Plant Molecular Biology 45: 41–49, 2001.
© 2001 Kluwer Academic Publishers. Printed in the Netherlands.
41
A recombinase-mediated transcriptional induction system in transgenic
plants
Tine Hoff, Kirk M. Schnorr1 and John Mundy∗
Department of Plant Physiology, University of Copenhagen, Oester Farimagsgade 2A, 1353 Copenhagen K, Denmark (∗ author for correspondence; e-mail: [email protected]); 1 Present address: Novo Nordisk A/S, 1BM1.05
Novo Alle, 2880 Bagsvaerd, Denmark
Received 17 February 2000; accepted in revised form 30 July 2000
Key words: Arabidopsis, Cre-lox, Cre-recombinase, heat shock, inducible, vector
Abstract
We constructed and tested a Cre-loxP recombination-mediated vector system termed pCrox for use in transgenic
plants. In this system, treatment of Arabidopsis under inducing conditions mediates an excision event that removes
an intervening piece of DNA between a promoter and the gene to be expressed. The system developed here uses
a heat-shock-inducible Cre to excise a DNA fragment flanked by lox sites, thereby generating a constitutive GUS
reporter gene under control of the CaMV 35S promoter. Heat-shock-mediated excision of several, independent
lines resulted in varying degrees of recombination-mediated GUS activation. Induction was shown to be possible
at essentially any stage of plant growth. This single vector system circumvents the need for genetic crosses required
by other, dual recombinase vector systems. The pCrox system may prove particularly useful in instances where
transgene over-expression, or under-expression by antisense, would otherwise affect embryo, seed or seedling
viability.
Abbreviations: BAR, phosphinothricin acetyltransferase; Basta, glufosinate; Cre, Cre-recombinase; GUS, βglucuronidase; HSP, heat shock promoter; Lfy, leafy; MUG, 4-methyl-umbelliferyl-β-D-glucuronide; NLS, nuclear
localisation signal; nos30 , nopaline synthase transcription terminator; npt, neomycin phosphotransferase; ocs30 ,
octopine synthase transcription terminator.
Introduction
Transgenic techniques are powerful tools in studies
of gene function and of biosynthetic and signalling
pathways in plants (Huang et al., 1999; Coles et al.,
1999). Most transgenic studies have been performed
using constitutive promoters, such as the 35S RNA
promoter from cauliflower mosaic virus, to ectopically over- or under-express sense or anti-sense RNAs.
While this is sufficient for some studies, inducible systems are preferable in cases where ectopic transgene
function is disruptive to normal growth and development. For example, induced expression of a mRNA
or antisense RNA at a specific temporal or developmental stage may differentiate between primary and
secondary effects of the transgene.
Several inducible systems have been reported for
plants (reviewed in Gatz and Lenk, 1998). Two recent, promising ones are an ethanol-inducible system
(Salter et al., 1998), and a steroid-inducible system (Aoyama and Chua, 1997). However, while the
ethanol-inducible system has been shown to work in
tobacco, it remains to be described in the preferred
model Arabidopsis. In addition, Kang et al. (1999)
report that the glucocorticoid-inducible system itself
may cause growth defects in Arabidopsis and induces
defence-related genes. Furthermore, chemical inducers such as steroids and alcohol may preferentially
concentrate in transpirationally active areas, leading
to spatially uneven inductions (Gatz and Lenk, 1998).
In an attempt to circumvent such problems, we
constructed an alternative inducible vector system
42
(pCrox) based on the Cre-lox site-specific recombination system from phage P1 (Sternberg and Hamilton,
1981). The 38 kDa Cre-recombinase catalyses recombination between two 34 bp lox sites. When these sites
occur as direct repeats, Cre-mediated recombination
results in excision of the intervening sequence. The
construct we describe here incorporates two features.
First, it was designed to excise an intervening piece
of DNA so that a promoter becomes proximal to the
coding region or antisense of a gene of interest, thus
activating sense or antisense RNA expression. Second,
the intervening piece of DNA contains the inducible
Cre-recombinase required for recombination. This design should stabilise the excision event by removing
the Cre-coding region, thus leading to loss of Crerecombinase protein. Such a design should eliminate
the need for genetic crosses to activate recombination
as has been done previously (Bayley et al., 1992).
Here we describe the use of this system in Arabidopsis and discuss its characteristics and potential
applications.
Material and methods
DNA constructions
The binary plasmid pPZP212 was used as a cloning
backbone (Hajdukiewicz et al., 1994). In addition
to the CaMV 35S promoter-driven Tn5 neomycin
phosphotransferase II gene from Tn5, conferring resistance to kanamycin in plants, pPZP212 also contains
spectinomycin resistance for selection in bacteria. The
phosphinothricin acetyltransferase gene (BAR) conferring plant resistance to the herbicide glufosinate
(Basta), under control of the CaMV 35S promoter
and with the octopine synthase ocs30 terminator, came
from plasmid SLJ2011 (Jones et al., 1992). Tn903
that confers bacterial kanamycin resistance came from
pUC4K (Taylor and Rose, 1988). The promoter of
the Arabidopsis heat-shock-responsive HSP81-1 gene
(451 bp 50 upstream of the first ATG) was from plasmid pTT102H which contains part of the HSP81-1
gene (kindly provided by Taku Takahashi; Takahashi
et al., 1992). The nuclear localisation signal (NLS) engineered into the Cre-coding region is the 96 bp B domain of the maize regulatory protein Opaque 2 (Varagona and Raikhel, 1994). The Cre-recombinase coding
region (cre) from bacteriophage P1 was amplified by
PCR using λKC as template (Elledge et al., 1991).
The E9 terminator poly(A) addition signal of the pea
ribulose-biphosphate carboxylase small subunit) was
a EcoRI-SmaI fragment from VIP11 (Morelli et al.,
1985). The lox sites were supplied as synthesised oligo
adapters by D. Bouchez, INRA-Versailles, France.
The CaMV 35S promoter used to drive expression
of the BAR or β-glucuronidase (GUS) gene and the
GUS construct with ocs30 terminator came from plasmid SLJ4D4 (Jones et al., 1992). All DNA fragments
amplified by PCR for vector construction were sequenced. Several restriction sites were destroyed in
different vector components to facilitate the use of the
multiple cloning site in the cloning version of the vector. Vector details can be obtained from the authors or
at http://biobase.dk/∼mundy.
Plant material and growth conditions
pCrox was introduced into Agrobacterium tumefaciens strain C58 containing the helper plasmid
pGV3101 by electroporation. Arabidopsis thaliana
ecotype Columbia-0 was used for transformation by
the in vivo infiltration method (Bechtold et al., 1993)
modified by adding 0.01% Silwet L-77 (Lehle Seeds)
to the infiltration medium. Primary transformants (T1 )
were selected in sand by watering with 10 mg/l Basta
for 10 days. Resistant seedlings were removed to soil,
allowed to self-fertilise and set seeds. T2 seeds were
plated on sterile plates containing MS medium supplemented with 1% sucrose, 0.7% agar and 50 µg/ml
kanamycin to analyse marker segregation. T3 lines
were selected with Basta in soil. Homozygous T3
plants were used for all further experiments. All
plants, in pots or on sterile plates, were grown at
21 ◦ C. For GUS analyses, plants were grown under
short-day light conditions (10 h light).
Heat induction experiments were performed with
two consecutive heat treatments such that plants were
incubated at 37 ◦ C for 16 h, allowed to recover for 32 h
at 21 ◦ C, and then incubated a second time at 37 ◦ C for
16 h. Plants in pots were incubated in a closed plastic
bag to avoid wilting during the heat treatment. Plants
were analysed 24–48 h after the end of the second heat
treatment. The HSP81-1::GUS transgenic line (HSPGUS, Yabe et al., 1994) expressing the GUS gene
under control of the HSP81-1 heat shock promoter,
was used as a control for heat induction experiments.
GUS assays
Histochemical GUS assay was performed with
100 mM sodium phosphate pH 7, 1 mM EDTA,
1% Triton X-100, 0.5 mM K3 Fe(CN)6 , 0.5 mM
43
K4 Fe(CN)6 , 2 mg/ml X-Gluc (5-bromo-4-chloro3-indolyl-β-glucuronide cyclohexylamine salt. 4methyl-umbelliferyl-β-D-glucuronide (MUG) was used
as substrate for GUS activity measurements in vitro.
Plant material was extracted in 0.1 M potassium phosphate pH 7.8, 1 mM DTT, then centrifuged for 10 min
at 4 ◦ C and proper dilutions of the supernatant used
for the assay. Equal volumes of plant extracts and
2× assay buffer (2 mM MUG, 50 mM sodium phosphate pH 7.8, 10 mM EDTA, 0.1% Triton X-100,
0.1% sodium lauryl sarkosyl, 10 mM DTT, 40% v/v
methanol) were mixed and incubated at 37 ◦ C. Measurements were made at time 0 and every 30 min in
a fluorometer (Wallac Victor II). Total protein contents in the extracts were determined by the protocol
of Bradford (1976). Relative GUS activities as fluorescence per minute per mg protein were calculated.
DNA and RNA analyses
Total RNA and genomic DNA extractions and Southern analysis were performed as described previously
(Hoff et al., 1995). The random-primed probe used
for Southern analysis was a fragment of the Tn5
kanamycin resistance gene. For northern blot analysis, total RNA was isolated from 20 plants 3–4 weeks
old without cotyledons after control or heat treatments.
The BAR probe was an ApaI fragment from pCrox
covering most of the BAR-coding region, and the GUS
probe was a PCR fragment containing 500 bp of the
GUS-coding region.
Semi-quantitative PCR was used to determine the
recombination status of in planta pCrox constructs.
Genomic DNA was isolated from 20 plants 3–4 weeks
old, from which the cotyledons were removed, after control or heat treatments. DNA was also isolated from the cotyledons of the uninduced samples
to analyse if excision of the block had occurred in
this tissue. Genomic DNA (10 ng) was used as template in PCR to quantify the relative levels of unexcised BAR DNA with the primers designed to yield
a 300 bp internal BAR gene fragment (BAR-545: 50 CTGCCAGAAACCCACGTCATG and BAR-260: 50 CGAATCGACCGTGTACGTCTC). As an endogenous control, an internal 250 bp fragment of leafy (Lfy;
Weigel et al., 1992) was amplified in the same PCR reaction (Lfy up: 50 -ACCTTCATGTGGGTTTGGAGC
and Lfy lo: 50 -GTGCAAGATCTCTTTCCCCGC).
PCR conditions were: initial denaturation at 94 ◦ C for
1 min, 22 cycles of 94 ◦ C for 45 s, 60 ◦ C for 45 s and
72 ◦ C for 1 min. The number of cycles was optimised
so the 4 primers together amplified bands of about the
same intensity when they were mixed in one tube as
when they were run pairwise in separate tubes. PCR
products were electrophoresed and Southern hybridisation performed as described above with BAR and
Lfy fragments as hybridisation probes. Filters were
autoradiographed and unsaturated exposures quantified with a CCD camera as relative intensity ratios of
BAR/Lfy.
Results
The pCrox vector
We constructed the Agrobacterium binary vector system incorporating the Cre-recombinase system from
phage P1 diagrammed in Figure 1. Two versions of
the vector were made, one with a multiple cloning
site and one with the GUS reporter construct as shown
here. In the absence of Cre-mediated recombination,
the GUS reporter should be transcriptionally silent.
Upon heat induction, the Cre-recombinase would be
expressed and excise the block between the two lox
sites. Excision of the block would bring the 35S promoter proximal to the GUS coding region on the right
side of the block and lead to constitutive GUS expression. The vector was constructed with the BAR marker
within the block and the kanamycin marker outside
the block. This allows selection for maintenance of the
promoter block containing the basta gene. A nuclear
localisation signal was engineered at the N-terminus
of the Cre-recombinase to increase Cre nuclear targeting, as studies with native Cre protein demonstrated
recombination frequencies of only 40–50% (Bayley
et al., 1992). The HSP81-1 heat shock promoter was
initially chosen because earlier work showed that it
confers no detectable expression in Arabidopsis tissues under normal, uninduced conditions except for
weak expression in secondary root hairs (Yabe et al.,
1994). In contrast, this promoter confers strong expression in all tissues upon heat induction. The GUS
gene was chosen as reporter gene to test the vector,
because GUS expression is easy to detect and quantify.
Induction of GUS activity in transgenic plants
The pCrox-GUS construct was introduced into Arabidopsis and forty independent transgenic lines were
selected with Basta. In the T2 generation, 9 lines
were selected with a segregation ratio close to 1:3 of
44
Figure 1. The pCrox vector system. Two lox sites (loxP) between the T-DNA borders flank a region (promoter block, shaded grey) containing
3 different gene cassettes. The first cassette contains the BAR-coding region with an octopine synthase transcription terminator (ocs30 ) whose
expression is driven by a 35S promoter located outside the block at the left. BAR confers plant resistance to the herbicide Basta. The second
cassette contains the neomycin phosphotransferase gene from TN903 that confers resistance to kanamycin in bacteria. The third cassette
contains Cre-recombinase modified to contain an amino-terminal nuclear localisation signal (NLS). Expression of this Cre-recombinase, with
the E9 terminator from the pea Rubisco gene, is driven by the heat shock-responsive promoter of the Arabidopsis HSP81-1 gene. To the right
of the LoxP flanking the block is a GUS gene with the terminator of the nopaline synthase gene from the Ti plasmid. A kanamycin-resistant
marker is included in the T-DNA providing selection in plants. Bottom: excision of the block brings the GUS gene under control of the 35S
promoter. For general cloning purposes, the GUS gene was substituted for a multiple cloning site as shown. The recognition sites for EcoRV
that were used for Southern analysis are marked (E), as is the Southern probe (npt-probe).
kanamycin-sensitive/resistant plants. The 1:3 segregation ratio indicates pCrox insertion at single loci, but
does not exclude multiple insertions at the same locus
or silent or truncated inserts at other loci. Homozygous T3 progeny were first analysed by histochemical
GUS assays of induced or uninduced whole seedlings.
We examined the effect of varying the number and
length of heat treatments, and found that while a
6 h heat treatment induced GUS expression, a longer
or repeated shorter heat treatments more markedly
increased GUS expression (results not shown). We
found that two heat treatments of 16 h, with a recovery period between and after, produced relatively
strong GUS activities, minimised physiological effects
of the heat treatment, and were practical to perform.
All uninduced lines showed some GUS expression in
the cotyledons and some also had GUS expression in
the roots (results not shown). In addition, individual
leaf cell, as well as sectors of true leaves showed GUS
expression in the uninduced state in some lines. The
extent of such uninduced sectoring varied with the individual transgenic line, with some lines having no
detectable expression (Figure 3). While this indicates
that the pCrox-GUS reporter vector was somewhat
leaky, GUS expression levels were significantly increased in all lines upon heat induction. We note that
although we used T3 lines homozygous for pCrox, T1 ,
T2 and T3 heterozygotes were also found to respond to
the induction conditions in the same manner (data not
shown).
Quantitative fluorometric GUS assays were then
performed on shoots of T3 homozygotes with true
leaves of induced and uninduced 2–3-week old plants.
The results of these MUG assays on 9 pCrox-GUS
lines, a control HSPGUS line and WT Arabidopsis are
shown in Figure 2a and b. Five of the nine lines (0,
29, 42, 65, 81) had low or undetectable GUS activity
before induction, while their levels of GUS activity
increased from 16 to 271 times after heat treatment
(Figure 2b). In contrast, four of the nine lines (2, 9,
67, 74) had high GUS levels prior to induction, but
all four had relatively low levels of GUS induction
(1.3 to 3.6 times). These results indicate that induction
worked best in lines with relatively low GUS expression, presumably due to low Cre expression in them.
This implies that the pCROX strategy may be best ap-
45
Figure 2. Quantitative, fluorometric GUS assays. a. GUS values
plotted for 9 homozygous pCrox-GUS lines, a control HSPGUS
line, and a wild-type Arabidopsis (Wt). Shoots with true leaves
of induced and uninduced 2–3-week old plants were used for the
assays. GUS activity is expressed in relative fluorescence per minute
per mg protein. Each value is the mean of measurements on 4 different plants. b. GUS values as determined in the fluorometric assay
with standard deviation in parenthesis and the calculated fold of
induction.
plied from transgene locations that are less favourable
for gene expression.
Four lines were then selected for more detailed
analysis on the basis of this MUG analysis: three lines
with no background expression (lines 42, 65 and 81)
and one line with high background expression (line 2).
The results of histochemical GUS assays on uninduced and induced 4–5-week old plants of these four
lines are shown in Figure 3. As inferred from the quantitative MUG assay, lines 2, 9 and 74 showed some
GUS expression in uninduced leaf sectors, and a more
uniform expression upon induction (Figure 3a–c). Pronounced staining was seen under both conditions in
cotyledons. Similarly, uninduced plants of lines 42,
65 and 81 showed some staining in cotyledons, but
no GUS expression in their uninduced leaves with the
exception of a few small leaf sectors (Figure 3a and b).
In contrast, uniform GUS expression was seen upon
Figure 3. Histochemical GUS assays of 4 pCrox-GUS lines. a. Upper row are uninduced lines, lower row are induced lines. From left
to right: lines 2, 42, 65, 81. b. Rosette of uninduced line 65 showing
sectoring. c. Rosette and flower of induced line 65. d. Germinating
seedlings of uninduced lines 2, 42, 65 and 81 (from left to right).
e. Cotyledons of uninduced line 42. f–i. GUS expression in line 42
(f, g) and line 65 (h, i) for 8-week old plants 3 weeks (f, h) and
1 week (g, i) after heat treatment. Somatic sectoring found in both
the induced and uninduced tissues is most clear in c and e.
46
heat induction in the leaves of lines 42, 65 and 81
(Figure 3a and c).
To further investigate GUS expression in cotyledons of uninduced plants, the same four lines were
GUS-stained at different stages during germination.
GUS stainings of the germinating seed, when the radicle had just penetrated the seed coat and when the
cotyledons were almost fully expanded, are shown in
Figure 3d and e. Line 2 stained strongly for GUS activity throughout the cotyledons (Figure 3d), while the
other 3 lines had only small sectors of GUS activity
throughout the cotyledons (Figure 3d and e). Since
these experiments were performed on seeds germinated at 21 ◦ C, germination of the four lines at lower
temperatures down to 4 ◦ C was also examined. Germination of the material at lower temperatures did
not reduce the GUS expression levels or patterns in
cotyledons (results not shown).
Several conclusions may be drawn from these experiments. First, while lines could easily be found
which inducibly express a transgene from the pCrox
vector, basal and inducible expression levels varied,
perhaps due to transgene position effects and/or the
number of T-DNA inserts per locus. Second, transgene
expression in cotyledons was apparently high, despite
the fact that the HSP81-1 promoter was not found to
be active in cotyledons of uninduced plantlets (Yabe
et al., 1994). Two explanations for the expression
of GUS in cotyledons may be noted. First, lowlevel expression of Cre from the HSP81-1 promoter
in cotyledons would presumably lead to Cre-mediated
excision to produce detectable levels of GUS from the
constitutive 35S promoter. Second, the GUS activity
could arise from a cryptic promoter driving GUS expression in an unexcised pCrox construct. This latter
possibility was effectively ruled out by subsequent,
semi-quantitative PCR experiments (below).
Whatever the cause, this ectopic, uninduced expression apparently did not occur at significant levels
in meristematic cells, as true leaves which developed
later did not express significant GUS activity. This
suggests that induced lines did not express Cre recombinase in the meristem. To examine this more closely,
GUS expression was examined one and three weeks
after heat treatments of 8-week old plants (lines 42 and
69 in Figure 3f-i). This showed that GUS expression
was confined to tissues already developed after eight
weeks of growth at the stage when the heat treatment
was applied.
Figure 4. Northern blot analysis of 4 pCrox-GUS lines. RNA
(15 µg) was isolated from uninduced (−) and induced (+) shoots of
2–3-week old plants of the lines as numbered. The same filter was
probed with the BAR gene (BAR) and with the GUS gene (GUS).
The ethidium bromide stained gel is shown below (Gel). The signals quantified with a CCD camera are presented with white bars
(uninduced) and grey bars (induced).
Effect of heat treatment on BAR and GUS RNA levels
Northern blot analysis was performed to examine the
effects of heat induction on GUS and BAR mRNA
levels in lines 2, 42, 65 and 81 (Figure 4). As expected, the BAR mRNA encoded on the pCrox block
was abundant in uninduced plants of all 4 lines (Figure 4, lanes 1, 3, 5 and 7). After heat treatment, BAR
mRNA levels were undetectable in line 2 (Figure 4,
lane 2) and significantly reduced in lines 42, 65 and
81 (Figure 4, lanes 4, 6 and 8). Hybridization of the
same blot with the GUS probe detected high levels of
GUS mRNA in uninduced samples of line 2 (Figure 4,
lane 1), but none or very little GUS mRNA in uninduced samples of lines 42, 65 and 81 (Figure 4, lanes
3, 5 and 7). After heat treatment, little change in GUS
mRNA levels was seen in line 2 (Figure 4, lane 2),
while GUS mRNA levels increased in lines 42, 65 and
81 (Figure 4, lanes 4, 6 and 8). These results indicate
that heat treatment caused excision of the block carrying the BAR gene and led to an increase in GUS
expression due to the proximity of the 35S promoter
(Figure 1).
47
Figure 5. Estimation of in planta levels of BAR DNA in 4
pCrox-Gus lines by semi-quantitative PCR and Southern blotting.
DNA from uninduced cotyledons, uninduced leaves, and induced
leaves were used to PCR amplify an internal BAR fragment (Bar)
and a control gene (Lfy). Following electrophoresis, Southern blotting with the Bar and Lfy probes yielded the autoradiograph shown
at the top. The signals were quantified and are presented below as
relative Bar/Lfy signals. White bars represent cotyledons, light grey
bars uninduced leaves and dark grey bars induced leaves.
pCrox structure in planta estimated by
semi-quantitative PCR and Southern blotting
Semi-quantitative PCR and Southern blotting was
used to estimate the in planta levels of BAR DNA
in pCrox relative to an endogenous, control gene.
This enabled an estimate of the extent of excision
of the block containing the BAR gene in the four
selected lines before and after heat treatment (Figure 5). As might be expected, line 2, which had high
background GUS levels in cotyledons and leaves of
untreated plants, had no detectable BAR product from
uninduced cotyledons or in induced leaves. This suggests that the block was excised in the majority of
the cells in these tissues of line 2. Nonetheless, since
the BAR product could be amplified from uninduced
leaves of line 2, the block was not completely excised
in this tissue. In DNA of lines 42, 65 and 81, relatively low levels of BAR product were detected in
uninduced cotyledons, with higher product levels in
uninduced plants and reduced BAR product levels in
induced plants.
Two conclusions may be drawn from this experiment. First, the lower levels of BAR products in
uninduced cotyledons compared to uninduced leaves
indicates that GUS expression in untreated cotyledons
was due to excision of the block, and not from leaky or
Figure 6. Southern blot analysis of 4 pCrox-GUS lines and a
wild-type (Wt) Arabidopsis. Genomic DNA (2 µg) from each line
was digested with EcoRV and a part of the nptII gene was used
as probe. Localisation of the EcoRV sites and the nptII probe in
pCrox-GUS is indicated in Figure 1. Band a of 2.8 kb indicates a
presumptive head-to-head insertion, while band b of 2.6 kb indicates
a head-to-tail insertion.
cryptic promoter activity driving GUS in the absence
of excision in cotyledons. Second, the relative BAR
levels in the different lines indicate that heat treatment
mediates excision of the block containing the BAR
gene in at least a portion of the cells and/or a portion
of the T-DNA inserts.
Further evaluation of the frequency of excision in
the pCrox lines was complicated by the finding that
the four selected lines contained multiple T-DNA repeat insertions (Figure 6). Southern blotting of DNA
digested with EcoRV and probed with a left-border
probe (Figure 1) identified both multiple head-to-head
and head-to-tail insertions which would, respectively,
yield the 2.8 kb and 2.6 kb bands seen in the four
lines (Figure 6, lanes 1–4). For example, line 2 appears to contain three copies of the insert that have
a head-to-head, tail-to-head configuration (Figure 6).
The organisation of T-DNA copies in the other three
lines was more complex. Although some of the bands
may be due to partial EcoRV digestion, similar patterns were obtained on several Southern blots with two
different sets of genomic DNA. This and the fact that
the selectable markers in all four lines segregated as
single loci suggests that pCrox insertion in them involved complex DNA rearrangements as is often seen
with T-DNA (Nacry et al., 1998).
48
Discussion
The nine transgenic lines, shown by segregation to
contain single pCrox T-DNA insertion sites, all exhibited heat-inducible increases in GUS reporter activity.
In four lines examined further, northern analysis and
semi-quantitative PCR showed that heat treatment resulted in excision of a DNA region (the block) between
the pCrox lox sites. This excision is apparently mediated by the Cre-recombinase encoded in the block,
and results in switching transcription driven by the
35S promoter from the BAR gene in the block to the
GUS reporter outside the lox sites. For three of the four
lines analysed in detail, this induction was very clear
at the histochemical level. These results indicate that
the basic pCrox vector design functions in planta.
The HSP81-1 promoter element was chosen as an
initial test of the pCrox system as this promoter has
been shown to be tightly regulated and because heat
treatments are easy to perform. This proved to be
the case in leaves in which GUS induction was seen
histochemically after a single 6 h heat treatment at
37 ◦ C. Stronger inductions were obtained using two
heat treatments of 16 h with a recovery period of
24 h between. Expression levels of a gene of interest can therefore be regulated by varying the length
and number of heat treatments. One potential concern
with heat treatment is that it induces the expression
of endogenous heat shock-responsive genes (Schöffl
et al., 1992). While this could interfere with phenotypic analyses, expression of the endogenous HSPs
would be transient while expression of a gene of interest would continue in induced tissues. We note
that we did not see long-term, adverse effects of heat
treatment on any of the plants examined. Nonetheless, the pCrox system might be improved by the
use of another inducible promoter, including chemically inducible ones (Aoyama and Chua, 1997; Salter
et al., 1998). Although these systems have met with
marginal success as stand-alone inducible systems in
Arabidopsis, they may be quite effective for short-term
activation of Cre-recombinase in a pCrox-like system.
Two other potential problems with the pCrox system
may be noted. First, block excision and resultant GUS
expression was apparently leaky in cotyledons, despite
the lack of evidence from other studies that the heatresponsive HSP81-1 promoter is active in cotyledons
at lower temperatures (Yabe et al., 1994). This activity may be explained by transcription of cre from
a cryptic promoter specifically in cotyledons, or by
an increased efficiency in cells of cotyledons of Cre-
recombinase expressed at basal, ‘leaky’ levels from
the HSP81-1 promoter. How big a problem this would
be in reverse genetic studies with other genes of interest remains to be determined. We note, however,
that the pCrox GUS reporter vector should be a very
sensitive monitor of basal promoter activity because
Cre expression from pCrox mediates a permanent genetic change that results in expression of the stable
GUS protein from the strong 35S promoter. Furthermore, we suspect that moderate expression levels of a
deleterious transgene only in cotyledons would not be
lethal to plantlets grown on nutritive media given that
the transgene would not be expressed in the meristem
and subsequent, true leaves. We note also that in three
of the lines examined with low basal levels of GUS,
the cell-autonomous action of Cre results in sectors
of GUS activity. Such chimerism may prove useful in
determining whether the activity of a gene of interest
is cell-autonomous.
A second difficulty with the system was apparent
in lines containing multiple T-DNA copies. Since Cre
activation may not lead to block excision in all copies,
some copies may retain the BAR resistance marker.
This is despite the introduction of a nuclear localisation signal into the Cre-recombinase in an attempt to
boost its activity (Varagona and Raikhel, 1994). As a
result, the pCrox BAR gene functions inefficiently as a
selectable marker against Cre-mediated recombination
as originally designed. This problem would be negligible in lines containing a single T-DNA copy. Such
lines should be obtainable by lowering the efficiency
of the vacuum infiltration protocol and the use of less
virulent Agrobacteria strains or helper plasmids (Hood
et al., 1993).
The pCrox system may have potential advantages
over other inducible systems. For example, pCrox induction results in a permanent genetic change such
that the induced gene is permanently expressed in
cells and their daughters. In other systems, expression is only triggered when the inducing agent is
applied and present. For example, chemically induced
systems may require repeated applications to identify the effect of a gene of interest on developmental
events. Chemical delivery may also be complicated
by various factors including stability and transpiration flow. Compared to other recombinase-activated
systems for plants, pCrox may also have the advantage of containing all necessary elements in a single
vector. For example, other recombinase systems have
required progeny analysis of crosses between Crerecombinase-expressing plants with plants containing
49
a lox-based target gene. In addition, although we have
only tested the pCrox system in Arabidopsis, it may
also be useful in other species since Cre-lox recombination functions in plants such as tobacco (Bayley
et al., 1992), tomato (Stuurman et al., 1996), petunia
(Que et al., 1998) and wheat (Srivistava et al., 1999).
Acknowledgements
This work was supported by a grant from the Danish Agricultural and Veterinary Research Council
(9502284) and Biotechnology 3 Program (9502002).
References
Aoyama, T. and Chua, N.H. 1997. A glucocorticoid-mediated transcriptional induction system in transgenic plants. Plant J. 11:
605–612.
Bayley, C.C., Morgan, M., Dale, E.C. and Ow, D.W. 1992. Exchange of gene activity in transgenic plants catalyzed by the
Cre-lox site-specific recombination system. Plant Mol. Biol. 18:
353–361.
Bechtold, N., Ellis, J. and Pelletier, G. 1993. In planta Agrobacterium mediated gene transfer by infiltration of adult Arabidopsis
thaliana plants. C.R. Acad. Sci. Paris, Life Sci. 316: 1194–1199.
Bradford, M. 1976. A rapid and sensitive method for the quantitation of microgram quantities of protein utilizing the principle of
protein-dye-binding. Anal. Biochem. 72: 248.
Coles, J.P., Phillips, A.L., Croker, S.J., Garcia-Lepe, R., Lewis,
M.J. and Hedden, P. 1999. Modification of gibberellin production and plant development in Arabidopsis by sense and antisense
expression of gibberellin 20-oxidase genes. Plant J. 17: 547–56.
Elledge, S.E., Mulligan, J.T., Ramer, S.W., Spottswood, M. and
Davis, R.W. 1991. Lambda YES: a multifunctional cDNA expression vector for the isolation of genes by complementation of
yeast and Escherichia coli mutations. Proc. Natl. Acad. Sci. USA
88: 1731–1735.
Gatz, C. and Lenk, I. 1998. Promoters that respond to chemical
inducers. Trends Plant Sci 3: 353–358.
Hajdukiewicz, P., Svab Z. and Maliga, P. 1994. The small, versatile pPZP family of Agrobacterium binary vectors for plant
transformation. Plant Mol. Biol. 25: 989–994.
Hoff, T., Schnorr, K.M., Meyer, C. and Caboche, M. 1995. Isolation
of two Arabidopsis cDNAs involved in early steps of molybdenum cofactor biosynthesis by functional complementation of
Escherichia coli mutants. J. Biol. Chem. 270: 6100–6107.
Hood, E.E., Gelvin, S.B., Melchers, L.S. and Hoekema, A. 1993.
New Agrobacterium helper plasmids for gene transfer to plants.
Transgenic Res. 2: 208–218.
Huang, N.C., Liu, K.H., Lo, H.J. and Tsay, Y.F. 1999. Cloning
and functional characterization of an Arabidopsis nitrate trans-
porter gene that encodes a constitutive component of low-affinity
uptake. Plant Cell 11: 1381–1392.
Jones, J.D., Shlumukov, L., Carland, F., English, J., Scofield,
S.R., Bishop, G.J. and Harrison K. 1992. Effective vectors for
transformation, expression of heterologous genes, and assaying transposon excision in transgenic plants. Transgenic Res. 1:
285–297.
Kang, H.G., Fang, Y. and Singh, K.B. 1999. A glucocorticoidinducible transcription system causes severe growth defects in
Arabidopsis and induces defence-related genes. Plant J. 20:
127–133.
Morelii, G., Nagy, F., Fraley, R.T., Rogers, S.G. and Chua,
N.-H. 1985. A short conserved sequence is involved in the
light- inducibility of a gene encoding ribulose 1,5-bisphosphate
carboxylase small subunit of pea. Nature 315: 200–204.
Nacry, P., Camilleri, C., Courtial, B., Caboche, M. and Bouchez D.
1998. Major chromosomal rearrangements induced by T-DNA
transformation in Arabidopsis. Genetics 149: 641–650.
Que, Q., Wang, H.-Y. and Jorgensen, R.A. 1998. Distinct patterns
of pigment suppression are produced by allelic sense and antisense chalcone synthase transgenes in petunia flowers. Plant J.
13: 401–409.
Salter, M.G., Paine, J.A., Riddell, K.V., Jepson, I., Greenland,
A.J., Caddick, M.X. and Tomsett, A.B. 1998. Characterisation of
the ethanol-inducible alc gene expression system for transgenic
plants. Plant J. 16: 127–132.
Schöffl, F., Diedring, V., Kliem, M., Rieping, M., Schröder, G.
and Severin, K. 1992. The heat shock response in transgenic
plants: the use of chimaeric heat shock genes. In: J.L.Wray (Ed.)
Inducible Plant Proteins, Society for Experimental Biology Seminars Series 49, Cambridge University Press, Cambridge, UK,
pp. 247–266.
Srivastava, V., Anderson O.D. and Ow, D.W. 1999. Single-copy
transgenic wheat generated through the resolution of complex integration patterns. Proc. Natl. Acad. Sci. USA 96: 11117–11121.
Sternberg, N. and Hamilton, D. 1981. Bacteriophage P1 site-specific
recombination between loxP sites. J. Mol. Biol. 150: 467–486.
Stuurman, J., de Vroomen M.J., Nijkamp, H.J.J. and van Haaren
M.J.J. 1996. Single-site manipulation of tomato chromosomes in
vitro and in vivo using Cre-lox site-specific recombination. Plant
Mol. Biol. 32: 901–913.
Takahashi, T., Naito S. and Komeda Y. 1992. Isolation and analysis
of the expression of two genes for the 81-kilodalton heat-shock
proteins from Arabidopsis. Plant Physiol. 99: 383–390.
Taylor, L.A. and Rose, R.E. 1988. A correction in the nucleotide
sequence of the Tn903 kanamycin resistance determinant in
pUC4K. Nucl. Acids Res. 6: 7762.
Varagona, M.J. and Raikhel, N.V. 1994. The basic domain in
the bZIP regulatory protein Opaque2 serves two independent
functions: DNA binding and nuclear localization. Plant J. 5:
207–214.
Weigel, D., Alavarez, J., Smyth, D.R., Yanofsky, M.F. and
Meyerowitz, E.M. 1992. LEAFY controls floral meristem identity in Arabidopsis. Cell 69: 843–859.
Yabe, N., Takahashi, T. and Komeda, Y. 1994. Analysis of tissuespecific expression of Arabidopsis thaliana HSP90-family gene
HSP81. Plant Cell Physiol. 35: 1207–1219.