N-terminal acetylation and methylation differentially affect the function of MYL9

Recombinant protein expression and in vitro enzyme assays

Human NRMT1 (rhNRMT1) and NAA10 (rhNAA10) were expressed as His₆-tagged recombinant proteins and purified as previously described [45,46]. Briefly, protein ORFs (GE Dharmacon, Marlborough, MA) were amplified and subcloned into the pET15b vector (EMD Millipore, Billerica, MA), expressed in BL21 Star (DE3) Escherichia coli (Thermo Scientific, Waltham, MA), and purified on Ni²⁺-NTA beads (Qiagen, Hilden, Germany).

In vitro methyltransferase experiments were carried out using the MTase-Glo Methyltransferase Assay (Promega, Madison, WI) following manufacturer guidelines. Briefly, 0.2 μM rhNRMT1 was incubated at room temperature (RT) with synthetic peptide substrates corresponding to the 14 N-terminal amino acids of wild type (WT) or mutant MYL9 (Figure 1A) (AnaSpec, Fremont, CA) in the presence of 40 μM s-adenosyl methionine (SAM). Peptide substrates were used at concentrations ranging from 0 to 160 μM. Reactions were stopped at 20 min by addition of trifluoroacetic acid. Methyltransferase activity was then measured by the accumulation of a luminescent signal, which was generated by reaction of a detection reagent with the methyltransferase reaction product s-adenosyl-L-homocysteine (SAH).

In vitro acetyltransferase experiments were performed using the Enzo Acetyltransferase Activity Kit (Enzo Life Sciences, Farmingdale, NY) following manufacturer guidelines. Briefly, 0.1 μM rhNAA10 was incubated with N-terminal peptides of WT or mutant MYL9 at RT in the presence of excess acetyl-CoA. The reaction was stopped after 30 min by addition of ice-cold isopropyl alcohol. Acetyltransferase activity was measured by reading fluorescence (Ex 380 nm/Em 520 nm) generated by the reaction of a detection reagent with the acetyltransferase reaction product CoA.

All readings were taken on a Cytation5 cell imaging multi-mode reader (BioTek, Winooski, VT). For both methyltransferase and acetyltransferase assays, background signal was measured by the inclusion of no substrate control reactions and was subtracted from experimental reactions. Background subtracted signal intensities at varying substrate concentrations were then fit to a model of Michaelis–Menten enzyme kinetics and a K_M determination was made using GraphPad Prism 7 software (San Diego, CA).

Molecular cloning, mutagenesis, and lentivirus production

To generate the C-terminally tagged MYL9-Dendra2 fusion protein, the MYL9 protein ORF (GE Dharmacon) was amplified and subcloned into the pDendra2-N vector (Clontech, Mountain View, CA) using XhoI and HindIII restriction sites. Site-directed mutagenesis was performed using the Quikchange system (Agilent Technologies, Santa Clara, CA) to generate S3P (SPK) and K4Q (SSQ) mutant MYL9-Dendra2 fusion proteins. WT, SPK, and SSQ MYL9-Dendra2, as well as Dendra2 alone, were then amplified and subcloned into the pCDH-EF1-MCS-IRES-Puro lentiviral expression vector (System Biosciences, Palo Alto, CA) using XbaI and NheI restriction sites. The following forward primers and their reverse compliments were used for mutagenesis: S3P: 5′-GATCTCGAGATGTCCCCCAAGCGGGCCAAAGC-3′ and K4Q: 5′-CGAGATGTCCAGCCAGCGGGCCAAAGCC-3′. To generate C-terminally tagged MYL9-FLAG proteins, WT, SPK, and SSQ MYL9 were amplified from the MYL9-pDendra2 constructs and subcloned into the pWPI lentiviral expression vector (Addgene, Cambridge, MA) using 5′ and 3′ PmeI restriction sites. The following 3′ primer included the FLAG nucleotide sequence followed by a stop codon: 5′-CGGTTTAAACTCATTTATCATCATCATCTTTATAATCGTCGTCTTT ATCCTTGGCGC-3′ Lentivirus was generated through co-transfecting HEK293T cells with 50 μg of expression plasmid (pCDH or pWPI from above), 37.5 μg psPAX2 packaging vector, and 15 μg pMD2.G envelope plasmid using calcium phosphate transfection as previously described [47]. Viral supernatants were collected forty-eight hours after transfection, concentrated with 100 KDa molecular mass cut-off filters (Millipore Sigma, Burlington, MA), and titered in HEK293T cells. GFP expressed by the pWPI vector and Dendra2 expressed by the pCDH vector was used for titering.

Cell culture, transgene expression, and small molecule treatment

HEK293T human embryonic kidney and NIH 3T3 mouse embryonic fibroblast cell lines were cultured in Dulbecco’s Modified Eagle Medium (DMEM; Life Technologies, Grand Island, NY) with 10% fetal bovine serum (FBS; Atlanta Biologicals, Atlanta, GA) and 1% penicillin–streptomycin (P/S; Life Technologies). HCT116 human colorectal carcinoma and CRISPR/Cas9-mediated NRMT1 knockout (KO) HCT116 cell lines were cultured in McCoy’s 5A Modified Medium (Life Technologies) supplemented with 10% FBS and 1% P/S. NRMT1 KO cells were previously generated and verified by sequencing and protein expression [45]. Maintenance of NRMT1 loss was verified by Western Blot (Figure 4E). All cells were grown on tissue culture treated plastic and maintained at 37°C and 5% CO₂. HEK293T, NIH 3T3, and HCT116 cell lines were a generous gift from Dr. Ian Macara, Vanderbilt University.

For experiments involving transgene expression, cells were harvested with Trypsin-EDTA (Life Technologies), plated in fresh media, and transduced with lentivirus at a multiplicity of infection (MOI) of 2. Transgene expression was confirmed and cells were used for experiments 3–4 days after transduction. Cells transduced with the pCDH vector were selected for by treatment with 2 μg/ml puromycin 48 h after transduction, and were used for experiments 48 h after treatment. For experiments involving small molecule treatment, cells were cultured to 80% confluence, and media was changed to fresh media containing the desired small compound. Calcium ionophore A23187 (calcimycin; Sigma–Aldrich, St. Louis, MO) was resuspended in DMSO and used at a final concentration of 0.1 μM. Cells were treated with calcimycin for 45 min to induce phosphorylation of MYL9 by MLCK. Recombinant human tumor necrosis factor alpha (TNFα; Life Technologies) was resuspended in sterile, distilled water and used at a final concentration of 20 ng/ml. Cells used in transcriptional studies were treated for 2 h, while cells used for localization and promoter binding studies were treated for 1 h.

Half-life measurement and modeling

25K HEK293 cells expressing Dendra2 or Dendra2-MYL9 fusion proteins were plated in triplicate in blackwalled, clear-bottom 96-well tissue culture plates and given 24 h to adhere. Experiments were then carried out in a Cytation 5 cell imaging multi-mode reader maintained at 5% CO₂ and 37°C. Cells were imaged prephotoconversion at 20× in the GFP and RFP channels, followed by photoconversion through exposure to the DAPI laser line (UV) at maximum intensity for 16 sec. Following photoconversion, cells were imaged in the RFP channel every 4 h for 48 h, imaging settings (LED intensity, integration time, and gain) were maintained throughout experiments. Sum RFP fluorescence intensity values were calculated from post-photoconversion images at each time point, and then were plotted relative to the 0 h intensity. Three to four experiments were performed for each group, and MYL9-Dendra decay curves were then generated in GraphPad Prism using a one-phase exponential decay model. Initial y values were set to 1 for all curves and curve plateaus (y-value at infinite time) were set to 0. Decay curves for each MYL9-Dendra variant were compared with WT in a pairwise fashion to determine if the data sets could be adequately fit by a single curve or if independent curves were required. An extra sum-of-squares F test was used to compare the goodness of fit of shared and independent curves, and subsequently determine the probability with which the shared curve represents both data sets. MYL9 half-life was calculated as ln(2)/k, with k being the rate constant (h⁻¹) generated by the exponential decay model.

For cycloheximide experiments, 1 × 10⁶ HCT116 cells or HCT116 cells transduced with lentivirus expressing WT-MYL9-FLAG, SPK-MYL9-FLAG, or SSQ-MYL9-FLAG were plated in 6-well tissue culture plates 24 h prior to cycloheximide treatment. 100 μg/ml cycloheximide (Millipore Sigma) was added, and cells were harvested in 100 μl cell lysis buffer (50 mM Tris–HCl pH 7.5, 300 mM NaCl, 5 mM MgCl₂, 1% NP-40, 7 mM BME, 1 mM PMSF, 25 μg aprotinin, and 25 μg leupeptin) at indicated time points. Protein was quantified with Pierce 660 nM Protein Assay (Thermo Scientific) and 20 μg total protein was loaded per lane. Samples were analyzed by Western blots as described below.

Cell spreading

Two-well chamber slides (Thermo Scientific) were coated with 3 μg/cm² fibronectin (Corning, Corning, NY) in Tris-buffered saline for 1 h at RT. The remaining solution was removed and slides were blocked with 5 mg/ml bovine serum albumin (BSA; Research Products International, Mt. Prospect, IL) in phosphate-buffered saline. Control and NIH 3T3 cells expressing WT, SPK, or SSQ MYL9-FLAG were grown to confluence, trypsinized, counted, and resuspended at 100K cells/ml in fresh media. Cells were then placed on ice for 20 min to promote rounding. 70K cells were then plated per chamber, and chamber slides were immediately placed in a Cytation 5 cell imaging multi-mode reader with CO₂ maintained at 5% in the sample chamber. The sample chamber was raised from RT to 37°C, with time 0 being defined as when 37°C was reached. Cells were then imaged with a 20× phase contrast objective every 15 min for 1 h. The surface area was quantified for the four largest cells per image at each time point by tracing cell outlines using the ImageJ 1.47v (NIH, Bethesda, MD) plug-in NeuronJ version 1.4.3 [48]. Twelve cell areas taken from three independent experiments were used to generate the mean cell surface area for each sample at each time point.

Transwell migration

Cell migration was measured using a 96-well basement membrane extract (BME) cell invasion assay (Trevigen, Gaithersburg, MD) as previously described [40]. Briefly, control and WT, SPK, or SSQ MYL9-FLAG expressing HCT116 cells were serum-starved for 24 h in McCoy’s 5A Modified Medium with 0.1% BSA. 50K cells in serum-free media were then plated on 0.25× BME in triplicate in the upper chamber of a transwell plate. Media supplemented with 1% FBS was plated in the lower chamber of transwells to serve as a chemoattractant. Migration was allowed to proceed for 24 h, nonmigrating cells were washed away, and migrating cells were stained using Calcein-AM. Fluorescence at 485 nm excitation, 520 nm emission was used to quantify the degree of cell migration.

Western blots

Twenty microgram of total protein per sample was separated on 10% SDS–PAGE gels and transferred to nitrocellulose membranes using a Trans-Blot SD Semi-Dry Transfer Cell (Bio-Rad, Hercules, CA). Nitrocellulose membranes were blocked for one hour in 5% w/v nonfat dry milk in TBS + 0.1% Tween 20 (TBST) at RT. Primary and secondary antibodies were also diluted in the 5% dry milk solution. Dilutions used for primary antibodies were: rabbit anti-FLAG (1:500; Invitrogen, Carlsbad, CA), HRP-conjugated mouse anti-FLAG (1:1000; Sigma–Aldrich), rabbit anti-β-actin (1:1000; Cell Signaling Technologies, Danvers, MA), rabbit anti-pS19 MYL9 (1:1000; Cell Signaling Technologies), rabbit anti-Nα-trimethyl RCC1 (1:10,000) [20], rabbit anti-PSMD7 (1:1000; Origene, Rockville, MD), rabbit anti-cofilin (1:1000; abcam, Cambridge, MA), and rabbit anti-HEXIM2 (1:1000; abcam). Membranes were rocked in primary antibody overnight at 4°C and washed three times for 15 min in TBST. Membranes were rocked in secondary antibody for 1 h at RT and washed three times for 15 min in TBST. The secondary antibody used was donkey anti-rabbit (1:5000; Jackson ImmunoResearch, West Grove, PA). Blots were developed using KPL LumiGLO Peroxidase Chemiluminescent Substrate (VWR, Radnor, PA) or SuperSignal West Femto Maximum Sensitivity Substrate (Thermo Scientific) and a ChemiDoc imaging system (Bio-Rad). Densitometry was performed using ImageJ 1.47v software (NIH).

Real time PCR analysis

Cells were lysed in TRIzol (Life Technologies) and RNA was then extracted in chloroform, pelleted in isopropanol, and washed with ethanol. The final pellet was air dried and resuspended in 100 μl sterile water. RNA concentration was determined with the Cytation 5 plate reader (BioTek). One microgram of RNA was taken through the SuperScript First-Strand Synthesis System (Life Technologies) to generate cDNA according to manufacturer guidelines. Two microgram of cDNA from each sample was used as template with SYBR Green PCR Master Mix (Bio-Rad) in a CFX96 Touch Real-Time PCR Detection System (Bio-Rad). Transcript expression levels were compared using the ΔΔCT quantification method, with GAPDH serving as a control. Melt curves were performed to confirm the presence of single reaction products. Primers (Integrated DNA Technologies, Coralville, IA) used were:

ICAM1 forward 5′-GTAGCAGCCGCAGTCATAAT-3′

ICAM1 reverse 5′-GGGCCTGTTGTAGTCTGTATTT-3′ GAPDH forward 5′-ACAGCCTCAAGATCATCAGCAA-3′ GAPDH reverse 5′-CCATCACGCCACAGTTTCC-3′.

Immunofluorescence

Dendra2 or MYL9-Dendra2 expressing HCT116 cells were analyzed for MYL9 cellular compartmentalization either before or 1 h post-TNFα treatment. At RT, cells were fixed with 4% paraformaldehyde for 10 min, quenched in 50 mM ammonium chloride for 10 min, permeabilized with 0.3% Triton X-100 for 5 min, and blocked with 3% BSA for 1 h. Mouse anti-Dendra2 primary antibody (Origene) was diluted 1:150 in 1% BSA and samples were incubated overnight at 4°C. Samples were washed three times in 1 ml of PBS. Alexa Fluor 488 conjugated goat anti-mouse secondary antibody (Thermo Scientific) was diluted 1:2000 in 1% BSA and added to the samples for 1 h at RT. Cells were washed three times in 1 ml of PBS and counterstained with Hoechst dye (AnaSpec) diluted 1:50,000 in PBS. Cells were washed three times in 1 ml of PBS and slides were mounted with Fluoromount-G (Southern Biotech, Birmingham, AL). Imaging was performed on a Cytation 5 cell imaging multi-mode reader with a 40× objective and image analysis was performed using Gen5 software (BioTek). Total Dendra2 signal was quantified by the sum green fluorescence intensity of images. To quantify nuclear Dendra2, the sum green fluorescence intensity overlapping with Hoechst signal was quantified.

Chromatin immunoprecipitation

Control and WT, SPK, or SSQ MYL9-FLAG expressing HCT116 cells were treated with TNFα for 1 h. 1 × 10⁷ cells were cross-linked with 1% formaldehyde for 10 min at RT, quenched with 1× glycine for 5 min at RT, resuspended in 600 μl ChIP lysis buffer [150 mM NaCl, 25 mM Tris–HCl, pH 7.5, 5 mM EDTA, 1% Triton X-100, 0.1% SDS, 0.5% sodium deoxycholate (NaDoc)], and sonicated 6 × 10 s at 30% amplitude to fragment DNA. 10% of total sheared lysate was saved as an input sample. MYL9-FLAG was then immunoprecipitated from lysates overnight at 4°C using 20 μl M2 agarose anti-FLAG beads (Millipore Sigma) that were pre-blocked in herring sperm DNA (Thermo Scientific). After immunoprecipitation, beads were washed one time in High salt ChIP wash buffer (500 mM NaCl, 20 mM Tris–HCl, pH 8.0, 0.1% SDS, 1% NP-40, 2 mM EDTA), one time LiCl ChIP wash buffer (0.25 M LiCl, 10 mM Tris–HCl, pH 8.0, 1% NaDoc, 1% NP-40, 1 mM EDTA) and two times Tris–EDTA (TE). Bound material was eluted 2× in 100 μl elution buffer (1% SDS, 0.1 M NaHCO₃) for 15 min at RT. De-cross-linking was performed by adding 8 μl 5M NaCl to each sample and incubating overnight at 65°C. Eluates were then treated with 1 μl of RNase A (24 mg/ml stock, Thermo Scientific) and 1 μl of proteinase K (10 mg/ml stock, Millipore Sigma) and incubated at 45°C for 1 h. The remaining immunoprecipitated DNA was isolated using a Qiagen PCR purification kit (Qiagen) and resuspended in 30 μl of TE buffer. Input samples were processed alongside ChIP samples after elution. Equal amounts of purified DNA were amplified by PCR and analyzed by running on 2% agarose gels. PLK-1 served as a control for nonspecific DNA binding. Primers for the PLK-1 (Integrated DNA Technologies) and ICAM1 (Invitrogen) promoters were as follows: ICAM1 forward 5′-CGCCCGATTGCTTTAGCTTG-3′, ICAM1 reverse 5′-GGCTGAGGTTGCAACTCTGA-3′, PLK-1 forward 5′-GGTTTGGTTTCCCAGGCTAT-3′, and PLK-1 reverse 5′-GCTGGGAACGTTACAAAAGC-3′.

Statistical analyses

Statistics were performed with GraphPad Prism 7 software. For experiments involving the comparison of only two groups, a two-tailed Student’s t-test was performed. For experiments involving the comparison of three or more groups, one-way ANOVA was performed with Tukey’s test for multiple comparisons between all groups. For experiments involving comparison of three or more groups at multiple time points, two-way random measures ANOVA with Tukey’s multiple comparisons test between all groups was performed. For experiments involving comparison of three or more groups under two conditions, regular two-way ANOVA with Bonferroni’s multiple comparisons test between all groups was performed. All error bars shown represent standard deviation.

Peptide arrays

Experiments were carried out at the MD Anderson Protein Array and Analysis Core under the direction of Dr. Mark Bedford as previously described [49]. Briefly, arrays of recombinant methyl or acetyl reader domains were immobilized on chips using a high-density arrayer. The methyl reader array consisted of 104 unique domain constructs and the acetyl reader array consisted of 30 unique domain constructs. Biotinylated N-terminal peptides of MYL9 were synthesized (Anaspec) and flowed over the reader chips. Unmodified, mono, and tri methylated MYL9 peptides were tested for binding on the methyl reader array. Unmodified and acetylated MYL9 peptides were tested for binding on the acetyl reader array. Peptides that bound reader domains were then detected using streptavidin–AlexaFluor488 and visualized using a Molecular Probes GenePix 4400A MicroArray Scanner.

SILAC LC–MS/MS

The SILAC LC–MS/MS screen was adapted from a previously described protocol to identify readers of trimethyl-lysine [50]. Briefly, HEK293 cells were grown in normal media (light) or media supplemented with ¹³C labeled lysine (heavy) for eight passages. This resulted in 98.9% incorporation of isotopically labeled lysine into proteins of cells grown in heavy media as determined by LC–MS/MS. For pulldowns, synthetic 14 amino acid N-terminal peptides of MYL9 were synthesized in an unmodified, Nα-acetylated, Nα-monomethylated, and Nα-trimethylated state (Anaspec). All peptides were synthesized with a C-terminal biotin moiety, allowing them to be linked to M280 streptavidin Dynabeads. Pulldowns with unmodified MYL9 peptide were performed using protein lysate harvested from cells grown in light media. Separate pulldowns were performed with acetylated, monomethylated, or trimethylated MYL9 peptides using heavy isotope labeled protein lysate. Each pulldown was performed with 20 μg of peptide and 300 μg of protein lysate. After pulldowns, beads were washed and bound protein was eluted in Laemmli buffer. For analysis, elutions from each modified peptide pulldown were mixed with an elution from an unmodified peptide pulldown. This generated three samples that each contained light proteins that interacted with unmodified MYL9 and heavy proteins that interacted with one of the forms of modified MYL9, allowing for a quantitative comparison of interactors based on modification state. Samples were resolved by SDS–PAGE, in-gel trypsin digested, and subjected to LC–MS/MS using a NanoAcquity HPLC system interfaced to a ThermoFisher Q Exactive mass spectrometer. LC–MS/MS and subsequent data analysis were performed by MS Bioworks (Ann Arbor, MI). Only proteins that had a SILAC ratio (peptides identified in both light and heavy states) were included for further analysis, resulting in ~650 proteins per sample pair. Interactions of interest were further selected by setting a cut-off of a three-fold change in the SILAC ratio (0.33 > heavy/light > 3). For validation of SILAC results, pulldowns were performed as above using only normal media. After being separated by SDS–PAGE, proteins were transferred to nitrocellulose and taken through Western blot analysis with anti-PSMD7, anti-HEXIM2, or anti-cofilin antibodies.

N-terminal mutants of MYL9 select for Nα-methylation or Nα-acetylation

MYL9 has previously been immunoprecipitated from cells or tissue in both Nα-methyl and Nα-acetyl modification states [21]. The N-terminal methionine of MYL9 is cleaved, revealing a Ser-Ser-Lys (SSK) amino acid sequence that conforms to the consensus sequences of both NRMT1 and NatA (Figure 1A) [19,21]. To confirm these enzymes catalyze the Nα-PTMs of MYL9, in vitro enzymatic assays were performed with rhNRMT1 and rhNAA10, the catalytic subunit of NatA. A peptide substrate corresponding to amino acids 2–15 of MYL9 was used at varying concentrations to generate a K_M determination for each enzyme. Both NRMT1 and NAA10 were found to catalyze the modification of MYL9. NRMT1 showed moderate activity with a K_M of 28.8 μM (Figure 1B), while NAA10 showed high activity with a K_M of 0.7 μM (Figure 1C).

To test if we could select for exclusive Nα-methylation or Nα-acetylation of MYL9, we designed N-terminal mutants of MYL9 based on the consensus sequence requirements for NRMT1 and NatA. The N-terminal amino acid sequence SPK is known to be a favorable substrate for NRMT1, while a proline in the second position of proteins is well described as blocking NatA activity [19,21,37]. Similarly, NRMT1 activity is restricted by a glutamine in the third position, while this amino acid is permissive for NatA activity [19–21]. Based on these data, we utilized 14 amino acid peptides of MYL9 containing either an S3P (SPK MYL9) or K4Q (SSQ MYL9) mutation as substrates in the in vitro enzymatic assays, to see if they could select for the activity of either NRMT1 or NatA (Figure 1A).

Matching the published consensus sequence requirements, NRMT1 showed increased affinity for SPK MYL9 with a K_M of 0.7 μM (Figure 1D), while NAA10 showed no activity with SPK MYL9 out to 40 μM of substrate (Figure 1E). SSQ MYL9 showed opposing results, with NRMT1 having no activity on this substrate out to 160 μM of substrate (Figure 1F), and NAA10 maintaining activity with a K_M of 5.0 μM (Figure 1G). Taken together, these results demonstrate that the N-terminal consensus sequence of MYL9 is permissive for modification by both NRMT1 and NatA, and that SPK and SSQ mutants of MYL9 select for either Nα-methylation or Nα-acetylation, respectively.

Nα-PTMs do not alter the half-life of MYL9

A major focus of the Nα-PTM field has been on how these modifications affect the stability of proteins. Early studies suggested that Nα-methylation protects proteins from degradation, as this modification was shown to increase the stability of cytochrome c and block degradation by aminopeptidases [38,39]. Nα-acetylation is known to decrease the stability of some proteins by acting as a docking site for E3 ubiquitin ligases, and thus promoting proteolytic degradation [33,51]. However, accumulating evidence suggests that these effects on stability are substrate specific, rather than intrinsic properties of Nα-acetylation and Nα-methylation [18,31].

To see if Nα-PTMs regulate the stability of MYL9, the photoswitchable fluorescent protein Dendra2 was C-terminally tagged to full-length WT, SPK, and SSQ MYL9. Dendra2 in its basal state fluoresces in the GFP channel, but after photoconversion fluoresces in the RFP channel [52,53]. This allows for tracking of half-life by following the rate of red fluorescence decay after photoconversion, and alleviates the need to treat with cytotoxic agents such as cycloheximide [54]. When Dendra2 was expressed alone and photoconverted, the sum red fluorescence intensity was found to be stable over 48 h, verifying the Dendra2 signal does not dissipate over this time frame (Figure 2A,B). When WT and mutant MYL9 were tagged with Dendra2, decay in red fluorescence related to the degradation of MYL9-Dendra2 was observable within 24 h of photoconversion for all variants (Figure 2C). This was confirmed by cycloheximide experiments that also showed observable protein turnover for all MYL9-FLAG variants within 24 h (Supplementary Figure S1).

The half-life of WT MYL9-Dendra2 was determined to be 16.4 h (95% confidence interval, 15.49–17.52 h) by fitting fluorescence decay data gathered over 48 h to a model of one-phase exponential decay. To determine if alterations in Nα-modification state could be statistically linked to changes in the rate of MYL9 decay, SPK and SSQ MYL9 fluorescence data were each fit to an individual model of decay as well as a model shared with WT MYL9. These models were compared for best fit, and in each case it was found that a model of decay shared with WT MYL9 was adequate to describe the decay of SPK and SSQ MYL9 (Figure 2D,E). This demonstrates that regulation of MYL9 stability is not a role for Nα-methylation or Nα-acetylation, and helps overturn the perception that the primary role of these modifications is as modifiers of protein half-life.

Nα-acetylation of MYL9 promotes cell spreading

We next turned our attention to how Nα-PTMs regulate the function of MYL9 in the cytoplasm. The primary role of MYL9 in the cytoplasm is regulation of actin cytoskeletal dynamics through modulation of NMII activity [22,24]. We analyzed cell spreading of NIH 3T3 mouse fibroblasts on fibronectin, as fibroblast adhesion to fibronectin is a well-studied process requiring active cytoskeletal rearrangement [55–57]. We hypothesized that Nα-acetylation would facilitate the cytoplasmic role of MYL9 as a cytoskeletal regulator. As such, we expected the Nα-acetylation exclusive SSQ mutant of MYL9 to show the most robust cell spreading.

Control NIH 3T3 cells or cells transduced with WT, SPK, or SSQ MYL9-FLAG were plated on 3 μg/ml fibronectin-coated chamber slides and imaged every 15 min for 1 h. Cells were placed on ice prior to plating to promote a uniform distribution of rounded, single cells (Figure 3A, first column). Untransduced, WT MYL9-expressing, and SPK MYL9-expressing cells were just beginning to spread after 1 h and covered very little surface area (Figure 3A,B). However, the Nα-acetylation exclusive SSQ MYL9-expressing cells covered a significantly greater surface area after 1 h (Figure 3B), and only they had readily observable lamellipodia and filopodia that stretched over the fibronectin-coated surface (Figure 3A, black arrows). In order to ensure that expression level differences were not driving our results, we confirmed equal expression of WT, SPK, and SSQ MYL9-FLAG by Western blotting (Figure 3C). As such, it appears that exclusive Nα-acetylation favors the participation of MYL9 in cytoskeletal rearrangement.

Nα-acetylation promotes cell migration and serine 19 phosphorylation

We next sought to confirm the impact of Nα-methylation on the cytoskeletal activity of MYL9 in HCT116 human colon carcinoma cells. These cells have a migratory phenotype, and the nuclear role of MYL9 in ICAM1 transcription has been reported in colon cells, making this cell line suited for the study of both cytoplasmic and nuclear functions of MYL9 [29,58]. To test the role of Nα-modifications of MYL9 in regulating cytoskeletal activity, we measured the migratory potential of nontransduced control cells and cells transduced with WT, SPK, or SSQ MYL9-FLAG. A Boyden chamber design was employed, in which cells were serumstarved, plated on an extracellular matrix extract, and migration towards serum containing media was quantified after 24 h. MYL9 variants were transduced into cells at equal levels so that protein abundance differences would not affect results (Figure 4A). Only HCT116 cells expressing the Nα-methylation deficient SSQ MYL9 mutant showed a significant increase in migration as compared with nontransduced cells (Figure 4B). This suggests that exclusive Nα-acetylation of MYL9 leads to increased cytoskeletal dynamics, as required for migration. This result is consistent with our findings in NIH 3T3 cells, and supports a model in which Nα-acetylation of MYL9 participates in the activation of NMII.

Multiple signaling pathways converge on phosphorylation of MYL9 at serine 19 (pS19) in order to increase NMII activity [22,24]. Because of the requirement for pS19 MYL9 in NMII related activities, including migration and cell spreading, we next tested if the activated cytoskeletal function of SSQ MYL9 is associated with increased levels of pS19. To accomplish this, we treated HCT116 control and MYL9-FLAG (WT and mutant) expressing cell lines with calcimycin, an MLCK activator, and evaluated pS19 levels. We found that after treatment, the nonmethylatable SSQ MYL9 mutant has increased pS19 as compared with WT and SPK MYL9 (Figure 4C). The ratio of pS19 to total MYL9-FLAG was quantified, and it confirmed that SSQ MYL9 shows enrichment of this activating PTM (Figure 4D). This result provides a biochemical basis for the cytoskeletal phenotypes seen in SSQ MYL9-expressing cells and suggests Nα-acetylation is more permissive for S19 phosphorylation than Nα-methylation.

As blocking of phosphorylation by neighboring methylation is a common phenomenon [59,60], we wanted to determine if in addition to Nα-acetylation being permissive to pS19, Nα-methylation could be inhibitory. We thereby assayed the phosphorylation of WT MYL9 in NRMT1 KO HCT116 cells [45]. These cells are deficient in Nα-methylation (Figure 4E), and if Nα-methylation is inhibitory, they should exhibit higher S19 phosphorylation of WT MYL9 than control cells. After calcimycin treatment, the pS19 to total WT MYL9-FLAG ratio was significantly higher in NRMT1 KO HCT116 cells, as compared with control cells (Figure 4E,F). This confirms that Nα-acetylation is permissive for S19 phosphorylation, a modification that is critical for proper regulation of cytoskeletal dynamics, while Nα-methylation can be inhibitory.

Nα-methylation of MYL9 promotes its nuclear function

We next looked at the role of Nα-PTMs in regulating the nuclear function of MYL9, where it serves as a transcriptional activator of ICAM1 [29]. As NRMT1 is found in the nucleus, and Nα-methylation has been shown to facilitate protein–DNA interactions, we proposed that Nα-methylation of MYL9 would facilitate its transcriptional activation of ICAM1 [20,37,42–44]. To test this, we treated nontransduced as well as WT and mutant MYL9-FLAG transduced HCT116 cells with TNFα in order to induce ICAM1 transcription, and then measured the induction of ICAM1 transcription for each cell line by qPCR. We found that cells expressing the Nα-methylation enriched SPK MYL9 mutant had significantly greater induction of ICAM1 transcription than all other cell lines (Figure 5A).

To confirm that enrichment of Nα-methylation was responsible for the heightened transcriptional activity of SPK MYL9, and not loss of Nα-acetylation, this mutant was transduced into control and NRMT1 KO HCT116 cells and ICAM1 transcription was again assayed by qPCR. Loss of Nα-methylation reduced the transcriptional activity of SPK MYL9, as seen by a decrease in ICAM1 transcription when expressed in NRMT1 KO cells as compared with control cells (Figure 5B). This confirms a requirement for Nα-methylation to promote the optimal transcriptional activity of MYL9 in its specialized nuclear role.

We next sought to clarify the mechanism by which Nα-methylation facilitates the transcriptional activity of MYL9. First, we used immunofluorescence to determine if Nα-methylation of MYL9 results in increased nuclear localization, which would increase the pool of MYL9 available to participate in transcription. HCT116 cells transduced with WT, SPK, or SSQ MYL9-Dendra2 were imaged before and post-TNFα treatment. A Dendra2-only expressing cell line was included as a control for nonspecific changes in localization. Dendra2-only did not show TNFα responsive localization (Figure 5C). All variants of MYL9 showed a significant increase in the nuclear to total ratio of MYL9-Dendra after TNFα treatment (Figure 5C), but there were no significant differences between MYL9 variants with regards to this increase. This experiment indicates that Nα-methylated MYL9 does not increase the pool of nuclear MYL9, and that the mechanism by which this PTM facilitates transcriptional function occurs downstream of nuclear localization.

Previous studies have shown Nα-methylation to be important for the association of chromatin regulating and DNA damage response proteins with DNA [20,42–44]. To see if Nα-methylation similarly facilitates the protein–DNA interaction of MYL9 with the ICAM1 promoter, we performed ChIP in nontransduced and WT, SPK, and SSQ MYL9-FLAG expressing HCT116 cells. Only the Nα-methylation enriched SPK mutant of MYL9 resulted in enriched immunoprecipitation of the ICAM1 promoter over input (Figure 5D). This indicates that Nα-methylation facilitates the protein–DNA interaction of MYL9 with the ICAM1 promoter, and provides a mechanism by which Nα-methylation supports the transcriptional function of MYL9.

Regulation of MYL9 protein–protein interactions by Nα-acetylation and Nα-methylation

As regulating protein–protein interactions is another important general function of both acetylation and methylation, we wanted to next determine if Nα-acetylation and Nα-methylation play a role in the protein–protein interactions of MYL9. It has been recently shown that protein domains that recognize histone methylation can also recognize methylation of nonhistone proteins [9,10]. To address whether any known reader domains recognize and bind the methylated or acetylated alpha-amino group of MYL9, an in vitro binding screen was performed with arrays of immobilized recombinant methyl or acetyl reader domains [49].

The methyl reader array consisted of 104 unique domain constructs (Supplementary Figure S2A) and the acetyl reader array consisted of 30 unique domain constructs (Supplementary Figure S2B). Biotinylated N-terminal peptides of unmodified, mono, and trimethylated MYL9 were tested for binding on the methyl reader array. Unmodified and acetylated MYL9 peptides were tested for binding on the acetyl reader array. No known methyl or acetyl reader domains showed specific affinity for modified MYL9 peptides (Figure 6A,B), indicating Nα-methylated and Nα-acetylated MYL9 are read by unique domains or are not read in a domain-specific way. While this result is specific to MYL9, it seems possible that it may hold true for Nα-PTMs in general, as the N-terminal α-amine presents a considerably different steric environment than is found at the ε-amine of lysine or at the guanidino group of arginine.

In attempt to identify unique protein domains that specifically recognize and bind Nα-methylation and Nα-acetylation, an unbiased stable isotope labeling of amino acids in cell culture (SILAC) liquid chromatography-mass spectrometry (LC–MS/MS) pulldown screen was performed. For pulldowns, 14 amino acid N-terminal peptides of MYL9 were synthesized in unmodified, Nα-acetylated, Nα-monomethylated, and Nα-trimethylated states. Pulldowns with unmodified MYL9 peptide were performed using protein lysate harvested from cells grown in light media. Separate pulldowns were performed with acetylated, monomethylated, or trimethylated MYL9 peptides using heavy isotope labeled protein lysate. Elutions from each modified peptide pulldown were mixed with an elution from an unmodified peptide pulldown to generate three samples that each contained light proteins that interacted with unmodified MYL9 and heavy proteins that interacted with one of the forms of modified MYL9. Proteins in these pools were identified and quantified by LC–MS/MS.

Each sample yielded ~1300 protein identifications, with 1611 total proteins identified across all three-sample pairs. Only proteins that had a SILAC ratio (peptides identified in both light and heavy states) were included for further analysis, resulting in ~650 proteins per sample pair. Of these 650, there were only six candidate interactors that showed increased binding to modified MYL9 (Figure 6C,D). Interestingly, Nα-modification of MYL9 predominantly reduced protein–protein interactions, and this was most striking for Nα-monomethylated and trimethylated MYL9. While Nα-acetylated MYL9 peptide only showed 22 downregulated protein interactions as compared with unmodified MYL9 peptide, mono and trimethylated MYL9 peptides each showed ~300 down-regulated interactions (Figure 6C). Interestingly, many of the genes that showed down-regulated binding to Nα-methyl-MYL9 were involved in cytoskeletal regulation (Figure 6E), which supports a model of Nα-methylation being able to block the cytoplasmic roles of MYL9. We selected up-regulated and down-regulated interactors for further validation by peptide pulldowns followed by Western blot. The increased binding of PSMD7 and HEXIM2 to Nα-acetyl-MYL9 did not reproduce but the decreased binding of Cofilin-1 to Nα-methyl-MYL9 was reproducible (Figure 6F). These results indicate that the small number of enhanced binding interactions seen with modified MYL9 as compared with unmodified MYL9 may have been artifactual and that blocking protein–protein interactions, as opposed to promoting protein–protein interactions, may be an important regulatory feature of MYL9 Nα-methylation.