A Retroviral Strategy That Efficiently Creates Chromosomal Deletions In Mammalian Cells

A Retroviral Strategy That Efficiently Creates Chromosomal Deletions In Mammalian Cells
of 6
All materials on our website are shared by users. If you have any questions about copyright issues, please report us to resolve them. We are always happy to assist you.
Related Documents
  A retroviral strategy that efficiently createschromosomal deletions in mammalian cells Me´lanie Bilodeau 1 , Simon Girard 1 , Jose´e He´bert 2,3 & Guy Sauvageau 1–3 Chromosomal deletions, as a genetic tool for functionalgenomics, remain underexploited for vertebrate stem cellsmostly because presently available methods are too labor-intensive. To address this, we developed and validated a setof complementary retroviruses that creates a wide range of nested chromosomal deletions. When applied to mouseembryonic stem cells (ESCs), this retrovirus-based methodyielded deletions ranging from 6 kb to 23 Mb (average 2.9 Mb),with an efficiency of 64% for drug-selected clones. Notably,several of the engineered ESC clones, mostly those with largedeletions, showed major alteration in cell fate. In comparisonto other methods that have also exploited retroviruses for chromosomal engineering, this modified strategy is moreefficient and versatile because it bypasses the need for homologous recombination, and thus can be exploited for rapid and extensive functional screens in embryonic andadult stem cells. Capitalizing on the reliability of Cre- loxP  –based recombination, agroup previously reported the generation of nested chromosomaldeletions in mouse ESCs by sequentially delivering two  loxP  sequences into the genome, followed by Cre-mediated excision of the chromosomal region between the  loxP   sites 1 . In this approach,the first  loxP   sequence was introduced into a particular locus of choice by homologous recombination using a targeting vectorthat included a nonfunctional ‘split’  Hprt1  cassette. The second loxP   and complementary   Hprt1  sequences were delivered usingretroviral gene transfer. Cre-mediated recombinants wereselected in HAT medium after reconstitution of the functional Hprt1  mini-gene 1 . Although this method considerably improvedour ability to generate high-resolution sets of nested deletionsaround a targeted locus,its extension to severalother lociremainedlabor-intensive and precluded large-scale functional screens inESCs. Moreover, by its nature, this method was limited to cellspermissive to homologous recombination, thus excluding mostmammalian cells. Here we sought to overcome these limitationsand elected to develop a strategy that would strictly rely on theuse of replication-defective retroviruses while exploiting theCre- loxP   recombination system and reconstitution of a functionalneomycin ( neo ) cassette for selection of recombination events.The first  loxP   sequence is delivered using a vector that werefer to as the anchor virus and the second by a saturatingvirus ( Fig. 1a , b ). RESULTSSelection of anchor and saturation proviruses From a series of 10 different retroviral constructs consistingof 5 anchor (A1–A5) and 5 saturation viruses (S1–S5), weselected viruses A1 and S1 based on the criteria listed in Supplementary Figure 1  online. We expected chromosomaldeletions to have occurred in geneticin-resistant (G418 R  )clones that have lost both puromycin- and hygromycin-resistancegenes ( Fig. 1a ).Using retroviral preparations adjusted to provide gene transferto mouse R1 ESCs of   o 1%, we first confirmed that mostclones infected with virus A1 and selected on puromycin (hereaftercalled primary clones) had a single integrated provirus (data notshown). We expanded 11 randomly selected puromycin-resistant(puro R  ) clones and infected them, in 1–4 independent experiments(A–D) per each clone, with low-titer S1 virus to generate aseries of hygromycin-resistant (hygro R  ) populations with a com-plexityof  B 20,000independent secondaryclones( Supplementary Table 1  online). After  cre  electroporation, we observed G418 R  recombinants (hereafter called tertiary clones) for each of the 11 primary clones analyzed with an average frequency of 2.5 ± 2.2    10 –4 ( Supplementary Table 1 ). We verified expectedCre-induced rearrangements between the integrated A1 and S1proviruses by Southern blot analyses for several tertiary clonesderived from each of the 11 families ( Fig. 1c  and  Supplementary Fig. 2  online). We never observed spontaneous G418 resistance inthe absence ofCre expression. We confirmed productive rearrange-ment leading to the expression of the neomycin gene for severaltertiary clones in which we detected a single messenger RNAof  B 1.0 kb ( Fig. 1d ). Fragments corresponding to the recombina-tion junctions ( Pgk  - loxP-neo ) were also PCR-amplified fromrepresentative G418 R  tertiary clones and sequenced, confirmingthe expected breakpoint ( n  ¼  5 clones selected in 4 families;data not shown). RECEIVED 19 SEPTEMBER 2006; ACCEPTED 5 JANUARY 2007; PUBLISHED ONLINE 4 FEBRUARY 2007; DOI:10.1038/NMETH1011 1 Laboratory of Molecular Genetics of Stem Cells, Institute for Research in Immunology and Cancer (IRIC), Universite´ de Montre´al, Montre´al, Que´bec, Canada, H2W1R7.  2 Department of Medicine, Montre´al, Que´bec, Canada, H3C 3J7.  3 Leukemia Cell Bank of Quebec and Division of Hematology, Maisonneuve-Rosemont Hospital,Montre´al, Que´bec, Canada, H1T 2M2. Correspondence should be addressed to G.S. ( NATURE METHODS  |  VOL.4 NO.3  |  MARCH 2007  |  263 ARTICLES    ©   2   0   0         7     N  a   t  u  r  e   P  u   b   l   i  s   h   i  n  g   G  r  o  u  p    h   t   t  p  :   /   /  w  w  w .  n  a   t  u  r  e .  c  o  m   /  n  a   t  u  r  e     m    e     t     h    o     d    s  Evaluation of chromosomal deletions Among the different chromosomal rearrangements obtained, wescreened deletions by testing for the concomitant loss of puromy-cin- and hygromycin-resistance genes (puro S and hygro S ; see Supplementary Table 1  for frequencies), which occurred in 9 of the 11 families. We used inverse-PCR (I-PCR), array-based com-parative genomic hybridization (aCGH) and spectral karyotypingto confirm deletions in several tertiary clones from 8 of these 9different families and to assess the genomic integrity of the alteredESCs ( Fig. 1e ,  Table 1  and  Supplementary Fig. 3  online). The sizedistribution of deleted DNA fragments varied according to thefamily studied ( Fig. 1f   and  Table 1 ), ranging from 6 kb to 23 Mb,with an average of 2.9 ± 5.2 Mb. Notably, as noticed with tertiary clonesderived from family9,deletion sizes did not follow a normaldistribution as they either ranged in the scale of kilobase pairs(6–317kb, n ¼ 8independentdeletions)ormegabasepairs(4.2–5.0Mb,  n ¼ 3 independent deletions;  Fig. 1f   and  Table 1 ). On average,I-PCR-confirmed deletions included 21 ± 46 genes, 15 ± 28 CpGislands, 1,689 ± 3,283 spliced expressed sequence tags (ESTs) and0 ± 1 microRNA ( Table 1  and data not shown). The frequency of clones with a deletion that were free of other rearrangements was0.71( Table1 ),indicatingthatthefrequencyofvaluabledeletionsina pool of G418 R  colonies was 0.15 (0.26_frequency of puro S clones   0.9_frequency of puro S hygro S clones    0.9_frequency of deletions confirmed out of the puro S hygro S clones with indepen-dent rearrangements   0.71 frequency of deletions without otherrearrangements, confirmed by aCGH and spectral karyotyping).This frequencyofdeletionsis probably an underestimation becausewe excluded from the analysis a subgroup of clones that showedambiguous sensitivity to puromycin or hygromycin (puro S+R  hygro S+R  ). DNA analyses suggested that these puro S+R  hygro S+R  cells represented at best a minor fraction of our G418 R  coloniesbecause, most of the time, we could not detect a signal to thesegenes in the selected clones ( Supplementary Fig. 2 ).We were able to engineer large chromosomal deletionsfor most of the regions tested in our study. Two anchorsites in primary clones 12 and 15, located on chromosomeX and 11, respectively, were not permissive for deletion.This might suggest the proximity of a haplolethal determinantfor ESCs or the presence of physical constraints, such as chromatinstructure, preventing recombination between  loxP   sites orientedfor deletions. KKKKHHA15 ′ SIN5 ′ SIN3 ′ SIN5 ′ LTR3 ′ LTR3 ′ LTR neo   ATG-less neo   ATG-less ATG  - Pgk ATG  - Pgk Pgk  -puro3.4 kbhygroS12.9 kbCreKKHHA1-S1 * 1.0 kb mRNA3.0 kb Family 91 ° 2 ° 3 °    7   5   7   4   4   9   4   8   4   7   1   8   1   7   1   6   R   1 neo  2.0 kb2.0 kb2.0 kb4.9 kb4.9 kbRRRRRSSSSHygroPuroShygropuropurohygroA1 (3.4 kb)A1 (3.4 kb)S1 (2.9 kb)A1-S1 (3.0 kb)A1    C   l  o  n  a   l  a  n  a   l  y  s  e  s   (    H   i  n   d    I   I   I   )   R  e  a  r  r  a  n  g  e  m  e  n   t  a  n  a   l  y  s  e  s   (    K  p  n    I   ) A1A1S1CreA1-S1+++RecombinationAnchoring(primary clone)Saturation(secondary clones)(tertiary clones) 18S28S*A1-S1 neo  R messenger(~1 kb)ClonesFamilies neo  0.24 kb1.35 kb10    1   8   1   8   1   7   1   0   4   1   0   7   3   1   6   8   9   3   9   0   6   2   2   9   3   7   3   5   7   1   R   1 1 ° 3 ° 3 ° 1 ° 962 Mb57 Mb9-189-37, 9-1049-68, 9-71, 9-90, 9-1079-17, 9-29, 9-31, 9-35,qE160 MbqD3I-PCR A1 anchor siteS1 site using aCGHS1 site using I-PCR    9  -   1   8   9  -   3   7   9  -   9   0   9  -   3   5   A  n  c   h  o  r   9   R   1 3 ° 1 ° aCGHAnchor 9,9-35, 9-909-379-18SKYAnchor 9 (8/15 40, XY)9-18 (10/15 40, XY)9-104 (8/15 40, XY) YX19181716 YX19181716 YX19181716 111213141511121314151112131415109876 109876 109876 123451234512345 14131097641 1Independent deletions confirmedin each family of clones101001,00010,000100,000    D  e   l  e   t   i  o  n  s   i  z  e   (   k   b   ) a cd fbe Figure 1  | Cre-induced chromosomal rearrangements in mouse ESCs. ( a ) Representation of the recombination between A1 and S1 proviruses. After Cretransfection, the coupling of the  Pgk  -ATG in S1 to the neomycin (ATG-less  neo ) gene in A1 allows the selection of recombinants. Deletions are identified bythe concomitant losses of puromycin (puro) and hygromycin (hygro) resistance genes. Symbols are detailed in  Supplementary Figure 1 . ( b ) Cartoon of nesteddeletions sharing the same endpoint (virus A1). ( c ) Southern blot analysis of DNA isolated from selected clones documents the integrity of provirus A1 (3.4 kb)and S1 (2.9 kb), and their successful recombination (A1-S1, 3.0 kb). Clonal diversity is shown in the two bottom panels. 1 1 , 2 1  and 3 1 , primary, secondary andtertiary clones, respectively. S, sensitive; R, resistant. Unmodified blots are presented in  Supplementary Figure 2 . ( d ) Neomycin resistance gene expression inpresence (lanes 3–15 and 17) or absence (lanes 2 and 16) of Cre treatment. Note: the 1-kb transcript is indicated by an asterisk in  a . ( e ) CGHAnalyzer  20 representation (middle) linked to the chromosomal localizations of confirmed deletions (Ensembl Karyoview 19 ; left) for selected clones in family 9. Red linesrepresent nested deletions for the indicated clones. Spectral karyotyping for selected clones (right). Scale bars, 10  m m. S1 proviral integration site of selecteddeletions confirmed by I-PCR or inferred from aCGH analyses. ( f  ) Size distribution of confirmed deletions in indicated families. Dots and strokes representindependent deletion and average deletion size per family, respectively. 264  |  VOL.4 NO.3  |  MARCH 2007  |  NATURE METHODS ARTICLES    ©   2   0   0         7     N  a   t  u  r  e   P  u   b   l   i  s   h   i  n  g   G  r  o  u  p    h   t   t  p  :   /   /  w  w  w .  n  a   t  u  r  e .  c  o  m   /  n  a   t  u  r  e     m    e     t     h    o     d    s  Interchromosomal recombination events Interchromosomal events are expected to give loss either of puromycin- or of hygromycin-resistance gene, or conservationof both resistance genes, but not their concomitant loss 2 .In the course of the aCGH and spectral karyotyping analyses, weunexpectedly observed two interchromosomal rearrangementsfrom a group of 22 puro S hygro S clones (believed to representdeletions), one of which was a confirmed translocation (clone14-27: t(2;16), translocation between chromosome 2 and 16; Supplementary Fig. 4  online). A possibility that could accountfor this phenomenon is the loss of a chromosome (for example theloss of the recombined chromosome bearing the puromycin- andhygromycin-resistancegenessrcinatingfromrecombinationinG1or in G2 with Z-segregation 3 ), accompanied by the duplication of the homologous chromosome 4 ( Supplementary Fig. 4 ).From a subgroup of 190 tertiary clones selected for furtheranalyses, eight had sensitivity to either puromycin or hygromycin,and represented seven independent rearrangements as assessed by clonal analysis of proviral integration. Of these seven clones, twocontained productive (that is, confirmed by I-PCR analysis)transchromosomal rearrangements (that is, 2-03 and 14-32;  Sup-plementary Fig. 4 ). Notably, the frequency of single loss of puromycin-resistance gene varied according to the family, showinghighest values for families 4 and 14 in which the anchor viruswas located close to the telomeric ends of the chromosome 2( Supplementary Table 1 ).Together these results suggest that recombination events in  trans occur at higher frequency for particular loci, but also at low frequency in the puro S hygro S clones, highlighting the importanceof complementary analyses (for example, spectral karyotyping) forthese types of studies.  In vitro  and  in vivo  differentiation of recombined clones To gain insights into the potential of our approach to generateclones that can be used in a functional screen  in vitro , we selected 43tertiary clones from9 families to cover awide range of deletion sizes(from 6 kb to 23 Mb) and differentiated them into embryonicbodies for identification of phenotypic anomalies ( Supplementary Table2 onlineand Fig.2a , b ).One-third(3/9)ofthefamiliesstudiedcontained clones that had major differentiation anomalies, repre-senting 11% (5/43) of our sample ( Supplementary Table2 ). Clonesin families 1 and 9 are particularly interesting because they cover awide range of deletions ( Supplementary Table 2 ), and only cloneswith larger deletions show phenotypical anomalies. For example,clone 1-03 included a 1.5-Mb deletion and differentiated normally,whereas clone 1-13, with an B 23-Mb deletion, did not differentiate in vitro . The correlation between deletion size and phenotypeis more striking for family 9 where all 8 clones having less than a318-kb deletion show normal  in vitro  differentiation, whereas 3 of the3cloneswithgreater than4.1-Mb deletionsdidnotdifferentiate.Reinforced by the observation that most of these clones lackedadditional DNA rearrangement as assessed by aCGH and spectralkaryotyping analyses ( Table 1 ), these results argue against othergenetic events being responsible for these phenotypes. Additionalevidence to support this argument includes the observationthat most of the clones with differentiation anomalies (for example, Table 1  |  Characteristics of independent deletions confirmed by I-PCR and aCGH Tertiaryclone a Datasource ChromosomeStartcoordinateEndcoordinateSize(kb)Number of Refseq genesNumber of spliced ESTsNumber of CpG islandsaCGHconfirmationGenomicanomaly b 1-03 I-PCR 14 22165099 23710534 1,545 17 1,080 14 (+) (–)1-13 I-PCR 14 22165099 44937742 22,773 206 14,400 126 (+) (–)4-2 I-PCR 2 167486681 168900222 1,414 19 1,819 30 n.d. n.d.6-36 I-PCR 17 26955839 27622141 666 13 1,272 14 (+) (+) d,e 7-30 I-PCR 16 35918443 36011960 94 3 155 1 (+) c (–)9-31 I-PCR 18 57155985 57162362 6 0 0 0 n.d. n.d.9-107 I-PCR 18 57155985 57166502 10 0 0 0 n.d. n.d.9-71 I-PCR 18 57155985 57174937 19 0 0 0 n.d. n.d.9-17 I-PCR 18 57155985 57174941 19 0 0 0 n.d. (-) d 9-68 I-PCR 18 57155985 57175174 19 0 0 0 n.d. n.d.9-29 I-PCR 18 57155985 57177486 22 0 0 0 n.d. n.d.9-35 I-PCR 18 57155985 57179077 23 0 0 0 (+) c (-)9-90 I-PCR 18 57155985 57473132 317 2 114 4 (+) c (+) f  9-104 I-PCR 18 57155985 61338307 4,182 20 3,289 18 n.d. (+) g 9-37 I-PCR 18 57155985 61468765 4,313 21 3,419 20 (+) (–) d 9-18 I-PCR 18 57155985 62204954 5,049 32 4,726 27 (+) (–)10-18 I-PCR 16 59749084 65165857 5,417 12 481 10 (+) (–)10-21 I-PCR 16 57307345 65165857 7,858 51 1,371 22 (+) (+) h 13-34 aCGH 4 78266064 82222600 3,956 14 792 10 (+) (–)14-16 I-PCR 2 156503387 157071542 568 9 869 8 (+) c (–)Average 2,914 21 1,689 15 (–) frequency: 0.71s.d. 5,244 46 3,283 28 Mapping and deletion analyses were done using the UCSC Genome Browser  17,18 .  a Tertiary clones are labeled according to their family number (same integration of virus A1), followed by a specific identifier number.  b Anomaly that was not present in the primary clone from which the tertiary clone was derived, as determined by aCGH or spectral karyotyping. (–), no additional anomaly; (+), additional anomaly; n.d., notdetermined.  c The deletion is not observed, in agreement with the resolution of aCGH.  d Normal except for the loss of chromosome Y.  e Amplification of chromosome 1.  f  Amplification of chromosome 8.  g Manychromosomes were lost according to spectral karyotyping.  h Amplification on chromosome 14. NATURE METHODS  |  VOL.4 NO.3  |  MARCH 2007  |  265 ARTICLES    ©   2   0   0         7     N  a   t  u  r  e   P  u   b   l   i  s   h   i  n  g   G  r  o  u  p    h   t   t  p  :   /   /  w  w  w .  n  a   t  u  r  e .  c  o  m   /  n  a   t  u  r  e     m    e     t     h    o     d    s  9-104) show embryonic body formation at low frequency. In thisclone (9-104), we conducted fluorescence  in situ  hybridization(FISH) analysis using a bacterial artificial chromosome (BAC)probe corresponding to the deleted region of chromosome 18.This analysis revealed that 94% and 6% of undifferentiated 9-104cells showed one versus two signals, respectively. Among the raredifferentiated cells derived from this clone, however, 60% displayedtwo signals ( Fig. 2c ). Clonal analysis of DNA extracted from theserare differentiated cells confirmed their srcin from clone 9-104( Fig. 2c  and  Supplementary Fig. 5  online), ruling out possiblecontaminants as an explanation for this complementation.It thus appears likely that these cells have reacquired the deletedsequence. This low frequency of revertants is consistent withthe chromosomal instability observed in ESCs 5 and wasconfirmed by the extensive aCGHand spectral karyotyping analysesreported herein. Most importantly, this observation documents thelow frequency of spontaneous revertants, thus strengtheningthe argument that differentiation is dependent on the presence of the deleted fragment.Consistent with the  in vitro  results, the ESCs from tertiary clone9-35 (23 kb deletion, no phenotype) contributed to the generationofchimeric14.5dayspostcoitum(d.p.c.)fetus andviablenewbornmice with an overall proportion of 75% and 30%, respectively (3/4 at E14.5, and 3/10 at birth and adulthood;  Table 2 ). Germlinetransmission of these ESCs was documented by coat color analysisalthough the perpetuation of the deleted chromosome was notdocumented. Of interest, embryos injected with ESCs fromclone 9-18 (limited potential to differentiation  in vitro ,  Fig. 2a )showed a high mortality rate at embryonic day 14.5 (46% viable, Table 2 ) with undetectable ESC-derived contribution (DNAanalysis and coat color) evaluated in six 14.5d.p.c. fetuses and in 31adults ( Table 2 , Fig. 2d – e  and data not shown). Thus,within the limit of these analyses, there isa good correlation between the embryonicbody formation competency   in vitro  andcontribution to chimerism  in vivo  for ourdeleted ESC clones. These results alsodocument that a subgroup of ESCs thatunderwent our procedures remain compe-tent for the creation of chimeras, thuspaving the way to use this strategy for in vivo  studies. loading5.1 kbloading5.1 kb neo neo  1433 0000 2233 0021444222442200 neo  +  (PCR)Total analyzedHeadsHematopoieticcoloniesES cells andfetal livers    0 2   5   5   0   7   5   1   0   0   1 3 4 5 6 7 8 9 1   0   1   1   2   5   5   0   7   5   1   0   0   4 5 6 7 2   5   5   0   7   5   1   0   0   3 5 6 8 1   1   1   2 ChimerasChimerasChimerasDNADNADNA% in R1% in R1% in R1Primary 9Tertiary 9-35Tertiary 9-189-18Number 10Abnormal9-18Number 5Abnormal9-18Number 4Abnormal9-35Number 6Normal In vivo In vitro  Kpn  l  Bgl  lI neo  5.1 kb    E   B   E   S   E   S   E   S   E   S   E   B   E   B   E   B    9  -   1   8   9  -   1   8   9  -   1   0   4   9  -   1   0   4 9-104 EB9-104 ESno. 9 ESno. 9 EB2 signals: 93%1 signal: 7%2 signals: 60%1 signal: 40%2 signals: 6%1 signal: 94%2 signals: 99%1 signal: 1% Differentiated (–LIF)Undifferentiated    T  e  r   t   i  a  r  y   P  r   i  m  a  r  y Tertiary 9-35Primary 9R1 Clone ID 9-18( n   = 4)9-104( n   = 3)9-35( n   = 1)9( n   = 4)R1( n   = 4)0255075100125150175200    N  u  m   b  e  r  o   f   E   B  s  p  e  r   6   0  m  m    d   i  s   h ** adeb c Figure 2  |  In vitro  and  in vivo  differentiation of ESC clones with deletions. ( a ) Day 7–8 embryoid body formation (mean ± s.e.m.) of parental R1 ESCs, primaryclone 9 and selected tertiary clones. *, low frequency of embryonic body formation is observed with 10–100  higher seeding density. ( b ) Examples of day 7–8embryoid bodies generated for selected clones. Scale bar, 250  m m. ( c ) FISH analysis using BAC RP23-109P21 as a probe. Note the relative signal distribution inundifferentiated and differentiated cells. Green signals were enhanced using Adobe Photoshop CS imaging tool to replicate visualization on LCD monitor. Scalebar, 10  m m. Recombination and clonal analyses ( Kpn I and  Bgl  II restriction enzymes, respectively) of DNA extracted from indicated ESCs (ES) and embryoidbodies (EB). ( d ) Pictures of 14.5 d.p.c. fetuses. Scale bar, 2 mm. ( e ) Southern blot analysis ( Bgl  II restriction digests;  neo  probe) or PCR studies of genomic DNAextracted from the indicated cells. These included ESC clones 9, 9-35, 9-18, R1 control and 14.5 d.p.c. chimeric fetus livers, heads or hematopoietic coloniesderived from fetal liver cells (numbers shown between upper and lower panels). Note the absence of contribution for clone 9-18 to the chimeric fetuses.For   c  and  e , unmodified blots are presented in  Supplementary Figure 5 . Table 2  |  Analysis of chimeras Reimplantation Fetus 14.5 d.p.c. MiceESCcloneNumber of embryosimplantedNumber of fetusesobserved/ expectedProportionnormal (%)Proportion of chimeric fetus a Number of neonatesobserved/ expectedProportion of chimeric mice b 9 173 19 / 32 89 10 / 18 56 / 141 7 / 529-35 36 7 / 8 100 3 / 4 11 / 28 3 / 109-18 67 15 / 16 46 0 / 6 31 / 51 0 / 31 a According to Southern blot or PCR analysis of DNA, or eye-pigmentation analysis.  b According to coat-color analysis. 266  |  VOL.4 NO.3  |  MARCH 2007  |  NATURE METHODS ARTICLES    ©   2   0   0         7     N  a   t  u  r  e   P  u   b   l   i  s   h   i  n  g   G  r  o  u  p    h   t   t  p  :   /   /  w  w  w .  n  a   t  u  r  e .  c  o  m   /  n  a   t  u  r  e     m    e     t     h    o     d    s  DISCUSSION In this study we developed a system, entirely based on retroviruses,to engineer chromosomal deletions in the genome of mammaliancells. As this technology relied on a pair of complementary retroviruses to deliver  loxP   sites, it bypassed the laborious step of homologous recombination, considerably accelerating thecreation of large deletions that can be easily mapped throughI-PCR. We tested 10 different viruses before we identified aneffective pair. We validated the functionality of this system throughthe analysis of 11 different families including several independentclones and provided evidence for an efficiency of deletionsnearing 64% for clones that are selected based on sensitivity to puromycin and hygromycin. We also showed that the averagedeletion in these clones is  B 2.9 Mb in size, suggesting thata complexity of 10 3 primary clones could cover a haploid genomein the mouse, providing that our anchor virus shows no preferencefor integration.The method described here complements other functionalgenomics strategies applicable to mammalian cells 6–11 . Althoughcertain limitations of the proposed method remain to be deter-mined (integration of retrovirus in gene-poor regions in ESCs,epigenetic changes in long-term cultured ESCs, ability to producehomozygous deletions using high G418 concentrations 12 ), ourprocedure should be easily amenable to high-throughput screens.We suspect that our complementary viruses and deletion strategy will be particularly useful for functional screens that involve cellsthat have poor frequencies of homologous recombination (forexample, human ESCs) and to identify fragments of DNA involvedin tumor progression. METHODS Retroviral constructs.  We generated the A1 retroviral construct(plasmid 1647) by inserting both a  loxP  –ATG-less– neo R  PCR cassette from pPNT 13 and a SV40 early mRNA polyadenylationsignal fragment from pDsRed2-N1 (Clontech) into pRETRO-SUPER  14 linearized by   Xho I- Eco RI (blunt) as indicated ( Supple-mentary Fig. 1 ). For S1 virus (plasmid 1643), we placed a Pgk-kozac- ATG -loxP   fragment in reverse orientation to a hygrocassette in  Hpa I-linearized MSCV vector from which we removedthe neomycin resistance gene. Inverse PCR, sequencing and mapping.  We linearized 0.5  m g of genomic DNAwith 20 U of a restriction enzyme (Invitrogen), in atotal volume of 20 m l. We used either  Eco RI or  Stu I for the primary clones, and  Bst  EII or the double-digest  Bgl  II- Bam HI for thetertiary clones. After ethanol precipitation in presence of 0.5  m lof linear polyacrylamide carrier 15 , DNA was resuspended in 24  m lof HPCL-grade water (J.T. Baker). We put aside 4  m l of this linearDNA to be used as the PCR negative control and circularized 20  m lusing T4 DNA ligase (Invitrogen, 4 U in a total volume of 45  m l,incubated overnight at 16  1 C). We precipitated the ligated productwith ethanol (and carrier) and resuspended again in 24  m l of HPCL-grade water. We carried out the first PCR round using4  m l of ligated DNA, 1   PC2 reaction buffer (AbPeptides),0.25 mM of each of the four dNTPs (Invitrogen), 2 mM MgCl 2 ,20 pmol of forward and reverse primers (BioCorp), 5 U of KlenTaq LA-16 DNA polymerase 16 (mix 15:1 of Klentaq1 fromAbPeptides and  Pfu  from Stratagene), in a total volume of 50  m l.We carried out PCR in a Perkin Elmer thermocycler using thefollowing parameters: 2 min at 94  1 C for one cycle, 20 s at 94  1 C,30 s at 63  1 C, 15 min at 68  1 C for 10 cycles, 20 s at 94  1 C, 30 s at 63 1 C, 15 min at 68  1 C with a 20-s autoextension for 20 cycles, andfinally an extension of 30 min at 68  1 C. We diluted thePCR product 1:10,000 to 1:50,000 and used this in a secondPCR round with nested primers. We used the same settingsexcept with the annealing temperature at 65  1 C. For bothprimary and tertiary clones, the PCR primers for the first PCR round were (longNEO-F2) 5 ¢ -TGGCCGCTTTTCTGGATTCATCGACTGTGG-3 ¢  and (long-NEO-R) 5 ¢ -AAGCGGCCGGAGAACCTGCGTGCAATC-3 ¢ . The second round of PCR was done withprimers (longPgk-R) 5 ¢ -GGCGCCTACCGGTGGATGTGGAATGTGTG-3 ¢  and (long-NEO-R) for primary clones, and with(longPack-R) 5 ¢ -GGCGGATGGAGGAAGAGGAGGCGGAGG-3 ¢ and (longPgk-R) for tertiary clones, respectively. We separatedPCR products on 0.8% agarose gel and purified them with theQIAEX II Gel Extraction Kit (Qiagen). We subcloned fragments inpBluescript (Stratagene, T3 and T7 sequencing primers) orsequenced directly using one of the nested primers. We processedsamples using a dideoxy chain termination method and the3730XL DNA Analyzer system (ABI), at the Genome QuebecInnovation Center (McGill University, Montreal, Canada).Mapping and deletion analyses were done using the UCSCGenome Browser (, NCBI mouseBuild 33) 17,18 . We used the Ensembl Genome Browser for sche-matic representations of deletions (, v.32-Jul 2005) 19 . Additional methods.  Descriptions of pCX-cre plasmid, cellculture, viral production and transfection, RNA and DNAanalyses, aCGH, spectral karyotyping and FISH, chimera produc-tion, equipment and settings are available in  Supplementary Methods  online. Accession codes.  Gene Expression Omnibus (GEO): GSE6706. Note: Supplementary information is available on the Nature Methods website. ACKNOWLEDGMENTS We thank P. Chartrand, M. Therrien and colleagues for critically reading themanuscript; V. Paradis, S. Harton and E. Milot from the transgenic facilityof IRIC; J. Cowell, M. Rossi and D. McQuaid for the aCGH service (Roswell ParkCancer Institute); J.-P. Laverdure from the bioinformatic service of IRIC;C. Charbonneau from the imaging service of IRIC; N. Fradet, M. Fre´chette,A. Fredette, E. St-Hilaire, T. MacRae, C. Rondeau and P. Lussier for technical assistance; A. Nagy for providing the R1 ESCs and the pCX-EYFP construct(Samuel Lunenfeld Research Institute, Mount Sinai Hospital, Toronto);A. Bradley for the pOG231 construct (Wellcome Trust Sanger Institute, WellcomeTrust Genome Campus, Hinxton); R.G. Hawley for the MSCV vectors (The GeorgeWashington University Medical Center); M. van Lohuizen for the pRETRO-SUPERconstruct (The Netherlands Cancer Institute) and R. Jaenisch for the DR-4 mousestrain (Whitehead Institute for Biomedical Research, Massachusetts Instituteof Technology). This work was mostly supported by a grant from Ge´nomeQue´bec to G.S. and in part by a grant from the Re´seau de Recherche enTransgene`se du Que´bec to J.H. M.B. is a recipient of a Canadian Institutesof Health Research (CIHR) studentship, and G.S. is a recipient of a CanadaResearch Chair in molecular genetics of stem cells and a scholar of theLeukemia Lymphoma Society of America. AUTHOR CONTRIBUTIONS M.B. performed all the experiments and the analyses described herein, exceptfor I-PCR (S.G.), spectral karyotyping and FISH analyses (J.H.). aCGH experimentsand chimeras production were conducted by the services mentioned above. M.B.wrote the manuscript, prepared all the figures and performed the experimentsunder the guidance of G.S. NATURE METHODS  |  VOL.4 NO.3  |  MARCH 2007  |  267 ARTICLES    ©   2   0   0         7     N  a   t  u  r  e   P  u   b   l   i  s   h   i  n  g   G  r  o  u  p    h   t   t  p  :   /   /  w  w  w .  n  a   t  u  r  e .  c  o  m   /  n  a   t  u  r  e     m    e     t     h    o     d    s
Related Search
We Need Your Support
Thank you for visiting our website and your interest in our free products and services. We are nonprofit website to share and download documents. To the running of this website, we need your help to support us.

Thanks to everyone for your continued support.

No, Thanks