Transgenic Xenopus laevis embryos can be generatedusing /C31 integraseBryan G Allen & Daniel LWeeks

Bacteriophage /C31 encodes an integrase that can mediate the

insertion of extrachromosomal DNA into genomic DNA. Here we

show that the coinjection of mRNA encoding /C31 integrase with

plasmid DNA encoding the green fluorescent protein (GFP) can be

used to generate transgenic X. laevis embryos. Despite integra-

tion into the genome, appropriate promoter expression required

modification of the reporter plasmid by bracketing the GFP

reporter gene with tandem copies of the chicken b-globin 5¢ HS4

insulator to relieve silencing owing to chromatin position effects.

These experiments demonstrate that the integration of insulated

gene sequences using /C31 integrase can be used to efficiently

create transgenic embryos in X. laevis and may increase the

practical use of /C31 integrase in other systems as well.

The ability to generate transgenic animals has revolutionizedstudies of gene function. Transgenic frogs of the genus Xenopushave been made using methods based on a restriction enzyme–mediated sperm integration approach1. This method has beenremarkably successful, but can be technically demanding. In addi-tion, many embryos contain several copies of the transgene insertedat random insertion sites1,2.

We tried an alternative approach, creating transgenic Xenopusembryos by using fC31 integrase–mediated recombination. fC31is a bacteriophage that encodes an integrase that mediates sequence-directed recombination between a 34-nucleotide bacterial attach-ment site (attB) and a 39-base-pair (bp) phage attachment site(attP). fC31 integrase has a high efficiency of recombination,requires no accessory factors3,4 and has been effectively used tointegrate genes into plant5 or mammalian6–12 cells and fruit flyembryos13. The fC31 integrase does not require perfect attP siteconservation to cleave the attP site8. These imperfect attP sites, orpseudo-attP sites, may have a similarity as low as 24% to an attP siteand still allow recombination8. It has been estimated that amammalian genome may contain 100–1,000 pseudo-attP sites8.As the X. laevis genome is approximately the same size as amammalian genome (3 � 109 bp), we hypothesized that therewould also be pseudo-attP sites in its genome that could be used tocreate transgenic Xenopus embryos. In the experiments that follow,all the plasmids injected are pEGFPB2, or derivatives thereof, whichcontain an attB site and the CMV promoter driving GFP expression.

We found that fC31 integrase can mediate integration ofplasmid DNA into the Xenopus genome. We also show that toachieve promoter-appropriate expression, the integrated DNAneeds to be insulated to prevent chromatin position effects.

RESULTS/C31 integrase integrates plasmid into the Xenopus genomeInjection of plasmids into Xenopus embryos can recapitulatepromoter-specific temporal and spatial expression; however, theirexpression is mosaic, randomly failing to transcribe in all of the cellsthat they should14. In addition, gene expression from plasmid DNArarely extends to late stages (stage 42 and beyond) of develop-ment14. Typically 50–100 pg of plasmid is used in plasmid injectionexperiments14. An example of a stage 39 embryo injected with100 pg of a CMV-gfp reporter plasmid demonstrating mosaic GFPexpression is shown in Figure 1a. The half-life of the GFP proteinhas been estimated to be as long as 80 h. Most of the embryos thatare initially GFP-positive no longer fluoresce by stage 46, but about10% still glow at this stage in a mosaic pattern (Fig. 1b).

We reduced the amount of plasmid injected such that we nolonger detected CMV-gfp expression in the embryo. We screenedhundreds of embryos from three different mating pairs of frogs,and all were negative for fluorescence when we injected 5 pg of theCMV-gfp reporter plasmid. We conclude that this concentration istoo low to routinely see expression using pEGFPB2 as an extra-chromosomal reporter plasmid. To test whether the fC31 integrasecan be used to insert DNA into the Xenopus genome, we injected5 pg of CMV-gfp reporter plasmid into single-cell Xenopus embryosalong with 1 ng of mRNA encoding the fC31 integrase protein(integrase mRNA). The earliest opportunity for expression of thereporter plasmid coincides with the activation of embryonictranscription at the 4,000-cell midblastula stage (stage 7.5, about8 h after fertilization)15. We were able to detect green fluorescenceas early as stage 14 (about 20 h post fertilization) and monitoredexpression through stage 46 (about 8 d post fertilization; Fig. 1).

We focused on embryos that appeared developmentally normalat stage 46. We note that although survival varied depending on thespecific mating pair of adults, overall there was no deleterious effectof plasmid or integrase mRNA on the health of the embryos whencompared to control injections (Table 1).

p

uor

G g

n ih si l

bu

P eru ta

N 500 2©

e rut a

n/m

oc.erut a

n.w

ww//:

ptth

sdoht

em

RECEIVED 4 AUGUST; ACCEPTED 17 OCTOBER; PUBLISHED ONLINE 18 NOVEMBER 2005; DOI:10.1038/NMETH814

Department of Biochemistry, Bowen Science Building, University of Iowa, Iowa City, Iowa 52242, USA. Correspondence should be addressed to D.L.W. (daniel-weeks@uiowa.edu).

NATURE METHODS | VOL.2 NO.12 | DECEMBER 2005 | 975

ARTICLES

Coinjection of 5 pg of the CMV-gfpreporter plasmid with integrase mRNAresulted in approximately one-third of theembryos scoring positive for green fluores-cence. GFP expression was limited to asubset of tissues, despite being driven by aCMV promoter (Fig. 1c–f). Embryos couldbe sorted into three groups based on GFPexpression. We observed that 31% hadexpression in the eyes and the forebrain(Fig. 1d), 57% showed expression insomatic muscle tissue along the tail(Fig. 1e), and 7% had expression in gillstructures (Fig. 1f). Although the embryosrepresented in Figure 1d,f have a verydifferent expression pattern than we haveobserved with reporter plasmid–aloneinjection, those in Figure 1e are reminiscentof the mosaic expression from the uninte-grated plasmid. We confirmed integrationof the reporter plasmid by Southern blotanalysis (Fig. 2). We did not routinely detectthe extrachromosomal reporter plasmid,indicating that it did not persist in a non-integrated state. It is worth noting that one-quarter of nonfluorescing embryos tested bySouthern blot analysis still had an integrated copy of the GFPplasmid. These experiments suggest that the fC31 integrase canmediate recombination between a plasmid carrying an attB site andthe Xenopus genome, but we did not obtain transgenic embryoswith the expected expression pattern.

Chromatin position effects are overcome using insulatorsThe CMV promoter should have been active in every tissue, so wesought to explain why GFP expression was not ubiquitous. Ascells differentiate, multicellular eukaryotes render regions ofchromosomes transcriptionally inactive by a process called silenc-ing16,17. Chromatin silencing can spread to neighboring loci, andthe insertion of a transgene into a silenced region of chromatin

prevents the expression of the transgene. The transcriptionalsilencing (or the inappropriate activation by insertion next to anenhancer) of a transgene based on where it inserts into the genomeis referred to as chromosome position effect18. Insulators are DNAsequences that protect genes from chromosome positioneffects16,17. One of the best characterized insulators is the chickenb-globin 5¢ HS4 insulator19,20. Experimentally it has been shownthat tandem copies of the 250-bp core region of the chickenb-globin 5¢ HS4 insulator are sufficient to block the spread ofchromatin silencing17,19,20.

We inserted duplicated copies of the chicken b-globin 5¢ HS4250-bp insulator flanking the CMV-driven gfp into the pEGFPB2reporter plasmid. We injected 5 pg of insulated CMV-gfp reporter

p

uor

G g

n ih si l

bu

P eru ta

N 500 2©

e rut a

n/m

oc.erut a

n.w

ww//:

ptth

sdoht

em

a

d

b

e

c

f

Figure 1 | Coinjection of fC31 integrase with CMV-gfp reporter plasmid allows detection of GFP

expression in stage 46 Xenopus embryos. (a) Stage 39 Xenopus embryo injected with 100 pg insulated

CMV-gfp reporter plasmid. (b) Stage 46 Xenopus embryo injected with 100 pg insulated CMV-gfp reporter

plasmid (c) Stage 46 Xenopus embryo injected 5 pg of pEGFPB2. (d–f) Stage 46 Xenopus embryos injected

with 1 ng of integrase mRNA and 5 pg of pEGFPB2. Depending on the embryo we noted GFP signal in eyes

and neural tissue (d), dorsal muscle (e) or in the gill structures of the embryos (f). All embryos were

anesthetized, and in b–f the tails were curled around toward the head for photodocumentation using

a Zeiss Axioplan 2 microscope, except a Zeiss fluorescence dissecting microscope was used in a.

In every case the insert shows a brightfield image of the embryo. Scale bars, 500 mm; The scale bar in brepresents the scale in b–f.

Table 1 | Assay of GFP fluorescence as an indicator of integrase-mediated gene insertion

Treatmenta

Number of embryos

treated

(# trials/mothers)b

Number of embryos

surviving to

stage 46

Number of embryos with

GFP fluorescence

at stage 46

Average % of

embryos with GFP

fluorescence (range)

Untreated (noninjected) 1,194 (12/7) 779 0 0

DEPC H2O 916 (7/5) 398 0 0

Integrase mRNA (1 ng) 125 (1/1) 44 0 0

CMV-gfp (5 pg) 337 (3/3) 165 0 0

Insulated CMV-gfp plasmid (5 pg) 300 (4/3) 118 0 0

CMV-gfp plasmid (5 pg) + integrase mRNA (1 ng) 452 (3/3) 258 95 36.82 (12–49)

Insulated CMV-gfp plasmid (5 pg) + integrase mRNA (1 ng) 300 (4/3) 148 47 31.76 (29–44)

Insulated CMV-gfp plasmid (5 pg) + integrase mRNA (500 pg) 100 (1/1) 34 4 11.76

Insulated CMV-gfp plasmid (5 pg) + integrase mRNA (50 pg) 100 (1/1) 52 2 3.85

Insulated crystallin lens–gfp plasmid (5 pg) 200 (4/2) 69 0 0

Insulated crystallin lens–gfp plasmid (5 pg) + integrase mRNA (1 ng) 200 (4/2) 73 29 39.73 (37–44)

aInjection of nucleic acids or DEPC (diethylpyrocarbonate) water (DEPC H2O), as indicated. bStarting numbers of one-cell embryos and the number of trials are indicated. Mothers refers to the number of differentfemale frogs that produced eggs used in the collected data.

976 | VOL.2 NO.12 | DECEMBER 2005 | NATURE METHODS

ARTICLES

plasmid along with 1 ng of mRNA encoding fC31 integrase intosingle-cell Xenopus embryos; these embryos began to expressGFP ubiquitously at stage 14 and continued expressing it untilstage 46 when we stopped the experiment (Fig. 3a–c). Embryosinjected only with the reporter plasmid did not fluoresce atany stage (Fig. 3b).

When embryos injected with 1 ng of integrase mRNA andinsulated reporter plasmid were assayed at stage 46, approximately34–42% expressed GFP uniformly (Fig. 3c and Table 1). Weconfirmed integration by Southern blot analysis (an example ofintegration is shown in Fig. 2). Embryos injected with less integrasemRNA (500 pg and 50 pg, respectively) produced tadpoles with theexpected GFP expression pattern, just at lower frequencies(Table 1). We note that the quality of the integrase mRNA iscritical to the success of this technique. Though further testingis needed, the expression patterns of GFP are consistent withintegration generally occurring before the first cleavage. Most of

the remaining embryos did not express GFP, and an occasionalembryo expressed GFP in a manner similar to that seen foruninsulated plasmid, or only on one side.

Promoter specificity is retained in insulated integrated DNAWe next asked whether a tissue-specific promoter derived fromXenopus would behave in an appropriate manner. Using theinsulated GFP reporter plasmid we replaced the CMV promoterwith 551 bp of the Xenopus gamma-crystallin lens promoter21

(Fig. 3d–f). Unlike the global expression observed with the CMVpromoter, embryos injected with integrase mRNA and the crystallinlens gfp plasmid expressed GFP only in the lens of the eye (Fig. 3f).GFP expression in the lens of the eye was uniform (insert inFig. 3f). Integration efficiency ranged from 37 to 41% (Table 1)and we confirmed it by Southern blot analysis (Fig. 2).

DISCUSSIONWe show here both by expression analysis of GFP and bySouthern blot analysis that genomic integration of plasmidcan be obtained if the plasmid has an attB site and fC31 integraseis expressed. We have assayed 39 embryos by Southern blotanalysis. Although we commonly found that a single integrationevent occurred, we have seen nine different-sized bandscontaining the integrated plasmid on Southern blots with noconsistent indication of insertion site preference. The Southernblot band sizes are not consistent with multiple repeats orextrachromosomal plasmid, but vary in size depending on thenearest common restriction site in the Xenopus genome. Wehave not yet been able to assign particular integration sites tothe different patterns of tissue expression observed with uninsu-lated reporter plasmids, but further analyses of the sites ofintegration are underway. Although we do not know whether theintegrations using this technique will be transmitted in the germline, we have analyzed buccal cells derived from a juvenile froggrown from an embryo coinjected with integrase mRNA andpEGFPB2. We find that there is still an integrated copy of thereporter gene in the buccal cell DNA (data not shown), so we arecautiously optimistic.

p

uor

G g

n ih si l

bu

P eru ta

N 500 2©

e rut a

n/m

oc.erut a

n.w

ww//:

ptth

sdoht

em

Figure 3 | Coinjection of fC31 integrase mRNA

with insulated CMV-gfp or crystallin lens–gfp

reporter plasmid generated Xenopus embryos with

tissue-appropriate expression. (a–c) Stage 46

embryos assaying the expression of insulated CMV-

gfp reporter: noninjected embryo (a), an embryo

injected with the insulated CMV-gfp reporter

without coinjection of integrase mRNA (b), and an

embryo coinjected with insulated CMV-gfp reporter

and integrase mRNA (c). (d–f) Stage 42 embryos

assaying the expression of insulated crystallin

lens–gfp reporter: noninjected embryo (d), an

embryo injected with the insulated crystallin lens–

gfp reporter without coinjection of integrase mRNA

(e), and an embryo coinjected with insulated

crystallin lens–gfp reporter and integrase mRNA (f).Images a–f are of embryos viewed under

fluorescent light. In every case the insert shows a

brightfield image of the embryo. The two grayscale inserts in f represent a magnified view of eyes typical of embryos transgenic for crystallin lens–gfp. The left

image was taken using brightfield and the right using fluorescence through a 10� objective on a Zeiss Axioplan 2 microscope. Scale bars, 500 mm. The scale bar

in a represents scale in a–c; the scale bar in d represents scale in d–f.

a

d

b

e

c

f

12

C 1 2 3 4 5

6

4

2

Figure 2 | Southern blot analysis of treated embryos indicates insertion of

the CMV-gfp reporter plasmid into the embryonic genome. Markers, in Kbp,

are indicted on the left side of a composite of data from several Southern

blots. Lane ‘C’ is linearized pEGFPB2, lane 1 and 2 used DNA isolated from

the embryos in Figure 1d,e, respectively. Lane 3 and 4 are from embryos

transgenic for the insulated crystallin lens-gfp reporter, and lane 5 is

from an embryo transgenic for insulated CMV-gfp reporter.

NATURE METHODS | VOL.2 NO.12 | DECEMBER 2005 | 977

ARTICLES

The insulated GFP reporters render the expression of theintegrated gene independent of chromosome position effects. Inaddition, we have shown that tissue-specific gene expression can beachieved, as is demonstrated by the studies using crystallin lens-gfpreporter plasmid. The plasmids successfully used in this study werebetween 5 and 6.2 Kbp. We are now constructing larger plasmids totest the limits of the system. We have reason to be optimistic that wewill be able to use plasmids as large as 41 Kbp because that is thesize of the fC31 phage22.

The goal of this study was to determine whether the fC31integrase system could generate transgenic Xenopus laevis embryosat a high enough efficiency to be able to study the F0 generation oftransgenes. This goal has been met. We feel that this technique hasthe potential to become a valuable tool for transgenic studies onXenopus and other systems.

METHODSDNA constructs. M. Calos (Stanford University) provided plas-mids pET11phiC31poly(A)23, used to generate the integrasemRNA in vitro, and reporter plasmid pEGFPB2 (ref. 23), whichcontains a CMV-driven gfp and an attB site. G. Felsenfeld(National Institute of Diabetes and Digestive and Kidney Diseases,National Institutes of Health) provided a plasmid encoding theduplicated 250-bp 5¢ HS4 insulator. We purchased restrictionenzymes from either Promega or New England Biolabs. We PCRamplified the 5¢ HS4 double insulator core fragment with eitherBmgBI- or PciI-digestible ends using the primer pairs 5¢-CCACGTCCGGGTACCGAGTTGGCGCG-3¢ and 5¢-CGACGTGTCGATGAATTCGAGTTGGC-3¢, or 5¢-CACATGTCCGGGTACCGAGTTGGCGCG-3¢ and 5¢-CACATGTCGATGAATTCGAGTTGGC-3¢. Wecloned the 550-bp PCR fragments into TOPO 4 (Invitrogen)and sequenced the resulting plasmid to ensure insulator sequencefidelity. We ligated the BmgBI-HS4 double insulator into theBmgBI site between the attB site and the 5¢ end of the CMV-gfpreporter and the PciI-HS4 double insulator ligated into the PciIsite 3¢ of the CMV-gfp reporter.

To create a plasmid containing the Xenopus minimal gamma-crystallin lens promoter21 driving GFP expression, we removed theCMV promoter from insulated CMV-gfp reporter plasmid bydigesting with AseI and AgeI. P. Krieg (University of Arizona)provided the plasmid with the full Xenopus gamma-crystallin lenspromoter. We PCR-amplified a 551-bp region of the gamma-crystallin lens promoter, digested the ends with AseI and AgeI andinserted into the reporter plasmid.

Integrase mRNA synthesis. We linearized the plasmid pET11-phiC31poly(A) with EcoRI. We synthesized mRNA in vitro usingthe T7 mMessage Machine Kit following the manufacturer’sinstructions (Ambion). We used a total of 1 mg of digested DNAas a template for mRNA production.

Xenopus injections. We obtained X. laevis males and females fromeither Xenopus I or Nasco. We obtained eggs from hormonallyinduced females and fertilized as previously described24. Weinjected single-cell embryos with either water, 5 pg of reporterplasmid, or 5 pg of reporter plasmid and 1 ng, 500 pg or 50 pg offC31 integrase mRNA in a total volume of 10 nl as previouslydescribed25. Embryonic developmental stages were identified usingcriteria as described26. All procedures using animals were reviewed

and approved by the University of Iowa Institutional Animal Careand Use Committee.

Embryo analysis. We analyzed embryos for GFP fluorescencethrough embryonic stage 46 with a fluorescence dissectingmicroscope (Stemi SV 11; Zeiss) or for higher magnificationswith a Zeiss Axioplan 2 microscope. Light passed through a GFPfilter with an emission filter limit of 535 nm. Color brightfieldand GFP embryo pictures were taken with a Diagnostic Instru-ments Spot Camera. Black-and-white images were taken with aZeiss Axiocam. Images were oriented and processed with AdobePhotoshop 7.0.

Genomic DNA extraction and Southern blotting. We extractedDNA from stage 46 tadpoles according to the manufacturer’sinstructions (Qiagen DNeasy Tissue Kit). We digested genomicDNA with HindIII (noninsulated integrant experiments) or Bam-HI (insulated integrant experiments) overnight at 37 1C, separatedon a 1% agarose gel, and transferred onto positively chargedHybond membrane (Amersham). For the noninsulated integrantDNA analysis, we linearized pEGFPB2 with HindIII and thenlabeled with [32P]dCTP using the Rediprime II Random PrimeLabeling System (Amersham) following the manufacturer’sinstructions. For the insulated integrant DNA analysis, we isolateda 790-bp gfp fragment from pEGFPB2 by digestion with AgeI andBamHI. We then labeled the fragment with [32P]dCTP usingRediprime II Random Prime Labeling System following themanufacturer’s instructions. We hybridized the filters with thelabeled probe and washed them at 68 1C to a stringency of 0.1�SSC, 0.1% SDS.

ACKNOWLEDGMENTSWe thank M. Calos, G. Felsenfeld and P. Krieg for their generosity in providingplasmids used in this study. We thank H. Bartlett, J. Dagle, S. Kolker, C. Fett,E. Hornick and E. Stokasimov for helpful discussions. Funding to D.L.W. fromthe National Institutes of Health (GM069944 and DC007481) supported thiswork. B.G.A. is a student in the Medical Scientist Training Program at theRoy J. and Lucille A. Carver College of Medicine, University of Iowa.

COMPETING INTERESTS STATEMENTThe authors declare competing financial interests (see the Nature Methods websitefor details).

Published online at http://www.nature.com/naturemethods/Reprints and permissions information is available online athttp://npg.nature.com/reprintsandpermissions/

1. Kroll, K.L. & Amaya, E. Transgenic Xenopus embryos from sperm nucleartransplantations reveal FGF signaling requirements during gastrulation.Development 122, 3173–3183 (1996).

2. Sparrow, D.B., Latinkic, B. & Mohun, T.J. A simplified method of generatingtransgenic Xenopus. Nucleic Acids Res. 28, E12 (2000).

3. Thorpe, H.M., Wilson, S.E. & Smith, M.C. Control of directionality in the site-specific recombination system of the Streptomyces phage fC31. Mol. Microbiol.38, 232–241 (2000).

4. Thorpe, H.M. & Smith, M.C. In vitro site-specific integration of bacteriophage DNAcatalyzed by a recombinase of the resolvase/invertase family. Proc. Natl. Acad.Sci. USA 95, 5505–5510 (1998).

5. Lutz, K.A., Corneille, S., Azhagiri, A.K., Svab, Z. & Maliga, P. A novel approach toplastid transformation utilizes the fC31 phage integrase. Plant J. 37, 906–913(2004).

6. Groth, A.C., Olivares, E.C., Thyagarajan, B. & Calos, M.P. A phage integrase directsefficient site-specific integration in human cells. Proc. Natl. Acad. Sci. USA 97,5995–6000 (2000).

7. Chalberg, T.W., Genise, H.L., Vollrath, D. & Calos, M.P. fC31 integrase confersgenomic integration and long-term transgene expression in rat retina. Invest.Ophthalmol. Vis. Sci. 46, 2140–2146 (2005).

p

uor

G g

n ih si l

bu

P eru ta

N 500 2©

e rut a

n/m

oc.erut a

n.w

ww//:

ptth

sdoht

em

978 | VOL.2 NO.12 | DECEMBER 2005 | NATURE METHODS

ARTICLES

8. Thyagarajan, B., Olivares, E.C., Hollis, R.P., Ginsburg, D.S. & Calos, M.P. Site-specific genomic integration in mammalian cells mediated by phage fC31integrase. Mol. Cell. Biol. 21, 3926–3934 (2001).

9. Thomason, L.C., Calendar, R. & Ow, D.W. Gene insertion and replacement inSchizosaccharomyces pombe mediated by the Streptomyces bacteriophage fC31site-specific recombination system. Mol. Genet. Genomics 265, 1031–1038(2001).

10. Olivares, E.C. et al. Site-specific genomic integration produces therapeutic FactorIX levels in mice. Nat. Biotechnol. 20, 1124–1128 (2002).

11. Held, P.K. et al. In vivo correction of murine hereditary tyrosinemia type I byfC31integrase–mediated gene delivery. Mol. Ther. 11, 399–408 (2005).

12. Belteki, G., Gertsenstein, M., Ow, D.W. & Nagy, A. Site-specific cassette exchangeand germline transmission with mouse ES cells expressing fC31 integrase. Nat.Biotechnol. 21, 321–324 (2003).

13. Groth, A.C., Fish, M., Nusse, R. & Calos, M.P. Construction of transgenic Drosophilaby using the site-specific integrase from phage fC31. Genetics 166, 1775–1782(2004).

14. Vize, P.D., Melton, D.A., Hemmati-Brivanlou, A. & Harland, R.M. Assays forgene function in developing Xenopus embryos. Methods Cell Biol. 36, 367–387(1991).

15. Newport, J. & Kirschner, M. A major developmental transition in early Xenopusembryos: II. Control of the onset of transcription. Cell 30, 687–696 (1982).

16. Kuhn, E.J. & Geyer, P.K. Genomic insulators: connecting properties to mechanism.Curr. Opin. Cell Biol. 15, 259–265 (2003).

17. West, A.G., Gaszner, M. & Felsenfeld, G. Insulators: many functions, manymechanisms. Genes Dev. 16, 271–288 (2002).

18. Bell, A.C. & Felsenfeld, G. Stopped at the border: boundaries and insulators.Curr. Opin. Genet. Dev. 9, 191–198 (1999).

19. Chung, J.H., Whiteley, M. & Felsenfeld, G. A 5¢ element of the chicken b-globindomain serves as an insulator in human erythroid cells and protects againstposition effect in Drosophila. Cell 74, 505–514 (1993).

20. Pikaart, M.J., Recillas-Targa, F. & Felsenfeld, G. Loss of transcriptional activity ofa transgene is accompanied by DNA methylation and histone deacetylation and isprevented by insulators. Genes Dev. 12, 2852–2862 (1998).

21. Offield, M.F., Hirsch, N. & Grainger, R.M. The development of Xenopus tropicalistransgenic lines and their use in studying lens developmental timing in livingembryos. Development 127, 1789–1797 (2000).

22. Smith, M.C., Burns, R.N., Wilson, S.E. & Gregory, M.A. The completegenome sequence of the Streptomyces temperate phage straight fC31:evolutionary relationships to other viruses. Nucleic Acids Res. 27, 2145–2155(1999).

23. Hollis, R.P. et al. Phage integrases for the construction and manipulation oftransgenic mammals. Reprod. Biol. Endocrinol. 1, 79–89 (2003).

24. Rebagliati, M.R., Weeks, D.L., Harvey, R.P. & Melton, D.A. Identification andcloning of localized maternal RNAs from Xenopus eggs. Cell 42, 769–777(1985).

25. Dagle, J.M., Walder, J.A. & Weeks, D.L. Targeted degradation of mRNA in Xenopusoocytes and embryos directed by modified oligonucleotides: studies of An2 andcyclin in embryogenesis. Nucleic Acids Res. 18, 4751–4757 (1990).

26. Nieuwkoop, P.D. & Faber, J. Normal table of Xenopus laevis (Daudin): asystematical and chronological survey of the development from the fertilized eggtill the end of metamorphosis. (Garland Pub., New York, 1994).

p

uor

G g

n ih si l

bu

P eru ta

N 500 2©

e rut a

n/m

oc.erut a

n.w

ww//:

ptth

sdoht

em

NATURE METHODS | VOL.2 NO.12 | DECEMBER 2005 | 979

ARTICLES