ABSTRACT
The dominant
kan
r marker gene plays an important role in gene disruption experiments in budding
yeast, as this marker can be used in a variety of yeast strains lacking the
conventional yeast markers. We have developed a
loxP
-
kanMX
-
loxP
gene disruption cassette
,
which combines the advantages of the heterologous
kan
r marker with those from the Cre-
lox P
recombination system. This disruption cassette integrates with high efficiency
via homologous integration at the correct genomic locus (routinely 70%). Upon
expression of the Cre recombinase the
kanMX
module is excised by an efficient recombination between the
loxP
sites, leaving behind a single
loxP
site at the chromosomal locus. This system allows repeated use of the
kan
r marker gene and will be of great advantage for the functional analysis of gene
families.
The yeast
Saccharomyces cerevisiae
will be the first eukaryotic organism for which the entire sequence of the
genome will be determined (
1
). This will result in the identification of all 6500-7000 genes from this organism. The next big challenge is the functional
characterization of the unknown gene products. The first step towards a
functional analysis of a protein is the complete deletion of the corresponding
gene on the chromosome (null mutant) using the one-step gene transplacement method (
2
). Classically, DNA fragments flanking the gene of interest are cloned left and
right of a yeast marker gene and, after transformation of this construct into
yeast, homologous recombination between the flanking regions results in a
deletion of the gene and the simultaneous integration of the marker gene.
Because yeast has very efficient mechanisms for homologous recombination, it
has been possible to reduce the length of the flanking DNA regions to 30-45 bp, allowing the construction of gene disruption cassettes by the
polymerase chain reaction (PCR) (
3
-
5
; see also Fig.
2
). This system allows construction of gene disruption cassettes without cloning
steps and only requires the DNA sequence of the relevant chromosomal locus. The
system could be improved significantly by using a heterologous marker gene.
This avoids the problem of gene conversion associated with the use of yeast
marker genes and recipient yeast strains not completely deleted for the marker
gene. The
kan
r
gene from the
Escherichia coli
transposon Tn
903
when expressed in yeast renders the transformants resistant to the
aminoglycoside antibiotic G418 (
6
,
7
). Very recently the
kan
r
gene was fused to the
TEF
promoter and terminator sequences from the filamentous fungus
Ashbya gossypii
yielding the
kanMX
expression module. PCR-mediated generation of disruption cassettes carrying this completely
hetero- logous expression module allows efficient gene disruptions (
8
).
An important result from the yeast sequence analysis is the finding that a
substantial portion of genes are duplicated in the genome (
1
). For example,
in silico
analysis of the available sequence data revealed the existence of at least 15
proteins belonging to the HXT family of hexose transporters (
9
). A second example are the flocculation genes, which constitute a new
subtelomeric gene family (
10
). Thus, for the functional analysis of an unknown gene all members of a gene
family might have to be disrupted before the real phenotype of a null mutant
can be studied. For example, only if all seven hexose transporter genes,
HXT1
-
HXT7
, are deleted can reduced growth on low and high glucose media be observed (
11
).
As the number of marker genes is limited, efficient procedures for marker rescue
will be very important for functional analysis projects. For marker rescue, in
the presently available disruption cassettes the marker genes are surrounded by
direct repeats of 40-1100 bp. After gene disruption homologous recombination between the two
repeats results in marker removal, leaving behind a single repeat at the
deleted gene locus (
8
,
12
,
13
). The disadvantage of the low mitotic recombination frequency of ~10
-4
has been overcome in the case of the
URA3
marker gene by selecting for the loss of the marker on 5-fluoroorotic acid (5-FOA) (
12
,
13
). Alternatively the Cre-
loxP
recombination system of bacteriophage P1 has been shown to mediate efficient
recombination between
loxP
sites flanking a marker gene in yeast, resulting in excision of the marker gene
(
14
).
We have constructed the gene disruption cassette
loxP
-
kanMX
-
loxP
, which combines the advantages of the heterologous
kan
r
marker with those of the Cre-
loxP
system. The PCR-generated gene disruption cassette integrates with high efficiency via
homologous recombination at the correct genomic locus (routinely 70%). In
addition, the fast marker rescue procedure allows repeated use of the
kan
r
marker.
Strain CEN.PK2 (MAT
a
/MAT[alpha]
leu2-3,112/leu2-3,112 ura3-52 /ura3-52 trp1-289/trp1-289 his3-
[Delta]
1/ his3-
[Delta]
1 MAL2-8
c
SUC2
) was obtained from Karl-Dieter Entian (University of Frankfurt, Germany). The
E.coli
strain XL1-Blue was from Stratagene (Heidelberg, Germany). Yeast and
E.coli
media were prepared as described (
15
). For selection for G418 resistance after yeast transformations the YPD plates were supplemented with 200 mg/l Geneticin
®
(G-418 sulphate from Gibco BRL, Germany; the activity of this chemical may be
batch dependent). Selection for the
kan
r
gene in
E.coli
was on YT plates supplemented with 50 mg/l kanamycin sulphate (Fluka, Germany) (
8
).
The
loxP
-
kanMX
-
loxP
module containing plasmid pUG6 was constructed by modification of plasmid pFA6-kanMX4 (
8
). Four oligonucleotides (319, 320, 321 and 322) were made, which after
annealing formed the two oligonucleotide duplexes 319/320 and 321/322, each
carrying the 34 bp
loxP
sequence and appropriate restriction enzyme sites at either end (Table
2
). The oligonucleotide duplex 319/320 was ligated into the
Sal
I/
Bgl
II-cleaved pFA6-kanMX4 plasmid. Next, this modified vector was cleaved with
Sac
I and
Eco
RV and the oligonucleotide duplex 321/322 was inserted, generating plasmid pUG6
(Fig.
1
A). The modified regions were sequenced using the SP6 and T7 primers to verify
correct insertion of the oligonucleotide duplexes.
We pooled two preparative PCR reactions for a gene disruption. A 100 [mu]l preparative PCR reaction contained 10 [mu]l 10* GoldStar PCR buffer [750 mM Tris-HCl, pH 9.0, 200 mM ammonium sulphate, 0.1% (w/v) Tween
20], 200 [mu]M dNTP mix (200 [mu]M each of dATP, dCTP, dGTP and dTTP), 30-60 ng pUG6 template DNA, 100 pmol each primer, 1.5 mM MgCl
2
and 0.25 U GoldStar polymerase (Eurogentec, Belgium). The PCR conditions were:
98oC for 1.0 min, 50oC for 1.0 min, 72oC for 2.5 min (40 cycles). Aliquots of 5 [mu]l of the reaction were quantified on an agarose gel to verify
the amount of DNA. The two PCR reactions were pooled and the DNA was extracted
with phenol/chloroform (1:1) and ethanol precipitated. The pellet was
resuspended in 13 [mu]l water and 0.5 [mu]l were analyzed on an agarose gel. Typically, the concentration of the
PCR product was 0.3-0.5 [mu]g/[mu]l. About 0.5 [mu]l DNA were used for cloning the disruption cassette into pUG7
(Güldener and Hegemann, unpublished results), while 12 [mu]l (~5 [mu]g) were used for yeast transformation.
The yeast transformation procedure used was a slightly modified version of the
protocol described by Gietz and colleagues (
18
). Cells from an overnight culture were resuspended in 50 ml YPD (start OD
600
0.2) and grown to an OD
600
of 0.7-1.0. The cells were harvested by centrifugation (Heraeus Minifuge RF; 4000 r.p.m., 5 min) and resuspended in 10 ml sterile distilled water. The cells were
harvested by centrifugation, resuspended in 1 ml water and transferred to a 1.5
ml Eppendorf tube. The cells were harvested by centrifugation (Heraeus Biofuge
13; 5000 r.p.m., 1 min) and resuspended in 1.5 ml freshly prepared sterile
TE/LiOAc (prepared from 10* concentrated stocks; 10 * TE = 0.1 M Tris-HCl, 0.01 M EDTA, pH 7.5; 10* LiOAc = 1 M LiOAc adjusted to pH 7.5 with dilute
acetic acid). The cells were harvested again and resuspended in 200 [mu]l TE/LiOAc (cell concentration should be ~2 * 10
9
cells/ml). For a gene disruption experiment ~5 [mu]g DNA (12 [mu]l) of the disruption cassette were mixed with 50 [mu]g (5 [mu]l) of freshly denatured salmon sperm DNA (10 mg/ml, boiled
for 20 min in a water bath, then chilled in ice/water; the stock solution was
prepared as described by Schiestl and Gietz;
19
) and 50 [mu]l cells in TE/LiOAc were added and mixed with caution (no vortexing!).
Immediately 300 [mu]l of freshly prepared sterile 40% PEG 4000 (prepared from stock solutions:
50% PEG 4000, 10* TE, 10* LiOAc, 8:1:1 v/v, pH 7.5) were added and carefully mixed (no
vortexing!). Cells were incubated for 30 min at 30oC with constant agitation. Cells were incubated for 15 min at 42oC, then 800 [mu]l sterile water were added, mixed and cells were collected by
centrifugation (Biofuge 13; 13 000 r.p.m., 10 s). Cells were resuspended in 1 ml YPD (no vortexing!)
and incubated for 2-3 h at 30oC. Cells were collected by centrifugation, resuspended in 200 [mu]l YPD and plated onto YPD plus G418 plates (200 [mu]g/ml G-418; Gibco BRL). Plates were incubated at 30oC until colonies appeared. In cases where the
background growth was too strong after 24-48 h, the microcolonies were replica-plated onto new YPD plus G418 plates. After re-streaking transformants onto YPD plus G418 plates only
transformants growing well were chosen for further analysis.
Transformation of yeast strain CEN.PK2 with the
loxP-kanMX-loxP
disruption cassette typically yields 10-100 kan
+
transformants/[mu]g DNA, while transformation of a control plasmid (pRS316, a
CEN
,
ARSH4
,
URA3
plasmid) routinely gives 2-10 * 10
4
transformants/[mu]g DNA.
Detection of the correct gene disruption of ORF
N2809
was done by either diagnostic PCR or Southern analysis. For PCR yeast cells
were taken directly from a YPD+G-418 plate. The PCR primers were a `start' primer and a `stop' primer,
located 400-500 nt upstream of the ATG and of the stop codon respectively of the
disrupted gene, and two primers located in the 5' (kanRE) and in the 3' regions (kanFW) of the
kan
r
gene respectively (Table
1
). Genomic integration of the disruption cassette was verified using the ORF
`start' and the kanRE primers as well as with the ORF `stop' and the kanFW
primers. After the
kan
r
gene was excised by the Cre recombinase the ORF `start' and `stop' primers were
used. The PCR reaction was as described for the generation of the disruption
cassette. The conditions were: 94oC for 2.0 min (hot start); then 40 cycles of 94oC for 1.5 min, 45oC for 2.0 min and 72oC for 2 min. About 15 [mu]l of the PCR reaction were loaded on an agarose gel.
Table 1
Analysis of the disruption events of ORF
N3265
was by Southern analysis using a PCR-generated probe (0.46 kb) located 5' of the ORF. The PCR was carried out as described above using
primers 400 and 424 (Table
1
). The probe was labeled with digoxigenin and the Southern analysis was
performed according to the supplier's instructions (Boehringer, Mannheim,
Germany).
The aim of this work was to combine the great advantages of the dominant
kan
r
marker system for gene disruption and the Cre-
loxP
recombination system for marker rescue. To this end the plasmid pFA6-kanMX4 (
8
) was modified by integrating two 34 bp
loxP
sequences as direct repeats left and right of the
kanMX
module using appropriate oligonucleotide duplexes (Fig.
1
A). The new vector, pUG6, has all the restriction sites present in pFA6-kanMX4 excluding those used for integration of the
loxP
sequences (Fig.
1
A). Thus pUG6 can be used in the classical way for gene disruption experiments.
For this, homologous DNA fragments are cloned left and right of the
loxP
-
kanMX
-
loxP
module, followed by a second restriction enzyme digest to generate the gene
disruption cassette. Alternatively, short homologous sequences needed for
homologous recombination can be fused to the
loxP
-
kanMX
-
loxP
module via a PCR reaction (
3
,
4
). Routinely we use two oligonucleotides of ~60 nt in length, comprised of a 20 nt long segment homologous to sequences
left or right of the
kanMX
module at their 3'-ends and of a 40 nt long segment homologous to sequences left or
right of the gene to be deleted at their 5'-ends (Fig.
1
B ). The oligonucleotides also hybridize to the pFA6-kanMX vectors and thus can be used to generate disruption cassettes
without the
loxP
sites if required.
To evaluate the efficiency of the
loxP
-
kanMX
-
loxP
module in gene disruption experiments the pUG6 vector was used as target for
the PCR-mediated generation of disruption cassettes for three different ORFs. The
ORFs
N0868
,
N0901
and
N3216
had been identified during systematic sequencing work on chromosome XIV (
20
; Sen-Gupta
et al
., unpublished results; Düsterhöft, unpublished results). The corresponding oligonucleotides for
generation of the disruption cassettes were 59-62 nt in length (Table
1
). Transformation into the diploid yeast strain CEN.PK2 yielded sufficient
numbers of kan
+
transformants (48-370 transformants; Table
2
). Meanwhile, another 10 gene disruptions have been performed using the
loxP
-
kanMX
-
loxP
module, yielding on average 10-100 kan
+
tranformants/[mu]g disruption cassette (data not shown). Diagnostic PCR and Southern analysis
of a selected number of transformants verified the correct integration of the
cassette into the chromosomal target in ~70% of the kan
+
clones on average (Table
2
). The data show that the percentage of correctly integrated disruption
cassettes is independent of the number of transformants obtained in a
particular experiment, suggesting that the transformation efficiency is not
affected by the integration efficiency at the correct locus. The results of
these disruption experiments demonstrate the power of the
loxP
-
kanMX
-
loxP
module in gene disruption experiments.
To use the
kan
r
marker repeatedly for several gene disruptions in one strain it is necessary to
eliminate the marker from the successfully disrupted gene. The heterozygous
disruption strain CEN.HE18, in which one of the two copies of ORF
N2809
was disrupted by the
loxP-kanMX-loxP
cassette (
N2809
/
N2809
::
loxP
-
kanMX
-
loxP
), was transformed with the
cre
expression plasmid pSH47, which carries the
URA3
marker gene and the
cre
gene under the control of the inducible
GAL1
promoter. Expression of the Cre recombinase was induced by shifting cells from
YPD (glucose) to YPG (galactose) medium (Fig.
2
). Growth for only 2 h in galactose medium after transfer from glucose medium was
sufficient to remove the
kan
r
marker gene in ~80-90% of the cells, as detected by plating cells on YPD and replica-plating the colonies onto YPD plus G418. Correct loss of the
kan
r
marker gene was verified by diagnostic PCR and Southern analysis, which
confirmed that in all kan
-
clones tested the
kan
r
marker gene had been excised, leaving behind a single
loxP
site at the chromosomal
N2809
locus (data not shown, but see Fig.
3
). The
cre
expression plasmid was removed from this strain by streaking cells on plates
containing 5-fluoroorotic acid to counterselect for the loss of the plasmid, yielding strain CEN.HE18-1 (relevant genotype
N2809
/-
N2809
::
loxP
).
The central question was, whether it would be possible to create a second gene
disruption in strain CEN.HE18-1 using the
loxP
-
kanMX
-
loxP
cassette and subsequently remove the
kan
r
marker again by induction of the Cre recombinase. To test this, strain CEN.HE18-1 was transformed with a PCR-generated disruption cassette with homology to 3'- and 5'-ends of ORF
N3265
. This ORF is also located on chromosome XIV and in opposite orientation ~93 kb away from the already disrupted ORF
N2809
. From 68 kan
+
transformants obtained, two of four clones tested by diagnostic PCR were found
to be correctly integrated into the genome (data not shown). Two ORF
N3265
-disrupted transformants (strain CEN.HE18-2) were transformed with the
cre
expression plasmid pSH47 to induce recombination and subsequent loss of the
kan
r
marker to generate
N3265
::
loxP
. The two plasmid-transformed strains were pre-grown on selective glucose plates to select for presence of the
plasmid and then shifted to liquid YPD (glucose) and grown overnight. The cells
were then shifted to liquid YPG (galactose) for a 30, 60, 90 or 120 min period
to follow the kinetics of Cre induction. This would indicate whether a longer
incubation with the Cre recombinase would be toxic to the cells due to
recombination between the
loxP
sites at the
N3265
locus and the
N2809
locus leading to loss of all the genetic material located between both loci.
Afterwards cells were re-shifted to YPD plates and analyzed for the presence or absence of the
kan
r
marker. Incubation in YPD showed that ~11% of cells tested had lost the
kan
r
marker, indicating that the galactose-regulated
GAL1
promoter is not completely shut off in glucose medium and thus a small amount
of Cre recombinase is produced in YPD medium. Thirty minutes after the shift to
YPG ~80% of the cells were kan
-
and this number did not change significantly at later time points. Southern
analysis of the kan
-
strains revealed that all 72 kan
-
clones tested (from time points 0 and 30 min, 24 clones each; from time points
60 and 90 min, 12 clones each) had lost the
kan
r
gene, yielding a 2.16 kb band (Fig.
3
) indicative of the expected
N3265
::
loxP
disruption. As the
N2809
::
loxP
locus could be a second potential target for a recombination event we analyzed
this locus by diagnostic PCR. Twenty-four kan
-
clones analyzed showed a band of 0.86 kb in length, indicating that this locus
was intact and that no recom- bination had occurred between this
loxP
site and the other site at
N3265
(Fig.
3
).
These results show that all Cre-induced recombination events analyzed involved only the two
loxP
sites flanking the
kanMX
module. The single
loxP
site located ~93 kb away from the second integration site is obviously too far away to be
relevant for the Cre-induced recombination process. Further studies are required to determine
precisely how close the
loxP
-
kanMX
-
loxP
cassette can be placed to a single
loxP
site without affecting the desired recombination process. Furthermore, it needs
to be determined whether location of the
loxP
-
kanMX
-
loxP
cassette and the single
loxP
site on the same chromosome (see below), on homologous chromosomes or on non-homologous chromosomes is relevant for the outcome of the recombination
process. In mouse embryonic stem cells the frequency of a Cre-mediated trans- location between
loxP
sites located on non-homologous chromosomes was determined to be 1 in ~1200-2400 cells expressing Cre recombinase (
21
).
The
kan
r
marker is the recommended marker for the European Project for the Functional Analysis of Unknown Genes (EUROFAN) (Guidelines for
the EUROFAN B0 program: ORF Deletants, Plasmid Tools, Basic Functional Analysis by A. Wach, A. Brachat and P.
Philippsen, personal communication). The results presented here show that the
kan
r
marker, which is very efficient in gene disruption experiments, can be
successfully combined with the Cre-
loxP
recombination system to generate the reusable disruption cassette
loxP
-
kanMX
-
loxP
. Repeated use of the
kan
r
marker is now possible and will be of great advantage for the analysis of gene
families. After a first gene disruption, tetrad analysis can be used to
determine whether the disrupted gene is essential or not. In cases where it is
known that the gene to be disrupted is non-essential, gene disruption and Cre-mediated recovery of the
kan
r
marker can be performed directly in haploid yeast strains. ORFs
N2809
and
N3265
are both non-essential genes and were also deleted in a two-step process using the
loxP
-
kanMX
-
loxP
cassette and the Cre recombinase in a haploid CEN.PK-derived strain. No reduction in viability was observed after Cre-mediated removal of the
kan
r
gene of the second deletion, indicating also that no recombination event had
taken place between the single
loxP
site at the
N2809
locus and the two
loxP
sites at
N3265
(data not shown).
Further improvements to this system are possible. For example, the high
efficiency of the disruption cassettes should allow simultaneous transformation
of two disruption cassettes for two different genes. It has been shown
previously that co-transformation efficiency is 30-40% in yeast (
22
). Such a `co-gene disruption' would be even more efficient when the two disruption
cassettes contain different markers. Very recently a new hetero- logous
HIS
marker has been constructed which might be suitable for this purpose (Wach,
Brachat, Alberti-Segui and Philippsen, personal communication). After Cre activation both
markers can be recovered and used again.
In summary, the
loxP
-
kanMX
-
loxP
cassette in conjunction with the Cre recombinase will significantly facilitate
multiple gene disruptions. It is particularly useful in yeast strains carrying
incomplete deletions of the normally used homologous marker genes, as well as
in industrially used yeast strains, which often lack the standard auxotrophic
markers present in laboratory strains.
ACKNOWLEDEGEMENTS
We thank Dr Ursula Fleig for critical reading of the manuscript. We thank Dr
Brasch for plasmid PBS39 and Drs Wach and Philippsen for the pFA plasmid
series. This work was supported by a grant of the Deutsche Forschungsgemeinschaft (Sonderforschungsbereich 272, T.P. A1) and of the BMBF (no. FKZ 0310577) to JHH.
REFERENCES
Return