LEU2

pINA532 (C. Gaillardin)

pH2

Hansenula polymorpha

14·0 MOX-inducible

2 ì

LEU2

YEp13 (Gleeson et al., 1986)

pP3

Schizosaccharomyces pombe 7·1 ADH-constitutive

ars and stb

URA3 pMB332 (Bro¨ker and Ba¨uml, 1989)

pY3

Yarrowia lipolytica

7·9 XPR2-inducible

ars18

LEU2

As pY1, except no SfiI site is

present in the polylinker

1269










1998

John

Wiley

&

Sons,

L

td.

Yeast

14,
1267–1283

(1998)

(‘BMMY’, Pichia Expression Kit, from Invitro-

gen) containing 2% methanol; Y. lipolytica was

grown in 0·2% yeast extract, 0·1% glucose and 10%

proteose peptone in 50 m

 NaHPO

4

, pH 6·8 (a

modification of the YPDm medium; Nicaud et al.,

1989) when the test gene expression was based on

the XPR2 promoter. When Y. lipolytica test gene

expression was based on the novel promoters,

transformants were grown in SC-leu medium

containing 2% glucose. All transformants were

cultivated in 0·5 l shake flasks at 30

C. Solid media

(see

Figure 5

) were made with 1% agarose and

0·1% azurine dyed and cross-linked (AZCL) HE-

cellulose or Birch-xylan substrate (MEGAZYME).

Plasmids, test genes and transformation procedures

Expression vectors of the various yeasts are

described in

Table 1

. Furthermore, plasmid pYES

2·0 (Invitrogen) was used for cloning of Y. lipo-

lytica cDNAs, pSJ1678 (Novo Nordisk A/S) for

cloning of Y. lipolytica genomic DNA and pUC19

(Yanish-Perron et al., 1985) was used to enable

sequence determination of positives from the Y.

lipolytica genomic library originally cloned in

pSJ1678.

Test genes and donor organisms are described

in

Table 2

. The test genes were cloned in the

pC4, pK2, pY1, pH2 and pP3 expression vectors

(

Table 1

) as 5

–SfiI 3–NotI fragments. The celII

and xylI reporter genes were cloned in pY3 (and

modified editions of this vector including the

new Y. lipolytica promoters) as 5

–BamHI 3–NotI

fragments.

Plasmids were transformed into electrocompe-

tent E. coli strains MC1061 and SJ2, as described

(Meissner et al., 1987). Electrocompetent DH10B

cells were transformed as recommended by the

manufacturer. Transformation of S. cerevisiae, K.

lactis and Y. lipolytica with self-replicative expres-

sion vectors was carried out by electroporation

(Becker and Guarente, 1991). H. polymorpha was

transformed by use of the protoplast method

(Gleeson et al., 1986). Intact S. pombe cells were

transformed as described (Bro¨ker, 1987).

Measurement of enzymatic activity

Activity determination of Cellulase I, Cellulase

II, Galactanase I and Xylanase I was performed in

liquid assays. The degradation of AZCL substrates

(HE-cellulose, Arabinogalactan and Birch-xylan

(MEGAZYME)) was measured spectrophoto-

metrically at OD

620

and related to the correspond-

ing activity of the native enzyme. Supernatants

were incubated, shaking, at 40

C in: 0·04  citrate/

phosphate, pH 5·5 (Cellulase I); 0·04

 Tris,

pH 7·5 (Cellulase II); 0·04

 citrate/phosphate,

pH 4·5 (Galactanase I); or 0·04

 Tris, pH 7

(Xylanase I) in the presence of 0·16% substrate.

Polygalacturonase I and Lipase I activity was

measured on substrate-containing agarose plates

on which the area of the clearing zones was related

to a titration of the respective native enzyme.

Polygalacturonase I plates: 1% agarose, 0·1



citrate/phosphate, pH 4·5 and 1% Obipektin (DE

35%, NN Switzerland). Supernatants were in-

cubated for 24 h at 30

C prior to precipitation

with 1% mixed alkyltrimethylammoniumbromide

(MTAB) (Sigma

) solution at RT. Lipase I plates:

2% agarose, 0·1

 Tris, pH 9·5, 0·1  CaCl

2

, 0·5%

olive oil and 1% Triton X-100. Supernatants were

incubated for 24 h at 30

C. For each construction,

supernatants from two or three independently

inoculated cultures were assayed in duplicate. Re-

producibility was generally better than 8% be-

tween duplicates. Data from repeated experiments

generally varied less than 15%.

Electrophoresis and Western blotting

SDS–PAGE was performed essentially as

described by Laemmli (1970). Molecular mass

markers ranging from 6 to 98 kDa were used

(Pharmacia). 25 ìl supernatant aliquots of the

samples used for measuring of enzymatic activity

were loaded.

SDS–PAGE gels were blotted to polyvinylidine

difluoride membranes (‘Immobilon’, Millipore

Corp.) by semi-dry electrobotting (Kyhse-

Andersen, 1984). Unreacted binding sites were

blocked by incubation for 30 min at RT in 0·05



Tris, pH 7·5, 0·15

 NaCl, 0·1% Tween 20 contain-

ing 2% bovine serum albumin. For immunoblot-

ting, the membranes were incubated for 60 min at

Table 2. Test genes used.

Test genes

Origin

Cellulase I

Aspergillus aculeatus

Cellulase II

Humicola insolens

Galactanase I

Aspergillus aculeatus

Xylanase I

Humicola insolens

Polygalacturonase I

Aspergillus aculeatus

Lipase I

Thermomyces lanuginosus

1270

. ¨

 

.

1998 John Wiley & Sons, Ltd.

Yeast 14, 1267–1283 (1998)

RT with a polyclonal rabbit antiserum raised

against the tested enzymes diluted 1:1000 in 0·05



Tris, pH 7·5, 0·15

 NaCl, 0·1% Tween 20. Bound

rabbit antibodies were visualized by incubation

for 60 min at RT with alkaline phosphatase

conjugated goat anti-rabbit antibodies (DAKO

A/S, Denmark), diluted in 0·05

 Tris, pH 7·5,

0·15

 NaCl, 0·1% Tween 20 to 0·6 ìg/ml and

development in 10 ml ALP bu

ffer (150 m Tris–

Cl, pH 9·5, 100 m

 NaCl, 5 m MgCl

2

), 66 ìl

NBT (50 mg/ml nitro blue tetrazolium in 70%

dimethylformamide) and 33 ìl BCIP (50 mg/ml

5-bromo-4-chloro-3-indolyl phosphate in 70%

dimethylformamide).

RNA for cDNA libraries and Northern blot

analysis

Y. lipolytica was grown in YP medium contain-

ing 2% glucose or 2% glycerol at 30

C. Cells were

harvested late in the logarithmic phase at an

optical density (OD

600

) of 5·5, frozen in liquid

nitrogen and powdered. Total RNA was iso-

lated by the guanidium thiocyanate method fol-

lowed by ultracentrifugation through a 5·7



CsCl cushion (Chirgwin et al., 1979). Poly(A

+

)

RNA

was

isolated

by

oligo(dT)-cellulose

a

ffinity chromatography (Aviv and Leder,

1972; Sambrook et al., 1989). Double-stranded

cDNA was synthesized as described below. For

Northern blot analysis the poly(A

+

) RNAs

(2·5 ìg/sample) were electrophoresed in 1%

agarose, 2·2

 formaldehyde gels (Thomas, 1983)

and blotted to a nylon membrane (Hybond-N,

Amersham Corp.) with 10

SSC as transfer

bu

ffer. PCR-generated copies of the L1·41, L1·45,

L2·7 or L2·17 cDNAs were used as probes.

100 ng of DNA (isolated using Qiagen) of the

respective cDNAs were used as templates with

50 pmol of pYES 2·0 specific forward and reverse

primers, a DNA thermal cycler and 2·5 units of

Taq polymerase (Perkin–Elmer). Thirty cycles

of PCR were performed using a cycle profile of

denaturation at 94

C for 30 s, annealing at 55C

for 30 s, and extension at 72

C for 1 min. The

PCR products were purified with QIAquick PCR

Purification Kit (Qiagen). The probes were

labelled with

32

P (>1

10

9

cpm/ìg), hybridized

to the membrane and washed as described

(Kauppinen et al., 1995). After autoradiography

at

80C, the L1·41 probe was removed from

the membrane according to the manufacturer’s

instructions, and the filter was rehybridized

sequentially to the remaining probes.

Construction of directional Y. lipolytica cDNA

libraries from YP glucose and YP glycerol cultures

Double-stranded cDNA was synthesized from

5 ìg Y. lipolytica poly(A)

+

RNA from each of the

YP glucose and YP glycerol cultures, as described

(Gubler and Ho

ffman, 1983; Sambrook et al.,

1989) except that 1·5 ìg of oligo(dT

18

) NotI primer

(Pharmacia) was used in the first strand reaction.

After synthesis, the cDNA was ligated to non-

palindromic BstXI adaptors (Invitrogen), using a

50-fold molar excess of the adaptors. The adapted

cDNAs were digested with NotI, size fractionated

by agarose gel electrophoresis, ligated into BstXI/

NotI cleaved pYES 2·0 vector (Invitrogen), and the

ligation mixtures were transformed into electro-

competent E. coli DH10B cells (Gibco BRL). The

libraries, consisting of 2

10

6

(L1) and 3

10

5

(L2) clones (glucose- and glycerol-based cultures,

respectively), were stored in 20% glycerol at

80C.

Southern blot analysis

Y. lipolytica total DNA was prepared from

20 ml YP glucose overnight cultures. Cells were

resuspended in 400 ìl 0·9

 Sorbitol, 0·1  EDTA

pH 7·5, 14 m

 â-mercaptoethanol and incu-

bated for 30 min at 37

C with 100 ìl Novozyme

(2 mg/ml) (Novo Nordisk A/S). Spheroplasts were

resuspended in 400 ìl TE and 5 ìl 10

RNase

A+T added, 90 ìl fresh-made 280 m

 EDTA

pH 8·0, 444 m

 Tris, and 2·2% SDS, and incu-

bated for 30 min at 65

C. The suspension was

added to 80 ìl 5

 potassium acetate, mixed and

cooled on ice for 30 min. The supernatant was

extracted with phenol/chloroform at 65

C and the

DNA was precipitated with ethanol.

The Y. lipolytica total DNA was digested to

completion with BamHI, EcoRI, HindIII or KpnI

(10 ìg/sample), fractionated on a 1% agarose gel,

denatured, and transferred to a nylon filter

(Hybond-N, Amersham) using 10

SSC as trans-

fer bu

ffer (Southern, 1975). PCR copies of the

L1·41, L1·45, L2·7 and L2·17 cDNAs were used

as probes.

32

P-labelling, hybridization, washing,

autoradiography and removal of the probe from

the membrane prior to rehybridization were

carried out as described for Northern blot analysis.

Construction and screening of a Y. lipolytica

genomic library using PCR generated probes

Total DNA was prepared as described for

Southern blot analysis. Partial Sau3AI digested

1271

    

1998 John Wiley & Sons, Ltd.

Yeast 14, 1267–1283 (1998)

DNA was fractionated on a 1% low melting point

agarose gel. Fragments of 1–2 kb were recovered

from the gel, ligated into the BamHI site of

pSJ1678 and transformed into E. coli SJ2. A

genomic library of >45,000 clones was obtained.

Sau3AI sites were shown to be present internally

in the L1·41 and L2·7 cDNAs (data not shown). In

order to increase the probability of identifying

sequences located upstream of the coding part of

the corresponding genomic sequences, 400 bp PCR

generated copies of the 5

end of the L1·41 and

L2·7 cDNAs were used as probes for screening

of the genomic library. PCR was performed as

described for RNA gel blot analysis. L1·41 cDNA

specific primers: forward 5

-CCTTCTGAGTATA

AGAATC-3

, reverse 5-GATACCAGCCTCGA

ACTC-3

. L2·7 cDNA specific primers: forward

5

-GATCGCAGCTCTCTCCCAC-3, reverse 5-

GACCTTGATGAGAGGGTC-3

.

The Y. lipolytica genomic library was screened

by colony hybridization (Sambrook et al., 1989)

using random-primed (Feinberg and Vogelstein,

1983)

32

P-labelled (>1

10

9

cpm/ìg) PCR prod-

ucts for L1·41 and L2·7 as probes.

32

P-labelling

and hybridization as described (Kauppinen et al.,

1995).

Cloning of PCR generated EF-1á and RPS7

promoter sequences in Y. lipolytica expression

vectors

PCR of the EF-1á and RPS7 genomic sequences

was performed using conditions described for

Northern blot analysis except that the annealing

temperature was raised to 60

C. EF-1á genomic

sequence specific primers: forward 5

-CCATCGA

TAGAGACCGGGTTGGCG-TATTTGTGTCC-

3

, reverse 5-CGCGGATCCTTCGGGTGTGG

AGTTGACAAGG-3

. RPS7 genomic sequence

specific primers: forward 5

-CCATCGATTACCT

GCTACTTG-TCTCAACACC-3

, reverse 5-CGC

GATCCTTGTGTTTGTTGAGTGAAGAAAAG-

ATTTGG-3

. For both promoter sequences the 3

terminal nucleotide was defined to the last nucleo-

tide in the genomic sequence that was not repre-

sented in the corresponding cDNA sequence 5

end. The PCR-generated promoter sequences were

digested with BamHI and ClaI and cloned as

5

–ClaI 3–BamHI fragments in pY3 derivates

(

Table 1

) from which the XPR2 promoter

sequence was deleted. The cloning sites were intro-

duced into the promoter sequences by the syn-

thetic oligonucleotides used in the PCR reaction.

The PCR-generated promoter sequences were

confirmed by DNA sequencing.

Nucleotide sequence analysis of the Y. lipolytica

cDNAs and genomic sequences

All nucleotide sequences were determined by the

dideoxy chain-termination method (Sanger et al.,

1977) using 500 ng of Qiagen-purified template

(Qiagen), ABI PRISM

dye terminator cycle

sequencing ready reaction kit with Amplitag

DNA polymerase, FS (Perkin–Elmer) and 5 pmol

of either pYES 2·0 polylinker primers (Invitrogen)

or synthetic oligonucleotide primers. Reactions

were analysed on ABI PRISM 377 according to

manufacturers’ instructions.

Two sets of 100 cDNA 5

tags from subpools

of each Y. lipolytica library were sequenced on

one strand using the pYES 2·0 polylinker primer

(forward). The complete sequence of the selected

L1·41 and L2·17 cDNAs were determined on both

strands using the pYES 2·0 polylinker primers and

synthetic oligonucleotide primers. Sequence com-

parisons were carried out using FASTA searches

(Pearson and Lipman, 1988) on the GenEMBL

database.

Insert DNA from positive clones from the

Y. lipolytica genomic library was subcloned

into pUC19 (Yanisch-Perron et al., 1985) and

sequenced on both strands using synthetic oligo-

nucleotide primers.

Nucleotide sequence accession numbers

The TEF and RPS7 gene promoter sequences

presented in this paper, the TEF and RPS7 gene

cDNA sequences, and the cDNA sequences of the

fungal test genes have been submitted to GenBank

and assigned the following Accession Num-

bers: Genomic sequences: TEF gene promoter,

AF054508; RPS7 gene promoter, AF054509.

cDNA sequences: TEF, AF054510; RPS7,

AF054511; Cellulase I, AF054512; Cellulase II,

A21793; Galactanase I, L34599; Xylanase I,

X76047; Polygalacturonase I, AF054893; Lipase I,

AF054513.

RESULTS AND DISCUSSION

Criteria for evaluation of the alternative yeasts

To examine the potential of the four alternative

yeasts as host organisms for expression cloning, six

test genes donated by three filamentous fungal

1272

. ¨

 

.

1998 John Wiley & Sons, Ltd.

Yeast 14, 1267–1283 (1998)

strains (

Table 2

) were introduced in autonomously

replicating expression vectors under the control of

strong homologous promoters (

Table 1

). The yeast

transformants were tested in various rich and

minimal media containing di

fferent carbon sources

in order to maximize the secretion of active en-

zyme (data not shown). The optimal media found

in these preliminary experiments are described

in Materials and Methods. The following were

examined for:
1. The amount of active enzyme secreted. This is

the most important of the listed parameters in

the search for an attractive expression cloning

host organism. A high expression level of active

enzyme will increase the probability of identify-

ing cloned genes by plate screenings. Further,

the e

fficiency of the expression cloning system

would be significantly improved if the organism

used for the initial cloning of new enzyme

genes subsequently was able to express the

gene products in su

fficient amounts to enable

characterization.

2. Glycosylation pattern. This was monitored by

Western blots. This analysis made it possible to

observe posttranslational modifications in the

various hosts compared to the native enzymes.

The blots could also reveal whether the

specific activity of the enzymes was significantly

altered by comparing the amount of enzyme on

the blot with the measured enzyme activity

(

Figures 1A

,

2

).

3. Transformation frequency and plasmid stability.

In the context of expression cloning, the trans-

formation frequency of the host organism

relates to the necessity of screening standard

cDNA libraries. When introduced in the hosts,

the plasmids have to be su

fficiently stable to

allow identification of positives, preliminary

production and, eventually, plasmid rescue.

Initial results demonstrated that variation in the

5

non-translated sequence had a significant effect

on the expression level in all the host systems. The

presence of an SfiI restriction site reduced the

expression level up to threefold (data not shown).

This observation could be explained by the poten-

tial for a loop structure formation in the GC-rich

SfiI sequence. In the attempt to level the e

ffect of

variation in the 5

non-translated sequence in the

various host systems, an SfiI site (N-terminal) and

a NotI site (C-terminal) was introduced in all the

test genes. This enabled swapping of the test genes

between the expression vectors, thus maintain-

ing identical 5

non-translated sequences in the

di

fferent yeasts.

In an experimental setup such as the one

described here, it is di

fficult to compare the differ-

ent systems because of factors such as the relative

strength of the chosen promoters, copy number of

the individual vectors and use of individual growth

media, but the systems are standardized as much

as possible and, in case significant variation in the

capacity of the various hosts was observed we

would be able to select candidates for further

optimization.

Enzyme activity in supernatants

Samples were taken from the cultures three

times during the logarithmic growth phase and

three times during the stationary growth phase.

The substrate degradation capacity present in

the supernatants was measured and related to the

degradation capacity of known amounts of the

corresponding native enzymes as described in

Materials and Methods. The enzyme activity

values presented in

Figure 1

arise from the samples

where the average activity per volume supernatant

was highest. For all transformants, maximum

activity was observed in one of the samples from

the stationary growth phase corresponding to

60–90 h of growth. The enzyme activity values are

presented both as activity per volume of super-

natant (

Figure 1A

) and as activity per cell (

Figure

1B

). In the experiments described, the cells were

grown in conical flasks where a significant strain-

dependent variation in cell density of the cultures

was observed, therefore actvity per cell is a more

accurate measurement for comparative purposes.

Plasmid loss has not been considered in the values

presented in

Figure 1

.

A comparison of the enzyme activities shows

that all the alternative yeast strains secrete active

forms of the test gene products at a significantly

higher level than S. cerevisiae. This result is not

surprising, considering the fact that S. cerevisiae

has been used for thousands of years by mankind

in brewing and baking at conditions where the

ability to secrete substrate degrading enzymes to

the environment has been needless. The activity

data further shows that the expression e

fficiency of

the di

fferent yeasts varies significantly in a gene-

dependent way and is independent of the donor

organism (

Figures 1A, 1B

;

Table 2

).

In general, H. polymorpha, S. pombe and Y.

lipolytica secrete most active enzyme/ml (

Figure

1273

    

1998 John Wiley & Sons, Ltd.

Yeast 14, 1267–1283 (1998)

Figure 1. Relative enzyme activity per volume (A), and the actvity divided by the OD

600

value of the respective

yeast culture (B). Samples for measurement of enzyme activity in culture supernatants shown here were taken at the

time of maximal activity per volume. For all transformants this maximum was reached in the stationary growth

phase. The highest activity obtained per volume (A) or per OD

600

(B) for the respective enzymes was set to 100.

Cel I, Cellulase I; Cel II, Cellulase II; Gal I, Galactanase I; Xyl I, Xylanase I; PG I, Polygalacturonase I; Lip I,

Lipase I.

1274

. ¨

 

.

1998 John Wiley & Sons, Ltd.

Yeast 14, 1267–1283 (1998)

1A

), but in the case of H. polymorpha this is due to

a high cell density (

Figure 1B

).

Western blot analysis, transformation frequency

and plasmid stability

Western blots were carried out to monitor

glycosylation. Examples of Galactanase I (A) and

Lipase I (B) secretion are given in

Figure 2

.

According to the Western blots, only H. poly-

morpha over-glycosylated significantly. As this

observation has not been reported elsewhere, it

might be a strain-dependent phenomenon. The

over-glycosylation seems to decrease the specific

actvity of Lipase I, as seen when the activity per

volume is compared to the amount of secreted

enzyme (

Figures 1A

,

2B

). Over-glycosylation

does not seem to influence the specific activity of

Galactanase I (

Figures 1A

,

2A

). The Western

blot also reveals that Galactanase I is exposed

to a minor processing when secreted from Y.

lipolytica.

In the preliminary examination of all the alter-

native yeasts, the lowest acceptable transformation

frequency was defined as 1

10

3

cfu/ìg plasmid

DNA. With all the yeasts we could obtain trans-

formation frequencies >1

10

3

, except for S.

pombe.

For all yeast expression systems, enzyme activity

levels increased continually over 72 h in non-

selective medium, indicating that plasmid loss was

not a significant problem.

Evaluation of the alternative host systems

The advantages and disadvantages of the tested

host systems concerning parameters other than the

expression capacity are shown in

Table 3

.

Based on the high capacity for secretion of

active forms of the test gene products H. poly-

morpha, S. pombe and Y. lipolytica appear to be

the most promising alternatives to S. cerevisiae as

host organisms for expression cloning. However,

there are also disadvantages, such as over-

Figure 2. Western blots of galactanase (A) and lipase (B) cultures. 25 ìl supernatant aliquots from the samples used for

measurement of enzymatic activity (

Figure 1

) was loaded in each lane. S.c.=S. cerevisiae; S.p.=S. pombe; K.l.=K. lactis;

H.p.=H. polymorpha; Y.l.=Y. lipolytica; M=molecular markers; S.t.=100 ng of the native enzyme; + Indicates the presence of

a test gene in the expression vector;

Indicates the absence of an insert in the expression vector. In the case of lipase I (B), the

upper band is N-glycosylated and the lower is not (Boel and Huge-Jensen, 1989).

Table 3. Advantages and disadvantages of the yeasts concerning expression cloning. Transformation frequency

and growth are illustrated with +–+ + + + +, where + is relative minimum and + + + + + relative maximum.

S. cerevisiae

S. pombe

K. lactis

H. polymorpha

Y. lipolytica

Transformation frequency

+ + + + +

+

+ + + +

+ + +

+ + +

Glycosylation problems

Apparently none Apparently none Apparently none

Up to

20 kDa increase

Apparently none

Plasmid stability

Satisfactory

Satisfactory

Satisfactory

Satisfactory

Satisfactory

Growth

+ + +

+

+ + +

+ + + + +

+ + +

1275

    

1998 John Wiley & Sons, Ltd.

Yeast 14, 1267–1283 (1998)

glycosylation of the heterologous products in H.

polymorpha (this was observed for all test gene

products except for Xylanase I) and di

fficulty in

transforming S. pombe.

Complications in the use of Y. lipolytica was

neither related to the character of the secretory

system or handling of the cells, but seemed to be

restricted to the peptone induced XPR2 promoter

employed in the expression vector (see below).

Although the XPR2-based expression system is

unfit in the context of expression cloning, it re-

vealed a high potential of Y. lipolytica as an

alternative to S. cerevisiae, yielding an increase in

expression of the tested gene products of 4·5, 4·8

and 40 times (Polygalacturonase I, Cellulase II

and Galactanase I, respectively) and, in case of

Cellulase I, Xylanase I and Lipase I, an increase of

more than two orders of magnitude.

As the XPR2 promoter revealed a high potential

of Y. lipolytica as an alternative candidate, we

decided to optimize the Y. lipolytica expression

system by identification and integration of more

suitable homologous promoters.

Identification of suitable Y. lipolytica promoters

In the context of the expression cloning method,

optimization of the expression vector, by replace-

ment of the employed XPR2 promoter, was neces-

sary for several reasons: the XPR2 promoter is

only active at pH>6·0 on media lacking preferred

carbon and nitrogen sources and full induction

requires high levels of peptone in the culture

medium (Ogrydziak et al., 1977; Ogrydziak and

Scharf, 1982). The presence of peptone in the

medium complicates product recovery and puri-

fication and hinders the direct screening for

transformants based on LEU2 selection.

In the attempt to identify new strong promoters,

we constructed cDNA libraries from cultures

grown under conditions suitable for use in expres-

sion cloning. Determination of sequence tags

facilitated identification of highly expressed genes

and the isolation of the corresponding promoter

sequences from a genomic Y. lipolytica library.

This strategy was based on the assumption that a

high level of a specific mRNA, under conditions

suitable for expression cloning, reflects the pres-

ence of a strong promoter which is active under

these conditions. cDNA libraries were constructed

from YP-glucose and -glycerol cultures. By

examination of the gene activity in two di

fferent

assimilable media we hoped to identify not only

strong but also inducible (e.g. catabolite-repressed)

promoters.

Initial sequence determination was performed

on 100 clones from each cDNA library in which

300–600 nucleotides of the 5

end of the inserts

were determined. The sequence data from each

library were aligned against each other. In the

following, the cDNA library from the YP-glucose

culture is referred to as L1 and the library from the

YP-glycerol culture as L2. The number following

L1 or L2 refers to a specific clone in the library.

The sequence alignment revealed that six di

fferent

sequences from L1 were represented twice and two

di

fferent sequences were represented as triplets.

Twelve di

fferent sequences from L2 were repre-

sented twice. Alignment of L1 and L2 showed that

several sequences from one library also were rep-

resented in the other. Four sequences were chosen

for further examination: one representing a se-

quence observed as a triplet in L1 and twice in L2

(L1·41), one representing a sequence observed as

a triplet in L1 but not observed in L2 (L1·45),

one representing a sequence observed twice in L1

and L2 (L2·7), and one representing a sequence

observed twice in L2 and not in L1 (L2·17).

The detection of a sequence in two or three

copies in only one of the cDNA libraries could

indicate that di

fferent promoter activity was

present in the YP-glucose and -glycerol media. To

test this, a Northern blot analysis was performed.

Equal intensities of signals using poly(A

+

) RNA

from both glucose and glycerol grown cultures

revealed that these genes were not under control of

a glucose-repressed promoter (data not shown).

In order to examine whether the high frequency

of the selected cDNAs was due to a high copy

Figure 3. Examination of copy number by Southern blot

analysis. Y. lipolytica total DNA was cut to completion with

BamHI (B), EcoRI (E), HindIII (H) or KpnI (K). M is a marker

and -E is a DNase contamination control in which no enzyme

was present during the incubation. PCR-generated copies of the

selected cDNAs were used as probes: A=L1·41; B=L1·45;

C=L2·7; D=L2·17.

1276

. ¨

 

.

1998 John Wiley & Sons, Ltd.

Yeast 14, 1267–1283 (1998)

number of the corresponding genes, a Southern

blot analysis in combination with a restriction

analysis was carried out (

Figure 3

). Probes based

on the L1·41 (A), L1·45 (B) or L2·17 (D) sequences

only hybridize at one or two areas on the mem-

brane and in the case of two signals, this is due to

the presence of the respective restriction enzyme

recognition sequences in the structural genes. This

strongly indicates that the P

0

1d genome contains

only one copy of each of L1·41, L1·45 and L2·17.

The L2·7-based probe hybridizes at several distinct

areas on the membrane. Digest of a PCR copy of

the L2·7 sequence with the employed enzymes did

not reveal an internal presence of these sites (data

not shown). This shows that the L2·7 sequence

is present in several copies of the P

0

1d genome.

For this reason no attempt was made to

identify and test the promoter matching the L2·7

sequence.

Cloning of Y. lipolytica TEF and RPS7 promoter

sequences

To identify the promoters controlling the

strongly expressed transcripts, a Y. lipolytica

genomic library was screened using the first 400 bp

of the genes as probes. Colony hybridization

resulted in the isolation of two genomic sequences

matching the L1·41 and L2·17 cDNAs, respect-

ively. The 1·5 kb L1·41 and 0·9 kb L2·17 cDNAs

were sequenced to completion and aligned by

FastA searches (Pearson and Lipman, 1988) to the

GenEMBL database. The L1·41 cDNA sequence

showed significant homology to the translation

elongation factor-1á (TEF) gene of various

sources, e.g. Arxula adeninivorans, Neurospora

crassa and Saccharomyces cerevisiae. GAP align-

ments (Needleman and Wunch, 1970) to the TEF

gene sequences of the yeasts A. adeninivorans

(accession no. Z47379) and S. cerevisiae (accession

no. X00779) showed 83·8% and 76·4% identity,

respectively. The L2·17 cDNA sequence showed

significant homology to the ribosomal protein S7

(RPS7) gene of e.g. S. cerevisiae and the corre-

sponding RPS4 gene of e.g. Drosophila mela-

nogaster and Homo sapiens. GAP alignments

showed 69·2% identity to the RPS4 gene of D.

melanogaster (accession no. D16257) and 68·5%

identity to exon 1 and 2 of the S. cerevisiae

(accession no. M64293) RPS7 gene. Both of the

identified genes encode proteins involved in trans-

lation. The elongation factor-1á plays an essential

role in protein synthesis in eukaryotic cells by

binding the amino-acyl tRNA to the ribosomes in

exchange for the hydrolysis of GTP. The ribo-

somal protein S7 is the largest protein of the 40 S

subunit and is essential for growth (Synetos et al.,

1992). Identification of the TEF gene by the

present approach is not that surprising, as its

product is an exceptionally abundant protein com-

prising 3–10% of the soluble protein in most cells

(Cavallius et al., 1993). Ribosomal proteins are

also abundant, each representing 0·1–0·5% of the

total cellular protein, and their mRNA are corre-

spondingly abundant (reviewed by Warner, 1989).

As both genes were found to be present in only one

copy on the genome, we expected the matching

promoters, especially the TEF gene promoter, to

be strong. This has been shown to be the case for

the TEF gene promoter of several other organisms

(Schirmaier and Philippsen, 1984; Axelos et al.,

1989; Kim et al., 1990; Ursin et al., 1991).

The L1·41 related genomic sequence was deter-

mined and alignment with the matching cDNA

revealed a 366 bp sequence in the genomic DNA

located upstream of the 5

end of the cDNA

(

Figure 4A

). Similarly, a 758 bp genomic DNA

sequence was present upstream of the matching

L2·17 cDNA 5

end (

Figure 4B

). Visual inspection

of the putative promoter sequences indicates

the presence of the cis-acting regulatory element

UAS

rpg

, which is found in the promoter region of

several genes encoding highly expressed ribosomal

and house-keeping genes (Leer et al., 1985; Mager,

1988). The 3

terminal nucleotide of the new

promoter sequences cloned in expression vectors

(see below) was defined to be equal to the last

nucleotide in the genomic sequence that was not

represented in the corresponding cDNA.

Examination of the e

ffect of the novel

Y. lipolytica promoters

In order to test the e

ffect of the Y. lipolytica

TEF- and RPS7 gene promoters, PCR generated

copies were cloned into pY3 derivates from which

the XPR2 promoter was removed and the hetero-

logous cellulase II or xylanase I genes (

Table 2

)

were present as reporters. The reporter gene

expression level of Y. lipolytica transformed with

these constructs was compared to the correspond-

ing expression level in cells transformed with the

original XPR2 containing pY3. Transformants

containing the new promoters were grown in

media in which glucose was present as the carbo-

hydrate source.

1277

    

1998 John Wiley & Sons, Ltd.

Yeast 14, 1267–1283 (1998)

To examine the potential of the new promoters

as screening tools, the transformants were inocu-

lated on to substrate-containing growth plates

(

Figure 5

). TEF and RPS7 gene promoter trans-

formants were grown on selective inducing SC-

leu+glucose plates, whereas XPR2 transformants

were grown on non-selective, inducing peptone-

containing plates (see Materials and Methods).

The substrate degradation e

ffected by the Y. lipo-

lytica transformants was compared with the

corresponding degradation caused by S. cerevisiae

transformants inoculated on to selective inducing

SC-ura+galactose plates. The plate assay does not

reflect the expression level of the individual trans-

formants as the amount of cells varies but it

mimics an expression cloning screening event.

Figure 5

shows that the TEF gene promoter

very e

ffectively causes substrate degradation

(A, pY5TACII; D, pY5TAXI) in both the HE-

cellulose and Birch-xylan enzyme assays. As a

screening tool the TEF gene promoter appears

more e

ffective than the XPR2 promoter (B and E)

as well as the original expression cloning system

based on S. cerevisiae (C and F). Furthermore, the

plate assay demonstrates that both the new pro-

moters, in contrast to the XPR2 promoter,

are active under selective conditions and thereby

enable the use of Y. lipolytica as a host for direct

screening of cDNA libraries.

The e

ffect of the new promoters was also com-

pared to the e

ffect of the XPR2 promoter in a

quantitative analysis by measuring the enzyme

activity in liquid culture supernatants. As in the

plate assay, Y. lipolytica transformed with the new

promoter constructs were grown in selective induc-

ing media, whereas the XPR2 promoter transform-

ants were grown in non-selective inducing medium.

The relative enzyme activity values shown in

Figure 4. (A) Nucleotide sequence of the EF-1á promoter. The positions are related to the A

in the ATG start codon (bold) defined as +1. The putative UAS

rpg

boxes HOMOL1 (positions

191 to 180) and RPG (179 to 168) and the T-rich sequence are underlined. The

putative TATA box (

111 to 106) and a pyrimidine-rich sequence (85 to 58) are

double underlined. The putative transcription initiation site (

56 to 53) is written in bold.

Nucleotides located from position

40 and downstream were also present in the cDNA

sequence. (B) Nucleotide sequence of the RPS7 promoter. The positions are related to the A in

the ATG start codon (bold) defined as +1. The putative UAS

rpg

boxes HOMOL1 (positions

273 to 262) and RPG (247 to 236) and the T-rich sequence (present on the opposite

strand) are underlined. Putative TATA boxes (

201 to 190), a TATA-like sequence (46

to

41) and transcription initiation consensus sequences (85, 55, 15 and 13) are

double underlined. Nucleotides located from position

2 were also present in the cDNA

sequence.

1278

. ¨

 

.

1998 John Wiley & Sons, Ltd.

Yeast 14, 1267–1283 (1998)

Figure 6

have not been corrected for the di

fference

in percentage loss of plasmid. As the maximum cell

density was approximately the same in all cultures,

only the relative activity per volume is illustrated.

As in the initial study of the alternative host

systems, the activity values are from samples in the

stationary growth phase where the activity per

volume supernatant was highest. It should be

noted that the samples from the TEF and RPS7

gene promoter cultures do not necessarily repre-

sent the highest activity level possible, as no stag-

nation in the activity level was observed among the

stationary growth phase samples taken (data not

shown). The samples from the TEF and RPS7

gene promoter cultures presented in

Figure 6

were

taken after 80 h of cultivation. The continued

increase in enzyme activity is not surprising con-

sidering the need of the TEF and RPS7 gene

products in the maintenance of a cell. Ro¨sel and

Kunze (1995) compared the TEF gene mRNA

levels at di

fferent times during yeast growth of the

dimorphic yeast Arxula adeninivorans, in which the

TEF gene also seems to be present in only one

copy on the genome, in a Northern blot analysis.

As in the present growth experiment, glucose was

used as the carbon source. The transcript level

decreased when the culture entered the stationary

growth phase but then stabilized, and no di

fference

was observed among the last-taken 60 and 75 h

samples.

In case of the reporter genes tested here the TEF

gene promoter appears as the most striking of the

novel promoters. In the Cellulase II samples com-

pared, the activity level of the enzyme under

control of the TEF gene promoter almost equals

that of the XPR2 promoter. The e

ffect of the RPS7

gene promoter is roughly 30% less than the e

ffects

of the TEF gene promoter. Interestingly, use of

the xylanase I reporter gene seems to a

ffect the

tested promoters very di

fferently. As observed for

Figure 5. The e

ffect of the Y. lipolytica EF-1á (pY5TACII/XI) and RPS7 (pY5RBCII/XI) promoters (A and D), as indicated

by use of plate assays. The novel promoters were introduced in pY3 derivatives in which the XPR2 promoter was absent and with

cellulase II or xylanase I as reporter genes. Positive cellulase II (A) and xylanase I (D) transformants were reinoculated on to

SC-leu plates containing 2% glucose, 0·1% AZCL HE-cellulose (A) or AZCL Birch-xylan (D) substrate. CV=control vectors, in

which no promoter sequence was present. Y. lipolytica XPR2 (pY3CII/XI) promoter-based expression of Cellulase II (B) or

Xylanase I (E). The cellulase II or xylanase I test genes were cloned in pY3 and positive transformants were reinoculated on to

inducing non-selective plates (0·2% yeast extract, 0·1% glucose and 10% proteose peptone in 50 m

 NaHPO

4

, pH 6·8) containing

the relevant AZCL substrate. Cellulase II (C) and xylanase I (F) expression from the original expression cloning host S.

cerevisiae. The test genes were cloned in pYES 2·0 downstream of the Gal1 promoter and positive transformants were restriked

on SC-ura plates containing 2% galactose and the relevant AZCL substrate. Cellulase II transformants (A, B and C) were

incubated at 30

C for 48 h. Xylanase I transformants (D, E and F) were incubated at 30C for 24 h.

1279

    

1998 John Wiley & Sons, Ltd.

Yeast 14, 1267–1283 (1998)

Cellulase II expression, the enzyme activity is

higher in the presence of the TEF gene promoter

than in the presence of the RPS7 gene promoter,

but in the case of Xylanase I activity the TEF gene

promoter appears four times more e

ffective.

Whereas the e

ffect of the TEF gene promoter in the

case of Cellulase II activity almost equals the e

ffect

of the XPR2 promoter, it only corresponds to

about 50% of the XPR2 promoter e

ffect in the case

of Xylanase I activity. The Western blot analysis

(

Figure 7

) shows a decrease in the amount of

secreted product as reflected in the enzyme

activities shown in

Figure 6

.

The tests of the new promoters on substrate-

containing plates (

Figure 5

) indicated that the new

promoters were significantly more e

ffective than

the XPR2 promoter. The activity assay on the

supernatants (

Figure 6

) showed that this was not

the case when the cells were grown in liquid

medium. A possible explanation of these conflict-

ing observations could be that the environmental

demands for full induction of the XPR2 promoter

(e.g. pH 6·8) are not maintained in the growth

plates—this would also explain why pYES 2·0 in

S. cerevisiae appears as a more e

ffective ex-

pression system than XPR2-based expression in

Figure 6. Maximum Cellulase II and Xylanase I activity in

supernatants from Y. lipolytica cultures in the presence of the

XPR2, EF-1á or RPS7 promoters. The highest activity ob-

tained for the respective enzymes was set to 100%. In case of

Cellulase II expression, 100% correspond to the activity of

7·6 mg/l native enzyme. 100% Xylanase I activity corresponds

to the activity of 1·9 mg/l native enzyme.

Figure 7. Western blot of Y. lipolytica Cellulase II (A) and Xylanase I (B) cultures based on the new promoters and XPR2.

(A) Lane 1=0·3 ìg, 2=0·15 ìg, 3=0·075 ìg native Cellulase II. Lanes 4, 7 and 8 are empty. Lane 5 is loaded with 25 ìl

supernatant from TEF gene promoter transformants, which correspond to the activity of 0·19 ìg native enzyme. Lane 6 is

loaded with 25 ìl supernatant from RPS7 gene promoter transformants, which correspond to the activity of 0·1 ìg native

enzyme. Lane 9 is loaded with 25 ìl supernatant from XPR2 transformants, which correspond to the activity of 0·19 ìg native

enzyme. (B) Lane 1=0·0375 ìg, 2=0·0075 ìg, 3=0·00375 ìg native Xylanase I. Lanes 4, 7 and 8 are empty. Lane 5 is loaded

with 25 ìl supernatant from TEF gene promoter transformants, which corresponds to the activity of 0·0123 ìg native enzyme.

Lane 6 is loaded with 25 ìl supernatant from RPS7 gene promoter transformants, which correspond to the activity of

0·0048 ìg native enzyme. Lane 9 is loaded with 25 ìl supernatant from XPR2 transformants, which corresponds to the activity

of 0·0317 ìg native enzyme.

1280

. ¨

 

.

1998 John Wiley & Sons, Ltd.

Yeast 14, 1267–1283 (1998)

Y. lipolytica in the case of the plate assay (

Figure

5

) and not in the liquid assay.

The initial examination of Hansenula poly-

morpha, Kluyveromyces lactis, Schizosaccharo-

myces pombe and Yarrowia lipolytica as possible

alternative host organisms in expression cloning

showed that all of them are able to secrete active

forms of the test gene products at a higher amount

than the original host organism S. cerevisiae. The

expression capacity of the various host systems

varied significantly in a gene-dependent manner.

These results show that S. cerevisiae is certainly

not the optimal choice for a host organism and

that improvement of the expression cloning

method probably implies handling of more than

one host organism.

Following the comparative yeast study, we

focused on optimization of the attractive alterna-

tive host Y. lipolytica by replacement of the

initially employed XPR2 promoter. Based on the

present results, the TEF gene promoter seems to be

the most promising of the new Y. lipolytica pro-

moters. As the TEF gene promoter of various

other organisms has proved to be strong, it ap-

pears as an attractive tool for various purposes,

e.g. homologous or heterologous gene expression.

So far, only a few groups have tested the TEF gene

promoter in vivo (e.g. Axelos et al., 1989; Nakari-

Seta¨la¨ and Penttila¨, 1995) and these results indicate

that the strength of the promoter depends on

the origin. Our comparative examination of the

Y. lipolytica TEF gene- and XPR2 promoters

strongly indicates that, in the case of Y. lipolytica,

the TEF gene promoter is a very e

ffective tool for

expression of heterologous proteins.

In the present work we have demonstrated that

one of the most attractive of the examined alter-

native yeasts, Y. lipolytica, by use of an XPR2

promoter-based expression vector, expresses active

forms of the six tested gene products at significant

higher amounts than S. cerevisiae—ranging from

4·5 times to more than two orders of magnitude.

Subsequently, we have cloned two strong pro-

moters from Y. lipolytica, of which especially the

TEF gene promoter seems almost as e

fficient as the

XPR2 promoter but at conditions more suitable

for use in expression cloning. Use of Y. lipolytica

as the host organism in a TEF gene promoter-

based expression system will definitely increase the

number of positive cDNA clones to be detected

and allow production of some of the cloned

genes in su

fficient amounts to enable initial

characterization.

ACKNOWLEDGEMENTS
We thank Dr Fiona Du

ffner and Dr Sakari

Kauppinen for useful discussions and Heidi

Heinsøe, Jannie Steinvig, Lars A. Petersen and

Mari-Ann Allerslev for skilful technical assistance.

REFERENCES

Axelos, M., Bardet, C., Liboz, T., Thai, A. L. V., Curie,

C. and Lescure, B. (1989). The gene family encoding

the Arabidopsis thaliana translation elongation fac-

tor EF-1á: Molecular cloning, characterization and

expression. Mol. Gen. Genet. 219, 106–112.

Aviv, H. and Leder, P. (1972). Purification of bio-

logically active globin messenger RNA by chroma-

tography on oligothymidylic acid-cellulose. Proc. Natl

Acad. Sci. USA 69, 1408–1412.

Becker, D. M. and Guarente, L. (1991). High-e

fficiency

transformation of yeast by electroporation. Methods

Enzymol. 194, 182–187.

Boel, E. and Huge-Jensen, I. B. (1989). Recombinant

Humicula lipase and process for the production of

recombinant humicula lipases. EP0305216A1.

Bro¨ker, M. and Ba¨uml, O. (1989). New expression

vectors for the fission yeast Schizosaccharomyces

pombe. FEBS Lett. 248, 105–110.

Bro¨ker, M. (1987). Transformation of intact Schizo-

saccharomyces pombe cells with plasmid DNA. Bio-

Techniques 5, 516–518.

Buckholz, R. G. and Gleeson, M. A. G. (1991). Yeast

systems for the commercial production of heterolo-

gous proteins. Bio/Technology 9, 1067–1072.

Cavallius, J., Zoll, W., Chakraburtty, K. and Merrick,

W. C. (1993). Characterization of yeast EF-1á:

non-conservation of post-translational modifications.

Biochim. Biophys. Acta 1163, 75–80.

Chirgwin, J. M., Przybyla, A. E., MacDonald, R. J. and

Rutter, W. J. (1979). Isolation of biologically active

ribonucleic acid from sources enriched in ribo-

nuclease. Biochemistry 18, 5294–5299.

Christgau, S., Kauppinen, S., Vind, J., Kofod, L. V. and

Dalbøge, H. (1994). Expression cloning, purification

and characterization of a â-1,4 mannanase from As-

pergillus aculatus. Biochem. Mol. Biol. Int. 33, 917–925.

Christgau, S., Sandal, T., Kofod, L. V. and Dalbøge, H.

(1995). Expression cloning, purification and charac-

terization of a â-1,4-galactanase from Aspergillus

aculatus. Curr. Genet. 27, 135–141.

Christgau, S., Kofod, L. V., Halkier, T., Andersen,

L. N., Hockauf, M., Do¨rreich, K., Dalbøge, H. and

Kauppinen, S. (1996). Pectin methyl esterase from

Aspergillus aculeatus: expression cloning in yeast and

characterization of the recombinant enzyme. Biochem.

J. 319, 705–712.

Dalbøge, H. and Heldt-Hansen, H. P. (1994). A novel

method for e

fficient expression cloning of fungal

enzyme genes. Mol. Gen. Genet. 243, 253–260.

1281

    

1998 John Wiley & Sons, Ltd.

Yeast 14, 1267–1283 (1998)

Draborg, H., Kauppinen, S., Dalbøge, H. and

Christgau, S. (1995). Molecular cloning and expres-

sion in S. cerevisiae of two exochitinases from

Trichoderma harzianum. Biochem. Mol. Biol. Int. 36,

781–791.

Draborg, H., Christgau, S., Halkier, T., Rasmussen, G.,

Dalbøge, H. and Kauppinen, S. (1996). Secretion of

an enzymatically active Trichoderma harzianum endo-

chitinase by Saccharomyces cerevisiae. Curr. Genet.

29, 404–409.

Feinberg, A. P. and Vogelstein, B. (1983). A technique

for radiolabeling DNA restriction endonuclease frag-

ments to high specific activity. Anal. Biochem. 132,

6–13.

Fleer, R., Yeh, P., Amellal, N., Maury, I., Fournier, A.,

Bacchetta, F., Baduel, P., Jung, G., L’Hoˆte, H.,

Becquart, J., Fukuhara, H. and Mayaux, J. F. (1991).

Stable multicopy vectors for high-level secretion of

recombinant human serum albumin by Kluyveromyces

yeasts. Bio/Technology 9, 968–975.

Fournier, P., Guyaneux, L., Chasles, M. and Gaillardin,

C. (1991). Scarcity of ars sequences isolated in a

morphogenesis mutant of the yeast Yarrowia lipo-

lytica. Yeast 7, 25–36.

Gellissen, G. and Hollenberg, C. P. (1997). Applications

of yeasts in gene expression studies: a comparison

of Saccharomyces cerevisiae, Hansenula polymorpha

and Kluyveromyces lactis—a review. Gene 190, 87–

97.

Gleeson, M. A., Ortori, G. S. and Sudbery, P. E.

(1986). Transformation of the methylotrophic yeast

Hansenula polymorpha. J. Gen. Microbiol. 132, 3459–

3465.

Gubler, U. and Ho

ffman, B. J. (1983). A simple and very

e

ffective method for generating cDNA libraries. Gene

25, 263–269.

Kauppinen, S., Christgau, S., Kofod, L. V., Halkier, T.,

Do¨rreich, K. and Dalbøge, H. (1995). Molecular

cloning and characterization of a rhamnogalacturo-

nan acetylesterase from Aspergillus aculeatus. J. Biol.

Chem. 270, 27,172–27,178.

Kim, D. W., Uetsuki, T., Kaziro, Y., Yamaguchi, N.

and Sugano, S. (1990). Use of the human elongation

factor 1á promoter as a versatile and e

fficient expres-

sion system. Gene 91, 217–223.

Kofod, L. V., Kauppinen, S., Christgau, S., Andersen,

L. N., Heldt-Hansen, H. P., Do¨rreich, K. and

Dalbøge, H. (1994). Cloning and characterization of

two structurally and functionally divergent rham-

nogalacturonases from Aspergillus aculatus. J. Biol.

Chem. 269, 29,182–29,189.

Kyhse-Andersen, J. (1984). Electroblotting of multiple

gels: a simple apparatus without bu

ffer tank for rapid

transfer of proteins from polyacrylamide to nitro-

cellulose. J. Biochem. Biophys. Meth. 10, 203–209.

Laemmli, U. K. (1970). Cleavage of structural proteins

during the assembly of the head of bacteriophage T4.

Nature 227, 680–685.

Leer, R. J., Van Raamsdonk-Duin, M. M. C., Mager,

W. H. and Planta, R. J. (1985). Conserved sequences

upstream of yeast ribosomal protein genes. Curr.

Genet. 9, 273–277.

Mager, W. H. (1988). Control of ribosomal protein gene

expression. Biochim. Biophys. Acta 949, 1–15.

Meissner, P. S., Sisk, W. P. and Berman, M. L. (1987).

Bacteriophage ë cloning system for the construction

of directional cDNA libraries. Proc. Natl Acad. Sci.

USA 84, 4171–4176.

Nakari-Seta¨la¨, T. and Penttila¨, M. (1995). Production of

Trichoderma reesei cellulases on glucose containing

media. Appl. Envir. Microbiol. 61, 3650–3655.

Needleman, S. B. and Wunsch, C. D. (1970). A general

method applicable to the search for similarities in the

amino acid sequence of two proteins. J. Mol. Biol. 48,

443–453.

Nicaud, J. M., Fabre, E., Beckerich, J. M., Fournier, P.

and Gaillardin, C. (1989). Cloning, sequencing and

amplification of the alkaline extracellular protease

(XPR2) gene of the yeast Yarrowia lipolytica. J.

Biotechnol. 12, 285–298.

Ogrydziak, D. M., Demain, A. L. and Tannenbaum,

S. R. (1977). Regulation of extracellular protease

production in Candida lipolytica. Biochim. Biophys.

Acta 497, 525–538.

Ogrydziak, D. M. and Scharf, S. J. (1982). Alkaline

extracellular protease produced by Saccharomyces

lipolytica CX161-1B. J. Gen. Microbiol. 128, 1225–

1234.

Pearson, W. R. and Lipman, D. J. (1988). Improved

tools for biological sequence analysis. Proc. Natl

Acad. Sci. USA 85, 2444–2448.

Raleigh, E. A., Murray, N. E., Revel, H., Blumenthal,

R. M., Westaway, D., Reith, A. D., Rigby, P. W. J.,

Elhai, J. and Hanahan, D. (1988). MrcA and McrB

restriction phenotypes of some E. coli strains and

implications for gene cloning. Nucleic Acids Res. 16,

1563–1575.

Ro¨sel, H. and Kunze, G. (1995). Cloning and charac-

terization of a TEF gene for elongation factor 1á

from the yeast Arxula adeninivorans. Curr. Genet. 28,

360–366.

Sambrook, J., Fritsch, E. F. and Maniatis, T. (1989).

Molecular Cloning: A Laboratory Manual, 2nd edn.

Cold Spring Harbor Laboratory Press, New York.

Sanger, F., Nicklen, S. and Coulson, A. R. (1977). DNA

sequencing with chain-terminating inhibitors. Proc.

Natl Acad. Sci. USA 74, 5463–5467.

Schirmaier, F. and Philippsen, P. (1984). Identification

of two genes coding for the translation elongation

factor EF-1á of S. cerevisiae. EMBO J. 3, 3311–3315.

Sherman, F. (1991). Getting started with yeast. Methods

Enzymol. 194, 3–21.

Southern, E. M. (1975). Detection of specific sequences

among DNA fragments separated by gel electro-

phoresis. J. Mol. Biol. 98, 503–517.

1282

. ¨

 

.

1998 John Wiley & Sons, Ltd.

Yeast 14, 1267–1283 (1998)

Steiner, S. and Philippsen, P. (1994). Sequence and

promoter analysis of the highly expressed TEF gene of

the filamentous fungus Ashbya gossypii. Mol. Gen.

Genet. 242, 263–271.

Synetos, D., Dabeva, M. D. and Warners, J. R. (1992).

The yeast ribosomal protein S7 and its genes. J. Biol.

Chem. 267, 3008–3013.

Thomas, P. S. (1983). Hybridization of denatured RNA

transferred or dotted to nitrocellulose paper. Methods

Enzymol. 100, 255–266.

Ursin, V. M., Irvine, J. M., Hiatt, W. K. and

Shewmaker, C. K. (1991). Developmental analysis of

elongation-factor-1á expression in transgenic tobacco.

Plant Cell 3, 538–591.

Vernis, L., Abbas, A., Chasles, M., Gaillardin, C. M.,

Brun, C., Huberman, J. A. and Fournier, P. (1997).

An origin of replication and a centromere are both

needed to establish a replicative plasmid in the yeast

Yarrowia lipolytica. Mol. Cell. Biol. 17, 1995–2004.

Warner, J. R. (1989). Synthesis of ribosomes in Sac-

charomyces cerevisiae. Microbiol. Rev. 53, 256–271.

Yanisch-Perron, C., Vieira, J. and Messing, J. (1985).

Improved M13 phage cloning vectors and host

strains: nucleotide sequences of the M13mp18 and

pUC19 vectors. Gene 33, 103–119.

1283

    

1998 John Wiley & Sons, Ltd.

Yeast 14, 1267–1283 (1998)