Published online 20 December 2007 Nucleic Acids Research, 2008, Vol. 36, No. 1 e9
doi:10.1093/nar/gkm1123
Site-specific recombination in Schizosaccharomyces
pombe and systematic assembly of a 400kb
transgene array in mammalian cells using the
integrase of Streptomyces phage fBT1
Zhengyao Xu, Nicholas C. O. Lee, Felix Dafhnis-Calas, Sunir Malla,
Margaret C. M. Smith and William R. A. Brown*
Institute of Genetics, Queen’s Medical Centre, University of Nottingham, Nottingham NG7 2UH, UK
Received August 28, 2007; Revised November 10, 2007; Accepted December 3, 2007
ABSTRACT
We have established the integrase of the
Streptomyces phage fBT1 as a tool for eukaryotic
genome manipulation. We show that the fBT1
integrase promotes efficient reciprocal and conser-
vative site-specific recombination in vertebrate cells
and in Schizosaccharomyces pombe, thus establish-
ing the utility of this protein for genome manipulation
in a wide range of eukaryotes. We show that
the fBT1 integrase can be used in conjunction with
Cre recombinase to promote the iterative integration
of transgenic DNA. We describe five cycles of
iterative integration of a candidate mouse centro-
meric sequence 80kb in length into a human mini-
chromosome within a human-Chinese hamster
hybrid cell line. These results establish the generality
of the iterative site-specific integration technique.
INTRODUCTION
Site-specific recombinases and occasionally telomeres have
been used as tools for manipulating eukaryotic genome
structure. Cre (1) and Flp (2) were the first site-specific
recombinases to be used for this purpose. Both catalyse
reversible reactions between identical sites of about 35 bp.
These recombinases have been most widely used to
promote deletion reactions, where the instability of one
product at cell division renders the reaction effectively
irreversible. Heterospecific sites (3) have been advocated
as a way of rendering such reactions irreversible and have
been successfully exploited to allow efficient integration
reactions (4). Cre and Flp belong to the tyrosine
recombinase family of site-specific recombinases. Other
members of this family of site-specific recombinases such
as the integrase of phage
promote unidirectional or
irreversible reactions between sites differing in sequence,
but do so only in the presence of additional proteins,
limiting their utility for genome manipulation. A second
mechanistically distinct family of site-specific recombi-
nases is the serine recombinase family. This family
includes the integrase of the Streptomyces phage fC31,
which promotes an irreversible site-specific recombination
reaction between the related but distinct attachment sites
of the phage and bacterial genomes: attP and attB. The
discovery that this reaction can be promoted by the
purified integrase protein in the absence of any additional
proteins (5) has led to fC31 integrase being used as a
genome manipulation reagent (6,7). The fC31 integrase is
a member of the ‘large serine recombinase sub-family’
which includes many different members with a similar
pattern of domain organization (8). Several of these
proteins have been studied and shown to be similar to the
fC31 integrase in promoting unidirectional reactions
between non-identical sites in the absence of additional
proteins. These others proteins should also be useful as
reagents for genome manipulation but only a few have
been used as such (9).
Site-specific recombinases have been used to approach a
range of different problems but they are most widely
employed in the generation of tissue-specific conditional
loss of function alleles in metazoan genomes. It is often
only necessary to generate and study one such allele in
isolation, but for the appropriate modelling of human
cancers it is necessary to be able to generate independently
loss and gain of function alleles at several different loci
in the same animal. This type of complex manipulation
requires a large palette of enzymes of different specificities.
The ‘large serine recombinase sub-family’ potentially
satisfies this requirement. A second related area in which
site-specific recombinases have been used has been in the
Present address:
Margaret C. M. Smith, Institute of Medical Sciences, University of Aberdeen, Aberdeen AB25 2ZD, UK
*To whom correspondence should be addressed: Tel: +44 (0)115 823 0386; Fax: +44 (0)115 823 0338; Email: William.brown@nottingham.ac.uk

2007 The Author(s)
This is an Open Access article distributed under the terms of the Creative Commons Attribution Non-Commercial License (http://creativecommons.org/licenses/
by-nc/2.0/uk/) which permits unrestricted non-commercial use, distribution, and reproduction in any medium, provided the original work is properly cited.

introduction of long segments of human genomic DNA
into the genome of an experimental organism such as a
mouse (10). Such humanization of an experimental
organism may be useful for the study of human genetic
disease. A major technical challenge confronting this
approach is that of integrating long tracts of DNA. We
have described a method termed ‘Iterative site-specific
integration’ (ISSI), which allows this challenge to be
overcome (11). ISSI combines the activity of a unidirec-
tional site-specific integrase with that of a reversible
recombinase to allow the serial integration of transgenic
DNA and thus potentially allows the assembly of
transgenic arrays of arbitrary size. We have previously
used the fC31 integrase in combination with Cre to
iteratively integrate two inserts of 70 kb and 80 kb in size
cloned in bacterial artificial chromosomes into a human
mini-chromosome in a chicken human hybrid cell line.
Here, we investigate the potential of an additional
member of the serine recombinase family; the integrase of
the Streptomyces phage fBT1 as a tool for genome
manipulation. We show firstly that the fBT1 integrase
functions efficiently in vertebrate cells and in cells of the
yeast Schizosaccharomyces pombe. The results obtained in
fission yeast indicate that the fBT1 integrase is likely to
function in most eukaryotic cell types. We then show that
the fBT1 integrase can be used in conjunction with Cre
recombinase to promote the iterative integration of
transgenic DNA. Thus, we describe five cycles of iterative
integration of a candidate mouse centromeric sequence
80 kb in length into a human mini-chromosome within a
human-Chinese hamster hybrid cell line. This allowed us to
assemble a transgenic tract of DNA 400 kb in length. The
demonstration that ISSI can be implemented using diffe-
rent combinations of site-specific recombinases and in diff-
erent cell lines establishes the versatility of the technique.
MATERIALS AND METHODS
Plasmid construction
Plasmids were constructed by standard techniques. The
sequences of the plasmids and of the integrases used in this
work can be obtained from http://www.nottingham.ac.uk/
genetics/staff/williambrown/reg.html. The nuclear locali-
zation signal used to tag the fBT1 and fC31 integrases
was derived from the large T antigen of SV40 virus and
included the residues MPKKKRKV. The sequences of
the fBT1attP and attB sites used were respectively:
TTCGGGTGCTGGGTT GTTGTCTCTGGACAGTG
ATCCATGGGAAACTACTCAGCACCA and GCCG
TCCTTGA CCAGGTTTTTGACGAAAGTGATCCA
GATGATCCAGCTCCACACCC. In each case, these
att sites were engineered into plasmids using synthetic
DNA provided by Invitrogen. The plasmids were checked
by restriction site mapping and in all cases by sequencing.
The blasticidin resistance (BSR) (12) gene was a kind
gift from Hiroshi Arakawa of the GSF, Munich. The
hygromycin resistance gene used was present in the counter
selectable hygromycin–thymidine kinase fusion (HyTk)
(13). The CCAG promoter (14) was a kind gift of Ian
Chambers (EdinburghUniversity). All plasmids, sequences
and vectors are available from William Brown (WRAB)
subject to a materials transfer agreement. Plasmid DNA
was purified by alkali lysis and precipitation with poly-
ethylene glycol(PEG). PACs were purified by alkali lysis
and PEG precipitation (experiments shown in Figure 2).
The CCAG fBT1 integrase IRES zeo was linearized at the
unique PstI site at position 3557 and the CCAG Cre IRES
Neo at the unique NotI site at position 3762.
Cell culture
DT40 were maintained and electroporated (11) as
described previously and as summarized in http://pheas-
ant.gsf.de/DEPARTMENT/dt40.html. CHO cells were
manipulated as described previously (15). The following
refinements were made: blasticidin was used at 30 mg/ml
to select for DT40 transformants and at 6 mg/ml to
select for Chinese hamster ovary (CHO) transformants.
Hygromycin was used at 2mg/ml to select for DT40
transformants and 0.6mg/ml to select for CHO transfor-
mants. Zeocin and neomycin were used at 1 mg/ml and
2mg/ml respectively for DT40 and 0.25 and 0.45mg/ml,
respectively for CHO cells The site-specific integration
experiments in DT40 cells were carried out as follows: cells
were electroporated using previously established condi-
tions with the indicated amount of DNA and immediately
after transfection were plated out in 96-well dishes.
Selection was applied 18 h later. DT40 colonies were
counted after 12–14 days. Efficiencies given in Table 1 are
the total number of resistant colonies recovered divided by
the total number of cells electroporated. We refer to
this figure as the absolute efficiency of integration in
the table headings and text. During the course of the ISSI
experiments, 500 mg P1 artificial chromosome (PAC)
DNA was transfected into 3
10
7
CHO cells by electro-
poration with settings of 400V, 250 mF and 25
. Typically,
between 40 and 60 colonies were recovered in each
experiment.
PCR, FISH and filter hybridization analysis
Pulsed field gels and fluorescent in situ hybridization
(FISH) were as described previously (16) except that
Alexafluor dyes were used for the FISH. The conditions
for the PCR of the attL are as follows for 35 cycles: 948C
for 30 s, 558C for 30 s and 728C for 60 s. The attR used
Table 1. Efficiencies of site-specific integration promoted by the fBT1 integrase in DT40 cells
Cell line Transfected DNA Amount (mg) Number of
cells transfected
% plated Resistant clones Efficiency
DTCCAGattP
00
BSD fBT1-5
0
nls-1 attB
00
HyTk 100 5.0E+0.7 25 405 3
10E-5
DTCCAGattP
00
BSD fBT1-5
0
nls-2 attB
00
HyTk 100 1.5E+0.7 25 476 12
10E-5
DTCCAGattP
00
BSD fBT1-5
0
nls-2 attB
00
HyTk 100 5.0E+0.7 25 362 2.9
10E-5
e9 Nucleic Acids Research, 2008, Vol. 36, No. 1 PAGE 2 OF 9

similar cycle times but an annealing temperature of 628C.
The buffer used was a 16mM (NH
4
)
2
SO
4
, 67mM Tris–
HCl, pH 8.8, 0.1% Tween-20, 1mM MgCl
2
and the Taq
polymerase was from Bioline. The primer sequences and
locations are given in Table 1 of the Supplementary data.
The primers were used at 200 nM final concentration
except for the BSDNotR, which was used at 400 nM.
Markers used in the PCR in Figure 1C were the 100 bp
ladder from New England Biolabs.
Yeast methods
The genotypes of the key strains are given in the Results
section. Media and culture were as described on the
comprehensive site of Professor Susan Forsburg available
at http://www-rcf.usc.edu/forsburg/index.html except
for the following specific departures. S. pombe were
transformed using the lithium acetate procedure as
described in the paper of Bahler and colleagues (17).
DNA was extracted from 10ml of saturated culture using
DNAzol. Briefly, cells were spheroplasted using lyticase,
concentrated by centrifugation and then lysed with 1ml of
DNAzol. After incubation and mixing in 1ml DNAzol,
insoluble material was cleared from the extract by
centrifugation at 900 g for 10min. Nucleic acids were
precipitated from the supernatant with 1ml of ethanol
and dissolved in 0.5ml of TE containing RNAase A at
CCAG Promoter
IRES
Nuclear localisation
signal
Zeocin resistance
gene
8419 bp
CCAG BSD
HyTk
A
attP"
attB″
attB″ HyTk
CCAG BSDattR″HyTk attL″
B
’ ’
1 2 3 4 5 6 7 8 9 1
0
1
1
1
0
0
b
p

l
a
d
d
e
r
1
0
0
b
p

l
a
d
d
e
r
1
0
0
b
p

l
a
d
d
e
r
attL
attR
C
Figure 1. fBT1 integrase-mediated site-specific integration into vertebrate chromosomes. (A) CCAG fBT1 integrase IRES zeo is the fBT1 integrase
expression plasmid. (B) Representation of the promoter trapping strategy used to measure fBT1 integrase mediated site-specific integration into
chicken chromosomes. The linearized expression plasmid CCAG fBT1 integrase IRES zeo was introduced into DT40 cells by electroporation, the
plasmid CCAG attP’’ BSD was then randomly integrated into DT40 cells and site-specific recombination measured by the number of hygromycin
resistant cells recovered following transfection with the promoter-less plasmid attB’’ HyTk. (C) fBT1 integrase mediates conservative and reciprocal
site-specific integration into chicken chromosomes. PCR across the indicated junctions was used to confirm the site specific nature of the integration
reaction described in (B). Details of the primers and predicted product sizes are given in Table 1 of the Supplementary data.
PAGE 3 OF 9 Nucleic Acids Research, 2008, Vol. 36, No. 1 e9

200mg/ml and SDS at 0.1%. The solution was incubated
at 378C for 45min and then pronase was added to 50 mg/
ml. The sample was then incubated for a further 30min at
558C, phenol chloroform extracted two or three times
until the interface was free of precipitate and then
precipitated with ethanol. The pellet was dissolved in
200ml TE and stored at 48C. For western blotting, we used
a protocol kindly described to us by Stepen Kearsey of
Oxford University. Fifty microlitre of cells were grown to
an OD of 0.2 in minimal medium, concentrated by
centrifugation, rinsed in 0.9M sorbitol and resuspended
in 100 ml of 20% trichloracetic acid (TCA) and transferred
to a 1.5ml microcentrifuge tube. A quantity of 0.5mm
glass beads was added to produce a total volume of 0.5ml.
The tube was then agitated for five 1min periods using a
Scientific Instruments Disruptor Genie. A total of 900 ml
of 5% TCA was added, the beads were allowed to settle
and the protein precipitate was collected from the super-
natant by centrifugation at 900g for 10min. The pellet was
dissolved in 250ml of SDS-PAGE loading buffer. Twenty-
five microlitre of this material corresponding to about
50 mg of protein was loaded onto a 10% standard SDS
polyacrylamide gel and analysed by western blotting using
a rabbit anti integrase first antibody and a peroxidase
conjugated donkey anti-rabbit IgG second antibody
(Amersham), the binding of which was detected using
the ECL kit from Amersham Pharmacia.
RESULTS
fBT1 integrase functions in vertebrate cells
First of all, we wanted to establish the fBT1 integrase
functions in vertebrate cells. We established a chicken
DT40 cell line expressing the fBT1 integrase using the
expression plasmid CCAG fBT1 integrase IRES Zeo
(Figure 1A). We then integrated a linearized plasmid in
which a fBT1 integrase attP (referred to as attP
00
) site is
placed between the CCAG promoter and the coding
region of a gene conferring resistance to blasticidin
(CCAG attP
00
BSD; Figure 1B) and selected for resistance
to blasticidin. These cells were then transfected with a
circular plasmid in which a fBT1 integrase attB (attB
00
)
site abutted a gene conferring resistance to hygromycin
(attB
00
HyTk: Figure 1B), but which lacked a promoter.
Site-specific recombination between the stably integrated
attP
00
site and the attB
00
would place the hygromycin
resistance gene under the control of the CCAG promoter
and, thus confer, hygromycin resistance upon the trans-
fected cells (Figure 1B). Following transfection, hygro-
mycin resistant cells were recovered efficiently (Table 1).
Eleven hygromycin resistant clones were picked and
site-specific recombination was confirmed by PCR
(Figure 1C). Sequencing the products as uncloned PCR
products from two such cell lines confirmed that the
reaction was reciprocal and conservative (data not
shown). These results thus demonstrate that the fBT1
integrase promotes efficient, reciprocal, conservative site-
specific recombination in vertebrate cells. The efficiency of
recombination seen with the fBT1 integrase in this type of
experiment was comparable to what we had detected in
analogous experiments carried out using the fC31
integrase (11).
fBT1 integrase functions in fission yeast cells
We wished to establish the general utility of the fBT1
integrase in a variety of eukaryotic cells. We therefore
tested the function in the model organism S. pombe.
Initially we tried a promoter trapping approach using
single recombining sites that was analogous to that used in
the experiments described above but, despite many
attempts, we could detect no site-specific recombination
activity. We therefore used the target replacement appro-
ach described in Figure 2A and which had been previously
used by Thomason and colleagues (18) in their work on
the fC31 integrase. We needed a positive control and so
we chose to compare the activities of the fBT1 integrase
and the fC31 integrase. We targeted a counter selectable
marker gene; URA4; flanked by attB sites for either of the
two integrases to the leu1
+
locus into strain NL16 (h+,
ade6-M216, ura4-D18, leu1
+
, arg3-D4). This sequence
served as a target for the site-specific recombinase-
mediated integration reaction (Figure 2A). We then
introduced a plasmid in which the integrase expression
was driven either by the strong REP1 or weak REP81
promoter (Figure 2B) into the cells harbouring the target
site. We (11) and others (19) have shown that integrase
activity in eukaryotic cells is enhanced by the presence of a
nuclear localization signal and thus in both cases we used
derivatives that had been tagged at the amino termini with
the nuclear localization signals of SV40 large T antigen.
Cells containing integrase were grown under leucine
selection for the expression plasmid, transformed with a
circular plasmid (p attP Arg3 attP) (Figure 2A) containing
an Arg 3 gene flanked by the appropriate attP sites and
selection applied for arginine prototrophy. Selection on
the Leu marker for the expression plasmid was relaxed
at the time of transformation and the ability of extant
integrase to promote integration of the Arg3 gene into the
genome was measured. We recovered the arginine proto-
trophs and tested whether they had lost the URA4 gene on
the basis of resistance to fluororotic acid (FOA). The
efficiency of the integration reaction was estimated by
transformation with a plasmid pARSV40, a plasmid that
contains an S. pombe autonomously replicating sequence
(ARS) sequence as well as the ARG3 gene and is capable of
extra-chromosomal replication. The data in Figure 2C and
D are normalized with respect to the number of colonies
obtained in the same experiment upon transformation with
this plasmid. In order to estimate the extent of non-specific
integration of the pattP Arg3 attP plasmid, we also
established derivatives of each leu1::attB-ura4-attB strain
containing an empty REP1 vector and this was also
transformed with the appropriate pattPArg3 attP plasmid.
The results in Figure 2C and D are presented as the
mean of two independent experiments carried out for each
integrase. The proportion of the transformants that are
ARG
+
URA
+
or ARG
+
URA

are presented in dark
tone and light tones, respectively in Figure 2B and C. The
numbers of transformants recovered in the individual
experiments are presented in Supplementary data Table 2.
e9 Nucleic Acids Research, 2008, Vol. 36, No. 1 PAGE 4 OF 9

010
20
30
40
50
60
70
80
90
100
0
10
20
30
40
50
60
70
80
90
100
P-Arg3-P P-Arg3-P
Rep1
P-Arg3-P PARSV40 no Plasmid P“-Arg3-P” P“-Arg3-P”
Rep1
P“-Arg3-P” PARSV40
Rep1
No Plasmid
%
%
Ura4
Arg3
attP attP
attBattB
Leu1
Arg3attL attR
Ura4 attBattB
ARG URA
ARG URA
+
++
−
ARG URA
++
A
B
C
C
B
6
-
2
0
n
g
R
E
P
1
φ
C
3
1
-
2
0
µ
g
R
E
P
8
1
φ
C
3
1
-
2
0
µ
g
R
E
P
1
φ
C
3
1
-
2
0
µ
g
R
E
P
1
φ
B
T
1
-
4
0
µ
g
R
E
P
8
1
φ
B
T
1
-
4
0
µ
g
φ
B
T
1

i
n
t
e
g
r
a
s
e
-
4
0
n
g
φ
B
T
1

i
n
t
e
g
r
a
s
e
-
4
n
g
φ
C
3
1

i
n
t
e
g
r
a
s
e
-
4
n
g
R
E
P
1
φ
C
3
1
-
2
µ
g
175
47.5
62
kDa 83
E
nmt
Promoter
nmt terminator
Leu2 gene
Site specific
integrase
ars1
Generic
Integrase yeast
expression
plasmid
nls
D
φC31 integrase φBT1 integrase
Figure 2. fBT1 and fC31 integrase activities in S. pombe.(A) Schematic representation of the strategy used to detect site-specific integration
promoted by the serine recombinases in S. pombe. The leu1 gene was replaced with a cassette encoding the counter selectable ura4 gene flanked by
attB sites. Site-specific recombination in the phase indicated between both of these attB sites and the attP sites flanking the Arg3 gene on an
incoming plasmid leads to the replacement of the ura4 gene and the insertion of the Arg3 gene. Several other outcomes arising as a result of a single
recombination event on one sister chromatid followed by site-specific recombination with the other sister are discussed in the text and described in
the supplementary data. (B) The generic integrase expression plasmid used in the experiments are carried out in S. pombe. The gene encoding the
integrase tagged with a nuclear localization signal was cloned into either the REP1 or REP81 expression plasmids to produce the indicated construct.
(C) Site-specific recombination mediated by fC31 integrase at attB ura4 attB. The indicated integrase expression plasmid was transformed into the
attB ura4 attB containing strain and then a plasmid pBS attP Arg3 attP introduced, arginine prototrophs recovered and tested for loss of the ura4
gene by growth on FOA. The ARS containing plasmid PARSV40 was used as a control for the transformation efficiency in the two reactions. (D)
Site-specific recombination mediated by fBT1 integrase at attB
00
ura4 attB
00
. The indicated integrase expression plasmid was transformed into the
attB
00
ura4 attB
00
containing strain and then a plasmid pBS attP
00
Arg3 attP
00
introduced, arginine prototrophs recovered and tested for loss of the
ura4 gene by growth on FOA. The ARS containing plasmid PARSV40 was used as a control for the transformation efficiency in the two reactions.
(E) Western blotting of fC31 and fBT1 integrase expression in S. pombe. Purified integrase proteins or proteins from the indicated strains that had
been extracted by disruption with glass beads in TCA, size fractionated, transferred and analysed by western blotting. CB6 is a control strain lacking
the integrase expression plasmid.
PAGE 5 OF 9 Nucleic Acids Research, 2008, Vol. 36, No. 1 e9

As mentioned earlier, we identify the fBT1 integrase
attachment sites by the use of the superscript
00
in both
Figure 2 and Supplementary data Table 2. The results in
Figure 2 suggest that both integrases promote site-specific
integration of the respective pattP Arg3 attP into the
target locus. Several aspects of the data bear comment.
Firstly, the fC31 integrase activity was undetectable when
expressed from the REP1 promoter but activity could be
detected when expression was driven by the weaker
REP81 promoter. We noted that the REP1 fC31
integrase containing colonies were very small and it was
often difficult to transform the yeast with this plasmid. We
confirmed the identity of the plasmid by sequencing and
thus excluded trivial explanations for the result. We
therefore examined the transformants for the presence of
the integrase by western blotting (Figure 2E). This
experiment confirmed that the REP1 expression plasmid
was functional. The inability to detect the integrase in the
REP81 transformants by western blotting was consistent
with what the relative activities of the REP series of
expression plasmids http://www-rcf.usc.edu/forsburg/
index.html. The fact that the REP1fC31integrase contain-
ing cells contain several orders of magnitude, more
integrase yet do not permit efficient plasmid integration
indicates that the integrase is inactive when present at high
concentrations in S. pombe. In contrast, integration
activity of the fBT1 integrase could only be detected
using a REP1 expression vector. Comparison of the
relative amounts of the two integrases in the respective
yeast strains indicted that the likely cause of this difference
was that the fC31 integrase was more efficiently expressed
than the fBT1 integrase by a factor of about 10-fold
(Figure 2E).
Although ARG
+
URA

transformants were detected
following transformation with either the REP81 fC31
integrase or the REP1 fBT1 integrase expression con-
structs, about 70% in each case were ARG
+
URA
+
.
Given the background of ARG
+
URA
+
clones seen with
the empty vector plasmid, it was important to determine
the proportion of the ARG
+
URA
+
clones that arose by
site-specific recombination and to understand their origin.
We checked the identity and determined the structure of
the transformants by using PCR, using primers adjacent
to the ends of either the incoming or target attachment
sites (Supplementary data, Tables 3 and 4 illustrates the
relationship between the structure of the integrant and
the predicted content of attachment sites). We confirmed
the accuracy of the PCR analysis by restriction enzyme
digestion with an enzyme that cuts in neither the ARG3 or
URA4 loci, gel electrophoresis, filter transfer and hybri-
dization with a probe corresponding to the deleted
segment of the URA4 (18) locus in the original NL16
strain (Supplementary data; Figure 1). With the exception
of the single most complicated of the URA
+
ARG
+
clones the sizes of the cognate fragments were consistent
with the structure predicted on the basis of the PCR
results. In the case of both the enzymes, > 90% of the
ARG
+
URA
+
clones arose by site-specific integration at
the target leu1::attB-ura4-attB locus. In 21 cases out of 29,
it was necessary to invoke recombination reactions
between sister chromatids at G2 during the cell cycle.
Two clones had structures that could be explained if
homologous recombination was also involved. A final
clone was likely to be random integration events.
Hypothetical explanations for the origin of these struc-
tures are set out in Figure 2 of the Supplementary data.
In previous experiments using the fC31 integrase in
vertebrate cells we had identified damage to participating
sites as a cause of incomplete reaction (20) and so we
checked the sequences of both the substrate (attB and
attP) and (attL and attR) product sites (10 clones) for
both the fC31 and fBT1 integrase. In all cases, these sites
were intact and undamaged. Similarly, we checked 30
ARG
+
URA

clones by PCR and sequenced six, which
confirmed that they were site-specific replacements of the
ura4
+
gene with the arg3
+
gene. In summary therefore
our data show that in all of the transformants the reac-
tions had gone to completion and we found no evidence of
incomplete reaction as a result of damage to the substrate
sites.
The results of these experiments demonstrate that
the fBT1 integrase promotes site-specific integration in
fission yeast cells with efficiencies that in practical terms
are similar to the fC31 integrase and suggest that the
fBT1 integrase will be generally useful. The fact that the
fBT1 integrase is expressed less efficiently than the fC31
integrase will be discussed.
Iterative site specific recombination (ISSI) inChinesehamster
cells using a combination of the fBT1 integrase andCre
recombinase
ISSI (11) is a method that allows the serial assembly of
long transgene arrays as a result of the combined action of
a unidirectional recombinase and a reversible recombinase
such as Cre. ISSI has previously been implemented only in
chicken DT40 cells using Cre and the fC31 integrase. If
ISSI is to be widely adapted as a method then it is
important that it be shown to be practical in other cell
types and to know how effectively it works when used with
different combinations of recombinase. Therefore, we set
out to establish the ISSI methodology using Cre and the
fBT1 integrase in Chinese hamster cells. In order to do
this, we used a mini-chromosome derivative of the human
Y chromosome termed XP4 that had previously been
engineered by sequence targeting and telomere directed
chromosome breakage in DT40 cells (20). This mini-
chromosome includes a CCAG loxP HyTK attP
00
cassette
integrated by homologous recombination into the DYZ5
array 700 kb distant from the centromeric array of alphoid
DNA (Figure 3A). This cassette acts as a substrate for the
subsequent integration reactions that constitute the ISSI
process.
We used microcell transfer to move this chromosome
into the CHO cells, checked the structure of the
chromosome by pulsed field gel electrophoresis and
confirmed its linear integrity by fluorescent in situ
hybridization (data not shown). We then introduced the
CCAG fBT1 integrase IRES Zeo (Figure 1A) and CCAG
Cre IRES Neo (Figure 3B) expression plasmids into the
hybrid cell line, in order to establish a cell line expressing
e9 Nucleic Acids Research, 2008, Vol. 36, No. 1 PAGE 6 OF 9

CCAG Promoter
CRE Recombinase
IRES
G418 resistance
gene
CCAG CRE IRES NEO
8370 bp
Kanamycin
resistance
gene
Blasticidin
resistance
gene
pBS linker
attB”
loxP
pBS linker
14109 bp
14887 bp
BamHI
NotI
BamHI
NotI
BamHI
NotI
BamHI
NotI
loxP
Kanamycin
resistance
gene
Hygromycin
resistance
gene
attP”
attP''
B
D
C
attB”
F
A
3
8
2
/
5
4
1
6
/
4

(
1
)
4
2
6
/
8
-
8

(
2
)
4
5
9
/
1
0

(
3
)
4
9
1
/
1
6

(
4
)
5
0
6
/
8
-
5

(
5
)
97
194
291
388
485
582
679
kb
EF
700 kb
Yp
160 kb
TEL
20 kb
DYZ5 ar ray
loxP attP″
CCAGHyTkeGFP
attP
520 kb
attB
BPintegrase att sites
XP4
Figure 3. fBT1 integrase and Cre recombinase-mediated ISSI into vertebrate chromosomes. (A) A diagram of the centromere adjacent DNA of the
mini-chromsome XP4 used in the ISSI reaction sequence is shown in (D). The mini-chromosome was engineered by sequence targeting in the chicken
cell line DT40 and then moved into Chinese hamster ovary cells by microcell transfer. (B) CCAG CRE IRES Neo is the Cre expression plasmid used
to establish the cell line used in the ISSI experiments shown in this figure. (C) fBT1PACattB BloxPBsr att B is one of the two large fragment
cloning vector used in the ISSI experiments shown in this figure. (D) fBT1PACattP loxPHyTk attP is one of the two the large fragment cloning
vector used in the ISSI experiments shown in this figure. (E) ISSI mediated by the fBT1 integrase and Cre recombinase in CHO cells. Cells
containing the mini-chromosome XP4 were transfected with the fBT1 integrase and Cre expression plasmids shown in Figure 1 and then serially
transfected with PAC vectors; fBT1PACattB BloxPBsr att B and fBT1PACattP loxPHyTk attP containing 80 kb inserts of a tandemly repeated
DNA sequence designed to resemble the centromeric mouse-minor satellite DNA. (F) Cytogenetic characterization of the mini-chromosome 459/10
from the ISSI sequence shown in (D). The mouse-minor satellite DNA sequence and the human Y alphoid DNA have been detected by fluorescence
in situ hybridization and are shown in red and green, respectively.
PAGE 7 OF 9 Nucleic Acids Research, 2008, Vol. 36, No. 1 e9

both site-specific recombinases necessary for the ISSI
reaction.
We wished to use ISSI to assemble a candidate
centromeric sequence on the DYZ5 array of XP4, which
we could subsequently test for functionality by excising
the pre-existing centromere using the fC31 integrase as
described earlier. We therefore, designed a sequence that
was based upon the mouse-minor satellite DNA. A
tandem array of this sequence was assembled by standard
methods of concatamerization using ligation of BamHI
and BglII cleaved DNA. The final product of 80 kb in size
was cloned into each of two PAC vectors that can
participate in the ISSI reaction (Figure 3C and D). One
PAC vector; fBT1 PAC attB
00
attB
00
loxP BSR is used for
the first, third and subsequent odd numbered steps of the
ISSI reaction, the other fBT1 PAC attP
00
attP
00
loxP
HyTk is used for the even numbered steps. We carried out
five steps of the ISSI reaction resulting in the assembly of a
400 kb array of the test DNA on the mini-chromosome.
At each step, we typically recovered many tens of
transformants. We characterized these using both mole-
cular (Figure 3E) and cytogenetic approaches (Figure 3F)
in order to select as substrates for the next step in the ISSI
sequence cell lines containing a single un-rearranged
chromosome that includes an integrated array of the
predicted size. Typically, each step gave between three and
five clones with such a structure. The integrated DNA in
the remainder was either deleted or empty as has been
observed and explained previously (11). The success of
this sequence thus establishes ISSI as a generally useful
approach to the assembly of transgenic DNA arrays and
also demonstrates the utility of the fBT1 integrase.
DISCUSSION
The results contained in this article establish a new serine
recombinase; fBT1 integrase, as a tool for genome
manipulation. We have shown that this novel recombinase
functions efficiently in three eukaryotic cell types:
S. pombe, chicken DT40 cells and Chinese hamster
ovary cells. We have now shown that ISSI, an iterative
technique for the assembly of long tracts of transgenic
DNA, works with two large serine integrases in two diff-
erent vertebrate cell lines; DT40 cells and Chinese hamster
ovary cells. This indicates that ISSI is a robust technique
effective in cells other than the hyper-recombinogenic DT
40 chicken cell line where it was originally established.
One practical merit of the characterization of a second
unidirectional recombinase is that when combined with
the fC31 integrase, it should allow one to confidently
attempt the substitution of long tracts of genomic DNA
with new sequences. Thus, one could use one unidirec-
tional recombinase in combination with Cre in a series of
ISSI reactions to assemble a sequence at a locus of choice
and then use the second enzyme to delete the correspond-
ing stretch of genomic DNA. These results in conjunction
with those of others, therefore, suggest that the large
serine recombinases will turn out to be useful tools for
manipulating precisely the long range structure of
eukaryotic genomes. It will, however, be important to
test the ISSI technique in mouse embryonal stem cells (ES)
cells and to determine whether the manipulated cells will
colonize the germ line. Cre recombinase (21) and possibly
fC31 integrase (22) have been shown to damage DNA
when expressed in vertebrate somatic cells. Therefore,
it will be important to establish conditions in which Cre
recombinase and a serine integrase can be used to drive an
ISSI reaction and retain germ line competence in ES cells.
In general, the range of tools available in yeast is such
that it may be thought that the ability to use site-specific
recombinases in these cells is of no practical worth. The
S. pombe centromeres (23) are large and complex and
currently impossible to manipulate in vivo. Thus, despite
many years of work the sequence requirements for the
efficient function of the CENH3 (CENP-A) binding
central core are not defined. Our results raise the
possibility that we will be able to manipulate the central
core sequence organization using these site-specific
recombinases and thereby establish tools to analyse the
relationship between kinetochore and heterochromatin
assembly. The demonstration that the fBT1 integrase is
less efficiently expressed than the fC31 integrase in fission
yeast may be a consequence of the differences in codon
usage between Streptomyces and S. pombe. Recent
experiments which studied the expression of fC31
integrase in mouse embryos (24) have suggested the
importance of using codon optimized fC31 integrase
expression constructs. It would seem on this basis
desirable for work in whole animals and probably yeast
to use codon optimized fBT1 integrase integrase. In
preliminary work, one of us (N.C.O.L.) have compared
the activities in S. pombe of the native and fC31 integrase
proteins codon optimized for expression in mammalian
cells. These experiments have shown that the proteins
function with similar activity, but the yeast containing the
codon optimized protein proliferate more quickly than
those containing the native protein and can be trans-
formed much more efficiently with DNA. We therefore
conclude that it will be preferable to use the fC31
integrase protein that is codon optimized for expression in
mammalian cells than the native protein even when
working in fission yeast.
In previous work (20) we demonstrated that the fC31
integrase reaction may be interrupted in vertebrate cells
by damage to the substrate sites, probably as a result of
the activities of host DNA repair activities. With this
observation in mind, we examined the products of
integration reactions promoted by both the fC31 and
fBT1 integrases in S. pombe and have failed to detect any
evidence that reaction intermediates are liable to such
damage in these cells. It would seem unwarranted,
however, to draw the conclusions that site damage may
not limit the efficiencies of these enzymes in S. pombe or
that the fBT1 integrase reaction may not be liable to
being interrupted by site damage in vertebrate cells. In the
experiments of Malla and colleagues (20) the reaction that
was being analysed was occurring between fC31 sites that
were  1Mb apart and thus it may have been particularly
slow due to the difficulties of maintaining the sites in
alignment. A slow reaction may increase the likelihood
that the double strand breaks in the reaction intermediates
are recognized by the host cell DNA repair activities.
e9 Nucleic Acids Research, 2008, Vol. 36, No. 1 PAGE 8 OF 9

Thus, further work will be required to investigate the
efficiency of the fBT1 integrase and other serine
recombinases for promoting long range re-arrangements
and to determine how liable the reaction intermediates are
to damage arising from the activities of host DNA repair
activity in different cell types.
In summary, the practical conclusion form our work is
that we have established a new serine recombinase as a
tool for genome manipulation. The availability of two well
characterized unidirectional recombinases creates the
possibility of creating long range substitutions in eukary-
otic genomes using genome manipulation techniques.
SUPPLEMENTARY DATA
Supplementary Data are available at NAR Online.
ACKNOWLEDGEMENTS
Funding to pay the Open Access publication charges for
this article was provided by the BBSRC under the
Exploiting Genomics Initiative and the EU (Mechanisms
of Gene Integration).
Conflict of interest statement. None declared.
REFERENCES
1. Sauer,B. (1987) Functional expression of the cre-lox site-specific
recombination system in the yeast Saccharomyces cerevisiae. Mol.
Cell Biol., 7, 2087–2096.
2. Golic,K.G. and Lindquist,S. (1989) The FLP recombinase of yeast
catalyzes site-specific recombination in the Drosophila genome.
Cell, 59, 499–509.
3. Hoess,R.H., Wierzbicki,A. and Abremski,K. (1986) The role of the
loxP spacer region in P1 site-specific recombination. Nucleic Acids
Res., 14, 2287–2300.
4. Soukharev,S., Miller,J.L. and Sauer,B. (1999) Segmental genomic
replacement in embryonic stem cells by double lox targeting.
Nucleic Acids Res., 27, e21.
5. Thorpe,H.M. and Smith,M.C.M. (1998) In vitro site-specific inte-
gration of bacteriophage DNA catalyzed by a recombinase of the
resolvase/invertase family. Proc. Natl Acad. Sci. USA, 95, 5505–5510.
6. Thyagarajan,B. et al. (2001) Site-specific genomic integration in
mammalian cells mediated by phage fC31 integrase. Mol. Cell Biol.,
21, 3926–3934.
7. Belteki,G. et al. (2003) Site-specific cassette exchange and germline
transmission with mouse ES cells expressing fC31 integrase. Nat.
Biotechnol., 21, 321–324.
8. Smith,M.C. and Thorpe,H.M. (2002) Diversity in the serine
recombinases. Mol. Microbiol., 44, 299–307.
9. Nkrumah,L.J. et al. (2006) Efficient site-specific integration in
Plasmodium falciparum chromosomes mediated by mycobacter-
iophage Bxb1 integrase. Nat. Methods., 3, 615–621.
10. Wallace,H.A. et al. (2007) Manipulating the mouse genome to
engineer precise functional syntenic replacements with human
sequence. Cell, 128, 197–209.
11. Dafhnis-Calas,F. et al. (2005) Iterative in vivo assembly of large
and complex transgenes by combining the activities of fC31
integrase and Cre recombinase. Nucleic Acids Res., 33, e189.
12. Izumi,M. et al. (1991) Blasticidin S-resistance gene (bsr): a novel
selectable marker for mammalian cells. Exp. Cell Res., 197,
229–233.
13. Lupton,S.D. et al. (1991) Dominant positive and negative selection
using a hygromycin phosphotransferase-thymidine kinase fusion
gene. Mol. Cell Biol., 11, 3374–3378.
14. Niwa,H., Yamammura,K. and Miyazaki,J. (1991) Efficient selection
for high-expression transfectants with a novel eukaryotic vector.
Gene, 108, 193–201.
15. Brown,K.E. et al. (1994) Dissecting the centromere of the human Y
chromosome with cloned telomeric DNA. Hum. Mol. Genet., 3,
1227–1237.
16. Yang,J.W. et al. (2000) Human mini-chromosomes with minimal
centromeres. Hum. Mol. Genet., 9, 1891–1902.
17. Bahler,J. et al. (1998) Heterologous modules for efficient and
versatile PCR-based gene targeting in Schizosaccharomyces pombe.
Yeast, 14, 943–951.
18. Thomason,L.C., Calendar,R. and Ow,D.W. (2001) Gene insertion
and replacement in Schizosaccharomyces pombe mediated by the
Streptomyces bacteriophage fC31 site-specific recombination
system. Mol. Genet. Genomics, 265, 1031–1038.
19. Andreas,S. et al. (2002) Enhanced efficiency through nuclear
localization signal fusion on phage fC31 -integrase: activity
comparison with Cre and FLPe recombinase in mammalian cells.
Nucleic Acids Res., 30, 2299–2306.
20. Malla,S. et al. (2005) Rearranging the centromere of the human Y
chromosome with fC31 integrase. Nucleic Acids Res., 33, 6101–6113.
21. Loonstra,A. et al. (2001) Growth inhibition and DNA damage
induced by Cre recombinase in mammalian cells. Proc. Natl Acad.
Sci. USA, 98, 9209–9214.
22. Ehrhardt,A. et al. (2006) Molecular analysis of chromosomal
rearrangements in mammalian cells after fC31 -mediated integra-
tion. Hum. Gene. Ther., 17, 1077–1094.
23. Wood,V. et al. (2002) The genome sequence of
Schizosaccharomyces pombe. Nature, 415, 871–880.
24. Raymond,C.S. and Soriano,P. (2007) High-efficiency FLP
and fC31 site-specific recombination in mammalian cells. PLoS
ONE, 2, e162.
PAGE 9 OF 9 Nucleic Acids Research, 2008, Vol. 36, No. 1 e9