Ability of the hydrophobic FGF and basic TAT peptides to promote cellular uptake of recombinant Cre recombinase: A tool for efficient genetic engineering of mammalian genomes

Michael Peitz, Kurt Pfannkuche, Klaus Rajewsky*, and Frank Edenhofer†

Institute for Genetics, University of Cologne, Weyertal 121, 50931 Cologne, Germany

Contributed by Klaus Rajewsky, February 5, 2002 Conditional mutagenesis is a powerful tool to analyze gene func- sive mouse breeding causing the experiments to be time con- tions in mammalian cells. The site-specific recombinase Cre can be suming and costly. The leakiness of the system represents a used to recombine loxP-modified alleles under temporal and spa- critical factor because a Cre recombinase that is undesirably tial control. However, the efficient delivery of biologically active active before induction often leads to unwanted side effects such Cre recombinase to living cells represents a limiting factor. In this as mosaic recombination and͞or selection of recombined or study we compared the potential of a hydrophobic peptide mod- nonrecombined cells both in vivo and in vitro (9, 17, 18). ified from Kaposi fibroblast growth factor with a basic peptide Moreover, the widely used inducers IFN, hydroxy-tamoxifen, derived from HIV-TAT to promote cellular uptake of recombinant and doxycycline are known to display toxic side effects (19, 20) Cre. We present the production and characterization of a Cre and͞or induce also unwanted physiological effects that may protein that enters mammalian cells and subsequently performs interfere with the experimental phenotype of the conditional recombination with high efficiency in a time- and concentration- mutation to be analyzed (21, 22). In cultured cells, Cre-mediated dependent manner. Histidine-tagged Cre recombinase transduced recombination is limited by transfection efficiencies and putative inefficiently unless fused to a nuclear localization signal (NLS). toxicity of the protein (13, 23, 24). Thus, traditional delivery of Fusion of NLS-Cre to the fibroblast growth factor transduction Cre—either by Cre transgenics, viral vectors, or transfection— peptide did not improve the transducibility, whereas fusion with represents a limiting step of conditional mutagenesis employing the TAT peptide significantly enhanced cellular uptake and subse- ͞ quent recombination. More than 95% recombination efficiency in Cre loxP technology. fibroblast cells, as well as murine embryonic stem cells, was Protein transduction is a recently developed method to intro- achieved after His-TAT-NLS-Cre transduction. Efficient recombina- duce biologically active proteins directly into mammalian cells tion could also be obtained in primary splenocytes ex vivo.We with high efficiency (for review, see refs. 25 and 26). Recombi- expect that application of His-TAT-NLS-Cre, which can be produced nant technology has been used to modify biophysical properties readily in large quantities from a bacterial source, will expand our of proteins, particularly with respect to their cell permeability, by abilities to manipulate mammalian genomes. employing so-called protein transduction domains (PTDs) (27– 29). It has been demonstrated that a basic 11-aa peptide, derived from HIV-TAT, renders ␤-galactosidase (␤-gal) into a cell- onditional mutagenesis in mammalian cells has become an GENETICS permeable form. ␤-gal activity can be detected in organs such as important means for the analysis of gene function in vivo (for C liver, kidney, lung, brain, and spleen of mice after i.p. injection review, see refs. 1–3). Site-specific recombinases such as the ␤ bacteriophage P1 recombinase Cre have been used to gain of TAT- -gal (28). Besides the basic TAT peptide, other pep- control over the mutation in a spatial (4–6) and͞or temporal tides or protein (fragments) also have been reported to enhance manner (7, 8). An increasing number of studies have demon- cellular uptake of proteins, such as a hydrophobic peptide strated the efficacy of Cre-mediated conditional mutagenesis in modified from Kaposi fibroblast growth factor (FGF) (29, 30). mice and cell lines (9–12). Usually a mouse or cell line is Recently it has been reported that a fusion protein of a nuclear generated, in which an essential part of the gene of interest is localization signal (NLS), Cre recombinase, and a FGF peptide flanked by two loxP sites. The loxP sites represent Cre recom- displays cell permeability. After direct application of the recom- binase recognition sites and can be used to delete the respective binant protein to cultured cells ofaTlymphocyte line containing gene segment upon Cre recombination, resulting in a condi- a loxP-modified substrate, Cre-mediated recombination was tioned inactivation or mutation of the gene of interest. To gain observed in Ϸ80% of the cells (31). However, it remained temporal control over this mutation event two different ap- unclear from this study whether the FGF peptide is essential for proaches have been applied: (i) Cre is delivered into cultured transduction or even contributes significantly to the cellular cells either by transfection (13) or adenoviral infection (14, 15); uptake of Cre. (ii) Cre recombinase activity is induced by application of an exogenous inducer. Induction can be carried out either at the transcriptional level [e.g., Mx-Cre (7) or tetracycline-controlled Abbreviations: aa, amino acid(s); ␤-gal, ␤-galactosidase; ES, embryonic stem cell(s); FGF, fibroblast growth factor; NLS, nuclear localization signal; PTD, protein transduction Cre expression (16)] or posttranslational level employing fusion domain. proteins of Cre with mutated ligand-binding domains of steroid *Present address: Center for Blood Research, Harvard Medical School, 200 Longwood receptors (8). Although a number of studies demonstrated the Avenue, Boston, MA 02115. efficacy of Cre-mediated inducible mutagenesis, still the system †To whom reprint requests should be addressed. E-mail: [email protected]. could profit substantially from technical advance concerning The publication costs of this article were defrayed in part by page charge payment. This three major aspects: leakiness of the system before induction, article must therefore be hereby marked “advertisement” in accordance with 18 U.S.C. efficiency of induced recombination, and requirement of exten- §1734 solely to indicate this fact.

www.pnas.org͞cgi͞doi͞10.1073͞pnas.032068699 PNAS ͉ April 2, 2002 ͉ vol. 99 ͉ no. 7 ͉ 4489–4494 Downloaded by guest on October 2, 2021 Our study was aimed at evaluating the actual potency of two Overexpression was induced by adding IPTG to a final concen- prominent PTDs, namely FGF and TAT peptides, to promote tration of 0.5 mM and an additional incubation of 3–4hat37°C. the translocation of biologically active Cre across the plasma Cells were harvested by centrifugation and frozen in dry ice͞ membrane of mammalian cells. We generated expression vectors ethanol and stored at Ϫ20°C. Cell pellets were thawed by ͞ encoding Cre recombinase fused to FGF and TAT peptides, resuspending in lysis buffer [100 mM NaH2PO4 10 mM Tris (pH respectively. The potentials of the recombinant proteins to 8.0)͞300 mM NaCl͞10 mM imidazole] at 5 ml per gram wet transduce and subsequently recombine loxP-flanked targets in weight. Cleared lysate was obtained after incubation with 2 mammalian cells were compared side by side with Cre lacking mg͞ml lysozyme (Sigma) and benzonase (Novagen) and centrif- any particular PTD. ugation for 25 min at 30,000 ϫ g at 4°C. One milliliter of 50% Ni-NTA slurry (Qiagen) was added to 5 ml of cleared lysate and Materials and Methods mixed gently by shaking at 4°C for 1 h. The slurry was packed into Plasmid Constructions. A TAT-encoding fragment was generated a column and washed with 10 bed volumes [100 mM ͞ ͞ ͞ by PCR, using primers NPTatU and NeuUS. This fragment was NaH2PO4 10 mM Tris (pH 8.0) 300 mM NaCl 20 mM imida- ͞ cloned into the pTriEx-1 vector (Novagen) via NcoI and XhoI zole]. Recombinant protein was eluted [100 mM NaH2PO4 10 restriction sites, resulting in pTriEx-TAT-U-H (F.E. and K.R., mM Tris (pH 8.0)͞300 mM NaCl͞250 mM imidazole] and either unpublished data). This vector encodes, in addition to TAT, a dialyzed against appropriate media (see below) for immediate second protein fragment designated as ‘‘U,’’ which was not used use or frozen at Ϫ20°C for storage up to 8 weeks without in this study. The corresponding DNA fragment was deleted by significant loss of activity. Protein concentrations were measured digestion with SpeI and subsequent religation, resulting in using Bradford reagent (Sigma) and͞or Warburg method. Cre pTriEx-TAT-H. The TAT-encoding fragment was removed, activities were determined by a cell-free assay as described (31, resulting in pTriEx-H by PstI digestion. To obtain pTriEx-CH,a 33), except for our use of 250 ng of substrate DNA (33). For Cre-encoding PCR product was generated from the template quantification reactions were transformed into E. coli and pPGK-Cre-bpA (32) by using primers 5H3cre and 3X1cre. This colonies counted. A commercially available Cre protein (New fragment was subsequently cloned into the pTriEx-H by using England Biolabs) was used as a standard. HindIII and XhoI sites. To generate pTriEx-NCH, a NLS-Cre- encoding fragment was amplified using primers 5H3NLScre and SDS͞PAGE and Immunoblot Analysis. Samples were separated by 3X1cre and cloned as above. To construct pTriEx-FNCH, an- SDS͞PAGE (10%), transferred to nitrocellulose membranes nealed oligonucleotides 5PstFGF and 3PstFGF were cloned into (Amersham Pharmacia), and probed with anti-Cre (Novagen) or the PstI restriction site of pTriEx-NCH. To generate pTriEx- anti-Penta-His (Qiagen) antiserum. Anti-Cre was incubated HTNC, a NLS-Cre-STOP encoding PCR product was generated 1:5,000 in PBS containing 5% dry milk, and anti-Penta-His was using primers 5H3NLScre and 3creStopXho and subsequently used 1:2,000 in PBS containing 3% BSA. Blocking was carried cloned into pTriEx-TAT-H via HindIII and XhoI restriction out in corresponding buffers without antibody for1hatroom sites; the His tag was introduced by cloning annealed oligonu- temperature. As secondary antibodies anti-rabbit-POD (Vector cleotides NcoHis5 and NcoHis3 into the NcoI site. pTriEx- Laboratories) or anti-mouse-POD conjugates (Amersham Phar- HNCF was constructed by cloning annealed oligonucleotides macia) were used both at a 1:10,000 dilution in PBS containing NcoHis5 and NcoHis3, as well as 5XhoFGF and 3XhoFGF, into 5% milk powder. Blots were washed using PBST (PBS contain- the NcoI and XhoI restriction sites, respectively, of pTriEx-NCH. ing 0.5% Tween 20). Detection was carried out using chemilu- From pTriEx-HTNC a 700-bp fragment was prepared using minescence (ECL, Amersham Pharmacia). BamHI͞XhoI restriction sites and cloned into pTriEx-HNCF, resulting in pTriEx-HNC. All PCR-generated fragments were Cell Culture, and Transfection of and Transduction into CV Fibroblasts. confirmed by sequencing. NPTatU (5Ј-CTTGGCCATGGG- CV1–5B (8) cells were cultured in DMEM containing glutamax CGCTGCAGGTCGCAAGAAACGTCGCCAACGTCGCC- and 10% FCS and 100 units͞ml penicillin and 0.1 mg͞ml GTCCGCCTGCAGGCACTAGTCAGATTTTCGTCAA- streptomycin. Transfection experiments were carried out in GACTTTGAC-3Ј); NeuUS (5Ј-CCATTACTAGTTGCGCCC- 24-well plates by using 1 ␮g of plasmid as described (32). For Cre ATACCACCACGTAGCCTTAGCACAAGATG-3Ј); 5H3cre transduction experiments, 2 ϫ 106 cells were plated on a 24-well (5Ј-CGTCCAAGCTTGTCCAATTTACTGACCGTACACC- plate and grown for 24 h, or until 90% confluency. Cre protein 3Ј); 3X1cre (5Ј-CTGAACTCGAGACCATCGCCATCTTCCA- was dialyzed against a 1:1 mixture of DMEM and PBS containing GCAGGC-3Ј); 5H3NLScre (5Ј-CATGGAAGCTTGAAGAA- 0.1% pluronic F-68 (Sigma) and sterilized by filtration, using a GAAGAGGAAGGTGTC-3Ј); 5PstFGF(5Ј-GTATTACTTC- 0.2-␮m filter disk (Millipore). Cells were incubated with Cre CGGTTCTGTTAGCGGCACCGGGTGCA-3Ј); 3PstFGF containing medium at indicated concentrations for indicated (5Ј-CCCGGTGCCGCTAACAGAACCGGAAGTAATA- times. After transduction cells were washed using PBS and CTGCA-3Ј); 3creStopXho (5Ј-CTAATCTCGAGCTAATCG- cultured 60 h in normal medium. For determination of ␤-gal CCATCTTCCAGCAG-3Ј); NcoHis5 (5Ј-CATGGGCCAT- activity, cells were fixed and stained as described (8). For CACCATCACCATCACGG-3Ј); NcoHis3 (5Ј-CATGCCGTG- Southern blot analysis genomic DNA was extracted from cells by ATGGTGATGGTGATGGCC-3Ј); 5XhoFGF (5Ј-TCGAG- proteinase K digestion and isopropanol precipitation. DNA was GGTGCAGCTGTATTACTTCCGGTTCTGTTAGCGG- digested overnight by using EcoRI restriction enzyme and CACCGTGAT-3Ј); 3XhoFGF (5Ј-TCGAATCACGGTGCCG- separated on an agarose gel. DNA was immobilized on a Hybond CTAACAGAACCGGAAGTAATACAGCTGCACCC-3Ј). nitrocellulose membrane (Amersham Pharmacia) and probed with a 32P-labeled lacZ fragment. Quantification of bands was Expression and Purification of Recombinant Proteins. pTriEx plas- carried out by phosphor imaging technique (FUJIX BAS 1000; mids were used to transform Escherichia coli strain TUNER TINA 2.09 software, Raytest, Straubenhardt, Germany). (DE3)pLacI (Novagen), allowing isopropyl ␤-D-thiogalactoside (IPTG)-inducible expression of His-tagged proteins. An en- Immunofluorescence. CV1–5B Fibroblasts were grown on 24-well riched medium of LB ϩ 1% Glucose containing 50 ␮g͞ml tissue culture plates until 90% confluency. After a 4-h incubation carbenicillin and 34 ␮g͞ml chloramphenicol was used to inoc- in Cre fusion protein containing DMEM͞PBS (1:1), cells were ulate overnight cultures. Usually, 500-ml cultures of LB con- washed twice with PBS, trypsinized, and plated on coverslips. taining 100 ␮g͞ml ampicillin and 34 ␮g͞ml chloramphenicol After 16 h in culture, cells were fixed with cold methanol Ϫ were inoculated 1:50 and grown at 37°C until an OD600 of 0.7. ( 20°C). The samples were washed three times with PBS and

4490 ͉ www.pnas.org͞cgi͞doi͞10.1073͞pnas.032068699 Peitz et al. Downloaded by guest on October 2, 2021 blocked for 30 min. Blocking and antibody incubations were carried out in PBS containing 1% dry milk. Antibody detection was performed using polyclonal rabbit anti-Cre antibody (Babco, Richmond, VA, 1:1,000 dilution; 90 min) and FITC-anti-rabbit IgG (Sigma, 1:50 dilution; 60 min). The nuclei were stained by adding 0.01 mg͞ml 4,6-di-amino-2-phenylindol (DAPI) in PBS before mounting on glass slides. Cells were observed in an epifluorescence microscope (Zeiss Axiophot) and documented on color slide film. For reproduction images were scanned directly from slides.

Cre Transduction into Mouse Embryonic Stem Cells. ES cell culture was carried out as described (13). Single-cell suspensions of subconfluent plates were prepared by trypsinization. Cells were washed in PBS and resuspended in Cre containing medium. Cre protein was dialyzed against a 1:1 mixture of DMEM and PBS and sterilized by filtration. After resuspension in Cre-containing Fig. 1. Design of expression cassettes and purification of Cre fusion proteins. medium, ES cells were plated on appropriate plates containing (A) Six expression constructs were generated. All constructs encode Cre re- mitomycin C treated embryonic feeder cells. In a typical exper- combinase and a His-tag as represented by white boxes. Black boxes represent iment 2.5 ϫ 105 cells were used in a volume of 0.5 ml plated in PTDs (aa sequence): FGF (AAVLLPVLLAAP) and TAT (GRKKRRQRRR). The gray a well of a 24-well plate. After transduction, cells were washed box represents a NLS (KKKRKV) derived from SV40. Abbreviations of the ͞ and cultured at least 3 days in normal ES medium. Genomic constructs are given on the left. (B) SDS PAGE analysis of the purification of HTNC employing Ni(II)-affinity chromatography. CL, cleared lysate; FT, flow DNA was extracted and used for Southern blot analysis as through; W, wash fraction, E, eluted fraction; M, marker. (C) Comparison of described above, except the use of EcoRV digestion and a purified fusion proteins on a Coomassie stained gel. (D) Immunoblot analysis ROSA26-specific probe. of purified Cre fusion proteins employing an anti-Cre antibody.

Ex Vivo Cre Transduction into Mouse Splenocytes. For transduction into primary splenocytes we used pol␤-flox (6) mice. Freshly whereas the other proteins showed expression levels of about 10 prepared splenic cells were washed in PBS and resuspended in mg͞l. All tested proteins were detected by antibodies directed Cre-containing medium. Cre protein was dialyzed first against against Cre (Fig. 1D) and the His-tag (data not shown), respec- PBS, then against protein-free medium (HyClone, catalog no. tively, and displayed the expected size of Ϸ41 kDa. All proteins, SH 30349.01), and sterilized by filtration. Cells (5 ϫ 106) were except HTNC, are soluble in physiological buffers such as PBS resuspended in 1 ml of Cre-containing medium and plated in a up to a concentration of Ϸ12 ␮M(Ϸ0.5 mg͞ml). HTNC can be well of a 24-well plate. After transduction cells were harvested, concentrated up to 16 ␮M(Ϸ0.65 mg͞ml). resuspended in normal B cell medium (DMEM containing 10% ϫ FCS, 1 mM sodium pyruvate, 2 mM L-glutamine, 1 nonessen- FGF Peptide Is Dispensable for Transduction of His-NLS-Cre. To assess ␤ tial aa, 0.1 mM 2- -mercaptoethanol; GIBCO) and plated in a the capacity of recombinant Cre fusion proteins to invade well of a 24-well plate. After incubation for 16 h, genomic DNA eukaryotic cells and subsequently perform recombination of a was extracted and used for Southern blot analysis as described ␤ loxP-flanked substrate, we used the well established reporter cell above, except for our use BamHI digestion and a pol -specific line CV1–5B (8). This fibroblast cell line contains a single copy probe (6). Splenocytes were stained with fluorochrome- of a stably integrated reporter construct. Cre activity is moni- conjugated monoclonal antibodies [FITC; phycoerythrin (PE)].

tored by deletion of a loxP-flanked stop segment resulting in the GENETICS Antibodies against CD19 (PharMingen) and Thy1.2(CFO1; in activation of a lacZ gene (Fig. 2A). Reporter cells were incubated house) were used. For determination of cell viability propidium for 16 h with medium containing 2.5 ␮M(Ϸ100 ␮g͞ml) of iodide (PI) staining was performed. FACS analysis was carried HNCF, HNC, NCH, CH, and HTNC. Indeed, ␤-gal activity was out using a FACScan (Becton Dickinson); for sorting a FAC- detected in cells incubated with fusion proteins containing a Starϩ was used (Becton Dickinson). PTD, namely HNCF and HTNC, indicative of cellular uptake Results and Discussion and subsequent recombination of the reporter gene. Unexpect- Design of Expression Vectors, and Expression and Purification of edly, cells incubated with HNCF displayed only a low percentage Recombinant Cre Fusion Proteins. To evaluate the potency of the of blue cells, whereas incubation of the control proteins lacking PTDs FGF and TAT to promote cellular uptake of Cre recom- the PTD, namely HNC and NCH, resulted in a significantly binase into mammalian cells we generated expression vectors higher percentage of stained cells (Fig. 2B). A dose–response encoding FNCH (FGF-NLS-Cre-His), HNCF (His-NLS-Cre- analysis revealed that transduction of HNCF and HNC, as well FGF), and HTNC (His-TAT-NLS-Cre) (Fig. 1A). All constructs as NCH and subsequent recombination, is strictly dependent on include a His-tag for single-step purification of recombinant concentration and reaches a level of saturation of more than 90% Ϸ ␮ proteins from a bacterial source. As controls we also generated efficiency at 5 M (Fig. 2C). Half-maximal recombination is ␮ Ϸ Cre expression constructs lacking any particular transduction achieved for NCH and HNC between 1 and 2 M, whereas 3–4 ␮ domain, but carrying the His-tag either at the N terminus M HNCF are needed for the same level of recombination. To (His-NLS-Cre, abbreviated HNC) or C terminus (NLS-Cre-His, demonstrate Cre-mediated recombination not only by activation NCH) of Cre. Finally we generated the expression construct of a reporter gene but also at the DNA level, we performed Cre-His (CH) which lacks NLS and encodes a C-terminally Southern blot analysis of Cre-transduced CV1–5B reporter cells. His-tagged Cre recombinase (Fig. 1A). Analysis of expression We compared HNCF side by side with HNC to directly assess the revealed that one recombinant protein, namely FNCH, displayed function of the FGF peptide. A time course experiment revealed poor solubility under native conditions (data not shown), that incubation of the cells with 12 ␮M HNCF for1hresults in whereas the other recombinant proteins could be highly enriched about 15% recombination efficiency. Half-maximum recombi- under native conditions (Fig. 1 B and C). HTNC displayed nation is achieved after 4 h of incubation (Fig. 2D). Application reproducibly the highest level of expression (Ϸ30 mg͞l culture), of 12 ␮M HNC protein is more efficient; half-maximal recom-

Peitz et al. PNAS ͉ April 2, 2002 ͉ vol. 99 ͉ no. 7 ͉ 4491 Downloaded by guest on October 2, 2021 Fig. 3. (A) Comparison of Cre recombinase activities of HNCF, NCH, and CH by using a cell-free activity assay. Eighty nanograms of either HNCF, NCH, or CH were used to recircularize a linearized substrate vector as described (31, 33). The specific activity of a commercially available control was set to 100% (Cϩ). Further controls were: 80 ng heat-inactivated Cre protein (CϪ) and no protein (Ϫ). Vertical lines represent ranges of values. (B) Quantitative analysis of recombinase activities of NCH and CH after transfection. CV1–5B cells were Fig. 2. Transduction of Cre fusion proteins into a fibroblast Cre reporter cell transfected with corresponding pTriEx expression vectors or a mock control ␤ line. (A) Schematic representation of the reporter gene construct of the and stained for -gal activity. Given are mean values of three transfections. fibroblast reporter cell line CV1–5B (8). Cre deletes a loxP-flanked segment, n.d., not detectable. (C) Comparison of the time dependency of HNC and HTNC ␮ ␮ thereby activating a lacZ gene. (B) CV1–5B cells were incubated for 16 h in transduction. CV1–5B cells were incubated with either 10 M HNC or 2 M medium containing 2.5 ␮M(Ϸ100 ng͞␮l) purified Cre fusion proteins as HTNC between 30 min and 16 h, as indicated. After incubation, cells were ␤ indicated; as a control, cells were incubated with normal medium. After washed, incubated with normal medium, stained for -gal activity, and incubation cells were washed, incubated with normal medium, fixed, and counted. Percentages of blue cells are given. stained for ␤-gal activity. (C) Quantification of ␤-gal activities of reporter cells incubated for 16 h with various concentrations of Cre fusion proteins as indicated. Percentages of blue cells are given. (D and E) Southern blot analysis transduction using CH did not increase over 5%, whereas NCH of transduction and subsequent recombination. (D) Time kinetics of HNCF and and HNC, the corresponding constructs carrying a NLS, resulted HNC transduction. Reporter cells were incubated with either 12 ␮M HNCF in efficiencies of more than 90% (Fig. 2C). There are three or HNC for indicated periods of time. (E) Concentration dependency of HNCF possible explanations for this observation: (i) the enzymatic and HNC transduction. Reporter cells were incubated for 16 h with increasing activity of HNC and NCH, respectively, is higher than that of concentrations of either HNCF or HNC, as indicated. Calculated deletion efficiencies in % are given at bottom (⌬%). flox, loxP-flanked gene; ⌬, deleted CH; this is unlikely because CH and NCH displayed comparable gene. specific activities in a cell-free assay (Fig. 3A). (ii) Nuclear translocation of CH is inefficient because of the lack of the NLS resulting in a low recombination efficiency; this explanation is bination is achieved already after 1–2 h, and 8 h are sufficient to also unlikely because recombination efficiencies of NCH and CH perform recombination in more than 95% of the cells. To assess after transfection were comparable (Fig. 3B), indicating that the concentration dependency of transduction, we incubated nuclear translocation is not a limiting factor. This observation is reporter cells in serial dilutions of HNCF and HNC down to 0.4 consistent with the previously published finding that addition of ␮M. Southern blot analysis demonstrated that applications of up a NLS to Cre confers no increase of recombination (34). (iii) The to 0.8 ␮M of both either HNCF or HNC did not result in any NLS itself might function as a PTD or at least contributes to the detectable recombination (Fig. 2E). However, application of 1.5 cellular uptake of Cre fusion proteins. To investigate the func- ␮M HNC resulted in substantial recombination (Ϸ30%), tion of NLS during cellular uptake independently of the recom- whereas at the same concentration HNCF is almost inactive bination event we performed an immunohistochemical analysis. (Ϸ2%). Recombination in almost every cell is achieved by using Localization of CH and NCH after transduction was monitored 6 ␮M HNC and 12 ␮M HNCF, respectively (Fig. 2E). Recently by fluorescence microscopy (Fig. 4). NCH was detectable in it has been reported that a cell permeable Cre, namely His- most of the cells both in the cytoplasm and nucleus, whereas in NLS-Cre-MTS, which is almost identical to the HNCF protein the case of the CH-treated cells only a low percentage was found used in our study, can enter eukaryotic cells and perform to be Cre-positive. Moreover, in the latter case the Cre-specific recombination (31). However, the actual ability of FGF to signals were mainly concentrated in extracellular spot-like struc- promote this process was not investigated. Our comparative tures, whereas NCH staining was more diffuse, dispersing over analysis of HNC and HNCF demonstrated that transduction of the whole cell body. His-NLS-Cre and subsequent recombination is independent of the hydrophobic FGF peptide. Therefore, the FGF peptide is TAT Peptide Enhances Transduction of His-NLS-Cre. The comparative dispensable for efficient transduction of His-NLS-Cre. More- analysis of all proteins analyzed in this study demonstrated that over, the analysis of the time and concentration dependency of the TAT peptide containing HTNC is the most effective protein HNCF as compared with HNC demonstrated that application of in transduction and subsequent recombination. One can achieve a NLS-Cre fusion protein lacking the transduction peptide FGF recombination in more than 90% of fibroblast reporter cells with is more effective. only 0.5 ␮M HTNC, whereas Ϸ5 ␮M of HNC is necessary to achieve the same level of recombination within the same time NLS Promotes Cellular Uptake of Recombinant Cre. All recombinant (Fig. 2C). Moreover, using HTNC shorter incubation times as Cre fusion proteins analyzed in this study resulted in highly compared with HNC are needed to obtain the same recombi- efficient transduction and subsequent recombination, with one nation efficiency (Fig. 3C). Incubation with 2 ␮M HTNC for 30 exception, namely CH (Fig. 2B). Recombination efficiency after min resulted in about 50% recombination; incubation with 10

4492 ͉ www.pnas.org͞cgi͞doi͞10.1073͞pnas.032068699 Peitz et al. Downloaded by guest on October 2, 2021 Fig. 5. HTNC transduction into (A) murine embryonic stem cells and (B and C) splenocytes ex vivo analyzed by Southern blot. (A) RDR-IB1N ES cells were incubated for 20 h in medium containing indicated concentrations of either HNC or HTNC. After 3 additional days of incubation in normal ES medium, genomic DNAs were examined by Southern blot analysis. (B) Splenocytes of a pol␤-flox͞flox mouse (6) were incubated for 60 min in medium containing indicated concentrations of either HNCF, HNC, or HTNC. After transduction, cells were incubated for 16 h in normal medium. Shown is Southern blot analysis of sorted B cells (CD19ϩ͞PIϪ), T cells (Thy1.2ϩ͞PIϪ), and total spleno- ␤ ͞ Fig. 4. Detection of Cre in cells after transduction employing an anti-Cre cytes. (C) Splenocytes of a pol -flox flox mouse were incubated with medium antibody. CV1–5B cells were incubated for 4 h with 5 ␮M of either CH or NCH containing 250 nM, 500 nM, and 1000 nM of HTNC for indicated periods of ⌬ or medium only as a control. After incubation, cells were examined for time. Calculated deletion efficiencies ( %) and cell viabilities in % are given ⌬ immunofluorescence by using anti-Cre antiserum or by phase contrast micros- at the bottom. wt, wild type allele; flox, loxP-flanked allele; , deleted allele. copy. Cells were counterstained using 4,6-di-amino-2-phenylindol (DAPI). See Materials and Methods for details. After 60 min of incubation cells were incubated for additional 16 h in normal medium providing time for Cre-mediated re- ␮M HNC induced recombination only in Ϸ15% of the cells combination. Cells were then sorted into B and T cells. Southern within the same time. Highly efficient recombination of about blot analysis of extracted genomic DNAs demonstrated that 80% was achieved after 2 h incubation with HTNC, whereas 16 h treatment with 10 ␮M HNCF results in 15% deletion in total were needed to obtain the same level of recombination when splenocytes (Fig. 5B Left). As already observed in fibroblasts, the using HNC (Fig. 3C). same concentration of HNC is more efficient (29%). However, the most efficient recombination was obtained with 1 ␮M HTNC Transduction into Embryonic Stem Cells and Splenocytes ex Vivo. To (63%). Similar values were obtained from sorted B and T cells evaluate further the use of HTNC for genetic engineering we (Fig. 5B Middle and Right). 10 ␮M HNC resulted in 20–30% tested transduction into murine ES cells. We used the RDR- recombination, whereas a 10-fold lower concentration of HTNC IB1N cell line containing a loxP-modified substrate in the induced Cre-mediated recombination in 52–67% of the B and T

ROSA26 locus (F. T. Wunderlich, F.E., and K.R., unpublished cells. Even treatment with only 500 nM of HTNC resulted in GENETICS data). We incubated ES cells in either HNC- or HTNC- more than 50% recombination in either total splenocytes, or containing medium for 20 h. Southern blot analysis of genomic sorted B or T cells (Fig. 5B). To assess the time dependency of DNA demonstrated that Cre-mediated recombination was HTNC transduction and the viability of the treated cells we achieved in almost every cell when using 2 ␮M HTNC (Fig. 5A). incubated splenocytes for 15, 30, 60, and 120 min in 250 nM, 500 Eighty-seven-percent deletion was achieved by using only 1 ␮M nM, and 1 ␮M HTNC (Fig. 5C). Southern blot analysis revealed HTNC. A 10-fold higher concentration of HNC was necessary to that substantial recombination occurred already after 15 min obtain a similar level of deletion. No difference in colony incubation (12–44% recombination, depending on the concen- numbers were observed 2 days after transduction between the tration). Recombination efficiency increased with longer incu- cell suspensions incubated with 10 ␮M HNC and 1 ␮M HTNC, bation times and reached saturation between 30 and 60 min. No respectively, as compared with control cells incubated in normal further increase was observed after 120 min. The highest re- medium. A significant decrease of colony numbers was observed combination efficiency was obtained after 60 min of incubation in the case of cells treated with more than 2 ␮M HTNC, with 1 ␮M HTNC (71%); cell viability of this population was indicative of a cytotoxic effect of HTNC at high concentrations. 38%. Higher viabilities were obtained with lower concentrations: Indeed, growth of cells was almost completely inhibited in the after 60 min in 500 nM HTNC 57% of the splenocytes were still presence of 5 ␮M HTNC. However, at concentrations that are viable and 70% viability was determined after incubation with sufficient to induce deletion in about 90% of the cells (1–2 ␮M 250 nM HTNC; this value is comparable to cells cultured in HTNC) no growth inhibition was observed; recombination normal medium displaying a viability of 75% (Fig. 5C). efficiencies of more than 95% after either HTNC (2 ␮M) or HNC (10 ␮M) transduction were reproducibly confirmed also by Conclusions employing a different ES-cell line containing a loxP-substrate The aim of this study was to evaluate the potency of two (data not shown). Additionally, we also wanted to assess the prominent PTDs, namely the hydrophobic FGF and the basic potential of HTNC transduction into resting cells such as pri- TAT peptide, to promote cellular uptake of recombinant Cre. mary splenocytes. We therefore incubated splenocytes of mice Transduction of His-NLS-Cre and subsequent recombination of carrying a loxP-flanked segment in the 5Ј region of the pol␤ gene a loxP-modified substrate in reporter cells turned out to be (6) either in HNCF-, HNC-, or HTNC-containing medium. independent of the FGF peptide. This may also be true for the

Peitz et al. PNAS ͉ April 2, 2002 ͉ vol. 99 ͉ no. 7 ͉ 4493 Downloaded by guest on October 2, 2021 His-NLS-Cre-MTS fusion protein, which has recently been significant Cre-mediated recombination by Southern blot after reported to allow Cre transduction (31) and is very similar to the i.p. injection of the same amount as described (31) of HTNC, HNCF protein used in the present experiments. If not the FGF HNC, and HNCF, respectively. We observed only a weak signal peptide, what else drives the translocation of the His-NLS-Cre representing Ϸ10% deletion in a single mouse in the peritoneum across the cellular membrane? The only recombinant protein close to the site of injection (data not shown). We thus failed to analyzed in this study displaying no or weak transduction po- induce substantial Cre-mediated recombination in mice by Cre tential was Cre-His—i.e., the only construct without NLS. From transduction—not even when using the very efficient HTNC this observation we conclude that the basic NLS (see legend of protein. It has recently been reported that i.p. administration of Fig. 1A) itself can function as a PTD at least in the context of His-NLS-Cre-MTS, almost identical to HNCF, results in effi- Cre recombinase. This conclusion is consistent with the obser- cient recombination in mice as well (31). In this study, recom- vation that basic peptides in general are able to enhance cellular bination was determined indirectly by activation of a LacZ uptake of heterologous proteins or peptides (35–37). We further reporter gene. We infer that there might be some recombination extended this observation by generating a fusion of NLS-Cre activity in vivo, especially close to the site of injection. This with the longer basic TAT-derived peptide, known to act effi- activity could be strong enough to activate reporter genes in ciently as a PTD in some experimental settings (26–28). Indeed, some experimental settings, but seems insufficient to result in our studies demonstrated that the TAT-modified variant His- recombination of the loxP targets in the majority of the cells. TAT-NLS-Cre transduced much more efficiently than His-NLS- In conclusion, in our collection of Cre fusions HTNC turned Cre. More than 90% of fibroblast cells undergo Cre-mediated out to be most efficiently transducible into mammalian cells. recombination after treatment with 500 nM HTNC; a 10-fold Every cell line examined in this study, either primary cells ex vivo higher concentration of HNC is needed to obtain the same level or stable cell lines, could be modified genetically by HTNC of recombination. HTNC transduction is more efficient also in transduction at low, seemingly nontoxic concentrations, with at murine embryonic stem cells (5–10-fold) and primary spleno- least a 50% (and up to 95%) efficiency, depending on the cell cytes (10–20-fold) as compared with HNC. Thus, application of type. We expect that HTNC transduction will be applicable for HTNC turns out to be a powerful tool to rapidly and efficiently a much larger variety of cells, thereby overcoming limitations of perform Cre-mediated recombination in cultured mammalian conventional techniques such as transfection and adenoviral cells. infection. Because HTNC can be readily produced in and To examine the potential of HTNC transduction in vivo as purified from E. coli in large quantities, we expect HTNC to well, we performed a series of experiments in mice employing serve as a powerful tool for genetic manipulation in mammalian HTNC, HNC, and HNCF. To rule out the possibility that an cells. indirect read-out system (e.g., Cre-mediated activation of a reporter gene) could lead to an overestimation of the actual We are indebted to G. Schu¨tz for providing the reporter cell line CV1–5B. We thank F. T. Wunderlich, R. Ku¨hn, F. Schwenk, M. recombination efficiency, we decided to directly assess the Alimzhanov, N. Uyttersprot, and A. Waisman for helpful discussions, recombination product at the DNA level by Southern blot and C. Goettlinger for technical help. This work was supported by funds analysis. Although three different mouse lines each containing from the Volkswagen Foundation (I͞76353), the European Union different loxP-modified loci were used, we did not detect any (QLG1-1999-00202), and Artemis Pharmaceuticals, Cologne, Germany.

1. Rajewsky, K., Gu, H., Ku¨hn, R., Betz, U. A. K., Mu¨ller, W., Roes, J. & Schwenk, 19. Vasioukhin, V., Degenstein, L., Wise, B. & Fuchs, E. (1999) Proc. Natl. Acad. F. (1996) J. Clin. Invest. 98, S51–S53. Sci. USA 96, 8551–8556. 2. Nagy, A. (2000) Genesis 26, 99–109. 20. Danielian, P. S., Muccino, D., Rowitch, D. H., Michael, S. K. & McMahon, 3. Lewandoski, M. (2001) Nat. Rev. Genet. 2, 743–755. A. P. (1998) Curr. Biol. 8, 1323–1326. 4. Orban, P. C., Chui, D. & Marth, J. D. (1992) Proc. Natl. Acad. Sci. USA 89, 21. Kuzin, I. I., Snyder, J. E., Ugine, G. D., Wu, D., Lee, S., Bushnell, T., Jr., Insel, 6861–6865. R. A., Young, F. M. & Bottaro, A. (2001) Int. Immunol. 13, 921–931. 5. Lakso, M., Sauer, B., Mosinger, B., Jr., Lee, E. J., Manning, R. W., Yu, S. H., 22. Lin, Q., Dong, C. & Cooper, M. D. (1998) J. Exp. Med. 187, 79–87. Mulder K. L. & Westphal, H. (1992) Proc. Natl. Acad. Sci. USA 89, 6232–6236. 23. Loonstra, A., Vooijs, M., Beverloo, H. B., Allak, B. A., van Drunen, E., Kanaar, 6. Gu, H., Marth, J. D., Orban, P. C., Mossmann, H. & Rajewsky, K. (1994) R., Berns, A. & Jonkers, J. (2001) Proc. Natl. Acad. Sci. USA 98, 209–214. Science 265, 103–106. 24. Silver, D. P. & Livingston, D. M. (2001) Mol. Cell 8, 233–243. 7. Ku¨hn, R., Schwenk, F., Aguet, M. & Rajewsky, K. (1995) Science 269, 25. Hawiger, J. (1999) Curr. Opin. Chem. Biol. 3, 89–94. 1427–1429. 26. Schwarze, S. R., Ho, A., Hruska, K. A. & Dowdy, S. F. (2000) Trends Cell Biol. 8. Kellendonk, C., Tronche, F., Monaghan, A. P., Angrand, P. O., Stewart, A. F. 10, 290–295. & Schu¨tz, G. (1996) Nucleic Acids Res. 24, 1404–1411. 27. Nagahara, H., Vocero-Akbani, A. M., Snyder, E. L., Ho, A., Latham, D. G., 9. Kellendonk, C., Tronche, F., Casanova, E., Anlag, K., Opherk, C. & Schu¨tz, G. Lissy, N. A., Becker-Hapak, M., Ezhevsky, S. A. & Dowdy, S. F. (1998) Nat. (1999) J. Mol. Biol. 285, 175–182. Med. 4, 1449–1452. 10. Kulkarni, R. N., Bruning, J. C., Winnay, J. N., Postic, C. Magnuson, M. A. & 28. Schwarze, S. R., Ho, A., Vocero-Akbani, A. & Dowdy, S. F. (1999) Science 285, Kahn, C. R. (1999) Cell 96, 329–339. 1569–1572. 11. Li, M., Indra, A. K., Warot, X. Brocard, J., Messaddeq, N., Kato, S., Metzger, 29. Rojas, M., Donahue, J. P., Tan, Z. & Lin, Y. Z. (1998) Nat. Biotechnol. 16, D. & Chambon, P. (2000) Nature (London) 407, 633–636. 370–375. 12. Maruyama, M., Lam, K. P. & Rajewsky, K. (2000) Nature (London) 407, 30. Lin, Y. Z., Yao, S. Y., Veach, R. A., Torgerson, T. R. & Hawiger, J. (1995) 636–642. J. Biol. Chem. 270, 14255–14258. 13. Torres, R. M. & Ku¨hn, R. (1997) Laboratory Protocols for Conditional Gene 31. Jo, D., Nashabi, A., Doxsee, C., Lin, Q., Unutmaz, D., Chen, J. & Ruley, H. E. Targeting (Oxford Univ. Press, Oxford). (2001) Nat. Biotechnol. 19, 929–933. 14. Rohlmann, A., Gotthardt, M., Willnow, T. E., Hammer, R. E. & Herz, J. (1996) 32. Wunderlich, F. T., Wildner, H., Rajewsky, K. & Edenhofer, F. (2001) Nucleic Nat. Biotechnol. 14, 1562–1565. Acids Res. 29, e47. 15. Shibata, H., Toyama, K., Shioya, H., Ito, M., Hirota, M., Hasegawa, S., 33. Cantor, E. J. & Chong, S. (2001) Protein Expression Purif. 22, 135–140. Matsumoto, H., Takano, H., Akiyama, T., Toyoshima, K., et al. (1997) Science 34. Le, Y., Gagneten, S., Tombaccini, D., Bethke, B. & Sauer B. (1999) Nucleic 278, 120–123. Acids Res. 27, 4703–4709. 16. Utomo, A. R., Nikitin, A. Y. & Lee, W. H. (1999) Nat. Biotechnol. 17, 35. Wender, P. A., Mitchell, D. J., Pattabiraman, K., Pelkey, E. T., Steinman, L. 1091–1096. & Rothbard, J. B. (2000) Proc. Natl. Acad. Sci. USA 97, 13003–13008. 17. Minamino, T., Gaussin, V., DeMayo, F. J. & Schneider, M. D. (2001) Circ. Res. 36. Matsushita, M., Tomizawa, K., Moriwaki, A., Li, S. T., Terada, H. & Matsui, 88, 587–592. H. (2001) J. Neurosci. 21, 6000–6007. 18. Fuhrmann-Benzakein, E., Garcia-Gabay, I., Pepper, M. S., Vassalli, J. D. & 37. Han, K., Jeon, M. J., Kim, S. H., Ki, D., Bahn, J. H., Lee, K. S., Park, J. & Choi, Herrera, P. L. (2000) Nucleic Acids Res. 28, e99. S. Y. (2001) Mol. Cells 12, 267–271.

4494 ͉ www.pnas.org͞cgi͞doi͞10.1073͞pnas.032068699 Peitz et al. Downloaded by guest on October 2, 2021