Tissue-specific expression of myosin phosphatase subunits and isoforms in smooth muscle of mice and humans

Abstract

Alternative splicing of exon24 (E24) of myosin phosphatase targeting subunit 1 (Mypt1) by setting sensitivity to nitric oxide (NO)/cGMP-mediated relaxation is a key determinant of smooth muscle function. Here we defined expression of myosin phosphatase (MP) subunits and isoforms by creation of new genetic mouse models, assay of human and mouse tissues, and query of public databases. A Mypt1-LacZ reporter mouse revealed that Mypt1 transcription is turned on early in development during smooth muscle differentiation. Mypt1 is not as tightly restricted in its expression as smooth muscle myosin heavy chain (Myh11) and its E6 splice variant. Mypt1 is enriched in mature smooth versus nonmuscle cells. The E24 splice variant and leucine zipper minus protein isoform that it encodes is enriched in phasic versus tonic smooth muscle. In the vascular system, E24 splicing increases as vessel size decreases. In the gastrointestinal system, E24 splicing is most predominant in smooth muscle of the small intestine. Tissue-specific expression of MP subunits and Mypt1 E24 splicing is conserved in humans, whereas a splice variant of the inhibitory subunit (CPI-17) is unique to humans. A Mypt1 E24 mini-gene splicing reporter mouse generated to define patterns of E24 splicing in smooth muscle cells (SMCs) dispersed throughout the organ systems was unsuccessful. In summary, expression of Mypt1 and splicing of E24 is part of the program of smooth muscle differentiation, is further enhanced in phasic smooth muscle, and is conserved in humans. Its low-level expression in nonmuscle cells may confound its measurement in tissue samples.

INTRODUCTION

Myosin phosphatase (MP) is a complex multisubunit enzyme mediating smooth muscle relaxation and a key target of signaling pathways that control smooth muscle tone (reviewed in Refs. 1–3). Tissue-specific expression of MP regulatory subunits imparts properties onto MP that are specific to smooth muscle contractile subtypes and distinct from MP in nonmuscle cells, in which it mediates shape changes and motility (4, 5). A great deal of phenotypic diversity in SMCs is generated by alternative exon usage (6–8), a process that is nearly universal in mRNA precursors (9). For only a few of the thousands of regulated alternative splicing events in SMCs is the functional significance understood. Isoforms of the MP regulatory subunit Mypt1 (PPP1R12a) are generated by alternative splicing of exons in the 3′ (E24) and central (E12; 13–14) portions of the transcript. The skipping of E24 of Mypt1 codes for a COOH-terminal leucine zipper (LZ) motif. Inclusion of the 31 nucleotide (nt) E24 shifts the reading frame and thus codes for a distinct COOH-terminal sequence with a premature termination codon (designated LZ−). The Mypt1 COOH-terminal LZ is required for its dimerization with PKG1α and cGMP-mediated activation of the enzyme (10–15). Thus, the alternative splicing of Mypt1 E24 functions as a toggle to tune SMC sensitivity to nitric oxide and other signals that relax smooth muscle through the second messenger cGMP and PKG1α (16, 17). The functional significance of the alternative splicing of the central alternative exons is unknown and less well-evolutionarily conserved. These alternative exons lay just upstream of serine/threonine phosphorylation sites, so it is possible that they influence the response to upstream kinases.

We and others have demonstrated that the splicing of Mypt1 E24 is tissue-specific: Mypt1 E24 is predominantly included in phasic smooth muscle such as bladder with more exon skipping in tonic smooth muscle such as aorta, and completely skipped in proliferative smooth muscle and nonmuscle cells (reviewed in Refs. 3, 18; see also Ref. 7). The splicing of Mypt1 E24 is also developmentally regulated and modulates in disease. Nonetheless the expression of Mypt1 and its E24/LZ^+/− isoforms within the context of SMC diversity and the larger SMC gene program remains incompletely characterized in laboratory animals and has received limited study in humans (19–21). A more complete understanding of the expression of MP subunit isoforms is necessary to postulate the extent to which these isoforms may function to tune sensitivity to nitric oxide and other relaxing signals in SMCs dispersed throughout the cardiovascular, gastrointestinal, and other organ systems.

METHODS

The animal protocols used in this study were reviewed and approved by the Institutional Animal Care and Use Committee (IACUC) at the University of Maryland and adhere to the National Institutes of Health guidelines. For tissue harvest, mice were euthanized via CO₂ inhalation followed by cervical dislocation. Tissues were quickly dissected and either placed into RNAlater (Invitrogen) for later purification of RNA, frozen in liquid nitrogen, and stored at −80°C for protein analysis, or placed on ice-cold PBS for imaging of fresh intact tissues.

Mypt1-LacZ Reporter Mouse

The mouse line with β-galactosidase (LacZ) integrated into the Mypt1 locus was generated using embryonic stem (ES) cells obtained from the Wellcome Trust Sanger Institute Gene Trap Resource (SIGTR). In the ES cell line AM0486, the β-geo cassette (a fusion of the β-galactosidase and neomycin resistance genes) with a stop codon had randomly integrated into Intron 9 of the Mypt1 locus. This was confirmed by PCR/RT-PCR with DNA and RNA from cultured ES cells. These cells were injected into blastocysts to generate chimeric mice, which were bred to heterozygosity and maintained in the heterozygous state in a mixed 129/C57B6/J genetic background for at least five generations. The presence of the β-geo cassette within the Mypt1 locus and Mypt1 mRNA and protein was again confirmed by PCR and Western blot, respectively. Male and female mice of different ages as indicated in the figure legends were euthanized, organs harvested, and LacZ activity as a surrogate for Mypt1 expression was detected with a histochemical stain in whole mount and section as previously described (22). Images in these and subsequent experiments were captured with Leica FLIII and DMB stereo and compound microscopes and a SpotRT digital camera. These heterozygous mice had normal appearance and survival within the first year of life. It has previously been established that homozygous loss of Mypt1 causes nonviability in early development (23).

Mypt1 E24 cKO Mice

Mypt1 E24 cKO mice were previously described (24). In brief, treatment of 3- to 6-wk-old male smmhcCre+/E24 flox mice with tamoxifen (50 mg/kg ip every day for 3 days) caused deletion of E24 with high efficiency. These mice were maintained in a C57B6/j genetic background. Samples from these mice used in previous studies were used for the current study as controls.

Mypt1 E24 GFP-RFP Reporter Mouse

As this reporter mouse was ultimately unsuccessful, its generation is described only in brief. More information is available upon request. An ∼360 nt fragment of the mouse Mypt1 gene containing E24 and the highly conserved flanking intronic sequence was amplified by PCR. This Mypt1 E24 fragment was subcloned into the pFlare9a green fluorescent protein-red fluorescent protein (GFP-RFP) splicing reporter plasmid [a kind gift of P. Stoilov (25)] (see Fig. 4) and then through multiple subcloning into the pROSA-DEST vector (Addgene). The validated Mypt1 E24 GFP-RFP construct was injected into ES cells followed by creation of chimeric mice at the University of California Irvine Transgenic Mouse Facility (Irvine, CA). Mice were selected for germline transmission and further bred to homozygosity for the Mypt1 E24 GFP-RFP reporter in the ROSA locus and crossed with SM22Cre mice (26) to obtain mice that were homozygous for Mypt1 E24 GFP-RFP at the ROSA locus and positive for SM22Cre. The SM22Cre causes recombination of the stop-floxed reporter resulting in reporter protein expression in smooth muscle cells. In this model, skipping of Mypt1 E24 would rejoin the GFP translational start codon sequence (ATG) resulting in expression of GFP protein (see Fig. 3A). Inclusion of Mypt1 E24 would interrupt GFP translational start codon so that RFP protein would be expressed. These mice appeared normal and healthy as adults.

Figure 4.

Mypt1 E24 mini-gene GFP-RFP splicing reporter mouse model. A: schematic diagram of Mypt1 E24 mini-gene (31 nt Mypt1E24+ 130 nt upstream and 230 nt downstream intronic sequence) inserted into the GFP-RFP splicing reporter as described in methods. Skipping of Mypt1 E24 should result in translation of GFP, whereas inclusion of E24 by interrupting the GFP start codon should result in translation of RFP. B: freshly isolated small intestine and attached mesenteric arcade from an 8-wk-old transgenic male mouse imaged by confocal microscopy at low (left) and high magnifications (right). Fluorescence in the transgenic tissue is well above the background (wild type) signal on the green channel but not on the red channel. Images shown are representative of n ≥ 3 mice. C: Mypt1 mini-gene mRNA was specifically amplified in PCR using a 3′ primer specific to the mini-gene combined with a 5′ primer for Mypt1. A representative gel from PCR of the Mypt1 mini-gene is shown. Dashed lines represent cropped gels. %E24+ was < 10% in all samples (n ≥ 3). E24, exon24; GFP, green fluorescent protein; Mypt1, myosin phosphatase target subunit 1; nt, nucleotide; RFP, red fluorescent protein.

Figure 3.

Myosin phosphatase subunit expression in single cells isolated from the mouse brain. The Betsholtz laboratory database (www.betsholtzlab.org) was interrogated for expression of myosin phosphatase subunits. The database is derived from experiments in which single cells from the mouse brain expressing different cell type-specific fluorescent reporters were purified by flow cytometry and RNA sequencing was performed as described previously (30) (n = 2 or 3 mice/cell type). Data for single cells purified from the mouse brain is shown here as computer screen shots with average counts on the y-axis and cell type on the x-axis. A: PPP1R12a, protein phosphatase 1 regulatory subunit 12 a (= Mypt1). B: Myh11, smooth muscle myosin heavy chain. C: Ppp1R14a, protein phosphatase 1 regulatory subunit 14a; PPP1R12c, protein phosphatase 1 regulatory subunit 12c (= p85); PPP1cb, protein phosphatase 1 catalytic subunit b. aSMC, arterial SMC; aaSMC, arteriolar SMC; EC1,2,3, endothelial cell type 1,2,3; FB1,2, fibroblast-like 1,2; MG, microglia; Mypt1, myosin phosphatase target subunit 1; OL, oligodendrocytes; PC, pericytes; vSMC, venous smooth muscle cell.

smMHCCre/mTmG Mice

Mice with cell membrane-targeted dTomato (mT) and EGFP (mG) (mTmG) expressed from the ROSA locus (27) were obtained from Jackson Laboratory (Stock No. 007576). These mice were crossed with mice expressing smMHCCre^ERT2 (28) to generate Cre+/mTmG homozygous mice, hereafter referred to as Cre+/mTmG mice. These mice were in a mixed 129/C57B6/J genetic background. At age 3–6 wk, male mice were treated with Tamoxifen 50 mg/kg ip × 3 days to activate Cre recombinase in SMCs thereby switching them from expression of mT to mG. This Cre transgene is integrated onto the Y-chromosome such that only males are used in these experiments. Three or more weeks after tamoxifen treatment the mice were euthanized, organs removed, and dissections performed in ice-cold PBS under a Leica DMB fluorescent stereomicroscope to separate green appearing smooth muscle tissue from red appearing nonmuscle tissues. RNA and protein were purified from these tissues.

Small Human Arteries from Human Skeletal Muscle Biopsies and Other Human Tissues

Three healthy volunteers (60–62 yr of age) provided muscle samples for this investigation. The research was approved by the Institutional Review Board at the University of Maryland School of Medicine. All subjects provided written informed consent and procedures were conducted in accordance with the Declaration of Helsinki. Human skeletal muscle biopsies were obtained from the right vastus lateralis under local anesthesia, ∼15 cm above the patella, using Bergstrom needles (Stille, Solna, Sweden). A small portion of the muscle was immediately placed in RNAlater. Small arteries were dissected from the muscle and stored in RNAlater at 4°C until further processing. Total RNA purified from human large blood vessels, uterus, and bladder were purchased from Amsbio (Lake Forest, CA).

RNA Isolation and Analysis

Dissected tissues were lysed by homogenization and total RNA purified with RNEasy or Purelink columns (Invitrogen) with on-column DNase treatment as per the manufacturer’s instructions and as previously described (16). RNA yield was quantified by optical absorbance (NanoDrop). RNA was reverse transcribed with Superscript III reverse transcriptase enzyme and oligo dT primers. Mypt1 E24 splice variants were amplified by conventional PCR (Applied Biosystems Verti Thermocycler) in a single PCR using the following primers, respectively: 5′-TGC AGT TGG AAA AGG CTA CC-3′ (Forward) and 5′-TCA AGG CTC CAT TTT CAT CC-3′ (Reverse) and 5′-CAG CAG GCT AGA AAA GG-3′ (Forward) and 5′-ACT CTG ATC AAG GCC CCA TT-3′ (Reverse). The exogenous Mypt1 E24 splice variants were specifically amplified with primers specific for the mini-gene sequence 5′- CTGAGGAGAAGTCTGCCGTT-3′ and 5′- TGAACTTCAGGGTCAGCTTGC-3′. Mypt1 Exon24 splice variant PCR products were separated with 2.5% agarose gel electrophoresis, visualized and quantified with Cyto60 staining and a LI-COR Odyssey digital imager, and data reported as %Mypt1 E24 inclusion. As a negative control, RT enzyme was omitted from the reverse transcription reaction mix (data not shown).

MYPT1 E24 inclusion, as well as smooth muscle myosin heavy chain (SMMHC = MYH11) E6 inclusion, were also measured via quantitative real-time PCR (qPCR) in an Applied Biosystems StepOnePlus thermocycler using predesigned TaqMan probes (Invitrogen) and a Fast Advanced TaqMan Master Mix (Applied Biosystems) as previously described (29). In brief, probe sets specific to the alternative and constitutive exons were used in amplification reactions and threshold cycle (Ct) values were determined. Percent SMMHC E6 inclusion was calculated by 2^−ΔCT using Ct values for the E6+ and constitutive targets, then dividing by 100 to obtain %MYH11 E6 inclusion, as previously described (29). For Mypt1 E24, because of differences in the efficiencies of PCR amplification of the E24+ versus constitutive exons, a slightly different calculation was used and 2^−ΔCT was calculated as above. This ΔCt value was then subtracted from the ΔCt value for bladder. The calculation of 2^−ΔΔCt gives the fold difference versus bladder in which Mypt1 is close to 100% E24+.

Protein Analysis by Western Blot

Tissues were homogenized using the Omni Bead Ruptor Elite in a lysis buffer containing 50 mM Tris·HCl (pH 7.6) and 1% protease inhibitor cocktail in a 2 mL tube with 1.4 mm ceramic beads at 3–4 m/s for two cycles of 20 s. The homogenates were mixed with a 1:10 volume of 10X RIPA buffer (EMD 20–188), clarified by centrifugation at 8,000 g for 10 min, and protein concentration determined (Pierce BCA Protein Assay Kit, Thermo Fisher). Lysates were diluted in sample buffer, heated to 95°C for 5 min, and Western blotting was performed as described previously (16). In brief, lysates were loaded onto a 4%–15% Tris-glycine gel (Mini-Protean TGX; Bio-Rad), separated at 200 V for 35 min, and transferred to a PVDF membrane at 100 V for 1.5 h. Blots were probed with rabbit polyclonal antibodies specific for the MYPT1 LZ− and LZ+ isoforms used at 1:8,000 and 1:3,000, respectively. Blots were incubated in a secondary IRDye antibody (800 CW). The LZ+ antibody recognizes identical or similar LZ motifs present in Mypt family members p85 and M21, which thereby serve as internal controls. All blots were reprobed with a mouse monoclonal antibody raised against rat Mypt1 residues 723–840 (BD Biosciences No. 612165) and detected with an IRDye-conjugated secondary antibody (680RD). Signals were detected by scanning on a LI-COR Odyssey infrared scanner. LZ− signal was divided by total Mypt1 signal in each sample and normalized to the bladder (which has nearly 100% Mypt1 E24+). LZ+ signal for Mypt1 was divided by LZ+ signal for p85 in each sample and reported as the ratio of the two signals.

Confocal Microscopy

Mouse intestines with attached mesenteric arcade isolated from Mypt1 E24 GFP-RFP reporter mice were dissected and pinned onto gelatin-coated dishes in ice-cold PBS. Tissues were imaged on a Zeiss 710 laser scanning confocal microscope equipped with an ×20 1.0 NA dipping objective or an ×10 0.3 NA air objective. Green and red fluorescent proteins were excited at 488 nm and 561 nm, respectively. Signals were collected at 500–567 nm and 580–643 nm, respectively.

Statistical Analysis

Data are presented as means ± SE. Data were analyzed and graphed using either SigmaPlot or GraphPad Prism software. Data were analyzed using a one-way ANOVA and a Bonferroni post hoc test. Student’s t test was used where applicable. Significance was accepted with P < 0.05.

RESULTS AND DISCUSSION

Mypt1 Expression In Situ as Defined by Mypt1-LacZ Activity

LacZ expressed from the Mypt1 locus served as a surrogate indicator of Mypt1 expression in situ. LacZ detected by histochemical stain in whole mount and tissue sections was restricted to the smooth muscle of the blood vessels, lungs, intestines, and bladder (Fig. 1). The smooth muscle-specific activity of Mypt1-LacZ was evident as early as embryonic day 14 (E14; Fig. 2). Immunostaining for α-smooth muscle actin (ASMA) as a marker of smooth muscle differentiation corresponded with the LacZ staining (compare Fig. 2, B and C). Thus, transcription of Mypt1 is turned on relatively early in mouse development as part of the smooth muscle gene program.

Figure 1.

LacZ staining is restricted to smooth muscle tissues of Mypt1-LacZ transgenic (TG) mice. LacZ activity was detected by histochemical staining in whole mount of and cross section of organs from transgenic mice (age: 8–12 wk) in which β-geo cassette is integrated into intron 9 of Mypt1. A: LacZ staining is absent in the heart of a control mouse lacking the β-geo cassette. In the Mypt1-β-geo TG mice, LacZ staining in blue is evident in (B) whole mount and (C) cross section in the great vessels [aorta (AO), pulmonary artery (PA), and coronary arteries (inset)] and absent in the heart muscle; (D) medial layer of the aorta and (E) in whole mount in the small intestine and mesenteric vessels and absent in the surrounding fat tissues; (F) in cross section in the smooth muscle layer of the intestine and (G) in the trachea and bronchi of the lungs and absent in the lung parenchyma; (H) in cross section in the smooth muscle layer of the bladder. Images shown are representative of n ≥ 3 mice. Mypt1, myosin phosphatase target subunit 1.

Figure 2.

Mypt1 is expressed as part of the program of smooth muscle differentiation at embryonic day 14 (E14). Timed pregnant transgenic mice were euthanized at E14 and LacZ activity detected in the fetuses by histochemical staining. LacZ activity is present in the E14 fetus in (A) the small intestine and mesenteric blood vessel (arrow) imaged in whole mount (B) the muscle layers of the intestine and adjacent blood vessel imaged in cross section. C: immunostaining of an adjacent section for the smooth muscle marker α-smooth muscle actin (ASMA) identifies these as smooth muscle tissues. D: the medial layer of the aorta (AO) is LacZ+. Images shown are representative of n ≥ 3 mice. Mypt1, myosin phosphatase target subunit 1.

Mypt1 is also transcribed at lower levels in nonmuscle cells (4). We reported 24-fold higher level of Mypt1 in mRNA isolated from the smooth muscle versus endothelial cell fractions of rat mesenteric arteries (24). This is in approximate agreement with the expression of Mypt1 determined by single-cell RNA sequencing (scRNASeq) of cells isolated from mouse brain (Fig. 3) and lung circulations [data not shown; see www.betsholtzlab.org and (30)]. Mypt1 expression in the nonmuscle cells was well above background and 10- to 20-fold lower than in the isolated vascular SMCs (Fig. 3A). In contrast, Myh11 expression was specific to the vascular SMCs, validating the purity of the cell populations and contrasting the patterns of expression of Mypt1 versus smooth muscle-specific myosin heavy chain.

Interrogation of the Betsholtz laboratory database reveals interesting findings regarding expression of other MP subunits. PPP1R12b (M21) (data not shown) and PPP1R12c (p85) (Fig. 3C) subunits show cell-specific patterns of expression similar to that of Mypt1. The inhibitory subunit PPP1r14a = CPI-17) is also enriched in VSMCs but is unique among the MP subunits in being most highly expressed in brain oligodendrocytes (2–5 times greater than VSM; Fig. 3C), consistent with a prior study (31). Of the other inhibitory subunits (PPP1R14b-d), only PPP1R14c (KEPI) was expressed, and only in brain VSM, at levels similar to that of CPI-17 (data not shown). Of phosphatase catalytic subunits (Ppp1ca-c), only the β subunit is enriched in in brain and lung VSMCs (three- to sixfold vs. other cell types; Fig. 3C), whereas the PPP1ca subunit is expressed at similar moderate levels across all cell types, and the PPP1cc subunit minimally expressed across all cell types (data not shown). The former is consistent with PPP1cb serving as the MP catalytic subunit in smooth muscle as determined by gene inactivation (32).

A Mini-Gene Construct for the Study of Mypt1 E24 Splicing

Isoforms of Mypt1 are generated by alternative splicing of exons 13, 14, and 24 (numbering for mammals). Analyses limited to tissue that can be isolated by gross dissection have demonstrated tissue-specific splicing of Mypt1 E24. We generated a mini-gene splicing reporter construct to report on Mypt1 E24 splicing in smooth muscle dispersed throughout the different organ systems (Fig. 4 and methods). When E24 is skipped, green fluorescent protein (GFP) is expressed; when E24 is included, the GFP translational start codon (ATG) is interrupted and red fluorescent protein (RFP) is expressed. We validated the splicing of the mini-gene construct in in vitro (data not shown) but in vivo with the reporter expressed from the ROSA locus E24 inclusion into the heterologous mRNA occurred in <10% of transcripts resulting exclusively in expression of the GFP reporter (Fig. 4). This contrasts with significant levels of splicing observed with a similar Mypt1 mini-gene construct tested by injection into the chicken gizzard (33). This suggests that in this experiment either 1) insufficient Mypt1 gene sequence is present to activate splicing of E24 or 2) the E24+ intronic fragment is unable to direct splicing in this heterologous context.

Mypt1 E24 Splice Variants in GFP+ SMC-Enriched Tissues

As an alternative approach we used fluorescent reporter mice to identify smooth muscle tissue, which could then be isolated for analysis of gene expression. Treatment of smMHCCre/mTmG mice with Tamoxifen resulted in conversion of SMCs from red to green fluorescence with high efficiency, e.g., ≥90% in the small mesenteric arteries and kidney blood vessels (Fig. 5). Smooth muscle-containing tissues were removed from the mouse and GFP+ smooth muscle tissue was dissected under fluorescent stereomicroscope. Splice variants were measured by either end-point PCR followed by gel separation of products or by qPCR as described in methods. In the vascular system, %Mypt1 E24+ was greater in the small resistance size mesenteric arteries as compared with the large conductance aorta (Fig. 5B) consistent with our previous studies (see Ref. 34). In contrast the third-order intralobar pulmonary artery expresses predominantly the Mypt1 E24− splice variant while the adjacent bronchial smooth muscle expresses higher levels of Mypt1 E24+ similar to that of the aorta. In the genitourinary system, the bladder expresses nearly exclusively the Mypt1 E24+ variant consistent with our prior reports (see Ref. 29). In the gastrointestinal system, the esophagus expresses nearly exclusively the Mypt1 E24− variant. The %E24+ increases modestly in the stomach (fundus) and then reaches 50%–80% throughout the small intestines with the highest level observed in ileal smooth muscle before decreasing to low levels in the smooth muscle of the rectum. PCR of portal vein samples from wild-type and E24 cKO mice validate this assay.

Figure 5.

Tissue-specific splicing of Mypt1 E24 in tissues dissected from smMHCCre/mTmG mice (Cre/mTmG). Male Cre+/mTmG mice aged 3–6 wk were treated with tamoxifen 50 mg/kg ip every day for 3 days and tissues harvested at age 8–12 wk. A: organs were dissected and immediately imaged by confocal microscopy as described in methods. Approximately 90% of the SMCs wrapping around the small mesenteric arteries show membrane localized GFP indicating Cre-mediated recombination at mTmG locus. In a dissected kidney, the glomerular cells (non-SMCs) show membrane localized Tomato (arrow) while an adjacent blood vessel (*) shows membrane localized GFP. Tissues were dissected from Cre/mTmG mice under a fluorescent stereomicroscope to separate the smooth muscle from the nonmuscle portions. RNA was purified from the isolated smooth muscle tissues and splice variants analyzed by conventional and/or real-time PCR. B: Mypt1 was amplified with primers flanking E24 and products separated by gel electrophoresis. A representative gel is shown. Vertical dashed lines indicate lanes from a single gel that were rearranged for sequential presentation of gastrointestinal tissues in an aboral direction. Breaks in the pictures represent separate cropped gels. Data are shown as %Mypt1 E24+; n = 3. *P < 0.05 vs. bladder; in other pairwise comparisons, P > 0.05 for AO vs. BRO; DUO vs. JEJ and PA vs. FUN; for all other pairwise comparisons P ≤ 0.05. C: separate sets of TaqMan probes amplifying Mypt1 E24+ mRNA splice variant and constitutive Mypt1 exons were used in real-time PCR. The expression of Mypt1 E24+ was calculated using 2^−ddCt vs. bladder to obtain fold difference in Mypt1 E24. There is generally good agreement between these two different methods with the values obtained by qPCR tending to be higher. n = 3, *P < 0.05 vs. bladder. In other pairwise comparisons, P > 0.05 for AO, BRO, and PA; DUO, JEJ, ILE, REC, and MA; for all other pairwise comparisons P ≤ 0.05. D: separate sets of TaqMan probes amplifying Myh11 E6+ mRNA and Myh11 constitutive exons were used in real-time PCR. The expression of Myh11 E6+ was calculated by 2^−dCt multiplied by 100 and shown as %MYH11 E6 inclusion. n = 3, *P < 0.05 vs. bladder. In other pairwise comparisons, P > 0.05 for AO vs. PA; MA vs. ESO; FUND vs. PA; DUO, JEJ, ILE, and BRO; ILE vs. REC; for all other pairwise comparisons P ≤ 0.05. All data are shown as means ± SE. AO, aorta; BLA, bladder; BRO, bronchus; DUO, duodenum; ESO, esophagus; E24, exon24; FUN, fundus; GFP, green fluorescent protein; ILE, ileum; JEJ, jejunum; KO, knock-out; MA, mesenteric artery; Mypt1, myosin phosphatase target subunit 1; PA, pulmonary artery; ; PV, portal vein; REC, rectum; SMCs, smooth muscle cells; WT, wild type.

We corroborated these results with a second assay, qPCR, using probe sets that specifically detect the Mypt1 E24+ transcript versus all Mypt1 transcripts (see methods). dCt values were normalized to bladder to obtain %Mypt1 E24+ as compared with bladder (Fig. 5C). There was generally good agreement between the two assays in estimating %Mypt1 E24+ with %Mypt1 E24+ tending to be higher when estimated using this qPCR method. One exception is in the rectum samples, where Mypt1 E24+ is significantly lower than in small intestine samples in conventional PCR but equivalent when measured by qPCR.

We compared expression of Mypt1 E24 splice variants with that of Myh11 E6 splice variants also measured by qPCR. Myh11 E6+ codes for the SM-B isoform of Myh11 while Myh E6− codes for the SM-A isoform of Myh 11. Myh11 has the advantage that its expression is highly specific for SMCs, i.e., not contaminated by nonmuscle components of tissue samples. The expression of Myh11 E6+ splice variant is much more tightly regulated in its expression as compared with Mypt1 E24 (Fig. 5D). Myh11 was nearly exclusively of the E6+ splice variant in the phasic smooth muscle of the bladder and nearly exclusively E6− in the tonic smooth muscle of the aorta and pulmonary artery, consistent with prior studies (35, 36) (reviewed in Refs. 2, 37). In the bronchus smooth muscle, the E6+ splice variant predominated consistently with a prior study (38). In contrast to Mypt1 E24, low level of the MyhE6 variant was present in the MA. Like Mypt1 E24, low levels of Myh E6+ variant were present in esophagus and fundus with greater levels in small intestinal smooth muscle and a lower level in the rectum. Thus, transcription of Mypt1 and splicing of E24 appears to be part of the program of SMC differentiation, but it is more relaxed as compared with Myh6. Mypt1 E24 is present at low but still significant levels in tonic smooth muscle, with its expression further enhanced in phasic smooth muscle. In contrast, Myh11 is also part of the program of SMC differentiation but much more tightly regulated, and its E6 splice variant is more tightly regulated according to tonic versus phasic smooth muscle phenotype. Nonetheless not all tissues fall into this simple dichotomy, as mesenteric arteries that have higher Mypt1 E24 but low Myh E6, and bronchial airway smooth muscle, which has high levels of Myh E6 and intermediate levels of Mypt1 E24, supporting complexity in smooth muscle phenotypes.

Mypt1 COOH-Terminal Leucine Zipper-Positive and Zipper-Negative Isoforms in GFP+ SMC-Enriched Tissues

Western blots with isoform-specific antibodies were performed to determine if the expression of Mypt1 E24+/− splice variants translated into expression of Mypt1 LZ−/+ isoforms that they encoded. These antibodies were previously validated (16, 39). Gels were run in duplicate and blots probed with either the rabbit polyclonal Mypt LZ− or LZ+ antibody followed by a mouse monoclonal Mypt1 antibody detecting all Mypt1 isoforms. This method allowed for on-blot estimates of Mypt1 isoforms. LZ−/Mypt1 ratios in different tissues were normalized to bladder since the PCR assays indicated that bladder was nearly 100%E24− coding for Mypt1 LZ− isoform. The LZ−:Mypt1 signal in aorta, bronchus, and PA was one-third to one-half of that in the bladder (Fig. 6A; shown are average values from n = 3 mice). In contrast, the LZ−:Mypt1 signal in MA and ileum was 70%–80% of that of the bladder. These ratios correspond well with the %Mypt1 E24+ measured by PCR. The rabbit polyclonal antibody raised against the Mypt1 COOH-terminal LZ sequence also recognizes identical or nearly identical LZ sequences present in Mypt family members p85 (85kD) and M21 (21kD) (Fig. 6B). So here we used the LZ+ Mypt1:p85 ratio without any additional normalization as another indicator of Mypt1 isoform expression in the different tissues. The LZ+ Mypt1-to-p85 ratio was high in aorta, bronchus, and pulmonary artery (∼3) and severalfold lower in the bladder, mesenteric artery and ileum (0.7–1.1; Fig. 6B, shown are average values n = 3). These values correspond well with the %Mypt1 E24− measured by PCR.

Figure 6.

MYPT1 LZ+/− protein isoform expression in mouse tissues. Tissues from Cre+/mTmG mice as mentioned earlier were homogenized to obtain protein lysates. Twenty micrograms of protein were assayed by Western blot as described in methods. Separate membranes were probed with rabbit polyclonal antibodies specific for Mypt1 (A) LZ− and (B) LZ+ isoforms. The LZ+ antibody also recognizes the identical or nearly identical LZ motif present in Mypt family members p85 and M21. The LZ+/− isoform rabbit polyclonal antibodies were detected with a secondary antibody and signals quantified on the green channel. Then the blots were probed with a mouse monoclonal antibody recognizing all Mypt1 isoforms and an anti-mouse IgG secondary antibody on the red channel. The bands below the Mypt1 band at ∼130 kD on the red channel may represent slight degradation of Mypt1, though they are recognized by neither the LZ− nor LZ+ antibodies. The doublet at ∼130 kDa most evident with the LZ+ antibody represents central splice variants of Mypt1. Shown are the individual scans from a representative blot of samples from a single mouse. Signals were quantified with a LI-COR Odyssey gel scanner. Below each blot are average values for each tissue (n = 3). A: LZ– signals were normalized to total Mypt1 and then to values for bladder since by PCR this tissue is near 100% E24+ (coding for LZ−). B: Mypt1 LZ+ signals are divided by p85 LZ+ signals and shown as mean absolute values for each tissue. The merged image gives a visual representation of LZ+ vs. total Mypt1 expression. Dashed line indicates that blot was cropped for ease of presentation. AO, aorta; BLA, bladder; BRO, bronchus; DUO, duodenum; ESO, esophagus; E24, exon24; FUN, fundus; GFP, green fluorescent protein; ILE, ileum; JEJ, jejunum; KO, knock-out; LZ, leucine zipper; MA, mesenteric artery; MWM, molecular weight marker in kilodaltons; Mypt1, myosin phosphatase target subunit 1; PA, pulmonary artery; ; PV, portal vein; REC, rectum; SMCs, smooth muscle cells; WT, wild type.

Tissue-Specific Splicing of Mypt1 E24 is Conserved in Humans

The E24 sequence is present within the human Mypt1 gene and phylogenetically conserved (8) but has not been annotated as a functional exon (see genome browsers). In the few reports on the expression of human Mypt1 isoforms, the E24+/− splice variant ratios were not directly measured (19, 20). Here we utilized dissected small arteries from muscle biopsies and commercially obtained total RNA samples to measure Mypt1 E24+/− splice isoform ratios ∼60% of Mypt1 mRNA was E24+ in the skeletal muscle small arteries contrasting with the much lower E24+ in a cephalic artery and caval vein (Fig. 7). Mypt1 E24+ was undetectable human uterus while it comprised ∼30% of Mypt1 transcripts in human bladder. We obtained additional human uterine samples in which total RNA was purified from the dissected myometrium (smooth muscle). In these samples, Mypt1 E24 was present at significant levels with a wide range of % inclusion (data not shown) perhaps reflecting an artifact of tissue dissection.

Figure 7.

Mypt1 E24 splice variant expression in human blood vessels and other tissues. Small arteries were dissected from human muscle biopsy samples (n = 3 males age 60–62 yr) and RNA purified as described in methods. RNA from large arteries and veins, uterus, and bladder were purchased from Amsbio. RNA was reverse transcribed and Mypt1 E24+/− splice variants amplified by PCR with primers flanking E24 as described in methods. E24+/− splice variants were separated by 2.5% agarose gel electrophoresis and imaged and quantified with a LI-COR D-DiGit gel imager. Data are presented as %Mypt1 E24+. E24, exon24; Mypt1, myosin phosphatase target subunit 1.

We queried the GtexPortal (www.gtexportal.org) to further examine Mypt1 expression in human tissues. This website under the direction of the Broad Institute contains complete gene expression data from 54 different human tissues with n = 20 (bladder) to n = 584 (tibial artery) individuals with most sample sizes in the range of several hundred. Mypt1 mRNA is ∼10- to 20-fold higher in blood vessels and other smooth muscle containing tissues (gastrointestinal, uterus) at ∼80 to150 transcripts per million reads (TPMs) versus nonmuscle tissues at ∼10 to 20 TPMs, a difference similar in magnitude to that of rodent tissues. At the level of individual exons, E24 is not detected in the many nonmuscle tissues and cells queried (TPM = 0), including brain, spleen, and EBV-transformed lymphocytes and cultured fibroblasts, in which Mypt1 is expressed at low levels. We also observed absence of the Mypt1 E24+ splice variant in human iPS cells differentiated into smooth muscle cells, where expression of the smooth muscle-specific marker Myh11 was also very low and near background levels (data not shown). In the Gtex database, Mypt1 E24 is detected at low levels in the aorta and moderate levels in smaller arteries, bladder, uterus, and throughout the gastrointestinal system (esophagus, small, and large intestines; Table 1). Thus, the undetectable level of Mypt1 E24 in RNA from intact human uterus and relatively low level of E24 in RNA from intact human bladder, as compared with dissected rodent bladders likely reflects contamination by nonmuscle cells. Otherwise, this pattern of Mypt1 E24 splicing in human tissues matches well with what our laboratory has reported in other species including pigs (40), rodents(11), and birds (reviewed in Ref. 18).

Table 1.

Median read counts per base for human Mypt1 exons 23 and 24 from www.gtexportal.org

Tissue	Mypt1 E23	Mypt1 E24	%E24-in
Tibial artery	3.67	1.14	31
Coronary artery	2.08	0.42	21
Aorta	1.86	0.11	6
Bladder	2.58	0.64	25
Uterus	2.07	0.27	13
Esophagus muscularis	3.8	1.65	30
Colon sigmoid	2.53	0.94	37

Percent exon inclusion of Mypt1 E24 was calculated by dividing E24 counts by E23 counts. Data shown are average values from tissues obtained from n = 20 to n = 584 individuals. E23, exon23; E24, exon24; Mypt1, myosin phosphatase target subunit 1.

The expression in humans of the other MP subunits and their splice variants, as derived from the GtexPortal, is similar to that of mice. One notable difference is the apparent high percentage (∼80% to 90%) of the E3 skipped splice variant of the human CPI-17 transcript (of 4 exons), confirming a prior report of this splice variant (41), which appears to be unique to humans.

In conclusion, we have demonstrated the tissue-specific expression of myosin phosphatase subunits and Mypt1 E24 splice variant isoforms in mice and humans. Tissue-specific splicing of Mypt1 E24 is a function of both smooth muscle cell differentiation and the adoption of the fast (phasic) smooth muscle gene program. This contrasts with Myh11, in which transcription is tightly restricted to SMCs and splicing of E6 is tightly restricted to the phasic (fast) sublineage. Age is another factor that must be considered. We have demonstrated induction of Mypt1 E24+ isoform during rodent postnatal maturation (39) while another group has shown a shift to the Mypt1 E24− isoform in femoral arteries of aged mice (42), all associated with the predicted changes in vasodilator responses. The small artery samples in the current study were from aged human males, whereas the larger arteries were from younger individuals. Whether age and sex play a role in expression of Mypt1 E24 isoforms in humans requires further study. The low-level expression of Mypt1 (E24-splice variants) in nonmuscle cells may confound assay of SMC E24-LZ+/− isoform ratios, particularly in tissues in which the smooth muscle component is not easily purified. This limitation may be addressed in future studies by purifying SMC populations through fluorescent tagging and cell sorting followed by RNA sequencing. In our initial attempts to perform this in the current study, we observed loss of the Mypt1 E24+ splice variant, perhaps related to the long processing times and selective degradation of this premature termination containing mRNA by nonsense-mediated decay. An attempt to generate a mini-gene splicing reporter for Mypt1 E24 was not successful. Knocking a reporter into the endogenous Mypt1 locus will likely be more effective. New approaches to defining alternative splicing in the generation of phenotypic diversity have been developed (43) and should be particularly useful when applied to the study of SMCs dispersed throughout the different organ systems.

Perspectives and Significance

Myosin phosphatase mediates smooth muscle relaxation and is a critical end target of signaling pathways that regulate smooth muscle function. Thus regulated expression of myosin phosphatase subunits is a critical determinant of organotypic smooth muscle function. The current study utilized new genetic mouse models, human tissues, and online scRNASeq/RNASeq databases to more extensively characterize the expression of MP subunits. The tissue-specific pattern of expression of the Mypt1 E24+ splice variant coding for the NO/cGMP resistant LZ− isoform is conserved between humans and rodents. Unlike smooth muscle-specific myosin heavy chain (Myh11), there is low but still significant levels of expression of myosin phosphatase subunits, including Mypt1, in nonmuscle cells. Its expression in nonmuscle cells may confound analyses of MP expression assayed in human and laboratory animal tissue lysates.

GRANTS

This research was supported in part by the University of Maryland Scholars Program, an initiative of the University of Maryland: MPowering the State, National Institutes of Health (NIH) Grant R01-HL066171 and HL142971-01A1 and Veterans Affairs (VA) MERIT award I01BX004443 (to S.A.F.); T32-HL072751 and T32-AR007592 (to J.J.R.); and VA Merit Review Award I01-CX000730, NIH K23-AG040775, and the Baltimore Veterans Affairs Medical Center Geriatric Research, Education and Clinical Center (GRECC) (to S.J.P.).

DISCLOSURES

No conflicts of interest, financial or otherwise, are declared by the authors.

AUTHOR CONTRIBUTIONS

J.J.R., Y.L., S.K., and S.A.F. conceived and designed research; K.O., J.J.R., Y.L., S.K., D.K., S.J.P., and S.A.F. performed experiments; K.O., J.J.R., Y.L., S.K., D.K., and S.A.F. analyzed data; K.O., J.J.R., Y.L., S.K., D.K., and S.A.F. interpreted results of experiments; K.O., J.J.R., Y.L., S.K., and S.A.F. prepared figures; S.A.F. drafted manuscript; K.O., J.J.R., Y.L., S.K., D.K., S.J.P., and S.A.F. edited and revised manuscript; K.O., J.J.R., Y.L., S.K., and S.A.F. approved final version of manuscript.