A novel isoform of myosin 18A (Myo18Aγ) is an essential sarcomeric protein in mouse heart

Whereas myosin 18B (Myo18B) is known to be a critical sarcomeric protein, the function of myosin 18A (Myo18A) is unclear, although it has been implicated in cell motility and Golgi shape. Here, we show that homozygous deletion (homozygous tm1a, tm1b, or tm1d alleles) of Myo18a in mouse is embryonic lethal. Reminiscent of Myo18b, Myo18a was highly expressed in the embryo heart, and cardiac-restricted Myo18a deletion in mice was embryonic lethal. Surprisingly, using Western blot analysis, we were unable to detect the known isoforms of Myo18A, Myo18Aα and Myo18Aβ, in mouse heart using a custom C-terminal antibody. However, alternative anti-Myo18A antibodies detected a larger than expected protein, and RNA-Seq analysis indicated that a novel Myo18A transcript is expressed in mouse ventricular myocytes (and human heart). Cloning and sequencing revealed that this cardiac isoform, denoted Myo18Aγ, lacks the PDZ-containing N terminus of Myo18Aα but includes an alternative N-terminal extension and a long serine-rich C terminus. EGFP-tagged Myo18Aγ expressed in ventricular myocytes localized to the level of A-bands in sarcomeres, and Myo18a knockout embryos at day 10.5 exhibited disorganized sarcomeres with wavy thick filaments. We additionally generated myeloid-restricted Myo18a knockout mice to investigate the role of Myo18A in nonmuscle cells, exemplified by macrophages, which express more Myo18Aβ than Myo18Aα, but no defects in cell shape, motility, or Golgi shape were detected. In summary, we have identified a previously unrecognized sarcomere component, a large novel isoform (denoted Myo18Aγ) of Myo18A. Thus, both members of class XVIII myosins are critical components of cardiac sarcomeres.

with recombinant proteins indicate that Drosophila Myo18 (CG31045) and mouse Myo18A do not exhibit motor activity but, instead, probably act as tether molecules (11,12). In accord, Taft et al. (13) found that the motor domain of human Myo18A lacks intrinsic or actin-activated ATPase activity. In addition to its putative role in shaping the Golgi, Myo18A has been implicated in modulating membrane protrusive activity and cell migration (14), whereas Yang et al. (15) identified Myo18A as a receptor for surfactant protein A.
Interestingly, Mori et al. (4) reported that Myo18A␤ is expressed in immature macrophage-like cells, whereas Myo18A␣ expression first emerges in mature cells, suggesting that Myo18A may have a special role in macrophage function. The authors also found in subsequent work that Myo18A␣ (or the N-terminal extension alone), but not Myo18A␤, localized to actin filaments and the plasma membrane (5). In contrast, Billington et al. (16) recently reported that both Myo18A␣ and Myo18A␤ can coassemble with NMHC-IIA (nonmuscle myosin (heavy chain) IIA) to form mixed bipolar filaments. Here, we used various Myo18a knockout mouse models, including reporter knockout and conditional knockout models, to elucidate the physiological function of Myo18A.
Next, we investigated which isoforms of Myo18A are expressed in the heart. We generated an affinity-purified rabbit antibody against the last 18 amino acids of mouse Myo18A  (Myo18a tm1c ) were crossed with Tnnt2-Cre mice, in which Cre recombinase is expressed in cardiac myocytes. Conditional knockout mice were further crossed with Cre reporter (R26R-EYFP) mice to derive triple mutant (cardiac-restricted Myo18a knockout with Cre reporter) mice. B, bright field and fluorescence images of ventricular myocytes isolated from a Myo18a ϩ/tm1c /Tnnt2-Cre/R26R-EYFP mouse. Green (EYFP) fluorescence is an indicator of Cre recombinase activity. Scale bar, 10 m. C, no homozygous cardiac-restricted Myo18a knockout (Myo18a tm1c/tm1c /Tnnt2-Cre) mice were obtained after crossing Myo18a tm1c/tm1c and Myo18a ϩ/tm1c /Tnnt2-Cre mice, indicating that conditional deletion of Myo18a in cardiac myocytes is lethal. No Myo18a tm1c/tm1c /Tnnt2-Cre embryos were detected at E13.5, and three dead embryos were detected at E14.5. D, photographs of HET and homozygous cardiac-restricted Myo18a knockout embryos at day E14.5. Scale bar, 0.5 mm. E, Western blotting analyses. A custom antibody against the last 18 amino acids of mouse Myo18A (anti-Myo18A-C-term) recognizes the Myo18A isoforms Myo18A␣ and Myo18A␤. Lysates were prepared from HeLa cells transiently expressing EGFP-Myo18A␣ (mouse) or EGFP-Myo18A␤ (mouse). The custom anti-Myo18A-C-term antibody did not detect Myo18A in mouse heart lysate, whereas an antibody against the coiled-coil (CC) domain of Myo18A detected a larger than expected band on the same membrane.

Myo18A is a critical sarcomeric protein
(anti-Myo18A-C-term) and confirmed using Western blotting (Fig. 3E) that it recognizes EGFP-tagged mouse Myo18A␣ and mouse Myo18A␤, overexpressed in HeLa cells (Fig. 3E). Nevertheless, we were unable to detect either Myo18A␣ or Myo18A␤ in lysates from adult mouse heart (Fig. 3E). Similarly, Yang et al. (15) were unable to detect Myo18A in either heart or skeletal muscle using anti-Myo18A-C-term antibodies. However, using the same membrane, we were able to identify a larger than expected band (see red arrows in Fig. 3E) with alternative anti-Myo18A coiled-coil domain (Anti-Myo18A-CC-domain) anti-bodies from Proteintech (Fig. 3E) and Santa Cruz (not shown). We suspected that a large novel isoform of Myo18A with an alternative C terminus is expressed in the heart. To investigate alternative exon expression in the heart, we isolated RNA from mouse ventricular myocytes and performed next-generation sequencing (Fig. 4A). Quantitative visualization of the raw data (splicing events) from a single mouse heart in the form of a Sashimi plot (20), aligned to reference genome sequences, revealed that mouse ventricular myocytes express two large exons (highlighted by red rectangles in Fig. 4A), as well as sev- Myo18A is a critical sarcomeric protein eral additional smaller exons, which are not found in the transcripts for Myo18A␣ or Myo18A␤. We cloned and sequenced the cardiac transcript, after performing 5Ј-and 3Ј-RACE (rapid amplification of cDNA ends) PCR and long-range PCRs, which confirmed that the cardiac Myo18A isoform, denoted Myo18A␥ (GenBank TM accession number MK268687), has novel N and C termini (see Myo18A␥ domain structure shown in Fig. 4B). This explains why our custom antibody directed against the last 18 amino acids of Myo18A␣ (and Myo18A␤), as indicated in Fig. 4B, was unable to detect the cardiac isoform of Myo18A, Myo18A␥. Notably, the unique N-terminal extension (containing KE-rich and PDZ domains) encoded by exon 1 of Myo18a␣ is absent in Myo18A␥. Instead, Myo18A␥ has an alternative N-terminal extension (Fig. S1), containing a polyproline region, and a long alternative C terminus (Fig. S2), relative to Myo18A␣ (Fig. 4B), which is rich in serine residues (Fig.  S2). Notably, the deduced N-terminal protein sequence of Myo18A␥, including the start, nicely matches a mouse protein sequence from the Ensembl database (Fig. 4C). Myo18A␥ has a predicted molecular mass of 267 kDa, considerably larger than Myo18A␣ (230 kDa) and Myo18A␤ (190 kDa).

Human homolog of mouse Myo18A␥
RNA sequence analysis of human atrial tissue revealed additional exons that are not found in reference human MYO18A transcripts (Fig. 5A). Sequence alignment confirmed that these exons code a protein homolog of mouse Myo18A␥ (Fig. 5B). Multiple sequence alignment generated by the program Clustal Omega (21,22), using predicted translated transcripts (denoted by the accession prefix XM_) identified by tblastn (protein to translated nucleotide BLAST (basic local alignment search tool)) as input, indicated that the N and C termini of mouse Myo18A␥ are well-conserved across mammalian species (Figs. S3 and S4), except no significant similarity to the C terminus was found in rabbit (Oryctolagus cuniculus). Comparison of mouse Myo18A␥ with the predicted sequences in nonmammalian species revealed high identity in the last 18 residues of the N terminus ( Fig. S5) and high identity in scattered regions in the C terminus (Fig. S6). No significant similarity to the N and C termini of mouse Myo18A␥ was found in Drosophila melanogaster or Caenorhabditis elegans.

Myo18A␥-EGFP localizes to the A-band of cardiac sarcomeres
To elucidate the subcellular localization of Myo18A␥, we transfected neonatal rat ventricular myocytes (NRVMs) with EGFP-tagged Myo18A␥ (Fig. 6). Myo18A␥-EGFP localized between Z-lines in NRVMs stained with antibodies against ␣-actinin, a Z-disc marker (first column in Fig. 6, A and B). Similarly, in NRVMs cotransfected with the Z-disc marker mScarlet-I-Cypher Ic, Myo18A␥-EGFP localized between Z-discs (middle column in Fig. 6, A and B), indicating an A-band pattern of distribution. The striated pattern of Myo18A␥-EGFP overlaid with the A-band marker Myl2-mScarlet-I (myosin light chain 2 tagged to mScarlet) (third column in Fig. 6, A and B), which gives a duplet fluorescence pattern in the A-bands due to dark H-zones (see inset), indicating that Myo18A␥-EGFP localizes to the level of A-bands. How-ever, Myo18A␥-EGFP poorly colocalized with Myl2-mScarlet-I, suggesting that Myo18A␥ may flank the myofibrils at the level of the A-band (third column in Fig. 6, A and B).

Ultrastructural analysis of Myo18a-deficient E10.5 embryo hearts
Transmission EM of E10.5 embryonic hearts from homozygous Myo18a knockout (Myo18a tm1b/tm1b ) mice revealed, compared with age-matched WT controls, overt disorganization of developing myofibrils, sarcomeric disarray, and disruptions of the A-bands (Fig. 7). WT cardiac embryonic myocytes were rich in developing sarcomeres with well-organized structure, including the presence of Z-discs (Z), I-bands (I), and A-bands (A) (Fig. 7A). The myofilaments recognizable in the A-bands were dense and oriented in parallel (see inset in Fig. 7A). In HET Myo18a knockout (Myo18a ϩ/tm1b ) cardiomyocytes, no differences from WT embryonic myofibrils were observed (Fig. 7, A and  B). In contrast, in HOM Myo18a knockout (Myo18a tm1b/tm1b ) myocytes, embryonic myofibrils were only rarely detected. Instead, thick myosin filaments together with I-Z-I brushes were widespread, and a few short sarcomeric formations with loosely packed A-band formations could be detected (Fig. 7C).

Expression of unconventional myosins in ventricular myocytes
Interestingly, analysis of the transcriptome of mouse ventricular myocytes indicated that class XVIII myosins (Myo18a and Myo18b) are the most highly expressed unconventional myosins (Fig. 8A). Moreover, the expression levels of class XVIII myosins were higher than those for class II (conventional) myosins (Fig. 8B), except for the extraordinarily high expression level of Myh6, which encodes the predominant myosin heavy chain of the thick filaments of cardiac sarcomeres in postnatal mice. Myo18B has been reported to be expressed in skeletal muscle, in addition to cardiac muscle (7). Using lysates from mouse skeletal (gastrocnemius) muscle, we could detect a protein band at the same level as cardiac Myo18A␥ (Fig. 8C).

Generation of myeloid-restricted Myo18a knockout mice and phenotype screen of Myo18a-deficient macrophages
Immune cells have been reported to predominantly express Myo18A␤, which lacks the PDZ-containing N-terminal extension (4,15). However, Myo18A␣ becomes detectable in mature macrophages, suggesting that Myo18A may be important in macrophage function (4). To explore the roles of Myo18A in mouse resident macrophages, triple mutant mice with conditional knockout (cKO) of Myo18a in myeloid cells and concomitant expression of a fluorescent Cre reporter gene were generated ( Fig. 9). First, floxed Myo18a mice were crossed with LysM-Cre mice, in which Cre is expressed under the control of the myeloid-restricted M lysozyme gene (23). Second, the offspring were further crossed with R26R-EYFP mice so that individual conditional knockout macrophages could be identified (Fig. 9A). EYFP-positive cells were purified from triple mutant (Myo18a tm1c/tm1c /LysM-Cre/R26R-EYFP) mice by cell sorting, and Western blot analysis confirmed that Myo18A protein was deleted in EYFP ϩ cells (Fig. 9B). We found that mature (resident peritoneal) macrophages predominantly express
Next, we investigated the impact of Myo18a deficiency on macrophage cell shape, motility, and chemotaxis. Using realtime chemotaxis assays, we observed no significant differences

Myo18A is a critical sarcomeric protein
in the morphological polarization, motility, or chemotactic efficiency of WT versus Myo18a-deficient macrophages in a complement C5a (chemoattractant) gradient (Fig. 9, E and F). These findings were surprising because Myo18A has been implicated in the regulation of cytoskeletal dynamics and cell motility (14).

Golgi morphology in WT and Myo18a conditional knockout macrophages
Myo18A has been reported to link the trans-Golgi to the actin cytoskeleton and provide the tension required for its characteristic morphology (9), a stack of elongated flattened cisternae (24). To image the Golgi apparatus, we used anti-GM130 (cis-Golgi matrix protein) and anti-Giantin (Golgi membrane protein) antibodies as markers of the cis-Golgi and Golgi complex, respectively. The Golgi complex was imaged by superresolution structured illumination microscopy (SR-SIM). To confirm that the system (Zeiss Elyra S.1 SIM microscope) could resolve structures beyond the diffraction limit, we measured point spread functions of subdiffraction-sized (100-nm diameter) fluorescent beads (Fig. 10A). The width of point spread functions at half-maximal intensity, full width at half-maximum, was used as an index of resolution (25). The median lateral resolution was x ϭ 108 nm/y ϭ 116 nm using the 488-nm laser and x ϭ 116 nm/y ϭ 131 nm using the 561-nm laser (Fig.  10A), whereas the axial resolution was 316 nm (488-nm laser) and 359 nm (561-nm laser), respectively (not shown). SR-SIM revealed no differences between WT and Myo18a conditional knockout macrophages in the geometry or location of the cis-Golgi or Golgi complex (Fig. 10, B and C). Furthermore, no localization of Myo18A to the Golgi complex could be observed in living HeLa cells expressing both Myo18A-EGFP and RFP-Golgi, a Golgi probe (Fig. 10D). Taken together, these data suggest that Myo18A is not critical for the development or maintenance of Golgi morphology.

Discussion
We investigated the expression of the unconventional myosin Myo18A and the phenotypes of various Myo18a knockout mouse models, which led to the identification of a novel cardiac isoform of Myo18A (Myo18A␥) as a critical sarcomere component. More specifically, we found that (i) homozygous deletion of Myo18a is embryonic lethal; (ii) Myo18a is highly expressed in the embryo heart; (iii) a novel isoform of Myo18A (Myo18A␥) with unique N and C termini, and larger than previously known isoforms, is expressed in the heart; (iv) EGFPtagged Myo18A␥ localizes to the sarcomere with A-band-like interruptions; (v) cardiac-restricted Myo18a deletion is embryonic lethal; and (vi) sarcomeres are disorganized in Myo18a knockout embryo hearts. In addition, we found that (vii) Myo18A␤ Ͼ Myo18A␣ is expressed in mouse resident macro-

Myo18A is a critical sarcomeric protein
phages, and (viii) macrophages isolated from myeloid-restricted Myo18a knockout mice have normal motility and Golgi morphology. Interestingly, the other class XVIII myosin, Myo18B (as distinct from Myo18A␤), has been previously reported to be highly expressed in mouse embryo heart and to localize to the sarcomeres of striated muscle, and Myo18b deletion is embryonic lethal (7). Thus, our findings i-vi underscore that both class XVIII myosins (Myo18A and Myo18B) are integral components of sarcomeres, and neither member can compensate for the loss of the other. The negative findings in Myo18a-deficient macrophages, as well as the recent work from Bruun et al. (10), suggest that Myo18A may not be essential for cell motility or Golgi shape, as deduced previously from knockdown studies (9,14,26). Interestingly, Cao et al. (26) observed that knockdown of myo18a (Myo18a) in zebrafish led to distorted myofibers, which the authors attributed to loss of Golgi binding to F-actin via the PDZ domain of Myo18A. Alternatively, the distortion of myofibers could, in principle, be due to impaired sarcomere assembly and/or stability. Surprisingly, we found that the cardiac isoform of Myo18A (Myo18A␥) lacks an N-terminal KE-rich region and PDZ domain, a hallmark of Myo18A. Instead, the N terminus of Myo18A␥ harbors a polyproline region (PKPPPPPPPPPAP-PKP), which probably adopts a polyproline II helix (27). This region, which contains both class I ([ϩ]-X-hydrophobic-P-X-X-P) and class II (P-X-hydrophobic-P-X-[ϩ]) consensus motifs (28), may bind to SH3 (Src homology 3) domains. Class I and class II consensus motifs were also found in the N termini of Myo18A␥ homologs identified in humans (Fig. 5) and other species (Fig. S3). The alternative C terminus of Myo18A␥ is 388 amino acids longer than in Myo18A␣ (see Fig. 4 and Fig. S2) and is rich in serines, exemplified by a stretch of 13 consecutive serine residues. Alignments indicated that the N and C termini of mouse Myo18A␥ are highly conserved in human heart (Fig.  5), as well as in inferred protein sequences from other mammalian species (Figs. S3 and S4). Multiple conserved blocks, segments of protein, can be identified in both the N and C termini ( Fig. 5 and Figs. S3 and S4). We speculate that Myo18A␥ may form higher-order structures and serve as a molecular scaffold for sarcomere assembly and maintenance. In support of this notion, the head domain (subfraction-1) of Myo18A␥ is similar to Myo18A␣ and Myo18A␤ and has been shown to lack motor activity (11)(12)(13). Future studies using antibodies specifically recognizing the novel N and C termini of Myo18A␥, combined with superresolution imaging and EM, as well as screening for protein interaction partners, may help to resolve how Myo18A␥ regulates sarcomere assembly and function.
In conclusion, we show using knockout mouse models, gene expression analyses, and imaging that the unconventional myosin Myo18A, which lacks motor activity, is an essential sarcomeric protein. We also show that a novel isoform of Myo18A, Myo18A␥ (267 kDa), rather than the known isoforms Myo18A␣ (230 kDa) and Myo18A␤ (190 kDa), is expressed in the heart, as well as in skeletal muscle. Myo18A␥ localizes to A-bands, and loss-of-function leads to disorganization of sarcomeres. Thus, both class XVIII myosins (Myo18A and Myo18B) have specialist functions in sarcomeres.
All animal work done at the Centre for Phenogenomics was performed in compliance with the Animals for Research Act of Ontario and the Guidelines of the Canadian Council on Animal Care. The Centre for Phenogenomics' Animal Care Committee reviewed and approved all procedures conducted on animals at the Centre for Phenogenomics. All work with animals at the

Myo18A is a critical sarcomeric protein
University of Münster was performed in accordance with the German Animal Welfare Act (Tierschutzgesetz) and approved by the local ethics committee of the University of Münster. All animal experiments performed at the Leiden University Medical Center were approved by the animal experiments committee and conformed to the Guide for the Care and Use of Laboratory Animals of the United States National Institutes of Health.

Genotyping
PCR for genotyping was performed in two phases. First, touchdown PCR was performed using the following thermocycling protocol: 94°C for 7 min and then 10 cycles of 94°C for 30 s, 65 to 55°C (starting at 65°C and subtracting 1°C each cycle) for 30 s, and 72°C for 60 s. This was followed by 30 cycles of 94°C for 30 s, 55°C for 30 s, and 72°C for 60 s. The following

DNA Southern blot analysis
Mouse genomic DNA, isolated from tail biopsies using the water/phenol method, was analyzed by Southern blotting DNA hybridization. Approximately 5 mg of genomic DNA was digested with EcoRI, fractionated on 0.8% agarose gels, and transferred to GeneScreen nylon membranes (NEN DuPont, Boston, MA). The membranes were hybridized with a 32 P-labeled probe containing sequences 5Ј or 3Ј to the targeted homology (HR probes) and washed with (final concentrations) 0.5ϫ SSPE (1ϫ SSPE is 0.18 M NaCl, 10 mM NaH 2 PO 4 , and 1 mM EDTA, pH 7.7) and 0.5% SDS at 65°C.

lacZ staining
Whole-mount lacZ staining and gene expression analysis was performed using embryos as described by Tuck et al. (35) and by Adams and Gale (46), cited therein. Briefly, embryos were dissected in 1ϫ PBS, and pieces of embryo-derived yolk sac were transferred to 96-well plates for genotyping. Embryos were dissected away from extraembryonic tissue and fixed in individual wells of a 12-well plate with 2.5 ml of fixative (1% neutral buffered formalin, 0.2% glutaraldehyde, 2 mM MgCl 2 , 5 mM EDTA, and 0.02% IGEPAL CA-630 in 1ϫ PBS) on a rocking platform for 30 min at room temperature. Embryos were washed three times for 10 min each in 1ϫ PBS, pH 8.5, 0.02% IGEPAL and then stained for 18 h at room temperature or 48 h at 4°C in the dark in 5 mM K 3 Fe(CN) 6 , 5 mM K 4 Fe(CN) 6 , 2 mM MgCl 2 , 0.02% IGEPAL, and 1 mg/ml X-gal in 1ϫ PBS (pH 8.5). After staining, embryos were washed two times for 10 min each in 1ϫ PBS with 0.1% Tween 20 at room temperature in the dark followed by 18 h of post-fixation in 10% neutral buffered formalin. Embryos were then cleared with sequential incubation in 20% glycerol in 1% KOH for 8 h at 37°C, 50% glycerol in 1% KOH for 24 -72 h at 37°C, followed by 70% glycerol in 1ϫ PBS for at least 1 week at room temperature prior to imaging.

Antibodies
Affinity-purified polyclonal (anti-mMyo18A) antibodies suitable for immunofluorescence imaging were produced in New Zealand White rabbits immunized with a synthetic peptide (HSYLSDSDTEAKLTETSA) corresponding to the last 18 amino acids of the C terminus of mouse Myo18A (a peptide antigen kindly recommended by John A. Hammer). An attempt to obtain polyclonal antibodies against a synthetic peptide (LMKKDKDKDGGRKEKK) specific for the N-terminal extension of mouse Myo18A was unsuccessful. Rabbit polyclonal anti-Myo18A antibody (14611-1-AP; Proteintech Europe, Manchester, UK), diluted 1:1000, was used to detect mouse Myo18A protein by Western blotting. Lysates from HeLa cells expressing EGFP-tagged mouse Myo18A␣ or Myo18A␤ were used as positive controls for the antibody. In addition, anti-Myo18A coiled-coil domain (tail domain) antibodies were obtained from Santa Cruz Biotechnology, Inc. (amino acids (human) 1656 -1780; catalogue no. sc-99127). Macrophages were purified by cell sorting using rat anti-mouse F4/80 antibodies conjugated to Alexa Fluor 488 (1:40, MCA497A488, Bio-Rad AbD Serotec). For immunofluorescence imaging of cis-Golgi and the Golgi complex (in fixed mouse macrophages), mouse monoclonal anti-GM130 antibodies (clone 35/GM130; 610822, BD Biosciences) and rabbit polyclonal anti-Giantin antibodies (ab24586; Abcam) were used, respectively.

Isolation of mouse ventricular myocytes
Perfusion medium (200 ml) contained 113 mM NaCl, 4.7 mM KCl, 0.6 mM KH 2 PO 4 , 0.6 mM Na 2 HPO 4 ⅐2H 2 O (dihydrate), 1.2 mM MgSO 4 ⅐7H 2 O, 12 mM NaHCO 3 , 10 mM KHCO 3 , 30 mM taurine, 10 mM 2,3-butanedione monoxime, 2 g/liter glucose, and 10 mM Hepes (pH titrated to 7.47 by the addition of 10 M NaOH). The digestion medium (25 ml) contained collagenase type II (190 units/ml), protease type XIV (0.33 units/ml), and 12.5 M CaCl 2 . The adult mouse heart was perfused with perfusion medium for 5 min at a rate of ϳ2.5 ml/min, followed by ϳ6-min perfusion with digestion medium. Subsequently, the heart was placed in 2.5 ml of digestion medium, and the remnant aorta, as well as the atria, were excised. Two tweezers were used to tear apart the heart, and cells were further dissociated Myo18A is a critical sarcomeric protein by repeated aspiration and dispersion via a 10-ml plastic pipette. Perfusion medium (2.5 ml) supplemented with 10% newborn calf serum (NBCS) was added (1:1) to inhibit collagenase and protease activity. After a 10-min sedimentation time, the supernatant was discarded, and cells were resuspended with 7 ml of perfusion medium supplemented with 5% NBCS. Resuspended cells were passed through a nylon filter, and an additional 3 ml of perfusion medium containing 5% NBCS was added, yielding 10 ml of suspended ventricular myocytes.

RNA-Seq
Total RNA of isolated primary ventricular myocytes and human atrial tissue was isolated using TRIzol (Thermo Fisher Scientific) and the Direct-zol RNA MiniPrep Kit (Zymo Research, Freiburg, Germany). Atrial appendages were received from patients with sinus rhythm undergoing open-heart surgery. Permission for experiments with atrial tissue was obtained from the ethics committee of the Technische Universität Dresden (reference number EK 446122011). The investigation conformed to the principles outlined in the Declaration of Helsinki.
Small RNA libraries were prepared for next-generation sequencing by Illumina sequencing using the following kits/ modules from New England Biolabs: the NEBNext Poly(A) mRNA Magnetic Isolation Module (used to isolate poly(A) ϩ RNA from total RNA), NEBNext Multiplex Oligos, and the NEBNext Ultra RNA Library Prep Kit. Quality control and quantification of the prepared libraries were performed on a 2100 Bioanalyzer (Agilent Technologies, Waldbronn, Germany). Libraries were sequenced on a HiScanSQ sequencing system (Illumina, San Diego, CA) in a single-end mode (50 bp) using a TruSeq SR Cluster version 3 and a TruSeq SBS version 3 kit (both Illumina). Validated sequences were mapped, and the Integrative Genomics Viewer (36) was used to visualize splice junctions of the Myo18a gene in a Sashimi plot.

Plasmids for transfection of rat neonatal cardiomyocytes
Expression plasmids were constructed using standard molecular cloning techniques. Restriction and DNA-modifying enzymes (Fermentas, Thermo Fisher Scientific (Breda, The Netherlands) or New England Biolabs, Bioké (Leiden, The Netherlands)) and oligodeoxyribonucleotides (Sigma-Aldrich) were used, following the manufacturer's instructions. For live sarcomeric Z-line imaging, we constructed an expression plasmid named pLV.HsUBC.mScarlet-IϳMmCypher1c, in which the coding sequence of the murine Cypher 1c protein (37) was preceded by a human ubiquitin C gene promoter (38) and extended at the N terminus with the coding sequence of mScarlet-I (39) (AddGene plasmid number 85044). For live-cell imaging of the sarcomeric A-band, the expression plasmid pLV.HsUBC.RnMyl2ϳmScarlet.WHVoPRE was made. In this construct, the coding sequence of rat regulatory myosin light chain 2 is C-terminally extended with the coding sequence of mScarlet-I. Expression of the resulting fusion protein is under the control of the human ubiquitin C gene promoter.
To determine the subcellular location of Myo18A␥, the expression plasmid pEGFP-N1-Myo18A␥ was generated. In this construct, the coding sequence of mouse Myo18A␥ protein was extended at its C terminus with the coding sequence of enhanced GFP. Expression of the resulting fusion protein is controlled by the human cytomegalovirus immediate-early gene promoter.

Isolation and culture of rat neonatal cardiomyocytes
Ventricular cardiomyocytes were isolated from the hearts of neonatal Wistar rats and cultured as described previously (40). Briefly, 2-day-old neonatal Wistar rats were anesthetized using 5% isoflurane inhalation. After adequate anesthesia was confirmed, the hearts were rapidly excised, atria and ventricles were separated, and the ventricular tissue was finely minced and dissociated by two subsequent treatments of 30 min with collagenase type I (Worthington) and DNase I (Sigma-Aldrich). Cells were pelleted by centrifugation (10 min, room temperature, 160 ϫ g) and resuspended in growth medium. The cell suspension was transferred to Primaria TM culture dishes (Corning Life Sciences, Amsterdam, The Netherlands) and incubated for 75 min at 37°C in a humidified atmosphere of 95% air and 5% CO 2 . The unattached cells, predominantly NRVMs, were collected and passed through a 70-m cell strainer (BD Biosciences). Subsequently, 5 ϫ 10 4 NRVMs were seeded in 500 l of growth medium on fibronectin-coated (Sigma-Aldrich) coverslips in a 24-well culture plate (Corning Life Sciences). To prevent the proliferation of noncardiomyocytes, ϳ20 h after seeding, the cultures were treated for 2 h with growth medium containing mitomycin-C (10 g/ml; Sigma-Aldrich, catalogue no. M4287) (41). After this treatment, the growth medium was replaced by culture medium.

Transfection and imaging of neonatal rat ventricular myocytes
NRVMs were lipofected immediately after seeding. In brief, cells in the wells of a 24-well culture dish were transfected with 50 l of lipoplexes consisting of 1.5 l of Lipofectamine 3000, 1 l of P3000 reagent (both from Thermo Fisher Scientific), and a total amount of 500 ng of plasmid DNA diluted in Opti-MEM I reduced serum medium (Thermo Fisher Scientific, catalogue no. 31985070). For co-transfection experiments, 250 ng of each expression plasmid was used. After 20 h, the transfection medium was replaced with growth medium containing mitomycin-C, which, after 2 h, was replaced with culture medium as described above.

Myo18A is a critical sarcomeric protein
Immunocytology of the NRVMs was performed as described previously (42). In short, 4 days after seeding, the cells were washed with PBS, fixed by incubation for 30 min in buffered 4% formaldehyde at 4°C (Added Pharma, Oss, The Netherlands), washed with PBS, and permeabilized by incubation for 5 min at room temperature with 0.1% Triton X-100 (Sigma-Aldrich) in PBS. Afterward, the cells were washed with PBS and incubated for 2 h at room temperature with PBS containing normal donkey serum (0.1%; Sigma-Aldrich) and the following primary antibodies: mouse anti-␣-actinin (Sigma-Aldrich, catalogue no. A7811, dilution 1:200) and rabbit anti-cardiac troponin T (Abcam (Cambridge, UK), catalogue no. ab45932, dilution 1:200). After three washes with PBS, the coverslips were incubated at room temperature for 45 min with appropriate 200-fold diluted secondary antibodies (i.e. Alexa Fluor 568 -conjugated donkey antimouse IgG (HϩL), Alexa Fluor 568 -conjugated donkey anti-rabbit IgG (HϩL), Alexa Fluor 488 -conjugated donkey anti-mouse IgG (HϩL), and Alexa Fluor 488 -conjugated donkey anti-rabbit 488 IgG (HϩL) (Thermo Fisher Scientific, catalogue nos. A10037, A10042, A21202, and A21206, respectively). Cells were washed with PBS, the nuclei were counterstained for 10 min with Hoechst 33342 (final concentration 10 g/l; Thermo Fisher Scientific), and cells were washed again with PBS. Cells were imaged via a ϫ63/1.4 oil objective lens using a Leica TCS SP5 confocal laser scanning microscope and analyzed using Leica Application Suite X software (Leica Microsystems, Rijswijk, The Netherlands).

Transmission EM
Ultrastructural analysis was performed on mouse embryonic heart samples (E10.5). Briefly, full-size embryos were fixed in PBS supplemented with 4% paraformaldehyde, 1% glutaraldehyde, and 15% picric acid, pH 7.4, at 4°C overnight. Embryos were rinsed twice in PBS and treated with 0.5% OsO 4 for 45 min following several washing steps in 100 mM phosphate buffer. Samples were counterstained with uranyl acetate, dehydrated via ethanol series, and embedded in Durcupan ACM epoxy resin (Sigma-Aldrich). Ultrathin sections of the embryonic heart were prepared from resin blocks using a Leica Ultracut S ultramicrotome (Mannheim, Germany) and adsorbed to glowdischarged formvar-carbon-coated copper single-slot grids. Electron micrographs were recorded using a Zeiss LEO 906 electron microscope; images were taken with a TRS Sharpeye CCD camera (Troendle, Moorenweis, Germany).

Isolation of resident peritoneal macrophages
Mice were killed by an overdose of isoflurane in air, and the peritoneal cavity was lavaged via a 24-gauge plastic catheter (B. Braun, Melsungen, Germany) using 2 ϫ 4.5-ml ice-cold Hank's balanced salt solution without Ca 2ϩ or Mg 2ϩ (PAA, Pashing, Austria). After centrifugation (300 ϫ g for 6.5 min), cells were resuspended in bicarbonate-free RPMI 1640 medium containing 20 mM Hepes (Biochrom AG, Berlin, Germany) and supplemented with 10% heat-inactivated fetal calf serum (FCS), 100 units/ml penicillin, and 100 g/ml streptomycin (pH 7.4). The cells were seeded into fibronectin-coated -slide I chambers or -slide chemotaxis chambers (Ibidi, Martinsried, Germany) and placed in a humidified incubator (37°C). After 2 h, -slide I chambers were filled with 1 ml of RPMI 1640 medium, including 10% FCS and antibiotics, and incubated overnight at 37°C with 5% CO 2 . Experiments were performed on the stage of an inverted microscope (LSM 510, Zeiss) equipped with a temperature-controlled incubator (incubator XL S, Zeiss) using the same bicarbonate-free RPMI 1640 medium as above.

2D chemotaxis assays
2D chemotaxis assays have been described previously (43,44). In brief, cells obtained by peritoneal lavage of a single mouse were resuspended in 125-175 l of medium, and 10 l of the suspension was seeded into the narrow (1000 ϫ 2000 ϫ 70 m) channel of an uncoated (IbiTreat) -slide chemotaxis chamber (Ibidi). The narrow channel (observation area) connects two 45-l reservoirs. After 3 h, the chemotaxis chamber was filled with bicarbonate-free RPMI 1640 medium containing 10% FCS, 100 ng/ml lipopolysaccharide, and antibiotics. Next, 15 l of medium containing chemoattractant (complement C5a) and 0.003% Patent Blue V (blue dye) was drawn into one of the reservoirs. The final concentration of complement C5a was 20 nM. The observation area was imaged by phasecontrast microscopy via a ϫ10/0.3 objective lens. The blue dye served as a visual indicator of gradient formation, and we have previously confirmed that it does not affect cell migration. Images were captured every 2 min for 14 h, and cell migration tracks between 6 and 12 h were analyzed with ImageJ (National Institutes of Health) using a manual tracking plugin and the chemotaxis and migration tool from Ibidi. Twenty-five randomly selected cells were manually tracked in each chemotaxis experiment.

Flow cytometry and cell sorting
Cell sorting of EYFP ϩ cells or macrophages stained with Alexa Fluor 488 -conjugated anti-F4/80 antibodies and cell analyses were performed using a BD FACSAria II (or FACSAria III) cell sorter (BD Biosciences).

Superresolution structured illumination microscopy
Superresolution structured illumination microscopy was performed using an Elyra S.1 inverted microscope system (Carl Zeiss MicroImaging), controlled by ZEN 2011 SP2 software (black edition; Zeiss). Cells were imaged via a Plan Apo ϫ63/1.4 oil objective lens, and images were captured with an Andor iXon EM-CCD. The following lasers and filters (in parentheses) were used: 405 nm (BP 420 -480 ϩ LP 750), 488 nm (BP 495-550 nm ϩ LP 750 nm), 561 nm (BP 570 -620 ϩ LP 750 nm), and 642 nm (LP 655 nm). Five grating positions and five phase shifts were used for each z-slice. Point spread functions were measured using TetraSpeck 100-nm fluorescent beads mounted on a slide (T14792; Thermo Fisher Scientific) and the protocol described by Cole et al. (25). Data were analyzed using the MetroloJ plug-in in the open-source platform Fiji (45).

Statistics
Normality and homoscedasticity were tested using the Shapiro-Wilk and Levene tests, respectively. A one-way analysis of variance was used to test for statistical differences at the 0.05 level of significance. When the assumed conditions of normality and homogeneity of variance were not fulfilled, as in most cases, we used the nonparametric Mann-Whitney U test or Kruskal-Wallis one-way analysis of variance on ranks (at the 0.05 level of significance). Post hoc multiple comparisons were made using Dunn's method. Statistical analyses were performed using SigmaPlot (version 12) software (Systat Software, Erkrath, Germany) or Origin 2015 SR2 (OriginLab), and data are presented as box plots or mean Ϯ S.E.