1,2-dilinolenoyl-3-(4-aminobutyryl)propane-1,2,3-triol

Similar image processing procedures were employed for the apo-state and the DAPT-bound data sets. We used MOTIONCORR (Li et al., 2013 (link)) for whole-frame motion correction, CTFFIND4 (Rohou and Grigorieff) for estimation of the contrast transfer function parameters, and RELION-1.4 (Scheres, 2012 (link)) for all subsequent steps. References for template-based particle picking (Scheres, 2015 (link)) were obtained from 2D class averages that were calculated from a manually picked subset of the micrographs. A 20 Å low-pass filter was applied to these templates to limit model bias. All low-pass filters employed were cosine-shaped and fell to zero within 2 reciprocal pixels beyond the specified frequency. To discard false positives from the picking, we used initial runs of 2D and 3D classification to remove bad particles from the data. The selected particles were then submitted to 3D auto-refinement, particle-based motion correction and radiation-damage weighting (Scheres, 2014 (link)). The resulting 'polished particles' were used for masked classification with subtraction of the residual signal as described in the main text, and the original particle images from the resulting classes were submitted to a second round of 3D auto-refinement. All 3D classifications and 3D refinements were started from a 40 Å low-pass filtered version of the high-resolution consensus structure. Fourier Shell Coefficient (FSC) curves were corrected for the effects of a soft mask on the FSC curve using high-resolution noise substitution (Chen et al., 2013 (link)). Reported resolutions are based on gold-standard refinement procedures and the corresponding FSC=0.143 criterion (Scheres and Chen, 2012 (link)). Prior to visualization, all density maps were corrected for the modulation transfer function (MTF) of the detector, and then sharpened by applying a negative B-factor that was estimated using automated procedures (Rosenthal and Henderson, 2003 (link)).
For the apo-state data set, the template-based algorithm picked 1.8 million particles from 2,925 micrographs, and 412,272 particles were selected after initial 2D and 3D classification. Subsequent 3D auto-refinement and particle polishing yielded a 3.5 Å map with fuzzy densities in the transmembrane region. Masked classification into eight classes with subtraction of the residual signal yielded three classes with good density as described in the main text. Poor reconstructed density was observed in the other five classes. Separate 3D auto-refinements of the corresponding particles in the original data set for the three best classes gave rise to reconstructions to 4.0– 4.3 Å resolution (also see Figures 2–3, Table 1).
For the DAPT-bound state, 1.4 million particles were picked from 2,206 micrographs, and initial classification selected 271,361 particles. After particle polishing, this subset gave rise to a 4.3 Å resolution map with relatively poor density in the transmembrane domain. Application of the masked classification procedure with residual signal subtraction into eight classes identified a single class with good density. After 3D auto-refinement, the corresponding 51,366 particles gave a map with a resolution of 4.2 Å, which showed improved density in the transmembrane domain.

ALDH⁻, ALDH⁺, and CD44⁺/CD24⁻ cells were first treated with vehicle control (DMSO) or treatment (ASR490, DAPT, MG132, or CQ) for prescribed doses and time points. Cell viability assays: Alamar blue (Life Technologies Corporation Eugene, OR) and EdU Cell Proliferation (using the EdU-Click 488 kit, cat# BCK-EdU488-1, Sigma), were then performed per the manufacturer’s instructions.

Human mammary immortalized cells (MCF10A), and the TNBC cell line MDA-MB-231 were purchased from American Type Culture Collection. MCF10A cells were grown in a 1:1 mixture of Dulbecco’s modified Eagle’s medium (DMEM) and Ham’s F12 medium with 20 ng mL⁻¹ human epidermal growth factor, 100 ng mL⁻¹ cholera toxin, 0.01 mg mL⁻¹ bovine insulin, 500 ng mL⁻¹ hydrocortisone, and 5% horse serum. MDA-MB-231 cells were grown in DMEM containing L-glutamine and sodium pyruvate, supplemented with 10% fetal bovine serum and 1% antibiotic and antimycotic solution in a humidified atmosphere of 5% CO2 at 37°C in an incubator. Human BCSC cells: ALDH⁺ and CD44⁺/CD24⁻, and BC cells: ALDH- were purchased from Celprogen (San Pedro, CA, United States) and maintained in human BCSC expansion and undifferentiation media. DAPT (γ-secretase), cycloheximide (CHX), chloroquine (CQ), and MG132 were purchased from Sigma (St. Louis, MO).

Cell lysates of ALDH⁻, ALDH⁺, and CD44⁺/CD24⁻ cells following treatment with vehicle and treatment (ASR490, DAPT, MG132, CHX, or CQ) for prescribed doses and time points, were prepared with RIPA buffer (Thermo Scientific, Rockford, IL, United States) per the manufacturer’s protocol. Western blotting was performed using specific antibodies against Notch1 (CST, #3608), HES1 (Sigma, #SAB2108472), Hey1 (Proteintech, #19929-1-AP), NFκB p65 (CST, #8242), Bcl-2 (CST, #15071), Bcl-xL (CST, #2764), Vimentin (CST, #46173), Slug (CST, #9585), E-Cadherin (CST, #3195), β-catenin (CST, #8480), Ubiquitin (CST, #3933), Cleaved-PARP (CST, #5625), Cleaved-caspase-9 (CST, #20750), BAX (CST, #41162), Notch2 (CST, #D76A6), Lamp1 (CST, #9091), and LC3B (Proteintech, #14600-1-AP). β-Actin (CST, #4970) was used as the loading control. Protein bands were visualized using the Bio-Rad ChemiDocTM imaging system. For IP experiments, protein samples were immunoprecipitated with Notch1 antibody as per the protocol described elsewhere (Chandrasekaran et al., 2020 (link)), and Western blots were performed with ubiquitin antibody.