Controlling transgene expression in subcutaneousimplants using a skin lotion containing the applemetabolite phloretinMarc Gitzingera, Christian Kemmera, Marie Daoud El-Babab, Wilfried Webera,1, and Martin Fusseneggera,2

aDepartment of Biosystems Science and Engineering, Eidgenossische Technische Hochschule Zurich, Mattenstrasse 26, CH-4058 Basel, Switzerland;and bUniversite de Lyon, F-69622, Lyon, France

Edited by Charles R. Cantor, Sequenom, Inc., San Diego, CA, and approved April 29, 2009 (received for review February 10, 2009)

Adjustable control of therapeutic transgenes in engineered cellimplants after transdermal and topical delivery of nontoxic triggermolecules would increase convenience, patient compliance, andelimination of hepatic first-pass effect in future therapies. Pseudo-monas putida DOT-T1E has evolved the flavonoid-triggered TtgRoperon, which controls expression of a multisubstrate-specificefflux pump (TtgABC) to resist plant-derived defense metabolitesin its rhizosphere habitat. Taking advantage of the TtgR operon,we have engineered a hybrid P. putida–mammalian genetic unitresponsive to phloretin. This flavonoid is contained in apples, and,as such, or as dietary supplement, regularly consumed by humans.The engineered mammalian phloretin-adjustable control element(PEACE) enabled adjustable and reversible transgene expression indifferent mammalian cell lines and primary cells. Due to the shorthalf-life of phloretin in culture, PEACE could also be used toprogram expression of difficult-to-produce protein therapeuticsduring standard bioreactor operation. When formulated in skinlotions and applied to the skin of mice harboring transgenic cellimplants, phloretin was able to fine-tune target genes and adjustheterologous protein levels in the bloodstream of treated mice.PEACE-controlled target gene expression could foster advances inbiopharmaceutical manufacturing as well as gene- and cell-basedtherapies.

synthetic biology � synthetic gene networks � transdermal gene regulation �gene therapy � biopharmaceutical manufacturing

Synthetic mammalian expression systems, which enable re-versible and adjustable transgene expression, have been

essential for recent advances in (i) functional genomic research(1), (ii) drug discovery (2, 3), (iii) manufacturing of difficult-to-produce protein therapeutics (4, 5), (iv) the design of syntheticgene networks replicas reaching the complexity of electroniccircuits (6–9), and (v) gene therapy applications (10–12).

To date, a multitude of heterologous transgene expressionsystems for use in mammalian cells and transgenic animals havebeen described (4). The prevailing design consists of a heterol-ogous small molecule-responsive transactivator engineered byfusing a prokaryotic repressor to a eukaryotic transactivationdomain and a transactivator-specific promoter containing thematching prokaryotic operator linked to a minimal eukaryoticpromoter. Inducer-triggered modulation of the affinity of thetransactivator to its cognate promoter results in adjustable andreversible transcription control of the specific target gene (13–16). In recent years, a panoply of such heterologous transcriptioncontrol modalities have been developed, which are responsive tovarious inducer molecules such as antibiotics (13, 14, 17), steroidhormones and their analogs (18, 19), quorum-sensing molecules(20, 21), immunosuppressive and antidiabetic drugs (22, 23),biotin (24), L-arginine (25), as well as volatile acetaldehyde (16).Apart from gaseous acetaldehyde, which can simply be inhaled,all other inducers need to be either taken up orally or beadministered by injection in any future gene therapy application.Transdermal and topical delivery of inducer molecules, which

would provide advantages over conventional injection-based ororal administration such as convenience, improved patient com-pliance, and elimination of hepatic first-pass effect, have not yetbeen developed.

Phloretin is mainly found in the root bark of apple trees andin apples where it acts as a natural antibacterial plant defensemetabolite (26). Phloretin has been studied as a possible pene-tration enhancer for skin-based drug delivery (27–31), attenu-ates inflammation by antagonizing prostaglandins (32), protectsthe skin from UV light-induced photodamage (33, 34), and iscurrently evaluated as a chemopreventive agent for cancertreatment (35). Because the plant rhizosphere is one of thenatural habitats of Pseudomonas putida (strain DOT-T1E), thisprokaryote has evolved the RND family transporter TtgABCwith multidrug recognition properties, which is controlled by itscognate repressor TtgR binding to a specific operator (OTtgR) inthe TtgR promoter (PTtgR). Phloretin has been shown to bind tothe TtgR-operator complex at a stoichiometric ratio of 1 effectormolecule per TtgR dimer, and to release TtgR from OTtgR, whichresults in induction of TtgABC production and effective pump-mediated efflux of the flavonoid from P. putida (26, 36).

Capitalizing on the phloretin-responsive TtgR-OTtgR interac-tion of P. putida DOT-T1E, we have assembled a syntheticmammalian phloretin-adjustable control element (PEACE),which was able to reversibly adjust product gene expression oftransgenic cells grown in culture, standard bioreactors, or im-planted into mice after addition of pure phloretin or topicaladministration of a phloretin-containing skin lotion.

ResultsDesign of a Synthetic Mammalian PEACE. With a half-life of 70 h inculture and no negative influence on viability, growth, orproduction of CHO-K1 cells, phloretin is a valid flavonoidcandidate for trigger-inducible transcription control in mamma-lian cells (SI Materials and Methods and Fig. S1). Living in theplant rhizosphere, P. putida DOT-T1E has evolved resistance tovarious plant-derived antimicrobials (37, 38), which is triggeredby phloretin-induced release of TtgR from the operator (OTtgR)of its target promoter and subsequent induction of a broadlyspecific TtgABC efflux pump (26, 36). By fusing TtgR (36) to theHerpes simplex-derived transactivation domain VP16 (39), wecreated a synthetic mammalian transactivator (TtgA1), which is

Author contributions: M.G., W.W., and M.F. designed research; M.G., C.K., and M.D.E.-B.performed research; M.G., W.W., and M.F. analyzed data; and M.G. and M.F. wrote thepaper.

The authors declare no conflict of interest.

This article is a PNAS Direct Submission.

Freely available online through the PNAS open access option.

1Present address: Albert-Ludwigs-University Freiburg i.Br., D-79108 Freiburg, Germany.

2To whom correspondence should be addressed. E-mail: fussenegger@bsse.ethz.ch.

This article contains supporting information online at www.pnas.org/cgi/content/full/0901501106/DCSupplemental.

10638–10643 � PNAS � June 30, 2009 � vol. 106 � no. 26 www.pnas.org�cgi�doi�10.1073�pnas.0901501106

able to bind and activate transcription from chimeric promoters(PTtgR1) harboring OTtgR linked to a minimal human cytomeg-alovirus immediate early promoter (PhCMVmin), in a phloretin-responsive manner (Fig. 1 A and B). Cotransfection of theconstitutive TtgA1 expression vector pMG11 (PSV40-TtgA1-pA)and pMG10 [PTtgR1-secreted alkaline phosphatase (SEAP)-pA]encoding a TtgA1-specific PTtgR1-driven SEAP expression unit,resulted in high-level SEAP expression [23.6 � 3.1 units (U)/L],which compares with an isogenic vector containing a constitutivePSV40-driven SEAP expression cassette (pSEAP2-Control;21.4 � 1.0 U/L). Addition of increasing concentrations ofphloretin (0–70 �M) to a culture of pMG10- and pMG11-cotransfected CHO-K1 cells resulted in dose-dependent reduc-tion of SEAP expression up to complete repression (Fig. 1C).These data suggest that PEACE-controlled transgene expressionis adjustable and enables complete repression within a nontoxicphloretin concentration range. We have also designed PTtgR1

variants with different tandem OTtgR modules and TtgA1 vari-ants harboring various transactivation domains, and provide adetailed combinatorial performance analysis in different celllines and different expression configurations, including auto-regulated ones that are known to be essential for the assemblyof complex synthetic gene networks (SI Materials and Methods,Fig. S2, Fig. S3, and Table S1) (6, 9, 40).

PEACE Control by Phloretin and Other Flavonoids. Because TtgR ofP. putida was shown to bind several plant-derived flavonoids withhigh affinity (26), we profiled their PEACE-controlling capac-ities in mammalian cells. CHO-K1 were transiently (co)trans-fected with either pMG10 (PTtgR1-SEAP-pA) and pMG11(PSV40-TtgA1-pA), to score regulation performance, or withpSEAP2-Control (PSV40-SEAP-pA), to assess compound-related cytotoxicity, and then cultivated for 48 h in mediumcontaining different concentrations (0, 25, and 50 �M) ofspecific f lavonoids (berberine, butylparaben, genistein, luteolin,�-naphthol, naringenin, phloretin, phloridzin, or quercetin) be-fore SEAP production was profiled (Fig. 2A). Althoughgenistein, luteolin, �-naphthol, naringenin, and quercetine werecytotoxic within the tested concentration range (genistein onlyat 50 �M), berberine, butylparaben, phloridzin, and phloretindid not reduce cell viability. However, berberine failed to controlPEACE, and butylparaben, as well as phloridzin, were able toregulate, but not fully repress SEAP production (Fig. 2B).Therefore, phloretin, which enabled maximum expression levels,as well as full transgene repression, was chosen as the idealPEACE inducer for all further experiments.

Fig. 1. Design and functionality of PEACE. The P. putida DOT-T1E-derivedbacterial repressor TtgR was fused to the VP16 transactivation domain of H.simplex virus, and the resulting transactivator TtgA1 (TtgR-VP16) was clonedunder control of the constitutive simian virus 40 promoter (PSV40) (pMG11).The phloretin-responsive promoter (PTtgR1; OTtgR-PhCMVmin) contains a chimericTtgR-specific operator sequence (OTtgR, CAGTATTTACAAACAACCATGAATG-TAAGTATATTC; TtgR binding sites in italics), which is located 5� of a minimalhuman cytomegalovirus immediate early promoter (PhCMVmin), and was set todrive expression of the human placental SEAP (pMG10). (A) ON status. TtgA1

is constitutively expressed and binds to PTtgR1 in the absence of phloretin; thus,inducing SEAP expression. (B) OFF status. Addition of phloretin releases TtgA1

from PTtgR1, which switches SEAP expression off. (C) SEAP expression profilesof CHO-K1 transiently transfected with pMG11 (PSV40-TtgA1-pA) and pMG10(PTtgR1-SEAP-pA), and cultivated for 48 h in the presence of different phloretinconcentrations (0–70 �M).

Fig. 2. PEACE responsiveness to different flavonoids. (A) Toxicity of fla-vonoids. CHO-K1 were transiently transfected with pSEAP2-Control, culti-vated in medium supplemented with different flavonoids (0, 25, and 50 �M)SEAP levels were scored after 48 h. (B) CHO-K1 cells transiently expressing allPEACE components (pMG10 and pMG11) were cultivated in the presence ofdifferent flavonoids and SEAP expression was profiled after 48 h.

Gitzinger et al. PNAS � June 30, 2009 � vol. 106 � no. 26 � 10639

BIO

PHYS

ICS

AN

DCO

MPU

TATI

ON

AL

BIO

LOG

Y

Phloretin-Controlled Transgene Expression Is Functional in DifferentMammalian Cell Lines and Human Primary Cells. To assess its ver-satility, we tested PEACE in several immortalized mammaliancell lines, as well as in human primary cells. Therefore, pMG10(PTtgR1-SEAP-pA) and pMG11 (PSV40-TtgA1-pA) were cotrans-fected into BHK-21, COS-7, HaCaT, HEK-293, HT-1080, andNIH/3T3 cell lines, as well as into primary human fibroblasts andkeratinocytes, and cultivated for 48 h in the presence (50 �M)and absence of phloretin, followed by scoring of SEAP levels(Table 1). PEACE-controlled transgene expression was func-tional in all tested cell lines, suggesting that this technology willbe broadly applicable.

Expression Kinetics, Adjustability, and Reversibility of PEACE-Con-trolled Transgene Expression in a Stable Transgenic CHO-K1 Cell Line.We have generated 5 double-transgenic cell lines (CHO-PEACE) by sequential transfection and clonal selection ofpMG11 and pMG10 into CHO-K1. All of these PEACE-transgenic cell lines showed phloretin-regulated SEAP expres-sion, but differed in their overall regulation performance (max-imum and leaky expression levels) as a result of differences intransgene copy number and integration sites that remains be-yond control using standard transfection technology (Fig. S4)(41). CHO-PEACE8, which was considered the best cell line andwas, therefore, used in all follow-up experiments, showed (i)unchanged maximum SEAP expression levels and unaffectedregulation performance in long-term cultures over 60 days (day0, ON: 70.9 � 3.1 U/L, OFF: 1.8 � 0.1 U/L; day 60, ON: 67.6 �2.7 U/L, OFF: 2.1 � 0.2 U/L; OFF at 50 �M phloretin), (ii)excellent adjustability (Fig. 3A), (iii) exponential SEAP expres-sion kinetics (Fig. 3B), (iv) full reversibility of transgene expres-sion (Fig. 3C), and (v) optimal compatibility with other trans-gene regulation systems (SI Materials and Methods and Tables S2and S3).

Time-Delayed Induction of Product Gene Expression in a PrototypeBiopharmaceutical Manufacturing Scenario. Because phloretin is anontoxic fruit component, it could be an ideal product gene inducerfor biopharmaceutical manufacturing scenarios that require precisetiming or dosing of difficult-to-express protein pharmaceuticals (4,42–44). Also, because phloretin has a determined half-live of 70 hin mammalian cell culture systems (SI Results), any productionculture can be programmed to start product gene expression at apredefined point in time as phloretin concentrations drop below arepressing threshold level. We have inoculated a 1-L BioWavebioreactor with 2 � 103 CHO-PEACE8 and different transgene-repressing phloretin doses of 60, 80, or 100 �M. Although CHO-PEACE8 grew exponentially from the start of the bioprocess, SEAP

production gradually increased once phloretin was degraded tononrepressive levels (40 �M; Fig. 4). Using PEACE, mammalianproduction cultures could indeed be programmed for timely induc-tion of product gene expression without any process intervention.

Phloretin-Mediated Transdermal Gene Expression in s.c. Implants inMice. Because phloretin has been suggested as a penetrationenhancer for transdermal drug delivery (27–31), and was shownto propagate systemically in rodents after local skin-basedadministration (28), we have evaluated phloretin as a potentialtransdermal therapeutic transgene expression inducer. There-fore, we have microencapsulated CHO-PEACE8 in coherentalginate-PLL-alginate capsules and implanted them s.c. intomice. The back of treated mice was partially shaved, andpetroleum jelly-based skin lotions (200 �L) containing differentamounts of phloretin (0–42 mg) were put on every day. The

Table 1. PEACE-controlled transgene expression in variousmammalian cells

SEAP production, U/L

Cell line0 �M

phloretin50 �M

phloretin

BHK-21 14.4 � 0.4 1.5 � 0.2COS-7 1.3 � 0.02 NDHaCaT 0.9 � 0.03 NDHEK-293 27.3 � 0.8 0.2 � 0.04HT-1080 7.8 � 0.2 NDNIH/3T3 8.5 � 0.3 NDPrimary human fibroblasts 1.0 � 0.04 NDPrimary human keratinocytes 1.5 � 0.05 ND

SEAP production was quantified 48 h after cotransfection pMG10 (PTtgR1-SEAP-pA) and pMG11 (PSV40-TtgA1-pA) into indicated cell lines and primarycells. ND, not detectable.

Fig. 3. Design and characterization of the stable CHO-PEACE8 cell linetransgenic for phloretin-responsive SEAP expression. (A) Dose-response pro-file of CHO-PEACE8. (B) SEAP expression kinetics of CHO-PEACE8 cultivated for72 h in the presence and absence of 50 �M phloretin. (C) Reversibility ofCHO-PEACE8-based SEAP production; 2 � 105 CHO-PEACE8 were cultivated for144 h in the presence or absence of 50 �M phloretin. Every 48 h, the cell densitywas readjusted to 2 � 105, and the phloretin status of the culture was reversed.

10640 � www.pnas.org�cgi�doi�10.1073�pnas.0901501106 Gitzinger et al.

SEAP levels detected in treated mice 72 h after implantationshowed phloretin-dependent dose-response characteristics akinto the ones observed with the same microencapsulated implantbatch cultivated and exposed to phloretin in vitro (Fig. 5 A andB). Control mice receiving CHO-K1 cells transgenic for consti-tutive SEAP expression were insensitive to any treatment withphloretin-containing skin lotion (0 mg phloretin: 6.2 � 0.43 U/L;42 mg phloretin: 6.02 � 0.58 U/L).

DiscussionFlavonoids such as phloretin are polyphenols widely distributedin the plant kingdom, and they are present in fruits andvegetables regularly consumed by humans. The recent findingsthat phloretin and some of its derivatives have potent antiin-f lammatory (32), antioxidative (33, 45), and even some antican-cer (35, 46) activities form the scientific basis for the commonsaying ‘‘an apple a day keeps the doctor away.’’ Phloretin has alsobeen successfully evaluated as a penetration enhancer for trans-dermal drug delivery (27–31) and as skin protectant reducingoxidative stress resulting from external insults, such as UVirradiation, which triggers skin cancer and photo aging (33, 34,47). Recent studies in rodents have confirmed that dermaladministration of phloretin will systemically spread in the animal(28), and that phloretin has a rather short half-life in vivo (48).All of the aforementioned characteristics corroborate phloretinto be a nontoxic natural compound with high metabolic turnoverthat could be ideal for reversible transdermal induction oftherapeutic transgenes. Transdermal and topical delivery ofdrugs and regulating molecules provide advantages over con-ventional oral or injection-based administrations, such as con-venience, improved patient compliance and elimination of he-patic first-pass effect. However, most molecules are notapplicable to dermal administration due to the excellent barrierproperties of the skin, which requires penetrating molecules topass the stratum corneum with its compact keratinized cell layersand the viable epidermis before reaching the papillary dermisand crossing the capillary walls into systemic circulation. Phlor-etin-containing skin lotions put on the skin of mice containingcell implants harboring a synthetic phloretin-responsive expres-sion system were able to precisely fine-tune target gene expres-sion in the animal. This pioneering transdermal transcriptioncontrol system may enable precise patient-controlled dosing ofprotein pharmaceuticals that are produced in situ by cell im-plants contained in clinically licensed devices (49–52). Thereshould be no risk of nutrition-based interference of PEACE-controlled therapeutic transgene expression in transgenic cellimplants, because a patient (70 kg) would need to eat �2,000apples to reach PEACE-modulating phloretin levels in hisbloodstream (48, 53). Besides this gene therapy-focused in vivoscope of phloretin-responsive transgene expression, the systemshowed excellent regulation performance, including adjustabil-ity and reversibility in vitro. Owing to its short half-life in culture,phloretin-responsive production cultures grown in bioreactorscould be preprogrammed for timely production initiation byinoculation with excessive phloretin concentrations. While theproduction cell cultures grow and phloretin levels decrease afterprecise kinetics, production will be initiated at a defined point intime, which is a function of the inoculum and the initial phloretinconcentration. Such a time-delayed production concept wouldbe particularly valuable for difficult-to-express protein thera-peutics like those that impair growth or are cytotoxic (4, 42–44).The fact that phloretin is a natural fruit component regularlyconsumed by humans and licensed for use in skin lotions (34)may facilitate approval of such biopharmaceutical manufactur-ing protocols by governmental agencies. Also, because thePEACE system has been assembled after a standard binarytransactivator/promoter design, it is conceivable that its perfor-mance can easily be adapted to specific control requirementsusing an established refinement program (54, 55).

Considering all facts, the pioneering phloretin-based trans-gene control technology is unique, and may foster advances inthe production of difficult-to-express protein pharmaceuticals,as well as in cell implant-based therapeutic applications.

Materials and MethodsExpression Vector Design. The pMG10 (PTtgR1-SEAP-pA) harbors a phloretin-responsive SEAP expression unit, and pMG11 (PSV40-TtgA1-pA) encodes con-

Fig. 4. Automatically programmed product gene expression in bioreactorsusing well-defined phloretin degradation profiles. The 2 � 103 cells/mL CHO-PEACE8 were cultivated in a bioreactor containing 1-L culture medium sup-plemented with either 60, 80, or 100 �M phloretin, and SEAP production wasprofiled for 211 h. Owing to a defined phloretin degradation profile in cultureand a precise induction threshold of 40 �M phloretin, the onset of SEAPproduction can be programmed to occur at a very precise point in time bydefining the cell density and the phloretin concentration at production start.

Fig. 5. PEACE-controlled transgene expression in mice. (A) Microencapsu-lated CHO-PEACE8 were implanted s.c. into female OF1 mice (2 � 106 cells permouse); 200 �L of a cream containing different amounts of phloretin (0, 5.25,10.5, 21, and 42 mg) was applied to a shaved skin area near the implant site.SEAP serum levels were quantified 72 h postimplantation. (B) SEAP expressionprofiles of the microencapsulated CHO-PEACE8 implant batch cultivated invitro for 72 h at different phloretin concentrations.

Gitzinger et al. PNAS � June 30, 2009 � vol. 106 � no. 26 � 10641

BIO

PHYS

ICS

AN

DCO

MPU

TATI

ON

AL

BIO

LOG

Y

stitutive expression of the phloretin-dependent transactivator. Detailed in-formation on expression vector design and plasmids used in this study isprovided in Table S4.

Cell Culture and Transfection. Wild-type Chinese hamster ovary cells (CHO-K1,ATCC CCL-61) and their derivatives were cultivated in standard medium:ChoMaster HTS (Cell Culture Technologies) supplemented with 5% (vol/vol)FCS (cat. no. 3302, lot no. P251110; PAN Biotech) and 1% (vol/vol) penicillin/streptomycin solution (cat. no. P4458; Sigma). Human embryonic kidney cells(HEK-293) (56), African green monkey kidney cells (COS-7, ATCC: CRL-1651),baby hamster kidney cells (BHK-21, ATCC: CCL10), human fibrosarcoma cells(HT-1080, ATCC: CCL-121), the human keratinocyte cell line HaCaT (57), andmouse fibroblasts (NIH/3T3, ATCC CRL-1658) were cultured in DMEM (cat. no.52100-39; Invitrogen) supplemented with 10% (vol/vol) FCS and 1% (vol/vol)penicillin/streptomycin solution. Primary human foreskin fibroblasts werecultivated in DMEM supplemented with 20% FCS (vol/vol) and 1% (vol/vol)penicillin/streptomycin solution, and primary human foreskin keratinocyteswere cultured in chemically defined serum-free keratinocyte medium (cat. no.10744019; Invitrogen), all kindly provided by Sabine Werner. All cell typeswere cultivated at 37 °C in a 5% CO2-containing humidified atmosphere. Fortransient transfection of CHO-K1, 1 �g of total plasmid DNA (for cotransfec-tion an equal amount of each plasmid) was transfected into 50,000 cells perwell of a 24-well plate according to an optimized calcium phosphate protocol,which resulted in typcal transfection efficiencies of 35 � 5% (58). Plasmid DNAwas diluted to a total volume of 25 �L of 0.5M CaCl2 solution, which was mixedwith 25 �L 2� HBS solution (50 mM Hepes/280 mM NaCl/1.5 mM Na2HPO4, pH7.1). After incubation for 15 min at room temperature, the precipitates wereimmediately added to the well and centrifuged onto the cells (5 min at 1,200 �g) to increase transfection efficiency. After 3 h, the cells were treated with 0.5mL glycerol solution (ChoMaster HTS medium containing 15% glycerol) for60 s. After washing once with PBS (cat. no. 21600-0069; Invitrogen), cells werecultivated in 0.5 mL of standard ChoMaster HTS medium in the presence orabsence of different concentrations of phloretin. For transfection of BHK-21,COS-7, and HEK-293, plasmid DNA-Ca3(PO4)2 precipitate was prepared andapplied to the cells as described above. HEK-293 and COS-7 cells were washedonce with PBS after 3-h incubation with the DNA-Ca3(PO4)2 precipitate andsubsequently cultivated in standard DMEM, whereas BHK-21 cells were incu-bated overnight with the precipitates and then cultivated in DMEM afterbeing washed once with PBS. HaCaT, HT-1080, NIH/3T3, as well as primaryhuman fibroblasts and keratinocytes, were transfected with Fugene 6 (cat. no.11814443001; Roche Diagnostics AG) according to the manufacturer’s proto-col, and cultivated in the cell culture medium specified above. After transfec-tion, all cells were cultivated in DMEM supplemented with various concen-trations of phloretin, and reporter protein levels were profiled 48 h aftertransfection, unless otherwise indicated.

Construction of Stable Cell Lines. The stable CHO-PEACE8 cell line, transgenicfor phloretin-controlled SEAP expression, was constructed in 2 steps: (i)CHO-K1 cells were cotransfected with pMG11 (PSV40-TtgA1-pA) and pSV2neo(cat. no. 6172-1; Clontech) at a ratio of 20:1, and clonal selection resulted in thecell line CHO-TtgA; and (ii) CHO-TtgA was cotransfected with pMG10 (PTtgR1-SEAP-pA) and pPur (cat. no. 6156-1; Clontech) at ratio of 20:1, and thephloretin-responsive SEAP-producing double-transgenic cell line CHO-PEACE8

was chosen after clonal selection. Phloretin-dependent dose-response char-acteristics of CHO-PEACE8 were analyzed by culturing 100,000 cells per mL for48 h in standard ChoMaster HTS medium at various phloretin concentrationsranging from 0 to 70 �M. Reversibility of phloretin-mediated SEAP productionwas assessed by cultivating CHO-PEACE8 (100,000 cells pr mL) for 144 h whilealternating phloretin concentrations from 0 to 50 �M every 48 h.

Quantification of Reporter Protein Production. Production of the humanplacental SEAP was quantified using a p-nitrophenylphosphate-based light

absorbance time course (59). 1 SEAP unit corresponds to the conversion of 1�mol pNPP per minute.

In Vivo Methods. CHO-PEACE8 and CHO-SEAP18 (60) were encapsulated inalginate-poly(L-lysine)-alginate beads (400 �m; 200 cells per capsule) using anInotech Encapsulator Research IE-50R (Recipharm) according to the manufac-turer’s instructions and the following parameters: 0.2-mm nozzle, 20-mLsyringe at a flow rate of 405 units, nozzle vibration frequency of 1,024 Hz, and900 V for bead dispersion. The backs of female OF1 mice (oncins France souche1; Charles River Laboratories) were shaved, and 300 �L of ChoMaster HTScontaining 2 � 106 encapsulated CHO-PEACE8 were injected s.c. Control micewere injected with microencapsulated CHO-K1. Shaving ensured direct con-tact of the phloretin-containing cream with the skin of the animal; 1 h afterimplantation, 200 �L of the phloretin-containing cream was applied to theskin around the injection site. The phloretin amounts in creams ranged from0 to 42 mg. The cream was applied once a day for up to 3 days. Thereafter, themice were killed, blood samples were collected, and SEAP levels were quan-tified in the serum, which was isolated by using microtainer SST tubes accord-ing to the manufacturer’s instructions (Beckton Dickinson). All of the exper-iments involving mice were performed according to the directives of theEuropean Community Council (86/609/EEC), approved by the French Republic(no. 69266310), and performed by Marie Daoud El-Baba at the Universite deLyon.

Bioreactor Operation. CHO-PEACE8 (inoculum of 2 � 103 cells per mL) werecultivated in a BioWave 20SPS-F bioreactor (Wave Biotech) equipped with a2-L Wave Bag optimized for optical pH and dissolved oxygen concentrationcontrol of the 1-L culture. The bioreactor was operated at a rocking rate of 15min�1, a rocking angle of 6°, and an aeration rate of 100 mL/min with inlet gashumidification (HumiCare 200; Gruendler Medical) to prevent evaporation ofthe medium. The medium (ChoMaster HTS, 5% FCS, 1% penicillin/streptomy-cin) was supplemented with 60, 80, or 100 �M phloretin.

Inducer Compounds and Formulation of the Skin Lotion. Berberine (cat. no.20425-0100; Acros) and luteolin (cat. no. L14186; Alfa Aesar) were prepared as10 mM stock solutions in 1:5 (vol/vol) DMSO/H2O. Butylparaben (cat. no.AV14043; ABCR), genistein (cat. no. A2202.0050; Axonlab), �-naphthol (cat.no. 185507; Sigma), naringenin (cat. no. N5893; Sigma), phloretin (cat. no.P7912; Sigma), phloridzin (cat. no. P3449; Sigma), and quercetin (cat. no.Q0125; Sigma) were prepared as 50 mM stock solution in DMSO, and wereused at a final concentration of 50 �M unless indicated otherwise. Thepetroleum jelly-based phloretin-containing creams were professionally for-mulated (Pharmacy Hoengg) and contained 25, 12.5, 6.25, and 3.125% (wt/wt)phloretin; 200 �L of skin lotion was topically applied per mouse and treat-ment, which corresponds to a respective total phloretin amount per dose of42, 21, 10.5, and 5.25 mg.

To quantify phloretin in cell culture medium, the samples were added to5 � 104 CHO-PEACE8 and incubated for 48 h before SEAP quantification.Phloretin levels were calculated by comparing SEAP production with a cali-bration curve (Fig. 3A), established using the same parameters, and definedphloretin concentrations. Similarly, the half-life of phloretin was estimatedbased on the degradation dynamics observed in cell culture (61, 62).

ACKNOWLEDGMENTS. We thank Martine Gilet for skilled technical assis-tance; Sabine Werner (Eidgenossische Technische Hochschule Zurich, Zurich,Switzerland) for providing primary human fibroblasts and keratinocytes; BeatKramer and Cornelia Fux (Eidgenossische Technische Hochschule Zurich, Zu-rich, Switzerland) for providing pBP99 and pCF59, respectively; and MarciaSchoenberg, Marcel Tigges, and William Bacchus for critical comments on themanuscript. This work was supported by the Swiss National Science Founda-tion Grant 3100A0-112549, and in part by the European Commission Frame-work Program 7 PERSIST.

1. Baumgartel K, et al. (2008) Control of the establishment of aversive memory bycalcineurin and Zif268. Nat Neurosci 11:572–578.

2. Sharpless NE, Depinho RA (2006) The mighty mouse: Genetically engineered mousemodels in cancer drug development. Nat Rev Drug Discov 5:741–754.

3. Weber W, et al. (2008) A synthetic mammalian gene circuit reveals antituberculosiscompounds. Proc Natl Acad Sci USA 105:9994–9998.

4. Weber W, Fussenegger M (2007) Inducible product gene expression technology tai-lored to bioprocess engineering. Curr Opin Biotechnol 18:399–410.

5. Ulmer JB, Valley U, Rappuoli R (2006) Vaccine manufacturing: Challenges and solu-tions. Nat Biotechnol 24:1377–1383.

6. Kramer BP, Fussenegger M (2005) Hysteresis in a synthetic mammalian gene network.Proc Natl Acad Sci USA 102:9517–9522.

7. Deans TL, Cantor CR, Collins JJ (2007) A tunable genetic switch based on RNAi and repressorproteins for regulating gene expression in mammalian cells. Cell 130:363–372.

8. Weber W, Daoud-El Baba M, Fussenegger M (2007) Synthetic ecosystems based onairborne inter- and intrakingdom communication. Proc Natl Acad Sci USA 104:10435–10440.

9. Tigges M, Marquez-Lago TT, Stelling J, Fussenegger M (2009) A tunable syntheticmammalian oscillator. Nature 457:309–312.

10. Gersbach CA, Le Doux JM, Guldberg RE, Garcia AJ (2006) Inducible regulation ofRunx2-stimulated osteogenesis. Gene Ther 13:873–882.

11. McLoughlin JM, et al. (2005) TNFerade, an adenovector carrying the transgene forhuman tumor necrosis factor alpha, for patients with advanced solid tumors: Surgicalexperience and long-term follow-up. Ann Surg Oncol 12:825–830.

10642 � www.pnas.org�cgi�doi�10.1073�pnas.0901501106 Gitzinger et al.

12. Weber W, Fussenegger M (2006) Pharmacologic transgene control systems for genetherapy. J Gene Med 8:535–556.

13. Fussenegger M, et al. (2000) Streptogramin-based gene regulation systems for mam-malian cells. Nat Biotechnol 18:1203–1208.

14. Gossen M, et al. (1995) Transcriptional activation by tetracyclines in mammalian cells.Science 268:1766–1769.

15. Urlinger S, et al. (2000) The p65 domain from NF-kappaB is an efficient human activatorin the tetracycline-regulatable gene expression system. Gene 247:103–110.

16. Weber W, et al. (2004) Gas-inducible transgene expression in mammalian cells andmice. Nat Biotechnol 22:1440–1444.

17. Weber W, et al. (2002) Macrolide-based transgene control in mammalian cells andmice. Nat Biotechnol 20:901–907.

18. Nordstrom JL (2002) Antiprogestin-controllable transgene regulation in vivo. CurrOpin Biotechnol 13:453–458.

19. Palli SR, Kapitskaya MZ, Potter DW (2005) The influence of heterodimer partnerultraspiracle/retinoid X receptor on the function of ecdysone receptor. FEBS J272:5979–5990.

20. Neddermann P, et al. (2003) A novel, inducible, eukaryotic gene expression systembased on the quorum-sensing transcription factor TraR. EMBO Rep 4:159–165.

21. Weber W, Link N, Fussenegger M (2006) A genetic redox sensor for mammalian cells.Metab Eng 8:273–280.

22. Pollock R, Clackson T (2002) Dimerizer-regulated gene expression. Curr Opin Biotech-nol 13:459–467.

23. Rollins CT, et al. (2000) A ligand-reversible dimerization system for controlling protein–protein interactions. Proc Natl Acad Sci USA 97:7096–7101.

24. Weber W, Bacchus W, Daoud-El Baba M, Fussenegger M (2007) Vitamin H-regulatedtransgene expression in mammalian cells. Nucleic Acids Res 35:e116.

25. Hartenbach S, Daoud-El Baba M, Weber W, Fussenegger M (2007) An engineeredL-arginine sensor of Chlamydia pneumoniae enables arginine-adjustable transcriptioncontrol in mammalian cells and mice. Nucleic Acids Res 35:e136.

26. Teran W, Krell T, Ramos JL, Gallegos MT (2006) Effector-repressor interactions, bindingof a single effector molecule to the operator-bound TtgR homodimer mediatesderepression. J Biol Chem 281:7102–7109.

27. Akimoto T, Kawahara K, Nagase Y, Aoyagi T (2001) Polymeric transdermal drugpenetration enhancer. The enhancing effect of oligodimethylsiloxane containing aglucopyranosyl end group. J Control Release 77:49–57.

28. Auner BG, Valenta C, Hadgraft J (2003) Influence of phloretin and 6-ketocholestanolon the skin permeation of sodium-fluorescein. J Control Release 89:321–328.

29. Auner BG, Valenta C, Hadgraft J (2003) Influence of lipophilic counter-ions in combi-nation with phloretin and 6-ketocholestanol on the skin permeation of 5-aminolevu-linic acid. Int J Pharm 255:109–116.

30. Valenta C, Cladera J, O’Shea P, Hadgraft J (2001) Effect of phloretin on the percuta-neous absorption of lignocaine across human skin. J Pharm Sci 90:485–492.

31. Valenta C, Nowak M, Hadgraft J (2001) Influence of phloretin and 6-ketocholestanolon the permeation of progesterone through porcine skin. Int J Pharm 217:79–86.

32. Blazso G, Gabor M (1995) Effects of prostaglandin antagonist phloretin derivatives onmouse ear edema induced with different skin irritants. Prostaglandins 50:161–168.

33. Auner BG, Wirth M, Valenta C (2005) Antioxidative activity and cytotoxicity of fourdifferent flavonoids for dermal applications. J Drug Deliv Sci Technol 15:227–232.

34. Oresajo C, et al. (2008) Protective effects of a topical antioxidant mixture containingvitamin C, ferulic acid, and phloretin against ultraviolet-induced photodamage inhuman skin. J Cosmet Dermatol 7:290–297.

35. Wu CH, et al. (2009) In vitro and in vivo study of phloretin-induced apoptosis in humanliver cancer cells involving inhibition of type II glucose transporter. Int J Cancer124:2210–2219.

36. Teran W, et al. (2003) Antibiotic-dependent induction of Pseudomonas putida DOT-T1E TtgABC efflux pump is mediated by the drug binding repressor TtgR. AntimicrobAgents Chemother 47:3067–3072.

37. Ramos JL, Duque E, Huertas MJ, Haidour A (1995) Isolation and expansion of thecatabolic potential of a Pseudomonas putida strain able to grow in the presence ofhigh concentrations of aromatic hydrocarbons. J Bacteriol 177:3911–3916.

38. Rojas A, et al. (2001) Three efflux pumps are required to provide efficient tolerance totoluene in Pseudomonas putida DOT-T1E. J Bacteriol 183:3967–3973.

39. Triezenberg SJ, Kingsbury RC, McKnight SL (1988) Functional dissection of VP16, thetrans-activator of herpes simplex virus immediate early gene expression. Genes Dev2:718–729.

40. Becskei A, Kaufmann BB, van Oudenaarden A (2005) Contributions of low moleculenumber and chromosomal positioning to stochastic gene expression. Nat Genet37:937–944.

41. Wurm FM (2004) Production of recombinant protein therapeutics in cultivated mam-malian cells. Nat Biotechnol 22:1393–1398.

42. Weber W, et al. (2007) A synthetic time-delay circuit in mammalian cells and mice. ProcNatl Acad Sci USA 104:2643–2648.

43. Boorsma M, et al. (2000) A temperature-regulated replicon-based DNA expressionsystem. Nat Biotechnol 18:429–432.

44. Fussenegger M, Schlatter S, Datwyler D, Mazur X, Bailey JE (1998) Controlled prolif-eration by multigene metabolic engineering enhances the productivity of Chinesehamster ovary cells. Nat Biotechnol 16:468–472.

45. Middleton E, Jr, Kandaswami C, Theoharides TC (2000) The effects of plant flavonoidson mammalian cells: Implications for inflammation, heart disease, and cancer. Phar-macol Rev 52:673–751.

46. Kobori M, Iwashita K, Shinmoto H, Tsushida T (1999) Phloretin-induced apoptosis inB16 melanoma 4A5 cells and HL60 human leukemia cells. Biosci Biotechnol Biochem63:719–725.

47. Hollman PC, Katan MB (1999) Dietary flavonoids: Intake, health effects and bioavail-ability. Food Chem Toxicol 37:937–942.

48. Crespy V, et al. (2001) Bioavailability of phloretin and phloridzin in rats. J Nutr131:3227–3230.

49. Edelstein ML, Abedi MR, Wixon J (2007) Gene therapy clinical trials worldwide to2007—an update. J Gene Med 9:833–842.

50. Drury JL, Mooney DJ (2003) Hydrogels for tissue engineering: Scaffold design variablesand applications. Biomaterials 24:4337–4351.

51. Ehrbar M, Schoenmakers R, Christen EH, Fussenegger M, Weber W (2008) Drug-sensinghydrogels for the inducible release of biopharmaceuticals. Nat Mater 7:800–804.

52. Kelm JM, Fussenegger M (2004) Microscale tissue engineering using gravity-enforcedcell assembly. Trends Biotechnol 22:195–202.

53. Tsao R, Yang R, Young JC, Zhu H (2003) Polyphenolic profiles in eight apple cultivarsusing high-performance liquid chromatography (HPLC). J Agric Food Chem 51:6347–6353.

54. Mullick A, et al. (2006) The cumate gene-switch: A system for regulated expression inmammalian cells. BMC Biotechnol 6:43.

55. Weber W, Fussenegger M (2009) Engineering of synthetic mammalian gene networks.Chem Biol 16:287–297.

56. Mitta B, et al. (2002) Advanced modular self-inactivating lentiviral expression vectorsfor multigene interventions in mammalian cells and in vivo transduction. Nucleic AcidsRes 30:e113.

57. Boukamp P, et al. (1988) Normal keratinization in a spontaneously immortalizedaneuploid human keratinocyte cell line. J Cell Biol 106:761–771.

58. Greber D, Fussenegger M (2007) Multi-gene engineering: Simultaneous expression andknockdown of six genes off a single platform. Biotechnol Bioeng 96:821–834.

59. Schlatter S, Rimann M, Kelm J, Fussenegger M (2002) SAMY, a novel mammalianreporter gene derived from Bacillus stearothermophilus alpha-amylase. Gene 282:19–31.

60. Fluri DA, Kemmer C, Daoud-El Baba M, Fussenegger M (2008) A novel system fortrigger-controlled drug release from polymer capsules. J Control Release 131:211–219.

61. Aubel D, et al. (2001) Design of a novel mammalian screening system for the detectionof bioavailable, non-cytotoxic streptogramin antibiotics. J Antibiot (Tokyo) 54:44–55.

62. Weber CC, et al. (2005) Broad-spectrum protein biosensors for class-specific detectionof antibiotics. Biotechnol Bioeng 89:9–17.

Gitzinger et al. PNAS � June 30, 2009 � vol. 106 � no. 26 � 10643

BIO

PHYS

ICS

AN

DCO

MPU

TATI

ON

AL

BIO

LOG

Y