8
Supporting Information Rothenberg et al. 10.1073/pnas.0803217105 SI Methods DNA and shRNA Constructs. The putative target sequences for the shRNA vectors are: shCTRL: 5-GCAAGCTGACCCTGAAGTTCAT-3 shEGFR 1: 5-GCTGGATGATAGACGCAGATA-3 shEGFR 2: 5-GCTGAGAATGTGGAATACCTA-3 shEGFR 3: 5-AGAATGTGGAATACCTAAGG-3 shEGFR 4: 5-GCTGCTCTGAAATCTCCTTTA-3 shErbB-3: 5-AATTCTCTACTCTACCATTG-3 The shCTRL vector encodes an shRNA that minimally affects the levels of Green Fluorescent Protein (GFP). The shEGFR1 vector targets a sequence within the EGFR coding region. The shEGFR vectors 2– 4 target sequences within the 3-untranslated region (UTR) of the primary EGFR mRNA. All shRNAs were expressed in pLKO, a self-inactivating lentiviral vector (William Hahn), from the U6 snRNP promoter. This vector also contains a puromycin resistance cassette expressed from the SV40 pro- moter. shRNA inserts were subcloned into the EcoRI-AgeI sites of pLKO, and the correct orientation and sequence were con- firmed by dideoxynucleotide sequencing. cDNA expression con- structs were generated in the lentiviral vector pLenti 6V5 DEST, which also contains a blasticidin resistance cassette expressed from an SV40 promoter, using the Gateway cloning system (Invitrogen), or in pWPI, a lentiviral cDNA expression vector containing a co-cistronic green fluorescent protein reporter (Didier Trono). The entire EGFR coding sequence and muta- tions therein for each construct was verified by dideoxynucle- otide sequencing. The HIV-1 packaging construct pCMV- deltaR8.91 encodes the viral GAG, PRO, POL, TAT and REV proteins. pMD.G expresses the VSV-G envelope protein from a CMV promoter. Lentivirus Production and Transduction. shRNA or cDNA lentiviral expression constructs were co-transfected into 293T cells with pCMVdeltaR8.91 and pMD.G using TransIT-LT-1 transfection reagent (Fisher Scientific). Viral supernatants were collected at 48 and 72 h posttransfection in DMEM with 30% FCS, filtered through 0.45 m syringe filters to remove cell debris and stored in aliquots at 80 °C. Transductions were carried out in the presence of 8 g/ml Polybrene (Sigma) by spinoculation at 1200 g and at 32 °C for 60 min in a Sorvall Legend RT table-top centrifuge. Viral supernatant was exchanged for fresh media 30–36 h after spinoculation. To generate stable EGFR variant- expressing cell lines, cells were transduced by serial spinocula- tion. 24 h after the second round of transduction, blasticidin was added (where appropriate) to 1 g/ml and the cells were cultured for an additional two passages. Individual shRNA vector super- natants were titered using the A549 human lung cancer cell line, which is relatively insensitive to the toxic effects of shRNA vectors targeting EGFR, by comparing cell survival in the absence or presence of puromycin. Vector titers on each PC9- derived cell population for each experiment was compared in the same manner using the lentiviral shCTRL construct at the same time as the cells were transduced with the lentviral shEGFR vectors. Viral supernatant was stable to multiple freeze-thaw cycles. Cell Lines and Reagents. NSCLC cell line PC-9 (derived from a patient with adenocarcinoma previously untreated) was pro- vided by Prof. Hayata (Tokyo Medical University Tokyo, Japan). NSCLC cell line A549 was obtained from the American Type Culture Collection. Cell lines were cultured with RPMI-1640 (GIBCO-BRL/Invitrogen) supplemented with 10% FBS (GIBCO-BRL/Invitrogen), penicillin and streptomycin (100 U/ml and 100 g/ml, respectively; GIBCO-BRL). 293-T cells were cultured in DMEM (GIBCO-BRL/Invitrogen) supple- mented with 10% FBS, penicillin, and streptomycin. All cell lines were cultured in a humidified atmosphere of 5% CO 2 at 37 °C. Erlotinib was obtained from the Massachusetts General Hospi- tal Pharmacy. PHA-665752 was obtained from Pfizer. PI-103 was obtained from Cayman Chemical. Immunoblotting. Cell lysis buffer consisted of 1% Nonidet P-40, 20 mM Tris, 150 mM NaCl, 10 mM MgCl 2 , 0.25% sodium deoxycholate, 10% glycerol, 50 mM sodium fluoride, 1 mM sodium vanadate, 4 g/ml leupeptin, 4 g/ml aprotinin, and 1 mM PMSF. Protein levels were determined by immunoblot of cell lysates after SDS/PAGE and transfer to PVDF membranes. Reactive bands were visualized using the SuperSignal West Pico Peroxide reagent (Pierce) after incubating membranes with anti-rabbit or anti-mouse HRP-linked IgG (Cell Signaling). Levels of EGFR, Akt, ERK1/2, and ErbB-3 were determined using rabbit polyclonal anti-EGFR (Santa Cruz Biotechnology), anti-Akt (Cell Signaling), anti-ERK 1/2 (Cell Signaling), or anti-ErbB-3 (Santa Cruz Biotechnology) antibodies. The phos- phorylated forms of EGFR, ERK1/2, and ErbB-3 were detected using rabbit polyclonal anti-EGFR [pY1068] (Abcam), anti- ERK1/2 [pT202/S204] (Cell Signaling), or anti-ErbB-3 [pY1289] (Cell Signaling) phosphospecific antibodies. Phosphorylated Akt was detected using rabbit monoclonal anti-Akt [pS473] phos- phospecific antibody (Biosource). Levels of polyADP ribose polymerase (PARP) and GAPDH were detected with mouse monoclonal anti-PARP (BD PharMingen) or anti-GAPDH (Chemicon) antibodies. Anti-p85 antibody was from Upstate Biotechnology. Phospho-Tyr antibody was from Cell Signaling. Flow Cytometry. Adherent cells were gently detached with tryp- sin, collected by centrifugation, resuspended in PBS containing 3% BSA and incubated at 37 °C for 30 min before addition of 0.5 g mouse monoclonal antibody 528 (Santa Cruz) for 30 min. Cells were washed 3 times with PBS and then incubated with 1 g Alexa Fluor 647 goat anti-mouse IgG (HL) (Molecular Probes/Invitrogen) for 30 min. Cells were washed 3 times with PBS, resuspended in 1% paraformaldehyde, and analyzed using an LSRII f low cytometer (Becton-Dickinson). For f luorescence- activated cell sorting, GFP-expressing cells were isolated using a FACSAria cytometer (Becton-Dickinson). 1. Rokudai S, Fujita N, Hashimoto Y, Tsuruo T (2000) Cleavage and inactivation of antiapoptotic Akt/PKB by caspases during apoptosis. J Cell Physiol 182(2):290 –296. 2. Fan QW, et al. (2006) A dual PI3 kinase/mTOR inhibitor reveals emergent efficacy in glioma. Cancer Cell 9(5):341–349. 3. Fan QW, et al. (2007) A dual phosphoinositide-3-kinase alpha/mTOR inhibitor coop- erates with blockade of epidermal growth factor receptor in PTEN-mutant glioma. Cancer Res 67(17):7960 –7965. Rothenberg et al. www.pnas.org/cgi/content/short/0803217105 1 of 8

Supporting Information - PNAS€¦ · The shCTRL vector does not significantly affect signaling output from EGFR or cause apoptosis (lanes 1–2). In control (GFP) cells treated

  • Upload
    others

  • View
    7

  • Download
    0

Embed Size (px)

Citation preview

Page 1: Supporting Information - PNAS€¦ · The shCTRL vector does not significantly affect signaling output from EGFR or cause apoptosis (lanes 1–2). In control (GFP) cells treated

Supporting InformationRothenberg et al. 10.1073/pnas.0803217105SI MethodsDNA and shRNA Constructs. The putative target sequences for theshRNA vectors are:

shCTRL: 5�-GCAAGCTGACCCTGAAGTTCAT-3�shEGFR 1: 5�-GCTGGATGATAGACGCAGATA-3�shEGFR 2: 5�-GCTGAGAATGTGGAATACCTA-3�shEGFR 3: 5�-AGAATGTGGAATACCTAAGG-3�shEGFR 4: 5�-GCTGCTCTGAAATCTCCTTTA-3�shErbB-3: 5�-AATTCTCTACTCTACCATTG-3�The shCTRL vector encodes an shRNA that minimally affects

the levels of Green Fluorescent Protein (GFP). The shEGFR1vector targets a sequence within the EGFR coding region. TheshEGFR vectors 2–4 target sequences within the 3�-untranslatedregion (UTR) of the primary EGFR mRNA. All shRNAs wereexpressed in pLKO, a self-inactivating lentiviral vector (WilliamHahn), from the U6 snRNP promoter. This vector also containsa puromycin resistance cassette expressed from the SV40 pro-moter. shRNA inserts were subcloned into the EcoRI-AgeI sitesof pLKO, and the correct orientation and sequence were con-firmed by dideoxynucleotide sequencing. cDNA expression con-structs were generated in the lentiviral vector pLenti 6V5 DEST,which also contains a blasticidin resistance cassette expressedfrom an SV40 promoter, using the Gateway cloning system(Invitrogen), or in pWPI, a lentiviral cDNA expression vectorcontaining a co-cistronic green fluorescent protein reporter(Didier Trono). The entire EGFR coding sequence and muta-tions therein for each construct was verified by dideoxynucle-otide sequencing. The HIV-1 packaging construct pCMV-deltaR8.91 encodes the viral GAG, PRO, POL, TAT and REVproteins. pMD.G expresses the VSV-G envelope protein from aCMV promoter.

Lentivirus Production and Transduction. shRNA or cDNA lentiviralexpression constructs were co-transfected into 293T cells withpCMVdeltaR8.91 and pMD.G using TransIT-LT-1 transfectionreagent (Fisher Scientific). Viral supernatants were collected at48 and 72 h posttransfection in DMEM with 30% FCS, filteredthrough 0.45 �m syringe filters to remove cell debris and storedin aliquots at �80 °C. Transductions were carried out in thepresence of 8 �g/ml Polybrene (Sigma) by spinoculation at1200 � g and at 32 °C for 60 min in a Sorvall Legend RT table-topcentrifuge. Viral supernatant was exchanged for fresh media30–36 h after spinoculation. To generate stable EGFR variant-expressing cell lines, cells were transduced by serial spinocula-tion. 24 h after the second round of transduction, blasticidin wasadded (where appropriate) to 1 �g/ml and the cells were culturedfor an additional two passages. Individual shRNA vector super-natants were titered using the A549 human lung cancer cell line,which is relatively insensitive to the toxic effects of shRNAvectors targeting EGFR, by comparing cell survival in theabsence or presence of puromycin. Vector titers on each PC9-derived cell population for each experiment was compared in thesame manner using the lentiviral shCTRL construct at the same

time as the cells were transduced with the lentviral shEGFRvectors. Viral supernatant was stable to multiple freeze-thawcycles.

Cell Lines and Reagents. NSCLC cell line PC-9 (derived from apatient with adenocarcinoma previously untreated) was pro-vided by Prof. Hayata (Tokyo Medical University Tokyo, Japan).NSCLC cell line A549 was obtained from the American TypeCulture Collection. Cell lines were cultured with RPMI-1640(GIBCO-BRL/Invitrogen) supplemented with 10% FBS(GIBCO-BRL/Invitrogen), penicillin and streptomycin (100U/ml and 100 �g/ml, respectively; GIBCO-BRL). 293-T cellswere cultured in DMEM (GIBCO-BRL/Invitrogen) supple-mented with 10% FBS, penicillin, and streptomycin. All cell lineswere cultured in a humidified atmosphere of 5% CO2 at 37 °C.Erlotinib was obtained from the Massachusetts General Hospi-tal Pharmacy. PHA-665752 was obtained from Pfizer. PI-103was obtained from Cayman Chemical.

Immunoblotting. Cell lysis buffer consisted of 1% Nonidet P-40,20 mM Tris, 150 mM NaCl, 10 mM MgCl2, 0.25% sodiumdeoxycholate, 10% glycerol, 50 mM sodium fluoride, 1 mMsodium vanadate, 4 �g/ml leupeptin, 4 �g/ml aprotinin, and 1mM PMSF. Protein levels were determined by immunoblot ofcell lysates after SDS/PAGE and transfer to PVDF membranes.Reactive bands were visualized using the SuperSignal West PicoPeroxide reagent (Pierce) after incubating membranes withanti-rabbit or anti-mouse HRP-linked IgG (Cell Signaling).Levels of EGFR, Akt, ERK1/2, and ErbB-3 were determinedusing rabbit polyclonal anti-EGFR (Santa Cruz Biotechnology),anti-Akt (Cell Signaling), anti-ERK 1/2 (Cell Signaling), oranti-ErbB-3 (Santa Cruz Biotechnology) antibodies. The phos-phorylated forms of EGFR, ERK1/2, and ErbB-3 were detectedusing rabbit polyclonal anti-EGFR [pY1068] (Abcam), anti-ERK1/2 [pT202/S204] (Cell Signaling), or anti-ErbB-3 [pY1289](Cell Signaling) phosphospecific antibodies. Phosphorylated Aktwas detected using rabbit monoclonal anti-Akt [pS473] phos-phospecific antibody (Biosource). Levels of polyADP ribosepolymerase (PARP) and GAPDH were detected with mousemonoclonal anti-PARP (BD PharMingen) or anti-GAPDH(Chemicon) antibodies. Anti-p85 antibody was from UpstateBiotechnology. Phospho-Tyr antibody was from Cell Signaling.

Flow Cytometry. Adherent cells were gently detached with tryp-sin, collected by centrifugation, resuspended in PBS containing3% BSA and incubated at 37 °C for 30 min before addition of 0.5�g mouse monoclonal antibody 528 (Santa Cruz) for 30 min.Cells were washed 3 times with PBS and then incubated with 1�g Alexa Fluor 647 goat anti-mouse IgG (H�L) (MolecularProbes/Invitrogen) for 30 min. Cells were washed 3 times withPBS, resuspended in 1% paraformaldehyde, and analyzed usingan LSRII flow cytometer (Becton-Dickinson). For fluorescence-activated cell sorting, GFP-expressing cells were isolated using aFACSAria cytometer (Becton-Dickinson).

1. Rokudai S, Fujita N, Hashimoto Y, Tsuruo T (2000) Cleavage and inactivation ofantiapoptotic Akt/PKB by caspases during apoptosis. J Cell Physiol 182(2):290–296.

2. Fan QW, et al. (2006) A dual PI3 kinase/mTOR inhibitor reveals emergent efficacy inglioma. Cancer Cell 9(5):341–349.

3. Fan QW, et al. (2007) A dual phosphoinositide-3-kinase alpha/mTOR inhibitor coop-erates with blockade of epidermal growth factor receptor in PTEN-mutant glioma.Cancer Res 67(17):7960–7965.

Rothenberg et al. www.pnas.org/cgi/content/short/0803217105 1 of 8

Page 2: Supporting Information - PNAS€¦ · The shCTRL vector does not significantly affect signaling output from EGFR or cause apoptosis (lanes 1–2). In control (GFP) cells treated

Fig. S1. Lentiviral shRNA vectors targeting EGFR demonstrate different efficiencies of knockdown and resulting toxicity. (A) PC9 cells transduced with theindicated lentiviral shRNA vectors were examined for EGFR by flow cytometry (Left). Cellular toxicity correlates with knockdown by each vector (Right). (B)Increasing the titer of the shEGFR3 vector causes progressively increased loss of EGFR protein (Left) and increased toxicity (Right).

Rothenberg et al. www.pnas.org/cgi/content/short/0803217105 2 of 8

Page 3: Supporting Information - PNAS€¦ · The shCTRL vector does not significantly affect signaling output from EGFR or cause apoptosis (lanes 1–2). In control (GFP) cells treated

Fig. S2. Sources of off-target toxicity for shEGFR vectors. (A) PC9 cells expressing GFP or DEL15-delUTR EGFR are resistant to the shCTRL vector across the entirerange of viral titers (black curves). Therefore, selection with puromycin gives a measure of the relative fraction of transduced cells, which is similar for each celltype (blue and red curves). (B) Following infection with the shCTRL or shEGFR vectors, levels of the indicated proteins were monitored by immunoblot every 12 h.The shCTRL vector does not significantly affect signaling output from EGFR or cause apoptosis (lanes 1–2). In control (GFP) cells treated with the shEGFR vector,the levels of phosphorylated Akt and ERK begin to decrease at 36 h posttransduction (lanes 3 vs. 5) but are preserved in DEL15-delUTR cells (lanes 4 vs. 6). Thedecrease in phosphorylated Akt clearly precedes the decrease in total Akt that is evident at 48 h posttransduction (lanes 7–8) as a result of caspase-mediatedcleavage (1). The loss of phosphorylated and total Akt even in the DEL15-delUTR cells is due to incomplete reconstitution of EGFR activity or an off-target effectof the shRNA vector. Despite this, the appearance of the cleaved 85kD PARP fragment indicative of apoptosis in control cells is significantly delayed in cellsexpressing DEL15-delUTR (lanes 7 vs. 8).

Rothenberg et al. www.pnas.org/cgi/content/short/0803217105 3 of 8

Page 4: Supporting Information - PNAS€¦ · The shCTRL vector does not significantly affect signaling output from EGFR or cause apoptosis (lanes 1–2). In control (GFP) cells treated

NOT NORMALIZED

-7.5 -5.0 -2.5 0.0 2.5 5.0

1.75E6

3.5E6

LOG 2 [VIRUS ADDED (ul)]

PIXE

LIN

TEN

SITY

WELLS

-5.0 -2.5 0.0 2.5 5.0

NORMALIZED TO UNTREATED

-7.50.0

0.5

1.0

LOG 2 [VIRUS ADDED (ul)]

REL

ATIV

EW

ELL

INTE

NSI

TY

NORMALIZED FOR TITER

-7.5 -5.0 -2.5 0.0 2.5 5.00.0

0.5

1.0

LOG 2 [VIRUS ADDED (ul)]

VIA

BIL

ITY

3

6

0AUC

(Arb

itrar

yun

its)

t-test = 0.24

DEL15-1 DEL15-2

A

B

C

DEL15-1 shCTRLDEL15-1 shEGFRDEL15-2 shCTRLDEL15-2 shEGFR

0

Fig. S3. Reproducibility of AUC analysis. Two PC9 cell populations (DEL15–1 and DEL15–2), each expressing exogenous DEL15-delUTR, were generatedindependently on different days, and AUC analysis for each population was carried out, also on different days for each cell line. (A) Raw well intensity data (frompuromycin-treated wells) before any normalization. (B) Intensity data for puromycin-treated wells normalized to untreated wells. (C) Intensity data (e.g.,viability) normalized for variations in shRNA viral titer as determined with the shCTRL virus. After normalization, the AUCs for each cell population are notstatistically different (bar graph). Data are presented as the mean of four replicates �1 STDEV.

Rothenberg et al. www.pnas.org/cgi/content/short/0803217105 4 of 8

Page 5: Supporting Information - PNAS€¦ · The shCTRL vector does not significantly affect signaling output from EGFR or cause apoptosis (lanes 1–2). In control (GFP) cells treated

LOG2 [VIRUS ADDED (ul)]

VIA

BIL

ITY

shCTRL

-7.5 -5.0 -2.5 0.0 2.5 5.00.0

0.5

1.0

GFPK745AWTT790ML858R

DEL15

DEL18

T790M+DEL15T790M+L858R

A

C

GAPDH

EGFR

P-EGFR

shC

TRL

shE

GFR

shC

TRL

shE

GFR

shC

TRL

shE

GFR

GFP K745A DEL15L858R

shC

TRL

shE

GFR

B

-7.5 -5.0 -2.5 0.0 2.5 5.0

0.0

0.5

1.0

LOG 2 [VIRUS ADDED (ul)]

VIA

BIL

ITY

shEGFR

GFP DEL15

shC

TRL

shE

GFR

shC

TRL

shE

GFR

shC

TRL

shE

GFR

T790MDEL15

EGFR (Long exposure)

GAPDH

DGFP

T790M DEL15

GAPDH

P-AKTAKT

ERK1/2P-ERK1/2

shC

TRL

shE

GFR

DEL15

shC

TRL

shE

GFR

shC

TRL

shE

GFR

DEL18

sCTR

Lsh

EG

FR

L858R

shC

TRL

shE

GFR

WT

shC

TRL

shE

GFR

Fig. S4. Relative activity of EGFR variants. (A) Levels of the dual EGFR mutants achieved through heterologous expression are less than the single mutants (butare still detectable compared to control cells). (B) EGFR levels were analyzed in PC9 cells engineered to stably express the indicated EGFR variants (minus theendogenous 3�-UTR) 2 days after transduction with the control shRNA vector or the shEGFR vector targeting the EGFR 3�-UTR. 20 �g of each extract was probedwith the indicated antibodies. All endogenous EGFR protein is eliminated in control GFP-expressing cells (lanes 1 vs. 2). Therefore, the levels of phosphorylatedEGFR remaining after knockdown reflect the relative kinase activity of the exogenous EGFR variants. As expected for the kinase-inactive K745A variant, althoughit is detectable after knockdown of endogenous EGFR, it is completely unphosphorylated (lanes 3 vs.4). (C) AUC curves for the complete panel of EGFR variants.(D) Lysates were examined by immunoblotting with the indicated antibodies after transduction of cells as in Fig. 2B.

Rothenberg et al. www.pnas.org/cgi/content/short/0803217105 5 of 8

Page 6: Supporting Information - PNAS€¦ · The shCTRL vector does not significantly affect signaling output from EGFR or cause apoptosis (lanes 1–2). In control (GFP) cells treated

GFPEGFR MUTANT LTR IRES EF1aLTR

A

B

GFP EGFR

C

GFP EGFRPARENTAL 1.00 1.00

UNSORTED 8.97 1.43SORTED 45.05 2.33

EGFRPARENTAL 1.00

PAR. KO 0.13SORTED KO 1.27

EV

EN

TS

EV

EN

TS

EV

EN

TS

EGFR

Fig. S5. Methodology for increasing exogenous EGFR levels. (A) EGFR variants were subcloned into the pWPI vector, which encodes a co-cistronic GFP reporter.(B) After transduction of PC9 cells, highly GFP-positive cells were collected by FACS (left panel), and levels of EGFR were determined by flow cytometry as describedin SI Methods (right panel). The table shows mean EGFR levels for the indicated cell populations. (C) The sorted cells retain as much cell-surface EGFR, afterknockdown of the endogenous protein (green histogram), as parental PC9 cells before knockdown (blue histogram). The decreased EGFR in a small fraction ofthe sorted cells is likely due to the approximately 10% GFP ‘‘low’’ cells that remain after sorting (Fig. S5B green histogram). The table shows the median EGFRlevels for the indicated cell populations.

Rothenberg et al. www.pnas.org/cgi/content/short/0803217105 6 of 8

Page 7: Supporting Information - PNAS€¦ · The shCTRL vector does not significantly affect signaling output from EGFR or cause apoptosis (lanes 1–2). In control (GFP) cells treated

Fig. S6. Truncated DEL15 mutants retain the ability to activate ErbB-3. (A) Deletion of the C-terminal portion of EGFR removes the epitope for the antibodyused for immunoblot. Therefore, the exogenous truncation mutants were detected by flow cytometry using an antibody against the EGFR ectodomain. InGFP-expressing control cells, the shEGFR vector results in significant loss of cell-surface EGFR (histogram of EGFR expression in shCTRL-treated cells [green] shiftsto the left in shEGFR-treated cells [red]). However, cells expressing either full-length or truncated DEL15 EGFR retain detectable EGFR after knockdown of theendogenous protein (green and red curves overlap). The levels of the truncated EGFR at the cell-surface are higher than full-length (compare the peak of thered histograms for ‘‘DEL15 957T’’ and ‘‘DEL 15’’ plots); however, the sorting approach described in Fig. S5 leads to equivalent levels of each protein (comparered histograms for ‘‘DEL15 957T’’ and ‘‘DEL 15 SORTED’’ plots). (B) Knockdown of EGFR decreases both phosphorylated ErbB-3 and phosphorylated Akt (lanes2–4), while knockdown of ErbB-3 (lanes 5–7) decreases phosphorylated Akt.

Rothenberg et al. www.pnas.org/cgi/content/short/0803217105 7 of 8

Page 8: Supporting Information - PNAS€¦ · The shCTRL vector does not significantly affect signaling output from EGFR or cause apoptosis (lanes 1–2). In control (GFP) cells treated

0.0

GFPDEL15+TRUNCGFP+PI-103DEL15+TRUNC+PI-103

VIAB

ILIT

Y

-7.5 -5.0 -2.5 0.0 2.5 5.0

0.5

1.0

LOG 2 [VIRUS (ul)]

-7.5 -5.0 -2.5 0.0 2.5 5.00.0

0.5

1.0

LOG 2 [VIRUS (ul)]

VIAB

ILIT

Y

B.shCTRL

shEGFR

A

P-ErbB-3

ErbB-3

P-AKT

AKT

P-ERK

ERK

GAPDH

shC

TRL

shE

GFR

shC

TRL

shE

GFR

DMSO PI-103

shE

GFR

shC

TRL

shC

TRL

shE

GFR

TRUNC TRUNCGFP GFP

P-EGFR

EGFR

GFP

TRUNC

shEGFRshCTRL

* *

Fig. S7. EGFR C-terminal deletion mutants are sensitive to PI3K inhibition. (A) PC9 cells stably expressing expressing GFP or the shRNA-resistant DEL15 C-terminaltruncation were transduced with control or EGFR shRNA vectors. 24 h after transduction, 1 �M PI-103, a highly specific and potent class IA PI3K inhibitor (2, 3),or DMSO was added. 48 h after transduction, extracts were prepared and immunoblot analysis was performed with the indicated antibodies (Left). The levelsof phosphorylated Akt (compare lanes 5–8 to 1–4) are significantly reduced as a result of PI-103 treatment. The expression of the C-terminal truncation EGFRwas verified by flow cytometry after knockdown of endogenous EGFR as in Fig. S6 (Right). (B) PC9 cells were plated at 1e4 per well in 96-well format andtransduced the following day with a dilutions series (0.12–12.5 �l) of the shCTRL or shEGFR vectors. 30 h after transduction, puromycin and 1 �M PI-103 or DMSOwere added. Cell viability was determined as in Fig. 2 C.

Rothenberg et al. www.pnas.org/cgi/content/short/0803217105 8 of 8