Counterselection and Co-Delivery of Transposon and
Transposase Functions for Sleeping Beauty-Mediated
Transposition in Cultured Mammalian Cells
Andrea D. Converse,1Lalitha R. Belur,1Jennifer L. Gori,1Geyi Liu,1Felipe Amaya,2
Estuardo Aguilar-Cordova,2,3Perry B. Hackett,1,4and R. Scott McIvor1,5
Sleeping Beauty (SB) is a gene-insertion system reconstructed from transposon sequences
found in teleost fish and is capable of mediating the transposition of DNA sequences from
transfected plasmids into the chromosomes of vertebrate cell populations. The SB system
consists of a transposon, made up of a gene of interest flanked by transposon inverted
repeats, and a source of transposase. Here we carried out a series of studies to further
characterize SB-mediated transposition as a tool for gene transfer to chromosomes and
ultimately for human gene therapy. Transfection of mouse 3T3 cells, HeLa cells, and human
A549 lung carcinoma cells with a transposon containing the neomycin phosphotransferase
(NEO) gene resulted in a several-fold increase in drug-resistant colony formation when co-
transfected with a plasmid expressing the SB transposase. A transposon containing a
methotrexate-resistant dihydrofolate reductase gene was also found to confer an increased
frequency of methotrexate-resistant colony formation when co-transfected with SB trans-
posase-encoding plasmid. A plasmid containing a herpes simplex virus thymidine kinase
gene as well as a transposon containing a NEO gene was used for counterselection against
random recombinants (NEO+TK+) in medium containing G418 plus ganciclovir. Effective
counterselection required a recovery period of 5 days after transfection before shifting into
medium containing ganciclovir to allow time for transiently expressed thymidine kinase
activity to subside in cells not stably transfected. Southern analysis of clonal isolates indi-
cated a shift from random recombination events toward transposition events when clones
were isolated in medium containing ganciclovir as well as G418. We found that including
both transposon and transposase functions on the same plasmid substantially increased the
stable gene transfer frequency in Huh7 human hepatoma cells. The results from these
experiments contribute technical and conceptual insight into the process of transposition in
mammalian cells, and into the optimal provision of transposon and transposase functions
that may be applicable to gene therapy studies.
KEY WORDS: Transposon; Sleeping Beauty; gene transfer; gene therapy.
1Gene Therapy Program and the Beckman Center for Transposon Research, Institute of Human Genetics,
Department of Genetics, Cell Biology and Development, University of Minnesota, 6-160 Jackson Hall
321 Church Street S.E., Minneapolis, MN, 55455, USA.
2Advantagene, Inc, San Diego, CA,USA.
3Harvard Gene Therapy Initiative, Harvard Medical School, Cambridge, MA, USA.
4Discovery Genomics, Inc., Minneapolis, MN, USA.
5To whom correspondence should be addressed. E-mail: email@example.com
Bioscience Reports, Vol. 24, No. 6, December 2004 (? 2005)
0144-8463/04/1200-0577/0 ? 2005 Springer Science+Business Media, Inc.
Long-term expression of gene sequences newly introduced into mammalian cells
requires a strategy for maintaining the presence of these sequences inside the cell.
For some cell and tissue populations, such as muscle, newly introduced plasmid
sequences are maintained intracellularly for as long as 1 year (Wolff et al., 1990,
1992). Viral or cellular replicative origins can also confer extrachromosomal main-
tenance of newly introduced sequences (Calos, 1996). However, stable chromosomal
integration of the newly introduced sequence provides the most reliable genetic
outcome for long-term expression. Moreover, the newly introduced gene is then
reliably transmitted to all subsequently-generated cell progeny. Stable integration
can be readily achieved through the use of retroviral and lentiviral vectors, which
harness the integration step after reverse transcription in the viral replicative cycle
(Miller, 1992; Kay et al., 2001). Adeno-associated virus vectors also mediate inte-
gration, although the mechanism by which this occurs has not yet been characterized
in detail (Carter and Samulski, 2000). In contrast, non-viral gene insertion has relied
primarily on integration via random recombination to achieve stable, long-term
A newly established, non-viral approach to gene insertion in mammalian sys-
tems employs the use of transposons. In the cut-and-paste mechanism of transpo-
sition (Fig. 1), a transposase enzyme recognizes short transposon termini that flank
the sequence of interest, excising the transposon from its pre-existing site and
inserting the transposon at some other chromosomal site. Significant transposition
of sequences from transfected plasmids into vertebrate cellular chromosomes was
not demonstrated until reconstruction of the Sleeping Beauty transposase from a
defunct transposon found in teleost fish resembling the Tc1/Mariner transposon
family first identified in C. elegans (Ivics et al., 1997). Sleeping Beauty thus consti-
tutes a strategy for insertion of defined DNA sequences into the chromosomes of
target cells and tissues for cellular engineering and gene therapy.
Fig. 1. Transposition mediated by the Sleeping Beauty
transposase. A transposon containing a gene of interest
flanked by inverted repeats (boxes with dark arrows) is
recognized at the transposon termini by the Sleeping
Beauty transposase (SB), which excises the transposon
from the plasmid in which it was presented, and then
catalyzes insertion of the transposon into a site in the
chromosomes of the target cell.
578 Converse et al.
Since its discovery, Sleeping Beauty has been demonstrated to mediate trans-
position in a variety of different vertebrate cell and tissue types (Izsvak et al., 2000),
including mouse embryos (Dupuy et al., 2002) and germ cells (Luo et al., 1998;
Dupuy et al., 2001; Fischer et al., 2001), and in mouse liver (Yant et al., 2000; Nakai
et al., 2001) and lung (Belur et al., 2003). Here we address several key questions
regarding Sleeping Beauty-mediated transposition: (i) What are the differences in
transposition between several cell lines of differing tissue origin? (ii) Under what
conditions can the HSV thymidine kinase gene be used as a counterselectable marker
in SB transfection studies? (iii) What is the relative effectiveness of transposition
when transposon and transposase functions are delivered as a part of the same
plasmid, thus providing both components in the same ratio? Results from these
studies provide several important tools for the application of SB-mediated trans-
position in more complex cellular targets, such as animal and human tissues in
preclinical and, ultimately, clinical gene therapy studies.
MATERIALS AND METHODS
Mammalian Cell Lines and Culture
Hela, NIH 3T3 (ATCC CRL)1658), A549 (ATCC CCL)185), and Huh7 (Pugh
et al., 1988), were routinely cultured in Dulbecco’s-modified Eagle medium supple-
mented with 10% fetal bovine serum (Summit Biotechnology), 2 mM glutamine and
100· antibiotic/antimycotic agent (10,000 U/ml penicillin, 10,000 U/ml streptomy-
cin, and 25 lg/ml Fungizone; Gibco BRL) at 37?C and 5% CO2in a humidified
tissue culture incubator. Selective conditions (see below) consisted of 400 lg/ml
G418 (Cellgro) for NIH3T3 cells, 700 lg/ml G418 for A549 and Huh7 cells, 5 lg/ml
Cytovene-IV (ganciclovir; Roche Laboratories, Inc.) and 200 nM methotrexate
Plasmid pT/neo, described by Ivics et al. (1997), contains the SV40 promoter/
enhancer transcriptionally regulating the neomycin phosphotransferase (NEO) gene
with a SV40 poly(A) signal, all flanked by transposon inverted repeat sequences
(Fig. 2). pT/neo/HSV-TK, constructed by blunt-end insertion of the herpes simplex
virus thymidine kinase gene into the AflIII site of pT/neo 397 bases downstream of
the NEO transposon cassette, was kindly provided by Izsvak et al. (2000). pT/BAD
contains a DHFR transcription unit consisting of the chicken ß-actin promoter, the
L22Y variant murine DHFR cDNA (Morris and McIvor, 1994) and a region of the
hepatitis B virus surface antigen gene spanning the polyadenylation signal (Morris
et al., 1996), all flanked by IR sequences. A 450 bp XhoI--SacI fragment containing
the ß-actin promoter and the 5¢ end of the DHFR gene isolated from pLBD (Zhao
et al., 1997) was ligated with a 900 bp SacI--SalI fragment containing the 3¢ end of
the DHFR cDNA and the HBV 3¢ UTR sequences (Morris et al., 1996) into
pBluescript between XhoI and SalI to form pßAD. The DHFR expression cassette
was then excised with XhoI (blunt) and EcoRI and ligated into the transposon
sequences of pTBH (Geurts et al., 2003) between HindIII (blunt) and EcoRI to form
579 Sleeping Beauty-Mediated Transposition
pT/ßAD (Fig. 2). pCMV(SB10), containing the CMV early promoter regulating the
SB10 transposase gene, was previously described as pSB10 (Ivics et al., 1997)
(Fig. 2). pT/neo/CMV(SB10) (Fig. 2) was constructed by insertion of a 2100 bp
EcoRI--SalI CMV-SB10 fragment and a 2250 bp XhoI--SalI T/neo fragment into
pLPBL1 (a derivative of pGEM7, Promega).
All adherent cell line transfections were carried out using the DNA-calcium
phosphate co-precipitation technique (Wigler et al., 1979) as previously described
(McIvor and Simonsen, 1990). Briefly, cells were seeded the day prior to transfection
at 2.5--5 · 105per 60 mm dish. DNA-calcium phosphate co-precipitate was pre-
pared by adding DNA (10 lg total per transfection) in 2 M CaCl2dropwise to
HEPES buffered saline, pH 7.2, containing 1.15 mM sodium phosphate, allowing
the precipitate to form for approximately 5 min before adding dropwise to recipient
Fig. 2. Transposon and transposase expression constructs.
Coding sequences are indicated by dark boxes; SB, Sleeping
Beauty transposase; NEO, neomycin phosphotransferase;
DHFR, murine L22Y dihydrofolate reductase variant. TK,
herpes simplex virus thymidine kinase, including the viral
regulatory elements. Striped boxes with dark arrows; trans-
poson inverted terminal repeat sequences. Open boxes repre-
sent the following transcriptional regulatory elements; (1) pA,
bovine growth hormone polyadenylation signal; (2) HBV-pA,
hepatitis B virus surface antigen gene including its 3¢ untrans-
lated region and polyadenylation signal; (3) cmv, promoter
from the cytomegalovirus early region; (4) sv40, promoter from
the simian virus 40 early region; and (5) ß-ac, promoter from
the chicken ß-actin gene.
580Converse et al.
cell cultures. Plasmids were purified by banding on CsCl twice. Within 3--24 h of
adding the precipitate to the cells, they were shocked for three minutes with 15%
glycerol in phosphate-buffered saline (PBS), washed twice, and provided fresh
growth medium. Two to four days post-transfection the cells were plated into
selective medium, and drug-resistant colonies were counted after 10--14 days. Re-
sults are expressed as the total number of drug-resistant colony-forming units gen-
erated in the originally transfected plate. The percentage of cells stably transfected
was calculated from the total number of cells giving rise to drug-resistant colonies
divided by the number of cells recovered post-transfection (approximately 2 · 106)
multiplied by 100.
Southern Hybridization Analysis
To distinguish between transposition and random recombination of stable
integrants, Southern analysis was carried out on DNA samples isolated from clonal
cell populations as previously described (Jonsson et al., 1995). Briefly, cells were
lysed in 0.5% sodium dodecyl sulfate/0.17 mg/ml Proteinase K solution and DNA
extracted with phenol/chloroform. DNA samples were digested with SacI and SalI
according to the manufacturer’s recommendation and electrophoresed through 1%
agarose in tris-acetate buffer (Sambrook et al., 1989). Gels were blotted onto Nytran
(Osmonics) and prehybridized in a solution containing 6· SSC (SSC is 175 g/l NaCl,
88.2 g/l sodium citrate, pH 7.0), 50% formamide, 5· Denhardt’s solution
(Sambrook et al., 1989), and 100 lg/ml sheared, denatured salmon sperm DNA.
Blots were probed with a 619 bp NcoI--HindIII NEO fragment isolated from plas-
mid pLXSN (Miller and Rosman, 1989) and radiolabelled with a)32P-dCTP by
random priming (Prima-a-Gene kit, Promega). After hybridization overnight, the
blots were washed twice at room temperature for 15 min and then twice for 30 min
each time at 50?C, with subsequent exposure of X-ray film at )80?C using an
Sleeping Beauty-mediated Transposition of Drug-resistance Genes in Cultured
The Sleeping Beauty transposon system consists of two components (Fig. 1).
The first component is a plasmid or other source of DNA containing a transposon,
which includes a sequence of interest flanked by inverted repeats that define the
termini of the transposon. The second component is SB transposase, which is usually
supplied by a plasmid that encodes SB transposase, although RNA has been used as
well (Izsvak et al., 1997; Dupuy et al., 2002; Wilber et al., submitted). The trans-
posase can excise the transposon at the ends of the IRs and insert the transposon at
some chromosomal site (Fig. 1). SB has previously been demonstrated to mediate
stable gene transfer in a variety of different cell types (Ivics et al., 1997; Izsvak et al.,
2000). Here we tested SB-mediated gene transfer into several different cell lines
representative of key target tissues for gene therapy. The plasmid pT/neo, containing
an SV40-regulated neomycin phosphotransferase gene (Fig. 2), was transfected with
581Sleeping Beauty-Mediated Transposition
or without the SB transposase-encoding plasmid pCMV(SB10) (Fig. 2) into mouse
3T3 cells, human HeLa cells, human A549 lung carcinoma cells, and human Huh7
liver hepatocarcinoma cells (see below). The transfected cells were then plated in
medium containing G418 to determine the frequency of drug-resistant colony for-
mation. The extent to which co-transfection with pCMV(SB10) causes an increase in
the frequency of G418)resistant colony formation provides an assessment of the
frequency of transposition (Ivics et al., 1997; Izsvak et al., 2000; Fischer et al., 2001;
Cui et al., 2002). In mouse 3T3 cells, we consistently observed a 3)fold increase in
pCMV(SB10) (Fig. 3b), for an estimated transposition frequency of 0.15%. Hela
cells were characterized by a much lower baseline transfection frequency of 0.025%
as defined by transfection with pT/neo alone (Fig. 3a). But, when co-transfected with
G418-resistant colonies per plate
0 20004000 6000 8000
0 2000 40006000 8000
Fig. 3. Transposition frequencies in various lines of cultured mam-
malian cells. (a) HeLa cells, (b) murine NIH 3T3 cells, and (c) human
A549 cells. Cells were transfected with 5 lg of pT/neo and with 5 lg
of pCMV(SB10) or with 5 lg of pCAT as a control. Transfected cells
were subcultured 2 days post-transfection into medium containing
0.4 (a) or 0.7 (b and c) mg/ml G418, and colonies enumerated after
2 weeks of selection. The total number of G418-resistant colonies
generated per transfected plate is shown.
582 Converse et al.
pCMV(SB10), a 10-fold increase in G418-resistant colony formation was observed,
for a transposition frequency of 0.25%. A low baseline pT/neo transfection fre-
quency of 0.025% was also observed for A549 cells (Fig. 3c). Co-transfection of
A549 cells with pT/neo plus pCMV(SB10) resulted in a 4-fold increase in G418-
resistant colony-formation, for a transposition frequency of 0.1%. We thus observed
an estimated transposition frequency of 0.1--0.25% in these three cell lines.
For molecular characterization of transposition events, several of the G)418-
resistant A549 clones were harvested and expanded in culture for DNA isolation and
Southern hybridization analysis (Fig. 4). Test samples were digested with SalI+
SacI, enzymes that cut just outside of the transposon in pT/neo (see Fig. 2), and then
probed for NEO-hybridizing sequences. Random recombination of plasmid
sequences should mostly result in the maintenance of a 2.3 kb NEO-hybridizing
SacI/SalI fragment (as seen for the positive plasmid controls in Fig. 4), although
random recombination proximal to either restriction site could result in a differently-
sized NEO-hybridizing fragment as well. All transposition events will result in an
increase in the size of the NEO-hybridizing SacI/SalI fragment. We observed that
only one (number 14) of seven clones isolated after co-transfection of pT/neo with
pCAT control plasmid exhibited a NEO-hybridizing fragment of increased size,
while six (clones 1,3,4,5,7 and 8) of nine clones co-transfected with pT/neo and
pCMV(SB10) exhibited an increase in the size of the NEO-hybridizing SacI/SalI
fragment. These results are consistent with transposition as the genetic means by
which these A549 clones acquired stable resistance to G418.
Mouse 3T3 cells were also used to test pT/ßAD, a transposon designed for
expression of drug-resistant dihydrofolate reductase (DHFR) activity (Fig. 2)
conferring resistance to methotrexate (Morris and McIvor, 1994). As for pT/neo,
cotransfection of 3T3 cells with pCMV(SB10) + pT/ßAD resulted in a 4-fold in-
crease in drug-resistant colony formation in comparison with pT/ßAD + pCAT,
yielding a transposition frequency of about 0.15% (Fig. 5). No increase in the fre-
quency of methotrexate-resistant colony formation over background was observed
when pCMV(SB10) was co-transfected with pßAD, a DHFR expression plasmid
lacking IR sequences. These results demonstrate the effectiveness of SB in mediating
transposition of DHFR as a selectable marker, and the necessity of flanking IR
sequences for this process. Sleeping Beauty thus mediated increased gene transfer
frequency of NEO and DHFR drug resistance genes in multiple cell lines.
Counterselection Against Random Recombination Favors Transposition Events
The standard method for unequivocal demonstration of transposition is to
examine junction sequences of the integration site (Ivics et al., 1997; Izsvak et al.,
2000; Yant et al., 2000; Cui et al., 2002; Dupuy et al., 2002). Isolation and sequencing
of these junctions is arduous. Consequently, we devised a positive-- negative selection
process to distinguish transposition from random recombination events. We used the
plasmid pT/neo/HSV-TK, which contains the herpes simplex virus thymidine kinase
gene inserted into pT/neo outside of the NEO transposon (Fig. 2) to test this strategy.
After transfection with pT/neo/HSV-TK, random recombination should yield colo-
nies that are G418-resistant but sensitive to ganciclovir. However, after co-transfec-
tion of pT/neo/HSV-TK with pCMV(SB10), transposition of the sequence away from
583Sleeping Beauty-Mediated Transposition
the HSV-TK sequence should yield colonies which are G418-resistant and ganciclovir
In the process of establishing this counterselective system, we found it necessary
to delay the imposition of ganciclovir selection due to transient expression of
Fig. 4. Southern hybridization analysis of G418-resistant A549 clones. Clones
were isolated after transfection and selection as described in the legend to Fig. 3.
Genomic DNA was isolated and subjected to Southern hybridization analysis as
described in Materials and Methods. DNA extracted from individual cell clones
was digested with SacI and SalI, enzymes which cut just outside of the IR’s of the
NEO transposon in pT/neo (see Fig. 2). In this analysis, Random recombination
events not interrupting the NEO transposon render a 2.3 kb NEO-hybridizing
SacI--SalI fragment (arrow), while transposition of the NEO transposon away
from the pT/neo plasmid to cellular chromosomal sites renders NEO-hybridizing
SacI--SalI fragments which are larger than 2.3 kb in length. pT/neo plasmid was
included as a positive control (P) for untransposed NEO-hybridizing material.
Clones 1 through 9 were isolated after cotransfection with pT/neo and
pCMV(SB10). Clones 11 through 18 were isolated after co-transfection with pT/
neo and pCAT. R; clones scoring for random recombination events (i.e., only the
2.3 kb NEO-hybridizing fragment observed). T; clones scoring for transposition
events (i.e. only NEO-hybridizing fragments larger than 2.3 kb observed). B; clones
scoring for both random recombination and transposition events.
584 Converse et al.
HSV-TK. In the experiment depicted in Fig. 6, mouse NIH 3T3 cells were trans-
fected with pT/neo/HSV-TK with or without pCMV(SB10). Two days post-trans-
fection, the cells were subcultured into replicate dishes containing selective medium
with G418. At various times thereafter, we switched selective conditions from G418
alone to G418 plus ganciclovir, and then stained for drug-resistant colonies 2 weeks
post-transfection. Figure 6 shows the proportion of G418-resistant colonies (i.e.,
total gene transfer frequency) that are resistant to both drugs as a function of the
time that selective conditions were switched from G418 only to G418 + ganciclovir.
When selectiveconditions were switched on
pCMV(SB10) had no effect on the proportion of colonies that were resistant to both
drugs. However, over the next few days the proportion of colonies generated by co-
transfection with pCMV(SB10) that were resistant to both drugs increased to 60%,
while the same proportion generated by transfection with pT/neo/HSV-TK alone
was only 25%. To the extent that the proportion of colonies resistant to both drugs
represents an estimate of the transposition frequency in cells co-transfected with
both pT/neo/HSV-TK and pCMV(SB10), this result (60%) is consistent with the 4-
fold increase in G418-resistant colony formation observed when either pT/neo
(Fig. 3) or pT/neo/HSV-TK (Fig. 6) are cotransfected with pCMV(SB10). Cells
transfected with pT/neo/HSV-TK alone that were resistant to both drugs must have
resulted from NEO gene integration in a way that excluded, interrupted or in some
other way impaired expression of the TK gene. Thus, not all random recombinants
can be eliminated in this counterselective procedure.
day1, co-transfection with
0 2000 40006000 8000
Mtx-resistant colonies per plate
Fig. 5. Transposition of a murine, drug-resistant L22Y DHFR gene
into murine NIH3T3 cells. Cells were transfected with 5 lg of pT/ßAD
(Fig. 2) or a plasmid containing a similar DHFR transcription unit but
without transposon inverted repeats (pßAD) along with 5 lg of
pCMV(SB10) or with 5 lg of pCAT as a control. Two days after
transfection, the cells were plated into medium containing 0.2 lM
methotrexate. Drug-resistant colonies were counted 2 weeks later. The
number of MTX-resistant colonies observed per transfected plate is
585 Sleeping Beauty-Mediated Transposition
The experiment described above indicates that counterselection of cells
co-transfected with pT/neo/HSV-TK+pCMV(SB10) results in a shift in scoring
from random recombination events to mainly transposition events. To verify the
validity of this procedure, we isolated individual drug-resistant clones and extracted
DNA for Southern hybridization analysis. Test samples were digested with SalI+
SacI, enzymes that cut just outside of the transposon in pT/neo/HSV-TK (see
Fig. 2), and then probed for NEO-hybridizing sequences. As described above for the
pT/neo transfected A549 clones, random integration of the transposon plus flanking
plasmid sequences should for the most part generate a 2.3 kb NEO-hybridizing
fragment, identical to that observed from control pT/neo/HSV-TK plasmid DNA
(P in Fig. 7a, d). Most clones isolated after transfection without pCMV(SB10)
contained the 2.3 kb NEO-hybridizing fragment indicative of random recombina-
tion, regardless of selection conditions (Fig. 7a, b). The existence of some fragments
larger than 2.3 kb most likely resulted from random recombination events that
interrupted the plasmid sequence between the NEO transcription unit and either the
SacI site or the SalI site. These integrants are thus indistinguishable from transpo-
sition events in this analysis. However, there was a greater tendency for the presence
of NEO-hybridizing fragments larger than 2.3 kb when pT/neo/HSV-TK was co-
transfected with pCMV(SB10), as seen in Fig. 7c, d. Furthermore, counterselection
with ganciclovir resulted in fewer clones containing random recombinants (clones
% (G418 + Ganc)
Fig. 6. Positive/negative selection of transposition
events in medium containing G418 and ganciclovir.
Mouse NIH 3T3 cells were transfected as indicated
with pT/neo/HSV-TK (Fig. 2) with or without
pCMV(SB10). Two days post-transfection, the cells
were subcultured into medium containing 0.4 mg/ml
G418. At various times thereafter, the selection
medium was switched in some of the plates from
medium containing G418 alone to medium contain-
ing G418 plus 5 lg/ml ganciclovir. After 2 weeks of
incubation under selective conditions, drug-resistant
colonies were enumerated, and the percentage of
G418-resistant colonies that were resistant to both
drugs was calculated.
586 Converse et al.
2,4,7,9 and 11 for a total of 5 out of 14 clones in Fig. 7c) in comparison with G418
selection alone (clones 1--5, 7, 10, 13 and 14 for a total of 9 out of 12 clones in
Fig. 7d). These results provide molecular evidence for the effectiveness of ganciclovir
counterselection to favor transposition over random recombination in pT/neo/HSV-
TK transfected cells.
Co-delivery of Transposon and Transposase Components on the Same Plasmid
Facilitates Gene Transfer in Huh7 Human Hepatocarcinoma Cells
In all of the experiments described above, transposase function was provided by
co-transfection of a separate plasmid containing an SB10 transposase gene. Under
these conditions, different ratios of transposon/transposase are delivered to different
cells in the transfected cell population. Transposition rates are sensitive to the ratio
of SB transposon/transposase: too little transposase or too much transposase can
inhibit transposition (Hartl et al., 1997; Geurts et al., 2003). One way to assure
coordinate delivery of both transposon and transposase components is to introduce
both as part of a single plasmid. We tested this approach in the human hepatocar-
cinoma cell line Huh7. Huh7 cells exhibited an extremely low transfection frequency
Fig. 7. Southern hybridization analysis of 3T3 cell clones isolated by positive or positive/negative
selection. NIH3T3 cell clones were isolated after co-transfection (as indicated) with subsequent
selection in medium containing G418 with or without ganciclovir as described in the legend to Fig.
6. DNA was isolated and subjected to Southern hybridization analysis after digestion with SacI
and SalI to identify random recombination vs. transposition events as described in the legend to
Fig. 4. Clones were isolated after co-transfection of pT/neo/HSV-TK with pCAT control plasmid
(a and b) or with pCMV(SB10) (c and d), with subsequent selection in medium containing either
G418 + Ganc (a and c) or G418 alone (b and d). pT/neo/HSV-TK plasmid DNA was included in a
and d as a positive control (P) for random recombination events. Clones were scored for random
recombination (R), transposition (T) or both (B) as described in the legend to Fig. 4.
587 Sleeping Beauty-Mediated Transposition
(less than 0.005%; Fig. 8), and interestingly co-transfection with pCMV(SB10) did
not bring about an increase in pT/neo-mediated G418-resistant colony formation. In
contrast, the frequency of drug-resistant colony-formation was increased 10-fold to
0.05% for Huh7 cells transfected with pT/neo/CMV(SB10), a plasmid containing
both the T/neo transposon as well as an SB transposase expression cassette.
WealsoconductedSouthern analysison G418-resistantHuh7clonesisolatedafter
transfection with pT/neo/CMV(SB10), pT/neo + pCMV(SB10), or pT/neo + pCAT
(Fig. 9). In contrast with our transposition results in A549 cells (Fig. 4) and 3T3 cells
(Fig. 7), Huh7 cells co-transfected with pT/neo plus pCMV(SB10) did not exhibit a
shift toward a greater number of NEO-hybridizing fragments larger than 2.3 kb (4 out
of 7 clones in Fig. 9b) in comparison with cells co-transfected with pT/neo plus pCAT
(6 out of 7 clones in Fig. 9a), perhaps because of the low gene transfer frequency in
these cell populations. The marked increase in gene transfer frequency observed for
Huh7 cells transfected with pT/neo/CMV(SB10) was associated with clones consisting
of an equal proportion of those exhibiting only the 2.3 kb NEO-hybridizing fragment
(7 out of 21 in Fig. 9c), those exhibiting only fragments larger than 2.3 kb (8 out of 21)
and those exhibiting both (6 out of 21). There were several clones (# 3, 15,16, 17, and
20) containing multiple integrants, which was not observed in Huh7 cells transfected
with pT/neo + pCAT or with pT/neo + pCMV(SB10). These results indicate that the
increased gene transfer efficiency brought about by introduction of both transposon
and transposase functions on the same plasmid was apparently associated with an
increase in both transposition and random recombination events (see Discussion).
10 100 100010000
Total G418-Resistant Colonies
Fig. 8. Effect of introducing NEO transposon sequences and SB
transposase sequences on the same plasmid into Huh7 cells. Huh7 cells
were transfected with 10 lg of pT/neo along with either 10 lg of
pCMV(SB10) or with 10 lg of pCAT, or they were transfected with
10 lg of pT/neo/CMV(SB10) + 10 lg of pCAT. Three to four days
post-transfection, the cells were subcultured into medium containing
0.7 mg/ml G418, enumerating drug-resistant colonies 2 weeks later and
calculating the total number of drug-resistant colony-forming units per
transfected plate. The data shown were pooled from three similarly
588Converse et al.
A series of studies was conducted to further characterize gene transfer mediated
by the Sleeping Beauty transposon system. The frequency of transposition varied
from 0.1 to 0.25% among the different cell lines tested, using the NEO gene and the
DHFR gene as selectable markers. Incubation in medium containing ganciclovir plus
G418 selected against random recombinants of a plasmid containing an HSV-TK
gene in addition to a NEO transposon, as determined by Southern hybridization
analysis. Introduction of transposon and transposase functions on the same plasmid
significantly improved transposition frequency in Huh7 cells. These results provide
Fig. 9. Southern hybridization analysis of Huh7 clones. Clones were isolated in
medium containing G418 after co-transfection of pT/neo/CMV(SB10) with pCAT
(C) or after co-transfection of pT/neo with either pCMV(SB10) (a) or with pCAT
(b) as described in the legend to Fig. 8. Genomic DNA was isolated and subjected to
Southern hybridization analysis after digestion with SacI and SalI to identify ran-
dom recombination vs. transposition events as described in the legend to Fig. 4. pT/
neo plasmid DNA was included as a positive control (P) for random recombination
events. The arrow shows the position of a 2.3 kb neo-hybridizing fragment gener-
ated from pT/neo (P) or pT/neo/CMV(SB10) plasmid, or by random recombina-
tion. Clones were scored for random recombination (R), transposition (T) or both
(B) as described in the legend to Fig. 4. The NEO-hybridizing fragments smaller
than 2.3 kb observed for clone 6 in part A and clones 3 and 20 in part C were most
likely generated by random recombination.
589 Sleeping Beauty-Mediated Transposition
several key insights and tools for analysis of SB-mediated transposition that are
important for application to somatic cell genetic studies and potentially for gene
Transposons have been used extensively for the introduction and expression of
new genes as well as for insertional mutagenesis in several model organisms, most
notably Drosophila (P-elements) and C. elegans (Tc1/Mariner) (Plasterk, 1993).
Although sequences resembling those of cut-and-paste type transposons are com-
mon in vertebrate genomes, transposition at quantitatively significant levels has not
been reported in vertebrate systems until recently, with the reconstruction of the
Sleeping Beauty transposase from telost fish sequences (Ivics et al., 1997). Tc1-type
and mariner-type transposons have been reported to mediate low level transposition
in cultured vertebrate cells, targeting both transfected sequences (Zhang et al., 1998)
as well as chromosomal sites (Li et al., 1998; Schouten et al., 1998). Further studies
of the SB transposon system has demonstrated its capability of mediating gene
insertion into cells from a variety of species (Izsvak et al., 2000), into mouse somatic
tissues (Yant et al., 2000; Nakai et al., 2001; Belur et al., 2003), and in the mouse
germ line (Fischer et al., 2001; Horie et al., 2001; Dupuy et al., 2002), including the
mobilization of resident transposon elements to new sites. The SB transposon system
thus has great potential utility for analyses benefiting from gene insertion into cul-
tured cells, somatic tissues, and the germ line of experimental animals. As an
effective, enzymatic way of integrating sequences into the chromosomes of somatic
tissues, SB also has great potential utility as a vector for gene therapy, using either
non-viral or viral delivery systems.
As in any gene transfer experiment, it may be anticipated that the frequency of
transposition will be limited by the efficiency of gene delivery, i.e. the maximum
transposition frequency is defined as the number of cells into which the transposon
and transposase functions have been delivered. Another key element in quantitating
transposition is distinguishing transposition events from a background level of
random recombination. In the transposition studies described in this paper and
elsewhere (Ivics et al., 1997; Izsvak et al., 2000; Fischer et al., 2001; Karsi et al.,
2001), assessment of the frequency of transposition is commonly made by comparing
the frequency of G418-resistant colony-formation after transfection with NEO
transposon plasmid with or without an SB transposase expression plasmid. The
effect of the expressed transposase is most apparent as a certain-fold increase in
overall gene transfer frequency. Transposase activity is thus most effectively
measured under conditions that optimize the -fold increase in gene transfer efficiency
(Cui et al., 2002; Geurts et al., 2003). However, to the extent that the increase in gene
transfer frequency brought about by co-transfection with transposase-encoding
plasmid is in fact the result of transposition, the frequency of transposition can be
calculated as the difference (rather than the ratio) in gene transfer frequency between
cells co-transfected vs. cells not co-transfected with transposase-encoding plasmid.
Thus, conditions that optimize the sensitivity of detecting transposase activity
(i.e. -fold increase in gene transfer efficiency over background) may differ from those
that maximize the frequency of gene transfer by transposition (i.e. calculated as the
difference between cell populations transfected with transposon + transposase-
encoding plasmid and those transfected without transposase-encoding plasmid).
590 Converse et al.
Expression from newly-introduced plasmid sequences may be the result of
sequence maintenance in extrachromosomal, episomal forms, random (illegitimate)
recombination, or, in the case of work presented here, by transposition of sequences
from the newly-introduced plasmid to host cell chromosomes. We utilized the genetic
segregation that occurs between transposon and plasmid sequences brought about
by cut-and-paste transposition to select against random recombination events and
thus better quantitate gene transposition frequency, employing the HSV-TK gene in
combination with ganciclovir as a negative selection marker. This counterselective
technique has been used extensively for the purpose of selecting for homologous
recombination events in embryonal stem cells (Mansour et al., 1988; Rubinstein
et al., 1993). This technique was for the most part an effective way to reduce
(although not eliminate) random recombinants generated as a part of the experi-
ment. We also found that it was necessary to delay imposition of ganciclovir
selection in this strategy: The fraction of total, G418-resistant cells that was also
resistant to ganciclovir was not maximized until cells were shifted to medium con-
taining both drugs approximately 5 days post-transfection. This sensitivity to
ganciclovir soon after transfection was most likely the result of transiently expressed
thymidine kinase activity, with approximately 5 days required to allow degradation
of transiently expressed TK enzyme. Successfully transposed cells, in which the NEO
gene on its own was stably integrated, would thus score a false negative for trans-
position (G418r, GANCs) when selective conditions are shifted to medium con-
taining both drugs prior to 5 days post-transfection. These results imply that in any
strategy employing the TK gene for negative selection after segregation from a
second sequence, such as that used for negative selection after homologous recom-
bination, recombinants will be underestimated (Izsvak et al., 2000) unless ganciclovir
selection is delayed until the activity of transiently expressed TK has subsided.
Transposase function can theoretically be provided in the form of protein or in
the form of RNA or DNA containing a transposase coding sequence. Although
purified forms of SB protein have been generated that exhibit transposase-like
character (DNA binding and nuclear localization (Ivics et al., 1997)), SB protein-
mediated transposition has not yet been reported. Transposase activity has been
provided by co-transfection either with a transposase-encoding plasmid, as described
here and elsewhere (Ivics et al., 1997; Izsvak et al., 2000; Yant et al., 2000), or with
transposase-encoding RNA generated by in vitro transcription (Izsvak et al., 1997;
Fadool et al., 1998; Dupuy et al., 2002; Wilber et al., submitted). We anticipate that
use of a plasmid containing both transposon and transposase functions will improve
the likelihood of localizing both functions to the nucleus of target cells, where the
transposase sequence may be transcribed for translation in the cytoplasm, and where
relocalized transposase may then excise the transposon, subsequently inserting it into
the chromosome. Co-delivery of transposon and transposase components on the
same plasmid has been demonstrated in mouse liver (Mikkelsen et al., 2003; Score
et al., submitted), and the effectiveness of this approach in our studies was apparent
for Huh7 cells, a human hepatoma cell line. We found this cell line difficult to
transfect in general, and increased gene transfer frequency was not observed when
transposon and transposase functions were delivered on separate plasmids (Fig. 8).
However, co-delivery of transposon and transposase functions on the same plasmid
provided a 10-fold increase in the frequency of stable gene transfer (G418 resistant
591 Sleeping Beauty-Mediated Transposition
colony-formation). The observed difference in transposition efficiency between Huh7
cells and the other tested cell lines may be related to the relatively low frequency of
stable gene transfer observed in general for Huh7 cells. Transit across the plasma
membrane and subsequent delivery to the nucleus have been characterized as rate-
limiting steps for gene transfer and expression in mammalian cells and tissues
(Nishikawa and Huang, 2001). If plasmid delivery is extremely limited for Huh7
cells, then delivery of both transposon and transposase components on separate
plasmids would be very unlikely, while including both functions on the same plasmid
would result in a substantial increase in the frequency of events in which both
functions are delivered. Interestingly, Southern analysis indicated that co-delivery of
transposon and transposase functions on the same plasmid resulted in an increase in
random recombination as well as transposition events in Huh7 cells (Fig. 9). The
DNA binding and nuclear localization activity of SB transposase could have pro-
vided increased delivery of transposon DNA to the nucleus for either transposition
or random recombination, although this possibility has yet to be addressed experi-
Overall, these results support use of the SB transposon system as a tool for gene
transfer in mammalian cells and, potentially, for gene therapy. The defined genetic
structure of transposon integrants (Ivics et al., 1997; Luo et al., 1998; Dupuy et al.,
2001; Horie et al., 2001; Cui et al., 2002) makes transposition an attractive alter-
native to random recombination in achieving stable, long-term gene expression.
Gene expression may be more reliable from one integrant to the next, and the
presence and effect of sequences flanking integrated transposons can more readily be
analyzed for identification of new genes by insertional mutagenesis or expression
trapping strategies (Clark et al., 2004). The effectiveness of these strategies may be
more readily assessed when carried out in conjunction with the counterselective
approach described in this paper, or in conjunction with the delivery of transposon
and transposase components on the same plasmid, particularly for cells in which the
transfection efficiency is limited. For the purposes of gene therapy, SB-mediated
transposition offers the advantage of a non-viral, DNA-mediated vector system with
the capability of stably integrating new gene sequences into host cell chromosomes.
For gene therapy applications requiring long-term expression in a particular target
tissue or cell population, this might be achieved without the use of an integrating
viral vector, thus providing a less risky therapeutic procedure and simplifying the
process of manufacturing the relevant gene transfer reagent.
We thank Zoltan Ivics and Zsuzsanna Izsvak for providing the pT/neo/HSV-
TK plasmid. This work was supported by the Arnold and Mabel Beckman Foun-
dation and by research grant DK 55571 from the National Institutes of Health.
Belur, L., Frandsen, J., Dupuy, A., Ingbar, D. H., Largaespada, D. A, Hackett, P. B., and McIvor, R. S.
(2003) Gene insertion and long-term expression in lung mediated by the Sleeping Beauty transposon
system. Mol. Ther. 8:501--507.
592Converse et al.
Calos, M. P. (1996) The potential of extrachromosomal replicating vectors for gene therapy Trends Genet.
Carter, P. J. and Samulski, R. J. (2000) Adeno-associated viral vectors as gene delivery vehicles. Int. J.
Mol. Med. 6:17--27.
Clark, K. J., Geurts, A. M., Bell, J. B., and Hackett, P. B. (2004) Transposon vectors for gene-trap
insertional mutagenesis in vertebrates. Gen. J. Genet. Dev. 39:225--233.
Cui, Z., Guerts, A. M., Liu, G., Kaufman, C. D., and Hackett, P. B. (2002) Structure-function analysis of
the inverted terminal repeats of the Sleeping Beauty transposon. J. Mol. Biol. 318:1221--1235.
Dupuy, A., Clark, C., Carlson, C., Fritz, S., Davidson, A. E., Markley, K. M., Finley, K., Fletcher, C. F.,
Ekker, S., Hackett, P., Horn, S., and Largaespada, D. A. (2002) Mammalian germline transgenesis by
transposition. Proc. Nat. Acad. Sci. USA 99:4495--4499.
Dupuy, A., Fritz, S., and Largaespada, D. A. (2001) Transposition and gene disruption using a mutagenic
transposon vector in the male germline of the mouse. Genesis 30:82--88.
Fadool, J. M., Hartl, D. L., and Dowling, J. E. (1998) Transposition of the mariner element from
Drosophila mauritiana in zebrafish. Proc. Nat. Acad. Sci. USA 95:5182--5186.
Fischer, S. E., Wienholds, E., and Plasterk, R. H. (2001) Regulated transposition of a fish transposon in
the mouse germ line. Proc. Nat. Acad. Sci. USA 98:6759--6764.
Geurts, A. M, Yang, Y., Clark, K. J., Cui, Z., Dupuy, A., Bell, J. L., Largaespada, D. A., and Hackett,
P. B. (2003) Gene transfer into genomes of human cells by the Sleeping Beauty transposon system.
Mol. Ther. 8:108--117.
Hartl, D. L., Lozovskaya, E. R., Nurminsky, D. I., and Lohe, A. R. (1997) What restricts the activity of
mariner-like transposable elements. Trends Genet. 13:197--201.
Horie, K., Kuroiwa, A., Ikawa, M., Okabe, M., Kondoh, G., Matsuda, Y., and Takeda, J. (2001) Efficient
chromosomal transposition of a Tc1/mariner-like transposon Sleeping Beauty in mice. Proc. Nat.
Acad. Sci. USA 98:9191--9196.
Ivics, Z., Hackett, P. B., Plasterk, R. H., and Izsvak, Z. (1997) Molecular reconstruction of Sleeping
Beauty, a Tc1-like transposon from fish, and its transposition in human cells. Cell 91:501--10.
Izsvak, Z., Ivics, Z., and Hackett, P. B. (1997) Repetitive elements and their genetic applications in
zebrafish. Biochem. Cell Biol. 75:507--523.
Izsvak, Z., Ivics, Z., and Plasterk, R. H. (2000) Sleeping Beauty, a wide host-range transposon vector for
genetic transformation in vertebrates. J. Mol. Biol. 302:93--102.
Jonsson, J. J., Habel, D. E., and McIvor, R. S. (1995) Retrovirus-mediated transduction of an engineered
intron-containing purine nucleoside phosphorylase gene. Hum. Gene Ther. 6:611--623.
Karsi, A., Moav, B., Hackett, P., and Liu, Z. J. (2001) Effects of insert size on transposition efficiency of
the Sleeping Beauty transposon in mouse cells. Mar. Biotechnol. 3:241--245.
Kay, M. A., Glorioso, J. C., and Naldini, L. (2001) Viral vectors for gene therapy: the art of turning
infectious agents into vehicles of therapeutics. Nature Med. 7:33--40.
Li, Z. H., Liu, D. P., Wang, J., Guo, Z. C., Yin, W. X., and Liang, C. C. (1998) Inversion and trans-
position of Tc1 transposon of C. elegans in mammalian cells. Somatic Cell Mol. Genet. 24:363--369.
Luo, G., Ivics, Z., Izsvak, Z., and Bradley, A. (1998) Chromosomal transposition of a Tc1/mariner-like
element in mouse embryonic stem cells. Proc. Nat. Acad. Sci. USA 95:10769--10,773.
Mansour, S. L., Thomas, K. R., and Capecchi, M. R. (1988) Disruption of the proto-oncogene int)2 in
mouse embryo-derived stem cells: a general strategy for targeting mutations to non-selectable genes.
McIvor, R. S. and Simonsen, C. C. (1990) Isolation and characterization of a variant dihydrofolate
reductase cDNA from methotrexate-resistant murine L5178Y cells. Nucleic Acids Res. 18:7025--7032.
Mikkelsen, J. G., Yant, S. R., Meuse, L., Huang, Z., Xu, H., and Kay, M. A. (2003) Helper-Independent
Sleeping Beauty transposon-transposase vectors for efficient nonviral gene delivery and persistent
gene expression in vivo. Mol. Ther. 8:654--665.
Miller, A. D. (1992) Retroviral vectors Curr. Topics Microbiol. Immunol. 158:1--24.
Miller, A. D. and Rosman, G. J., (1989) Improved retroviral vectors for gene transfer and expression.
Biotechniques 7:980--982, 984--986, 989--990.
Morris, J. A., May, C., Kim, H. S., Ismail, R., Wagner, J. E., Gunther, R., and McIvor, R. S. (1996)
Comparative methotrexate resistance of transgenic mice expressing two distinct dihydrofolate
reductase variants. Transgenics 2:53--67.
593Sleeping Beauty-Mediated Transposition
Morris, J. A. and McIvor, R. S. (1994) Saturation mutagenesis at dihydrofolate reductase codons 22 and Download full-text
31. A variety of amino acid substitutions conferring methotrexate resistance. Biochem. Pharmacol.
Nakai, H., Yant, S. R., Storm, T. A., Fuess, S., Meuse, L., and Kay, M. A. (2001) Extrachromosomal
recombinant adeno-associated virus vector genomes are primarily responsible for stable liver trans-
duction in vivo. J. Virol. 75:6969--6976.
Nishikawa, M. and Huang, L. (2001) Nonviral vectors in the new millennium: delivery barriers in gene
transfer. Hum. Gene Ther. 12:861--870.
Plasterk, R. H. (1993) Molecular mechanisms of transposition and its control Cell 74:781--786.
Pugh, J. C., Yaginuma, K., Koike, K., and Summers, J. (1988) Duck hepatitis B virus (DHBV) particles
produced by transient expression of DHBV DNA in a human hepatoma cell line are infectious in
vitro. J. Virol. 62:3513--3516.
Rubinstein, M., Japon, M. A., and Low, M. J. (1993) Introduction of a point mutation into the mouse
genome by homologous recombination in embryonic stem cells using a replacement type vector with a
selectable marker. Nucleic Acids Res. 21:2613--2617.
Sambrook, J., Fritsch, E. F., and Maniatis, T. (1989) Molecular Cloning, A Laboratory Manual. 2 edn.
CSHL Press, Cold Spring Harbor.
Schouten, G. J., van Luenen, H. G., Verra, N. C., Valerio, D., and Plasterk, R. H. (1998) Transposon Tc1
of the nematode Caenorhabditis elegans jumps in human cells. Nucleic Acids Res. 26:3013--3017.
Score, P. R., Belur, L. R., Frandsen, J. L., Geurts, J. L., Hackett, P. B., Largaespada, D. A., and McIvor,
R. S., Sleeping Beauty-mediated transposition and long-term expression in vivo: Use of the LoxP-Cre
recombinase system for transposition-specific expression (Submitted).
Wigler, M., Pellicer, A., Silverstein, S., Axel, R., Urlaub, G., and Chasin, L. (1979) DNA-mediated
transfer of the adenine phosphoribosyltransferase locus into mammalian cells. Proc. Nat. Acad. Sci.
Wilber, A. C., Frandsen, J. L., Geurts, J. L., Largaespada, D. A., Hackett, P. B., and McIvor, R. S., RNA
as a source of transposase for Sleeping Beauty-mediated gene insertion and expression in somatic cells
and tissues (Submitted).
Wolff, J. A., Ludtke, J. J., Acsadi, G., Williams, P., and Jani, A. (1992) Long-term persistence of plasmid
DNA and foreign gene expression in mouse muscle. Hum. Mol. Genet. 1:363--369.
Wolff, J. A., Malone, R. W., Williams, P., Chong, W., Acsadi, G., Jani, A., and Felgner, P. L. (1990)
Direct gene transfer into mouse muscle in vivo. Science 247:1465--1468.
Yant, S. R., Meuse, L., Chiu, W., Ivics, Z., Izsvak, Z., and Kay, M. A. (2000) Somatic integration and
long-term transgene expression in normal and haemophilic mice using a DNA transposon system.
Nat. Genet. 25:35--41.
Zhang, L., Sankar, U., Lampe, D. J., Robertson, H. M., and Graham, F. L. (1998) The Himar1 mariner
transposase cloned in a recombinant adenovirus vector is functional in mammalian cells. Nucleic
Acids Res. 26:3687--3693.
Zhao, R. C., McIvor, R. S., Griffin, J. D., and Verfaillie, C. M. (1997) Gene therapy for chronic mye-
logenous leukemia (CML): a retroviral vector that renders hematopoietic progenitors methotrexate-
resistant and CML progenitors functionally normal and nontumorigenic in vivo. Blood 90:4687--4698.
594 Converse et al.