Titin’s P-zone domains A164–167 are essential for thick filament structural arrangement

The striated muscle sarcomere is a fine-tuned lattice-like structure composed of thin filaments, thick filaments, and titin filaments (Wang et al., 2021). The myosin-based thick filament, into which titin incorporates, is tightly regulated structurally and functionally, such that the myosin heads maintain precise spacing, orientation, and intermolecular interactions throughout the filament (Houmeida et al., 1995; Jung et al., 2008). Titin’s A-band contains super-repeating units of immunoglobulin-like (Ig-like or Ig) and fibronectin type III (FnIII) domains connected by short, 3–4 residue linkers (Fleming et al., 2023; Bang et al., 2001; Bucher et al., 2010). At the distal region of the thick filament, titin’s D-zone contains 6 super-repeats of 7 domains each, while the C-zone contains 11 super-repeats of 11 domains each (Labeit and Kolmerer, 1995) (see Fig. 1 A for details). In the C-zone, the length of the titin super-repeats matches that of the myosin helical repeat and the spacing between myosin binding protein-C (MyBP-C) molecules, 43 nm (Tonino et al., 2019; Bennett et al., 2020; Squire et al., 2004; Craig and Offer, 1976b). Titin has long been considered a “molecular ruler” or “template” for the thick filament and has been shown to dictate thick filament length by its C-zone super-repeats (Tonino et al., 2017), but not its I/A junction at the outer edges of the thick filament (Granzier et al., 2014). Whether the D-zone contributes to thick filament length control remains to be tested.

The C-terminal M-band segment of titin, on the other hand, contains Ig-like domains and seven unstructured “insertion sequences” named M-is1 through M-is7 (Bang et al., 2001). This segment of titin is essential for M-band structure and sarcomere formation (Musa et al., 2006; Peng et al., 2007; Radke et al., 2019) and interacts with other structural proteins like myomesin, obscurin, and obscurin-like 1, as well as signaling proteins such as muscle creatine kinase, four and a half LIM domains 2, and myospryn (Sarparanta et al., 2010; Hornemann et al., 2003; Weinert et al., 2006; Ackermann et al., 2009). Titin’s kinase domain (titin kinase) is implicated in titin mechanosensing and interacts with p62 and Nbr1 when ubiquitinated (Bogomolovas et al., 2014; Bogomolovas et al., 2021; Lange et al., 2005; Stahl et al., 2011; Puchner et al., 2008). The currently accepted model in the field for titin in the M-band indicates that titin molecules extend into the M-band to overlap with antiparallel titins by ∼120 nm (Obermann et al., 1996).

Linking titin’s C-zone and M-band segments is the proximal zone (P-zone). Titin’s P-zone segment, spanning only seven domains (A164–A170), is often overlooked or simply considered to be part of the M-band segment; however, its unique structural features and important interactions warrant examination. The domain pattern of this segment—Ig-Ig-FnIII-FnIII-Ig-Ig-FnIII—is unique from that of the D- and C-zones and lacks the M-is found in the M-band. While the A-band of titin contains many sets of 2–3 consecutive FnIII domains, there are no other pairs of Ig-like domains, except in the P-zone, which contains two tandem Ig-like pairs (Bang et al., 2001; Labeit et al., 1992). Furthermore, the linkers between the Ig domains in the A164–A165 and A168–A169 tandems are beta (β) strands, which are not present elsewhere in the tandem Ig segments of titin and are indicative of a need for rigidity, although this is untested (Steward et al., 2012). The C-terminal three domains of the P-zone, A168–A170, comprise a binding site for E3 ubiquitin ligases, including the MuRFs, and are thought to be important for protein degradation signaling (Mrosek et al., 2007; Müller et al., 2007; Lange et al., 2005; Bogomolovas et al., 2021).

Two groups recently revealed the structure of the native cardiac thick filament C-zone using cryo-electron tomography (Tamborrini et al., 2023) and cryo-electron microscopy (Dutta et al., 2023). Of particular interest was the finding that in the C-zone to P-zone transition region, titin adopts two conformations, termed titin alpha (α) and titin beta (β) (Tamborrini et al., 2023). Titin α extends to the M-band as was previously established. At domain A158, titin β bends to form a triangular loop (referred to as a “bowtie” in Tamborrini et al. [2023]) of 14 domains, through domain A167, including the first 4 P-zone domains (A164–167). Domains A168 through M1 were found to lie linearly along the filament alongside the P-zone myosin heads, but the titin domains beyond M1 were not detected in titin β. Since the β bowtie loop offsets domains A158-M1 from their normal locations in titin α, these domains exist in two locations in each half-thick filament in the model provided by Tamborrini et al., while domains beyond M1 exist in one location in the M-band of each half-thick filament (see Fig. 1 A for a schematic of this structure). This novel finding challenges the current paradigm for titin’s C-terminal structure and raises the question of what might be the functional purpose of the titin β loop and the domains that comprise it, particularly the P-zone domains.

It is unknown whether the first four P-zone domains, A164–A167, have binding partners outside of their interactions with myosin, and a role in thick filament structure and function has not been established. However, their high conservation across species suggests importance (Obermann et al., 1996; Labeit et al., 1992). Herein, we investigated the function of the first four domains of titin’s P-zone, A164–A167, using a novel mouse model in which they are removed, Ttn^{ΔA164–A167}. Functional studies revealed mild cardiac diastolic functional changes and remodeling, along with transcriptional activation of stress and remodeling pathways. Extensor digitorum longus (EDL) muscle showed myosin fiber-type shifts and transcript-level changes in fast- and slow-associated genes, along with depressed contractile function. Structural studies in cardiac muscle indicate altered incorporation of titin and cardiac myosin binding protein-C (cMyBP-C) into the thick filament along with shorter A-bands. Taken together, the data indicate an important role of titin’s P-zone domains A164–167 in thick filament templating.

All procedures were performed according to the NIH Guide for the Care and Use of Laboratory Animals and approved by the Institutional Animal Care and University Committee of the University of Arizona. All animal experiments were performed at 3 mo of age.

The Ttn^ΔA164–167 mouse model was generated by the Genetically Engineered Mouse Model Core (Arizona Core Network). CRISPR guide RNAs (gRNAs) for deleting the sequence encoding mouse titin domains A164–167 (the mouse titin gene P-zone domains A164–A167) were designed using the CRISPOR Design tool (https://crispor.gi.ucsc.edu/). Four guides were selected, and the successful ones were as follows: 5′-ACC​CAT​GCA​GAG​AGA​ACT​ATAGG-3′ (5′gRNA1) and 5′-TGG​AGC​CTG​AAT​CGG​GAC​GTCGG-3′ (3′gRNA1) both on the reverse strand (PAM sequence in italic type). Synthego synthesized the chosen single guide RNAs (sgRNA). eSpCas9 recombinant protein was ordered from Millipore (ESPCAS9PRO-250UG). A single-stranded oligo (ssODN) with short homology arms for sequences on each side of the gRNA-mediated double-stranded break was designed and ordered from IDT (sequence: 5′-CAG​ACA​GTT​CAC​CAT​CGG​CGG​TTT​GCT​AGA​AGC​TAC​TGA​ATA​TGA​GTT​CCG​AGT​ATT​TGC​TGA​GAA​CGA​GAC​TGG​GCT​CAG​CCG​ACC​ACG​TAG​AAC​AGC​TAT​GTC​TGT​CAA​GAC​TAA​ACT​AAC​ACC​GAT​TCA​GGC​TCC​ACA​CTT​CAA​GGA​GGA​GCT​GAG​AAA​CCT​GAA​TGT​AAG​GTA​TCA​GAG​CAA​CGC​CAC​TCT​GGT​CT-3′). Ribonucleoprotein complexes were assembled by incubating recombinant Cas9 protein with sgRNAs for 10 min at room temperature. Then, the ss oligo was added to the mixture, followed by 10 min of centrifugation at 10,000 rpm. The final concentrations used for the microinjections were 50/30/50 ng/μl, respectively. Fertilized eggs were collected from the oviducts of super-ovulated BL6/NJ females. Microinjection was performed by continuous-flow injection of the Cas9/gRNA/ssODN mixture into the pronucleus of 1-cell zygotes. Tail tipping of the newborn mice was utilized to purify DNA for genotyping by PCR, employing three screening primers: P1 5′-AGA​CAG​TGC​TTG​GAA​GAA​GAG​C-3′, P2 5′-GTT​GGT​GGC​TCT​GAC​TTG​G-3′, and P3 5′-ATT​GAA​GGT​CTT​GAA​TAT​GAG​TTC​C-3′ producing a 351-bp band for the wild type (WT) and 415-bp band in the mice where the deletion occurred. Two positive founders were obtained, F8 and M12, and bred to WT Black 6/NJ mice to confirm transmission of the desired modification in F1 generation and establish stable colonies (see Fig. S1 A for details).

Body weight data from WT (11 males, 10 females) and Ttn^ΔA164–167 (11 males, 4 females) mice were collected from 1 to 12 wk after birth. Dissection of muscles was performed on mice anesthetized with isoflurane (USP, Phoenix Pharmaceuticals, Inc.) and killed by cervical dislocation. The hearts were removed, and both atria, the right ventricle, and the left ventricle (LV) were rapidly dissected and weighed. The following skeletal muscles were dissected and rapidly weighed: tibialis cranialis, EDL, soleus, plantaris, gastrocnemius, and quadriceps. Both tibias were removed, and the mean tibia length was used for normalization of muscle weight data. All muscles were quick-frozen in liquid nitrogen and stored at −80°C.

3-mo-old male mice (5 WT, 5 Ttn^ΔA164–167) were held by the tail and allowed to grip the pull bar of a Columbus Instruments grip strength meter (Chatillon Force Measurement DFEII, Columbus Instruments). With both front paws grasping the grip bar, the mice were gently pulled away at a constant speed, until they released the bar from their paws. The maximum force generated during the pull (in grams) was recorded, and this was repeated two times, for a total of three pulls per mouse. For each mouse, the average of the three pulls is reported, and normalized to body weight.

Mouse tissues from LV and EDL were collected from 3-mo-old mice (female for EDL and mixed male and female for LV) and stored in RNAlater (Thermo Fisher Scientific) to preserve RNA integrity. For RNA extraction, 600 μl prechilled buffer RLT (RNeasy Fibrous Tissue Mini Kit, Qiagen) with 1% β-mercaptoethanol was added to muscle tissue stored in RNAlater in a 4-ml cryovial. Tissue was disrupted by using a rotor–stator homogenizer for 30 s. A protein digest was performed by adding 600 μl RNase-free water containing 6 mAU Proteinase K and incubating at 55°C for 10 min. Samples were transferred to a 1.5-ml microfuge tube and centrifuged for 3 min at 14,000 g. The supernatant was pipetted to a 2-ml tube with 600 μl ethanol and transferred to an RNeasy mini spin column. Thereafter, RNA extraction was performed following the manufacturer’s instructions and quantified using a NanoDrop ND-1000 spectrophotometer (Thermo Fisher Scientific). RNA integrity was checked by running the samples on 2100 Bioanalyzer (Agilent), and all RIN scores were confirmed to be ≥ 8. Directional libraries were prepared using NEBNext UltraExpress RNA Library Prep Kit after ribosomal RNA removal with NEBNext rRNA Depletion Kit v2.

Sequencing was performed on an Illumina NovaSeq X plus sequencer using 150-bp paired-end sequencing. Raw data are available under BioProject Accession ID PRJNA1412858. Adapters and low-quality reads were removed with fastp (Chen et al., 2018), and reads were mapped to the mouse genome (GRCm39/M36) using STAR (Dobin et al., 2013) with default settings.

DESeq2 (Love et al., 2014) was used to determine differentially expressed genes (DEGs) between WT and Ttn^ΔA164–167 for EDL and LV tissues based on read counts determined with STAR. Heatmaps of genes associated with fast or slow fiber types were generated by selecting a subset of genes that are well known to be associated with type 1, 2A, or 2X fibers as determined by single-fiber proteomics (Murgia et al., 2021). Gene expression values were scaled by z-score normalization (z = (x − μ)/σ) to make gene expression differences more comparable regardless of the gene’s overall expression level or variability. Visualization was done using the pheatmap R package (Kolde R [2025]). pheatmap: Pretty Heatmaps. R package version 1.0.13, https://github.com/raivokolde/pheatmap).

Gene ontology (GO)–term analysis was performed using NIH’s functional annotation tool, DAVID (Sherman et al., 2022), using the top 300 DEGs (by adjusted P value).

For calculation of inclusion percentages from titin exons, inclusion reads (IRs) and exclusion reads (ERs) were counted for each exon based on titin transcript ENSMUST00000099981. IRs are reads overlapping the exon being investigated, normalized by exon length. ERs are reads either upstream or downstream that support exclusions of the read. From these factors, the following equations were used to calculate the PSI index using the ASpli R package (Mancini et al., 2021): $IRi,n=IRilength exoni+read length−1, ERi,n=ERiread length−1$⁠,where i is the exon number and n the normalized read counts. The PSI_i of exon_i could then be calculated based on normalized counts as follows: $Ii=IRi,nIRi,n+ERi,n%$⁠.

Titin exon numbering in splice graphs is based on equivalent human exons from isoform ENST00000589042.

Protein expression was quantified as previously described (Karimi et al., 2024; Li et al., 2020). Flash-frozen LV and EDL tissues from 8 WT to 8 Ttn^ΔA164–167 3-mo-old male mice were pulverized in liquid nitrogen and then solubilized in urea buffer ([in mol/L]: 8 urea, 2 thiourea, 0.05 Tris-HCl, 0.075 dithiothreitol [DTT] with 3% SDS, and 0.03% bromophenol blue, pH 6.8) and 50% glycerol with protease inhibitors ([in mmol/L]: 0.04 E64, 0.16 leupeptin, and 0.2 phenylmethylsulfonyl fluoride [PMSF]) at 60°C for 10 min. Samples were centrifuged at 13,000 RPM for 5 min, aliquoted, flash-frozen in liquid nitrogen, and stored at −80°C. Titin isoform analysis was performed on solubilized samples (from n = 8 male; n = 8 male Ttn^ΔA164–167 90- to 100-day-old mice) using a vertical SDS-agarose gel system as previously described (Warren et al., 2003). Then, 1% gels were run at 15 mA per gel for 3:20, then stained using Coomassie brilliant blue, scanned using a commercial scanner, and analyzed with One-D scan (Scanalytics Inc.). The integrated optical density (IOD) of titin and myosin heavy chain (MHC) was determined as a function of loading volume (in a range of 5 volumes). The slope of the linear relationship between IOD and loading was obtained for each protein to quantify expression ratios.

For western blotting, solubilized samples were run on a 12% agarose gel, then transferred onto polyvinylidene difluoride membranes using a semi-dry transfer unit (Trans-Blot Cell, Bio-Rad). Blots were stained with Ponceau S to visualize the total protein transferred. Blots were then probed with primary antibodies followed by secondary antibodies conjugated with infrared fluorescent dyes. Blots were scanned using Odyssey Infrared Imaging System (Li-COR Biosciences). Primary antibodies included anti-ANT1/2 (Proteintech) and anti-GAPDH (Santa Cruz, Cell Signaling). Secondary antibodies used were CF790 goat anti-mouse and CF680 goat anti-rabbit (Biotium).

For MHC protein quantification, EDL and soleus samples from the n = 5–6 WT and n = 5–6 Ttn^ΔA164–167 3-mo-old mice used for whole-muscle mechanics (below), as well as n = 8 WT and n = 8 Ttn^ΔA164–167 LV samples, were solubilized as above. MHC isoform analysis was performed as previously described (Li et al., 2015). Briefly, MHC isoforms were separated using 8% (7% for LV samples) acrylamide gels, run for 24 h at 15°C and a constant voltage of 275 V, and stained with Coomassie blue. Gels were scanned and analyzed with ImageJ (v1.49, NIH, Bethesda, MD, USA). MHC type I and IIB are well separated on gels, but the IIA overlaps with IIX due to insufficient separation on most gels. We refer to this band as IIA/IIX. The identity of the bands was established by running standard lysates consisting of a mixture of EDL and soleus muscle lysates, containing all isoforms in a well-established order (i.e., IIA/X at top, IIB in middle, and I at bottom, e.g., see Fig. 2 D). For LV samples, a standard lysate known to contain both α-myosin and β-myosin was used.

The mouse titin A164–167 sequence (UniProt entry A2ASS6, amino acid positions 32322–32710) was codon-optimized for Escherichia coli expression by GenScript. The ALFA tag sequence (5′-SRLEEELRRRLTE-3′) (Götzke et al., 2019) was also codon-optimized, and both sequences were cloned into the pET-52b(+) expression plasmid between the BamHI and SacI restriction enzyme sites (see Fig. S2 A for plasmid map).

NiCO21 cells were transformed with the pET52b(+) plasmid encoding the titin A164–A167 segment. Clones were grown in 300 ml of LB broth supplemented with 0.1 mg/ml carbenicillin overnight at 30°C. The next day, the bacteria were pelleted by centrifugation and resuspended in 300 ml fresh LB broth supplemented with 0.1 mg/ml carbenicillin and 1 mM isopropyl β-D-thiogalactoside to induce protein expression overnight at 16°C. Cells were pelleted, followed by lysis and extraction using an EmulsiFlex-C3 (Avestin, Inc.). Cell extracts were purified using Ni-NTA resin (Qiagen). The purified protein was dialyzed into a HEPES-based storage buffer (50 mM HEPES, pH 7.5, 150 mM NaCl, 0.5 mM EDTA, 10% glycerol, 2 mM DTT).

Pulldowns from muscle tissue lysates were performed using the ALFA Selector ST system (Götzke et al., 2019). Tissue lysates were prepared by adding fresh tissue (gastrocnemius or LV) to 20 volumes of cold lysis buffer (50 mM HEPES, 150 mM NaCl, 20 mM NaPO₄, 20 mM β-glycerophosphate, 10 mM NaF, 2 mM EDTA, 1% IGEPAL CA-630, 10% glycerol, 1 mM MgCl₂, 1 mM CaCl₂, 2.5 mM PMSF, 2 mM Na₃VO₄, 20 μg/ml leupeptin, 10 μM E-64) and homogenizing using a Tissue Tearor handheld homogenizer (Cole-Parmer) in 10-s pulses. The homogenate was incubated on ice for 30–60 min, followed by clarifying by centrifuge at 25,000 × g.

Pulldowns were performed in mini spin columns (Enzymax LLC). The sdAb anti-ALFA ST resin (NanoTag Biotechnologies) was vortexed to suspend evenly, and 20 μl of resin slurry was transferred to each mini spin column. Resin was washed three times with lysis buffer by centrifuging for 1 min at 1,000 × g. ALFA-tagged recombinant protein (30 μg) (titin fragment A164–167) and an excess of tissue lysate were added to the mini spin columns. Control columns contained lysate-only or recombinant protein-only. The columns were incubated overnight at 4°C with gentle agitation. Mini spin columns were then centrifuged for 1 min at 1,000 × g to remove unbound contents and washed five times with wash buffer (0.01 M phosphate-buffered saline [PBS], 20 mM β-glycerophosphate, 10 mM NaF, 2 mM EDTA, 2.5 mM PMSF, 2 mM Na₃VO₄, 20 μg/ml leupeptin, 10 μM E-64, 0.5% IGEPAL CA-630). A final wash was performed with 0.01 M PBS. Two denaturation/elution steps were performed. First, 10 μl SDS-PAGE buffer (4% SDS, 0.0625 M Tris-HCl, 10% glycerol, 0.2% bromophenol blue, 8 M urea, 5% β-mercaptoethanol) was added to each mini spin column, columns were incubated in a 95°C water both for 4 min, and elution into a fresh 2-ml tube (Eppendorf) was performed by centrifugation for 1 min at 1,000 × g. The second elution step was performed as above using 20 μl SDS–urea buffer and by centrifuging for 30 s at 10,000 × g.

An entire eluate was loaded onto a Criterion 4–20% TGX Precast Midi Protein Gel (#5671095; Bio-Rad) and resolved by SDS-PAGE. Gels were stained with Coomassie brilliant blue (2% H₃PO₄, 10% (NH₄)₂SO₄, 0.1% CBB-G250) and destained with 25% methanol. Gel bands of interest were sliced from the gels and processed as below.

The gel slices were subjected to trypsin digestion, and the resulting peptides were purified by C18-based desalting exactly as previously described (Kruse et al., 2017; Parker et al., 2019). In brief, the SDS-PAGE gel slices were placed in a 0.6-ml LoBind polypropylene tube (Eppendorf), destained twice with 375 μl of 50% acetonitrile (ACN) in 40 mM NH₄HCO₃, and dehydrated with 100% ACN for 15 min. After removal of the ACN by aspiration, the gel pieces were dried in a vacuum centrifuge at 60°C for 30 min. Trypsin (250 ng; Sigma-Aldrich) in 20 μl of 40 mM NH₄HCO₃ was added, and the samples were maintained at 4°C for 15 min prior to the addition of 50–100 μl of 40 mM NH₄HCO₃. The digestion was allowed to proceed at 37°C overnight and was terminated by the addition of 10 μl of 5% formic acid (FA). After further incubation at 37°C for 30 min and centrifugation for 1 min, each supernatant was transferred to a clean LoBind polypropylene tube. The extraction procedure was repeated using 40 μl of 0.5% FA, and the two extracts were combined and dried down to ∼5–10 μl followed by the addition of 10 μl of 0.05% heptafluorobutyric acid/5% FA (vol/vol) and incubation at room temperature for 15 min. The resulting peptide mixtures were loaded on a solid-phase C18 ZipTip (Millipore) and washed with 35 μl 0.005% heptafluorobutyric acid/5% FA (vol/vol) followed by elution first with 4 μl of 50% ACN/1% FA (vol/vol) and then a more stringent elution with 4 μl of 80% ACN/1% FA (vol/vol). The eluates were combined and dried completely by vacuum centrifugation, and 6 μl of 0.1% FA (vol/vol) was added followed by sonication for 2 min. 2.5 μl of the final sample was then analyzed by mass spectrometry.

HPLC-ESI-MS/MS was performed in positive ion mode on a Thermo Fisher Scientific Orbitrap Fusion Lumos Tribrid mass spectrometer fitted with an EASY-Spray source (Thermo Fisher Scientific) as previously described (Parker et al., 2019). In brief, NanoLC was performed using a Thermo Fisher Scientific UltiMate 3000 RSLCnano System with an EASY-Spray C18 LC column (cat. # ES803; Thermo Fisher Scientific, 50 cm × 75 mm inner diameter, packed with PepMap RSLC C18 material, 2 mm); loading phase for 15 min; mobile phase, linear gradient of 1–47% ACN in 0.1% FA for 106 min, followed by a step to 95% ACN in 0.1% FA over 5 min, hold for 10 min, and then a step to 1% ACN in 0.1% FA over 1 min and a final hold for 19 min (total run 156 min); Buffer A = 100% H₂O in 0.1% FA; Buffer B = 80% ACN in 0.1% FA; and flow rate, 300 nl/min. All solvents were of liquid chromatography–mass spectrometry grade. Spectra were acquired using Xcalibur, version 2.3 (Thermo Fisher Scientific). A “TopSpeed” data-dependent MS/MS analysis was performed (acquisition of a full scan spectrum followed by collision-induced dissociation mass spectra of the Top N most intense precursor ions within the 3-s cycle time). Dynamic exclusion was enabled with a repeat count of 1, a repeat duration of 30 s, an exclusion list size of 500, and an exclusion duration of 40 s. Tandem mass spectra were extracted from Xcalibur “RAW” files, and charge states were assigned using the ProteoWizard 3.0 msConvert script using the default parameters. The fragment mass spectra were searched against the Swiss-Prot Mus musculus database (17,097 entries) using Mascot (Matrix Science; version 2.4) using the default probability cutoff score. The search variables that were used were as follows: 10 ppm mass tolerance for precursor ion masses and 0.5 Da for product ion masses; digestion with trypsin; a maximum of two missed tryptic cleavages; variable modifications of oxidation of methionine and phosphorylation of serine, threonine, and tyrosine. Cross-correlation of Mascot search results with X! Tandem was accomplished with Scaffold (Proteome Software). Probability assessment of peptide assignments and protein identifications was made using Scaffold. Only peptides with ≥95% probability were considered.

The EDL muscles were excised from both hindlimbs (3–8 fibers from each of 3 WT and 3 Ttn^ΔA164–167 male and female mice) and demembranated overnight in skinning solution at 4°C (see the Solutions section). On the following day, the EDL muscles were thoroughly washed with relaxing solution (see the Solutions section) and stored at 4°C in fresh relaxing solution, which was refreshed every 24 h, for up to 3 days for experiments.

For experiment, the EDL muscles were transferred to fresh relaxing solution, single muscle fibers were carefully isolated, and both ends of each fiber were attached to T-clips allowing mounting onto a force transducer (403A; Aurora Scientific) and a high-speed length controller (802D-120-322-TJ; Aurora Scientific). Sarcomere length was simultaneously measured using a high-speed VSL camera integrated with ASI 900B software (Aurora Scientific). The fiber cross-sectional area (CSA) was determined at a sarcomere length of 2.4 µm by measuring width and height (via a right-angle prism) at three different locations along its length. The measurements were then averaged, and CSA was calculated by approximating the cross-section as an ellipse. Mechanical experiments were performed at 15°C.

Passive force, including peak, elastic, and viscous components, was measured at sarcomere lengths between 2.4 and 3.6 µm, as previously described (Han et al., 2025). Initially, the fiber was positioned at slack length to establish zero force and then adjusted to 2.4 µm to determine the CSA. The fiber was subjected to a passive stretch of 5% at a stretching velocity of 1%/s, followed by a 20-s hold. This stretch protocol was repeated 10 times, with sarcomere length progressively increased up to 3.6 µm before being positioned back to 2.4 µm.

Active force, cross-bridge kinetics, and calcium sensitivity were evaluated as previously described (Han et al., 2025). Following a 5-min rest at slack length, the fiber was stretched to a sarcomere length of 2.6 µm and activated (see the Solutions section). Activation was maintained until maximum steady-state force was reached for >5 s. Deactivation was then induced using relaxing solution. The obtained active forces were normalized to the fiber’s CSA measured at 2.6 µm to calculate stress (expressed in mN/mm²).

To prevent the possibility of slippage of the fibers from the rig during activation, micro-knots were tied at both ends of each fiber using 7-0 silk suture, and the knots were secured within T-clips. The T-clips were mounted onto the rig by hooks through the hole in each T-clip. To prevent the T-clip from moving during activation, a small amount of high-vacuum grease was applied to the hook tip, which fixes the T-clips to the rig.

Cross-bridge kinetics were quantified by measuring the rate of tension redevelopment (K_tr) (Han et al., 2025). Briefly, the muscle fiber was first set at 2.4 µm and activated to achieve maximum steady-state active force. Once the active force reached its steady state, the fiber was rapidly shortened by 20%, held at the shortened sarcomere length for 0.02 s, and then quickly returned to its initial sarcomere length of 2.4 µm. K_tr was determined by fitting a single-exponential function to the force redevelopment curve.

Calcium sensitivity of skeletal muscle fibers was examined by measuring active force at various calcium concentrations (expressed in pCa, −log[Ca²⁺]), ranging from 4.5, 5.7, 5.9, 6.0, 6.25, to 9.0 pCa, with one force–pCa curve per fiber (Han et al., 2025). Each fiber was first set at a slack sarcomere length to establish zero force and stretched to a sarcomere length of either 2.6 µm or 3.0 µm. The measurements began in pCa 9.0 solution, and the fiber was sequentially transferred to solutions of increasing calcium concentration once the steady-state active force was obtained. The active force was first converted to stress by using the CSA of the fibers and normalized to its maximum stress achieved at pCa 4.5. Stress–pCa data were then fit with a Hill equation to determine both the EC₅₀ (level of [Ca²⁺] where 50% of maximum active force was achieved) and the Hill coefficient.

Relaxing solution contains 40 mM BES, 10 mM EGTA, 6.56 mM MgCl₂, 5.88 mM Na-ATP, 46.35 mM potassium propionate, 15 mM creatine phosphate, 1 mM DTT, and protease inhibitors including 1 mM E64, 1 mM leupeptin, and 1.25 mM PMSF (Ogut et al., 1999). Activating solution contains 40 mM BES, 10 mM CaCO₃-EGTA, 6.29 mM MgCl₂, 6.12 mM Na-ATP, 45.3 mM potassium propionate, 15 mM creatine phosphate, 1 mM DTT, and protease inhibitors (same as in the relaxing solution). For skinning solution, 1% Triton X-100 was mixed into the relaxing solution. Submaximal activating solutions, ranging from 5.7 to 6.25 pCa for calcium sensitivity measurements, were obtained by mixing relaxing and activating solutions based on Fabiato Program, set at 15°C (Ogut et al., 1999). All solutions were set at pH 7.0.

The compositions of all solutions have been reported previously (Pappas et al., 2018). Activating solution and relaxing solution were mixed to obtain activating solutions containing between 0.64 and 46.8 μM [Ca²⁺] (pCa 6.2–4.3). Experiments were performed at 15°C.

Single-cell experiments were performed on an inverted microscope stage using an Aurora Scientific 803B permeabilized myocyte apparatus with a 406A force transducer. Permeabilized single cells were obtained by first thawing a piece of frozen LV wall tissue in relaxing solution containing 1% Triton X-100 (Thermo Fisher Scientific) and then homogenizing the tissue. The obtained cells were gently pelleted with a 3-min centrifugation at 500 rpm. This pellet was resuspended in relaxing solution, and this pelleting/resuspending procedure was repeated two more times. 3–4 cells from each of four WT and four Ttn^ΔA164–167 male mice were studied.

For passive stress measurements, single cells were set to their slack sarcomere length (∼1.85 μm) and the cell was then lengthened at 1 length/s to an amplitude that varied from ∼1.9 to ∼2.4 µm. Once stretched, the cell was held at this new length for 10 s to allow for stress relaxation to occur, and the cell was then quickly released back to the slack length and held there for 10 s. This stretch–hold–release protocol was repeated until a range of SLs was achieved. From each stretch–hold–release protocol, the peak passive stress was obtained, the elastic stress value was obtained from the last 100 ms of the 10-s hold, and the viscous stress was obtained from the difference between peak stress and elastic stress. Six WT and six Ttn^ΔA164–167 male mice were used, with six cells from each mouse tested.

(Prosser et al., 2011) where P_0min gives the minimum specific force, P_0max gives the maximum specific force, F₅₀ defines the frequency where P₀ = 0.5 of P_0max, and 1/k is a measure of the steepness of the P₀ vs. F relationship. The curves for the different genotypes were also tested for significance using an extra sum-of-squares F test. For fatigue, an index was used, where the average of the last five values measured was divided by the average of the values from stimulations 2–6 (the first stimulation tends to produce highly variable force levels, with consistency being established at stimulation 2).

EDL muscles were dissected from three WT and three Ttn^ΔA164–167 female mice and placed in Cryomolds. Before dissection from the leg, the approximate length of the muscle was measured (from its insertion point in the knee to the distal tendon). Once in the Cryomold, the muscle was stretched to a length as close to the in situ length as possible and pinned through the Cryomold. The muscles were then embedded in the Cryomolds in optimal cutting temperature (OCT) compound and immediately frozen in 2-methylbutane precooled in liquid nitrogen. Then, 10-mm cross-sections were cut from the center of the muscle and mounted onto microscope slides. Tissue sections were permeabilized in 0.2% Triton X-100/PBS for 20 min at room temperature, blocked with 2% bovine serum albumin (BSA) and 1% normal donkey serum in PBS for 1 h at 4°C, and incubated overnight at 4°C with rabbit anti-laminin antibody (L9393; Sigma-Aldrich) diluted in 50% blocking solution. Sections were washed with PBS two times for 5 min and incubated with Alexa Fluor 594 goat anti-rabbit secondary antibody (A11037; Invitrogen) diluted in 50% blocking solution for 3 h at room temperature. The sections were washed two times for 5 min with PBS, and coverslips were mounted with Vectashield Mounting Media with DAPI. Slides were imaged on a Zeiss Axio Imager M.1 microscope utilizing a Zeiss AxioCam MRc and 10 × magnification. The total number of fibers was counted using the MyoSight plugin for FIJI (Babcock et al., 2020), which also generates an image mask of the nuclei. The number of fibers containing central nuclei was then manually identified.

Male WT (n = 8) and Ttn^ΔA164–167 (n = 12) mice at 90 days of age were anesthetized with isoflurane, then placed in dorsal recumbence on a heated platform (body temperature: 37°C). Transthoracic echo images were obtained with a Vevo 2100 High-Resolution Imaging System (VisualSonics, Inc.) using the model MS550D scan head for cardiac. Care was taken to avoid animal contact with excessive pressure, which could induce bradycardia. Imaging was performed at a depth setting of ∼1 cm. Images were collected and stored as a digital cine loop for offline calculations. Standard imaging planes and functional calculations were obtained according to American Society of Echocardiography guidelines. The parasternal long-axis view and mid-wall cross-sectional view of the LV were used to guide calculations of percent fractional shortening, percent ejection fraction, and ventricular dimensions and volumes. The left atrial dimension was measured in the long-axis view directly below the aortic valve leaflets. Passive LV filling peak velocity, E (cm/s), and atrial contraction flow peak velocity, A (cm/s), were acquired from the images of mitral valve Doppler flow from tilted parasternal long-axis views. A sweep speed of 100 mm/s was used for M-mode and Doppler studies. The heart rates of animals during the echocardiographic study were maintained in the range of 500–550 beats/min for M-mode, 450–500 beats/min for B-mode, and 400–450 beats/min for Doppler studies (for details, see Table S2).

Hearts from three WT and three Ttn^ΔA164–167 male and female mice were fixed in situ by formaldehyde perfusion to maintain all hearts at near diastasis. Mice were injected with 50–100 units of heparin 15–30 min before tissue collection. Mice were anesthetized with isoflurane and killed by cervical dislocation, followed by immediately opening the chest cavity. The femoral artery was then exposed and cut to open the circulatory system, and then, the apex of the heart was slowly injected with 3–4 ml of cardiac arrest solution (30 mM KCl, 30 mM 2,3-butanedione monoxime in 1X HEPES), followed by injection of 10 ml 10% buffered formalin by a syringe pump at a rate of 3 ml/min. Once the formalin injection finished, the heart was carefully removed and stored in 10% buffered formalin. The fixed hearts were processed for histological staining using a Leica ASP300S Tissue Processor. Each heart was cut into four equal transverse slices, and all four slices for each heart were embedded in one paraffin block. 5-μm sections were cut and mounted on glass microscope slides, followed by staining with Masson’s trichrome stain. Slides were imaged on a Zeiss Axio Imager M.1 microscope utilizing a Zeiss Axio Cam MRC and 20× magnification.

Ultrastructural labeling of the C-zone (cMyBP-C domains C5C7), titin P-zone (titin A168–170), and titin M-band (titin M8–M10) epitopes was performed on skinned LV papillary muscle from 3-mo-old male and female WT (2 mice per epitope) and Ttn^ΔA164–167 mice (2 mice per epitope). Papillary fiber bundles were dissected and skinned in relaxing solution (in mM): 40 BES, 10 EGTA, 6.56 MgCl₂, 5.88 Na-ATP, 1 DTT, 46.35 K-propionate, 15 creatine phosphate, pH 7.0, containing 1% Triton X-100, and protease inhibitors ([in mM]: 0.1 E64, 0.47 leupeptin, and 0.25 PMSF), then secured with aluminum T-clips, stretched by ∼20% from the slack length, and processed by the immunoelectron microscopy (IEM) pre-embedding technique previously described (Tonino et al., 2017). Cardiac fiber bundles were skinned twice, rinsed in relaxing solution with inhibitors, and immediately fixed with 3% paraformaldehyde in 10 mM PBS, pH 7.2, for 30 min at 4°C, followed by rinses with PBS and PBS containing protease inhibitors (0.04 mM E64 and 0.16 mM leupeptin), and incubation with glycine 50 mM in the same buffer. Blocking of muscle bundles was then performed with a solution of 0.5% BSA in PBS containing protease inhibitors and 0.05% Tween-20 for 1 h at 4°C. Afterward, muscle fiber bundles were incubated for 48 h at 4°C with rabbit polyclonal antibodies against the following titin domains: anti-titin A168–170 (TTN-8; Myomedix), anti-titin M8–M10 (TTN-9; Myomedix), and anti-cMyBP-C (domains C5C7), continued with three rinses in PBS containing protease inhibitors, and incubated with secondary goat anti-rabbit IgG antibody conjugated to Alexa Fluor 568 (ab175471; Invitrogen), overnight at 4°C. Negative controls were performed in fiber bundles by replacing each primary antibody with 0.5% BSA in PBS containing protease inhibitors. Subsequently, muscle bundles were fixed with 3% glutaraldehyde in 10 mM PBS, pH 7.2, for 30 min at 4°C, postfixed in 1% osmium tetroxide in the same buffer, and processed for routine transmission electron microscopy (TEM), as follows. Briefly, cardiac fiber bundles were dehydrated in an increased ethanol-graded series, infiltrated with propylene oxide and a mix of 1:1 propylene oxide: Araldite 502/Embed 812 resin, then transferred to three changes of fresh resin, embedded, and polymerized upside down in Beem capsules for 48 h at 60°C. Ultrathin (90 nm) longitudinal sections were obtained with a PowerTomeXL ultramicrotome (RMC Boeckeler Instruments, Inc.), collected on formvar-coated copper grids, and contrasted with 1% potassium permanganate and lead citrate. Images were registered in a Tecnai Spirit G2 BT transmission electron microscope with a side-mounted AMT Image Capture Engine V6.02 (4Mpix) digital camera operated at 100 kV. Digital images were calibrated with Fiji/ImageJ software 2.14.0v (NIH), and the density plot profiles were analyzed to determine the width of A-band, and titin A168–170, titin M8M10, and cMyBP-C epitope distances were measured across the M-band, as well as the distance between cMyBP-C stripes.

The sample preparation process for TEM is well known to cause shrinkage at the sarcomere level of around 10% (Sjöström and Squire, 1977; Sosa et al., 1994; Tonino et al., 2019); therefore, some fiber bundles were colabeled with an antibody against cMyBP-C (domains C5–C7) and its known spacing of 43 nm was used to generate a correction factor to apply to each sarcomere. However, for practical reasons, not all fiber bundles were dual-labeled, and in those sarcomeres, a correction factor based on sarcomere length, using a linear regression, was generated from the sarcomere length vs. correction factor plot from the experiments that did use cMyBP-C colabeling (Fig. S6 A). All length measurements obtained from IEM were corrected for shrinkage.

For cMyBP-C measurements, the distance from stripe 1 on one side of the sarcomere to the corresponding stripe 1 across the M-band was measured and halved to estimate its distance from the center of the sarcomere. The same process was done for stripe 9 so that the width and overall location of the C-zone could be estimated. The distance between at least 4 other stripes was measured and used to calculate the percent shrinkage of the sample based on the known distance between stripes of 43 nm. The known 43-nm distance was divided by the average measured distance to calculate a correction factor, by which all measurements for each sarcomere were multiplied.

WT (n = 3 per epitope) and Ttn^ΔA164–167 (n = 3 per epitope) male and female mice aged 3 mo were used in the super-resolution structured illumination microscopy (SR-SIM) study. Experiments were performed as previously described (Tonino et al., 2019). Skinned cardiac fiber bundles (prepared as described above) were embedded in OCT compound and immediately frozen in 2-methylbutane precooled in liquid nitrogen. Then, 5-μm-thick cryosections were cut and mounted onto microscope slides. Tissue sections were permeabilized in 0.2% Triton X-100/PBS for 20 min at room temperature, blocked with 2% BSA and 1% normal donkey serum in PBS for 1 h at 4°C, and incubated overnight at 4°C with primary antibodies diluted in 50% blocking solution. The primary antibodies included (Table S1) the following (dilution: 1:200-1:300): a mouse monoclonal anti-titin “Ti102” antibody (domains I111–I112) (Jin, 1995), a rabbit polyclonal titin “MIR” antibody (domains I109–I111) (Myomedix) (Centner et al., 2000), a mouse monoclonal anti-α-actinin antibody (EA-53; Sigma-Aldrich), and a rabbit polyclonal anti-titin A77–A78 antibody (Bucher et al., 2010; Muhle-Goll et al., 2001). Sections were then washed with PBS two times for 30 min and incubated with secondary antibodies diluted in 50% blocking solution for 3 h at room temperature. The secondary antibodies (dilution: 1:200–1:300), obtained from Invitrogen and Abcam, included the following: Alexa Fluor 488–conjugated goat anti-mouse IgG, Alexa Fluor 568–conjugated goat anti-rabbit IgG, Alexa Fluor 568–conjugated goat anti-mouse IgG, and Alexa Fluor 568–conjugated donkey anti-rabbit IgG. The sections were then washed with PBS two times for 15 min and covered with number 1.5H coverslips using ProLong Glass Antifade Mountant (Thermo Fisher Scientific, Inc.). A Zeiss ELYRA S1 SR-SIM microscope was used with ultraviolet and solid-state laser (488/561/642 nm) illumination sources, a 100 × oil immersion objective (NA = 1.46) for 100× magnification, and a sCMOS camera. Typical imaging was performed on a 49.34 × 49.34 μm² area with 1,280  ×  1,280 pixel dimensions. Image stacks comprising 20 slices were acquired with 0.101-μm Z-steps, five angles, and five phases/angle for each slice. Image reconstruction and fluorescence intensity plot profile generation were performed with ZEN 2 software (Zeiss). Plot profiles were fit with Gaussian curves to determine the epitope peak position using Fityk 1.3.0 software. The Ti102 and A77–78 epitope positions were measured across the M-band and divided by 2 to approximate distance from the center of the sarcomere. There is negligible shrinkage of frozen samples prepared for immunofluorescence (Tonino et al., 2019), so no correction factor is applied to those measurements, which are expressed as average ± SD.

Data were analyzed and visualized with GraphPad Prism 10.2.1. Results are expressed as mean ± SD. Two-tailed, unpaired t tests were used to compare two groups (titin and myosin protein ratios, muscle and body weight measurements, echocardiography data). Where multiple data points from the same mouse were considered, nested t tests or nested ANOVA (with Sidak’s multiple comparisons test) was used (SIM and IEM data, single EDL fibers, and cardiomyocytes). Two-way ANOVA or mixed-effects analyses (where appropriate, with repeated-measures correction and Sidak’s multiple comparisons rest) and/or curve fitting compared by an extra sum-of-squares F test was used to compare data that were analyzed as a relationship, i.e., sarcomere length–stress relationship, stress–calcium relationship, force–frequency relationship. P values <0.05 were considered statistically significant.

Fig. S1 shows details of the generation and characterization of the Ttn^ΔA164–167 mouse model. Fig. S2 displays the plasmid used to generate the bait protein for pulldown assays and results of the assays. Fig. S3 shows analysis of centrally located nuclei in EDL muscle. Fig. S4 shows the soleus muscle functional data. Fig. S5 shows the WLC modeling results and intact EDL absolute force data. Fig. S6 shows the linear regression used to calculate IEM correction factors and epitope distance vs. sarcomere length relationships. Table S1 lists antigen, host species, concentration, and source details of the antibodies used for titin and cMyBP-C mapping. Table S2 lists all relevant measured echocardiography parameters.

The Ttn^ΔA164–167 mouse model was generated by making a deletion from the end of exon 337 to within exon 340 from the mouse titin gene using CRISPR/Cas9 to remove the sequence encoding domains A164–A167 (Fig. 1 A and Fig. S1 A). The A168–A170 segment of the P-zone, which houses a binding site for E3 ubiquitin ligases (Müller et al., 2021), was not disrupted. Homozygous Ttn^ΔA164–167 mice are viable, are born to heterozygous parents at expected Mendelian genotype ratios, and have normal body weights and skeletal muscle weights (Fig. S1, B–E). RNA-sequencing analysis of the titin gene in adult LV myocardium and EDL muscle revealed that the correct exons are missing from titin transcripts in homozygous mice (Fig. 1, C and D). Titin splicing in LV was otherwise normal. In EDL muscle, we found slightly higher inclusion (14%) of M-band exon 5, and decreased exon inclusion in the I-band (Fig. 1 D). This is reflected in SDS-PAGE protein analysis, which demonstrates normal titin protein levels in adult LV and EDL muscles, but with a minor band of increased mobility in EDL, referred to as N2A-2 (Fig. 1 E). Altered I-band splicing in EDL muscle is a common phenomenon in mutant titin mice (van der Pijl et al., 2020), and could impact titin-based tension in myocytes (see below). The total amount of titin protein and its major degradation product T2 were measured in LV and EDL homogenates by optical density analysis of SDS-PAGE gels and normalized to the optical density of MHC (Fig. 1, F and G, respectively). These ratios in Ttn^ΔA164–167 tissues were not different from WT tissues. As an initial evaluation of muscle function in the Ttn^ΔA164–167 mouse model, we performed a front-limb grip strength assay and found that Ttn^ΔA164–167 mice gripped with ∼30% less force relative to body weight than WT mice (Fig. 1 H), indicating that further investigation was warranted. Overall, characterization of the Ttn^ΔA164–167 model at the RNA and protein level indicated it to be well suited for studying the role of titin’s P-zone domains A164–A167.

Since it is unknown whether titin’s P-zone domains A164–167 have binding partners, we performed pulldown assays to probe protein–protein interactions. We used a recombinantly expressed titin A164–167 bait protein with C-terminal ALFA and 10× histidine tags (Fig. S2 A) and performed the pulldown using the ALFA Selector ST system (Götzke et al., 2019). The bait protein was incubated with tissue lysate from WT LV or skeletal muscle (gastrocnemius complex due to its large size). Specific bands on the resulting SDS-PAGE gels following the pulldown were analyzed by mass spectrometry. A band that consistently appeared enriched in the skeletal muscle pulldown lanes and that was ∼100 kDa in molecular weight was identified by mass spectrometry to be most enriched in the protein SERCA1 (Fig. S2, B and C). Since we did not find mechanical changes indicative of altered calcium reuptake (see below) and it is known that SERCA1 localizes to the sarcoplasmic reticulum membrane (Wu et al., 1995), we interpreted this to be an artifact of the lysate conditions in which the pulldown assay is performed. Highly enriched in the pulldown lane from LV muscle was a band of ∼25 kDa, identified to contain adenine nucleotide transporter 1 (ANT1, also known as ADP/ATP translocase 1) (Fig. S2, B and C), which localizes to the inner mitochondrial membrane and exchanges ADP and ATP across the membrane (Willis et al., 2018); it is therefore unlikely to come into contact with titin. Western blots indicated no change in the overall levels of ANT1/2 in the Ttn^ΔA164–167 LV and EDL lysates (Fig. S2, D and E). Thus, we did not pursue these hits as they are unlikely interactors of titin’s P-zone in the sarcomere.

We performed RNA sequencing to analyze transcriptomic changes in the LV and EDL tissue of Ttn^ΔA164–167 mice. There were 1,964 DEGs in the LV tissue; 768 were upregulated and 1,196 were downregulated. We performed GO enrichment analysis of the top 300 DEGs (Fig. 2 A). In the molecular function category, significantly enriched terms were related to protein–protein and protein–ATP binding. In the biological process category, significantly enriched terms were related to metabolic processes, including glycolysis, fatty acid metabolism, and glycogen metabolism. This category also included positive regulation of the ERK1/2 signaling cascade, which is often associated with cardiac remodeling and response to stress (Nomura et al., 2018). The cellular component category included enrichment of terms related to mitochondrial components and extracellular matrix. Visualization of DEGs by volcano plot showed modest expression changes in Ttn^ΔA164–167 LV relative to WT by log₂ fold change (Fig. S1 F). Genes of interest are annotated and include a subset of markers of stress and hypertrophy. The most highly upregulated genes were Rny3 and Rny1, which are noncoding Y RNAs associated with inflammation and coronary artery disease (Repetto et al., 2015). Also upregulated are Col3a1 and Prkca, which are linked to fibrosis and hypertrophic remodeling (Herum et al., 2022; Braz et al., 2004; Muth et al., 2001). Other notable upregulated genes include Myh7, Myh7b, and Mybpc2, which align with expression patterns in diseased hearts (Nandi and Mishra, 2015; Dirkx et al., 2013; Del Gaudio et al., 2023). To verify whether there was a MHC expression switch at the protein level consistent with fetal gene reprogramming, we performed SDS-PAGE analysis, and found that Ttn^ΔA164–167 LV, like WT, expresses exclusively α-myosin (Fig. 2 B). Together, the RNA-seq results in the Ttn^ΔA164–167 LV indicate the differential expression of genes associated with stress and remodeling, as well as energetics and metabolism.

We also performed RNA sequencing on EDL tissue and analyzed the results similarly. Ttn^ΔA164–167 EDL had 826 DEGs: 425 were upregulated and 401 were downregulated. Examination of the upregulated genes revealed strong increases in the expression of genes associated with slow muscle fiber types, including MYH (MyHC), troponin I, troponin C, myosin light chains, and cMyBP-C (Kedlian et al., 2024; Talbot and Maves, 2016). In conjunction with this, we found downregulation of the corresponding fast muscle fiber–associated genes. These results are represented as a heatmap in Fig. 2 C. Also upregulated were Ankrd1 and Ankrd2, which are typically associated with mechanical stress (Witt et al., 2005; Kojic et al., 2004; Kojic et al., 2011) (Fig. S1 G). Overall, the transcriptomic changes in the Ttn^ΔA164–167 EDL were consistent with stress and compensation processes.

The marked upregulation of slow-twitch–associated genes and downregulation of fast-twitch–associated genes in the Ttn^ΔA164–167 EDL prompted us to perform MyHC fiber typing by SDS-PAGE. This confirmed a shift toward slower MyHC paralogs in Ttn^ΔA164–167 mice. WT EDL muscle contained 12% type IIA/IIX fibers and 88% type IIB fibers. Ttn^ΔA164–167 EDL showed an increase in IIA/IIX myosin and decrease in IIB myosin, containing 26% IIA/IIX and 74% IIB fibers (Fig. 2 D). Thus, a mild but significant fiber-type switch (12% increase in slower IIA/IIX fiber content) is evident at the protein level in the EDL tissue, consistent with transcriptional changes.

In EDL muscle of Ttn^ΔA164–167 mice, inclusion of titin’s I-band exons was significantly reduced. This change is predicted to increase strain on the extensible I-band segment for a given sarcomere stretch, thereby elevating passive stress. To quantify this effect, we applied a serially linked WLC model of titin’s extensible I-band segment (see Materials and methods). Modeling predicted a marked increase in passive stress, particularly at sarcomere lengths exceeding ∼3.0 µm (Fig. S5 A). An additional phenotype in Ttn^ΔA164–167 mice was a ∼43-nm reduction in thick filament length per half-sarcomere (see below), which is likewise predicted to further raise passive stress (Fig. S5 A). A similar WLC-based simulation was performed for cardiac muscle, where no change in I-band exon splicing was detected; here, the ∼43-nm half-filament shortening alone was predicted to increase passive force (Fig. S5 B).

Mechanical experiments were conducted on skinned single fibers dissected from EDL muscle. To enable direct comparison across samples, all fibers used for mechanical testing were fiber-typed by SDS-PAGE, and only the results from type IIB fibers were analyzed. Passive Stress Measurements—Passive stress was assessed by imposing a ramp stretch–hold–release protocol and recording the resulting stress. Ttn^ΔA164–167 fibers generated greater peak passive stress than WT (Fig. 3 A), particularly at long sarcomere lengths (3.4–3.6 μm), when analyzed by nested ANOVA. The elastic component of passive stress was also elevated at long sarcomere lengths in Ttn^ΔA164–167 fibers (Fig. 3 B), while the viscous component trended higher (Fig. 3 C). Note that the exponential curve fits, when tested by an extra sum-of-squares F test, were different between WT and Ttn^ΔA164–167 for peak, elastic, and viscous passive stress. These findings align with the reduced exon inclusion in titin’s I-band segment, which is predicted to increase passive stress. Active Stress Measurements—Maximal Ca²⁺ activation revealed no difference in active stress (force normalized to CSA) between Ttn^ΔA164–167 and WT fibers (Fig. 3 D). However, the CSA of Ttn^ΔA164–167 fibers was ∼30% smaller than WT, and a ∼30% decrease in absolute force was measured in association with this (Fig. 3 E and Fig. S5 C). The K_tr following a slack–stretch during activation was unchanged (Fig. 3 F). Furthermore, the stress–Ca²⁺ relationship was similar between genotypes, with no shift in EC₅₀ (Fig. 3 G). In summary, single-fiber mechanical experiments demonstrate that type IIB EDL fibers from Ttn^ΔA164–167 mice have normal active mechanical properties, reduced CSA, and significantly elevated passive stress compared with WT fibers.

We next performed single skinned cardiomyocyte mechanical experiments. Passive Stress Measurements—Relaxed cardiomyocytes were subjected to a ramp stretch–hold–release protocol to measure passive stress. Peak passive stress was measured at a range of sarcomere lengths, and the elastic and viscous components were calculated. The sarcomere length–stress relationship was plotted for each component and fit to an exponential growth curve. Ttn^ΔA164–167 cardiomyocytes generated slightly higher passive stress as indicated by significantly different curve fits tested by an extra sum-of-squares F test (Fig. 4, A–C). Active Stress Measurements—The stress produced by skinned cardiomyocytes maximally activated with calcium was also measured but was not different between controls and Ttn^ΔA164–167 mice (Fig. 4 D). We tested the stress–Ca²⁺ relationship and found that calcium sensitivity, as indicated by EC₅₀, was not different in Ttn^ΔA164–167 mice compared with controls (Fig. 4 E). Cardiomyocyte studies therefore indicated that active stress and calcium sensitivity are unaffected in Ttn^ΔA164–167 mice, but that passive stress is mildly increased.

Front-limb grip strength assays indicated that Ttn^ΔA164–167 mice were weaker than controls (Fig. 1 H), so we proceeded with ex vivo whole-muscle mechanical experiments to test the function of a representative fast-twitch muscle, EDL. The EDLs from WT and Ttn^ΔA164–167 mice were stimulated at frequencies ranging from 1 Hz (twitch) to 250 Hz (maximal tetanus). Ttn^ΔA164–167 EDLs produced less specific force (force normalized to PCSA) than WT controls at all stimulation frequencies. The maximal specific force (250 Hz) generated by Ttn^ΔA164–167 EDLs was, on average, 28% less than WT EDLs (Fig. 5 A), but the frequency at which 50% of the maximal force was produced, F₅₀, was not changed in Ttn^ΔA164–167 EDLs (Fig. 5 B). We noted that the PCSA of the Ttn^ΔA164–167 EDLs was significantly smaller than WT (Fig. 5 A inset), and therefore also compared the absolute EDL force generation (not normalized to PCSA). Maximal absolute force was decreased to a larger degree in Ttn^ΔA164–167 EDLs than specific force, with an average decrease of 35% compared with WT EDLs (Fig. S5 C). Twitch (1 Hz) stimulations were used to study contraction and relaxation kinetics and showed that maximal rate of contraction (max dx/dt) was decreased in Ttn^ΔA164–167 EDLs compared with WT, but maximal relaxation rate (minimum dx/dt) was unaffected (Fig. 5, C and D). The EDL muscles also underwent a fatigue experiment, in which they were rapidly stimulated 75 times at a frequency of 60 Hz. Fatigue index, defined here as the ratio of the average force produced by the last five stimulations to the average force produced by the first five stimulations, was used to indicate fatigue resistance. Ttn^ΔA164–167 EDLs had an increased fatigue index, indicating increased resistance to fatigue compared with WT EDLs (Fig. 5 E). Because muscle weakness with fiber-type shifting can be a myopathic phenotype, we examined EDL cross-sections for the presence of central nuclei, a prominent myopathy marker. There was no difference in the proportion of fibers with central nuclei between the genotypes, with an average of 2% of Ttn^ΔA164–167 and 0.8% of WT fibers containing central nuclei (Fig. S3). Thus, the EDL muscles of Ttn^ΔA164–167 mice are weaker and have kinetic and fatigability changes consistent with the gene and protein expression shift toward a slow-like phenotype, but with no overt signs of myopathy.

To determine whether the functional effects measured in the EDL are consistent across skeletal muscles, we tested soleus function by whole-muscle mechanics. In contrast to our findings in the EDL muscle, the Ttn^ΔA164–167 soleus muscles generate force at levels similar to WT, with no significant differences as determined by two-way ANOVA (Fig. S4). However, the force–frequency curve fit was statistically different between WT and Ttn^ΔA164–167. Therefore, the slightly lower forces produced by the Ttn^ΔA164–167 when stimulated at higher frequencies (60 Hz and above) are relevant and suggest that, although very minor, soleus function is impacted by the deletion of domains A164–167. As with the EDL muscles, we measured MyHC content of the mechanically tested soleus muscles and found a more drastic fiber-type switch between WT and Ttn^ΔA164–167 muscles, with type I fiber content increasing by ∼45% in Ttn^ΔA164–167 soleus relative to WT.

Echocardiography showed that systolic function, as indicated by both fractional shortening and ejection fraction, was normal in Ttn^ΔA164–167 mice (Fig. 6 A and Table S2). However, several parameters suggest the presence of mild changes to diastolic function. Specifically, isovolumic relaxation time (normalized to RR interval) and myocardial performance index showed strong increasing trends (P = 0.072 and P = 0.0824, respectively) in Ttn^ΔA164–167 mice, suggesting a tendency toward delayed relaxation. Mitral valve E-wave deceleration time, which correlates inversely with myocardial stiffness, was significantly decreased by 23%. Mitral valve E/A ratio trended higher in Ttn^ΔA164–167 mice (P = 0.0789), driven by a trending decrease in A-wave velocity. Collectively, these findings suggest potential early-stage diastolic filling impairment (Mishra et al., 2007; Mandinov et al., 2000) (Fig. 6 A). In line with the increased LV tissue weights of Ttn^ΔA164–167 mice, we found a trending increase in LV chamber volume during both systole and diastole (P = 0.130 and P = 0.135, respectively) (Table S2). Although LV volume was slightly larger in Ttn^ΔA164–167 mice, the eccentricity index remained normal, suggesting enlargement of the LV without true dilation. Upon dissection and weighing of the chambers, we found significant increases in the mass of all cardiac chambers of both male and female Ttn^ΔA164–167 mice. LV mass increased by 17%, RV mass increased by 21%, and combined atrial mass increased by 45% compared with WT mice (Fig. 6 B). Thus, despite having preserved systolic function, Ttn^ΔA164–167 mice exhibit signs of mildly impaired diastolic filling and remodeling.

To determine whether myocardial fibrosis was a contributing factor to the organ-level diastolic functional changes detected by echocardiography, we examined Ttn^ΔA164–167 hearts histologically. Qualitative assessment of Masson’s trichrome-stained WT and Ttn^ΔA164–167 hearts found no overt fibrosis (Fig. 6 C), indicating this is not a driving factor of the functional changes observed at 3 mo of age.

We mapped titin’s position in the thick filament using immunolabeling. The four-domain deletion of domains A164–167 is expected to cause a 16-nm length reduction in titin, as each Ig-like and FnIII-like domain spans ∼4 nm when folded (Trinick et al., 1984; Pfuhl and Pastore, 1995; Dutta et al., 2023). Four antibodies that label titin epitopes along the A-band segment were chosen to determine titin’s position on either side of the deleted domains (see Fig. 7 A for details). IEM and SR-SIM were used to visualize antibody-labeled muscle sections.

We first probed the location of titin’s C-terminal M-band domains using the M8–M10 antibody (Fig. 7 B). This epitope measured an average of 41 nm from the center of the sarcomere in WT LV sarcomeres. In Ttn^ΔA164–167 LV sarcomeres, localization of this epitope was normal with an average distance of 43 nm from the sarcomere center, indicating that titin’s M-band likely incorporates normally in Ttn^ΔA164–167 cardiac sarcomeres (Fig. 7 B and Table 1).

Titin’s P-zone domains A168–170 lie immediately adjacent to the deleted domains A164–167 (Fig. 7 A), and immunolabeling with an antibody against these domains in WT mice yields two stripes on either side of the M-band when visualized by EM (Fig. 7 C, left). We interpret these two stripes to correspond to titin’s two conformations in this region, titin α and titin β, as described by Tamborrini et al. (2023). In our experiments, the titin α A168–170 epitope measured an average of 87 nm from the sarcomere center in WT mice, while the titin β epitope measured 125 nm from the center (a separation of 38 nm). In Ttn^ΔA164–167 mice, however, only one stripe was labeled by the A168–170 antibody, at a mean distance of 70 nm from the center (Fig. 7 C [right] and Table 1). This is an average shift toward the M-band of 17 nm from the WT titin α location and 55 nm from the WT titin β location (Fig. 7 C and Table 1).

We next probed titin’s location in the C-zone with an antibody against domains A77–78 (C-zone super-repeat 4) (Fig. 7 A). In WT LV, this antibody label was measured to localize an average of 443 nm from the center of the sarcomere, while in Ttn^ΔA164–167 LV, it localized an average of 407 nm from the center. The location of titin domains A77–78 is therefore shifted 37 nm toward the M-band in Ttn^ΔA164–167 mice (Fig. 7 D and Table 1).

We lastly probed the location of the edge of titin’s A-band with the Ti102 and MIR antibodies (Fig. 7 A). The Ti102 antibody binds to domains I111 and I112, and the MIR antibody binds to domains I109–I111 in the I/A junction, near the edge of the thick filament (Jin, 1995; Skeie et al., 1997). Since their epitopes overlap, both antibodies were used in experiments. In WT LV sarcomeres, the average epitope distance from the center of the sarcomere was 776 nm. In Ttn^ΔA164–167 sarcomeres, this distance was 736 nm (Fig. 7 E and Table 1), a shift of 40 nm toward the center of the sarcomere. Together, the titin epitope measurements show that titin’s C terminus is incorporated normally, that its α and β conformations are disrupted, and that epitopes distal to the A164–167 deletion are shifted uniformly by ∼40 nm in Ttn^ΔA164–167 mice (Fig. 7 E and Table 1).

Since titin is considered a thick filament template, we next examined whether the thick filament itself was altered. We measured A-band length in the immunoelectron micrographs from which titin epitopes were measured above (see Materials and methods for shrinkage correction details). The average thick filament length in WT samples was 1,531 nm, while in Ttn^ΔA164–167 samples, it was 1,428 nm (Fig. 8 A). Per half-sarcomere, these values are 765 nm (WT) and 714 nm (Ttn^ΔA164–167), close to the titin A-band segment length of 776 and 735 nm, respectively, as indicated by the Ti102 and MIR antibodies.

Lastly, we examined the location of a prominent thick filament accessory protein, cMyBP-C, using an antibody against domains C5–C7 (Luther et al., 2008), visualized by EM (Fig. 8 B). There are nine cMyBP-C stripes per half A-band, spaced 43 nm apart (in correlation with the length of a myosin helical repeat and a titin C-zone super-repeat), aligning with titin C-zone super-repeats 2 through 11 (Tonino et al., 2019; Bennett et al., 2020; Dutta et al., 2023; Tamborrini et al., 2023). Our findings indicate that the width of the C-zone (distance from stripe 1 to stripe 9) does not differ between WT and Ttn^ΔA164–167 mice (Fig. 8 C). Although the spacing of the cMyBP-C stripes is normal, their location within the thick filament differs from WT in Ttn^ΔA164–167 mice. While stripe 1 localized to 165 nm from the center of the sarcomere in WT mice, it localized to 126 nm in Ttn^ΔA164–167 mice, a difference of 39 nm (Fig. 8 D and Table 1). Stripe 9 measured 509 nm from the center of the sarcomere in WT mice but was shifted toward the M-band by 46 nm in Ttn^ΔA164–167 mice to an average of 463 nm from the center (Fig. 8 D and Table 1). These shifts are similar to the titin epitope shifts found in the C-zone and at the edge of the A-band (around 40 nm) and are consistent across a range of sarcomere lengths (Fig. S6). All epitope measurements are summarized in Table 1, and all epitope distance–sarcomere length relationships are plotted in (Fig. S6).

Various regions of titin have been systematically studied using deletion mouse models (Chung et al., 2013; Buck et al., 2014; Granzier et al., 2009; Granzier et al., 2014; Brynnel et al., 2018; van der Pijl et al., 2020; Radke et al., 2007; Lee et al., 2010; Tonino et al., 2017; Tonino et al., 2019; Gotthardt et al., 2003; Biquand et al., 2021), but the P-zone (domains A164–170) has not received much attention yet. The A164–167 segment’s unique domain sequence (Ig-Ig-FnIII-FnIII) and conservation across vertebrate titins and invertebrate twitchins (Obermann et al., 1996; Labeit et al., 1992) suggest that it is a necessary element of the titin molecule. In the sarcomere, titin incorporates into the thick filament and is thought to serve as a molecular template for thick filament assembly (Dutta et al., 2023; Tonino et al., 2017; Bennett et al., 2020; Fleming et al., 2023; Bennett and Gautel, 1996) and support regulation of thick filament activation (Park-Holohan et al., 2021) (reviewed in Granzier and Labeit [2025] and Brunello and Fusi [2024]). A high-resolution structure of the relaxed cardiac thick filament C-zone was recently published, which revealed that titin maintains precise structural patterns (e.g., kinks at each Ig-like domain in the C-zone that facilitate contacts between titin dimers and between titin and myosin) throughout the filament that likely contribute to its templating and regulatory functions (Dutta et al., 2023; Tamborrini et al., 2023). At the C-zone-to-P-zone transition, titin adopts two conformations: the α conformation, in which titin runs linearly through the P-zone and into the M-band, with domains A164–A167 positioned alongside myosin crowns P3 and P2; and the β conformation, in which a loop comprising domains A158–A167 wraps around the neck of myosins in crown A2 (Tamborrini et al., 2023). Following the loop, domains A168 through M1 are arranged linearly alongside myosin crown A1. Each thick filament contains three pairs of titin molecules, and within each pair, one molecule is thought to adopt the α conformation, while the other adopts the β conformation. Therefore, while six titin molecules run through the C-zone of the sarcomere, only three titin molecules might extend into the M-band.

The four titin P-zone domains (A164–A167) missing in the Ttn^ΔA164–167 mouse correspond to ∼16 nm of titin’s length. Thus, Ttn^ΔA164–167 titin is ∼16 nm shorter than WT titin, and we expected that IEM studies would reveal titin epitope shifts of that magnitude. We first probed the location of titin’s C-terminal end using the M8-M10 antibody and found that it is normally incorporated in the Ttn^ΔA164–167 sarcomere. The location of the remaining P-zone domains, A168–170, was studied with an antibody against this three-domain segment. In WT mice, two epitopes were detected, ∼40 nm apart (87 and 125 nm from the center of the sarcomere). This is the first antibody-based evidence of two titin conformations in the P-zone, and we interpret these two epitopes to correspond to the A168–170 location in titin α (87-nm stripe) and titin β (125-nm stripe), in agreement with the layout of titin established by Tamborrini et al. (2023). In Ttn^ΔA164–167 LV, only one stripe was labeled by the A168–170 antibody, which was closer to the M-band than the WT α titin epitope by ∼17 nm (70 nm from the sarcomere center). The presence of only one epitope suggests that all titin molecules adopt the same conformation in the Ttn^ΔA164–167 sarcomere, and, thus, that domains A164–167 are necessary for the β loop to form. In WT titin α, where titin domains are linearly arranged, the measured distance from M8–M10 to A168–170 is ∼43 nm. In Ttn^ΔA164–167 titin, the measured distance from M8-M10 to A168–170 is ∼30 nm. This M-ward shift of the A168–170 epitope was unexpected considering the M8–M10 epitope was positioned normally. This shift could be accounted for by rearrangements of titin’s M-band M-is, particularly the large and flexible M-is2 segment (which lies between the M8–M10 and A168–170 epitopes) (Lange et al., 2002; Bang et al., 2001), in response to the misalignment caused by the deletion.

Titin’s C-zone was studied using the A77–78 antibody and the I/A junction studied using the Ti102 and MIR antibodies. Both the C-zone and I/A junction epitopes were located ∼40 nm closer to the M-band in Ttn^ΔA164–167 sarcomeres. In WT mice, the Ti102/MIR epitope localizes at the edge of the thick filament and the distance across the M-band to the corresponding epitope represents thick filament length. In the case of the Ttn^ΔA164–167 model, it seemed plausible that the thick filament remained the same length despite that titin’s A-band segment had shifted, or that the thick filament shortened such that the Ti102/MIR epitope remained at the edge. We tested this by measuring the thick filament length in electron micrographs and found slightly shorter thick filaments in Ttn^ΔA164–167 sarcomeres, by ∼50 nm per half-thick filament, which is in relatively good agreement with the Ti102/MIR epitope shift. Thus, in the Ttn^ΔA164–167 mouse, titin’s distal A-band epitopes localize closer to the M-band by ∼40 nm and the thick filament shortened accordingly.

We also measured the location of cMyBP-C stripes by IEM and found that in Ttn^ΔA164–167 sarcomeres, the C-zone, with edges demarcated by cMyBP-C stripes 1 and 9, had shifted toward the M-band by ∼43 nm. Interestingly, the location of cMyBP-C stripe 1 in Ttn^ΔA164–167 sarcomeres occupied the location of the titin β A168–170 epitope in WT sarcomeres, and the Ttn^ΔA164–167 C-zone shift corresponds closely to the length of a C-zone super-repeat and a myosin helical repeat (3 crowns, 43 nm). This shift matches closely with the ∼40-nm shifts in titin A77–78 and Ti102/MIR epitope and suggests that the Ttn^ΔA164–167 thick filament components in the C-zone and beyond rearranged such that they remained in register with each other, as opposed to titin shifting independently of the other thick filament components as initially hypothesized. This leads us to propose a structural model of the thick filament in Ttn^ΔA164–167 mice in which the 3 myosin crowns in the P-zone are missing (crowns P1, P2, and P3), which would shorten the half-thick filament by 43 nm, thereby placing the first C-zone myosin crowns (crowns A1, A2, and A3) in the location normally occupied by the P-zone crowns (Fig. 9). Knowing titin’s role as a thick filament template, it seems likely that missing domains in the P-zone impair the ability of the myosin molecules in that region to assemble, resulting in the crown-bearing region of the thick filament starting with the C-zone crowns instead.

The reduction in thick filament length observed in Ttn^ΔA164–167 mice is predicted to influence both active and passive stress generation. Loss of a single myosin helical repeat (three crowns, nine myosin molecules) represents ∼6% of the 147 myosin molecules comprising a half-thick filament (Craig and Offer, 1976a), suggesting an expected ∼6% decrease in maximal active stress generated by Ttn^ΔA164–167 sarcomeres. However, no reduction was detected in skinned cardiac myocytes or skeletal muscle fibers (Fig. 3 E and Fig. 4 F). This absence of a measurable effect may reflect the inherent variability of maximal active stress measurements, stemming from factors such as variation in sarcomere length during contraction and uncertainty in fiber CSA, which could obscure a change of this magnitude. Alternatively, this could reflect inherently different regulation and/or activation of the Ttn^ΔA164–167 thick filament in compensation for the shortened length (discussed further below).

At the intact muscle level, EDL muscles from Ttn^ΔA164–167 mice showed a ∼30% reduction in maximal active stress (Fig. 5 A), far greater than the ∼6% expected from thick filament shortening. The underlying cause of this disproportionate decrease remains unknown. Note that the ∼30% reduction in CSA of the skinned IIB fibers used for mechanical experiments was consistent in magnitude with the ∼30% decrease in absolute whole-muscle force generation, in line with the possibility of the muscle having less myofibrillar force-generating area. Fiber-type switching toward slower IIA/IIX fibers likely also contributes to the reduction in whole EDL force generation, which would not have been detected in skinned fiber mechanics, which focused solely on IIB fibers. Intact soleus contractile function was also examined and showed a minimal decrease in force generation at high stimulation frequencies, along with a significant upregulation of type I fibers (and concomitant decrease in the content of type IIA and IIX fibers) (Fig. S4). We hypothesize that the differential impact of the titin A164–167 deletion on EDL vs. soleus muscles may stem from the inherent difference in thick filament activation and regulation in fast- vs. slow-twitch muscles that may allow the soleus muscle to better compensate for structural changes than the EDL (Gong et al., 2022). Furthermore, if increased energetic demands due to compensatory mechanisms are a driver of the phenotype, it is possible that the increased mitochondrial content of type I fibers (Leary et al., 2003) allows them to perform better than type II fibers in Ttn^ΔA164–167 muscles.

At the whole heart level, echocardiography analysis revealed no difference in systolic performance between genotypes (Fig. 6 A), implying that the heart may compensate for the expected 6% loss in force due to thick filament shortening. As mentioned above, one possible mechanism for this could be differences in the regulatory state of the myosin heads, which titin is known to influence (Park-Holohan et al., 2021; Squarci et al., 2023). Myosin heads can be positioned in an “ON” (available to form cross-bridges) or “OFF” (folded against the thick filament backbone in the interacting head motif and acting as a reserve) state (Ma et al., 2023; Grinzato et al., 2023; Mann et al., 2020). Perturbations to the regulatory state of the thick filament (e.g., myosin mutations that cause hypercontractility) can trigger compensatory responses, such as cardiac remodeling, as seen in the present study (reviewed in Spudich [2019]). In the case of the Ttn^ΔA164–167 mouse model, we observed organ-level morphological remodeling. Echocardiography revealed LV enlargement, and cardiac chamber weights were significantly increased. The structural changes caused by deletion of domains A164–167 might result in an increase in the ON state of the thick filament, preserving contractile function in the short term. This increased demand likely underlies the enlarged heart observed in the Ttn^ΔA164–167 mouse model. Additionally, greater myosin activity elevates mitochondrial ATP demand, potentially driving mitochondrial remodeling (reviewed in Botella et al. (2023), Lazaropoulos and Elrod [2022], and Abel and Doenst [2011]). Consistent with this, RNA-seq analysis of the Ttn^ΔA164–167 LV showed enrichment for GO terms associated with mitochondrial components, fatty acid metabolism, glycogen metabolism, and glycolysis. Metabolic pathway shifts from oxidative to glycolytic metabolism have been documented in DCM and HCM, as well as changes to ERK/MAPK signaling, which were also detected at the transcriptional level (Muchir et al., 2010; Dávila-Román et al., 2002; Gallo et al., 2019; Schafer et al., 2017; Lumish et al., 2024). Fibrotic remodeling, another common feature of pathological adaptation, is also suggested by enrichment of the collagen-containing extracellular matrix GO term, although this was not detected by cardiac histology at the 3-mo age point.

We also examined the passive stress-generating properties of both the EDL and the heart muscle in response to the deletion of domains A164–167 and the resulting shorter thick filaments. In sarcomeres with shortened thick filaments, titin’s extensible I-band segment is longer at any given sarcomere length, and passive stress is therefore expected to be elevated in Ttn^ΔA164–167 mice. WLC modeling, which incorporates both the shorter thick filament and I-band splicing effects, predicted a large, combined increase in EDL passive stress, ∼100% at a sarcomere length of 3.0 μm (Fig. S5 A). Single-fiber measurements confirmed significant elevations in passive stress but to a lesser extent than predicted (∼30% increase at SL 3.0 μm). This smaller-than-expected increase may reflect deviations from WLC model assumptions, particularly the absence of Ig domain unfolding. Notably, unfolding of a single Ig domain increases its contour length by ∼25 nm (Watanabe et al., 2002a), which could substantially offset the effect of shortening of the thick filament.

In the LV tissue, titin splicing analysis revealed no differences between genotypes, indicating that any change in passive stress arises solely from increased strain on titin’s I-band due to shorter thick filaments. The WLC model predicted increased titin-based passive stress in Ttn^ΔA164–167 cardiac sarcomeres, ∼50% higher than WT in the 2.2–2.3 µm sarcomere length range (the upper physiological limit; Fig. S5 B). However, skinned myocyte measurements showed only small increases in passive stress, significant only when comparing fitted sarcomere length–stress curves between genotypes (Fig. 4, A–D). The smaller-than-predicted effect may reflect Ig domain unfolding in the proximal or distal tandem Ig segments (Minajeva et al., 2001) or variability inherent to skinned myocyte measurements, as discussed above. Furthermore, posttranslational modifications play a significant role in modulating titin stiffness and are frequently imbalanced during cardiac disease states (reviewed in Hidalgo and Granzier (2013) and Loescher et al. (2022)), which adds another layer of complexity that is not accounted for in the WLC model. At the organ level, echocardiography revealed mild diastolic filling impairment, evidenced by a shorter mitral valve deceleration time and increased atrial weight (Fig. 6). We probed the Ttn^ΔA164–167 hearts for fibrosis as a possible explanation for the organ-level diastolic impairments (Conrad et al., 1995) but found no overt fibrosis at 3 mo of age. While diastolic dysfunction can have multiple causes including fibrosis, increased titin-based passive stiffness is a major contributor (Granzier and Irving, 1995; Chung and Granzier, 2011; Lin et al., 2022; Makarenko et al., 2004), and our findings overall are consistent with elevated titin-based passive stress in the Ttn^ΔA164–167 model. Further research probing titin posttranslational modifications will be required to fully understand the mechanism of the mild diastolic filling changes.

Future work aims to better understand the mechanism of both cardiac and skeletal functional changes, with a particular emphasis on understanding how the phenotype of this mouse model may evolve with age. In cardiac muscle, our findings are consistent with the notion that the Ttn^ΔA164–167 heart is undergoing active remodeling in response to increased energetic demands imposed by thick filament structural perturbations. We hypothesize that at 3 mo of age, the heart is in a state of compensation, where contractility is rescued by overactivation of the thick filament. As is the case in many cardiomyopathies, we hypothesize that the compensated state will eventually devolve into heart failure due to an inability to meet the energetic demand required to maintain myosin overactivation and increased size of the myocardium. Note that a limitation of this study is the echocardiographic evaluation of heart function in only male animals.

This study identifies titin’s previously uncharacterized P-zone domains A164–167 as critical determinants of thick filament organization and function. Using the Ttn^ΔA164–167 mouse model, we demonstrated that loss of these domains disrupts the spatial arrangement of titin and cMyBP-C in the thick filament, shortens thick filament length, and eliminates titin’s α and β conformations, resulting in a new titin structural state in the P-zone. These structural rearrangements led to functional consequences in both cardiac and skeletal muscle. We propose that these changes arise from perturbation of the fine-tuned structural relationship among thick filament proteins, compounded by compensatory adaptations. Mechanistically, the data support a model in which deletion of domains A164–167 eliminates a myosin repeat (three crowns, nine molecules) in each half-thick filament, specifically crowns P1, P2, and P3, resulting in P-zone disorganization. At the start of the C-zone, alignment between thick filament components is re-established. The loss of the titin β conformation further implicates domains A164–167, and the β loop in particular, as essential templating or stabilizing factors for thick filament assembly near the M-band. Together, these findings establish titin domains A164–167 as elements critical for the structure and function of the thick filament.

The data underlying Fig. 1, C and D; and Fig. 2, A and C, are openly available in BioProject (Accession ID PRJNA1412858). All other data are available in the published article and its online supplemental material. Raw data are available upon request.

Olaf S. Andersen served as editor.

We thank Teodora Georgieva, PhD, director of the Bio5 ORP GEMM Core Facility at the University of Arizona for the generation of the Ttn^ΔA164–167 mouse model; Douglas Cromey, MS, comanager of the ORP Imaging Core—Optical (RRID: SCR_023355) at the University of Arizona for training and assistance with SR-SIM imaging; and Marloes van den Berg, MD, PhD, research scientist for the Small Animal Phenotyping Core Facility at the University of Arizona for echocardiography analysis. Electron microscopy was performed at the University of Arizona ORP Imaging Cores—Electron (RRID: SCR_023279) (P.T.), mass spectrometry was performed in the University of Arizona Quantitative Proteomics Laboratory (P.R.L.), and cardiac histological processing, sectioning, and staining were performed by the University of Arizona Cancer Center’s Tissue Acquisition and Cellular/Molecular Analysis Core Facility. We are grateful to Samantha Harris, PhD, for the cMyBP-C antibody.

This work was supported by the National Heart, Lung, and Blood Institute (NHLBI) grant 5R35HL144998-07 (H. Granzier), National Institute of Arthritis and Musculoskeletal and Skin Diseases (NIAMS) grant 5R01AR083233-03 (H. Granzier), Interdisciplinary Training in Cardiovascular Research grants T32HL007249-46 and T32HL007249-47 (C. Hoover Browne), and NIAMS grant F31AR085369 (C. Hoover Browne). Open Access funding was provided by the University of Arizona.

Author contributions: Catherine Hoover Browne: conceptualization, data curation, formal analysis, funding acquisition, investigation, methodology, project administration, validation, visualization, and writing—original draft, review, and editing. Seong-won Han: data curation, formal analysis, and writing—review and editing. Gerrie P. Farman: data curation, formal analysis, and writing—original draft, review, and editing. John E. Smith: conceptualization, investigation, methodology, supervision, and writing—review and editing. Justin Kolb: investigation, project administration, and writing—review and editing. Jochen Gohlke: data curation, formal analysis, methodology, resources, software, and writing—review and editing. Paul R. Langlais: data curation, formal analysis, and writing—original draft, review, and editing. Paola Tonino: investigation, methodology, resources, validation, and writing—original draft, review, and editing. Mei Methawasin: investigation and writing—review and editing. Robbert van der Pijl: conceptualization, project administration, supervision, and writing—review and editing. Henk Granzier: conceptualization, funding acquisition, project administration, resources, supervision, and writing—review and editing.