Post-translational regulation of CD133 by ATase1/ATase2-mediated lysine acetylation

Abstract

The CD133 cell-surface protein expresses the AC133 epitope that is associated with cancer progenitor cells and cancer resistance to traditional anticancer therapies. We report that the endoplasmic reticulum Golgi intermediate compartment residing acetyltransferases, ATase1 (NAT8B) and ATase2 (NAT8), can physically interact with CD133 to acetylate the protein on three lysine residues predicted to reside on the first extracellular loop of CD133. Site-directed mutagenesis of these residues mimicking a loss of acetylation and downregulation or inhibition of ATase1/ATase2 resulted in near-complete abolishment of CD133 protein expression. We also demonstrate that targeting ATase1/ATase2 results in apoptosis of CD133 expressing acute lymphoblastic leukemia cells. Taken together, we suggest that lysine acetylation on predicted extracellular residues plays a key role in expression and trafficking of CD133 protein to the cell surface and can be targeted to disrupt CD133 regulation and function.

Keywords: ERGIC; membrane protein; post-translation modifications; protein processing; trafficking.

Figure 1. CD133 can be lysine acetylated

a) MS analysis of CD133 Flag immunoprecipitated from HEK 293/CD133-VA lysates identified CD133 to be acetylated at K216, K248 and K255. MS analysis was performed as previously described (12) with the addition that post-translational modifications for acetylation were also detected. b) Mapping of acetylated lysine sites on the predicted topology of the CD133 protein at the plasma membrane. c) Endogenous immunoprecipitation of CD133 from Caco-2 cell lysate was analyzed by Western Blot for lysine acetylation of endogenous CD133 protein. Acetylated CD133 was detected using an anti-Ac-lysine antibody (7F8, Santa Cruz Biotechnology Inc.). d) Lysates from Flag-tagged wild type CD133-F and the CD133 K-to-Q Caco-2 stable cells were Flag immunoprecipitated and analyzed by Western Blot to determine CD133 lysine acetylation. IP of CD133 was performed as previously described (12) with exception of 0.5% Triton X-100 was used for IP of CD133 for MS analysis and RIPA buffer was used for IP of CD133 for in vitro acetylation experiments in an effort to disrupt any protein-protein interactions. CD133 K-to-Q mutant was generated as previously described (3) with the following primer setsCD133(K255Q) sense 5′-CAT GGC AAC AGC GAT CCA AGA GAC CAA AGA GGC G-3′ and antisense 5′-CGC CTC TTT GGT CTC TTG GAT CGC TGT TGC CAT G-3′; CD133(K248Q) sense 5′-CCT GTT CTT GAT GAG ATT CAA TCC ATG GCA ACA GCG ATC-3′ and antisense 5′-GAT CGC TGT TGC CAT GGA TTG AAT CTC ATC AAG AAC AGG-3′; CD133(K216Q) sense 5′-CAA CAC TAC CAA GGA CCA AGC GTT CAC AGA TCT GAA C-3′ and antisense 5′-GTT CAG ATC TGT GAA CGC TTG GTC CTT GGT AGT GTT G-3′

Figure 2. CD133 requires lysine acetylation for its stability

a) Flag immunoblotting of lysates from Caco-2 cells and AC133 immunoblotting of lysates from HEK293 cells stably expressing either CD133-VA wild type or K-to-R mutant, or an empty vector. Immunoblotting of actin was used as loading control. CD133 K-to-R mutant was generated as previously described (3) with the following primer sets: CD133(K255R) sense 5′-GGC AAC AGC GAT CAG AGA GAC CAA AGA GGC G-3′, antisense 5′-CGC CTC TTT GGT CTC TCT GAT CGC TGT TGC C-3′; CD133(K248R) sense 5′-GTT CTT GAT GAG ATT CGG TCC ATG GCA ACA GCG-3′ and antisense 5′-CGC TGT TGC CAT GGA CCG AAT CTC ATC AAG AAC-3′; CD133(K216R) sense 5′-CAC TAC CAA GGA CCG GGC GTT CAC AGA TC-3′ and antisense 5′-GAT CTG TGA ACG CCC GGT CCT TGG TAG TG-3′;. b) Caco-2 stables were analyzed for versatile affinity tagged CD133 transcript levels by quantitative PCR with the following primers CD133-F: sense, 5′-GGA CGT GTA CGA TGA TGT TG-3′ and antisense, 5′-CAC CGT CAT GGT CTT TGT AG-3′. Primers do not recognize endogenous CD133. Transcript levels were normalized to actin and relative to CD133 VA wild type. Error bars represent the standard deviation of three independent replicates (n=3). p-value was calculated against the control using a two-tailed Student’s t-test. c) FACS analysis of HEK293 cells stably expressing either CD133-VA wild type or the K-to-R mutant, or an empty vector stained with AC133-APC. Cell staining for flow cytometry and analysis was performed as previously described (9). d) Flag (against CD133-VA) and CANX immunofluorescence of HEK293 cells stably expressing either CD133-VA wild type or the K-to-R mutant. Cells were stained with Hoechst 33342 to visualize nuclei. Immunofluorescence were performed as previously described (3). Scale bar, 25μm. e) Wild type and K-to-R mutant CD133 expressing HEK293 cells were treated with either vehicle only or with MG132 for 24 hours as previously described (4).

Figure 3. ATase1 and ATase2 physically interact with CD133 in the ERGIC compartment

a) Membrane yeast two-hybrid (MYTH) was performed as previously described (4, 8). Generation of NubG fused ATase1 or ATase2 was performed by PCR amplification from pDONR223-NAT8 and pDONR223-NAT8B (Open Biosystems Inc.), respectively, with the following primers for N’terminal tagged NubG: sense 5′-GGT GGT CCA TAC CCA TAC GAT GTT CCA GAT TAC GCT GCT CCT TGT CAC ATC CGC AAA TAC-3′ and antisense 5′-GTA AGC GTG ACA TAA CTA ATT ACA TGA CTC GAG TCA CAG ACT CCC TAC C-3′ and C’terminal tagged NubG: sense 5′ CAA TAT TTC AAG CTA TAC CAA GCA TAC AAT CAA CTC AAT GGC TCC TTG TCA CAT CCG-3′ and antisense 5′-GGA GCG TAA TCT GGA ACA TCG TAT GGG TAC ATA TCC AGA CTC CCT ACC TTA GAA G-3′. PCR products were introduced into the pPR3-N or pPR3-C MYTH prey vectors by yeast homologous recombination as previously described (8). MFα-CD133-Cub-T or the control MFα-CD4-Cub-T were expressed in the yeast strain THY.AP4 along with NubG-ATase1, NubG-ATase2, ATase1-NubG or ATase2-NubG and assayed for growth on selective media. MFα-CD133-Cub-T and MFα-CD4-Cub-T were also expressed in THY.AP4 along with positive controls Ost1-NubI and Fur4-NubI and the negative controls Ost1-NubG and Fur4-NubG to demonstrate absence of self-activation and proper membrane integration. b) CD133 (pDONR223-CD133) (3), ATase1 (pDONR223-NAT8B) and ATase2 (pDONR223-NAT8) (Open Biosystems Inc.) were introduced into Gateway-compatible protein complementation assay (PCA) vectors. Vectors were transfected into HEK293 cells and immunofluorescence was performed as previously described (4), using the anti-ERGIC-53 antibody (C-6, Santa Cruz Biotechnology Inc.). CD133-VF1 and either VF2 only, VF2-ATase1 or VF2-ATase2 were co-transfected into HEK293 cells. 48 hours post-transfection, immunofluorescence was performed to determine co-localization between ERGIC-53, CD133-VF1 and either VF2 only, VF2-ATase1 or VF2-ATase2. Cells were stained with Hoechst 33342 to visualize nuclei. Scale bar, 10μm.

Figure 4. ATase1 and ATase2 regulate CD133 stability

a) An in vitro acetylation assay was performed as previously described (6). Briefly, affinity-purified wild-type or K-to-Q mutant CD133 was incubated with [³H]acetyl-CoA (American Radiolabeled Chemicals) in the presence of the acetyl-CoA:lysine acetyltransferase, ATase1 or ATase2, purified using the ProFound kit (Pierce-Thermo Scientific). The reaction was performed in 200 μl of acetylation buffer (50 mM Tris-HCl (pH 8.0), 0.1 mM EDTA, 1 mM dithiothreitol, 10% glycerol, 20 μM acetyl-CoA) for 1 h at 30°C. The reaction was stopped by adding an equal volume of ice-cold acetylation buffer and immediate immersion in ice. Following the reaction, CD133 was immunoprecipitated using anti-FLAG magnetic beads (Sigma-Aldrich) and then counted on a liquid scintillation counter. As a control, CD133 was incubated with [³H]acetyl-CoA in the absence of the enzyme (ATase1 or ATase2). Error bars represent the standard deviation of three independent replicates (n=3). b) HEK293/CD133-F and Caco-2 cells were treated with two independent shRNAs targeting either ATase1 (shATase1-1: TRCN0000035450, ATase1-2:TRCN0000035451, Sigma-Aldrich Inc.) or ATase2 (shATase2-1: TRCN000035564 and shATase2-2: TRCN0000035568 Sigma-Aldrich Inc). Lentiviral production and infection was performed as previously described (3). Lysates were analyzed by immunoblotting for CD133 using the anti-AC133 antibody. ATase1 and ATase2 knockdown at the protein level was monitored by immunoblotting using the anti-NAT8 (Ap4967c, Abgent Inc.) and anti-NAT8B (ab97885, Abcam Inc.) antibodies. Immunoblotting of actin was used as loading control. c) ATase1, ATase2, and CD133 transcript levels were monitored by quantiative PCR in the presence of shATase1 and shATase2 knockdown using primers for ATase1 or ATase2 (6). Transcript levels were normalized to actin and are relative to the control shRNA. Error bars represent the standard deviation of four independent replicates (n=4). *p<0.05, #p<0.01. p-value was calculated against the control using a two-tailed Student’s t-test. d) FACS analysis of Caco-2 cells transduced with either a control shRNA, shATase1-1 or shATase2-1. Cell staining for flow cytometry and analysis was performed as previously described (9).

Figure 5. Targeting ATase1/ATase2 results in CD133 downregulation and apoptosis

a) Caco-2 cells were treated with either DMSO or 10μM of compound 19 for 24 hours. Lysates were analyzed by immunoblotting for CD133 using the anti-AC133 antibody. b) SEM-K2 cells were treated for 24 hours with either DMSO or 10μM compound 19. Cells were co-stained with AC133-APC and Annexin-488 and analyzed by flow cytometry as previously described (9).