Supplementary Material and Method Details
Reagents
CellTiter-Glo® Luminescent Cell Viability Assay was from Promega, bortezomib was clinical grade, recombinant PDIA1 from Enzo Life Sciences Inc., PDIA3 was from Biovision Inc. Biotin labeled molecules were visualized using Vectastain Elite ABC, (Vector Laboratories). di-E-GSSG was from IMCO Corp. Ltd. PACMA 31 was from Sigma Aldrich Inc. and LOC14 was from Enamine Ltd.
Chemical library
A consortium library set of 30,335 diverse small molecules was obtained from SPECS USA, Hopkinton, RI. The only strict criterion applied to compound selection was size (MW<500, range 116.21-499.86). Kier and Hall topological (valence) connectivity indices were then calculated on the remaining SPECS stock of small molecules. After removing near-constant descriptors, a principal component analysis was applied. K-means clustering on the 10 most relevant components yielded 30,498 clusters from which the SPECS global diversity algorithm picked approximately one compound per cluster for the consortium library set. Generation of Bemis-Murcko frameworks on this selection revealed 12,912 unique frameworks corresponding to an average of 2.36 structures per framework. Most scaffolds only occur a few times (12,132 out of 12,912 frameworks occur 1-5 times) confirming diversity of the library. Although not restricted by drug-like property calculations, 26,613 compounds had no or one violation of the following rules: Rotational bonds <10, H-bond donor < 5, H-bond acceptor < 10, tPSA < 140 Å, logP < 5, logS > -7.
Primary anti-myeloma screen and confirmation of hits
We screened a library of 30,335 chemically diverse small molecules for anti-MM activity using the human MM1.S cell line with a single step cellular ATP quantitation assay (Supplementary Table S1). Since we were only interested in compounds that were active in the low micromolar range and were reconstituting the dry library manually, we decided to use one volume of DMSO for all compounds that allowed suspension, guaranteed at least 10μM drug exposure in the primary screen accounting for possible variation in pipetting, and yielded final DMSO concentrations that did not affect MM1.S growth or survival. This resulted in a median small molecule concentration of 21μM, ranging from 15 to 65 μM for the largest (MW 499.9) and smallest molecule (MW 116.2), respectively, in the primary screen. Only the 1219 compounds that inhibited MM1.S growth or survival by 95% after 3-4 days of incubation were moved to confirmatory secondary screening that used defined concentrations (1,3,5,10 μM) and yielded 225 confirmed active hits with IC70 5μM in the ATP assay.
Liver homogenate preparation for multilayered (sandwich) in vivo model
Livers were collected from C57BL/KaLwRij male mice, washed twice in ice-cold PBS and placed in a dounce tissue grinder (7ml volume, Sigma-Aldrich) with 1ml of ice cold tissue culture medium and crushed with 10 strokes, then aliquoted and frozen at -80°C. Right before use aliquots were thawed, diluted (1:50) with tissue culture medium and pushed through 0.22μm filters (EMD Millipore).
Drug formulations for mouse treatments
CCF642 formulations used sterile filtered (0.1μm) 10% bovine serum albumin (Sigma Aldrich, cat. # A7030) dissolved in water with NaCl addition to achieve final concentrations of 133mM after dissolving CCF642 stock (5mg/ml in DMSO) in aqueous BSA. This solution was further enriched for albumin dissolved CCF642 by 20 min centrifugation at 2000g in Amicon Ultra-15 protein purification tubes (EMD Millipore, Merck KGaA) yielding final albumin concentrations of 14% and CCF642 concentrations of 0.7mg/ml. Vehicle control substituted DMSO for CCF642 stock but was otherwise prepared identically. BTZ was diluted to 0.125mg/ml in PBS.
Immunoblotting
This was done according to standard procedures with one refinement; to enhance detection of low abundant proteins by antibodies, PVDF membranes, after protein transfer, were air dried at room temperature until visibly dry (20-30 min) followed by brief (2-10 seconds) immersion in 100% methanol, then washed in 1xPBS with 0.05% Tween 20. Subsequent steps starting with blocking in 5% non-fat milk dissolved in 1xPBS with 0.05% Tween-20 were as standard in most laboratories.
Antibodies
The following antibodies were from Cell Signaling, Inc. (Danvers, USA): PARP rabbit pAb No. 9542, caspase 3 rabbit mAb No. 9665, XBP-1s rabbit mAb #12782, Actin rabbit (HRP conjugated) mAb No. 12620, histone H3 rabbit mAb No. 4499, PDI rabbit mAb No. 3501, IRE-1α rabbit mAb No. 3294, PERK rabbit mAb No. 5683, CHOP rabbit mAB No. 5554. Phospho PERK rabbit pAb (Thr 981) was from Santacruz, sc-32577 and anti-his mouse mAb from Pierce, cat. # MA1-21315.
Scheme for synthesis of CCF642 and CCF642-COOH
Scheme for synthesis of biotinylated CCF642
Proteomic Analysis of biotinylated CCF642 targets
MM1.S cells were treated with biotinylated CCF642 or vehicle and 2h later lysed followed by purification of biotinylated proteins using streptavidin magnetic beads. After SDS-PAGE resolution, the area around 60kD which contained the putative main target based on biotin stain of B-CCF642 treated MM1.S cell lysates (Supplementary Fig. S4) was cut out and a sample at the same molecular weight obtained from controls in our proteomics core facility, which also performed subsequent analyses. Gel samples were washed with water and dehydrated in acetonitrile. Reduction with DTT, cysteine alkylation with iodoacetamide, and digestion with trypsin were done in-gel. After extraction of peptides using acetonitrile and formic acid LC-MS analysis was performed using Finnigan LTQ-Obitrap Elite hybrid mass spectrometer system and Mascot software for identification of peptides.
PDI activity
PDI was desalted using DyeEx 2.0 columns (Qiagen, Valencia, USA). 150nM of PDI were mixed with inhibitor at different concentrations in buffered (pH7.1), 50mM K2HPO4/KH2PO4 solution and kept on ice for 30 minutes. One sample was prepared without PDI to monitor background, non-enzymatic reduction of Di-E-GSSG during assessment of PDI activity. Di-E-GSSG was added into the reaction mix at a final concentration of 100nM and samples were transferred into white multiwell plates (Becton Dickson Labware, Franklin Lakes, USA). To start reactions freshly prepared DTT was added at a final concentration of 100µM using a multichannel pipettor immediately before placing into a Synergy H1 plate reader (Biotek, Winoosky, USA) for kinetic analysis using excitation at 518nm, emission at 545nm, reads of 0.1sec/well, every minute, at 25 °C, for 30 minutes. Baseline fluorescence was determined from Di-E-GSSG reactions without PDI and without DTT.
PDIA1 mutation
Both active sites of PDIA1 (NM_000918.3) were mutated the same way, from lysine to glutamine (K57Q and K401Q) or to glutamic acid (K57E and K401E) and in cysteine to serine mutants both cysteines in both active sites were mutated (C53S, C56S and C397S, C400S). The following primers were used for site directed mutagenesis:
Active site a:
M1: 1_KtoQ_R ; GCCAGAGCCTGGCAGTGGCCACACCAAGGGG;
1_KtoQ_F; CCCCTTGGTGTGGCCACTGCCAGGCTCTGGC;
M2: 1_CtoS_CtoS_R; GCCAGAGCCTTGCTGTGGCCACTCCAAGGGG;
1_CtoS_CtoS_F; CCCCTTGGAGTGGCCACAGCAAGGCTCTGGC;
M3: 1_KtoE_R; GCCAGAGCCTCGCAGTGGCCACACCAAGGGG;
1_KtoE_F; CCCCTTGGTGTGGCCACTGCGAGGCTCTGGC;
Active site a’:
M1: 2_KtoQ_R GCCAACTGTTGGCAGTGACCACACCATGGGG;
2_KtoQ_F; CCCCATGGTGTGGTCACTGCCAACAGTTGGC;
M2: 2_CtoS_CtoS_R; GCCAACTGTTTGCTGTGACCACTCCATGGGG;
2_CtoS_CtoS_F; CCCCATGGAGTGGTCACAGCAAACAGTTGGC;
M3: 2_KtoE_R; GCCAACTGTTCGCAGTGACCACACCATGGGG;
2_KtoE_F; CCCCATGGTGTGGTCACTGCGAACAGTTGGC;
Page | 2