Targeted isolation of cloned genomic regions
by recombineering for haplotype phasing and
Marta Nedelkova1, Marcello Maresca1, Jun Fu1, Maria Rostovskaya2, Ramu Chenna3,
Christian Thiede4, Konstantinos Anastassiadis2, Mihail Sarov5and A. Francis Stewart1,*
1Technische Universitaet Dresden, Genomics, BioInnovationsZentrum,2Center for Regenerative Therapies,
Technische Universitaet Dresden,3Applied Bioinformatics Group, BioInnovations Zentrum, Am Tatzberg 47,
4Medizinische Klinik und Poliklinik I, Universitaetsklinikum Carl Gustav Carus der Technischen Universitaet,
Fetscherstrasse 74 and5Max Planck Institute for Cell Biology and Genetics, Pfothenauerstrasse 108,
D-01307 Dresden, Germany
Received February 11, 2011; Revised July 21, 2011; Accepted July 29, 2011
Studying genetic variations in the human genome is
complex traits, including rare personal variations
and their associations with disease. The interpret-
ation of polymorphisms requires reliable methods to
isolate natural genetic variations, including combin-
ations of variations, in a format suitable for down-
stream analysis. Here, we describe a strategy
for targeted isolation of large regions (?35kb)
from human genomes that is also applicable
to any genome of interest. The method relies on
recombineering to fish out target fosmid clones
from pools and thereby circumvents the laborious
need to plate and screen thousands of individual
clones. To optimize the method, a new highly
recombineering-efficient bacterial host, including
inducible TrfA for fosmid copy number amplifica-
tion, was developed. Various regions were isolated
from human embryonic stem cell lines and a
personal genome, including highly repetitive and
duplicated ones. The maternal and paternal alleles
at the MECP2/IRAK 1 loci were distinguished based
on identification of novel allele-specific single-
nucleotide polymorphisms in regulatory regions.
Additionally, we applied further recombineering to
construct isogenic targeting vectors for patient-
specific applications. These methods will facilitate
work to understand the linkage between personal
variations and disease propensity, as well as
possibilities for personal genome surgery.
Recent progress in single-nucleotide polymorphism (SNP)
mapping, genome-wide association studies and mas-
sively parallel sequencing is revealing the diversity of
genetic variation within the human genome (1–5). They
encompass SNPs, insertions, deletions, inversions and du-
plications, which can be linked with disease (1,6).
Understanding the genetic architecture of complex traits
requires knowledge about the polymorphisms in different
parts from the genome, including non-coding regions (6,7)
as well as information about the haplotype phasing, that is
the combination of polymorphisms at the maternal and
paternal alleles (8). SNPs in intergenic and intronic
elements like enhancers have been shown to regulate
gene expression (9,10) and to contribute to human dis-
orders (7,11). Recently, it was demonstrated that the
activity of long interspersed elements contributes to inter
individual genetic variations and can be associated with
disease phenotypes (12,13).
Various methods exist for genome-wide identification of
SNPs and structural variations (1). Recent advances in
high-throughput DNA sequencing technologies have
enabled rapid progress in the field (14) and in the near
future their detection in personal genomes will be per-
formed routinely (15,16). However, the variations lying
in duplicated and highly identical sequences are still diffi-
cult to resolve and extensive bioinformatic analysis is
needed to map the short next-generation sequencing
reads in such regions (17,18).
Although the detection of structural variations is
very important, base pair resolution of their breakpoints
and further functional analysis is usually required
to define their potential impact (19,20). The existing
*To whom correspondence should be addressed. Tel: +49 351 463 40129; Fax: +49 351 463 40143; Email: firstname.lastname@example.org
Marcello Maresca, Novartis Institutes for BioMedical Research, Inc., 250 Massachusetts Avenue, Cambridge, MA 02139, USA.
Published online 18 August 2011Nucleic Acids Research, 2011, Vol. 39, No. 20 e137
? The Author(s) 2011. Published by Oxford University Press.
This is an Open Access article distributed under the terms of the Creative Commons Attribution Non-Commercial License (http://creativecommons.org/licenses/
by-nc/3.0), which permits unrestricted non-commercial use, distribution, and reproduction in any medium, provided the original work is properly cited.
target-enrichment strategies, based on polymerase chain
reaction (PCR) (21), hybridization or molecular inversion
probes (15) merely detect variations, without isolation of
the intact allele as a clone that can be further analyzed to
link polymorphisms over large regions or to be genetically
manipulated for downstream functional analysis. Allele
linkages can be achieved using whole genome bacterial
artificial chromosome (BAC) or fosmid DNA clone
libraries (12,22) but the costs and time required to
generate and map them are often not justified when only
a specific region of the genome needs to be investigated.
In this study, we present a simple approach, based on
recombineering (23,24) for targeted isolation of genomic
regions in a vector format, suitable for downstream
analysis. Recombineering is a DNA engineering technol-
ogy, based on homologous recombination in Escherichia
coli, mediated by the ? phage proteins Reda/Redb or
their functional counterparts RecE/RecT from the Rac
prophage (23,25). We and others have shown that
subcloning by gap repair (25), point mutagenesis in
BACs (24), oligonucleotide directed mutagenesis (26),
BAC engineering for gene targeting (27,28) or protein
tagging (29–31). The high efficiency and fidelity of
recombineering permits high-throughput DNA engineer-
ing at genome scale (30,31).
Here, we demonstrate an application of recombineering
for selective isolation of large genomic fragments of choice
from complex genomes. It circumvents the need for the
classical method of library screening using hybridization
to filters or individually picking and end-sequencing tens
of thousands of clones for indexing. The method is applic-
able to duplicated and repetitive regions and allows for
breakpoint resolution of structural variations at single nu-
cleotide level. The approach further allows the generation
of isogenic targeting constructs with homology arms
carrying the combination of SNPs characteristic for the
source genome. Such constructs will facilitate genome en-
gineering in embryonic stem cells (ESCs) and induced
pluripotent stem cells (iPSCs) for disease studies. We dem-
onstrate the utility of the approach through isolation of
several loci from H7 and Shef4 hES cell lines and from a
cancerous genome and their subsequent haplotype vari-
MATERIALS AND METHODS
Escherichia coli strains
All the strains used in this study are derived from E. coli
DH10B. The strains GB05, GB05Red and DY380 as well
as the low copy, temperature-sensitive pSC101gbaA
plasmid were described previously (32–34). The pSC101b
plasmid is derivative of pSC101gbaA plasmid and encodes
the Redb protein instead of the RedgbaRecA operon. The
E. coli strain GB05RedTrfA was constructed by insertion
of the double operon PBADTrfA-PRharedgbarecA at the
ybcC locus of GB05 (33). For development of the
cassette the PRhapromoter was amplified from pRedFlp
(30). The PBAD promoter from the PBADredgbarecA
operon was replaced with PRha by recombineering.
The PBADTrfA was amplified from the genome of E. coli
EPI300 (Epicentre Biotechnologies, Madison, WI, USA)
and added by recombineering to the PRharedgbarecA.
For the stability test a minimal BAC clone containing two
558bp direct repeats was constructed from pBeloBAC11
vector [New England Biolabs (NEB), Boston, MA, USA].
The repeats are part of the chloramphenicol resistance
gene (cat), which is split into two and is not functional.
The minimal BAC clone contains also neomycin/kanamy-
cin (neo) and zeocin (zeo) genes conferring antibiotic re-
sistance. For the stability assay the strains were grown
overnight at 30?C in LB supplemented with kanamycin
(km) 10mg/ml. From the overnight culture, 106cells
were inoculated in 1ml LB containing zeo 25mg/ml and
grown ON at 30 or 37?C. To estimate the number of spon-
taneous recombinants, the
DNA isolation and shearing
The H7 hES DNA was prepared from cells grown in our
laboratory under standard conditions. The primary bone
marrow sample PS-37027 is from an acute myeloid
leukemia (AML) patient. DNA was isolated applying
cell lysis treatment followed by phenol–chloroform extrac-
tion, isopropanol precipitation and ethanol washing. The
Shef4 hES DNA was kindly provided by Andrew Smith.
The DNA was sheared using the HydroShear device
(Digilab Genomic Solutions, MA, USA) and shearing
assembly 4–40kb (Zinsser Analytic, Frankfurt/Main,
Germany) following the protocol for preparation of
fosmid libraries (35). The sheared DNA was end-repaired
and ethanol precipitated according to the metagenomic
DNA isolation protocol (Epicentre Biotechnologies).
Fosmid library construction and DNA isolation from pools
copy control library kit following the manufacturing
protocol (Epicentre Biotechnologies). The host used
GB05RedTrfA+pSC101b. For library ligations between
0.4 and 1.8mg end-repaired and precipitated DNA was
used (Supplementary Table S1). The titer of the library
was determined and on average 3500 clones were plated
per 15-cm culture dish containing LB agar+cm (10mg/ml)
and tetracycline (tet, 5mg/ml). Plates were incubated at
30?C for 18–24h. To generate the pools, colonies from
each dish were washed off with 2ml LB+cm+tet,
glycerol was added to 20% and 100ml aliquots were
stored at ?80?C in 96-well plates. For DNA isolation
from the pools, 25ml aliquots were inoculated in 1ml
LB+cm at 37?C. The fosmids were induced to high
copy overnight with 0.2% L(+)-arabinose and DNA was
isolated using 96-well filter plate A (VWR International,
Darmstadt, Germany). The DNA was combined in pools
from one row or one column of a 96-well plate for the
e137Nucleic Acids Research, 2011,Vol.39, No. 20PAGE 2 OF 11
PCR pre-screen of the library
The PCR primers for pre-screening of the library
(Supplementary Table S2) were designed using the
oligos were chosen to be in close proximity to the site of
cassette insertion. Their sensitivity was tested with
Ensembl BlastN search tool with search-sensitivity of
near-exact matches and in silico PCR (http://genome
50–100ng DNA from each plate row or plate column
was used. PCR amplification was performed using
Eppendorf Mastercycler CP 534X. Thermal cycling par-
ameters for Taq DNA polymerase (5 prime, Hamburg,
Germany) were 95?C for 4min followed by 35 cycles of
95?C for 15s, annealing for 15s (temperature indicated in
Supplementary Table S2) and extension at 68?C for 15s
with a final extension of 10min at 68?C. All the oligos in
this study were purchased from Biomers (http://www
Recombinogenic cassette design and modification
(Supplementary Table S3) were designed according to
version CRCh37 release 54–58. The cassettes were
generated by PCR using the blasticidin resistance gene
(bsd) and oligonucleotides that contain the flanking
50bp homology regions. The bsd selectable marker was
amplified from the genomic ara-leu locus of strain GB05
(previously recombined with this cassette) to prevent
phosphorylated at one 50-end but not to the other 50-end
to generate PO or OP cassettes, where O means hydroxyl
(36). The cassettes were purified from the PCR reaction
Germany). The cassette for testing the recombineering
efficiency of the E. coli strains was also phosphorylated
at one of the 50-end. In addition two phosphorothioate
linkages (S) were inserted in the first and second bond at
the other 50-end (PS cassette) (36).
arm (HA)for the capturingcassettes
The cassettes were
To screen the library by recombineering, aliquots (25ml)
from the PCR positive pools were grown in 1ml LB sup-
plemented with tet (5mg/ml) and cm (10mg/ml) overnight
at 30?C. The overnight culture was diluted 1/50 and grown
in 25ml at 30?C for 2h, followed by addition of L(+)-
arabinose (Sigma A-3256) and L(+)-rhamnose (Sigma
R3875) to 0.2% and growth for 45min at 37?C. The
cells were centrifuged, transferred to an Eppendorf tube
and washed twice with 1ml of ice-cold 10% glycerol,
followed by resuspension in 80ml. About 600ng cassette
was added to 40ml competent cells. For each electropor-
ation, a pre-chilled 1mm electroporation cuvette (BTX,
Harvard apparatus) was used at settings 1350V, 10mF,
600V (Eppendorf Electroporator 2510). After electropor-
ation the cells were resuspended in 1ml SOC medium and
incubated for 1h at 37?C before plating on low-salt LB
agar supplemented with 40mg/ml blasticidin S (BSD)
(InvivoGen, San Diego, CA, USA). The plates were
incubated at 37?C for 18–24h.
Characterization of the isolated recombinant fosmids
Between 1 and 16 clones per captured region were
inoculated in 1ml low salt LB supplemented with BSD
40mg/ml and grown overnight at 37?C then 30ml were
inoculated in 0.5ml TB supplemented with BSD and
grown overnight at 37?C. To the rest, glycerol was
added to 20% and stored at ?80?C. Fosmid DNA was
isolated by using Invisorb spin plasmid mini two (Invitek,
Berlin, Germany) or 96-well filter plate A (VWR
International). The clones were end-sequenced with
pCC2Fos vector primers. Around 0.7mg DNA was used
for the restriction digestion experiments in a 40-ml reaction
volume. All enzymes were supplied by NEB.
Next-generation sequencing parameters and
Fosmid DNA was mixed in five pools at final concentra-
tion of ?3.5mg/6ml so that overlapping clones were kept
in different pools. The DNA was sheared using the
Covaris S2 (Covaris, Inc. Massachusetts, MA, USA) to
an average fragment size of 200bp. The fragmented pools
of DNA were indexedand
sequencing library for Illumina platform was prepared
(NEB, NEBNext?DNA Sample Preparation). After
flow cell generation on the cBOT (Illumina) standard
single read sequencing (51 bases) was performed on the
HiSeq 2000 platform (Illumina). A total of 1.2?108reads
were obtained from which 75% were mappable. Mapping
was done with Bowtie (version 0.12.7 64-bit) against
UCSC_GRCh37/hg19 human genome assembly. Initial
SNP calling was carried out with samtools and subse-
quently custom software was written and used for the
SNP analysis. The latest snp132 database was used to
annotate the variations and bambino and IGV 1.5
(Broad Insititute) software was used to identify the
genomic regions for polymorphisms.
Isogenic targeting constructs generation
Gene Construction Kit (TEXTCO BioSoftware). The
recombineering experiments were performed in the library
host GB05RedTrfA, which had lost the temperature-
sensitive pSC101b plasmid by culture at 37?C. The
recombineering protocol was the same as described for
screening the libraries but in the subsequent steps the in-
duction was only with L(+)-rhamnose. The capturing cas-
settes contain 40bp sequences flanking the bsd that serve
as homology arms for sequential recombineering with the
reporter cassette lacZneo (sA-T2A-LacZ-T2A-Neo-pA-
loxP). The rest of the cassettes for generation of condi-
tional knockout targeting construct were designed as
already published (33). The oligos for attachment of
homology arms by PCR to the capturing cassette, the
sub cloning vector p15A-pTK-DTA-ampR and the down-
stream cassette rox-BSD-PGK-rox-loxP are given in
Supplementary Table S4.
constructs werein silico designed using
PAGE 3 OF 11 Nucleic Acids Research, 2011,Vol.39, No. 20 e137
Generation of recombineering proficient host for fosmid
Our goal was to develop an assay that can capture by
recombineering large regions of interest from human
genomes in a fosmid clone format suitable for sequencing
and genetic engineering. We generated a new fosmid
library host (GB05RedTrfA) (Figure 1), which carries in
its genome the gbaRecA recombineering operon (32)
under the rhamnose inducible promoter (PRHA) (37) as
well as the TrfaA protein (38) under the arabinose indu-
cible promoter (PBAD) (39). The TrfA protein is required
for initiation of the replication from the bidirectional
origin OriV and subsequent increase in the fosmid copy
number. The strain is highly stable (Supplementary
Figure S1) with rates of spontaneous rearrangements
in the absence of induction comparable with the previ-
GB05(BAD)Red (33) or DY380 (34). We optimized the
recombineering conditions using a blasticidin resistance
cassette insertion assay into a single fosmid clone
(Figure 1A). One of the strands of the dsDNA cassette
was phosphorylated at the 50-end and phosphothioate
linkages were added to the 50-end of the other strand, to
facilitate the enzymatic conversion to ssDNA in vivo,
which improves the recombineering frequencies (36).
We tested if the recombineering efficiencies can be
further promoted by the helper plasmids pSC101b or
pSC101gbaA (32), in which the recombineering genes
are also under PBADcontrol. The additional transient ex-
pression of the strand annealing protein Redb alone from
the helper plasmid pSC101b increased the frequency of
recombination almost twice as much as the additional
complete recombineering operon from pSC101gbaA
(Figure 1B), indicating that overexpression of some of
the other proteins in the operon may be detrimental to
the overall efficiency.
More than 3-fold increase in the number of recombin-
ants was observed after high copy fosmid induction in
GB05RedTrfA in comparison with the GB05Red strain
where oriV cannot be induced (Figure 1B). Using the
Figure 1. Fosmid library host optimization. (A) Recombineering assay with GB05RedTrfA+pSC101b. The strain carries in the genome the modified
red operon (32) (gam, beta, exo, recA) (red) under the control of the rhamnose inducible Rha promoter and has the TrfA gene (blue), which
promotes high copy fosmid replication under the control of the arabinose inducible BAD promoter. The BAD promoter also drives expression of the
Redb protein (red) located on the helper plasmid pSC101b. A random fosmid clone was chosen for the insertion of modified blasticidin (bsd) cassette
via recombineering. The cassette is flanked with 50bp homology arms, identical to the region of choice (green). After the selection step the
temperature-sensitive plasmid pSC101b is lost at 37?C. OriV—bidirectional origin of replication, Fori—unidirectional origin of replication. (B)
Strain history. All strains were derived from DH10B. The strains GB05, GB05Red and DY380 as well as the pSC101gbaA plasmid were described
previously (32–34). The BADTrfA and RhagbaA cassettes in GB05RedTrfA were inserted at ybcC locus as the BADgbaA cassette in GB05Red. (C)
Comparison of the recombineering efficiencies of the host strains. To test recombineering efficiency, a modified blasticidin cassette was inserted in a
randomly selected fosmid from the H7 library. The number of recombinants was normalized to the number of cells surviving electroporation. Higher
recombineering frequencies were obtained using the pSC101b helper plasmid. High copy fosmid induction further promoted recombineering
e137 Nucleic Acids Research, 2011,Vol.39, No. 20PAGE 4 OF 11
GB05RedTrfA+pSC101b and transient high copy fosmid
replication induction, we achieved up to 6.8?103recom-
binants per million viable cells after transformation, an
efficiency which allows for recombineering mediated tar-
geting of a specific clone in a complex fosmid library.
Targeted isolation of genomic regions by recombineering
The general outline of our approach is shown in the
flowchart of Figure 2. First, a fosmid library is con-
structed from mechanically
(Figure 2A). Next, the library is split into pools of about
3500 clones, which are then screened by PCR. Finally, the
target clones are fished out by recombineering through the
insertion of a modified blasticidin cassette flanked by
50-bp long homology arms (Figure 2B).
We optimized the method using genomic DNA isolated
from H7 human embryonic stem (hES) cell line (40).
Based on the recombineering efficiencies determined with
single fosmids (6.8?103recombinants/106cells) and given
thatthe numberof surviving
recombineering reaction in the absence of selection is
about 109cells/ml, we estimated that the recombineering
efficiency of the new host should allow us to isolate 10–100
recombinants of a specific clone in a mixture of 104clones.
In a pilot experiment, a defined fosmid was added to pools
of different complexities to determine that the optimal
performance was achieved with pools of 3.5?103
fosmids (data not shown). At that complexity, a library
of over 3-fold coverage of the haploid human genome can
fit in a single 96-well plate, and any region of interest can
be isolated within 2 days, saving time and effort involved
in screening entire libraries.
cells ina typical
Application of the method to retrieve various regions
We applied the approach to capture the OCT4 locus from
the H7 hES cell line. After recombineering, blasticidin-
resistant colonies were obtained from five PCR positive
pools from two independent libraries (Supplementary
Table S5). End sequencing from the vector and restriction
analysis established that the captured fosmids covered the
OCT4 locus and surrounding regions (Figure 3A;
Supplementary Figure S2 and Supplementary Table S5).
Five further regions were retrieved from the H7 hES
cells. For the adenosine kinase (AK), methyl CpG
binding protein2 gene (MECP2) and paired box 6
(PAX6) transcriptional factor we isolated the genomic
regions, required for isogenic targeting construct gener-
ation (Figure 3B–D). The entire MYCN and NANOG
genes and their surrounding regions were also successfully
captured (Figure 3E and F). NANOG has several pseudo-
genes and one of them, NANOG P1, arose through local
duplication of the NANOG gene (41). In order to isolate
the gene, 100bp of homology sequence unique to the
NANOG locus was chosen. The captured fosmid covers
the whole locus, an intergenic region and part of the
neighboring gene, which is also duplicated (41). Large
parts of the 36kb genomic fragment contain repeats
from which 66% belong to different classes of Alu
elements. Restriction analysis confirmed that the highly
repetitive fosmids were not rearranged (Supplementary
Figure S2 and Supplementary Table S5).
In further exercises, we used the male hES cell line Shef4
(42) and a primary leukemic sample. With the available
cassettes, we isolated Shef4 MECP2, OCT4, PAX6 and
GATA4 regions (Supplementary Table S5). For the
leukemic sample, we focused on potential disease-related
regions of chromosome 2 and isolated two independent
clones for each of the regions of interest (TP53I3,
ASXL2 and MYCNOS loci).
All target regions from both hES cells lines and the per-
sonal genome were captured successfully (Supplementary
Table S5). As with other recombineering applications, we
have not found any sequence limitation in the choice of
homology arms except for the need to avoid repeats.
Figure 2. Recombineering strategy for fishing out genomic regions. (A) Fosmid library preparation. High molecular weight DNA was isolated from
hES cell line or patient tissue sample. The DNA was sheared to ?40kb fragments. The fragmented DNA was ligated to pCC2Fos (dark grey lines) as
concatamers and packaged in ? phage particles. The DNA was transduced into the E. coli host strain GB05RedTrfA+pSC101b. (B) Screening of the
library by recombineering. On average 3500 cfu were plated per petri dish and then collected as a pool to a single well of a 96-well plate, which were
pre-screened in super pools of rows and columns by PCR. Positive wells were cultured to induce the fosmids to high copy and express the Red
proteins for recombineering before electroporation with a bsd. Recombination into the fosmid of choice conveyed blasticidin resistance after plating.
PAGE 5 OF 11 Nucleic Acids Research, 2011,Vol.39, No. 20 e137
Hence the approach appears to be applicable to a diverse
spectrum of genomic regions. No incorrect insertions were
observed and the restriction digest analysis showed a very
low number of rearranged clones. The number of recom-
binants varied for each of the targeted regions but was
within the expected range (1–728 recombinants per
reaction). Addition of more than 500ng of the cassette
did notincrease the
(Supplementary Figure S3).
We used single-strand DNA recombineering as it
provides higher efficiency and fidelity (36). Either strand
can be used, but the strand annealing to the lagging strand
of the replication fork is favored by the recombineering
reaction (43). In our experiments, the efficiencies between
the two strands varied several fold (Supplementary
Table S6), indicating that testing both strands can be
beneficial for the isolation of difficult regions.
Haplotype phasing and identification of allelic differences
in H7 loci
Regions from the H7 cell line for which more that one
fosmid was fished out (Figure 3) were sequenced with
Illumina in order to reconstruct the haplotype phase of
the genomic regions. Indexed libraries, containing the
overlapping clones were sequenced to a mean depth of
11071 reads per base pair. Bioinformatic analyses
indicated two positions on chromosome X with potential
allelic differences that were supported with similar number
of unique readsbetween theoverlappingclones
Figure 3. Genomic regions of interest isolated from the H7 hES cell line. Depicted are the fished out clones (red), the capturing cassettes (black
triangles) and the characteristics of the genomic region showing exons (thick yellow lines), introns (thin yellow lines), repetitive elements (all repeats)
and G+C content (%GC). (A) Clones containing the OCT4 locus. (B) The AK locus with the capturing cassette inserted in front of exon 5.
(C) Methyl CpG binding protein 2 locus (MECP2) and two different probes used for fishing out regions of interest. (D) The 30-end of the PAX6
locus. (E) Isolation of the NANOG locus from the highly similar NANOG P1 pseudogene (depicted in lilac). (F) Clone covering the MYCN gene
and the probe used, which has 69% G+C content.
e137Nucleic Acids Research, 2011,Vol.39, No. 20PAGE 6 OF 11
(Supplementary Table S7). These include differences at the
MECP2/IRAK1 loci that are not annotated in SNP132
database. The observed allelic polymorphisms are G/A
at the 30-UTR of MECP2 and C/G at the promoter
region of IRAK1 located 5325-bp downstream on the
dinucleotides and are located in regulatory regions—a
DNase1 hypersensitive site in the 30-UTR of MECP2
and the CpG island upstream of IRAK1 (USCS genome
browser GRSh37/hg19). The SNP at the 30-UTR of
MECP2 was validated by PCR and sequencing (data not
shown). The second SNP is located in an extremely GC
rich region and we failed to amplify it by PCR with several
sets of primers.
In addition to the allele-specific SNPs we reconstructed
the combination of SNPs across the sequenced regions of
chromosome X, 6 and 10 (Supplementary Table S7).
As expected more SNPs were found in the highly poly-
morphic region of chromosome 6 than at the other loci. In
addition several non-synonymous mutations in CCHCR1
and TCF19 genes and small-scale indels were scattered
across the 35kb genomic region from chromosome 6
(data not shown). The indels for the OCT4 loci from the
H7 and Shef4 cell line were validated by PCR and
sequencing (Supplementary Table S8).
Generation of isogenic targeting constructs
We used the retrieved fosmids to generate allele-specific
targeting constructs for MECP2, AK and OCT4 by the
following method. The blasticidin cassette used for fishing
from the pools was designed to contain additional 40bp
(Figure 5A). After isolation of the isogenic clones, the
blasticidin cassettes were replaced by recombineering
with a lacZneo stop cassette that is flanked by the same
40bp homology arms (Figure 5B). For MECP2, the
blasticidin cassette was targeted to the intron upstream
of exon 4, which was selected because its later removal
by Cre recombinase will cause a frame shift in the
addition of a 30loxP site after the frame-shifting exon
were done following the established pipeline for condition-
al targeting constructs generation (Figure 5C and D) (33).
All recombineering steps after clone isolation were
mediated by the rhamnose inducible redgbaRecA operon
present in the genome of the GB05RedTrfA. The expected
products were validated by restriction mapping and
sequencing of the recombineering junctions. They have
been successfully used for targeting in H7 hES cells
(data not shown).
Studying genetic variations in the human genome is im-
portant for the understanding of phenotypes, diseases,
drug responsiveness and the mechanisms of complex
traits (6). For many applications, only a small part of
the genome, such as specific genes or regulatory regions,
are of interest (44,45). The current methods for selected
enrichment of genomic regions followed by next gener-
ation sequencing are based on PCR or hybridization
approaches (15). These methods encounter size limitations
particularly to link variations separated by more than a
few hundred base pairs, as well as limitations in duplicated
and repetitive regions.
The recombineering strategy presented here is useful for
targeted isolation of genomic regions in a vector format
that allows for rapid adaptation to functional analysis
based on gene targeting (27,28) or transgenesis (30). A
similar approach to isolate genomic regions in BACs has
been published recently (46). We use fosmids, because they
are easy to handle, stable, suitable for genomic structural
variation studies (2,5,22) and preparation of targeting
constructs. Most importantly, compared to BAC libraries,
fosmid library construction requires much less genomic
DNA, which is a major consideration when the source
of DNA is a patient sample.
Figure 4. Allele-specific SNPs at the MECP2/IRAK1 locus on the X chromosome in H7 hES cells. In the upper part of the diagram the distribution
of uniquely mappable Illumina reads from fosmid clones H7-F and H7-C02 is shown as grey lines. Gaps indicate repetitive sequence and the average
corresponds to 11000 reads. The colored lines indicate the SNP positions in comparison with the human reference genome as follows: green-A;
red-T; blue-C and orange-G. The lower part of the diagram shows a 14kb region containing two allele-specific SNPs among seven other SNPs and
one newly identified indel, which is present on both alleles. One of the alleles has A instead of G at position 153290956 and G instead of C at
position 153285631. The SNPs are in DNaseI hypersensitive site (DHS) and in CpG island, respectively as indicated. The indel indicates an insertion
of a G onto a poly G track at 153287325.
PAGE 7 OF 11Nucleic Acids Research, 2011,Vol.39, No. 20e137
To increase the targeting efficiency and thereby the
complexity of the pools from which a specific region can
be retrieved, we engineered a new strain that allows for
switching from unidirectional to bidirectional fosmid rep-
lication. In that way, we exploit an additional increase in
copy-number after TrfA induction. This improved the iso-
lation of genomic regions of choice from complex fosmid
pools. The very low levels of illegitimate recombination
reduced the need to screen through a large number of
clones to obtain the desired region. The number of recom-
binants varied between the captured loci, possibly reflect-
ing the different replication speeds of the individual clones
within the pools. Variability in the number of recombin-
ants for several E. coli chromosomal locations has previ-
ously been correlated with the rate of replication of the
Previously a method to screen genomic libraries by
recombineering was reported (47). However, this method
does not appear to have been subsequently utilized,
possibly because the complex counter selection strategy
imposed practical difficulties. Similarly our previous ex-
perience with genomic cloning by recombineering (25),
indicated certain practical limits to lambda Red recombin-
ation in complex backgrounds. Hence, we adapted a
recombineering method to optimally sized pools of
cloned genomic regions.
bineering proteins not only improved the recovery of
target clones but also likely contributed to the successful
isolation of intact, highly repetitive, regions. Indeed,
previous work has shown that overexpression of Redg
from a plasmid can increase the total number of
colonies, but the frequency of correct recombinant
BACs was low (48). Transient RecA co-expression from
a plasmid has been previously shown to enhance the total
number of colonies surviving electroporation (32), but
leaky expression of RecA could cause increased basal
That is why we expressed RecA from the genome,
together with the Red operon, using the tightly controlled
Figure 5. Workflow for the generation of isogenic conditional targeting constructs. All recombineering steps after the first one were performed in
GB05RedTrfA after rhamnose induction of the RedgbaA operon. All the genes conveying antibiotic resistance have prokaryotic promoters (data not
shown). The antibiotics used for selection of the recombinants at each stage are indicated at the right of each step: cm, chloramphenicol; bsd,
blasticidin; km, kanamycin; amp, ampicillin. (A) The fosmid clone was identified by insertion of the bsd (lilac) as described above using 50bp
homology arms to the genomic region of interest (orange) and thereby inserting the homology arms for the next recombineering step (blue). The
insertion site was chosen to be nearby a frame-shifting exon (orange). Fori and oriV, fosmid replication origins; cat, chloramphenicol resistance gene.
(B) The bsd gene is replaced by a lacZneo reporter/stop cassette that conveys kanamycin resistance by recombination through the homology arms
(blue). The lacZneo cassette contains our standard design (27,28,33) including frt sites for FLP recombinase and a loxP site for Cre recombinase at
the 30-end. The cassette also contains splice acceptor (sA); ribosomal-skipping peptides (t2A) and the SV40 polyadenylation signal (SVpA). Neo is the
kanamycin/G418 resistance gene and lacZ is the b-galactosidase gene. (C) A region containing the lacZneo cassette, the frame shifting exon and
flanking ?4.5kb homology arms is subcloned by gap repair into a p15A plasmid vector. The plasmid vector also includes the HSVtk promoter (TK)
driven diphtheria toxin A chain gene (DTA) for counter selection and the ampicillin resistance gene (bla). (D) A second loxP site is added on the 30
side of the frameshifting exon using blasticidin selection. The PGK–BSD gene is flanked by rox sites for later removal by Dre recombinase.
e137Nucleic Acids Research, 2011,Vol.39, No. 20PAGE 8 OF 11
The extent of variation within human genomes is now
being revealed by SNP maps and massively parallel
sequencing (1–4). However, knowledge about the ‘haplo-
type phasing’ in different genomes has been scarce (8).
Two recently published methods for genome-wide reso-
lution of the haplotypes (49,50) pave the way to system-
atically study haplotype phasing in individual genomes
and cell lines. Our approach is complementary to these
studies and allows for the determination of SNP linkage
and therefore the disease susceptibility throughout the
selected regions covered by fosmid clones. Thereby, we
reconstructed haplotypes at loci from chromosome 6, X
and 10 from the H7 hES cell line. Comparative analysis
between the H7 and Shef4 OCT4 haplotypes revealed dif-
ferences in 12 SNP positions and most of the identified
indels were cell line specific (13 of 16). These variations
were found in more than one independent clone and there-
fore represent true polymorphisms of the cell lines.
Whole-genome sequencing shows that structural vari-
ations smaller than 50kb account for the large portion
genomes (1,5). Most of these events are enriched near or
in repeated and segmental duplicated regions and
difficulties to resolve them have been reported by different
investigators (5,17). Using the targeted retrieval of clones,
we were able to distinguish between highly similar se-
quences like NANOG and its pseudogene NANOG P1.
Once isolated, such regions can be further characterized
by sequencing at very high depth. This allows the descrip-
Exploring the impact of the mutations and their char-
acterization as benign or disease associated can be
achieved through gene targeting in stem cells (51,52)
with isogenic constructs. Our approach permits generation
of such constructs with personal genome specific combin-
ation of variations. The isogenicity of the flanking hom-
ologous sequences is an important issue. First, it could
promote the targeting efficiency in human ES cells as
was shown for mouse ES cells (48,53). Second, bearing
in mind that SNPs may influence transcription factor
binding and gene expression (9,10), targeting with
isogenic vectors should not disturb the existing genomic
context. This will be useful for gene editing in stem
We identified two novel allele-specific SNPs located in
regulatory regions on one of the X chromosome in the H7
cell line at the MECP2/IRAK1 loci. The biological signifi-
cance of these polymorphisms is not known. The
whole-genome ENCODE analysis on the male H1 hES
cell line indicates that the two SNPs are located in an
enhancer and a promoter where c-Myc and Pol2 bind,
respectively. The SNPs are in CpG dinucleotides thus
they may influence the binding of regulatory proteins or
the methylation status of the two alleles.
demonstrated in this and previous studies allows the
further scale up of the method to high-throughput liquid
format (30,31) for simultaneous isolation of multiple loci.
For example, the method can be used to develop screening
assays for isolation of regions affected by mobilized
at single nucleotide
retrotransposons or other repetitive elements in personal
genomes. Recently, numerous novel active retrotrans-
posons were identified in the human genome (12,13).
Although they are underrepresented in the reference
sequence, they exist at low allele frequencies in the popu-
lation and can be a source for disease-producing
This method can also simplify the acquisition of DNA
regions from model organisms or metagenomic studies of
environmental samples. The approach is straight forward
anddoes notrequire any
complicated computational analysis. Because it is flexible
with many potential applications, we recommend it to a
wide range of researchers.
Supplementary Data are available at NAR Online.
The authors thank Andrew Smith for providing the DNA
from Shef4 hES cell line. The authors are grateful to
Andreas Dahl of the Deep Sequencing Facility at Biotec
Dresden, for Illumina sequencing and the primary data
analysis. M.N. designed and performed the experiments
for all the main figures and most of the Supplementary
Supplementary Fig. 1. M.R. cultured the H7 hES cell
Supplementary Table 7. C.T. provided the leukemic
sample. M.S., K.A., M.M., J.F. and A.F.S. contributed
with ideas and discussions throughout the project. M.N.
and A.F.S. prepared the manuscript.
NGFN plus (Nationales Genomforschungsnetz of the
01GS0872) (to C.T. and A.F.S.); European Commission
6th Framework Program, ESTOOLS and 7th Framework
Program, EUCOMMTOOLS (to A.F.S.). Funding for
open accesscharge: NGFN
Leukemia grant (to C.T. and A.F.S.).
Conflict of interest statement. The primary patents for
recombineering are held by Gene Bridges GmbH, a
company which AFS founded and is a major shareholder.
1. Feuk,L., Carson,A.R. and Scherer,S.W. (2006) Structural
variation in the human genome. Nat. Rev. Genet., 7, 85–97.
2. Kidd,J.M., Cooper,G.M., Donahue,W.F., Hayden,H.S.,
Sampas,N., Graves,T., Hansen,N., Teague,B., Alkan,C.,
Antonacci,F. et al. (2008) Mapping and sequencing of structural
variation from eight human genomes. Nature, 453, 56–64.
3. Lander,E.S. (2011) Initial impact of the sequencing of the human
genome. Nature, 470, 187–197.
4. Mills,R.E., Walter,K., Stewart,C., Handsaker,R.E., Chen,K.,
Alkan,C., Abyzov,A., Yoon,S.C., Ye,K., Cheetham,R.K. et al.
PAGE 9 OF 11Nucleic Acids Research, 2011,Vol.39, No. 20 e137
(2011) Mapping copy number variation by population-scale
genome sequencing. Nature, 470, 59–65.
5. Tuzun,E., Sharp,A.J., Bailey,J.A., Kaul,R., Morrison,V.A.,
Pertz,L.M., Haugen,E., Hayden,H., Albertson,D., Pinkel,D. et al.
(2005) Fine-scale structural variation of the human genome.
Nat. Genet., 37, 727–732.
6. Frazer,K.A., Murray,S.S., Schork,N.J. and Topol,E.J. (2009)
Human genetic variation and its contribution to complex traits.
Nat. Rev. Genet., 10, 241–251.
7. Visel,A., Rubin,E.M. and Pennacchio,L.A. (2009) Genomic
views of distant-acting enhancers. Nature, 461, 199–205.
8. Bansal,V., Tewhey,R., Topol,E.J. and Schork,N.J. (2011) The
next phase in human genetics. Nat. Biotechnol., 29, 38–39.
9. Bandele,O.J., Wang,X., Campbell,M.R., Pittman,G.S. and
Bell,D.A. (2011) Human single-nucleotide polymorphisms alter
p53 sequence-specific binding at gene regulatory elements.
Nucleic Acids Res., 39, 178–189.
10. Kasowski,M., Grubert,F., Heffelfinger,C., Hariharan,M.,
Asabere,A., Waszak,S.M., Habegger,L., Rozowsky,J., Shi,M.,
Urban,A.E. et al. (2010) Variation in transcription factor binding
among humans. Science, 328, 232–235.
11. Cookson,W., Liang,L., Abecasis,G., Moffatt,M. and Lathrop,M.
(2009) Mapping complex disease traits with global gene
expression. Nat. Rev. Genet., 10, 184–194.
12. Beck,C.R., Collier,P., Macfarlane,C., Malig,M., Kidd,J.M.,
Eichler,E.E., Badge,R.M. and Moran,J.V. (2010) LINE-1
retrotransposition activity in human genomes. Cell, 141, 1159–1170.
13. Iskow,R.C., McCabe,M.T., Mills,R.E., Torene,S., Pittard,W.S.,
Neuwald,A.F., Van Meir,E.G., Vertino,P.M. and Devine,S.E.
(2010) Natural mutagenesis of human genomes by endogenous
retrotransposons. Cell, 141, 1253–1261.
14. Pushkarev,D., Neff,N.F. and Quake,S.R. (2009) Single-molecule
sequencing of an individual human genome. Nat. Biotechnol., 27,
15. Mamanova,L., Coffey,A.J., Scott,C.E., Kozarewa,I., Turner,E.H.,
Kumar,A., Howard,E., Shendure,J. and Turner,D.J. (2010)
Target-enrichment strategies for next-generation sequencing.
Nat. Methods, 7, 111–118.
16. Wheeler,D.A., Srinivasan,M., Egholm,M., Shen,Y., Chen,L.,
McGuire,A., He,W., Chen,Y.J., Makhijani,V., Roth,G.T. et al.
(2008) The complete genome of an individual by massively
parallel DNA sequencing. Nature, 452, 872–876.
17. Medvedev,P., Stanciu,M. and Brudno,M. (2009) Computational
methods for discovering structural variation with next-generation
sequencing. Nat. Methods, 6, S13–20.
18. Sudmant,P.H., Kitzman,J.O., Antonacci,F., Alkan,C., Malig,M.,
Tsalenko,A., Sampas,N., Bruhn,L., Shendure,J. and Eichler,E.E.
(2010) Diversity of human copy number variation and multicopy
genes. Science, 330, 641–646.
19. Conrad,D.F., Bird,C., Blackburne,B., Lindsay,S., Mamanova,L.,
Lee,C., Turner,D.J. and Hurles,M.E. (2010) Mutation spectrum
revealed by breakpoint sequencing of human germline CNVs.
Nat. Genet., 42, 385–391.
20. Kidd,J.M., Newman,T.L., Tuzun,E., Kaul,R. and Eichler,E.E.
(2007) Population stratification of a common APOBEC gene
deletion polymorphism. PLoS Genet., 3, e63.
21. Tewhey,R., Warner,J.B., Nakano,M., Libby,B., Medkova,M.,
David,P.H., Kotsopoulos,S.K., Samuels,M.L., Hutchison,J.B.,
Larson,J.W. et al. (2009) Microdroplet-based PCR enrichment
for large-scale targeted sequencing. Nat. Biotechnol., 27,
22. Eichler,E.E., Nickerson,D.A., Altshuler,D., Bowcock,A.M.,
Brooks,L.D., Carter,N.P., Church,D.M., Felsenfeld,A., Guyer,M.,
Lee,C. et al. (2007) Completing the map of human genetic
variation. Nature, 447, 161–165.
23. Zhang,Y., Buchholz,F., Muyrers,J.P. and Stewart,A.F. (1998) A
new logic for DNA engineering using recombination in
Escherichia coli. Nat. Genet., 20, 123–128.
24. Muyrers,J.P., Zhang,Y., Benes,V., Testa,G., Ansorge,W. and
Stewart,A.F. (2000) Point mutation of bacterial artificial
chromosomes by ET recombination. EMBO Rep., 1, 239–243.
25. Zhang,Y., Muyrers,J.P., Testa,G. and Stewart,A.F. (2000) DNA
cloning by homologous recombination in Escherichia coli. Nat.
Biotechnol., 18, 1314–1317.
26. Ellis,H.M., Yu,D., DiTizio,T. and Court,D.L. (2001) High
efficiency mutagenesis, repair, and engineering of chromosomal
DNA using single-stranded oligonucleotides. Proc. Natl Acad. Sci.
USA, 98, 6742–6746.
27. Testa,G., Zhang,Y., Vintersten,K., Benes,V., Pijnappel,W.W.,
Chambers,I., Smith,A.J., Smith,A.G. and Stewart,A.F. (2003)
Engineering the mouse genome with bacterial artificial
chromosomes to create multipurpose alleles. Nat. Biotechnol., 21,
28. Skarnes,W.C., Rosen,B., West,A.P., Koutsourakis,M., Bushell,W.,
Iyer,V., Mujica,A.O., Thomas,M., Harrow,J., Cox,T. et al. (2011)
A conditional knockout resource for the genome-wide study of
mouse gene function. Nature, 474, 337–342.
29. Hofemeister,H., Ciotta,G., Fu,J., Seibert,P., Schulz,A.,
Maresca,M., Sarov,M., Anastassiadis,K. and Stewart,F. (2011)
Recombineering, transfection, Western and ChIP methods for
protein tagging via gene targeting or BAC transgenesis. Methods,
30. Sarov,M., Schneider,S., Pozniakovski,A., Roguev,A., Ernst,S.,
Zhang,Y., Hyman,A.A. and Stewart,A.F. (2006) A
recombineering pipeline for functional genomics applied to
Caenorhabditis elegans. Nat. Methods, 3, 839–844.
31. Poser,I., Sarov,M., Hutchins,J.R., Heriche,J.K., Toyoda,Y.,
Pozniakovsky,A., Weigl,D., Nitzsche,A., Hegemann,B., Bird,A.W.
et al. (2008) BAC TransgeneOmics: a high-throughput method
for exploration of protein function in mammals. Nat. Methods, 5,
32. Wang,J., Sarov,M., Rientjes,J., Fu,J., Hollak,H., Kranz,H.,
Xie,W., Stewart,A.F. and Zhang,Y. (2006) An improved
recombineering approach by adding RecA to lambda Red
recombination. Mol. Biotechnol., 32, 43–53.
33. Fu,J., Teucher,M., Anastassiadis,K., Skarnes,W. and Stewart,A.F.
(2010) A recombineering pipeline to make conditional targeting
constructs. Methods Enzymol., 477, 125–144.
34. Lee,E.C., Yu,D., Martinez de Velasco,J., Tessarollo,L.,
Swing,D.A., Court,D.L., Jenkins,N.A. and Copeland,N.G. (2001)
A highly efficient Escherichia coli-based chromosome engineering
system adapted for recombinogenic targeting and subcloning of
BAC DNA. Genomics, 73, 56–65.
35. Donahue,W.F. and Ebling,H.M. (2007) Fosmid libraries for
genomic structural variation detection. Curr Protoc Hum Genet,
Chapter 5, Unit 5.20.
36. Maresca,M., Erler,A., Fu,J., Friedrich,A., Zhang,Y. and
Stewart,A.F. (2010) Single-stranded heteroduplex intermediates in
lambda Red homologous recombination. BMC Mol. Biol., 11, 54.
37. Cardona,S.T. and Valvano,M.A. (2005) An expression vector
containing a rhamnose-inducible promoter provides tightly
regulated gene expression in Burkholderia cenocepacia. Plasmid,
38. Wild,J., Hradecna,Z. and Szybalski,W. (2002) Conditionally
amplifiable BACs: switching from single-copy to high-copy
vectors and genomic clones. Genome Res., 12, 1434–1444.
39. Guzman,L.M., Belin,D., Carson,M.J. and Beckwith,J. (1995)
Tight regulation, modulation, and high-level expression by
vectors containing the arabinose PBAD promoter. J. Bacteriol.,
40. Thomson,J.A., Itskovitz-Eldor,J., Shapiro,S.S., Waknitz,M.A.,
Swiergiel,J.J., Marshall,V.S. and Jones,J.M. (1998) Embryonic
stem cell lines derived from human blastocysts. Science, 282,
41. Booth,H.A. and Holland,P.W. (2004) Eleven daughters of
NANOG. Genomics, 84, 229–238.
42. Aflatoonian,B., Ruban,L., Shamsuddin,S., Baker,D., Andrews,P.
and Moore,H. (2010) Generation of Sheffield (Shef) human
embryonic stem cell lines using a microdrop culture system.
In Vitro Cell. Dev. Biol. Anim., 46, 236–241.
43. Zhang,Y., Muyrers,J.P., Rientjes,J. and Stewart,A.F. (2003)
Phage annealing proteins promote oligonucleotide-directed
mutagenesis in Escherichia coli and mouse ES cells. BMC Mol.
Biol., 4, 1.
44. Nijman,I.J., Mokry,M., van Boxtel,R., Toonen,P., de Bruijn,E.
and Cuppen,E. (2010) Mutation discovery by targeted genomic
enrichment of multiplexed barcoded samples. Nat. Methods, 7,
e137 Nucleic Acids Research, 2011,Vol.39, No. 20PAGE 10 OF 11
45. Chmielecki,J., Peifer,M., Jia,P., Socci,N.D., Hutchinson,K., Download full-text
Viale,A., Zhao,Z., Thomas,R.K. and Pao,W. (2010) Targeted
next-generation sequencing of DNA regions proximal to a
conserved GXGXXG signaling motif enables systematic discovery
of tyrosine kinase fusions in cancer. Nucleic Acids Res., 38,
46. Nefedov,M., Carbone,L., Field,M., Schein,J. and de Jong,P.J.
(2011) Isolation of specific clones from nonarrayed BAC libraries
through homologous recombination. J. Biomed. Biotechnol., 2011,
47. Zhang,P., Li,M.Z. and Elledge,S.J. (2002) Towards genetic
genome projects: genomic library screening and gene-targeting
vector construction in a single step. Nat. Genet., 30, 31–39.
48. Yang,Y. and Seed,B. (2003) Site-specific gene targeting in mouse
embryonic stem cells with intact bacterial artificial chromosomes.
Nat. Biotechnol., 21, 447–451.
49. Fan,H.C., Wang,J., Potanina,A. and Quake,S.R. (2011)
Whole-genome molecular haplotyping of single cells. Nat.
Biotechnol., 29, 51–57.
50. Kitzman,J.O., Mackenzie,A.P., Adey,A., Hiatt,J.B.,
Patwardhan,R.P., Sudmant,P.H., Ng,S.B., Alkan,C., Qiu,R.,
Eichler,E.E. et al. (2011) Haplotype-resolved genome sequencing
of a Gujarati Indian individual. Nat. Biotechnol., 29, 59–63.
51. Buecker,C., Chen,H.H., Polo,J.M., Daheron,L., Bu,L.,
Barakat,T.S., Okwieka,P., Porter,A., Gribnau,J., Hochedlinger,K.
et al. (2010) A murine ESC-like state facilitates transgenesis and
homologous recombination in human pluripotent stem cells.
Cell Stem Cell, 6, 535–546.
52. Song,H., Chung,S.K. and Xu,Y. (2010) Modeling disease in
human ESCs using an efficient BAC-based homologous
recombination system. Cell Stem Cell, 6, 80–89.
53. Zhou,L., Rowley,D.L., Mi,Q.S., Sefcovic,N., Matthes,H.W.,
Kieffer,B.L. and Donovan,D.M. (2001) Murine inter-strain
polymorphisms alter gene targeting frequencies at the mu opioid
receptor locus in embryonic stem cells. Mamm. Genome, 12,
PAGE 11 OF 11 Nucleic Acids Research, 2011,Vol.39, No. 20e137