www.sciencemag.org/content/354/6317/1296/suppl/DC1

Supplementary Materials for

β-cell–mimetic designer cells provide closed-loop glycemic control

Mingqi Xie, Haifeng Ye, Hui Wang, Ghislaine Charpin-El Hamri, Claude Lormeau, Pratik Saxena, Jörg Stelling,* Martin Fussenegger*

*Corresponding author. Email: fussenegger@bsse.ethz.ch (M.F.); joerg.stelling@bsse.ethz.ch (J.S.)

Published 9 December 2016, Science 354, 1296 (2016)

DOI: 10.1126/science.aaf4006

This PDF file includes:

Materials and Methods Supplementary Text Figs. S1 to S18 Tables S1 to S12 References

2

Materials and Methods

Vector Design

Comprehensive design and construction details for all expression vectors are provided

in Table S1. Some expression vectors were constructed by Gibson assembly using the

GeneArt® Seamless Assembly Cloning Kit (Obio Technology, Shanghai, China; cat. no.

BACR(C)20144001). Plasmids encoding KATP-subunits (pCMV Human SUR1 and pCMV6c

hKir6.2(BIR)) were kindly provided by Susumu Seino (Kobe University, Kobe, Japan).

Plasmids encoding Cav2.2-, Cav1.2- and Cav1.3-subunits (CaV1.3e[8a,11,31b,Δ32,42a],

CaV1.2, Cav2.2e[Δa10, Δ18a, Δ24a, 31a, 37a, 46], Cavb3 and CaVα2δ1) were kindly

provided by Diane Lipscombe (Brown University, RI, USA).

Cell Culture and Transfection

Human embryonic kidney cells (HEK-293T, ATCC: CRL-11268) were cultured in

Dulbecco’s modified Eagle’s medium (DMEM, Invitrogen) supplemented with 10% (v/v)

fetal bovine serum (FBS; Sigma-Aldrich, Buchs, Switzerland; cat. no. F7524, lot no.

022M3395) and 1% (v/v) penicillin/streptomycin solution (PenStrep; Biowest, Nuaillé,

France; cat. no. L0022-100) at 37°C in a humidified atmosphere containing 5% CO2. The

human 1.1E7 β-cell line (Sigma-Aldrich, cat. no. EC10070101) was cultured in RPMI 1640

medium (ThermoFisher Scientific, cat. no. 11875085) supplemented with 10% FBS and 1%

PenStrep. Human islets (HIR; Prodo Laboratories, Irvine, CA; lot. no. HP-16161-01) were

transferred from Prodo Transport medium (PIM(T)®; Prodo Laboratories; cat. no. PIM-

T001GMP) into CMRL medium (ThermoFisher Scientific; cat. no. 11530037) supplemented

with 10% FBS, 1% PenStrep, 1% ITS, 5mM D-Glucose, 2mM GlutaMAX, 1mM pyruvate,

10mM nicotinamide and 2.5mM HEPES, and cultivated for seven days prior to encapsulation

while changing fresh CMRL medium every third day. For passaging, cells of pre-confluent

HEK-293 and 1.1E7 cultures were detached by incubation in 0.05% Trypsin-EDTA (Life

Technologies, CA, USA; cat. no. 25300-054) for 3min at 37°C, collected in 10ml cell culture

medium, centrifuged for 3min at 290g and resuspended in fresh culture medium at standard

cell densities (1.5x105 cells / mL), before seeding into new tissue culture plates. Cell number

and viability were quantified using an electric field multichannel cell counting device (Casy

Cell Counter and Analyzer Model TT, Roche Diagnostics GmbH). For transfection, a solution

containing 2-3µg plasmid DNA and 6-9µg polyethyleneimine (PEI; Polysciences, Eppelheim,

Germany; cat. no. 24765-2) was incubated in 300µl serum- and antibiotics-free DMEM for

30min at 22°C and subsequently added dropwise to 3x105 cells seeded per well of a 6-well

plate. 12h after addition of PEI, transfected HEK-293 were detached by incubation in

Trypsin-EDTA, centrifuged (3min at 290g) and resuspended in low/no-glucose medium

(glucose-free DMEM [Life Technologies, CA, USA; cat. no. 11966-025] supplemented with

10% FBS, 1% PenStrep, 0-2mM D-glucose and 0.7mM CaCl2) and reseeded at a cell density

of 2x105 / mL. Unless stated otherwise, D-glucose or other control compounds were added to

transfected cells after cultivation under low/no-glucose conditions for another 12h.

Quantification of target gene expression

Expression levels of human placental secreted alkaline phosphatase (SEAP) in culture

supernatants were quantified according to a p-nitrophenylphosphate-based light absorbance

3

time course (22). SEAP levels in mouse serum were profiled using a chemiluminescence-

based assay (Roche Diagnostics GmbH, Mannheim, Germany; cat. no. 11 779 842 001).

Gaussia Luciferase (GLuc) levels in culture supernatants were profiled using the BioLux®

Gaussia Luciferase Assay Kit (New England Biolabs, Ipswich, MA; cat. no. E3300L). Human

insulin levels secreted by 1.1E7 cells and human islets were quantified using an Ultrasensitive

C-peptide ELISA kit (Mercordia, Uppsala, Sweden; cat. no. 10-1141-01). mINS levels in

culture supernatants and mouse serum were quantified with a Mouse Insulin ELISA kit

(Mercordia, Uppsala, Sweden; cat. no. 10-1247-01). Short human glucagon-like peptide 1

(shGLP-1) levels in culture supernatants were quantified with a Mouse IgG ELISA Kit

(Immunology Consultants Laboratory Inc., Portland, OR; cat. no. E-90G). Bioactive GLP-1

levels in mouse serum were quantified with a High Sensitivity GLP-1 Active ELISA Kit

(Merck Millipore, Schaffhausen, Switzerland; cat. no. EZGLPHS-35K). TurboGFP was

visualized by fluorescence microscopy using a Nikon Ti-E base Wide Field microscope

(Nikon AG, Egg, Switzerland) equipped with a Hammamatsu Orca Flash 4 digital camera, a

20×objective, a 488 nm/509 nm excitation and emission filter set and NIS Elements AR

software (version 4.3.0).

Generation of stable cell lines

The monoclonal HEK-293NFAT-SEAP cell line, transgenic for depolarization-stimulated

SEAP expression, was constructed by co-transfecting HEK-293 cells with a 20:1 (w/w)

mixture of pMX57 (PNFAT3-SEAP-pA) and pZeoSV2(+) (PSV40-zeo-pA), followed by selection

in culture medium containing 1mg/mL zeocin (Life Technologies, CA, USA; cat. no. R250-

05) and FACS-mediated single-cell cloning. Sixteen cell clones were selected and the best-in-

class HEK-293NFAT-SEAP1 was used for all follow-up studies.

The polyclonal HEKGLP1R population, transgenic for high-level GLP1R expression,

was constructed by cotransfecting 3x106

HEK-293 cells with 9500ng pMX250 (ITR-PhEF1α-

GLP1R-pA:PRPBSA-dTomato-P2A-PuroR-pA-ITR) and 500ng of the Sleeping Beauty

transposase expression vector pCMV-T7-SB100 (PhCMV-SB100X-pA, (38)). After selection

with 1µg/mL puromycin for two passages, the surviving population HEKMX250 was FACS-

sorted into three different subpopulations according to different red-fluorescence intensities.

The subpopulation with top 10% dTomato intensity HEKGLP1R showed highest sensitivity to

GLP-1 and was used for all follow-up studies.

The polyclonal HEKMX252 population, transgenic for stable expression of the α1D

subunit of Cav1.3 (Cacna1d), was constructed by co-transfecting 3x106 HEK-293 cells with

9500ng pMX252 (ITR-PhEF1α-Cacna1d-pA:PRPBSA-BFP-P2A-PuroR-pA-ITR) and 500ng

pCMV-T7-SB100 and selecting with 0.5µg/mL puromycin for two passages.

The polyclonal HEKCav1.3 population, transgenic for stable expression of the full

Cav1.3 channel componentry, was constructed by co-transfecting 3x106 HEKMX252 cells with

9500ng pMX251 (ITR-PhEF1α-Cacna2d1-P2A-Cacnb3-pA:PRPBSA-dTomato-P2A-BlastR-pA-

ITR) and 500ng pCMV-T7-SB100 and selecting with 10µg/mL of blasticidin for three

passages.

The monoclonal HEK-β cell line, transgenic for glucose-stimulated SEAP- and

insulin-expression, was constructed by cotransfecting 3x106 HEKCav1.3 cells with 9500ng

pMX256 (ITR-PNFAT5-SEAP-P2A-mINS-pA:PRPBSA-EGFP-P2A-ZeoR-pA-ITR) and 500ng

pCMV-T7-SB100. After selection with 100µg/mL zeocin for three passages, 5% of the

surviving population with highest EGFP expression levels were subjected to FACS-mediated

single-cell cloning. 50 cell clones were selected and clone no. 4 showing optimal glucose-

inducible insulin expression was used for all follow-up studies.

4

FACS-mediated cell sorting

HEK-293 cells expressing EGFP (488nm laser, 505nm long-pass filter, 530/30

emission filter) or dTomato (561nm laser, 570nm long-pass filter, 586/15 emission filter)

were sorted using a Becton Dickinson LSRII Fortessa flow cytometer (Becton Dickinson,

Allschwil, Switzerland) while excluding dead cells and cell doublets. Untreated HEK-293

cells or parental polyclonal populations were used as negative controls.

qRT-PCR

Total RNA of untreated HEK-293 cells was isolated using the ZR RNA MiniPrepTM

kit (Zymo Research, CA, USA; cat. no. R1064), treated with DNaseI (Thermo Scientific, cat.

no. EN0521) and cDNA was synthesized using the Applied Biosystems High Capacity cDNA

Reverse Transcription Kit (Life Technologies, CA, USA; cat. no. 4368814). For quantitative

analysis, PCR reaction (2min at 50°C, 20s at 95°C and 60 cycles of 1s at 95°C followed by

1min at 60°C) was performed on the Eppendorf Realplex2 Mastercycler (Eppendorf GmbH,

Hamburg, Germany) using the SYBR® Green PCR Master Mix (Life Technologies, CA,

USA; cat. no. 4309155) and the primers listed in Table S2 . The relative cycle threshold (CT)

was determined and normalized against the endogenous human glyceraldehyde 3-phosphate

dehydrogenase (GAPDH) gene.

Glucose-stimulated insulin secretion (GSIS)

Encapsulated human islets and 1.1E7 cells were washed (incubation for 30 min) in

0.25mL Krebs-Ringer Bicarbonate Buffer (Sigma-Aldrich, cat. No. K4002; 129mM NaCl,

5mM NaHCO3, 4.8mM KCl, 1.2mM KH2PO4, 1.2mM MgSO4, 2.5mM CaCl2, 10mM

HEPES, 0.1% BSA, pH7.4) and incubated for 30min in low-glucose (2.8mM) Krebs-Ringer

Bicarbonate Buffer. The culture was then switched to high-glucose (30mM) Krebs-Ringer

Bicarbonate Buffer for another 30min. The secreted isoform of the connecting peptide (C-

peptide) produced during proinsulin processing was quantified using the Ultrasensitive

Human C-peptide ELISA and the capsules were then transferred to fresh culture medium and

cultivated until the next GSIS assay.

Chemicals and Soft drinks

Acetic acid (cat. no. A6283), calcium chloride dihydrate (stock solution 0.5M in

ddH2O; cat. no. C7902), D-glucose (stock solution 1M in ddH2O; cat. no. G-7021), D-

mannitol (stock solution 0.1M in ddH2O; cat. no. M4125), D-mannose (stock solution 1M in

ddH2O; cat. no. M6020), D-galactose (stock solution 0.1M in ddH2O; cat. no. 48263), ethanol

(EtOH; cat. no. 02860), magnesium sulfate (MgSO4; cat. no. M2643), nicotinamide (stock

solution 0.5M in ddH2O; cat. no. N0636), potassium phosphate monobasic (K2HPO4; cat. no.

P5655), sodium bicarbonate (NaHCO3; cat. no. S5761), sucrose (stock solution 0.1M in

ddH2O; cat. no. S0389), D-maltose monohydrate (stock solution 0.1M in ddH2O; cat. no.

M9171), D-xylose (stock solution 0.1M in ddH2O; cat. no. X1500), L-glutamine (stock

solution 0.15M in ddH2O; cat. no. G3126), 3-(N-Morpholino)propanesulfonic acid (MOPS;

stock solution 0.1M in ddH2O; cat. no. M1254), palmitic acid (stock solution 0.02M in EtOH;

cat. no. P0500) and alloxan monohydrate (cat. no. A7413) were purchased from Sigma-

Aldrich (Buchs, Switzerland). Blasticidin S HCl (cat. no. A1113903), GlutaMAXTM

Supplement (cat. no. 35050061), Insulin Transferrin Selenium liquid media supplement (ITS;

cat. no. 41400045), N-2-hydroxyethylpiperazine-N-2-ethane sulfonic acid (HEPES; stock

5

solution 1M; cat no. 15630080), puromycin dihydrochloride (cat. no. A1113803), sodium

pyruvate (stock solution 100mM; cat no. 11360070) and ZeocinTM

selection reagent (cat. no.

R25005) were purchased from ThermoFisher Scientific (Reinach, Switzerland). Recombinant

human GLP-1 (stock solution 1mM in ddH2O; cat. no. 130-08; lot no. 0108358), IL-2 (stock

solution 10µM in 10mM aqueous acetic acid; cat. no. 200-02; lot no. 051512-1), IL-12 p70

(stock solution 1µM in DMEM; cat. no. 200-12; lot no. 0909S96) and IL-15 (stock solution

1µM in ddH2O; cat. no. 200-15; lot no. 061024) were purchased from PeproTech EC Ltd

(London, UK). D-fructose (stock solution 0.1M in ddH2O; cat. no. 161350010), L-leucine

(stock solution 0.01M in glucose-free DMEM; cat. no. 125121000) and linoleic acid (stock

solution 0.02M in EtOH; cat. no. 215040050) were purchased from Acros Organics (Geel,

Belgium). L-glucose anhydrous (stock solution 1M in ddH2O; cat. no. AB116919) was

purchased from abcr GmbH (Karlsruhe, Germany). Potassium chloride (KCl; stock solution

4M in ddH2O; cat. no. A3582) and sodium chloride (NaCl; stock solution 5M in ddH2O; cat.

no. A2942) were purchased from AppliChem (Darmstadt, Germany). Citric acid anhydrous

(stock solution 0.1M in ddH2O; cat. no. sc-211113) was purchased from Santa Cruz

Biotechnology (Dallas, USA). Poly(L-lysine) hydrobromide (cat. no. PLKB50) was

purchased from Alamanda Polymers (Alabama, USA). Trisodium citrate 2-hydrate (stock

solution 0.1M in ddH2O; cat. no. 6448) was purchased from Merck Millipore (Schaffhausen,

Switzerland). Bovine serum albumin (BSA; stock solution 10g/L; cat. no. B9000S) was

purchased from NEB Biolabs (Ipswich, MA). Streptozotocin (cat. no. 1621) was purchased

from Tocris Bioscience (Bristol, UK). Coca-Cola®

was purchased at local supermarkets,

degassed by extensive shaking and directly administered to mice (4 × 200 µl).

Animal experiments

The type-1 diabetes mouse model (T1D) was generated as described previously (21).

In brief, fasted mice (2x 18h/day) were injected with a single dose of freshly diluted alloxan

monohydrate (ALX; 200 mg/kg in 300µl phosphate buffered saline) and persistent fasting

hyperglycaemia (>20mM) developed after 48h. The type-2 diabetes mouse model (T2D) was

generated as described in (39, 40). In brief, fasted mice (20h/day) were injected with daily

doses of freshly diluted streptozotocin (STZ; 40mg/kg in 250µl ice-cold sodium citrate buffer

[pH 4.5, 0.01M, 0.11g/L NaCl]) for five consecutive days and chronic fasting hyperglycaemia

(>10mM) developed after 3 weeks. Glycaemia of mice was measured with a commercial

glucometer (Contour®

Next; Bayer HealthCare, Leverkusen, Germany; detection range: 0.5-

35mM). Intraperitoneal implants were produced by encapsulating transgenic HEK-293 cells,

1.1E7 cells or human islets into coherent alginate-poly-(L-lysine)-alginate beads (400µm; 500

cells or 1-10 IEQs per capsule) using an Inotech Encapsulator Research Unit IE-50R

(EncapBioSystems Inc., Greifensee, Switzerland) set to the following parameters: a 200-µm

nozzle with a vibration frequency of 1025Hz, a 25-mL syringe operated at a flow rate of 410

units and 1.12-kV voltage for bead dispersion (22). 5-9-weeks old female wild-type or

ALX/STZ-pretreated CD-1 Swiss albino mice (Janvier Labs, Le Genest-Saint-Isle, France)

were intraperitoneally injected with 1mL of glucose-free DMEM containing 1x104

microcapsules. Blood serum was isolated using microtainer serum separating tubes (SST)

according to the manufacturer’s instructions (centrifugation for 5 min at 10 000 x g; Becton

Dickinson, Plymouth, UK; cat. no. 365967). Most experiments involving animals were

performed according to the directive of the European Community Council (2010/63/EU),

approved by the French Republic and carried out by Ghislaine Charpin-El Hamri (No.

69266309; project No. DR2013–01 (v2)) and Marie Daoud-El Baba (No. 69266310; project

No. DR2013–01 (v2)) at the Institut Universitaire de Technologie, UCB Lyon 1, F-69622

6

Villeurbanne Cedex, France. Animal experiments related to Fig. 4B, 4C, 4D and 4G were

performed according to the protocol (Protocol ID: m20140301) approved by the East China

Normal University (ECNU) Animal Care and Use Committee and in direct accordance with

the Ministry of Science and Technology of the People's Republic of China on Animal Care

Guidelines.

Description of the Mathematical Model

Aims and architecture of the mathematical model

We developed the dynamic mathematical model for the following purposes:

1. helping the analysis of in vitro results by estimating the dynamics of unobservable internalstates;

2. predicting the implant’s long-term performance in vivo, which is difficult to assess inanimal experiments;

3. predicting the implant’s short-term performance in vivo, which is difficult to assess fortype 1 diabetic mouse models because glucose pulses would be lethal for these mice;

4. assessing the possible risks and advantages of the implant in different situations, by vary-ing in silico the types of diabetes, cell densities in the implant, and physiological condi-tions of the animals.

The model was adapted from a previous model developed for a similar implant, with de-signer cells that sense glycaemic state indirectly via pH and produce insulin accordingly (21).This previous model is based on ordinary differential equations and involves four modules: thedesigner cells synthetic circuit, the body insulin response, the body metabolism, and the bodypH. In the present work, glucose is directly sensed, which enables to keep the second module,and to replace the first module by a circuit-adapted representation. The architecture of the newmodel is described in Fig. S6. All equations relative to the body insulin response based on(41) were kept identical. In contrast, the designer cells synthetic circuit equations had to be sig-nificantly modified because the new circuit comprises both cell metabolism and ion transportmechanisms. Our new model for the synthetic circuit consists of the following compartments:

1. Cells: Intracellular compartment of designer cells;

2. Synthetic compartment: Medium in the dish for in vitro experiments, capsules for invivo experiments;

3. Blood: Compartment consisting of the mouse blood; its volume is set to zero for in vitroexperiments;

and the following modules in the synthetic compartment:

1. A minimal model of glycolysis (42) that summarizes core cell metabolism in three dy-namic equations, linking ATP to glucose concentration by lumping all intermediate metabo-lites into one variable, see equations (3) to (5). After adding equations for glucose uptake,the whole glycolysis module is comprised of equations (2) to (5).

2. A simplified model of ion transport across the membrane, linking intracellular calciumto ATP. The module includes equations (6) to (9). Most of the current-related equationswere based on a theoretical study of beta cell electrical activity (43). We estimated allparameters with our data sets since HEK-293 cells probably have different ion channelconductances than beta cells. Moreover, sodium currents, sodium-potassium pumps, andcalcium pumps endogenously present in HEK-293 cells were taken into account. Thecorresponding equations were extracted from an atrial cell model (44) and parameterswere estimated with our data set.

7

3. Calcium dependent gene expression, for which the same minimal model was used as in(21). The module includes equations (10) to (12).

4. For the HEK-βGLP system (Fig. 4E), an extended ODE system was used. All equations,parameters and rates relative to this extended model are noted with a *. The calciumdependent gene expression module was adjusted to GLP-1 production instead of Insulinor SEAP, and a new module was added for GLP-1-dependent gene expression. For theGLP-1 receptor GLP1R, we used the same minimal model for human GPCRs that coupleto cAMP signalling as in (21). In the new module, equation (10) is replaced by equations(10a*) to (10d*).

States of the model are defined in Table S4. Rates, currents, and fluxes are given in Ta-ble S5. The following tables summarize model parameters related to: general model properties(Table S6), metabolism (Table S7), ion transport (Table S8), gene expression (Table S9), and invivo conditions (Table S10).

Detailed description of new modules

This section will focus on the description of the glycolysis module and the ion transport module,which are the two new developments compared to the pH-sensor model from (21).

Glycolysis module. Glucose is either sensed in the cell culture dish, for in vitro experiments,or in the capsule, for in vivo experiments. For convenience, the capsule and the cell culture dishare called the synthetic compartment. Their volume and cell density are set to different valuesdepending on the type of experiment that is simulated, in vitro or in vivo.

Glucose in the synthetic compartment (concentration GS) decreases at the rate NSR1 of glu-cose uptake by the cells, where NS is the cell number and R1 the specific glucose uptake rate percell (equation (2)). The rate νGS accounts for diffusion of glucose from the blood to the syntheticcompartment in the case of in vivo experiments; it is set to zero for in vitro experiments.

Intracellular glucose increases correspondingly and it is immediately used to form phospho-rylated intermediate metabolites 6CP (42), at the rate R2 (equation (3)). This rate is dependenton several mechanisms such as allosteric inhibition of ATP consumption by ATP itself andbinding of ATP to phosphofructokinase.

Intermediate metabolites are then used to produce ATP with rate R3, which accounts forallosteric inhibition of pyruvate kinase flux by ATP. The factor

(1− A

Atot

)ensures that the ATP

concentration does not exceed Atot=5mM, the experimentally determined total concentrationof adenosine phosphates in the cell (43). In equation (4), q molecules of ATP are used byphosphofructokinase in R2 to enable intermediate metabolite production, and q+1 molecules ofATP are produced by pyruvate kinase with R3, resulting in the ATP production rate in equation(5). In order to keep our model as compact as possible, we used a simple Hill function to modelATP consumption by the cell.

Ion transport module. The ion transport module links ATP concentrations to calcium ionconcentrations. ATP and calcium are linked by the cell electrophysiology, which can be rep-resented by a Hodgkin-Huxley model of the cell membrane (see Fig. S7). Our model waslargely based on previous work by (43). In our model the resting membrane potential is mainlydetermined by outward currents of potassium through potassium channels, inward currents ofsodium through sodium channels and the sodium-potassium pumps that counteract these flows

8

by importing two potassium ions for each export of three sodium ions (see equations (6) and(7)). Potassium currents are partly due to channels sensitive to ATP concentrations. WhenATP concentrations are high, these channels close and outward potassium currents are lowered,which results in depolarization of the membrane (due to Kirchoff’s law, see equation 9). Thekey component of this module is the voltage gated calcium channel carrying the current ICaV .Intracellular calcium levels are normally kept low by a calcium pump. Upon depolarization,calcium will flow into the cell through calcium channels (equation 8).

The experimental data suggest that all components of the circuit are endogenously presentin HEK-293 cells, except the voltage-gated calcium channel, which upon transfection increasesKCl-induced (Fig. 1E) and glucose-induced calcium-dependent gene expression (Fig. 2).

Equations for calcium and potassium currents were found in (43). Expressions for sodium-potassium and calcium pumps were found in (44). ATP-sensitive potassium current in equation(25) was adapted from (43) by adding a Hill exponent. This exponent enabled to give moreflexibility to this equation and adapt it to our minimal, approximate model of glycolysis. Allparameters were estimated with our own experimental data since these models from literatureare related to beta cells and not HEK-293 cells. Moreover, we assume that ion transport acrossthe cell membrane does not have a significant influence on the external ion concentrations inthe synthetic compartment: dCaS

dt ≈dKSdt ≈

dNaSdt ≈ 0.

Simulation and parameter estimation

We used the same simulation approach as in (21): a single set of parameters was used forall simulations, except for parameters depending on the in vitro or in vivo setup, which wereswitched according to the specifications in Table S6, and input parameters such as KCl, glucoseinitial concentrations, or insulin production rates in the different mouse models.

Plasmid copy numbers can vary between experiments which were not performed on thesame day, depending on the transfection efficiency and the type of transfection (stable or tran-sient). We therefore allowed for experiment-specific plasmid copy numbers as listed in Ta-ble S11. Plasmid copy numbers are relative to SEAP copy number in the HEK-293NFAT-SEAP1cell line for HEK-β (see Fig. 2), and relative to GLP-1 receptor copy number in the HEKGLP1Rcell line for HEK-βGLP (see Fig. S16D-E), both with unity value. They were estimated for allexperiments according to the following assumptions:

1. Plasmid copy numbers always fulfill the transfection ratios of the experimental setup.

2. When Cav1.3 is not transfected, it is still present in the genome with always the same(low) estimated copy number of 0.4.

3. When Cav1.3 is transfected in lower than standard amounts, a linear scale was used be-tween 0.4 and the estimated copy number for 100% transfection to estimate the copynumber.

4. The SEAP (resp. GLP1R) copy number is identical for all experiments carried out withthe stable SEAP-expressing (resp. GLP1R expressing) cell line.

The same evolutionary strategy and objective function as in (21) were used for the estima-tion of parameters that were not available from literature. Briefly, we performed least-squaresminimization over all data and time points considered, with weights according to the inverse ofthe measurement standard deviation. To account for biological variability instead of technicalreplicate variances, the following minimal experimental errors were assumed:

9

1. Minimal s.d. of 1 U/L for in vitro SEAP assays

2. Minimal s.d. of 10 mU/L for in vivo SEAP assays

3. Minimal s.d. of 0.1 µg/L for in vivo insulin assays

4. Minimal s.d. of 1 mM for in vivo glucose assays

For parameter optimization, we employed all experimental data shown in the figures listedin Table S3, using the corresponding simulation protocols, except Fig. 4B,4C, 4D, 2C (33%and 66 % curves) which were not used for calibrating the model.

We fitted first all the in vitro data for the HEK-β system (Fig. 1E, 2C, 2D, S8B, S9) withcorresponding metabolic, ion transport and gene expression parameters from tables S9, S10 andS11, and then these parameters were fixed for further estimation of all other parameters withthe rest of the available data.

10

Equations of the Mathematical Model

Synthetic compartment

dNS

dt= µr(NS)NS (1)

dGS

dt=−NSR1(GS)+νGS (2)

dGSi

dt=

1VC

R1(GS)−R2(GSi,A) (3)

d6CPdt

= R2(GSi,A)−R3(6CP,A) (4)

dAdt

=−qR2(GSi,A)+(q+1)R3(6CP,A)−δ (A) (5)

dKSi

dt=−α

[IK(V,KSi)+ IKV (V,KSi)+ IKAT P(V,KSi,A)

]+2αINaK(NaSi) (6)

dNaSi

dt=−αINa(V,NaSi)−3αINaK(NaSi) (7)

dCaSi

dt=−α

2fCa

[ICa(V,CaSi)+ ICaV (V,CaSi)+ ICaP(V,CaSi)

](8)

dVdt

=− 1C

[IK(V,KSi)+ IKV (V,KSi)+ IKAT P(V,KSi,A)+ INa(V,NaSi) (9)

+ INaK(V,NaSi)+ ICa(V,CaSi)+ ICaV (V,CaSi)+ ICaP(V,CaSi)]

dMS

dt= nNFAT nSR4(CaSi)− (kdm +µr(NS))MS (10)

dSS

dt= ktlNSMS− kdsSS +νSS (11)

dIS

dt= FSCIktlNSMS− kdiIS +νIS (12)

11

Synthetic compartment for the HEK-βGLP model

Equations (1) to (9) and (11) to (12) were unchanged. Equation (10) was replaced by:

dMGLP1

dt= nNFAT nGR4(CaSi)− (kdm +µr(NS))MGLP1 (10a*)

dGLP1Sdt

= FSCGktlNSMG− kdGGLP1S+νGLP1S (10b*)

dGLP1Rdt

= nRkR− (kdR +µr(NS))GLP1R (10c*)

dMS

dt= nSR5(GLP1S)− (kdm +µr(NS))MS (10d*)

Growth and metabolic rates:

µr(NS) = µNmax

SNmax

S +NS(13)

R1(GS) =Vmax1GS

GS +Km1(14)

R2(GSi,A) =K2b−a

m2A Aa

K2bm2A +A2b

× Vmax2GSi

GSi +Km2G(15)

R3(6CP,A) =Vmax3×K2c

m36CPK2c

m3 +A2c×(

1− AAtot

)(16)

R4(CaSi) =Vmax4(CaSi)

d4

Kd4m4 +(CaSi)d4

(17)

δ (A) = uA2

K2δ+A2 (18)

R5(GLP1S) =Vmax5GLP1R(

kLCRE +(1− kLCRE)GLP1Sd5

GLP1Sd5 +Kd5m5

)(19*)

Currents and potentials:

12

VK(KSi) = 103 RTF

ln( KS

KSi

)(19)

VNa(NaSi) = 103 RTF

ln( NaS

NaSi

)(20)

VCa(CaSi) = 103 RTF

ln(CaS

CaSi

)(21)

IK(V,KSi) = gK(V −VK(KSi)) (22)

IKV (V,KSi) = gKV1

1+ e(−15−V )/5.61

1+ e(−43−V )/(−4.1)(V −VK(KSi)) (23)

IKAT P(V,KSi,A) = gKAT P1+(D/K1)

dK

1+(D/K1)dK +(A/K2)dK(V −VK(KSi)) (24)

D = Atot−A

INa(V,NaSi) = gNa(V −VNa(NaSi)) (25)

ICa(V,CaSi) = gCa(V −VCa(CaSi)) (26)

ICaV (V,CaSi) = nC ·gCaV (V −VCa(CaSi)) (27)

INaK(V,NaSi) = ImaxNaK

KS

kK +KS

NaSi

NaSi + kNa

V +150V +200

(28)

ICaP(V,CaSi) = ImaxCaP

CaSi

CaSi +2 ·10−4 (29)

In vivo model

dGdt

= gp− (gr + ksiIa)G−VS

VBνGS−

VB

VGνGI (30)

dGI

dt= νGI (31)

dIa

dt= I− kiarIa (32)

dIi

dt= FI(G− kiirIi) (33)

dIdt

= FI

[kpG+ kiIi + kdξ (I)

dGdt

]− VS

VBνIS− kirI (34)

dSdt

=−VS

VBνSS− kdsivS (35)

dGLP1dt

=−VS

VBνGLP1S− kdGivGLP1 (36*)

As for the in vivo insulin response module in (21), we used an auxiliary Hill function toprevent negative insulin concentrations in the case of rapid depletion of glucose at low insulin

13

levels:ξ (I) =

InHI

KnHIMI + InHI

Transfer fluxes:

νIS = γI(I− IS) (36)νGS = γG(G−GS) (37)νGI = DG(G−GI) (38)νSS = γS(S−SS) (39)

νGLP1S = γGLP1(GLP1−GLP1S) (40*)

14

Supplementary Figures

GLUT1 GLUT2 GLUT3 Kir6.2 SUR1 Kir6.1 SUR210 -7

10 -6

10 -5

10 -4

10 -3

10 -2

10 -1

100

qRT-PCR

Rel

ativ

e E

xpre

ssio

n (

GA

PD

H)

2 5 11 25 330

2

4

6

8

10

pMX57

Cav1.3/pMX57

T1R2/T1R3/pMX57

D-Glucose (mM)

SE

AP

(U

/L)

A) B)

Fig. S1Glucose-sensor control experiments. (A) Quantitative RT-PCR-based expression profiling ofendogenous glucose transporters and KATP-channels in HEK-293 using specific primers shownin Table S2. Transcription levels were normalized to glyceraldehyde 3-phosphate dehydroge-nase (GAPDH) transcripts by setting undetermined values to a maximum Ct of 40 cycles. Dataare mean ± SD, n=3. (B) Comparison of different glucose sensors in mammalian cells. HEK-293 cells were co-transfected with 1000ng pMX57 (PNFAT3-SEAP-pA) and 1000ng of eitherpcDNA3.1(+) , Cav1.3 (pCav1.3 / pCaVb3 / pCaVα2δ1; 333ng each) or the full sweet tastereceptor componentry (pT1R2 / pT1R3 / pGNAT3; 1:1:1, w/w). 24h after transfection andcultivation in low glucose medium (2mM), D-glucose was added to final concentrations as in-dicated on the x-axis. SEAP levels in the culture supernatants were scored at 48h after additionof D-glucose. Data presented are mean ± SD, n≥5.

15

0 5 8 11 18 25 33 40 47 540

2

4

6

8

10D-Glucose

L-Glucose

D-Mannitol

Concentration (mM)

SE

AP

(U

/L)

0 5 11 250

2

4

6

8 D-Glucose

D-Fructose

D-Galactose

Sucrose

D-Maltose

D-Xylose

Concentration (mM)

SE

AP

(U

/L)

A) B)

C)

2mM

D-G

lc

40m

M D

-Glc

6mM

L-G

ln

10m

M L

-Gln

5mM

L-L

eu

10m

M L

-Leu

0.1m

M P

A

0.1m

M L

A

0

2

4

6

8

SE

AP

(U

/L)

0 0.25 0.5 1 2 5 10 KCl0

25

50

75

100IL-2

IL-12

IL-15

Cav1.3 / PNFAT3

Concentration (nM)

SE

AP

(U

/L)

0 0.25 0.5 1 2 5 10 KCl0.0

2.5

5.0

7.5

10.0

12.5

15.0

17.5IL-2

IL-12

IL-15

PNFAT3 only

Concentration (nM)

SE

AP

(U

/L)

D) E)

Fig. S2Characterization of Cav1.3/PNFAT3-constituted excitation-transcription coupling systems. (A-B) Substrate specificity of Cav1.3/pMX57-transgenic mammalian cells. HEK-293 cells werecotransfected with Cav1.3 (pCav1.3 / pCaVb3 / pCaVα2δ1; 1:1:1, w/w) and pMX57 (PNFAT3-SEAP-pA), and the cells were cultured in glucose-free medium for 12 h before different (A) glu-cose isomers and osmotic controls or (B) nutritional sugar compounds were added. 48h after theaddition of control compounds, the SEAP levels in the culture supernatants were scored. Datapresented are mean ±SD, n≥3. (C) Substrate specificity of Cav1.3/pMX57-transgenic mam-malian cells. HEK-293 cells were co-transfected with Cav1.3 (pCav1.3 / pCaVb3 / pCaVα2δ1;333ng each) and pMX57 (PNFAT3-SEAP-pA; 1000ng) and cultivated in glucose-free medium for12h before different concentrations of D-glucose (D-Glc), glutamine (L-Gln), leucine (L-Leu),palmitic acid (PA) or linoleic acid (LA) were added. 48h after addition of control compounds,SEAP levels in the culture supernatants were scored. Data presented are mean ± SD, n≥3.(D, E) Insensitivity of Cav1.3/pMX57-transgenic HEK-293 cells to cytokine signaling. HEK-293 cells were cotransfected with pMX57 (PNFAT3-SEAP-pA; 1000ng) and either (D) Cav1.3(pCav1.3 / pCaVb3 / pCaVα2δ1; 333ng each) or (E) pcDNA3.1(+) (1000ng) and cultivated inglucose-free medium for 12h before different concentrations of recombinant human interleukin2 (IL-2), interleukin 12 (IL-12) and interleukin 15 (IL-15) or potassium chloride (40mM KCl;positive control) were added and SEAP levels were profiled in the culture supernatants after48h. Data presented are mean ±SD, n≥3.

16

0 12 24 36 48 600

1

2

3

4

Time (h)

GLu

c (1

06 x R

LU)

2mM 5mM 40mM KCl

Fig. S3Glucose-stimulated GLuc expression kinetics. HEK-293 cells were cotransfected with Cav1.3(pCav1.3 / pCaVb3 / pCaVα2δ1; 333ng each) and pWH29 (PNFAT3-GLuc-pA; 1000ng) andcultivated in low-glucose medium (2mM) for 12h before different concentrations of D-Glucose(5 or 40mM) or KCl (40mM) were added. GLuc levels in the culture supernatants were profiledat different time points after the addition of inducer compounds as indicated on the x-axis. Datapresented are mean ±SD, n≥3.

Fig. S4Glucose-stimulated tGFP expression kinetics. Right: Fluorescence micrographs profiling rep-resentative TurboGFP expression in HEK-293 cells cotransfected with Cav1.3 and pFS119(PNFAT3-TurboGFP:dest1-pA) and cultured in medium containing different concentrations of D-glucose (D-Glc) or potassium chloride (KCl). Left: Control cells transfected with pcDNA3.1(+)and pFS119.

17

1 2 3 4 5 6 7 8 9 101112131415160

5

10

15

Clone no.

SEA

P (U

/L)

1 3 6 80

5

10

15

Passage number

SEA

P (U

/L)

HEK-293NFAT-SEAP1

A)

B)

0mM KCl 50mM KCl

Fig. S5Design and construction of the stable HEK-293NFAT-SEAP1 cell line. (A) Depolarization-stimulated SEAP expression of different PNFAT-IL4-transgenic cell clones (HEK-293NFAT-SEAP1).HEK-293 cells were stably transfected with pMX57 (PNFAT3-SEAP-pA) and 16 randomlyselected cell clones were profiled for their depolarization-stimulated SEAP regulation per-formance by cultivating them for 48h in the presence (50mM) or absence (0mM) ofpotassium chloride (KCl). (B) Stable depolarization-stimulated SEAP expression of theHEK-293NFAT-SEAP1 cell line. 5x 104 HEK-293NFAT-SEAP1 cells from different generations werecultivated for 48h in the presence (50mM) or absence (0mM) of potassium chloride (KCl) be-fore SEAP levels in the culture supernatants were scored. All data presented are mean ± SD,n≥3.

18

Fig. S6Architecture of the mathematical model. In vitro and in vivo conditions are represented by de-signer cells being placed into a synthetic compartment representing the culture dish (left) and acapsule compartment for the implant (right), respectively. The designer cell model (bottom) isidentical in both cases, comprising representations of glucose input and metabolism, signal pro-cessing by ATP-dependent electrophysiological responses, and calcium-dependent expressionof SEAP or insulin as output.

19

Fig. S7Hodgkin-Huxley representation of the ion transport module. Cm is the membrane capacitance.For every ion i, Ii is the corresponding outward background current, IiV is the correespondingoutward voltage-dependent current, Vi is the reversal potential of this ion. IKATP is the outwardATP-dependent potassium current. INaK is the outward sodium-potassium pump current andICaP the outward calcium pump current.

20

Fig. S8Cav1.3/ HEK-293NFAT-SEAP1 control experiments. (A) Time-delayed glucose responsiveness.Cav1.3-transgenic Cav1.3/HEK-293NFAT-SEAP1 cells were cultured in low-glucose medium(2mM) for 0-36 h before D-glucose was added at the indicated final concentrations. 48h afterthe addition of D-glucose, the SEAP levels in the culture supernatants were profiled. Data pre-sented are mean± SD, n≥3. (B) Reversibility of the synthetic excitation-transcription couplingsystem. Cav1.3-transgenic Cav1.3/HEK-293NFAT-SEAP1 cells were cultured in high-D-glucosemedium (40mM) or low-D-glucose medium (5mM) for 72h while resetting the cell density to0.75x106 cells/mL and alternating D-glucose concentrations every 24h followed by extensivewashing over 12h. The SEAP levels in the culture supernatants were profiled every 12h within24h intervals of exposure to high/low glucose. Solid and dashed curves show correspondingmodel-based simulations. Data presented are mean ± SD, n≥4.

21

0 20 40 60 80 1000

20

40

60

80

100

KCl (mM)

SEA

P (U

/L)

- Cav1.3+ Cav1.3

0 12 24 36 48 600

5

10

15

20

Time (h)

SEA

P (U

/L)

2mM Glucose33mM Glucose

0 5 10 15 20 25 30 350

2

4

6

8

10

12

D-Glucose (mM)

SEA

P (U

/L)

- Cav1.3+ Cav1.3

0 12 24 36 48 600

20

40

60

80

Time (h)

SEA

P (U

/L)

Transient Stable40mM KCl0mM KCl

A) B)

C) D)

Fig. S9Experimental data (symbols) and simulation results (lines) for D-glucose- and KCl-stimulatedSEAP expression in vitro. (A) Cav1.3-dependent excitation-transcription coupling. HEK-293cells were co-transfected with pMX57 and either pcDNA3.1(+) (-Cav1.3) or Cav1.3 (+Cav1.3),and the cells were grown for 48 h in cell culture medium containing different KCl concen-trations before SEAP levels in the culture supernatants were profiled (see also Fig. 1E). Alldata presented are mean ± SD, n≥5. (B) D-glucose activated PNFAT3-activation. HEK-293cells were co-transfected with Cav1.3 and pMX57, and the cells were cultured in low-glucosemedium (2 mM) for 12 h before D-glucose was added to the indicated final concentrations.Forty-eight hours after the addition of control compounds, the SEAP levels in the culture super-natants were scored. Data presented are mean ± SD, n≥3. (C, D) SEAP expression kinetics.Twelve hours after transfection of 9 x 105 HEK-293 cells with pMX57 (PNFAT3-SEAP-pA)and Cav1.3 (pCav1.3/pCaVb3/pCaVα2δ1; 1:1:1, w/w), culture supernatants were exchangedby fresh medium containing different D-glucose- (C) and KCl- (D) concentrations. SEAP lev-els in the culture supernatants were profiled every 12 h. Grey color represents equally treatedCav1.3-transgenic HEK-293NFAT-SEAP1 cells. Data presented are mean ± SD, n≥3

22

0 30 60 90 1200

5

10

15

20

Time (h)

Gly

cem

ia (m

M)

WT (5x106 cells)Experiment

Fig. S10Model simulations for glucose tolerance in healthy WT-mice. Forty-eight hours after implanta-tion of 5 x 106 Cav1.3 transgenic HEK-293NFAT-SEAP1 cells, CD-1 Swiss albino mice received anintraperitoneal injection of aqueous 2 g/kg D-glucose, and the glycaemic profile of each animalwas tracked every 30 min. The orange curve shows a corresponding model-based simulation.All data shown as the mean ± SD, n = 8 mice.

0 12 24 36 48 60 720

100

200

300

Time (h)

SEA

P (m

U/L

)

WT (- Cav1.3)

WT (+ Cav1.3)

T1D (- Cav1.3)

T1D (+ Cav1.3)

Fig. S11Cav1.3-dependent SEAP-expression kinetics in vivo. Wild-type (WT, black) or type 1 dia-betic (T1D, orange) mice were implanted with 1 x 104 microencapsules (500 cells /capsule)containing HEK-293NFAT-SEAP1 cells transfected with either pcDNA3.1(+) (-Cav1.3) or Cav1.3(+Cav1.3), and SEAP levels in the bloodstream were quantified every 24h (opaque or hollowcircles). Solid or dashed curves show corresponding model-based simulations. The data areshown as the mean ± SD, n=8 mice.

23

5mM Glc 40mM Glc 40mM KCl0

10

20

30

40

50

mou

se F

c (n

g/m

L)

PNFAT3-shGLP1PNFAT4-shGLP1PNFAT5-shGLP1

BDLBDLBDL BDL

5mM Glc 40mM Glc 40mM KCl0.00.40.81.21.62.02.42.83.23.64.0

mIN

S (µ

g/L)

PNFAT3-mINSPNFAT4-mINSPNFAT5-mINS

BDLBDL BDLBDL BDL

A) B)

Fig. S12Optimization of the PNFAT-IL4-promoter for glucose- and depolarization-stimulated (A)shGLP1- and (B) mINS-expression. (A) HEK-293 cells were co-transfected with 1000ngCav1.3 and 1000ng of pMX61 (PNFAT3-shGLP1-pA), pMX117 (PNFAT4-shGLP1-pA) orpMX115 (PNFAT5-shGLP1-pA) and cultivated in low-glucose medium (2mM) for 12h beforedifferent concentrations of D-glucose (Glc) or potassium chloride (KCl) were added. 48h afteraddition of control compounds, murine IgG levels in the culture supernatants were quantified(BDL: below detection limit). (B) HEK-293 cells were co-transfected with 1000ng Cav1.3and 1000ng of pMX68 (PNFAT3-mINS-pA), pMX99 (PNFAT4-mINS-pA) or pMX100 (PNFAT5-mINS-pA) and cultivated in low-glucose medium (2mM) for 12h before different concentrationsof D-glucose (Glc) or potassium chloride (KCl) were added. 48h after addition of control com-pounds, murine insulin levels in the culture supernatants were quantified (BDL: below detectionlimit). All data presented are mean ± SD, n≥3.

24

0 24 48 72 960.0

0.2

0.4

0.6

0.8

1.0

Time (h)

Blo

od In

sulin

(g/

L)

0 30 60 90 1200

5

10

15

20

25

30

35

Time (min)

Gly

cem

ia (m

M)

0 24 48 72 96 120 144 168 192 216 2400

5

10

15

20

25

30

35

Time (h)

Gly

cem

ia (m

M)

0 24 48 72 96 120 144 168 192 216 2400.0

0.2

0.4

0.6

0.8

1.0

1.2

Time (h)

Blo

od In

sulin

(g/

L)

0 24 48 72 960

5

10

15

20

25

30

35

Time (h)

Fast

ing

Gly

cem

ia (m

M)

****

**

0 30 60 90 1200

1

2

3

Time (min)

Blo

od In

sulin

(g/

L)

0 24 48 72 96 120 144 168 192 216 2400

5

10

15

20

25

Time (h)

Gly

cem

ia (m

M)

0 24 48 72 96 120 144 168 192 216 2400.0

0.2

0.4

0.6

0.8

1.0

1.2

Time (h)

Blo

od In

sulin

(g/

L)

A) B)

C) D)

E) F)

WT (INS)

T1D (INS)

T1D (GLP-1)

- implant + implant 0.1 %1.6 %3.2 %6.3 %

13 % 25 % 50 %100 %

native insulin production

0 2.5 5 7.5 10 12.5 15 17.5 20 22.5 25----------- Number of transgenic Cav1.3/pMX100 cells ( x 106) ----------

Fig. S13Self-sufficient glucose-induced insulin expression in type-1 diabetic mice. (A) Self-sufficientinsulin expression in wild-type and type-1 diabetic mice. HEK-293 cells were cotransfectedwith Cav1.3 and pMX100 (PNFAT5-mINS-pA), and the cells were then microencapsulated into

25

alginate-poly-(L-lysine)-alginate beads (INS). Control implants consisted of equally encapsu-lated Cav1.3/pMX115-transgenic HEK-293 cells (GLP-1). Capsules (1x104; 500 cells/capsule)were implanted into wild-type (WT) or type-1 diabetic (T1D) mice. Serum insulin levels (de-tection limit: 0.2µg/L) were profiled at 72h after capsule implantation and 4h after food intake,with corresponding model simulations shown as lines. (B) Self-sufficient glycemic controlof Cav1.3/pMX100-transgenic implants. The fasting glycemia of the same groups of mice asin (A) was tracked for 96h after capsule implantation. All data in (A-B) are shown as themean ±SD, and the statistical analysis was performed with a two-tailed t-test (n=8 mice).*P<0.05, **P<0.01, ***P<0.001 vs. control. Orange curves show corresponding model-basedsimulations. (C-E) Model-based predictions of long-term effects on glucose homeostasis ofCav1.3/pMX100-based T1D-treatments. (C) (Left panel) Intraperitoneal glucose tolerance test(IPGTT) of T1D mice after INS-capsules-based treatment of life-threatening hyperglycemia.Solid curves show predicted glucose tolerance of mice implanted with 1x104 capsules contain-ing Cav1.3/pMX100-transgenic HEK-293 cells (500 cells/capsule). Dashed curves show thecorresponding behaviors without implants. Colours indicate simulated percentages of nativeinsulin levels (right panel) ranging from 0% (T1D) to 100% (WT). Physiological ranges areindicated by red background colour, (D and E) Robustness of Cav1.3/pMX100-based therapy.Solid curves show predicted profiles of glycemia (top panels) and blood insulin levels (bottompanels) in T1D mice implanted with different amounts of capsules containing Cav1.3/pMX100-transgenic HEK-293 cells (500 cells/capsule) (see colour code at the top of the figure). (D)Simulated scenario of the body’s glycemia and blood insulin response to a complete restora-tion of native insulin production (from 0% to 100%) at 120h after implantation. (E) Simulatedscenario of the body’s glycemia and blood insulin response to a complete loss of native insulinproduction (from 100% to 0%) at 120h after implantation.

26

2 5 11 25 40 KCl0

20

40

60

80

Glucose (mM)

SEA

P (U

/L)

0% (w/w) pMX2510.05% (w/w) pMX2515% (w/w) pMX25110% (w/w) pMX251

2 5 11 25 40 KCl0

10

20

30

40

D-Glucose (mM)

SEA

P (U

/L)

95% (w/w) pMX57 in HEKCav1.3

95% (w/w) pMX57 in HEK-29395% (w/w) pMX57 + 5% pMX252 in HEK-293

1 2 3 4 5 6 7 8 9 10111213141516171819202122232425262728293031323334353637383940414243444546474849500

200

400

600

Clone no.

SEA

P (U

/L)

5mM40mM

2 4 6 8 11 14 15 20 26 27 30 34 35 36 39 42 43 48 49 500

2

4

6

8

Clone no.

mIN

S (µ

g/L)

5mM40mM

2 5 11 25 40 KCl0

10

20

30

40

Glucose (mM)

SEA

P (U

/L)

0% (w/w) pMX2510.5% (w/w) pMX251

5 40 KCl0.0

0.5

1.0

1.5

2.0

2.5

mIN

S (µ

g/L)

50% (w/w) pMX25666.7% (w/w) pMX256100% (w/w) pMX256

A) B)

C) D)

E)

F)

Fig. S14Construction and characterization of the stable HEK-β cell line. (A, B) Characterization of

27

HEKMX252 stably expressing the Cav1.3 α1D subunit. (A) 3x106 HEK-293 cells were cotrans-fected with pMX252 (ITR-PhEF1α-Cacna1d-pA:PRPBSA-BFP-P2A-PuroR-pA-ITR; 9500ng) andpCMV-T7-SB100 (PhCMV-SB100X-pA; 500ng), selected with 0.5µg/mL puromycin for twopassages and 3x105 cells of the surviving population (HEKMX252) were then cotransfected withpMX57 (PNFAT3-SEAP-pA; 1000ng) and different amounts of pMX251 (ITR-PhEF1α-Cacna2d1-P2A-Cacnb3-pA:PRPBSA-dTomato-P2A-BlastR-pA-ITR; 0-200ng filled to 1000ng withpcDNA3.1(+)). 24h after transfection and cultivation in low glucose medium (2mM), D-glucoseor potassium chloride (KCl; 50mM) was added to the indicated final concentrations. SEAP lev-els were profiled in the culture supernatants 48h after addition of D-glucose. Data presented aremean ±SD, n≥5. (B) Control experiment of HEK-293 cells transfected with pMX57 (1000ng)and different amounts of pMX251 (0-10ng filled to 1000ng with pcDNA3.1(+)). Data presentedare mean±SD, n≥5. (C) Characterization of HEKCav1.3 stably expressing the full Cav1.3 chan-nel componentry (Cacna1d/Cacnb3/Cacna2d1). 3x106 HEKMX252 cells were cotransfected withpMX251 (9500ng) and pCMV-T7-SB100 (500ng), selected with 10µg/mL blasticidin for threepassages and 3x105 cells of the surviving population (HEKCav1.3) were then cotransfected withpMX57 (PNFAT3-SEAP-pA; 1900ng) and pcDNA3.1(+) (100ng). HEK-293 cotransfected witheither pMX57 alone (1900ng) or in combination with pMX252 (100ng) were used as negativecontrols. 24h after transfection and cultivation in low glucose medium (2mM), D-glucose orpotassium chloride (KCl; 50mM) was added to the indicated final concentrations. SEAP expres-sion levels were profiled in the culture supernatants 48h after the addition of D-glucose. Datapresented are mean ±SD, n≥5. (D) Glucose- and depolarization-stimulated insulin expressionin HEKCav1.3. 3x105 HEKCav1.3 cells were cotransfected with different amounts of pMX256(ITR-PNFAT5-SEAP-P2A-mINS-pA:PRPBSA-EGFP-P2A-ZeoR-pA-ITR, 1000-2000ng filled to2000ng with pcDNA3.1(+)) and cultivated in low glucose medium (2mM) for 12h before dif-ferent concentrations of D-glucose and KCl were added. 48h after the addition of the controlcompounds, mINS levels were profiled in the culture supernatants. Data presented are mean±SD, n≥3. (E, F) Clonal selection of HEK-β cells. (E) 3x106 HEKCav1.3 cells were cotrans-fected with pMX256 (9500ng) and pCMV-T7-SB100 (500ng), selected with 100µg/mL zeocinfor three passages and 5% of the surviving population showing highest EGFP expression levelswere subjected to FACS-mediated single-cell cloning. 50 expanded cell clones were profiled forglucose-stimulated SEAP expression by cultivating 5x104 cells in high-glucose (40mM) or low-glucose medium (5mM) for 48h before SEAP levels were profiled in the culture supernatants.Data presented are mean ±SD, n=3. (F) The 20 clones showing highest glucose-stimulatedSEAP inductions in (E) were profiled for glucose-stimulated insulin expression by cultivating5x104 cells in high-glucose (40mM) or low-glucose medium (5mM) for 48h before mINS lev-els were profiled in the culture supernatants. HEK-β (cell clone no. 4) was chosen for furtheranalysis. Data presented are mean ±SD, n=3.

28

0 12 24 36 480.00.40.81.21.62.02.42.83.23.64.0

Time (h)

mIN

S (

g/L)

5mM 40mM

2 5 11 25 33 400.0

0.5

1.0

1.5

2.0

D-Glucose (mM)

mIN

S (

g/L)

1st 2nd 3rd

0.0

0.5

1.0

1.5

2.0

2.5

Week after Encapsulation

Insu

lin (

g/L)

Human Islets (2000IEQ)1.1E7 (5mio.)HEK- (5mio.)

- - - - - - - - - +++++++++

******

***

A) B) C)

BD BD BD BD BD BDBD BD

Fig. S15Characterization of the monoclonal HEK-β cell line. (A) 3x104 HEK-β cells were cultivated inhigh-glucose (40mM) or low-glucose medium (5mM) for 48h and mINS levels were profiled inthe culture supernatants every 12h after addition of D-glucose. Data presented are mean ±SD,n≥5. (B) 5x104 HEK-β cells were cultivated in low-glucose medium (2mM) for 12h, beforedifferent concentrations of D-glucose were added and mINS levels were profiled in the culturesupernatants after 24h. Data presented are mean ±SD, n≥5. (C) Reversible glucose-stimulatedinsulin secretion. Identical capsule batches used for implantation into mice (human islets, Fig.S18; HEK-β and 1.1E7, Fig. 4B) were also maintained in cell culture medium for 3 weeks andglucose-stimulated (-, 2.8mM; +, 30mM) insulin production was profiled for 24h once everyweek. (BD: Below Detection limit) Data presented are mean ±SD, statistics were performedusing two-tailed t-test (n=3 independent experiments). *P<0.05, **P<0.01, ***P<0.001 HEK-β vs. 1.1E7.

29

0 0.01 0.1 1 100

50

100

150

200

GLP-1 (nM)

SEA

P (U

/L)

0.5% (w/w) pMX25010% (w/w) pMX25025% (w/w) pMX25050% (w/w) pMX250

0 0.1 0.5 1 50

5

10

15

20

GLP-1 (nM)

SEA

P (U

/L)

HEKGLP1R

HEKGLP1Rmedium

HEKGLP1Rlow

0.001 0.01 0.1 10

10

20

30

40

50

GLP-1 (nM)

SEA

P (U

/L)

no Cav1.3, no shGLP1no Cav1.3no shGLP1

Cav1.3 + shGLP1 + GLP1R

no GLP1R

0 2 4 6 8 10 12 14 16 18 20 22 240

10

20

30

Time (h)

SEA

P (U

/L)

0pM50pM

0 0.01 0.05 0.1 10

10

20

30

GLP-1 (nM)

SEA

P (U

/L)

pCK53 in HEKGLP1R

pCK53 in HEK-293

8 16 24 32 400

5

10

15

D-Glucose (mM)

SEA

P (U

/L)

no Cav1.3, no shGLP1no Cav1.3no shGLP1

Cav1.3 + shGLP1 + GLP1R

no GLP1R

0 2 4 6 8 10 12 14 16 18 20 22 240.0

0.5

1.0

1.5

2.0

2.5

Time (h)

mIN

S (µ

g/L)

0pM50pM

A) B)

C) D)

E) F)

G) H)

0 12 24 36 48 60 720

5

10

15

Time (h)

SEA

P (U

/L)

5mM40mM

Fig. S16Engineering of HEK-βGLP. (A) GLP-1 triggered SEAP expression in HEK-293 cells. 3x105

HEK-293 cells were cotransfected with pCK53 (PCRE-SEAP-pA; 200ng) and different amountsof pMX250 (ITR-PhEF1α-GLP1R-pA:PRPBSA-dTomato-P2A-PuroR-pA-ITR; 10-1000ng filledto 1800ng with pcDNA3.1(+)) before different concentrations of recombinant human GLP-1was added. 24h after the addition of GLP-1, SEAP levels were profiled in the culture

30

supernatants. Data presented are mean±SD, n≥5. (B, C) Characterization of HEKGLP1R stablyexpressing the human GLP-1 receptor (GLP1R). (B) 3x106 HEK-293 cells were cotransfectedwith pMX250 (9500ng) and pCMV-T7-SB100 (500ng), selected with 1µg/mL puromycin fortwo passages and the surviving population was FACS-sorted into three populations with dif-ferent red-fluorescence intensities (HEKGLP1R, HEKGLP1Rmedium, HEKGLP1Rlow). Each pop-ulation (1x105 cells) was transfected with pCK53 (100ng filled to 2000ng with pcDNA3.1(+))before different concentrations of recombinant human GLP-1 were added. 24h after the additionof GLP-1, SEAP levels were profiled in the culture supernatants. Data presented are mean±SD,n≥3. (C) HEK-293 cells transfected with pCK53 (100ng filled to 2000ng with pcDNA3.1(+))were used as negative control. Data presented are mean ±SD, n≥5. (D, E) Validation of theHEK-βGLP circuit. HEKGLP1R was cotransfected with Cav1.3 (pCav1.3/pCaVb3/pCaVα2δ1;333ng each), pMX61 (PNFAT3-shGLP1-pA; 1000ng) and pCK53 (PCRE-SEAP-pA; 250ng) andcultivated in low-glucose medium (2mM) for 12h before different concentration of (D) recombi-nant human GLP-1 or (E) D-glucose were added. HEKGLP1R cotransfected with pMX61/pCK53or Cav1.3/pCK53 and HEK-293 cotreansfected with Cav1.3/pMX61/pCK53 were used as neg-ative controls. (D) 24h after addition of GLP-1 and (E) 72h after addition of D-Glucose,SEAP levels were profiled in the culture supernatants. Data presented are mean ± SD, n≥5.Dashed curves show corresponding model-based simulations. (F, G) SEAP expression kinetics.HEKGLP1R cells were cotransfected with Cav1.3 (pCav1.3/pCaVb3/pCaVα2δ1; 333ng each),pMX61 (PNFAT3-shGLP1-pA; 1000ng) and pCK53 (PCRE-SEAP-pA; 250ng), cultivated in low-glucose medium (2mM) for 12h before different concentrations of (F) recombinant human GLP-1 or (G) D-glucose were added. SEAP levels were profiled in the culture supernatants (F) 24hor (G) 72h after addition of the respective compounds. Data presented are mean ±SD, n≥5.Dashed curves show corresponding model-based simulations. (H) mINS expression kineticsof HEK-βGLP. HEKGLP1R cells were cotransfected with Cav1.3 (pCav1.3/pCaVb3/pCaVα2δ1;333ng each), pMX61 (PNFAT3-shGLP1-pA; 1000ng) and pDA145 (PCRE-mINS-pA; 1000ng),cultivated in low-glucose medium (2mM) for 12h before different concentrations of recombi-nant human GLP-1 were added. mINS levels were profiled in the culture supernatants for 24hafter addition of control compounds. Data presented are mean ±SD, n≥5. Dashed curves showcorresponding model-based simulations.

31

Fig. S17Model-based analysis of HEK-βGLP’s effect on oral glucose tolerance. Solid curves showmodel-based simulations of blood (A) GLP-1, (B) insulin and (C) glucose levels of T1D miceimplanted with HEK-βGLP during the same oral glucose tolerance test shown in Fig. 4G. Cor-responding experimental data of glycemia (Fig. 4G) are shown as black spheres.

32

0 15 30 60 90 1200

10

20

30

40

Time (min)

Gly

cem

ia (

mM

)OGTT (7days after implantation)

mouse 1mouse 2

mouse 3mouse 4

0 15 30 60 90 1200

10

20

30

40

Time (min)

Gly

cem

ia (

mM

)

OGTT (14 days after implantation)

Fig. S18Oral glucose tolerance test (OGTT) of type-1 diabetic mice treated with encapsulated humanislets. 2000 IEQs of human islets were microencapsulated in alginate-poly-(L-lysine)-alginatebeads and injected into each of four type-1 diabetic mice. 7 and 14 days after implantation theanimals received oral D-glucose (2g/kg) and their glycemic excursions were recorded over 2h.

33

34

Table S1 Plasmids used and designed in this study.

Plasmid Description and Cloning Strategy Reference or Source

pCaV1.2 pcDNA3.1(+)-derived constitutive Cacna1c expression vector (PhCMV-Cacna1c-pA) (Addgene no. 26572). (45)

pCaV1.3 pcDNA3.1(+)-derived constitutive Cacna1d expression vector (PhCMV-Cacna1d-pA) (Addgene no. 26576). (46)

pCavb3 pcDNA3.1(+)-derived constitutive Cacnb3 expression vector (PhCMV-Cacnb3-pA) (Addgene no. 26574). Prof. Lipscombe Lab

(unpublished)

pCaVα2δ1 pcDNA3.1(+)-derived constitutive Cacna2d1 expression vector (PhCMV-Cacna2d1-pA) (Addgene no. 26575). (47)

pCav2.2 pcDNA3.1(+)-derived constitutive Cacna1b expression vector (PhCMV-Cacna1b-pA) (Addgene no. 26569). (48)

pcDNA3.1(+) Mammalian expression vector (PhCMV-MCS-pA). Life Technologies, CA

pcDNA3.2/v5-

DEST

Mammalian Gateway®

-compatible destination vector (PhCMV-attR1-Cmr-ccdB-attR2-pA). Life Technologies, CA

pcDNA3.2/v5-

DEST-hGlut2

pcDNA3.2/v5-DEST containing a constitutive hGlut2 expression unit (PhCMV-hGlut2-pA) (Addgene no. 18086). (49)

pCK53 PCRE-driven SEAP-expression vector (PCRE-SEAP-pA). (21)

pCMV6 Mammalian expression vector (PhCMV-MCS-pA). (50)

pCMV6-hSUR1 pCMV6 containing a constitutive hSUR1 expression unit (PhCMV-hSUR1-pA). (51)

pCMV6-hKir6.2 pCMV6 containing a constitutive hKir6.2 expression unit (PhCMV-hKir6.2-pA). (52)

pCMV-T7-SB100 Constitutive SB100X expression vector (PhCMV-SB100X-pA) (Addgene no. 34879). (38)

pDA43 Tetracycline-responsive GLuc expression vector (PhCMV*-1-GLuc-pA). (53)

pDA145 PCRE-driven mINS-expression vector (PCRE-mINS-pA). (21)

pDONR Gateway®

-compatible cloning vector. Life Technologies, CA

pDONR-hGCK pDONR containing hGCK (Addgene no. 23750). (54)

pEGFP-N1 Constitutive EGFP-expression vector (PhCMV-EGFP-pA). Clontech, CA

pGNAT3 pCMV6 containing a constitutive hGNAT3 expression unit (PhCMV-GNAT3-pA). OriGene, MD

pHY30 PNFAT1-driven SEAP expression vector (PNFAT1-SEAP-pA; PNFAT1, (NFATIL2)3-Pmin). (18)

pHY57 PNFAT1-driven shGLP1 expression vector (PNFAT1-shGLP1-pA; PNFAT1, (NFATIL2)3-Pmin). (18)

pSBbi-BP SB100X-specific transposon containing a constitutive BFP and PuroR expression unit (ITR-PhEF1α-MCS-pA:PRPBSA-

BFP-P2A-PuroR-pA-ITR) (Addgene no. 60512).

(55)

pSBbi-RB SB100X-specific transposon containing a constitutive dTomato and BlastR expression unit (ITR-PhEF1α-MCS-

pA:PRPBSA-dTomato-P2A-BlastR-pA-ITR) (Addgene no. 60522).

(55)

pSBbi-RP SB100X-specific transposon containing a constitutive dTomato and PuroR expression unit (ITR-PhEF1α-MCS-

pA:PRPBSA-dTomato-P2A-PuroR-pA-ITR) (Addgene no. 60513).

(55)

pSBtet-Pur SB100X-specific transposon containing a tetracycline-inducible luciferase expression unit and a constitutive rtTA and

PuroR expression unit (PhCMV*-1-Luc-pA:PRPBSA-rtTA-P2A-PuroR-pA) (Addgene no. 60507).

(55)

pSP16 PCREm-driven SEAP expression vector (PCREm-SEAP-pA). (34)

pUC57 pUC19-derived prokaryotic expression vector. GeneScript, NJ

pZeoSV2(+) Constitutive mammalian expression vector conferring zeocin resistance (PhCMV-ZeoR-pA). Life Technologies, CA

35

pFS119 PNFAT3-driven TurboGFP:dest1 expression vector (PNFAT3-TurboGFP:dest1-pA; PNFAT3, (NFATIL4)5-Pmin). Sedlmayer et al.,

unpublished

pKR32 PNFkB-driven SEAP expression vector (PNFkB-SEAP-pA). Rössger et al., unpublished

pKK56 Constitutive Cacna2d1 and Cacnb3 expression vector (PhEF1α-Cacna2d1-P2A-Cacnb3-pA). Krawczyk et al., unpublished

pMM195 Constitutive GLP1R expression vector (PSV40-GLP1R-pA). Müller et al., unpublished

pT1R2 Constitutive hT1R2 expression vector (PhCMV-hT1R2-pA). Wieland et al., unpublished

pT1R3 Constitutive hT1R3 expression vector (PhCMV-hT1R3-pA). Wieland et al., unpublished

pHY101 PCRE-driven expression vector for SEAP and mINS (PCRE-SEAP-P2A-mINS-pA).

SEAP was PCR-amplified from pMX57 using oligonucleotides OHY701 (5’-

gccacggggatgaagcagaagctgaattcgCCACCATGCTGCTGCTGCTGCTGCTGCTG-3’) and OHY702 (5’-

ggaaaagttggttgctccTGTCTGCTCGAAGCGGCCGGCCGCCC-3’), P2A-mINS was PCR-amplified from pMX155

using oligonucleotides OHY703 (5’-gggcggccggccgcttcgagcagacaGGAGCAACCAACTTTTCC-3’) and OHY704 (5’-

cgaagcggccggccgccccgactctagaaagcttTCAGTTGCAGTAGTTCTCCAGT-3’) and both fragments were joined and

cloned into the EcoRI/XbaI sites of pCK53 using the GeneArt® Seamless Cloning and Assembly Kit.

This work

pMX53 PcFOS-driven SEAP expression vector (PcFOS-SEAP-pA; PcFOS, (c-fos)4-Pmin).

Pmin-SEAP was PCR-amplified from pSP16 using oligonucleotides OMX59 (5’-

cgcgtgctagcagcctgacgtttcagagactgacgtttcagagactgacgtttcagagactgacgtttcagatctctcga

ggtcgacagcggAGACTCTAGAGGGTATATAATG-3’) and OMX24 (5’-

CTTGAGCACATAGCCTGGACCGTTTCCGTA-3’), restricted with NdeI/NheI and cloned into the corresponding

sites (NdeI/NheI) of pHY30.

This work

pMX56 PNFAT2-driven SEAP expression vector (PNFAT2-SEAP-pA; PNFAT2, (NFATIL4)3-Pmin).

Pmin-SEAP was PCR-amplified from pSP16 using oligonucleotides OMX63 (5’- cgcgtg-

[ctagctacattggaaaattttatacacgtt]3AGACTCTAGAGGGTATATAA-3’) and OMX24 (5’-

CTTGAGCACATAGCCTGGACCGTTTCCGTA-3’), restricted with NdeI/NheI and cloned into the corresponding

sites (NdeI/NheI) of pHY30.

This work

pMX57 PNFAT3-driven SEAP expression vector (PNFAT3-SEAP-pA; PNFAT3, (NFATIL4)5-Pmin).

PNFAT3 (5’- cgcgtg-[ctagctacattggaaaattttatacacgtt]5-agactc-Pmin-cttggcaatccggtactgttggtaaagaattcgcccacc-3’) was

excised from pMX570 with NheI/EcoRI and cloned into the corresponding sites (NheI/EcoRI) of pMX56.

This work

pMX58 PNFAT4-driven SEAP expression vector (PNFAT4-SEAP-pA; PNFAT4, (NFATIL4)7-Pmin).

PNFAT4 (5’- cgcgtg-[ctagctacattggaaaattttatacacgtt]7-agactc-Pmin-cttggcaatccggtactgttggtaaagaattcgcccacc-3’) was

excised from pMX580 with NheI/EcoRI and cloned into the corresponding sites (NheI/EcoRI) of pMX56.

This work

pMX59 PNFAT5-driven SEAP expression vector (PNFAT5-SEAP-pA; PNFAT5, (NFATIL4)9-Pmin).

PNFAT5 (5’- cgcgtg-[ctagctacattggaaaattttatacacgtt]9-agactc-Pmin-cttggcaatccggtactgttggtaaagaattcgcccacc-3’) was

excised from pMX590 with NheI/EcoRI and cloned into the corresponding sites (NheI/EcoRI) of pMX56.

This work

pMX61 PNFAT3-driven shGLP1-expression vector (PNFAT3-shGLP1-pA).

shGLP1 was PCR-amplified from pHY57 using oligonucleotides OMX70 (5’-

ctgttggtaaagaattcgcccaccATGAAGATCATCCTGTGGCTGTGTG-3’) and OMX71 (5’-

ggagtcgacgcgtGAAGCGGCCGGCCTCATTTACCAGGAGAGTGGG-3’), restricted with EcoRI/FseI and cloned into

the corresponding sites (EcoRI/FseI) of pMX57.

This work

pMX68 PNFAT3-driven mINS-expression vector (PNFAT3-mINS-pA). This work

36

mINS was PCR-amplified from pDA145 using oligonucleotides OMX72 (5’-

ctgttggtaaagaattcgcCCACCATGGCCCTGTGGATGCGCTTCCTGC-3’) and OMX73 (5’-

CTGAAACATAAAATGAATGCAATTGTTGTTG-3’), restricted with EcoRI/MfeI and cloned into the corresponding

sites (EcoRI/MfeI) of pMX57.

pMX90 Constitutive hGCK expression vector (PhCMV-hGCK-pA).

GCK was PCR-amplified from pDONR-hGCK using oligonucleotides OMX100 (5’-

ctggaattccaccATGCTGGACGACAGAGCCAGG-3’) and OMX101 (5’-

ctctagatgcatgctcgagTCACTGGCCCAGCATACAGGCCTTC-3’), restricted with EcoRI/XhoI and cloned into the

corresponding sites (EcoRI/XhoI) of pcDNA3.1(+).

This work

pMX99 PNFAT4-driven mINS-expression vector (PNFAT4-mINS-pA).

mINS was excised from pMX68 using EcoRI/HpaI and cloned into the corresponding sites (EcoRI/HpaI) of pMX58.

This work

pMX100 PNFAT5-driven mINS-expression vector (PNFAT5-mINS-pA).

mINS was excised from pMX68 using EcoRI/HpaI and cloned into the corresponding sites (EcoRI/HpaI) of pMX59.

This work

pMX115 PNFAT5-driven shGLP1-expression vector (PNFAT5-shGLP1-pA).

shGLP1 was excised from pMX61 using EcoRI/HpaI and cloned into the corresponding sites (EcoRI/HpaI) of

pMX59.

This work

pMX117 PNFAT4-driven shGLP1-expression vector (PNFAT4-shGLP1-pA).

shGLP1 was excised from pMX61 using EcoRI/HpaI and cloned into the corresponding sites (EcoRI/HpaI) of

pMX58.

This work

pMX155 PNFAT5-driven expression vector for EGFP and mINS (PNFAT5-EGFP-P2A-mINS-pA).

EGFP-P2A was PCR-amplified from pEGFP-N1 using oligonucleotides OMX193 (5’-

ctaacgaattcgcATGGTGAGCAAGGGCGAGGAG-3') and OMX192 (5’-

cacagggccatgggtccaggattctcctccacgtcgcctgcctgcttcagcagggaaaagttggttgctccCTTGTACAG

CTCGTCCATGCCGAGAG-3'), restricted with EcoRI/NcoI and cloned into the corresponding sites (EcoRI/NcoI) of

pMX100.

This work

pMX245 SB100X-specific transposon containing a tetracycline-inducible luciferase expression unit and a constitutive ZeoR

expression unit (ITR-PhCMV*-1-Luc-pA:PRPBSA-ZeoR-pA-ITR).

ZeoR was PCR-amplified from pZeoSV2(+) using oligonucleotides OMX248 (5’-

ctgcacctgaggccaccATGGCCAAGTTGACCAGTGCCG-3’) and OMX249 (5’-

caagcttcacgacaggccttcgaaTCAGTCCTGCTCCTCGGCCACGAAG-3’), restricted with Bsu36I/BstBI and cloned into

the corresponding sites (Bsu36I/BstBI) of pSBtet-Pur.

This work

pMX248 SB100X-specific transposon containing a PNFAT5-driven EGFP and mINS expression unit and a constitutive ZeoR

expression unit (ITR-PNFAT5-EGFP-P2A-mINS-pA:PRPBSA-ZeoR-pA-ITR).

GFP-P2A-mINS was PCR-amplified from pMX155 using oligonucleotides OMX251 (5’-

catttctctatcgataactagtGAGCTCTTACGCGTGCTAGC-3’) and OMX252 (5’-

cggggtaccGGTCGACGGATCCTTATCGATTTTACC-3’) and restricted with SpeI/KpnI. pMX245 was then restricted

with KpnI/PciI into a 4476bp fragment A and a 1736bp fragment B. Fragment A was further restricted with

XbaI/NcoI/HindIII to yield a 2182bp fragment C. Finally, fragment C (PciI/XbaI), fragment B (KpnI/PciI) and GFP-

P2A-mINS (SpeI/KpnI) were assembled by ligating compatible SpeI/XbaI ends.

This work

pMX250 SB100X-specific transposon containing a constitutive dTomato and PuroR expression unit and a constitutive GLP1R This work

37

expression unit (ITR-PhEF1α-GLP1R-pA:PRPBSA-dTomato-P2A-PuroR-pA-ITR).

GLP1R was PCR-amplified from pMM195 using oligonucleotides OMX253 (5’-

ctaccccaagctggcctctgaggccaccatggctcgagATGGCCGTCACCCCCAGCCTGCTGCGCCTG-3’) and OMX245 (5’-

tgggctgcaggtcgactctagagTCAGCTGCAGGAATTTTGGCAGGTGGCTGC-'3), restricted with NcoI/XbaI and cloned

into the corresponding sites (NcoI/XbaI) of pSBbi-RP.

pMX251 SB100X-specific transposon containing a constitutive dTomato and BlastR expression unit and a constitutive

Cacna2d1 and Cacnb3 expression unit (ITR-PhEF1α-Cacna2d1-P2A-Cacnb3-pA:PRPBSA-dTomato-P2A-BlastR-pA).

Cacna2d1-P2A-Cacnb3 was PCR-amplified from pKK56 using oligonucleotides OMX254 (5’-

aaggtgtcgtgaaaactaccccaagctggcctctgaggccaccATGGCTGCTGGCTGCCTGCTGGCCTTGACTCTG-3’) and

OMX255 (5’-

gagaattgatccccaagcttggcctgacaggccCTAGACTCGAGCGGCCGCTCATCAGTAGCTGTCCTTAGGCCAAGGCC

GG-3’), restricted with SfiI and cloned into the corresponding site (SfiI) of pSBbi-RB.

This work

pMX252 SB100X-specific transposon containing a constitutive BFP and PuroR expression unit and a constitutive Cacna1d

expression unit (ITR-PhEF1α-Cacna1d-pA:PRPBSA-BFP-P2A-PuroR-pA-ITR).

Cacna1d was PCR-amplified from pCaV1.3 using oligonucleotides OMX256 (5’-

aaggtgtcgtgaaaactaccccaagctggcctctgaggccaccATGCAGCATCAACGGCAGCAGCAAGAGGAC-3’) and OMX257

(5’-gagaattgatccccaagcttggcctgacaggccCTAGACTCGAGCGGCCGCTCATCAGAGCATCCGTTCAAGCATCTGTA

GGGCGATC-3’), restricted with SfiI and cloned into the corresponding site (SfiI) of pSBbi-BP.

This work

pMX256 SB100X-specific transposon containing a PNFAT5-driven SEAP and mINS expression unit and a constitutive EGFP and

ZeoR expression unit (ITR-PNFAT5-SEAP-P2A-mINS-pA:PRPBSA-EGFP-P2A-ZeoR-pA-ITR).

PRPBSA-EGFP-P2A-ZeoR-pA was excised from pMX257 using KpnI/SphI and cloned into the corresponding sites

(KpnI/SphI) of pMX260.

This work

pMX257 SB100X-specific transposon containing a PCRE-driven SEAP and mINS expression unit and a EGFP and ZeoR

constitutive expression unit for (ITR-PCRE-SEAP-P2A-mINS-pA:PRPBSA-EGFP-P2A-ZeoR-pA-ITR).

EGFP-P2A was PCR-amplified from pEGFP-N1 using oligonucleotides OWH107 (5’-

attgaattcgcgaggccaccaaggccaccATGGTGAGCAAGGGCGAGGAGCT-3’) and OWH74 (5’-

aggtccaggattctcctccacgtcgcctgcctgcttcagcagggaaaagttggttgctccagatccCTTGTACAGCTCGT

CCATGC-3’), P2A-ZeoR was PCR-amplified from pMX248 using oligonucleotides OWH108 (5’-

acttttccctgctgaagcaggcaggcgacgtggaggagaatcctggacctATGGCCAAGTTGACCAGTGCCGTT-3’) and OWH29 (5’-

AACAACAGATGGCTGGCAACTAGAAG-3’), and both fragments were joined by overlapping PCR using OWH107

and OWH29. PCRE-SEAP-P2A-mINS was then excised from pHY101 using MluI/SalI and pMX248 was restricted with

MluI/SalI/SfiI/DraIII resulting in a 2942bp fragment A and a 635bp fragment B. Finally, EGFP-P2A-ZeoR was

restricted with SfiI/DraIII and ligated with fragment A (DraIII/MluI), PCRE-SEAP-2A-mINS (MluI/SalI) and fragment

B (SalI/SfiI).

This work

pMX258 SB100X-specific transposon containing a PCRE-driven SEAP and mINS expression unit and a constitutive ZeoR

expression unit (ITR-PCRE-SEAP-P2A-mINS-pA:PRPBSA-ZeoR-pA-ITR).

PCRE-SEAP-P2A-mINS-pA was excised from pHY101 using MluI/SalI and cloned into the corresponding sites

(MluI/SalI) of pMX248.

This work

pMX260 SB100X-specific transposon containing a PNFAT5-driven SEAP and mINS expression unit and a constitutive ZeoR

expression unit (ITR-PNFAT5-SEAP-P2A-mINS-pA:PRPBSA-ZeoR-pA-ITR).

This work

38

SEAP-P2A-mINS-pA was excised from pHY101 using EcoRI/SalI and cloned into the corresponding sites

(EcoRI/SalI) of pMX248.

pMX570 pUC57-derived vector containing custom-designed PNFAT3. This work

pMX580 pUC57-derived vector containing custom-designed PNFAT4. This work

pMX590 pUC57-derived vector containing custom-designed PNFAT5. This work

pWH29 PNFAT3-driven GLuc expression vector (PNFAT3-GLuc-pA).

PNFAT3 was excised from pMX57 using MluI/EcoRI and cloned into the corresponding sites (MluI/EcoRI) of pDA43.

This work

Abbreviations: attR1/2, Gateway®

-compatible recombination sites; BFP, blue fluorescent protein; BlastR, gene conferring blasticidin resistance; Cav1.2, member 2 of the

Cav1 family of L-type voltage-gated Ca2+

channels; Cav1.3, member 3 of the Cav1 family of L-type voltage-gated Ca2+

channels; Cav2.2, N-type voltage-gated Ca2+

channel;

c-fos, human proto-oncogene from the Fos family of transcription factors; Cacna1b, α1-subunit of rat Cav2.2 (NCBI Gene ID: 257648); Cacna1c, α1-subunit of mouse Cav1.2

(NCBI Gene ID: 12288); Cacna1d, α1-subunit of rat Cav1.3 (NCBI Gene ID: 29716); Cacnb3, β3-subunit of rat Cav1.3 (NCBI Gene ID: 25297); Cacna2d1, α2δ-subunit of

rat Cav1.3 (NCBI Gene ID: 25399); ccdB, DNA gyrase toxin for positive selection; Cmr, chloramphenicol-resistance gene for negative selection; CRE, cAMP-response

element; dTomato, destabilized red fluorescent protein variant; EGFP, enhanced green fluorescent protein (Genbank: U55762); GLP-1, glucagon-like peptide 1; GLP1R,

human GLP-1 receptor; GLuc, Gaussia princeps luciferase; hGCK, human pancreas glucokinase (also known as hexokinase 4; NCBI Gene ID: 2645); hGlut2, human

facilitated glucose transporter member 2 (also known as SLC2A2; NCBI Gene ID: 6514); hGNAT3, human guanine nucleotide binding protein alpha transducing 3 (NCBI

Gene ID: 346562); hKir6.2, human inwardly rectifying KATP-channel subunit (also known as KCNJ11; NCBI Gene ID: 3767); hSUR1, regulatory sulphonylurea receptor

subunit of human KATP-channels (also known as ABCC8; NCBI Gene ID: 6833); hT1R2, member 2 of the human taste receptor type 1 family (NCBI Gene ID: 80834);

hT1R3, member 3 of the human taste receptor type 1 family (NCBI Gene ID: 83756); IL2/4, interleukin 2/4; ITR, inverted terminal repeats of SB100X; KATP, adenosine

triphosphate-sensitive potassium channel; Luc, Firefly Luciferase; MCS, multiple cloning site; mINS, modified insulin variant for optimal expression in HEK-293 cells;

NFAT, nuclear factor of activated T-cells; NF-κB, nuclear factor kappa-light-chain-enhancer of activated B cells; P2A, picornavirus-derived ribosome skipping sequence

optimized for bicistronic expression in mammalian cells; pA, polyadenylation signal; PcFOS, synthetic mammalian promoter containing a tetrameric c-fos response element

((c-fos)4-Pmin; (56)); PCR, polymerase chain reaction; PCRE, CRE-containing synthetic mammalian promoter; PCREm, modified PCRE variant; PhEF1α, human elongation factor

1α promoter; PEST, peptide sequence rich in proline, glutamic acid, serine and threonine; PhCMV, human cytomegalovirus immediate early promoter; PhCMV*-1, tetracycline-

responsive promoter (tetO7-PhCMVmin); PhCMVmin, minimal version of PhCMV; Pmin, minimal eukaryotic TATA-box promoter (5’-

TAGAGGGTATATAATGGAAGCTCGACTTCCAG-3’); PNFAT1, synthetic mammalian promoter containing three tandem repeats of a human IL2 NFAT-binding site

((NFATIL2)3-Pmin; (23)); PNFAT2, synthetic mammalian promoter containing three tandem repeats of a human IL4 NFAT-binding site ((NFATIL4)3-Pmin; (23)); PNFAT3, synthetic

mammalian promoter containing five tandem repeats of a human IL4 NFAT-binding site ((NFATIL4)5-Pmin); PNFAT4, synthetic mammalian promoter containing seven tandem

repeats of a human IL4 NFAT-binding site ((NFATIL4)7-Pmin); PNFAT5, synthetic mammalian promoter containing nine tandem repeats of a human IL4 NFAT-binding site

((NFATIL4)9-Pmin); PNFkB, synthetic mammalian promoter containing a NF-κB-response element (Genbank: EU581860.1); PRPBSA, constitutive synthetic mammalian

promoter; PSV40, simian virus 40 promoter; PuroR, gene conferring puromycin resistance; rtTA, reverse TetR-dependent mammalian transactivator; SB100X, optimized

Sleeping Beauty transposase; SEAP, human placental secreted alkaline phosphatase; shGLP1, short human glucagon-like peptide 1; tetO, TetR-specific operator; TetR,

Escherichia coli Tn10-derived tetracycline-dependent repressor of the tetracycline resistance gene; TurboGFP:dest1, PEST-tagged TurboGFP variant (Evrogen); ZeoR,

gene conferring zeocin resistance.

Oligonucleotides: Restriction endonuclease-specific sites are underlined and annealing base pairs are indicated in capital letters.

39

Table S2 RT-PCR Primers. Related to Figure S1A.

Target Sequences Amplicon

size

Reference

GAPDH 5’-ACATCGCTCAGACACCATG-3’

5’-TGTAGTTGAGGTCAATGAA-3’

143bp (57)

GLUT1 5’-TGAACCTGCTGGCCTTC-3’

5’-GCAGCTTCTTTAGCACA-3’

399bp (58)

GLUT2 5’-TCCAGCTACCGACAGCCTATTC-3’

5’-AGATGGCACAAACAAACATCCC-3’

253bp (59)

GLUT3 5’-AAGGATAACTATAATGG-3’

5’-GGTCTCCTTAGCAGG-3’

411bp (58)

SUR1 5’-TCACACCGCTGTTCCTGCT-3’

5’-AGAAGGAGCGAGGACTTGCC-3’

412bp (60)

SUR2 5’-CATTGCCTACTTATTTCTCTCAG-3’

5’-ACCATTCTGAAGAAAGCCAG-3’

474bp (60)

Kir6.1 5’-CTGGCTGCTCTTCGCTATC-3’

5’-AGAATCAAAACCGTGATGGC-3’

234bp (60)

Kir6.2 5’-CCAAGAAAGGCAACTGCAACG-3’

5’-ATGCTTGCTGAAGATGAGGGT-3’

449bp (60)

Table S3Simulation protocols

ID Description Protocol Fig.

P1 in vitro seeding and12h incubation of trans-fected cells in 2 mMglucose

in vitro states set to initial values described inTable S4 (GS=2mM), in vivo states set to 0,plasmid copy numbers set according to TableS11→ simulation for 12h

-

P2 in vitro kinetics afterKCl/glucose/GLP-1 in-duction

P1 → setting initial KCl/glucose/GLP-1concentration→ simulation for 48h

2D, S9C,S9D, S16F,S16G,S16H

P3 in vitro dose responseto KCl/glucose/GLP-1

Repeating P2 for different initialKCl/glucose/GLP-1 concentrations andsaving SEAP value at final time described(24, 48 or 72h depending on the experi-ment). Only the dose response with 100 %CaV transfection was used for parameterestimation

1E, 2C,4A, S9A,S9B S16D,S16E

P4(X ,t) in vitro seeding, induc-tion and incubation forreversibility steps

set external states including NS according toTable S4 and GS to X mM, simulation fortime t

-

P5 in vitro reversible kinet-ics with 2mM/33mMglucose induction low→ high→ low

P1 for 24h instead of 12 → P4(2,12) →P4(33,24)→ P4(2,12)→ P4(2,24)

S8B

P6 reversible kinetics with2mM/33mM glucoseinduction high → low→ high

P1 for 24h instead of 12 and GS set ini-tially to 33mM → P4(2,12) → P4(2,24) →P4(2,12)→ P4(33,24)

S8B

P7(FI) in vivo kinetics forwild-type /T1D mice

Initialization without implant (γi = 0 for alli), setting native insulin production FI , simu-lation for 3 days→ adding implant: set syn-thetic compartment states and plasmid copynumbers according to Tables S4 and S11,GS = 0 , switch insulin/SEAP production bythe implant with FSCI , simulation for 21 days

S11 S13A,S13B, 4B,4C, 4H

P8 in vivo intraperitonealglucose tolerance

P7→ set GI to 50 mM→ simulation for 2h 4D, S10,S13C,S13D

40

Table S3continued

ID Description Protocol Fig.

P9(n) in vivo reversible kinet-ics low→ high with im-plant cell number mul-tiplied n times

P7(0) with NS set to n×NS when adding im-plant→ FI set to 1 and simulation for 5 days

S13E

P10(n) in vivo reversible kinet-ics high→ low with im-plant cell number mul-tiplied n times

P7(1) with NS set to n×NS when adding im-plant→ FI set to 0 and simulation for 5 days

S13F

P11(n) in vivo oral glucose tol-erance

P7 → set GI to 63.16 mM and GLP-1 to123pM (values estimated for oral test) →simulation for 6h

4G

41

Table S4State variables used in the mathematical model. For in vivo experiments, the initial cell densityNS was calculated according to the number of cells per capsule (500) and a factor FNS accountingfor encapsulation efficiency.

Name Definition Unit Initialvalue

NS Cell density in the synthetic compartment in vitro cells L−1 2 ·108

Cell density in the synthetic compartment in vivo cells L−1 7.84 ·108

GS Concentration of glucose in the synthetic compartment mM from 2to 40

GSi Concentration of glucose in the cells mM 0

6CP Intermediate metabolites (phosphorylated six-carbonsugars) concentration in the cells

mM 0

A Concentration of ATP in the cells mM 0.1

CaS Concentration of calcium ions in the synthetic compart-ment

mM 1.8

CaSi Concentration of calcium ions in the cells mM 2 ·10−5

KS Concentration of potassium ions in the synthetic com-partment

mM 5.3

KSi Concentration of potassium ions in the cells mM 200

NaS Concentration of sodium ions in the synthetic compart-ment

mM 155

NaSi Concentration of sodium ions in the cells mM 8

V Transmembrane voltage Φin−Φout mV -90

MS Concentration of SEAP or insulin mRNA in the cells mM 0

MGLP1 Concentration of GLP-1 mRNA in the cells* mM 0

SS Concentration of SEAP protein in the synthetic compart-ment

U L−1 0

IS Concentration of insulin in the synthetic compartment µg L−1 0

I Concentration of insulin in the blood µg L−1 0

Ia Concentration of insulin in an action compartment µg L−1 0

Ii Concentration of insulin in a virtual compartment µg L−1 0

G Concentration of glucose in the blood mM 0

GI Concentration of glucose in an additional virtual com-partment

mM 0

GLP1S Concentration of GLP-1 in the synthetic compartment* mM 0

GLP1 Concentration of GLP-1 in the blood* mM 0

GLP1R Concentration of GLP-1 receptor in the cells* mM 042

Table S5Rates, currents, and fluxes used in the mathematical model

Name Definition Unit

R1 Rate of cellular glucose uptake mediated by Glut1 mM min−1 cell−1

R2 Rate of intracellular glucose and ATP consumption byphosphofructokinase with allosteric inhibition by ATP (qmolecules of ATP for one molecule of glucose)

mM min−1

R3 Rate of intermediate metabolites consumption by pyru-vate kinase producing q+1 molecules of ATP and withallosteric inhibition by ATP

mM min−1

R4 Rate of Ca2+-dependent intracellular SEAP production mM min−1

R5 Rate of GLP1S-dependent intracellular SEAP produc-tion*

mM min−1

δ Rate of global ATP consumption by the cell mM min−1

µr Cell-density dependent growth rate mM min−1

IK Outward background potassium current pA

IKV Outward voltage-dependent potassium current pA

IKAT P Outward ATP-dependent potassium current pA

INa Outward background sodium current pA

ICa Outward background calcium current pA

ICaV Outward voltage-dependent calcium current pA

INaK Outward sodium-potassium pump current pA

ICaP Outward calcium pump current pA

νG Glucose transfer flux from the blood to the syntheticcompartment

0.2986 mMmin−1

νS SEAP transfer flux from the blood to the synthetic com-partment

U L−1 min−1

νI Insulin transfer flux from the blood to the synthetic com-partment

µg L−1 min−1

νGLP1 GLP-1 transfer flux from the blood to the synthetic com-partment*

mM min−1

νGI Diffusion coefficient of glucose from the blood to thevirtual glucose compartment

mM min−1

43

Table S6General parameters used in the mathematical model. For in vivo experiments, the volume of thesynthetic compartment VS was calculated according to the number of capsules (1000) and thecapsule diameter (400 µm).

Name Definition Value Unit Source

VS Volume of the synthetic com-partment for in vitro experi-ments

1.5 ·10−4 L Fixed

VS Volume of the synthetic com-partment for in vivo experi-ments

3.35 ·10−4 L Fixed

VB Volume of the mouse blood 1.6 ·10−3 L Fixed

VG Volume of artificial glucosecompartment

3 ·10−3 L (21)

FSCI Scaling factor insulin 0.0229 µg U−1 Estimated

FSCG Scaling factor GLP1* 7.94 · 10−10

mmol U−1- Estimated

µ Maximal specific growth ratein vitro

3.848 ·10−4 min−1 (21)

µ Maximal specific growth ratein vivo

0 min−1 (21)

NmaxS Michaelis-Menten constant

cell density108 cells L−1 Estimated

r Radius of HEK-293 cell 6.5 µm Invitrogen specifica-tions sheet

VC Volume of HEK-293 cell 103 43π(10−6r)3 L Fixed

44

Table S7Metabolic parameters used in the mathematical model.

Name Definition Value Unit Source

q Autocatalytic stoichiometry 1 - Estimated

Vmax1 Maximum specific rate of cel-lular glucose uptake

10−13 mmol min−1 cell−1 (61)

Km1 Michaelis-Menten constantfor the uptake of glucose

0.31 mM Estimated

Vmax2 Max. rate of intracellular glu-cose and ATP consumption

10 mM min−1 Estimated

Km2A Affinity constant for al-losteric inhibition of ATPconsumption by ATP

1.61 mM Estimated

Km2G Michaelis-Menten constantfor glucose consumption

1.95 mM Estimated

a Hill coefficient for binding ofATP to phosphofructokinase

1 - Estimated

b Hill coefficient for allostericinhibition of ATP consump-tion by ATP

1 - Estimated

c Hill coefficient for allostericinhibition of intermediatemetabolites consumption byATP

1 - Estimated

Vmax3 Max. intermediate consump-tion rate

9.96 mM min−1 Estimated

Km3 Affinity for intermediary con-sumption inhibition by ATP

1.83 mM Estimated

Atot Total concentration of ATP +ADP

5 mM (43)

u Max. global rate of ATP con-sumption

0.12 mM min−1 Estimated

Kδ Michaelis-Menten constantfor ATP consumption

2 mM Estimated

45

Table S8Ion transport parameters used in the mathematical model

Name Definition Value Unit Source

R Ideal gas constant 8.3144 J mol−1 K−1 Fixed

T Temperature 310 K Fixed

F Faraday constant 9.6485 ·104 C mol−1 Fixed

C Capacitance of HEK-293cell (scaled for time unitsin min.)

7.2/60 ·10−3 nF (62)

α Coefficient converting cur-rent in pA to concentrationper unit time in mM/min

60 ·10−9

VCFmM min−1 pA−1 Fixed

gKV Conductance of potassiumions through voltage-activated potassiumchannels

1.24 nS Estimated

gK Conductance of potassiumions through backgroundpotassium channels

1.97 ·10−2 nS Estimated

gKAT P Conductance of potassiumions through ATP-dependent potassiumchannels

1.96 nS Estimated

K1 ADP binding constant,ATP-sensitive potassiumchannels

3 mM Estimated

K2 ATP binding constant,ATP-sensitive potassiumchannels

2.21 mM Estimated

dK Hill coefficient, ATP-sensitive potassiumchannels

7 Estimated

gNa Conductance of sodiumions through backgroundsodium channels

4.37 ·10−2 nS Estimated

gCaV Conductance of calciumions through voltage-activated calcium channels

0.80 nS Estimated

46

Table S8continued

Name Definition Value Unit Source

gCa Conductance of calciumions through backgroundcalcium channels

6.43 ·10−3 nS Estimated

ImaxNaK Maximum sodium-

potassium pump current42.72 pA Estimated

kK Rate constant of potassiumuptake from the mediumby NaK pump

1 min−1 (44)

kNa Rate constant of sodiumuptake from the cell byNaK pump

11 min−1 (44)

fCa Fraction of cytosolic cal-cium that is free

10−3 - (43)

ImaxCaP Maximum calcium pump

current9.40 pA Estimated

47

Table S9Gene expression parameters used in the mathematical model

Name Definition Value Unit Source

Vmax4 Max. calcium-dependent gene expression 2.67 ·10−6 mM min−1 Estimated

Km4 Affinity for calcium-dependent gene ex-pression

1.1 ·10−3 mM Estimated

d Hill coefficient for calcium-dependentgene expression

1.6 - Estimated

Vmax5 Max. GLP-1-dependent gene expression* 4.5 ·10−3 mM min−1 Estimated

Km5 Affinity for GLP-1-dependent gene ex-pression*

3.4 ·10−8 mM Estimated

d5 Hill coefficient for GLP-1-dependentgene expression*

1.0432 - Estimated

kLCRE CRE promoter basal expression level rel-ative to maximal expression*

0.0135 - Estimated

kR constitutive GLP-1 receptor expression* 9.76 ·10−6 mM min−1 Estimated

ktl Common translation constant 5.00 ·10−8 U mM−1

min−1 cell−1Estimated

kdm Degradation constant mRNA 2.8 ·10−3 min−1 Estimated

kds Degradation constant SEAP in vitro 9 ·10−4 min−1 Estimated

kdi Degradation constant insulin in the syn-thetic compartment

1.1 ·10−3 min−1 Estimated

kdR Degradation constant GLP-1 receptor* 3.5 ·10−3 min−1 Estimated

kdG Degradation constant GLP-1 in the syn-thetic compartment*

1.5 ·10−3 min−1 Estimated

nC plasmid copy number for Cav1.3 transfec-tion

defined intable S11

- Estimated

nS plasmid copy number for SEAP or Insulintransfection

defined intable S11

- Estimated

nNFAT Number of NFAT repeats for calcium-dependent SEAP

5 - Estimated

Number of NFAT repeats for calcium-dependent Insulin expression

9 - Estimated

nG plasmid copy number for GLP-1 transfec-tion

defined intable S11

- Estimated

nR plasmid copy number for GLP-1 receptortransfection

defined intable S11

- Estimated

48

Table S10Parameters for the mathematical model in vivo

Name Definition Value Unit Source

DG Vascular exchange constantglucose

0.2986 min−1 Estimated

γG Vascular exchange constantglucose

0.7047 min−1 Estimated

γI Vascular exchange constantinsulin

0.9137 min−1 Estimated

γS Vascular exchange constantSEAP

9.8254 ·10−5 min−1 Estimated

γGLP1 Vascular exchange constantGLP-1

0.9228 min−1 Estimated

gp Glucose production 0.1061 mM min−1 Estimated

gr Glucose effectiveness,insulin-independent glucoseremoval

4.1 ·10−3 mM min−1 Estimated

ksi Insulin sensitivity 1.2 ·10−3 L µg−1 min−1 Estimated

kiar Insulin clearance from actioncompartment

0.0566 min−1 Estimated

kiir Insulin clearance from virtualcompartment

0.0351 min−1 Estimated

kp Basal insulin production 4.0338 ·10−5 µg L−1 mM−1 min−1 Estimated

ki Integral (insulin-mediated)insulin production

1.05 ·10−10 µg L−1 mM−1 min−1 Estimated

kd Differential (glucose-mediated) insulin production

0.2340 µg L−1 mM−1 min−1 Estimated

kir Insulin clearance 3.1439 ·10−4 µg L−1 mM−1 min−1 Estimated

FI Factor wild-type / diabetic in-sulin production

1 or 0 - Fixed

kdsiv Degradation constant SEAPin vivo

1.3 ·10−3 min−1 Estimated

kdGiv Degradation constant GLP-1in the blood*

2.5 ·10−3 min−1 Estimated

49

Table S11Plasmid copy numbers estimated for the mathematical model of HEK-β. nC is the plasmidcopy number for Cav1.3 transfection and nS is the plasmid copy number for SEAP or Insulintransfection. T means transient transfection and S means stable

Fig. Description Cav1.3 nC SS/IS nS

2C stable dose-response to glucose 48h after induc-tion 100% CaV

T 3.75 S 1

2C stable dose-response to glucose 48h after induc-tion 66% CaV

T 2.61 S 1

2C stable dose-response to glucose 48h after induc-tion 33% CaV

T 1.51 S 1

2D stable kinetics with 33mM glucose induction T 3.2 S 1

S11B reversible kinetics with 2mM/33mM glucoseinduction

T 3.2 S 1

S9A,1E

transient dose-response to KCl 48h after induc-tion

T 0.9 T 0.9

S9A,1E

control for previous experiment none 0.4 T 0.9

S9B transient dose-response to glucose 48h after in-duction

T 2.5 T 2.5

S9B control for previous experiment none 0.4 T 2.5

S9C transient kinetics with glucose induction T 2.85 T 2.85

S9D transient kinetics with KCl induction T 1.1 T 1.1

S9D stable kinetics with KCl induction T 0.92 S 1

S11 in vivo SEAP expression kinetics for wild-type/ type 1 diabetes mice

T 3.75 S 1

S11 control for previous experiment (without cal-cium channel)

none 0.4 S 1

S13,S16

in vivo insulin transient expression experimentsand predictions

T 3.13 T 3.13

4G in vivo insulin stable expression, oral glucosetolerance test

S 3.6 S 3.6

50

Table S12Plasmid copy numbers estimated for the mathematical model of HEK-βGLP. nC is the plasmidcopy number for Cav1.3 transfection and nS is the plasmid copy number for SEAP or Insulintransfection. T means transient transfection and S means stable. Plasmid ratios correspond tothe experimental setup and experiments done at the same time are considered to have the sametransfection efficiencies.

Fig. Description Cav1.3 nC GLP1 nG SS/IS nS GLP1R nR

S16D,S16E

in vitro SEAP exper-iments without NFAT-GLP1 and without CaV

T 0.4 T 0 T 0.66 S 1

S16D,S16E

in vitro SEAP experi-ments without CaV

T 0.4 T 2.66 T 0.66 S 1

S16D,S16E

in vitro SEAP exper-iments without CRE-GLP1

T 2.66 T 0 T 0.66 S 1

S16D,S16E,S16G

in vitro SEAP exper-iments with all con-structs

T 2.66 T 2.66 T 0.66 S 1

S16D,S16E

in vitro SEAP exper-iments with GLP1R-deficient cell line

T 2.66 T 2.66 T 0.66 S 0

S16F in vitro SEAP kineticresponse to GLP-1 withGLP1R-deficient cellline

T 5.14 T 5.14 T 1.29 S 1

S16H in vitro Insulin kineticresponse to GLP-1 withGLP1R-deficient cellline

T 5.14 T 5.14 T 5.14 S 1

4H in vivo long-termglycemia

T 2.82 T 2.82 T 2.82 S 1

4G in vivo oral glucose tol-erance test

T 3.25 T 3.25 T 3.25 S 1

51

References and Notes 1. S. Cornell, Continual evolution of type 2 diabetes: An update on pathophysiology and

emerging treatment options. Ther. Clin. Risk Manag. 11, 621–632 (2015). doi:10.2147/TCRM.S67387 Medline

2. J. Tuomilehto, S. Bahijri, Epidemiology: Lifetime risk of diabetes mellitus—how high? Nat. Rev. Endocrinol. 12, 127–128 (2016). doi:10.1038/nrendo.2015.227 Medline

3. K. S. Polonsky, The past 200 years in diabetes. N. Engl. J. Med. 367, 1332–1340 (2012). doi:10.1056/NEJMra1110560 Medline

4. S. J. Cleland, Cardiovascular risk in double diabetes mellitus—when two worlds collide. Nat. Rev. Endocrinol. 8, 476–485 (2012). doi:10.1038/nrendo.2012.47 Medline

5. G. I. Shulman, Ectopic fat in insulin resistance, dyslipidemia, and cardiometabolic disease. N. Engl. J. Med. 371, 1131–1141 (2014). doi:10.1056/NEJMra1011035 Medline

6. S. M. Grundy, Drug therapy of the metabolic syndrome: Minimizing the emerging crisis in polypharmacy. Nat. Rev. Drug Discov. 5, 295–309 (2006). doi:10.1038/nrd2005 Medline

7. P. M. Rao, D. M. Kelly, T. H. Jones, Testosterone and insulin resistance in the metabolic syndrome and T2DM in men. Nat. Rev. Endocrinol. 9, 479–493 (2013). doi:10.1038/nrendo.2013.122 Medline

8. J. E. Bruin, N. Saber, N. Braun, J. K. Fox, M. Mojibian, A. Asadi, C. Drohan, S. O’Dwyer, D. S. Rosman-Balzer, V. A. Swiss, A. Rezania, T. J. Kieffer, Treating diet-induced diabetes and obesity with human embryonic stem cell-derived pancreatic progenitor cells and antidiabetic drugs. Stem Cell Rep. 4, 605–620 (2015). doi:10.1016/j.stemcr.2015.02.011 Medline

9. A. Rezania, J. E. Bruin, P. Arora, A. Rubin, I. Batushansky, A. Asadi, S. O’Dwyer, N. Quiskamp, M. Mojibian, T. Albrecht, Y. H. C. Yang, J. D. Johnson, T. J. Kieffer, Reversal of diabetes with insulin-producing cells derived in vitro from human pluripotent stem cells. Nat. Biotechnol. 32, 1121–1133 (2014). doi:10.1038/nbt.3033 Medline

10. J. C. Pickup, Insulin-pump therapy for type 1 diabetes mellitus. N. Engl. J. Med. 366, 1616–1624 (2012). doi:10.1056/NEJMct1113948 Medline

11. D. J. Drucker, M. A. Nauck, The incretin system: Glucagon-like peptide-1 receptor agonists and dipeptidyl peptidase-4 inhibitors in type 2 diabetes. Lancet 368, 1696–1705 (2006). doi:10.1016/S0140-6736(06)69705-5 Medline

12. G. M. Church, M. B. Elowitz, C. D. Smolke, C. A. Voigt, R. Weiss, Realizing the potential of synthetic biology. Nat. Rev. Mol. Cell Biol. 15, 289–294 (2014). doi:10.1038/nrm3767 Medline

13. F. W. Pagliuca, J. R. Millman, M. Gürtler, M. Segel, A. Van Dervort, J. H. Ryu, Q. P. Peterson, D. Greiner, D. A. Melton, Generation of functional human pancreatic β cells in vitro. Cell 159, 428–439 (2014). doi:10.1016/j.cell.2014.09.040 Medline

14. B. E. Tuch, B. Szymanska, M. Yao, M. T. Tabiin, D. J. Gross, S. Holman, M. A. Swan, R. K. B. Humphrey, G. M. Marshall, A. M. Simpson, Function of a genetically modified human liver cell line that stores, processes and secretes insulin. Gene Ther. 10, 490–503 (2003). doi:10.1038/sj.gt.3301911 Medline

52

15. C. Aguayo-Mazzucato, S. Bonner-Weir, Stem cell therapy for type 1 diabetes mellitus. Nat. Rev. Endocrinol. 6, 139–148 (2010). doi:10.1038/nrendo.2009.274 Medline

16. L. Marzban, C. J. Rhodes, D. F. Steiner, L. Haataja, P. A. Halban, C. B. Verchere, Impaired NH2-terminal processing of human proislet amyloid polypeptide by the prohormone convertase PC2 leads to amyloid formation and cell death. Diabetes 55, 2192–2201 (2006). doi:10.2337/db05-1566 Medline

17. S. A. Stanley, J. E. Gagner, S. Damanpour, M. Yoshida, J. S. Dordick, J. M. Friedman, Radio-wave heating of iron oxide nanoparticles can regulate plasma glucose in mice. Science 336, 604–608 (2012). doi:10.1126/science.1216753 Medline

18. H. Ye, M. Daoud-El Baba, R. W. Peng, M. Fussenegger, A synthetic optogenetic transcription device enhances blood-glucose homeostasis in mice. Science 332, 1565–1568 (2011). doi:10.1126/science.1203535 Medline

19. F. M. Ashcroft, P. Rorsman, K(ATP) channels and islet hormone secretion: New insights and controversies. Nat. Rev. Endocrinol. 9, 660–669 (2013). doi:10.1038/nrendo.2013.166 Medline

20. M. D’Arco, A. C. Dolphin, L-type calcium channels: On the fast track to nuclear signaling. Sci. Signal. 5, pe34 (2012). doi:10.1126/scisignal.2003355 Medline

21. D. Ausländer, S. Ausländer, G. Charpin-El Hamri, F. Sedlmayer, M. Müller, O. Frey, A. Hierlemann, J. Stelling, M. Fussenegger, A synthetic multifunctional mammalian pH sensor and CO2 transgene-control device. Mol. Cell 55, 397–408 (2014). doi:10.1016/j.molcel.2014.06.007 Medline

22. H. Ye, G. Charpin-El Hamri, K. Zwicky, M. Christen, M. Folcher, M. Fussenegger, Pharmaceutically controlled designer circuit for the treatment of the metabolic syndrome. Proc. Natl. Acad. Sci. U.S.A. 110, 141–146 (2013). doi:10.1073/pnas.1216801110 Medline

23. J. W. Rooney, M. R. Hodge, P. G. McCaffrey, A. Rao, L. H. Glimcher, A common factor regulates both Th1- and Th2-specific cytokine gene expression. EMBO J. 13, 625–633 (1994). Medline

24. M. Elsner, M. Tiedge, B. Guldbakke, R. Munday, S. Lenzen, Importance of the GLUT2 glucose transporter for pancreatic beta cell toxicity of alloxan. Diabetologia 45, 1542–1549 (2002). doi:10.1007/s00125-002-0955-x Medline

25. R. B. Robey, N. Hay, Mitochondrial hexokinases, novel mediators of the antiapoptotic effects of growth factors and Akt. Oncogene 25, 4683–4696 (2006). doi:10.1038/sj.onc.1209595 Medline

26. H. J. Jang, Z. Kokrashvili, M. J. Theodorakis, O. D. Carlson, B.-J. Kim, J. Zhou, H. H. Kim, X. Xu, S. L. Chan, M. Juhaszova, M. Bernier, B. Mosinger, R. F. Margolskee, J. M. Egan, Gut-expressed gustducin and taste receptors regulate secretion of glucagon-like peptide-1. Proc. Natl. Acad. Sci. U.S.A. 104, 15069–15074 (2007). doi:10.1073/pnas.0706890104 Medline

27. F. Macian, NFAT proteins: Key regulators of T-cell development and function. Nat. Rev. Immunol. 5, 472–484 (2005). doi:10.1038/nri1632 Medline

28. A. J. Vegas, O. Veiseh, M. Gürtler, J. R. Millman, F. W. Pagliuca, A. R. Bader, J. C. Doloff, J. Li, M. Chen, K. Olejnik, H. H. Tam, S. Jhunjhunwala, E. Langan, S. Aresta-Dasilva, S. Gandham, J. J. McGarrigle, M. A. Bochenek, J. Hollister-Lock, J.

53

Oberholzer, D. L. Greiner, G. C. Weir, D. A. Melton, R. Langer, D. G. Anderson, Long-term glycemic control using polymer-encapsulated human stem cell-derived beta cells in immune-competent mice. Nat. Med. 22, 306–311 (2016). doi:10.1038/nm.4030 Medline

29. G. G. Holz 4th, W. M. Kühtreiber, J. F. Habener, Pancreatic beta-cells are rendered glucose-competent by the insulinotropic hormone glucagon-like peptide-1(7-37). Nature 361, 362–365 (1993). doi:10.1038/361362a0 Medline

30. S. J. Russell, F. H. El-Khatib, M. Sinha, K. L. Magyar, K. McKeon, L. G. Goergen, C. Balliro, M. A. Hillard, D. M. Nathan, E. R. Damiano, Outpatient glycemic control with a bionic pancreas in type 1 diabetes. N. Engl. J. Med. 371, 313–325 (2014). doi:10.1056/NEJMoa1314474 Medline

31. T. J. Kieffer, Closing in on Mass Production of Mature Human Beta Cells. Cell Stem Cell 18, 699–702 (2016). doi:10.1016/j.stem.2016.05.014 Medline

32. J. T. McCluskey, M. Hamid, H. Guo-Parke, N. H. McClenaghan, R. Gomis, P. R. Flatt, Development and functional characterization of insulin-releasing human pancreatic beta cell lines produced by electrofusion. J. Biol. Chem. 286, 21982–21992 (2011). doi:10.1074/jbc.M111.226795 Medline

33. D. W. Scharp, P. Marchetti, Encapsulated islets for diabetes therapy: History, current progress, and critical issues requiring solution. Adv. Drug Deliv. Rev. 67-68, 35–73 (2014). doi:10.1016/j.addr.2013.07.018 Medline

34. P. Saxena, B. C. Heng, P. Bai, M. Folcher, H. Zulewski, M. Fussenegger, A programmable synthetic lineage-control network that differentiates human IPSCs into glucose-sensitive insulin-secreting beta-like cells. Nat. Commun. 7, 11247 (2016). doi:10.1038/ncomms11247 Medline

35. A. N. Zaykov, J. P. Mayer, R. D. DiMarchi, Pursuit of a perfect insulin. Nat. Rev. Drug Discov. 15, 425–439 (2016). doi:10.1038/nrd.2015.36 Medline

36. M. Kastellorizios, N. Tipnis, D. J. Burgess, Foreign Body Reaction to Subcutaneous Implants. Adv. Exp. Med. Biol. 865, 93–108 (2015). doi:10.1007/978-3-319-18603-0_6 Medline

37. F. M. Wurm, Production of recombinant protein therapeutics in cultivated mammalian cells. Nat. Biotechnol. 22, 1393–1398 (2004). doi:10.1038/nbt1026 Medline

38. L. Mátés, M. K. L. Chuah, E. Belay, B. Jerchow, N. Manoj, A. Acosta-Sanchez, D. P. Grzela, A. Schmitt, K. Becker, J. Matrai, L. Ma, E. Samara-Kuko, C. Gysemans, D. Pryputniewicz, C. Miskey, B. Fletcher, T. VandenDriessche, Z. Ivics, Z. Izsvák, Molecular evolution of a novel hyperactive Sleeping Beauty transposase enables robust stable gene transfer in vertebrates. Nat. Genet. 41, 753–761 (2009). doi:10.1038/ng.343 Medline

39. S. Arora, S. K. Ojha, D. Vohora, Characterisation of Streptozotocin Induced Diabetes Mellitus in Swiss Albino Mice. Glob. J. Pharmacol. 3, 81–84 (2009).

40. E. R. Gilbert, Z. Fu, D. Liu, Development of a nongenetic mouse model of type 2 diabetes. Exp. Diabetes Res. 2011, 416254 (2011). doi:10.1155/2011/416254 Medline

41. E. M. Watson, M. J. Chappell, F. Ducrozet, S. M. Poucher, J. W. Yates, A new general glucose homeostatic model using a proportional-integral-derivative controller.

54

Comput. Methods Programs Biomed. 102, 119–129 (2011). doi:10.1016/j.cmpb.2010.08.013 Medline

42. F. A. Chandra, G. Buzi, J. C. Doyle, Glycolytic oscillations and limits on robust efficiency. Science 333, 187–192 (2011). doi:10.1126/science.1200705 Medline

43. J. Keizer, G. Magnus, ATP-sensitive potassium channel and bursting in the pancreatic beta cell. A theoretical study. Biophys. J. 56, 229–242 (1989). doi:10.1016/S0006-3495(89)82669-4 Medline

44. D. S. Lindblad, C. R. Murphey, J. W. Clark, W. R. Giles, A model of the action potential and underlying membrane currents in a rabbit atrial cell. Am. J. Physiol. 271, H1666–H1696 (1996). Medline

45. T. D. Helton, W. Xu, D. Lipscombe, Neuronal L-type calcium channels open quickly and are inhibited slowly. J. Neurosci. 25, 10247–10251 (2005). doi:10.1523/JNEUROSCI.1089-05.2005 Medline

46. W. Xu, D. Lipscombe, Neuronal Ca(V)1.3alpha(1) L-type channels activate at relatively hyperpolarized membrane potentials and are incompletely inhibited by dihydropyridines. J. Neurosci. 21, 5944–5951 (2001). Medline

47. Y. Lin, S. I. McDonough, D. Lipscombe, Alternative splicing in the voltage-sensing region of N-Type CaV2.2 channels modulates channel kinetics. J. Neurophysiol. 92, 2820–2830 (2004). doi:10.1152/jn.00048.2004 Medline

48. T. J. Bell, C. Thaler, A. J. Castiglioni, T. D. Helton, D. Lipscombe, Cell-specific alternative splicing increases calcium channel current density in the pain pathway. Neuron 41, 127–138 (2004). doi:10.1016/S0896-6273(03)00801-8 Medline

49. H. Takanaga, W. B. Frommer, Facilitative plasma membrane transporters function during ER transit. FASEB J. 24, 2849–2858 (2010). doi:10.1096/fj.09-146472 Medline

50. Y. Yamada, S. R. Post, K. Wang, H. S. Tager, G. I. Bell, S. Seino, Cloning and functional characterization of a family of human and mouse somatostatin receptors expressed in brain, gastrointestinal tract, and kidney. Proc. Natl. Acad. Sci. U.S.A. 89, 251–255 (1992). doi:10.1073/pnas.89.1.251 Medline

51. P. Béguin, K. Nagashima, M. Nishimura, T. Gonoi, S. Seino, PKA-mediated phosphorylation of the human K(ATP) channel: Separate roles of Kir6.2 and SUR1 subunit phosphorylation. EMBO J. 18, 4722–4732 (1999). doi:10.1093/emboj/18.17.4722 Medline

52. N. Inagaki, T. Gonoi, J. P. Clement 4th, N. Namba, J. Inazawa, G. Gonzalez, L. Aguilar-Bryan, S. Seino, J. Bryan, Reconstitution of IKATP: An inward rectifier subunit plus the sulfonylurea receptor. Science 270, 1166–1170 (1995). doi:10.1126/science.270.5239.1166 Medline

53. M. Müller, S. Ausländer, D. Ausländer, C. Kemmer, M. Fussenegger, A novel reporter system for bacterial and mammalian cells based on the non-ribosomal peptide indigoidine. Metab. Eng. 14, 325–335 (2012). doi:10.1016/j.ymben.2012.04.002 Medline

54. C. M. Johannessen, J. S. Boehm, S. Y. Kim, S. R. Thomas, L. Wardwell, L. A. Johnson, C. M. Emery, N. Stransky, A. P. Cogdill, J. Barretina, G. Caponigro, H. Hieronymus, R. R. Murray, K. Salehi-Ashtiani, D. E. Hill, M. Vidal, J. J. Zhao, X. Yang, O. Alkan, S. Kim, J. L. Harris, C. J. Wilson, V. E. Myer, P. M. Finan, D. E. Root, T. M. Roberts,

55

T. Golub, K. T. Flaherty, R. Dummer, B. L. Weber, W. R. Sellers, R. Schlegel, J. A. Wargo, W. C. Hahn, L. A. Garraway, COT drives resistance to RAF inhibition through MAP kinase pathway reactivation. Nature 468, 968–972 (2010). doi:10.1038/nature09627 Medline

55. E. Kowarz, D. Löscher, R. Marschalek, Optimized Sleeping Beauty transposons rapidly generate stable transgenic cell lines. Biotechnol. J. 10, 647–653 (2015). doi:10.1002/biot.201400821 Medline

56. M. Sheng, S. T. Dougan, G. McFadden, M. E. Greenberg, Calcium and growth factor pathways of c-fos transcriptional activation require distinct upstream regulatory sequences. Mol. Cell. Biol. 8, 2787–2796 (1988). doi:10.1128/MCB.8.7.2787 Medline

57. X. Zhu, F. Ozturk, C. Liu, G. G. Oakley, A. Nawshad, Transforming growth factor-β activates c-Myc to promote palatal growth. J. Cell. Biochem. 113, 3069–3085 (2012). doi:10.1002/jcb.24184 Medline

58. M. A. Castro, C. Angulo, S. Brauchi, F. Nualart, I. I. Concha, Ascorbic acid participates in a general mechanism for concerted glucose transport inhibition and lactate transport stimulation. Pflugers Arch. 457, 519–528 (2008). doi:10.1007/s00424-008-0526-1 Medline

59. C. Limbert, G. Päth, R. Ebert, V. Rothhammer, M. Kassem, F. Jakob, J. Seufert, PDX1- and NGN3-mediated in vitro reprogramming of human bone marrow-derived mesenchymal stromal cells into pancreatic endocrine lineages. Cytotherapy 13, 802–813 (2011). doi:10.3109/14653249.2011.571248 Medline

60. Q. Du, S. Jovanović, A. Sukhodub, E. Barratt, E. Drew, K. M. Whalley, V. Kay, M. McLaughlin, E. E. Telfer, C. L. R. Barratt, A. Jovanović, Human oocytes express ATP-sensitive K(+) channels. Hum. Reprod. 25, 2774–2782 (2010). doi:10.1093/humrep/deq245 Medline

61. E. Wertheimer, S. Sasson, E. Cerasi, Y. Ben-Neriah, The ubiquitous glucose transporter GLUT-1 belongs to the glucose-regulated protein family of stress-inducible proteins. Proc. Natl. Acad. Sci. U.S.A. 88, 2525–2529 (1991). doi:10.1073/pnas.88.6.2525 Medline

62. B. Farrell, C. Do Shope, W. E. Brownell, Voltage-dependent capacitance of human embryonic kidney cells. Phys. Rev. E Stat. Nonlin. Soft Matter Phys. 73, 041930 (2006). doi:10.1103/PhysRevE.73.041930 Medline

56