CHAPTER 2
A Diversity of Serine Phage Integrases Mediate
Site-Specific Recombination in Mammalian Cells
Annahita Keravala, Amy C. Groth,

Sohail Jarrahian,

Bhaskar Thyagarajan,
Jason J. Hoyt, Patrick J. Kirby,

and Michele P. Calos
2006. Molecular Genetics and Genomics. 275:135-46
Reprinted with permission from Springer Science and Business Media
In this study, I performed all recombination characterization experiments for
phiRV1 and phiC31 comparisons.

17
ABSTRACT
This study evaluated the ability of five serine phage integrases, from phages
A118, U153, Bxb1, phiFC1, and phiRV1, to mediate recombination in mammalian
cells. Two types of recombination were investigated, including the ability of an
integrase to mediate recombination between its own phage att sites in the context of a
mammalian cell and the ability of an integrase to perform genomic integration pairing a
phage att site with an endogenous mammalian sequence. We demonstrated that the
A118 integrase mediated precise intra-molecular recombination of a plasmid containing
its attB and attP sites at a frequency of ~50% in human cells. The closely related U153
integrase also performed efficient recombination in human cells on a plasmid
containing the attB and attP sites of A118. The integrases from phages Bxb1, phiFC1,
and phiRV1 carried out such recombination at their attB and attP sites at frequencies
ranging from 11 - 75%. Furthermore, the A118 integrase mediated recombination
between its attP site on a plasmid and pseudo attB sites in the human genome, i.e.
native sequences with partial identity to attB. Fifteen such A118 pseudo att sites were
analyzed, and a consensus recognition site was identified. The other integrases did not
mediate integration at genomic sequences at a frequency above background. These site-
specific integrases represent valuable new tools for manipulating eukaryotic genomes.

18
INTRODUCTION
Based on evolutionary and mechanistic relationships, most site-specific
recombinases can be classified into the tyrosine or serine recombinase families. The
site-specific recombinases from both families are structurally and functionally diverse
(1). Microbial recombinases such as Cre and FLP use a catalytic tyrosine in their strand
exchange mechanism (2) and recognize DNA sequences that are typically 30-40 bp in
length. Cre and FLP require no host-specific co-factors and have been successfully
used in mammalian cells, providing important and widely used tools for genome
manipulation (3, 4, 5). Other site-specific recombinases, such as , use a catalytic
serine and are members of the evolutionarily unrelated serine recombinase family (2).
Cre, FLP, and are resolvases and catalyze bi-directional recombination between
identical sites. They are most useful for efficiently carrying out site-specific deletions
(3, 4). By contrast, phage integrases mediate unidirectional recombination between two
dissimilar attachment (att) sequences, the bacterial attB and phage attP sites, resulting
in hybrid sites, attL and attR (6). Because these integrases cannot carry out the reverse
excision reaction without additional cofactors, these enzymes are especially helpful for
catalyzing integration reactions.
Only a subset of phage integrases, including those from phages phiC31, R4, and
TP-901-1, belong to the serine recombinase family (7, 8, 9). phiC31 integrase, isolated
from a Streptomyces phage (10, 11), was shown to mediate intra-molecular
recombination of plasmids in E. coli and in vitro, requiring no host-specific co-factors
(7). We have shown that the ~605-amino acid phiC31 integrase can perform
recombination between minimal 34-bp attB and 39-bp attP sites in human cells (12) and
mediates stable, site-specific integration of plasmids bearing attB into attP sites
randomly integrated into the genomes of cultured human and mouse cells (13).
Furthermore, phiC31 integrase facilitates integration of attB-bearing plasmids at

19
endogenous sequences with partial identity to attP, termed pseudo attP sites (13).
Transgenes integrated with phiC31 produce long-term, robust expression (13, 14). This
ability of phiC31 integrase to integrate into endogenous genomic sites has been used in
gene therapy experiments to produce therapeutic levels of Factor IX and
fumarylacetoacetate hydrolase in mouse liver (15, 16), to correct genetic deficiencies in
human keratinocytes (17, 18), and to produce dystrophin in mouse muscle-derived stem
cells, human myoblasts, and mouse muscle (19, 20), and luciferase in rat retina (32). In
addition, phiC31 integrase has been used in the construction and manipulation of
transgenic Drosophila (22), Xenopus (23), and mice (24, 25).
Based on these favorable results with phiC31, two other serine integrases, from
phages R4 (469 amino acids, 8) and TP901-1 (485 amino acids, 9), were evaluated in
mammalian cells. Both enzymes could perform intra-molecular recombination between
their attB and attP sites in the human cell environment (21, 27), but were less useful for
recombination into chromosomal pseudo att sites. The R4 integrase could integrate
efficiently at endogenous sites, but frequently appeared to produce aberrant
chromosomal events, while the TP901-1 integrase did not perform integration at native
sequences at a detectable frequency. The value of phiC31 integrase as a new tool for
genetics and gene therapy prompted us to explore additional prokaryotic serine
integrases for their utility in mammalian cells. We carried out the present study to test
the hypothesis that the serine integrases that have been identified in phages A118,
U153, Bxb1, fFC1, and phiRV1 might also be useful genetic engineering tools for
mammalian cells.
These further integrases were isolated from phages of a variety of gram-positive
bacteria. The integrase from Listeria monocytogenes phage A118 (28) is a 452 amino
acid protein that is 50% similar to the TP901-1 integrase from Lactococcus lactis subsp
cremoris (29). U153 is also a bacteriophage of L. monocytogenes (30) and possesses a

20
closely related 452 amino acid serine integrase (31). The U153 and A118 integrases
shares 89% DNA sequence identity and 93% identity at the protein level (31). These
two similar integrases share a common attB site, and their attP sites are 94% identical
(31).
Mycobacteriophage Bxb1 is a temperate phage of Mycobacterium smegmatis
(32, 33). Integration of the Bxb1 genome into the mycobacterial genome is mediated by
the 500-amino acid Bxb1 integrase that catalyzes site-specific recombination between
its attP and attB sites (33, 34, 35).
Bacteriophage phiFC1 infects Enterococcus faecalis KBL 703 and integrates
into the host chromosome by a site-specific mechanism (36). The phiFC1 integrase is
465 amino acids long (36) and shares 57% overall similarity with the integrases from
phages A118 (28) and TP901-1 (29). The phiFC1 integrase requires only its attB and
attP sites to carry out integration in E.coli as well as in E. faecalis (36).
The prophage element phiRV1 resides within the Mycobacterium tuberculosis
genome (37). A “resurrected” non-replicative form of the phage was shown to
efficiently integrate within the non-host strain Mycobacterium bovis BCG (37). phiRV1
contains a 469 amino acid serine integrase that shares 27% identity with the Bxb1
integrase (38, 39).
It was previously unknown if any of these five serine integrases were capable of
mediating site-specific integration in mammalian cells. That question was answered in
these studies.

21
MATERIALS AND METHODS
Plasmids. The A118 integrase gene and attP were amplified by PCR from A118
phage genomic DNA (generous gift of Richard Calendar, U.C. Berkeley). The
integrase gene was amplified by using the following primers: 5’-
GTCTAGAATGAAGGCAGCTATTTATATAC-3’ and 5’-
GGATCCTAGAGCCATTCAATAG-3’. Primers used to amplify attP were 5’-
GTTGATTTTGACTATTTAGAGTTC-3’ and 5’-
TTATTTGCACAAGAAGTGAAATATAAATC-3’. The attB site was amplified by
PCR from Listeria monocytogenes genomic DNA (R. Calendar). The PCR primers for
attB were 5’-TAAACCCGCACACGATG-3’ and 5’-
GATAATCTTTAATATGTTGTGGG-3’. All three sequences were TOPO-cloned
using the TOPO-cloning kit from Invitrogen (Carlsbad, CA), creating the plasmids
pTA-A1, pTA-A1attP, and pTA-A1attB. The integrase gene was removed by digestion
with SpeI and BamHI and cloned into the SpeI and BamHI sites of pInt (12), replacing
the fC31 integrase gene and creating the plasmid pIntA1 (Fig. 2-1A). The plasmids
pBCPB-A1+ (Fig. 2-1B) and pBCPB-A1- (Fig. 2-1C), containing the lacZ gene flanked
by the att sites in a direct or inverted orientation, respectively, were created as follows.
The attB was removed from the plasmid pTA-A1attB by digestion with BamHI and
XhoI and cloned into the BamHI and XhoI sites of pBCbGal (12), creating the plasmid
pBCB-A1. The attP was then removed from pTA-A1attP by EcoRI digestion, blunted
with T4 DNA polymerase, and cloned into the SmaI site of pBCB-A1 in both
orientations, to create pBCPB-A1+ and pBCPB-A1-.
The plasmid pCMV-A1 (Fig. 2-1D) was created by removing the integrase gene
from pTA-A1 by BamHI and SpeI digestion and cloning the fragment into the BamHI
and XbaI sites of pCMVInt (12), replacing the phiC31 integrase gene. The A118 attB
and attP were removed from their respective pTOPO plasmids by EcoRI digestion,

22
blunted and cloned into the blunted BamHI site of pMSEUZ1 (13) to create the
hygromycin resistant donor plasmids for rescue experiments, pA1-B and pA1-P (Fig. 2-
1E).
The U153 integrase gene was isolated by PCR from the pPL1 plasmid (31, kind
gift of Richard Calendar, U.C. Berkeley) by using the following primers: U153- 5`-
GAAATGAAGGCAGCTATTTATATACGCGTAT and U153- 3` -
CTAGAGCCATTCAATAGTAACTTGTTCA. The PCR product was TOPO-cloned to
give the plasmid pTA-U1. pTA-U1 was digested with EcoRI to release the U153
integrase gene. In parallel, the pCS (15) plasmid was digested with EcoRI, and the
fragment containing the U153 integrase gene was ligated into the EcoRI-digested pCS
to give pCMV-U1 (similar to Fig. 2-1D).
The U153 attP site was also isolated by PCR from the pPL1 plasmid, by using
the following primers: U153attP-5` - TAGCTTGTTTATTTAGATTGTTTAGTT and
U153attP-3` - GCATGCGATAAAAAGCAATCTATAGAAAAACA. This PCR
product was TOPO-cloned to generate the plasmid pTA-U1attP. This plasmid was
digested with EcoRI to release the attP-containing fragment that was then blunted and
ligated into the HindIII/SacI-digested and blunted pA1-P to give the pU1-P plasmid
(similar to Fig. 2-1E).
The Bxb1 integrase gene was amplified from Bxb1 phage DNA (gift of W.R.
Jacobs, Albert Einstein College of Medicine) by using the primers: Int-F – 5`-
ATGAGAGCCCTGGTAGTCATC-3` and Int-R – 5`-
CTACGACATCCCGGTGTGTAG-3`. The fragment was then TOPO-cloned with the
TOPO-cloning kit from Invitrogen, creating the plasmid pTA-Bx. This plasmid was
then digested using EcoRI to release the Bxb1 integrase gene. In parallel, pCS (15) was
also digested with EcoRI to linearize it. The 1500 bp fragment containing the Bxb1

23
integrase gene was ligated to the linearized pCS creating pCMV-Bx (similar to Fig. 2-
1D). The integrase gene sequence was confirmed by DNA sequencing.
The plasmid pBx-P was generated by cloning the Bxb1 attP site into the pHZ-
attB plasmid (13). A 78-bp attP fragment was synthesized by using the following
oligos with HindIII/SacI overhangs: attP-F - 5`-
CTATGGCCGTGATGACCTGTGTCTTCGTGGTTTGTCTGGTCAACCACCGCGG
TCTCAGTGGTGTACGGTACAAACCCA-3`; attP-R – 5`-
AGCTTGGGTTTGTACCGTACACCACTGAGACCGCGGTGGTTGACCAGACAA
ACCACGAAGACACAGGTCATCACGGCCATAGAGCT-3`. The annealed attP was
then digested with HindIII /SacI, and in parallel pHZ-attB was digested with
HindIII/SacI. The Bxb1 attP was cloned into the pHZ-attB plasmid replacing the
phiC31 attB, generating pBx-P (similar to Fig. 2-1E).
Similarly, pBx-B was cloned by inserting the Bxb1 attB into the pHZ-attB
plasmid. A 78-bp attB fragment was synthesized by using oligos with HindIII/ SacI
overhangs: attB-F – 5`-
CTGGCCGTGGCCGTGCTCGTCCTCGTCGGCCGGCTTGTCGACGACGGCGGTC
TCCGTCGTCAGGATCATCCGGGCCAC-3` and attB-R - 5`-
AGCTGTGGCCCGGATGATCCTGACGACGGAGACCGCCGTCGTCGACAAGCC
GGCCGACGAGGACGAGCACGGCCACGGCCAGAGCT-3`. The annealed attB was
digested with HindIII/SacI to generate the overhangs, and pHZ-attB was digested with
HindIII/SacI. The Bxb1 attB was then cloned into pHZ-attB, replacing the phiC31 attB
to create pBx-B (similar to Fig. 2-1E).
The plasmid pBCPB-Bx+, containing the lacZ gene flanked by theBxb1 attP
and attB sites in a direct orientation, was created as follows. pBC Gal (12) was digested
with SmaI to linearize it. In parallel pBx-P was digested with MscI to release the attP
fragment. This attP fragment was ligated into the SmaI site on pBC Gal to generate

24
pBCP-Bx. The plasmid pBCPB-Bx+ (similar to Fig. 2-1B) was generated by cloning
the MscI fragment of pBx-B containing the Bxb1 attB site into NaeI-digested pBCP-Bx.
Orientation of attP and attB was confirmed by DNA sequencing.
A human codon-optimized phiFC1 integrase gene was synthesized by DNA 2.0,
Inc. (Menlo Park, CA) using the published phiFC1 amino acid sequence (36, Genbank
accession number AF124258). This gene was then cloned into the SalI/BamHI sites of
pCMVSportbGal (Invitrogen) to generate pCMV-FC. The plasmid pBCPB-FC+
(similar to Fig. 1B) was generated by cloning phiFC1 attB and attP sequences into the
plasmid pBCbGal (12) in direct orientation, such that a recombination event between
the two sites would result in the excision of the lacZ gene. The oligonucleotides used
were 5’-
ttatccaacagaaaacgattttcaaaggttcactgaatcaaAGTattgctcatccacgcgaaatttttcgggaagcggttaaat
at-3’ and 5’-aatgtaatattttaggtatatgattttgtttattAGTgtatataacactatgtactaaaatttatatttaat-3’ for
the phiFC1 attB and attP respectively. The letters in upper case indicate the central
cores of the att sites. The plasmids pEGFP-FCB and pEGFP-FCP (Fig. 1F) were
generated by cloning in the phiFC1 attB or attP oligonucleotide into the MluI site of
pEGFP-C1 (Clontech BD Biosciences).
The phiRV1 integrase gene was amplified by PCR from H37RV genomic DNA
(ATCC, Manassas, VA) by using the Rv1586c forward and reverse primer set (37).
This fragment was cloned into the EcoR1 site of pCMV-Sport vector (Invitrogen) to
create pCMV-RV. The intramolecular integration assay plasmid pBCPB-RV+ was
constructed as follows. A 370-bp fragment containing the attP site for phiRV1 was
recreated from the M. tuberculosis genome (ATCC) in two steps. First, the primer pairs
5’-GGTTGGCCGTGGACTGCTG and 5’-AGCCGAACGAGCTCTTCCCTCAC and
primer pairs 5’-GCGCTGAATTCGTTGTCGAGGTAC and 5’-
CCACGGCCAACCGTGGACCTG were used. These two products then served as

25
templates for the second PCR amplification step by using primer pairs 5’-
GCGCTGAATTCGTTGTCGAGGTAC and 5’-AGCCGAACGAGCTCTTCCCTCAC.
This PCR product was TOPO-cloned into pCR2.1 (Invitrogen) to create pTA-RVattP.
The attB site was amplified by PCR from the M. tuberculosis genome by using primers
5’-GCATTAATAGCTCTCGTGGTGGTGGAAGG and 5’-
GCCCGCGGCGGTCGGGGTACCAAGCTCGA. This fragment was ligated into
pCR2.1 to create pTA-RVattB. The attP site was removed from pTA-RVattP by
digesting with BamHI and XhoI and cloned into pBC Gal [10], to create pBC Gal-
RVattP. The attB portion was obtained by digesting pTA-RVattB with AseI and SacII
and cloning into pBC Gal-RVattP to create the final assay plasmid pBCPB-RV+, in
which the two phiRV1 att sites are in direct orientation (similar to Fig. 2-1B).
Bacterial intra-molecular integration assay. Electrocompetent bacteria were
created carrying the pIntA1 plasmid in DH10B E. coli. 2 ng of pBCPB-A1+ or
pBCPB-A1- were transformed into the integrase-expressing cells, spread on LB-agar
plates containing chloramphenicol (Sigma, MO) and 5-bromo-4-chloro-3-indolyl- -D-
galactopyranoside (X-gal, Invitrogen), and incubated at 37° C. White and blue colonies
were counted at ~16 hours. The percent recombination was calculated by dividing the
number of white colonies by the total number of colonies and multiplying by 100.
Transient intra-molecular integration assay (excision assay) in human cells.
Human 293 cells were grown to ~60-80% confluency in Dulbecco’s Modified Eagle
Medium (Gibco, Carlsbad, CA) supplemented with 9% fetal bovine serum (Gibco) and
1% penicillin-streptomycin (Gibco). Cells were transfected with 5 µg pCMV-A1
and/or 200 ng pBCPB-A1+ and salmon sperm DNA to 5.2 µg. DNA was combined
with 100 µl serum-free medium and 15.6 µl FuGENE6 (Roche, Indianapolis, IN),
allowed to incubate for 15-30 minutes, then added to a 60-mm dish of cells. Cells were

26
fed with fresh medium containing 50 units/ml DNase at 24 hours and harvested at 72
hours. Low molecular weight DNA was recovered by Hirt extraction, (40) transformed
into electrocompetent bacteria, and spread on plates supplemented with
chloramphenicol and X-gal.
Similar experiments were done with the U153 integrase, using the A118
excision assay plasmid. The att sites for the two integrases are similar, and an inversion
assay using U153 integrase had been successful with an inversion assay plasmid that
carried the A118 attB and attP sites (data not shown). 60-80% confluent 293 cells in 6-
well dishes were transfected with 2 mg pCMV-U1 and 100 ng pBCPB-A1+ using
FuGENE6. DNA was combined with 100 ml serum-free media and 6 ml FuGENE6
and allowed to incubate for 15 minutes at room temperature. At the end of incubation,
the complexed DNA was added to the cells that were in complete DMEM containing L-
glutamine, 9% fetal calf serum and 1% penicillin-streptomycin and incubated for 24 h.
Medium was replaced with DMEM supplemented with 50 units/ml DNase and the cells
were incubated for another 48 h, after which they were harvested by using the Hirt
extraction buffer. Cellular DNA was isolated from the cells using Qiagen DNeasy kits
(Qiagen, Valencia, CA). The DNA isolated using this method results in recovery of
both genomic DNA and plasmid DNA. This DNA was then transformed into strain
DH10B of E. coli and plated on LB-agar plates containing chloramphenicol and X-gal.
The ratio of white colonies (generated by a recombination event) to the total number of
colonies was the extrachromosomal recombination frequency.
For the Bxb1 and phiRV1 excision assays, 293 cells in 6-well plates were
transfected with 100 ng of pBCPB-Bx+ or pBCPB-RV+ and 1 µg of either pCMV-Bx
and pCMV-RV, respectively, or pCS by using FuGENE6. Cells were treated with
DNase at 24 h and then harvested at 72 hours after transfection. DH10B cells were

27
transformed with the harvested DNA and plated onto LB-agar plates containing
chloramphenicol and X-gal.
For the phiFC1 excision assay, 293 cells in 6-well plates were co-transfected
with 20 ng of pBCPB-FC+ and 1 mg pCMV-FC or pCS by using FuGENE6. The cells
were similarly treated with DNase, harvested, and the DNA transformed into E.coli
cells and plated onto LB-chloramphenicol X-gal agar plates. As a positive control for
all these enzymes, a similar experiment was done with phiC31, using pCMVInt and
pBCPB+ (12).
Assay for integration at chromosomal pseudo sites. Human 293 cells were co-
transfected with 5 µg pCMV-A1 or 5 µg of salmon sperm carrier DNA and 1 µg of
either pA1-B or pA1-P. Cells were split to four 100-mm plates 24 hours after
transfection and put under selection with 200 µg/ml hygromycin (Sigma) at 48 hours
post-transfection. Cells were selected for ~3 weeks, then assessed for colony number.
Similar experiments were done in mammalian cells using the U153, Bxb1, and phiRV1
integrases. Donor plasmids expressing the hygromycin resistance gene and having the
attP or attB of the respective enzyme were co-transfected with the plasmid expressing
the respective integrase, using FuGENE6.
For the phiFC1 integration assay, the 293 cells were co-transfected using
FuGENE6, with pCMV-FC and pEGFP-FCB or pEGFP-FCP, donor plasmids
containing the attB or attP site and the gene expressing neomycin resistance. Twenty-
four hours after transfection, the cells were split into 4 100-mm plates, and 48 h post-
transfection the medium on the cells was replaced with DMEM containing 350 µg/ml
Geneticin (Gibco). Selective pressure was maintained for ~3 weeks, after which the
number of neo-resistant colonies were counted.
As a positive control for integration in all these experiments, parallel
transfections were done with the phiC31 plasmids, pCMVInt and pTB2 (a plasmid

28
derived from pHZ-attB (13) containing the phiC31 attB and the hygromycin resistance
gene) or, in the case of pCMV-FC, pDB2, a plasmid containing the phiC31 attB and the
neomycin resistance gene.
Pseudo att site sequence determination. Plates of hygromycin-resistant colonies
generated by the A118 integrase were pooled, and genomic DNA was prepared using
the Qiagen Blood and Cell DNA Maxi Kit (Qiagen, Inc., Valencia, CA). Pools were
analyzed for integration sites by either the plasmid rescue or GenomeWalker method.
DNA from selected cells transfected with 5 µg of pCMV-A1 and either 50 ng or 1 µg of
pA1-P was analyzed. Plasmid rescue was performed as follows. Genomic DNA was
digested with AatII, which cuts outside of the att site and plasmid rescue primer
sequences. Digested DNA was then ligated overnight at a concentration of 20 ng/µl.
Primary PCR was performed using primers that amplify across the ligated junction,
A118 rescue forward 1 (ACTATGCCGATGATTAATTGTC) and A118 rescue reverse
1 (CTAAACCATGGCCAAGTTGAC). Secondary PCR was then performed using
nested primers A118 rescue forward 2 (CGGGGTACCAAGCTTATCGATG) and
A118 rescue reverse 2 (GCCAAGTTGACCAGTGCCGTTCC). Secondary PCR
products were separated by electrophoresis, extracted, and TOPO cloned. DNA from
individual TOPO colonies was then prepared, sequenced, and subjected to BLAT
(http://genome.ucsc.edu/) and/or BLAST (http://www.ncbi.nlm.nih.gov/BLAST/)
searches.
Other pseudo att site sequences were characterized by GenomeWalker analysis.
Experiments were conducted as described in the GenomeWalker manual (BD
Biosciences, Palo Alto, CA). The gene-specific primer sequences were as follows:
GSP1 (CTATGCCGATGATTAATTGTCAACACG), GSP2
(GGGTACCAAGCTTATCGATGGATC). Individual TOPO clones were subjected to
sequence analysis and BLAT and/or BLAST searches.

29
Pseudo site analysis was not performed for the remaining four integrases, U153,
Bxb1, phiFC1, and phiRV1, because the numbers of hygromycin or neomycin-resistant
colonies were not above the background without the integrase.
Bioinformatics. The stand-alone version of the MEME/MAST motif-discovery
program was used to search for motifs of length 6–40 bp on both strands
(http://meme.nbcr.net). The settings “zero or one occurrence per sequence” (zoops) was
used. Wild-type A118 attB and 12 pseudo attB sites were used as sequence inputs.
Nucleotides were then pseudo-colored to represent their probability of occurrence in the
consensus motif. WebLogo version 3.0 was used to generate the motif sequence logo
(http://weblogo.berkeley.edu/logo.cgi).

30
RESULTS
Bacterial intra-molecular recombination. To test how well the A118 integrase
functioned when removed from the L. monocytogenes environment, an intra-molecular
recombination assay was performed in E. coli strain DH10B. The plasmid pIntA1 (Fig.
2-1A), expressing the A118 integrase under the control of the lac promoter, was
transformed into DH10B cells. These cells carrying pIntA1 were transformed with
pBCPB-A1+, which contains the lacZ gene flanked by the A118 attP and attB sites in
direct orientation (Fig. 2-1B). A recombination event would excise out the lacZ gene,
resulting in a white colony. Unrecombined plasmid would result in a blue colony. In
this assay, the A118 integrase recombined 65% of the pBCPB-A1+ plasmids (1977
white/3034 blue). Thirty of the resulting white colonies were analyzed by PCR, and all
had the expected recombinant junction between the A118 attB and attP sites. Plasmid
DNA extracted from three of these colonies was sequenced in the area of the
recombination junction. All had the precise expected crossover event. Control
experiments were done with pBCPB-A1+ in DH10B cells or in DHIntA1 cells with the
inversion plasmid pBCPB-A1-, in which a recombination event would merely invert
rather than excise the lacZ gene (Fig. 2-1C). No white colonies resulted from the
transformation of pBCPB-A1+ into DH10B or the transformation of pBCPB-A1- into
DHIntA118 cells (0 white/1739 blue). We concluded that the A118 integrase
functioned well in the E. coli environment by the assay used, although not as well as
related integrases from phages phiC31 and TP901-1, which recombined their att sites at
~100% efficiency in a similar assay (12, 27).
Mammalian intra-molecular recombination. In order to determine whether the
serine integrases from phages A118, U153, Bxb1, phiFC1, and phiRV1 could function
in the mammalian cell environment, a transient intra-molecular recombination assay
was performed for each of them. Each integrase expression plasmid, pCMV-A1 (Fig.

31
2-1D), pCMV-U1, pCMV-Bx, pCMV-FC, or pCMV-RV, was co-transfected into
human tissue culture cells with the corresponding recombination assay plasmid,
pBCPB-A1+ (Fig. 2-1B), pBCPB-A1- (Fig. 2-1C), pBCPB-Bx+, pBCPB-FC+, or
pBCPB-RV+. The cells were harvested at 72 hours and the low molecular weight DNA
was transformed into E. coli. As in the bacterial assay, a recombination event would
result in a white colony, while an unrecombined DNA molecule would yield a blue
colony.
For A118, eight separate transfections were performed in 293 cells, resulting in
an average recombination frequency of ~50% (Table 2-1). These numbers were likely
to be underestimates, because untransfected donor plasmid still contributes to the total
number of blue colonies, but could not have undergone a recombination event. By PCR
screen, 29 of 30 white colonies had the expected recombinant band. DNA derived from
four of these bands was sequenced, and all four had the perfect crossover sequence,
precise to the base (data not shown). In order to ensure that the recombination occurred
in the mammalian cells and not the bacteria, a control transformation was performed.
Plasmids pBCPB-A1+ and pCMV-A1 were co-transformed into DH10B cells at the
same ratio used for the mammalian transfections. Plates were then scored for blue and
white colonies. No white colonies were found in 2,133 colonies, suggesting that the
pCMV-A1 plasmid did not express in bacteria and that the white colonies observed in
the 293 experiments were the result of recombination in the mammalian cells. The low
level of white colonies seen in the donor-only transformation was likely due to
mutations in transfected DNA caused by the mammalian cell environment and has been
seen in comparable experiments with other serine integrases (12, 26, 27).
Similar experiments using U153, Bxb1, phiFC1, and phiRV1 integrases revealed
that they too were capable of catalyzing extra-chromosomal recombination between
their att sites in 293 cells. In three separate transfections for each enzyme, U153, Bxb1,

32
phiFC1 and phiRV1 integrases showed average recombination frequencies of
approximately 31%, 75%, 46% and 11% respectively. The numbers of recombination
events in the absence of the respective integrases were negligible (< 1%), as shown in
Table 2-1. PCR screens showed that plasmid DNA extracted from several white
colonies for each enzyme had the expected recombinant band, and DNA sequencing of
these bands showed a perfect crossover sequence in each case (data not shown).
Therefore, serine integrases A118, U153, Bxb1, phiFC1, and phiRV1 mediated intra-
molecular recombination between their att sites in the human cell environment.
Mammalian inter-molecular recombination into pseudo att sites. We tested
whether A118 integrase could mediate integration at pseudo att sites in the human
genome. Several experiments were conducted in which hygromycin-resistant plasmids
carrying the A118 attB site (pA1-B; Fig. 2-1E) or the A118 attP site (pA1-P; Fig. 2-1E)
were transfected into human 293 cells with or without the A118 integrase-expressing
plasmid pCMV-A1 (Fig. 2-1D). Each transfection was subjected to hygromycin
selection, and colony number was assessed. In one such experiment, 3-fold more total
colonies were observed when the pA1-B plasmid was transfected with the integrase
plasmid, compared to pA1-B with control salmon sperm DNA lacking the integrase
gene (Fig. 2-2A). 37-fold more colonies were observed when pA1-P was transfected
with the A118 integrase-expressing plasmid, compared to pA1-P transfected with
control salmon sperm DNA. Although the total colony numbers were relatively low,
these results suggested that the A118 integrase recombined an attP-containing plasmid
with pseudo attB sites in the 293 genome, while integration of an attB-containing
plasmid into pseudo attP sites was less efficient. Transfections were done with the
phiC31 integrase-expressing plasmid pCMVInt and attB-bearing, hygromycin-resistant
plasmid pTB2 alongside the A118 transfections. The phiC31 transfections yielded
higher total colony numbers and a 13-fold increase with integrase compared to without

33
integrase. This value for phiC31 was similar to that reported in previous experiments
(13). In other experiments, an increase above background in colony numbers with A118
integrase, using either pA1-B or pA1-P, was not seen.
Using plates of human 293 cells that had been transfected with pA1-P and
selected with hygromycin, plasmid rescue and GenomeWalker methods were used to
determine the attR junction sequences between the P (5’) portion of the plasmid attP
site and the B’ (3’) portion of the putative genomic pseudo attB integration sites. These
sequences were then used to search the human genome database and determine the
sequence of the full genomic pseudo site. Fifteen A118 pseudo attB sites were
characterized (Fig. 2-3). Eleven of these pseudo attB sites had perfect crossovers,
meaning that they had the perfect attP sequence up to and including the GG core
leading directly into human genomic sequence. Forty base pairs of each of these eleven
sites were aligned to the wild-type A118 attP, with the GG core in the center (Fig. 2-3).
Along the entire 40 base pairs, these pseudo attB sites ranged in identity from 35% to
52.5% with the wild-type A118 attB, with an average identity of 45%. In addition to
the GG core, two other base pairs in the pseudo attB sites were invariant. Fourteen
other base pairs were highly conserved, recurring in at least 8 of the 12 sequences (the
wild-type attB plus the eleven pseudo attB sites with perfect crossovers). A consensus
sequence was derived from these invariant and conserved bases (Fig. 2-3). The
remaining four pseudo attB sites were not included in the alignment. Because the
crossovers were imperfect, it was unclear which bases the enzyme was recognizing.
Using the MEME and MAST motif finder, we performed a statistical analysis of
the consensus motif. MEME and MAST were created for the de novo discovery of
motifs in genomic sequences (41). Wild-type A118 attB and the pseudo attB sites
shown in figure 2-3 were used as the input sequences in the MEME/MAST motif
finder. A significant motif (E-value: 1.4e-36) was found in 12/13 sequences analyzed.

34
The E-value is an estimate of the expected number of motifs with the given log
likelihood ratio (or higher), and with the same width and number of occurrences, that
one would find in a similarly sized set of random sequences. Pseudo site 15q26.2 was
the only site to not match the derived consensus motif and is not shown (Fig. 2-4).
To determine if the serine integrases from phages U153, Bxb1, phiFC1, and
phiRV1 could recombine at pseudo att sites in the human genome, we used a similar
chromosomal integration strategy. Donor plasmids containing a hygromycin or
neomycin drug-resistance marker and either the respective attB or the attP site (Fig. 2-
1E, F) were transfected into mammalian cells with a plasmid expressing the integrase or
the pCS control plasmid lacking an integrase gene. If the integrase could target pseudo
att sites, the expectation was that the number of drug-resistant colonies would be higher
in the presence of the integrase. However, as shown in Fig. 2-2B, this was not the case.
There was no significant difference in the number of colonies obtained in the absence or
presence of these integrases, suggesting that these enzymes did not mediate integration
at pseudo att sites in the human genome at a frequency above background.

35
DISCUSSION
This work investigated a group of serine integrases, from phages A118, U153,
Bxb1, phiFC1, and phiRV1, for their utility as tools for site-specific recombination in
mammalian cells. All five enzymes performed efficient and precise recombination
between their attB and attP sites in mammalian cells, while only A118 integrase
showed evidence of mediating integration into a native unmodified mammalian genome
at a measurable frequency above background.
Recombination between attB and attP sites present on an extrachromosomal
assay vector introduced into human cells occurred at easily detectable frequencies for
each integrase, ranging between 11% and 75%. These frequencies were comparable to
those seen for three other phage integrases from the serine site-specific recombinase
family (12, 26, 27). It is thus likely that all serine integrases are free of a requirement
for host-specific co-factors and can function autonomously in a wide variety of species
and cell types. phiC31 integrase, the best-studied member of this family, has been
shown to function in E. coli (7, 12) and in eukaryotes ranging from fission yeast (42), to
insects (22), amphibians (23), and a variety of mammalian species including human,
mouse, hamster, and rat (13, 14, 21). These additional integrases are likely to share this
broad host range for recombination activity.
The ability of a recombinase to perform an intra-molecular integration reaction
on an extrachromosomal plasmid predicts that the enzyme will also have the ability to
perform the reaction when both att sites are embedded in a mammalian chromosome, as
demonstrated by our work with the phiC31, R4, and TP901-1 integrases (24).
Therefore, it is likely that the A118, U153, Bxb1, phiFC1, and phiRV1 integrases will
have utility in performing intra-molecular deletion reactions. These reactions, currently
most commonly carried out with Cre recombinase, are frequently used to catalyze
controlled rearrangements of integrated cassettes (3). The availability of the A118,

36
U153, Bxb1, phiFC1, and phiRV1 integrases for this purpose provides additional
alternatives to Cre, creating new options when more complex rearrangements are
required and more than one recombinase is needed. Further, the unidirectional aspect of
these serine integrases, as opposed to the bi-directional nature of reactions carried out at
loxP sites by Cre resolvase, may be an advantage in some experimental settings, such as
cassette exchange (25).
An efficient integration reaction between an incoming plasmid and native
genomic sequences is useful in gene therapy and construction of transgenic vertebrates.
The phiC31 integrase has been a valuable tool for such reactions (15, 16, 17, 18, 19, 20,
21, 24, 23). In search of further enzymes with this capability, we analyzed here whether
constructs bearing an antibiotic resistance gene and either an attB or an attP site for the
A118, U153, Bxb1, phiFC1, and phiRV1 integrases, when co-introduced with the
corresponding integrase, could bring about genomic integration at a frequency higher
than the background integration frequency in 293 cells, as measured by the number of
antibiotic resistant colonies.
Unlike the phiC31 integrase, the A118, U153, Bxb1, phiFC1, and phiRV1
integrases were generally negative in such an assay. However, in some experiments we
found that the A118 integrase could mediate integration into human genomic pseudo
attB sites partially identical to attB at frequencies significantly higher than random
integration. It is unclear why the A118 integrase functioned well in some
chromosomal integration experiments and showed no increase above background in
stable integrants at other times. This result suggests that the enzyme may be sensitive to
cell culture conditions. For example, if proliferating cells are favored for integration,
more rapidly dividing cultures may provide more opportunities for integration. The
genomic integration sites we identified here (Fig. 2-3) provide evidence that the A118
integrase possesses the capacity for integration at native human sequences.

37
Furthermore, we used the genomic DNA sequences at two of the identified integration
sites to develop PCR primers and used these primers to verify by PCR that integrants at
these sites were present in transfected populations of 293 cells that had received A118
integrase and A118 attP-neo plasmids (data not shown).
When A118 integrase catalyzed genomic integration, it preferentially mediated
recombination of a plasmid-borne attP site into pseudo attB sites. By contrast, two
other serine integrases that have been found to mediate efficient genomic integration,
phiC31 (13) and R4 (26), favor recombination of plasmids bearing an attB site into
genomic pseudo attP sites. A possible explanation involves the unusual sequence
structure of the A118 attP site. The attP site is almost all purines (18/19) in the 3’
direction from the GG core and has 18/19 pyrimidines in the 5’ direction from the core
(Fig. 2-3). There may be few stretches of ~20 pyrimidines followed by ~20 purines in
the human genome. The attB sequence is less extreme in these respects, so it is possible
that for A118 integrase it is easier to find near matches to attB than to attP in the
genome.
Most of the A118 integration sites (12/15; 80%) were in introns, which is
significantly more than the ~24% expected on the basis of random integration and
suggests a bias in favor of integration into genes. By contrast, ~37% of phiC31
integration sites were in introns (43). Furthermore, most A118 crossover sites were
perfect (11/15; 73%), whereas, the majority of phiC31-mediated integration events at
pseudo att sites were accompanied by small deletions or insertions (13, 43). Compared
to the phiC31 integrase, A118 integrase seemed to have a stronger sequence identity
requirement for recombination. There were four invariant base pairs among the eleven
A118 pseudo sites and the wild-type site, including the GG core and two bases in the
left arm. By contrast, among the 19 most frequently used pseudo attP sites for the
phiC31 integrase in human cell lines, there were no positions that were totally

38
conserved, even in the TT core sequence (43). Therefore, A118 appears to be a more
precise, but less robust integration tool than phiC31 integrase.
The other serine recombinases, from phages U153, Bxb1, phiFC1, and phiRV1,
analyzed in this paper did not produce stable integrants above background in the
presence of an attP or attB bearing plasmid in any of the chromosomal integration
experiments performed. Most likely, these enzymes have a lesser affinity for or ability
to complete recombination at mismatched pseudo att sites in the genome, relative to
phiC31 integrase. It may be possible to increase integration frequency of these serine
recombinases into pseudo att sites by a directed evolution strategy to enhance reactivity
with a particular genomic sequence, as has been done for phiC31 integrase (44). Such
mutants could have excellent attributes for engineering of genomes and for gene
therapy. Short of that goal, the ability of all five of these serine integrases, from phages
A118, U153, Bxb1, phiFC1, and phiRV1, to carry out efficient and precise
recombination between their attB and attP sequences in mammalian cells makes them
valuable new genetic tools for the engineering of complex genomes.

39
Table 2-1. Intra-molecular recombination frequencies in 293 cells.
In this assay, an active recombinase was able to excise the lacZ gene flanked by the enzyme’s attB and
attP. Plasmid extracts were then transformed into E. coli and the number of white colonies (recombined
plasmids) and blue colonies (unrecombineed) were counted to mquantify recombination activity.
Plasmids
Transfected
White Colonies Total Colonies % Recombination
pBCPB-A1+ +
pCS
101 5389 1.9
pBCPB-A1+ +
pCMV-A1
6198 5389 49.9
pBCPB-A1+ +
pCMV-U1
468 1504 31.1
pBCPB-Bx+ +
pCS
1 146 0.7
pBCPB-Bx+ +
pCMV-Bx
304 408 74.5
pBCPB-FC+ +
pCS
1 518 0.2
pBCPB-FC+ +
pCMV-FC
242 528 45.8
pBCPB-RV+ +
pCS
2 776 0.3
pBCPB-RV+ +
pCMV-RV
64 604 10.6

40
Figure 2-1. Schematic diagrams of plasmids used to assay various serine
recombinases. A. pIntA1 expresses the A118 integrase in E. coli under the control of
the lacZ promoter. B. pBCPB-A1+ contains the lacZ gene flanked by the attB and attP
of A118 in direct orientation. Parallel constructions were made carrying the att sites of
the Bxb1, phiFC1, and phiRV1 integrases. C. pBCPB-A1- contains lacZ flanked by the
attB and attP sites of A118 in inverted orientation. D. pCMV-A1 expresses A118
integrase in mammalian cells under the control of the CMV promoter. Parallel
constructions were generated to express the integrases from phages U153, Bxb1,
phiFC1, and phiRV1. E. pA1-B and pA1-P contain the A118 attB or attP site and
express the hygromycin resistance gene. Parallel constructions were made with the
U153 attP site and the Bxb1 and phiRV1 attB or attP site. F. pEGFP-FCB and
pEGFP-FCP contain the phiFC1 attB or attP site and express the neomycin resistance
gene.

41
Figure 2-2. Analysis of chromosomal integration at pseudo att sites. A. The
average colony numbers from three experiments in human 293 cells are shown. pA1-B
and pA1-P, which contain the A118 attB and attP, respectively, were co-transfected
with the A118 integrase in the +integrase transfections. pTB2, which contains the
phiC31 attP, received the phiC31 integrase expressing plasmid pCMVInt in the
+integrase transfections. The fold difference for each plasmid + integrase over the
A.
B.

42
plasmid + carrier is shown in parentheses. B. The average number of hygromycin- or
G418-resistant colonies from three experiments are shown for the serine integrases from
phages phiC31, U153, Bxb1, phiFC1, and phiRV1. Human 293 cells received the
donor plasmid carrying either the attB or the attP of the respective integrase and were
maintained under selective pressure for 2 weeks. Values are + SEM.

43
Site
Sequence
% identity
with attB
Crossover Site Location
attB TTTCGGATCA AGCTATGAAG GACGCAAAGA GGGAACTAAA 100
attP TTCCTCGTTT TCTCTCGTTG GAAGAAGAAG AAACGAGAAA 35.0
1q21.1 ATATGTGTCC CACCATGGGG GTAGAAGGGA AGGGACAAAG 50.0 Perfect AF379632 intron
1q42.3 CTCCGGCACT ACCTCCCCTG GGCCGTGGGG TGGCAGCAAA 45.0 Perfect GNG4.c intron
2p21 AGTGGGTTCA AGCATCCTGG GGCAATGAGA AGGTCCAAAA 52.5 Perfect AK094343 intron
2q36.3 ATTTGGACCT GGGGTGGAAG GGTAGAGAGC AGGGTCAAAT 50.0 Perfect D28476 intron
4p16.3 CTGGGTCCCC TCTCTCCCCG GAAGGAGACC TGGAGGAAAA 37.5 Perfect AK056302 intron
5q21.3 GCTTGGTTCA TAGCAACAGG GGAGAAAAGG AGGCGCCCAT 47.5 Perfect U26403 intron
11p15.2 ATAGGGTTCA CTCTCCCTAG GGTCCAGAGA AGGGCAAAAG 50.0 Perfect X95520 intron
14q11.2 TTTGGCACCT CTCCTCCTGG GTCCCTACTA GGGGCCAGGG 42.5 Perfect Ak074167 intron
15q26.2 ATTAGTGCCT TTTCACAGTG GCAGGATAGA AAAGGTTTTA 35.0 Perfect Integenic
18q21.1 CTTGGGGTCT TTCTCTGTGG GGCAGAGAGT AGGGCCAAAG 52.5 Perfect CGBP intron
Xp22.22 GTCCGGCCCG GCCACCCCGG GGAGGGGTGG AGGTACAGGA 37.5 Perfect SEDL intron
consensus NTNNGGNNCN NNCNNNNNNG GGNNNAGAGN AGGNNCAAAN 37.5
11p15.4 TTCCGGGTCT CTGCGAGCTG GGAAGAGCCA GGGACCGAAT 47.5 Imperfect 5’ UTR Weel-a
1q21.2 Imperfect HPRP3P intron
12q22 Imperfect AK003118 intron
15q26.1 Imperfect integenic
Figure 2-3. A118 pseudo attB sequences and site locations. The sequence of the
wild-type A118 attB and attP sites are depicted, as well as 11 human pseudo attB
sequences from the human genome, named for their chromosomal positions. Underlined
bases match the wild-type A118 attB base at that position. A consensus sequence is
derived on the lower line. In instances where at least 8 of the 12 sequences (wild-type
attB or pseudo attB) have the same base at a position, it is included in the consensus
(see fig. 2-4 for detailed motif analysis). Positions that do not meet this criterion are
depicted by the letter N. Four positions that are invariant between the A118 attB site
and all the pseudo attB sites are shown in bold. Four pseudo sites with imperfect
crossovers where the sequence recognized was uncertain are shown at the bottom.

44
Figure 2-4. Consensus alignment of A118 pseudo attB sites. Wild-type A118 attB
and 12 human pseudo attB sequences were searched for a common motif by using the
MEME/MAST motif finder. A. The multilevel consensus sequence (E=1.4e-36) derived
from the motifs is shown at the top. The E-value is an estimate of the expected number
of motifs with the given log likelihood ratio (or higher), and with the same width and
number of occurrences, that one would find in a similarly sized set of random
Consensus
TTGGGATCT CTCCCCCCGG GGAGGAGAGA AGGGCCAAAA
CC CC A ACGTTTGTA CAATA G G AA G
T G TG A A CC T
T
Site P-value
18q21.1 9.15e-16
CTTGGGGTCT TTCTCTGTGG GGCAGAGAGT AGGGCCAAAG
2p21 1.31e-13
AGTGGGTTCA AGCATCCTGG GGCAATGAGA AGGTCCAAAA
11p15.2 2.35e-13
ATAGGGTTCA CTCTCCCTAG GGTCCAGAGA AGGGCAAAAG
attB 8.27e-12
TTTCGGATCA AGCTATGAAG GACGCAAAGA GGGAACTAAA
1q22.1 9.12e-11
ATATGTGTCC CACCATGGGG GTAGAAGGGA AGGGACAAAG
2q36.3 2.29e-10
ATTTGGACCT GGGGTGGAAG GGTAGAGAGC AGGGTCAAAT
11p15.4 1.20e-09
TTCCGGGTCT CTGCGAGCTG GGAAGAGCCA GGGACCGAAT
5q21.3 3.33e-09
GCTTGGTTCA TAGCAACAGG GGAGAAAAGG AGGCGCCCAT
1q42.3 3.57e-09
CTCCGGCACT ACCTCCCCTG GGCCGTGGGG TGGCAGCAAA
Xp22.22 9.43e-09
GTCCGGCCCG GCCACCCCGG GGAGGGGTGG AGGTACAGGA
14q11.2 1.00e-08
TTTGGCACCT CTCCTCCTGG GTCCCTACTA GGGGCCAGGG
4p16.3 3.25e-08
CTGGGTCCCC TCTCTCCCCG GAAGGAGACC TGGAGGAAAA
A.
B.

45
sequences. Nucleotides are color-coded, such that yellow indicates nucleotides with the
highest probability of occurring at that sequence position. Nucleotides highlighted in
green, blue, and gray represent decreasing probabilities of occurring in the motif. Bases
that are not color-coded are below the probability threshold of occurrence. Sites are
ranked in order of significance. The P-value of an occurrence is the probability of a
single random subsequence the length of the motif, generated according to the 0-order
Markov background model, having a score at least as high as the score of the
occurrence. Pseudo attB site 15q26.2 is not shown as it did not score above the motif
threshold.
B. A sequence logo diagram is shown for the MEME consensus motif. The
probability of a given base occurring at a position is represented by the size of the letter.
Positions 20 and 21 represent the GG dinucleotide core observed in all pseudo sites.

46
REFERENCES
1. Smith MC and Thorpe HM. 2002. Diversity in serine recombinases. Mol.
Microbiol. 44:299-307.
2. Stark WM, Boocock MR and Sherratt DJ. 1992. Catalysis by site-specific
recombinases. Trends Genet. 8:432-9.
3. Sauer B. 1994. Site-specific recombination: developments and applications. Curr.
Opin. Biotechnol. 5:521-7.
4. O'Gorman S, Fox DT and Wahl GM. 1991. Recombinase-mediated gene
activation and site-specific integration in mammalian cells. Science 251:1351-5.
5. Sorrell DA and Kolb AF. 2005. Targeted modification of mammalian genomes.
Biotechnol. Adv. 23:431-69.
6. Landy A. 1989. Dynamic, structural, and regulatory aspects of lambda site-specific
recombination. Annu. Rev. Biochem. 58:913-49.
7. Thorpe HM and Smith MC. 1998. In vitro site-specific integration of
bacteriophage DNA catalyzed by a recombinase of the resolvase/invertase family.
Proc. Natl. Acad. Sci. USA 95:5505-10.
8. Matsuura M, Noguchi T, Yamaguchi D, Aida T, Asayama M, Takahashi H and
Shirai M. 1996. The sre gene (ORF469) encodes a site-specific recombinase
responsible for integration of the R4 phage genome. J. Bacteriol. 178:3374-6.

47
9. Christiansen B, Brondsted L, Vogensen FK and Hammer K. 1996. A resolvase-
like protein is required for the site-specific integration of the temperate lactococcal
bacteriophage TP901-1. J. Bacteriol. 178:5164-73.
10. Kuhstoss S and Rao RN. 1991. Analysis of the integration function of the
Streptomycete bacteriophage phiC31. J. Mol. Biol. 222:897-908.
11. Rausch H and Lehmann M. 1991. Structural analysis of the actinophage phiC31
attachment site. Nucleic Acids Res. 19:5187-9.
12. Groth AC, Olivares EC, Thyagarajan B and Calos MP. 2000. A phage integrase
directs efficient site-specific integration in human cells. Proc. Natl. Acad. Sci. USA
97:5995-6000.
13. Thyagarajan B, Olivares EC, Hollis RP, Ginsburg DS and Calos MP. 2001.
Site-specific genomic integration in mammalian cells mediated by phage phiC31
integrase. Mol. Cell Biol. 21:3926-34.
14. Thyagarajan B and Calos MP. 2005. Site-specific integration for high-level
protein production in mammalian cells. Methods Mol. Biol. 308:99-106.
15. Olivares EC, HollisRP, Chalberg TW, Meuse L, Kay MA and Calos MP. 2002.
Site-specific genomic integration produces therapeutic Factor IX levels in mice.
Nat. Biotechnol. 20:1124-8.
16. Held PK, Olivares EC, Aguilar CP, Finegold M, Calos MP and Grompe M.
2005. In vivo correction of murine hereditary tyrosinemia type I by phiC31
integrase-mediated gene delivery. Mol. Ther. 11:399-408.

48
17. Ortiz-Urda S, Thyagarajan B, Keene DR, Lin Q, Fang M, Calos MP and
Khavari PA. 2002. Stable nonviral genetic correction of inherited human skin
disease. Nat. Med. 8:1166-70.
18. Ortiz-Urda S, Thyagarajan B, Keene DR, Lin Q, Calos MP and Khavari PA.
2003. PhiC31 integrase-mediated nonviral genetic correction of junctional
epidermolysis bullosa. Hum. Gene Ther. 14:923-8.
19. Quenneville SP, Chapdelaine P, Rousseau J, Beaulieu J, Caron NJ, Skuk D,
Mills P, Olivares EC, Calos MP and Tremblay JP. 2004. Nucleofection of
muscle-derived stem cells and myoblasts with phiC31 integrase: stable expression
of a full-length-dystrophin fusion gene by human myoblasts. Mol. Ther. 10:679-87.
20. Bertoni C, Jarrahian S, Wheeler TM, Li Y, Olivares EC, Calos MP and Rando
TA. 2006. Enhancement of plasmid–mediated gene therapy for muscular dystrophy
by directed plasmid integration. Proc. Natl. Acad. Sci. USA 103:419-24.
21. Chalberg TW, Genise HL, Vollrath D and Calos MP. 2005. phiC31 integrase
confers genomic integration and long-term transgene expression in rat retina. Invest.
Ophthalmol. Vis. Sci. 46:2140-6.
22. Groth AC, Fish M, Nusse R and Calos MP. 2004. Construction of transgenic
Drosophila by using the site-specific integrase from phage phiC31. Genetics
166:1775-82.
23. Allen BG and Weeks DL. 2005. Transgenic Xenopus laevis embryos can be
generated using phiC31 integrase. Nat. Methods 2:975-9.

49
24. Hollis RP, Stoll SM, Sclimenti CR, Lin J, Chen-Tsai Y and Calos MP. 2003.
Phage integrases for the construction and manipulation of transgenic mammals.
Reprod. Biol. Endocrinol. 1:79.
25. Belteki, G, Gertsenstein M, Ow DW and Nagy A. 2003. Site-specific cassette
exchange and germline transmission with mouse ES cells expressing phiC31
integrase. Nat. Biotechnol. 21:321-4.
26. Olivares EC, Hollis RP and Calos MP. 2001. Phage R4 integrase mediates site-
specific integration in human cells. Gene 278:167-76.
27. Stoll SM, Ginsburg DS and Calos MP. 2002. Phage TP901-1 site-specific
integrase functions in human cells. J. Bacteriol. 184:3657-63.
28. Loessner MJ, Inman RB, Lauer P and Calendar R. 2000. Complete nucleotide
sequence, molecular analysis and genome structure of bacteriophage A118 of
Listeria monocytogenes: implications for phage evolution. Mol. Microbiol. 35:324-
40.
29. Christiansen B, Johnsen MG, Stenby A, Vogensen FK and Hammer K. 1994.
Characterization of the lactococcal temperate phage TP901-1 and its site-specific
integration. J. Bacteriol. 176:1069-76.
30. Hodgson DA. 2000. Generalized transduction of serotype 1/2 and serotype 4b
strains of Listeria monocytogenes. Mol. Microbiol. 35:312-23.
31. Lauer P, Chow MY, Loessner MJ, Portnoy DA and Calendar R. 2002.
Construction, characterization, and use of two Listeria monocytogenes site-specific
phage integration vectors. J. Bacteriol. 184:4177-86.

50
32. Barletta RG, Kim DD, Snapper SB, Bloom BR and Jacobs WR Jr. 1992.
Identification of expression signals of the mycobacteriophages Bxb1, L1 and TM4
using the Escherichia-Mycobacterium shuttle plasmids pYUB75 and pYUB76
designed to create translational fusions to the lacZ gene. J. Gen. Microbiol. 138:23-
30.
33. Mediavilla J, Jain S, Kriakov J, Ford ME, Duda RL, Jacobs WR Jr, Hendrix
RW and Hatfull GF. 2000. Genome organization and characterization of
mycobacteriophage Bxb1. Mol. Microbiol. 38:955-70.
34. Kim AI, Ghosh P, Aaron MA, Bibb LA, Jain S and Hatfull GF. 2003.
Mycobacteriophage Bxb1 integrates into the Mycobacterium smegmatis groEL1
gene. Mol. Microbiol. 50:463-73.
35. Ghosh P, Kim AI and Hatfull GF. 2003. The orientation of mycobacteriophage
Bxb1 integration is solely dependent on the central dinucleotide of attP and attB.
Mol. Cell 12:1101-11.
36. Yang HY, Kim YW and Chang HI. 2002. Construction of an integration-
proficient vector based on the site-specific recombination mechanism of
enterococcal temperate phage phiFC1. J. Bacteriol. 184:1859-64.
37. Bibb, LA and Hatfull GF. 2002. Integration and excision of the Mycobacterium
tuberculosis prophage-like element, phiRV1. Mol. Microbiol. 45:1515-26.

51
38. Cole ST, Brosch R, Parkhill J, Garnier T, Churcher C, Harris D, Gordon SV,
Eiglmeier K, Gas S, Barry CE 3rd, Tekaia F, Badcock K, Basham D, Brown D,
Chillingworth T, Connor R, Davies R, Devlin K, Feltwell T, Gentles S, Hamlin
N, Holroyd S, Hornsby T, Jagels K, Krogh A, McLean J, Moule S, Murphy L,
Oliver K, Osborne J, Quail MA, Rajandream MA, Rogers J, Rutter S, Seeger
K, Skelton J, Squares R, Squares S, Sulston JE, Taylor K, Whitehead S and
Barrell BG. 1998. Deciphering the biology of Mycobacterium tuberculosis from the
complete genome sequence. Nature 393:537-44.
39. Hendrix RW, Smith MC, Burns RN, Ford ME and Hatfull GF. 1999.
Evolutionary relationships among diverse bacteriophages and prophages: all the
world's a phage. Proc. Natl. Acad. Sci. USA 96:2192-7.
40. Hirt B. 1967. Selective extraction of polyoma DNA from infected mouse cultures.
J. Mol. Biol. 26:365-9.
41. Bailey TL and Elkan C. 1994. Fitting a mixture model by expectation
maximization to discover motifs in biopolymers. Proc. Int. Conf. Intell. Syst. Mol.
Biol. 2:28-36.
42. Thomason LC, Calendar R and Ow DW. 2001. Gene insertion and replacement
in Schizosaccharomyces pombe mediated by the Streptomyces bacteriophage
phiC31 site-specific recombination system. Mol. Genet. Genomics 265:1031-8.
43. Chalberg TW, Portlock JL, Olivares EC, Thyagarajan B, Kirby PJ, Hillman
RT, Hoelters J and Calos MP. 2006. Integration specificity of phage phiC31
integrase in the human genome. J. Mol. Biol. 357:28–48.

52
44. Sclimenti CR, Thyagarajan B and Calos MP. 2001. Directed evolution of a
recombinase for improved genomic integration at a native human sequence. Nucleic
Acids Res. 29:5044-51.