|Home | About | Journals | Submit | Contact Us | Français|
Autophagy controls and executes the turnover of abnormally aggregated proteins. MAP1S interacts with the autophagy marker LC3 and positively regulates autophagy flux. HDAC4 associates with the aggregation-prone mutant huntingtin protein (mHTT) that causes Huntington's disease, and colocalizes with it in cytosolic inclusions. It was suggested HDAC4 interacts with MAP1S in a yeast two-hybrid screening. Here, we found that MAP1S interacts with HDAC4 via a HDAC4-binding domain (HBD). HDAC4 destabilizes MAP1S, suppresses autophagy flux and promotes the accumulation of mHTT aggregates. This occurs by an increase in the deacetylation of the acetylated MAP1S. Either suppression of HDAC4 with siRNA or overexpression of the MAP1S HBD leads to stabilization of MAP1S, activation of autophagy flux and clearance of mHTT aggregates. Therefore, specific interruption of the HDAC4-MAP1S interaction with short peptides or small molecules to enhance autophagy flux may relieve the toxicity of mHTT associated with Huntington's disease and improve symptoms of HD patients.
Mammalian histone deacetylases (HDAC) are lysine deacetylases, which are classified into three main groups based on their homology to yeast proteins.
HDAC4 belongs to group II subgroup A of the family . Huntington's disease (HD) is a fatal progressive neuro-degenerative disorder caused by an autosomal dominant mutation with expansion of more than 36 trinucleotide CAG repeats (which codes for polyglutamine) in exon 1 of the huntingtin (HTT) gene that encodes huntingtin (HTT) protein . HDAC4 associates with the mutant HTT (mHTT) and colocalizes with it in cytoplasmic inclusions . In mouse models of Huntington's disease, HDAC4 reduction delays cytoplasmic formation of mHTT aggregates and rescues neuronal and cortico-striatal synaptic function, but does not repair the global transcriptional dysfunction . However, the mechanism by which HDAC4 reduction delays cytoplasmic formation of mHTT aggregates is unknown.
Autophagy is a process that begins with the formation of isolation membranes that recognize and engulf substrates such as aggregated proteins to form autophagosomes. These autophagosomes migrate along acetylated microtubules and fuse with lysosomes to generate autolysosomes in which autophagosomal cargos are degraded [4, 5]. Defects in autophagy in neurons cause accumulation of aggregate-prone proteins such as mHTT whose toxicity results in neurodegeneration . Microtubule-associated protein 1S (MAP1S, previously called C19ORF5) associates with microtubules [7, 8]. Like its sequence homologues MAP1A and MAP1B, MAP1S interacts with mammalian autophagy marker LC3 [9–11], and bridges components involved in autophagy with microtubules to affect autophagosomal biogenesis and degradation . Depletion of MAP1S results in decreased levels of Bcl-2 and P27 and was proposed to reduce initiation of autophagy. Both the direct function of MAP1S association with microtubules through LC3 and indirect positive impact of MAP1S on autophagy initiation through Bcl-2 and P27 affect the overall rate of autophagy flux . A potential interaction between HDAC4 and MAP1S revealed in a yeast two-hybrid screen  triggered us to investigate whether HDAC4 regulates autophagy to reduce mHTT aggregates via modification of MAP1S.
We discovered that MAP1S interacts with HDAC4 through a HDAC4-binding domain (HBD) within the overlapping region between the short chain (SC) and the heavy chain (HC) of MAP1S. MAP1S had no effect on levels and subcellular distribution of HDAC4. However, HDAC4 decreased the stability of MAP1S by catalyzing deacetylation of acetylated MAP1S and further led to suppression of autophagy flux and accumulation of mHTT aggregates. Inhibition of HDAC4 or overexpression of HBD promoted stabilization of MAP1S and restored the MAP1Sregulated autophagy flux and degradation of mHTT aggregates. This reveals a new potential to treat Huntington's disease by interrupting the specific interaction between HDAC4 and MAP1S.
GFP-HTT72Q, a GFP-tagged mHTT variant, includes a polypeptide encoded by exon 1 of the Huntingtin gene plus 72 expanded polyglutamine [poly(Q)] repeats in the N-terminus [13, 14]. The acetylation of huntingtin at residue K444 promotes autophagic degradation of huntingtin itself . The K444 residue is out of the sequence covered by HTT72Q so that HTT72Q degradation is not affected by the acetylation of K444. Overexpression of HDAC4 in cells expressing GFPHTT72Q led to enhancement of GFP-HTT72Q fluorescence (Figure 1A,B) and levels of GFP-HTT72Q aggregates (Figure 1C,D). Increasing levels of HDAC4 distributed surrounding the HTT72Q aggregates (Figure (Figure1A).1A). Suppression of HDAC4 levels with HDAC4-specific siRNA led to reduction of mHTT aggregates (Figure 1E,F). Analyses with Agarose Gel Electrophoresis for Resolving Aggregates (AGERA) revealed the same results as in the normal immunoblot analyses of aggregates in stacking gel. Accumulation of another mHTT (GFP-HTT74Q) with a similar role as the GFPHTT72Q in Huntington's disease was observed when the GFP-HTT74Q and HDAC4 were transiently coexpressed in neuroblastoma Neuro-2a (N2a) cells (Figure 1G-I). Thus, the inhibition of HDAC4 greatly reduces the severity of aggregation of mHTT.
Consistent with a previous report , GFP-HTT72Q was degraded in lysosomes. Inhibition of lysosomal activity with Bafilomycin A1 (BAF) led to accumulation of mHTT aggregates together with the autophagic marker LC3-II (Figure (Figure1J).1J). The percentage of GFP-HTT72Q aggregates that overlapped with LAMP2-labeled lysosomes was significantly increased in the presence of BAF (Figure 1K,L). This suggested that small aggregates were efficiently degraded through lysosomes in the absence of Bafilomycin A1, but large aggregates accumulated because of the compromised lysosomal degradation.
We then tested whether inhibition of HDAC4 enhanced autophagy that would consequently promote degradation of mHTT aggregates. Overexpression of HDAC4 led to a reduction in levels of MAP1S, an enhancer of autophagy flux . Impairment of autophagy flux due to the HDAC4 overexpression was confirmed by reduced levels of LC3-II in HeLa cells in the presence of BAF (Figure 2A,B). Suppression of HDAC4 with siRNA (Figure 2C,D) resulted in an increase in levels of MAP1S. This was accompanied by increased levels of LC3-II in the presence of BAF. The HDAC4 overexpression-triggered impairment of autophagy flux and HDAC4 suppression-induced activation of autophagy flux were further confirmed in N2a cells (Figure 2E-H). Autophagy flux measured at the cellular level by punctate foci of RFP-LC3 observed by fluorescent microscopy was impaired by HDAC4 overexpression (Figure 2I,J). Autophagy flux measured by autophagy vacuoles observed under the electron microscope was impaired by HDAC4 overexpression (Figure 2K,L) or enhanced by HDAC4 suppression (Figure 2M,N). Interestingly, the vacuoles in the MAP1S suppressed cells in the presence of BAF relatively contained less substrate contents (Figure (Figure2M),2M), indicating a deficiency in autophagy cargos when initiation of autophagy is exceedingly activated as others described . Thus, HDAC4 activity inhibits autophagy flux and inhibition of the activity enhances autophagy flux.
To determine whether the regulation of autophagy flux by HDAC4 depended on MAP1S, we enhanced the expression of HDAC4 in wild-type and MAP1S−/− MEF cells or suppressed the expression of HDAC4 with siRNA in wild-type and MAP1S−/− HeLa cells generated with the CRISPR/Cas9 system. We found that impairment of autophagy flux by overexpression of HDAC4 or the activation of autophagy flux by HDAC4 suppression was only evident in wild-type cells where MAP1S was present; and effects of alteration of HDAC4 levels were abrogated in MAP1S−/− cells (Figure 3A-D). Suppression of MAP1S with siRNA in the HeLa cell line stably expressing mHTT led to an inhibition in autophagy initiation (Figure 3E,F), further confirming our previous report that MAP1S impacts autophagy on both autophagy initiation and autophagosome-lysosome fusion . Consequently, the defects in autophagy triggered by silencing MAP1S caused an accumulation of GFP-HTT72Q aggregates (Figure 3G,H). Accumulation of GFP-HTT74Q was also observed in MAP1S−/− MEF cells transiently expressing the GFP-HTT74Q. The MAP1S-deficiency-triggered accumulation of HTT74Q aggregates was confirmed when the same cell lysates were analyzed by AGERA (Figure 3I-L). Enhancement of mHTT aggregation by overexpression of HDAC4 was only observed in wild-type and not in MAP1S−/− MEF cells (Figure 3M,N). These results indicate that HDAC4-mediated suppression of degradation of mHTT aggregates by autophagy is MAP1S-dependent.
Full-length MAP1S (FL) is processed by posttranslational modification to multiple isoforms that include heavy chain (HC), short chain (SC) and light chain (LC) (Figure (Figure4A)4A) [7, 11, 18]. Using a MAP1Sspecific monoclonal antibody 4G1 for immunoprecipitation, we detected complexes of endogenous HDAC4 and MAP1S in HeLa cells (Figure (Figure4B).4B). Further examination of the interaction in brain tissue lysates from wild-type and MAP1S−/− mice and N2a cells revealed that HDAC4 was co-precipitated with MAP1S (Figure 4C,D). The interaction was characterized in more detail with either HA- or MAP1S-specific antibody in 293T cells transiently overexpressing both HDAC4 and HA-tagged MAP1S (HA-MAP1S) (Figure 4E,F). Interaction of HDAC4 with MAP1S HC and SC products in addition to FL was apparent (Figure (Figure4G).4G). This suggested that a domain located between R653 and Q855 (Figure (Figure4A)4A) was necessary and sufficient for the interaction between HDAC4 and MAP1S. This is the region of overlap between HC and SC. MAP1S LC, which lacks this domain, exhibited no interaction with HDAC4 (Figure (Figure4H).4H). The isoform-specific interaction was further confirmed using purified MAP1S SC and LC variants tagged with GST (GST-SC or GST-LC) to pull down HDAC4 from lysates of 293T cells overexpressing HDAC4. Notably, the GST-SC pulled down HDAC4, but GST-LC did not (Figure 4I,J). Using the HA-tagged R653-Q855 fragment (HA-HBD) alone further indicated HBD located within the fragment (Figure (Figure4K).4K). Deletion of HBD in MAP1S led to the abolishment of its interaction with HDAC4 (Figure (Figure4L).4L). Taken together these results indicate that MAP1S interacts with HDAC4 within cells and the interaction occurs through HBD.
To assess potential function of the HDAC4-MAP1S interaction, we tested the effect of cellular levels of one on the other. The siRNA-mediated knockdown of MAP1S in HeLa cells, transgenic deletion of MAP1S in MEF cells, or overexpression of MAP1S isoforms in HeLa cells had no obvious impact on the levels of HDAC4 (Figure 5A-C). Transfection of 293T cells with increasing amounts of HDAC4 plasmid caused a dose-dependent decrease in MAP1S levels (Figure 5D,E). In contrast, suppression of HDAC4 expression with siRNA caused an elevation of MAP1S levels in both HeLa and COS7 cells (Figure (Figure5F).5F). HDAC4-dependent changes in levels of MAP1S occurred strictly on levels of the protein. No change in MAP1S mRNA levels was observed (Figure (Figure5G).5G). MAP1S stability was increased when HDAC4 was suppressed (Figure 5H,I). This and the sequence-specific direct interaction of HDAC4 with MAP1S suggests HDAC4 may control levels of MAP1S protein by direct modifications affecting stability.
To demonstrate that HDAC4 activity directly affected degree of acetylation of MAP1S, immunoprecipitates of HA-MAP1S were incubated with immuno-purified active HDAC4. Levels of acetylated MAP1S were inversely proportional to increasing levels of HDAC4 (Figure 6A,B) as well as time of incubation with the enzyme (Figure 6C,D). In contrast, no dramatic decrease in levels of acetylated MAP1S was observed when the immunoprecipitates were incubated with heatinactivated HDAC4 (Figure 6C,D).
We then investigated the degree of MAP1S acetylation when incubated with gain or loss of function mutants of HDAC4 [19, 20]. As expected, mutant H976Y that has a reported gain in activity relative to wild-type HDAC4 exhibited significantly higher MAP1S-deacetylation activity and lower levels of acetylated MAP1S than wild-type HDAC4. HDAC4 mutant H803A and D840N that have compromised catalytic activity each exhibited significantly lower MAP1S-deacetylation activity and higher levels of acetylated MAP1S than wild-type HDAC4. Unexpectedly, incubation of MAP1S with the H976F mutant resulted in decreased deacetylation and increased levels of acetylated MAP1S relative to wildtype HDAC4 (Figure 6I,J). The H976F mutant has been reported to be similar to wild-type HDAC4 using other substrates. These results show clearly that acetylated MAP1S is regulated by HDAC4.
To further confirm that the impact of HDAC4 on autophagy and degradation of mHTT aggregates is specifically through its association with MAP1S, we compared the impacts of HBD with HBDΔ on MAP1Smediated autophagy turnover of mHTT aggregates. When either HBD or HBDΔ was overexpressed in HeLa cells, HBD but not HBDΔ reduced the amount of endogenous MAP1S co-precipitated with HDAC4 (Figure 7A,B), suggesting that HBD competed with full length MAP1S for binding with HDAC4. In such a way, overexpression of HBD but not HBDΔ enhanced the stability of endogenous MAP1S (Figure 7C,D) and increased the levels of endogenous MAP1S protein (Figure 7E,F). Consequently, autophagy flux represented by LC3-II levels in the presence of Bafilomycin A1 was enhanced by HBD but not HA-HBDΔ and such enhancive effect was only observed in the presence of endogenous MAP1S (Figure 7E,G). The HDAC4 overexpression-induced deceases in MAP1S levels were prevented when cells expressed HBD but not HAHBDΔ (Figure 7H,I), and HDAC4 overexpression-induced impairment of autophagy flux was reactivated in the presence of HBD but not HA-HBDΔ (Figure 7H,J). Overexpression of HBD but not HA-HBDΔ protected stability of endogenous MAP1S and led to a significant decrease in levels of mHTT aggregates (Figure 7K-M).
Therefore, the HBD interrupts MAP1S-HDAC4 interaction and promotes MAP1S-mediated autophagy turnover of mHTT aggregates (Figure (Figure7N7N).
HDAC-mediated histone deacetylation is a key epigenetic modification that has attracted enormous attention since histone deacetylases emerged as a druggable class of enzymes . Non-histone protein acetylation has also attracted attention because of the demonstration that histone deacetylase 6 (HDAC6) is involved in microtubule-deacetylation and regulation of autophagy and mitophagy [22–24]. Originally, HDAC4 was characterized as a histone deacetylase regulating transcription factors MEF2  and Runx2 . It was reported that HDAC4 alone does not show deacetylase activity on histone substrate but regulates histone deacetylation through its interaction with HDAC3 . Deletion of HDAC4 in mouse brain was reported to have no effect on histone acetylation profiles and global transcription . It seems that HDAC4 regulates transcription of a specific set of genes by affecting the stability of the related transcriptional factors instead of the general epigenetic modification of genomeassociated histones. Recently, HDAC4 was found to be associated with aggregate-prone mutant Huntington's disease–associated products of the huntingtin gene (HTT) that have long polyglutamine stretches . The significance of the HDAC4-huntingtin association has not been assessed at the mechanistic level. Here we present evidence that HDAC4 impairs the degradation of mHTT aggregates, interacts directly with autophagy activator MAP1S, reduces MAP1S stability, consequently suppresses the autophagy flux mediated by MAP1S, and impairs the degradation of mHTT aggregates. Inhibition of HDAC4 resulted in stabilization of MAP1S, activation of MAP1S-mediated autophagy flux and fast degradation of mHTT aggregates. No matter HDAC4 directly or indirectly through HDAC3 exerts its deacetylase activity on MAP1S, this provides a mechanism by which therapeutic reduction of HDAC4-associated activity could reduce accumulation of cytoplasmic mHTT aggregates and ameliorate neurodegeneration .
In this study, we implicated and established a mechanism for the role of MAP1S in the development of Huntington's disease through its role in multiple steps of autophagy. Our results show that it is the inhibition of HDAC4 and its associated deacetylase activity that enhances the stability of MAP1S, increases autophagy flux and improves clearance of mHTT aggregates. HDAC4 interacts with mHTT but not wildtype HTT, and is found enriched in mHTT aggregates . Wild-type HTT protein normally is found in association with microtubules  and thus may regulate vesicle trafficking in neurons . The homology between regions of HTT and the yeast autophagy regulatory proteins predicted a normal regulatory function in autophagy , which has been confirmed [31, 32]. The mHTT disrupted the normal HTT-mediated regulation of autophagosomal dynamics and caused defects in cargo degradation . We reason that soluble mHTT protein may function similar to wild-type HTT in association with motor protein complexes on microtubules. However, mHTT sequesters HDAC4 on microtubules, HDAC4 impairs the stability and microtubule-associated functions of MAP1S through deacetylation and consequently interrupts autophagy flux. This defect leads to accumulation of mHTT and consequent formation of mHTT aggregates. The mHTT aggregates further sequester HDAC4 that further protects the aggregates from degradation by MAP1S-mediated autophagy.
Although there is no major treatment has been developed specifically for Huntington's disease, enhancing clearance of mHTT aggregates has been considered one of the feasible approaches to slow the neurodegeneration in Huntington's disease and has revealed some beneficial effects in animal models . HDAC4 reduction in mouse model by transgenic approach has been shown to delay cytosolic mHTT aggregate formation and improve motor coordination, neurological phenotypes and longevity . This presents a novel strategy to target huntingtin aggregates using small molecules to inhibit HDAC4 activity. Our results provide a detail mechanism for HDAC4 to impact on the degradation of mHTT. Since a series of cellular processes have been reported to be regulated by HDAC4, any small molecule inhibitor of HDAC4 will inhibit the general activity of HDAC4 and is surely expected to cause unnecessary side effect besides the clearance of mHTT aggregates. Here we have identified a HDAC4-binding domain from MAP1S that specifically interrupts the HDAC4-MAP1S interaction and protects MAP1S from being deacetylated by HDAC4-associated deacetylase activity. Thus, further development of short peptides to disrupt the HDAC4-MAP1S interaction will specifically enhance the MAP1S-mediated autophagy clearance of mHTT aggregates.
Antibody against MAP1S (4G1) was a gift from Precision Antibody (AG10006). Antibody against acetylatedtubulin (6–11B-1) was from abcam. Antibodies against HDAC4 (7628), acetylated lysine (9441), Myc (2276) and Huntingtin (2773) were from Cell Signaling Technology. Antibody against HA-tag (MMS-101P) was from Covance. Antibody against acetyl-lysine (05-515) and anti-acetyl lysine agarose conjugate (16–272) were from EMD Millipore. The siRNAs specific to human HDAC4 (sc-35540) and mouse HDAC4 (sc-35541) were from Santa Cruz Biotechnology. Protease inhibitor cocktail and anti-Flag M2-agarose were from Sigma. Other antibodies, siRNAs and reagents were described by Zou et al. .
RFP-LC3 was a gift from Dr. Mizushima . HA-MAP1S isoforms (HA-LC, HA-SC, HA-HC and HA-FL) have been described earlier . HA-MAP1S R653-Q855 fragment, representing the HDAC4-binding domain (HBD) of MAP1S was encoded by a PCR-amplified MAP1S insert in HA-PCMV plasmid (631604, Clontech Laboratories Inc.). Plasmids encoding GFP–HDAC4, Flag-HDAC4, Flag-HDAC4 H976Y, Flag-HDAC4 H976F, Myc-HDAC4 H803A, or Myc-HDAC4 D840N were previously reported [19, 20, 37]. Plasmid encoding GFP-HTT74Q (#40262) was purchased from Addgene. To delete the HBD fragment from the FL MAP1S to generate the HA-HBDΔ, a pair of primers (5′-CTCGCTGCCCTCTGCGGGGCT-3′ and 5′-ACGGA GAACGTCAGCCGCACC-3′) were phosphorylated with T4 polynucleotide kinase (NEB, M0201) and mixed with HA-FL plasmid template to amplify the HA-HBDΔ by PCR using the KOD hot start DNA polymerase from TOYOBO. The Restriction enzyme DpnI was added to the PCR reaction mixture to digest the template. T4 DNA Ligase (NEB, M0202) was used to ligate the PCR product to become a circular plasmid that was verified by DNA sequence.
Lenti-X™ Tet-Off® Advanced Inducible Expression System (Clontech, Cat: 632163) was used according to the manufacturer's instruction to generate the inducible GFP-HTT72Q stable cell lines. Briefly, an EGFP-tagged HTT fragment encoded by human HTT exon 1 plus 72Q repeat were amplified from the pUAST-Httex1-Q72-eGFP construct as previously described  and subcloned into the pLVX-Tight-Puro vector. Lenti-viruses containing Httex1-Q72 plasmid (pLVX-tight-Q72) and the regulator plasmids (pLVX-Tet-OFF advanced) were produced using the Lenti-X HTX packaging system (Clontech) and used to infect HeLa cells as instructed. Puromycin (2 ug/ml) and Neomycin (200 ug/ml) were applied to screen for positive clones, which were being maintained in the “off” state in the presence of 100 ng/ml Dox during the whole selection process to turn off the expression of potentially toxic HTT72Q proteins.
Guide RNAs targeting human MAP1S gene were designed using Optimized Crispr Design (http://crispr.mit.edu/). Synthesized DNA oligos were inserted into crispr/cas9 vector pSpCas9(BB)-2APuro (PX459) (Addgene, #48139). HeLa cells were transiently transfected with a pool of five plasmids encoding Cas9 nuclease and guide RNAs targeting for MAP1S or the vector for wild-type control. Cells were selected with 1.5 μg/ml Puromycin starting at 48 hours after transfection. Multiple monoclonal single cell clones were picked and cultured individually in separate wells. Immunoblot analysis of MAP1S levels was used to determine the knockout efficiency. The sequences of DNA oligos for gRNAs are
Cell lines used for transfection included HeLa, HEK (human embryonic kidney)-293T, COS7 cells, HeLa cells stably expressing ERFP–LC3 (HeLa-RFP-LC3), N2a, or MEF cells that were established as described [11, 36, 39]. Cell transfection and immunoblot analysis were performed as previously described .
HeLa or HeLa-RFP-LC3 cells were fixed and processed for fluorescence microscopy analysis as previously described . For transmission electron microscopy, HeLa cells transfected with Flag-HDAC4 plasmid or HDAC4-specific siRNA were treated with 10 nM BAF for 12 hrs. As previously described , cells were fixed and processed for examination with a JEM 1010 transmission electron microscope (JEOL). ImageJ software was used to measure percentages of areas occupied by autophagy vacuoles.
As described previously , cells were lysed in isolation buffer (50 mM Tris-HCl, pH 8.0, 100 mM NaCl, 5 mM MgCl2, 0.5% NP-40, and protease inhibitors). Insoluble pellets were isolated after 10 min centrifugation at 14000 rpm at 4°C and resuspended in buffer containing 20 mM Tris-HCl, pH 8.0, 15 mM MgCl2 and 0.5 mg/ml DNase, and then incubated at 37°C for 1 hr. Insoluble fractions were diluted in loading buffer and boiled for 5 min for immunoblot analysis. Agarose Gel Electrophoresis for Resolving Aggregates (AGERA) was performed following a described protocol . Briefly, 100 μg of total protein was loaded to a 1.5% agarose gels containing 0.1% SDS for AGERA. Gels were run at 100 V and semi-dry transfers were conducted at 200 mA for 1 h. After transfer, PVDF membranes were used for immunoblot analyses.
Cell lysates were subjected to immunoprecipitation with antibody against MAP1S, HDAC4, or their respective IgG control antibody as described before [35, 42]. For GST pull-down assay, 293T cells overexpressing Flag-HDAC4 or a control were lysed with lysis buffer containing 2 mM Dithiothreitol. Purified GST or GSTtagged proteins (7.5 μg) bound to 25 μl GSH-Sepharose 4B beads  were mixed with 200 μg pre-cleaned cell lysates, and each mixture was rotated for 1 hr at 4°C and then washed three times with lysis buffer. The pellets were resuspended in 50 μl of lysis buffer containing loading buffer and boiled for 10 min for immunoblot analyses.
Lysis buffer for in vitro deacetylation assays contained 50 mM HEPES, pH 7.5, 150 mM NaCl, 1 mM EDTA, 2.5 mM EGTA, 0.1% triton X-100 and 10% glycerol. As described previously , Flag-HDAC4, Flag-HDAC4 H976Y, Flag-HDAC4 H976F, Myc-HDAC4 H803A, or Myc-HDAC4 D840N was overexpressed in 293T cells and purified with conjugated anti-Flag M2-agarose or anti-Myc antibody and Protein G-agarose beads. HA-MAP1S was over-expressed in 293T cells and purified with anti-MAP1S antibody (4G1). The agarose beads containing immunoprecipitated HDAC4 were mixed with agarose containing HAMAP1S in deacetylation buffer (10 mM Tris-HCl, pH 8.0, 150 mM NaCl and 10% glycerol). Each mixture was incubated at 37°C for an indicated amount of time; each reaction was terminated with loading buffer.
Total RNA was extracted from HeLa cells transfected with HDAC4-specific siRNA or HDAC4 expression plasmid with Trizol reagent (Invitrogene#15596–026) according to the manufacturer's instructions. The Invitrogene SuperScript III First-Strand System was used for reverse transcription with random primers. Real-time PCRs were performed with SYBR Premix ExTaq (TaKaRa RR820A). Primers for human MAP1S included a forward primer 5′-CGCTGGAAGAACTCCTCATC-3′ and a reverse primer 5′-GAGTGAGCCCAGTGAGAAGG-3′ and those for human β-Actin included a forward primer 5′-ACTCTTCCAGCCTTCCTTCC-3′ and a reverse primer 5′-CAGTGATCTCCTTCTGCATCC-3′. The relative mRNA levels of MAP1S were quantified by normalizing the amount of MAP1S mRNA to amount of β-actin mRNA.
We thank Drs. Tony Kouzarides from University of Cambridge, UK for providing plasmids that express HDAC4 mutants H803A and D840N, Noboru Mizushima from Tokyo Medical and Dental University, Japan for the LC3 cDNA, Joe Corvera from A&G Pharmaceuticals, Inc., Columbia, MD, USA for anti-MAP1S mouse monoclonal antibody 4G1 which is now sold by Precision Antibody™ with catalog number of AG10006. TEM service was provided by the High Resolution Electron Microscopy Facility in the M. D. Anderson Cancer Center.
This work was supported by a National Institute of Health NCI R01CA142862 to Leyuan Liu, the National Natural Science Foundation of China (No: 81472382) to Hai Huang, Natural Science Foundation of Guangdong Province and Science of China (2015A030310530) to Guibin Xu. The funders had no role in study design, data collection and analysis, decision to publish, or preparation of the manuscript.
Conflict of interest statement
The authors declare no competing financial interests.