Attenuated lamin A–prohibitin2 interaction leads to mitochondrial dysfunction in LMNA 289 A>G–mediated dilated cardiomyopathy

TL;DR

A mutation in lamin A disrupts its interaction with prohibitin2, leading to mitochondrial dysfunction and contributing to dilated cardiomyopathy.

Contribution

This study is the first to elucidate lamin A's role in cellular bioenergetics and mechanotransduction in dilated cardiomyopathy.

Findings

Reduced lamin A K97E and prohibitin2 interaction causes mitochondrial fragmentation and ATP deficiency.

Impaired RhoA/ERK/FAK signaling disrupts actin assembly and promotes mitochondrial fission.

Mitochondrial dysfunction leads to depolarization, reduced glycolytic capacity, and elevated superoxide levels.

Abstract

Lamins are critical in maintaining nuclear homeostasis, chromosome positioning, and modulating mechanotransduction. Recent studies indicated the involvement of lamin A in mitochondrial homeostasis and the regulation of superoxide. Missense mutations in LMNA are linked to a spectrum of diseases known as laminopathies, which include conditions, such as dilated cardiomyopathy (DCM), muscular dystrophy, and progeria. K97E is one such mutation, which leads to DCM with severe phenotypes. In this study, we established direct reduction of interaction between lamin A K97E and prohibitin 2. As a sequel, mitochondria exhibited reduced fusion, elevated fragmentation, and ATP deficiency. On the other hand, impaired RhoA/extracellular signal–regulated kinase/focal adhesion kinase signaling cascade disrupted filamentous actin assembly, thereby promoting actin–mitochondria association, further facilitating mitochondrial fission. This feedback loop led to mitochondrial depolarization, and global metabolic derangement, in particular, reduced glycolytic capacity and incomplete fatty acid oxidation accompanied by elevated superoxide levels. In cardiomyocytes, such dysfunction may be correlated with contractile defects and arrhythmias. Thus, our findings elucidated for the first time the pivotal role of lamin A in cellular bioenergetics and mechanotransduction, offering novel insights into DCM pathophysiology, which could open newer vistas for developing targeted therapeutic strategies.

Figures

Full text

Lamins are type-V intermediate filament proteins that form the fundamental structural component of the ∼14 nm lamina underlying the inner nuclear membrane. It plays major roles in maintaining nuclear integrity, genome organization, and epigenetic regulations (1). Encoded by the LMNA gene (1q22 in humans), lamin A and its splice variant lamin C contribute toward the dynamic structural framework of the nucleus, ensuring that nuclear shape and rigidity are preserved under various mechanical stresses. Lamin A is not only implicated in the maintenance of nuclear architecture and mechanical properties but also modulates cellular responses to the extracellular cues. The relative abundance of lamin A varies according to tissue type and matrix stiffness through feedback from mechanotransduction signaling (2, 3). More than 500 LMNA mutations have been linked to 14 different human diseases, collectively known as “laminopathies,” including diseases like dilated cardiomyopathy (DCM), Emery–Dreifuss muscular dystrophy, Hutchinson–Gilford progeria syndrome, metabolic-associated fatty liver disease, familial partial lipodystrophy type 2, and mandibuloacral dysplasia (1). Laminopathic cells are characterized by an alteration of nuclear stiffness, fragile/misshapen nuclei, as well as altered chromosomal organization as visualized by 3D-FISH experiments (4, 5, 6). In addition, epigenetic aberrations have also been reported in several studies as one of the primary underlying causes of laminocardiomyopathy disorders (7, 8, 9). Among the laminopathies, DCM is the most prevalent laminopathic cardiac condition reported across the world, with over 200 associated LMNA mutations identified to date (http://www.umd.be/LMNA/). Lamin A-mediated DCM accounts for the pathogenesis of around 8% to 10% of overall DCM patients (10, 11, 12), primarily characterized by left ventricular dilation with impaired systolic functioning of the heart, resulting in sudden cardiac arrest and death in patients (13). Although the precise cellular mechanisms underlying DCM remain largely elusive, it is characterized by a gamut of pathophysiological alterations, which include mitochondrial dysfunction, elevated oxidative stress, cardiomyocyte apoptosis/necrosis, myocardial fibrosis, as well as lipotoxicity (14). Among these, mitochondrial dysfunction emerges as a critical factor, particularly in the context of heart failure, where an inadequate supply of ATP directly affects the contractile function of the heart. Mitochondria serve as a central hub for cellular metabolism, integrating multiple metabolic pathways to supply essential building blocks for macromolecule synthesis and the maintenance of metabolic homeostasis. Specifically, in cardiac tissue, the generation of ATP primarily depends on the fatty acid oxidation (FAO) and the activity of the tricarboxylic acid (TCA) cycle: processes that are indispensable for sustaining the high energy demands of continuous cardiac contraction (15, 16, 17). Mitochondrial homeostasis is crucial for these processes and is maintained by the interplay of fission, fusion, mitophagy, and intracellular transport, which regulate mitochondrial morphology, integrity, abundance, and spatial distribution. In the event of fission, the outer mitochondrial membrane (OMM) constricts because of actin polymerization within the endoplasmic reticulum (ER) contact sites, recruiting dynamin-related protein 1 (Drp1) through adapter proteins, such as mitochondrial fission 1 (Fis1) and mitochondrial fission factor (Mff) (18). This process segregates damaged components, allowing their selective removal via mitophagy, thereby preventing further cellular damage. On the other hand, mitochondrial fusion is a multistep process initiated by the activation of dynein-associated GTPases, including mitofusin (Mfn) 1/2 on the OMM and optic atrophy protein 1 (Opa1) on the inner mitochondrial membrane (IMM) (19). Following GTP hydrolysis–driven OMM and subsequent IMM fusion, mitochondrial contents are intermingled, diluting dysfunctional proteins and mitochondrial DNA to maintain homogeneity and functional stability (20). These processes are tightly regulated by various pathways and are activated according to cellular proliferation, differentiation, and stress. Among their molecular gatekeepers, prohibitins (PHBs) play a pivotal role in maintaining fission–fusion balance and are also involved in mitochondrial DNA maintenance, oxidative phosphorylation (OXPHOS) assembly, cristae organization, mitophagy, and apoptosis (21, 22, 23). By stabilizing Opa1, this complex promotes mitochondrial fusion, whereas its depletion results in mitochondrial fragmentation through improper Opa1 processing (24). In addition, proper assembly of the PHB complex alongside m-AAA protease, Oma1, and cardiolipin maturation is critical for preserving mitochondrial dynamics, which can be adversely affected by alterations in the levels of either PHB1 or PHB2 (25, 26). Interestingly, PHB2 is also transported to the nucleus, where it interacts with various transcription regulators and chromatin remodeling proteins to modulate transcriptional activity, apoptosis, growth, and differentiation in a tissue-specific manner (27, 28, 29), although its interaction with lamin A has not been established. On the other hand, recent studies have demonstrated that depletion of lamin A as well as its mutations disrupts mitochondrial metabolic processes through multiple intermediary mechanisms (30, 31, 32). These findings highlight the importance of nuclear lamins in regulating mitochondrial function and conjecture potential novel avenues for delineating the link between metabolic abnormalities and the pathogenesis of DCM.

In this study, we focused on the LMNA 289 A>G or K97E mutation in lamin A, which has been implicated in severe forms of DCM, reported among several Italian families where affected individuals exhibited pronounced clinical symptoms (33). Our investigation revealed a marked alteration in the lamin A interactome, highlighting PHB2, one of the key mitochondrial scaffold proteins. This led to mitochondrial dysfunction, manifesting as a deviation of cellular metabolism and bioenergetics. Specifically, we observed significant reductions in intracellular ATP levels, mitochondrial depolarization, and elevated production of reactive oxygen species (ROS). We validated the lamin A–PHB2 interaction and established its functional relevance in the pathogenesis of laminocardiomyopathy with reference to mitochondrial dysfunction. In parallel, we identified disruptions in actin cytoskeletal organization, characterized by diminished actin polymerization and altered stress fiber morphology, driven by dysregulation of the Rho/extracellular signal–regulated kinase (ERK)/focal adhesion kinase (FAK) signaling axis. These cytoskeletal defects further reinforced mitochondrial and mechanotransduction dysfunction in the pathogenic cascade. Collectively, our findings demonstrate that the K97E lamin A mutation leads to interconnected mitochondrial and cytoskeletal abnormalities, establishing a mechanistic link between bioenergetic collapse and impaired mechanotransduction in the progression of DCM.

Results

Identification of PHB2–lamin A interaction and its attenuation in the K97E mutation

K97E maps in the rod 1B domain (Fig. 1A) of lamin A and is one of the most deleterious mutations of LMNA, causing DCM (33) with conduction defects. In retrospect, this mutant exhibited an increased propensity for self-association and a gradual disappearance from the lamina (34, 35). On logical grounds, we first questioned whether this enhanced self-association impacts the lamin A interactome. To address it, we employed proximity-based BioID2 labeling followed by mass spectrometric (MS) analyses (Fig. 1B), where the BioID2 tag is capable of biotinylating interacting proteins within a range of 10 nm. It is important to note that BioID has been acknowledged to be one of the key tools for studying the interactome of WT lamin A, as well as complex multimeric protein structures, such as the nuclear pore complex, and has promising applications in studying protein interactions in crowded environments (36, 37). The WT and mutant lamin A-BioID2 plasmids were expressed in the human embryonic kidney 293T (HEK293T) cell line, expressing a minimal amount of endogenous lamin A proteins, and the cellular localization was confirmed through immunofluorescence (Fig. 1, C and D). Along with the event of colocalization of lamin A and myc-BioID2-lamin A variants (Fig. 1E), it was confirmed that the tag was not affecting the physiological properties of lamin A, which showed similar distribution as reported in our earlier studies (1). An equal amount of lysate was used for streptavidin pull-down (Fig. 1F), followed by MS identification of the peptides, which revealed a gross reduction in the number of interacting partners in the case of K97E by half of the value compared with the WT (Fig. 1G). Among these, the mutant protein K97E exhibited a gain in interaction concerning 19 proteins and a loss of 100 proteins (Fig. 1H). Upon enriching the set of proteins losing their interaction with mutant lamin A, we uncovered that the most prominent one was the PHB2 complex (Fig. 1I). Furthermore, we confirmed the loss of interaction of PHB2 with lamin A through BioID2-immunoblot (Fig. 1J) as well as coimmunoprecipitation (Fig. S1). We chose C2C12 cells for all studies, as it has remained a faithful and established model system for studying cardiomyopathies (38). From immunoblot (IB) analysis, we observed a downregulation of PHB2 but not PHB1 in cells expressing the mutant in both mRNA (Fig. 1K) and protein levels (Fig. 1, L and M). Confocal images also revealed the same for PHB2, especially in terms of nuclear localization (Fig. 1, N and O). To further validate these observations, we performed nuclear fractionation and found that indeed PHB2 levels were reduced to minuscule amounts in the nuclear extract of K97E-expressing cells (Fig. 1P), pointing to a plausible perturbation of the interaction between K97E and PHB2. Theoretically, such a reduction in the amount of PHB2 in the nucleus should be compensated by an increase in the cytoplasmic fraction. To this end, the cytoplasmic extract of K97E-expressing cells was checked for PHB2 expression, which showed a modest increase (Fig. 1Q). We asked ourselves whether it is a perturbation in the interaction between the mutant lamin A protein with PHB2, which in turn could retain PHB2 in the nucleus with much lower efficacy compared with the WT lamin A scenario. To check this hypothesis, we performed molecular docking studies using HADDOCK. For this, we utilized homology-modeled full-length PHB2 and lamin A proteins based on multiple Protein Data Bank IDs, including 8J4I and 6JLB, for PHB2 (domain 1–190) and lamin A (domain 1–300), respectively (Fig. S2). The energy parameters are shown in Table S1, which points to a distinctly weaker interaction of K97E with PHB2 (Fig. 1R). To confirm the simulation data, we investigated any possible direct interaction between lamin A and PHB2 by biochemical methods. In order to do so, first we performed blot overlay using purified lamin A proteins (Fig. S3, A and B) and whole cell lysate (Fig. 1S) and observed that PHB2 shows affinity to bind with WT lamin A, and the interaction is drastically perturbed by the introduction of the K97E charge shift. As the technique relies on denatured proteins, thereby altering native protein–protein interaction parameters, we subsequently employed an in-blot renaturation strategy using a guanidine hydrochloride gradient, followed by far Western blotting to assess protein interactions, and observed similar results altogether (Fig. 1T), thus establishing for the first time that lamin A directly interacts with PHB2.Figure 1**Loss of interaction with PHB2 complex in lamin A mutant K97E.**A, schematic representation of the K97E mutation in the 1B domain of lamin A (not to scale). B, schematic overview of proximity-dependent biotinylation and streptavidin pulldown (BioID2) strategy. C, Z-projected immunofluorescence images showing localization of myc-BioID2-lamin A variants (the scale bar represents 2 μm). D, intensity profiles across the nuclear axis depicting myc-BioID2-lamin A and overall lamin A distribution. E, Pearson's correlation coefficient analysis showing colocalization between myc and lamin A signals. F, immunoblot of lamin A/C and GAPDH showing equal BioID2 input loading and comparable expression of BioID2-lamin A variants in HEK293T cells. G, number of interacting partners of WT and K97E lamin A as observed from protein identification through BioID2-pulldown followed by MS. H, Venn diagram indicating independent and overlapping interactome of WT and K97E lamin A. I, enrichment analysis of protein–protein interactions lost in the K97E mutant compared with WT lamin A. J, BioID2-based immunoblot analysis showing reduced interaction between lamin A and PHB2 in the K97E mutant compared with WT. K, quantitative PCR (qPCR) analysis of PHB1 and PHB2 mRNA levels, revealing differential expression in response to the K97E mutation in C2C12 cells. L, immunoblot analysis of PHB1, PHB2, GAPDH, and lamin A in C2C12 cells. M, densitometric quantification of PHB1 and PHB2 immunoblots, normalized to GAPDH. N, Z-projected immunofluorescence images of PHB2 showing altered subcellular distribution upon K97E lamin A expression (the scale bar represents 4 μm). O, quantitative analysis of PHB2 fluorescence intensity from immunofluorescence imaging. P, subcellular fractionation of cells expressing WT and K97E lamin A variants, followed by immunoblot against GAPDH (cytosolic marker), lamin A (nuclear marker), and PHB2. Q, densitometric quantification of PHB2 levels in the cytoplasmic fraction, normalized to GAPDH. R, in silico molecular docking of PHB2 with lamin A variants highlighting lamin A LYS97 and PHB2 GLU231 as potential perturbed interaction. S, blot overlay of purified lamin A proteins overlayed with whole cell lysate and probed with PHB2 to detect interaction. T, far Western of purified lamin A proteins with cell lysate and probed with PHB2 to detect interaction. HEK293T, human embryonic kidney 293T cell line; PHB2, prohibitin 2.

K97E lamin A alters mitochondrial dynamics via PHB2 and actin cytoskeleton remodeling

It is also established that PHB2—the IMM protein—interacts with the autophagosomal membrane–associated LC3 protein through an LC3-interacting region upon mitochondrial depolarization. So we focused, at this point, on mitochondrial homeostasis and how it might be adversely affected by PHB2. Abnormalities in mitochondrial dynamics are major hallmarks of DCM, leading to long-term ailment of tissues/organs having large energy demands, such as muscles and specifically the heart (39). The mitochondrial fission–fusion dynamics are tightly regulated by several pathways involving PHBs, which act as central proteins to regulate this balance (Fig. 2A). Taking cues from the previous result, we followed the trail to investigate whether the alteration of the PHB2 axis would affect the morphology and dynamics of the mitochondria adversely. Using confocal and structured illumination microscopy, we observed reduced mitochondrial length in the case of K97E compared with the WT (Fig. 2, B and C). It is worthwhile to mention that mitochondrial fragmentation is one of the key signs of overall cellular stress (40, 41, 42). We performed experiments to precisely calculate the mitochondrial fission–fusion rate to account for the mechanism of length shortening through time-lapse imaging of mitochondria tagged with MitoTracker, followed by analysis using Mitometer (43). We observed an increased rate of fission with a concomitant decrease in the fusion rate in the case of the mutant protein (Fig. 2, D and E), indicating that the fission–fusion balance was perturbed. Due to this reduced size of the mitochondria, we also observed a sharp increase in mitochondrial dynamics through an elevated mitochondrial velocity (Fig. 2F).Figure 2**PHB2-mediated mitochondrial anomalies and role of actin destabilization.**A, schematic representation of the mitochondrial fission–fusion cycle highlighting the role of PHB2 (not to scale). B, Z-projected confocal images of mitochondrial morphology visualized with MLS-RFP (the scale bar represents 5 μm); super-resolution microscopy (SIM) images in the fourth panel reveal distinct mitochondrial morphology (the scale bar represents 3 μm). C, quantified mitochondrial length from SIM images. D, quantified mitochondrial fusion odds via time-lapse imaging of CMXRos-tagged mitochondria (n = 998). E, quantified mitochondrial fission odds (n = 998). F, quantified mitochondrial velocity (n = 998). G, qPCR analysis of Drp1 and Opa1 mRNA levels, revealing differential expression in response to the K97E mutation. H, immunoblot analysis of Opa1, Drp1, GAPDH, and INF2. I, densitometric quantification of Drp1 and Opa1 immunoblots, normalized to GAPDH. J, live-cell confocal fluorescence imaging showing colocalization of RFP-LifeAct–tagged actin filaments with MitoTracker deep red (MTDR)–labeled mitochondria; arrows indicate regions of colocalization (the scale represents 1 μm). K, Pearson's correlation coefficient analysis showing colocalization between MTDR and RFP-LifeAct signals. L, densitometric quantification of INF2 immunoblots, normalized to GAPDH. G, qPCR analysis of Drp1 and Opa1 mRNA levels, revealing differential expression in response to the K97E mutation. M, immunoblot analysis of pDrp1(p616) and GAPDH. N, densitometric quantification of pDrp1(p616), normalized to GAPDH. O, Z-projected confocal images showing morphology and architecture of phalloidin-tagged actin (the scale bar represents 10 μm). P, Z-projected SIM images of phalloidin-labeled actin filaments; dotted circles indicate regions of interest (ROIs) used for radial profile quantification (the scale bar represents 5 μm). Q, radial distribution profile of actin filaments quantified from a defined ROI. R, representation of the quantification scheme for radial profile and actin filament angles. S, quantified distribution of actin filament angles; inset graph depicts mean actin angles. T, time-lapse imaging of cells expressing lamin A variants and RFP-LifeAct; the first three panels represent Z-projected images for overall actin distribution (the scale bar represents 10 μm), whereas the fourth panel depicts the confocal plane time spectrum of actin dynamics within the marked ROI (the scale bar represents 5 μm). MLS, mitochondrial localization signal; PHB2, prohibitin 2; qPCR, quantitative PCR; RFP, red fluorescent protein.

To validate the claim that the fission–fusion balance was indeed perturbed, we checked the expression profile of two major fusion and fission proteins, Opa1 and Drp1, respectively. Quantitative PCR (qPCR) analysis revealed a reduced Opa1 expression but no significant increase of Drp1 (Fig. 2G). Subsequently, we quantified the corresponding protein levels through IB, which also elucidated a significant reduction in Opa1 levels, whereas Drp1 remained much the same (Fig. 2, H and I). Furthermore, it is interesting to note that mitochondrial fusion protein Opa1, through a concerted post-translational processing and functionalization mediated by the PHB complex, helps in mitochondrial fusion through inner membrane fusion and is closely associated with the pathogenesis of DCM (44). Hence, its downregulation and corresponding reduction in mitochondrial fusion are justified by PHB2 depletion as reported earlier. Although it was previously reported that irrespective of Drp1 levels in cells, downregulation of Opa1 is enough to reduce mitochondrial length (45, 46), which does not completely explain the increased mitochondrial fission rate with sufficient clarity in this scenario. Plagued by this apparent anomaly at this point, we turned to another important player in fission—mitochondrial-associated actins (MAAs), which not only regulates mitochondrial transportation for proper mitochondrial distribution and aids in fusion but also plays a vital role in mitochondrial fission through the formation of an actin “ring” in ER-associated mitochondrial fission sites through INF2 (47). Upon checking the distribution of MAA, we found out that indeed, the cells harboring the laminopathic mutant showed an increased association of mitochondria and actin inside cells (Fig. 2, J and K), thus explaining the conundrum. Interestingly, INF2, one of the actin regulators that, in tandem with the ER, mediates actin ring formation during mitochondrial fission (47), was also overexpressed in K97E (Fig. 2L), thereby strengthening the probability of elevated fission events. As the MAAs are reported to recruit Drp1 to the site of fission, where it gets phosphorylated (48), we checked for the levels of activated Drp1(p616) and observed an elevated activated pDrp1 in the case of the K97E mutation (Fig. 2, M and N), thus indicating that MAA recruited pDrp1 being one of the contributors of enhanced mitochondrial fission.

This led us to investigate the actin and mitochondrial dynamics. We surmised that actin dynamics and/or polymerization were modulating the mitochondrial fission–fusion equilibrium as a sequel to the mutation K97E. We observed through phalloidin staining that cells harboring the lamin A K97E displayed disrupted filamentous-actin (F-actin) architecture as seen (Fig. 2O). To obtain a clear perspective of the events, we employed structured illumination microscopy to scrutinize the F-actin filaments (Fig. 2P) and observed that the actin filaments bundled more prominently in the cortex of the cell while displaying discontinuous/broken filaments throughout the cytoplasm as inferred from the radial profile (Fig. 2Q). The angle that the axis of the actin filaments subtends to the major axis of the K97E-expressing cell (Fig. 2R) exhibited an anisotropic distribution with higher values compared with the WT scenario (Fig. 2S), thereby implying a possible perturbation in cellular mechanical tension. Time-lapse imaging using red fluorescent protein (RFP)-LifeAct showed that the actin filaments were highly dynamic or less stable in the mutant-containing cells (Fig. 2T). In summary, the destabilized actin filaments could be a potential modulator of mitochondrial homeostasis. A lowered actin stress filament with increased dynamic of F-actin arises from intense cellular stress and is an indication of destabilized actin, which has been reported to directly affect mitochondrial morphology and functioning (49, 50, 51). Although it is now clear that the destabilized actin stress fibers were indeed affecting the mitochondrial dynamics in K97E-expressing cells, a clear link between the lamin A mutation and actin destabilization as a sequel to that remained to be explained.

Mechanistic insights into actin cytoskeletal dysregulation induced by K97E mutation

To assess the origin of severed actin filaments in the mutant-expressing cells, we conducted F/G actin fractionation (Fig. 3A) and found that the F-actin/globular actin (G-actin) ratio, or the polymerized-actin level, indeed decreased in the mutant scenario (Fig. 3B). To further elucidate the mechanism behind the decreased F-actins, we examined the proteins responsible for depolymerizing actin filaments, notably cofilin. At first, we observed a straight downregulation of cofilin through checking its gene expression levels (Fig. 3C), which was counterintuitive. Thus, we took a step back in its respective pathway to check the levels of ERK1/2, a previously known lamin A-regulated actin destabilizer, which activates both cofilin and Arp2/3 complex (52, 53) and mediates actin depolymerization through cofilin activation. Upon checking gene expression of ERK1, we observed a lowered mRNA level in K97E (Fig. 3D). Furthermore, we checked the expression of activated pERK1/2 and found that the amount of pERK1/2 was also lowered in K97E-containing cells (Fig. 3, E and F). Upon imaging of pERK1/2 (Fig. 3G), we observed an expected lowered overall intensity with the majority of nuclear pERK sequestered in K97E aggregates (Fig. 3, H and I). This phenomenon might preclude ERK1/2 from activating cofilin and Arp2/3. Thus, combining reduced activated ERK1/2 levels in cells with cofilin downregulation indicates that depolymerization of actin filaments might be lowered in K97E and might not be the cause of reduced F-actins. But then the question remained: why is there more G-actin present in the mutant lamin A condition?Figure 3**Actin destabilization through the Rho-ERK–FAK axis also leads to defects in cellular mechanotransduction.**A, actin fractionation followed by immunoblot to quantify globular (G) and filamentous (F) actins. B, quantification of F/G actin ratio from actin fractionation. C, qPCR analysis to check the mRNA levels of cofilin1. D, qPCR analysis to check the mRNA levels of ERK1. E, immunoblot against phospho-ERK1/2 along with GAPDH. F, densitometric quantification of pERK1/2 from immunoblots, normalized to GAPDH. G, Z-projected immunofluorescence images showing localization of pERK1/2 (the scale bar represents 5 μm). H, quantitative analysis of pERK1/2 fluorescence intensity from immunofluorescence imaging. I, Pearson's correlation coefficient analysis showing colocalization between pERK1/2 and lamin A variant signals. J, qPCR analysis of Arp2/3 and profilin mRNA levels, revealing differential expression in response to the K97E mutation. K, immunoblot against Arp2 with respect to GAPDH. L, densitometric quantification of Arp2 from immunoblots, normalized to GAPDH. M, quantitative analysis of Arp2 fluorescence intensity from immunofluorescence imaging. N, Z-projected immunofluorescence images of Arp2 showing subcellular distribution upon K97E lamin A expression (the scale bar represents 5 μm). O, qPCR analysis of RhoA and ROCK1 mRNA levels, revealing differential expressions in response to the K97E mutation. P, immunoblot against RhoA and ROCK1. Q, densitometric quantification of RhoA from immunoblots, normalized to GAPDH. R, qPCR analysis of mRNA levels of various actin polymerization–promoting proteins, including formins, revealing differential expression in response to the K97E mutation. S, schematic representation of focal adhesion, highlighting its role in the actin polymerization cycle (not to scale). T, immunoblot against FAK and phosphopaxillin with respect to GAPDH. Q, densitometric quantification of FAK and paxillin from immunoblots, normalized to GAPDH. V, quantification of vinculin puncta area marking focal adhesion complex (n = 100). W, representative images from traction force microscopy showing force distribution in cells. X, quantified constrained traction forces exerted by cells. H and E of this figure share the same loading control (GAPDH), as protein samples were run in duplicates on the same gel. ERK, extracellular signal–regulated kinase; FAK, focal adhesion kinase; qPCR, quantitative PCR.

Investigating further, we quantified the expression of Arp2 and profilin. Arp2 is an actin nucleator, which localizes throughout the cell at actin branching sites, whereas profilin is a G-actin–binding protein, which helps in initiating actin polymerization through ATP hydrolysis (54). Upon examining the gene expression of Arp2, we noted a slight downregulation, which was also reflected in protein expression levels in the case of K97E (Fig. 3, J–L), which, along with downregulated pERK1/2, suggests that the branching of actin filaments may not be influencing actin dynamics significantly. This was validated by confocal images showing similar downregulation, particularly inside the nucleus (Fig. 3, M and N). Conversely, profilin exhibited a significant upregulation of gene expression (Fig. 3J), which, in tandem with the previous observations, accounted for a larger pool of G-actin present in K97E-expressing cells compared with WTs. Subsequently, we checked the expression of RhoA and ROCK1, two major regulatory entities of actin polymerization as well as a mitochondrial health determinant (55, 56). Gene expression supplemented by protein expression profile revealed a downregulation of RhoA (Fig. 3, O–Q), which is crucial for the activation of ROCK1 and, consequently, actin polymerization through formins. Following RhoA downregulation, we checked the expression of several groups of formins and noticed that FHOD3, formin2, and DIAPH2 were downregulated in K97E (Fig. 3R). Therefore, we can summarize that it may be due to reduced actin polymerization and not increased depolymerization, which led to an increased G-actin pool in the cells.

Both RhoA and ERK pathways are interconnected with a third entity—FAK, which is also affected by lamins and formins like DIAPHs and FHODs (57). FAK is an integral part of the focal adhesion complex and interacts with both ERKs and RhoA in mechanotransduction (58, 59, 60), where it phosphorylates paxillin to initiate the signaling cascade (61, 62, 63). Along the same line, FAK also partakes in actin polymerization through feedback from mechanical signals received from focal adhesion (64) (Fig. 3S). While checking the expression of FAK and p-paxillin, we found out that both the proteins were severely depleted in K97E-induced cells (Fig. 3, T and U), which was reported earlier to cause gross actin destabilization and mechanotransduction defects (65). Depletion of FAK also indicates lowered focal adhesion stability, which, when checked through staining vinculin, a marker for focal adhesion strength and turnover (Fig. S4), elucidated a lowered vinculin area per focal adhesion (Fig. 3V), amounting to weakened focal adhesions in K97E-containing cells. Along with its role in mitochondrial homeostasis, actin plays a primary role in cellular mechanotransduction. While it transmits the mechanical signals from focal adhesions to the nucleus via the linker of the nucleoskeleton and cytoskeleton complex (57, 61), it is also responsible for maintaining cellular stiffness and the generation of traction forces. Following observation on gross defects in actins and focal adhesion, we hypothesized mechanotransduction defects via alteration of traction forces exerted by the cell. Thus, to confirm that we employed traction force microscopy (TFM) and observed that the K97E cells indeed generated ∼50% lower traction forces compared with WT cells (Fig. 3, W and X).

In summary, we concluded that actin destabilization was caused by a lack of proper actin polymerization through downregulation of the Rho–ERK–FAK axis in the cell, which not only contributed toward increased mitochondrial fission but also a defective mechanotransduction pathway. Furthermore, it must be mentioned that PHB2 is one of the proteins that also helps in the stabilization and activation of FAK through the Akt pathway, along with its interaction with ERK and RhoA to maintain muscle homeostasis (66). Thus, through PHB2, we can correlate both actin and mitochondrial defects in cells having the K97E lamin A mutation.

K97E mutation disrupts mitochondrial redox homeostasis and metabolic output

After concluding from our previous results that mitochondrial destabilization was mediated largely by PHB2 expression and perturbations in actin stress fibers, we sought to investigate its functional implications. By and large, mitochondrial homeostasis correlates with its membrane potential (Ψ_m_), which is crucial for the functioning of mitochondrial enzymes and subsequent OXPHOS pathways (67). The increased mitochondrial fragmentation often points toward unhealthy mitochondria, characterized by mitophagy and reduction of Ψ_m_ (40). To check the mitochondrial membrane potential, we used an indirect analysis by measuring CMXRos (68), which can only effectively stain the healthy functional mitochondria, having a differential membrane potential, leaving the damaged mitochondria unstained. On the other hand, mitochondrial localization signal (MLS)-RFP is a direct readout for both healthy and unhealthy mitochondria. Therefore, comparison of the mitochondrial footprints from MLS-RFP and those of CMXRos yielded an approximate measure of unhealthy mitochondria. We observed a significant reduction of Ψ_m_ of mitochondria in K97E-expressing cells (Fig. 4, A and B). We also noticed that the mitochondrial length was further shortened while stained with CMXRos (Fig. 4C), compared with MLS-RFP, in K97E. Thus, it can be inferred that both the mitochondrial length and Ψ_m_ decreased in the case of the mutant pointing toward impaired metabolic quotient.Figure 4**Altered metabolic profile leads to decreased ATP levels and elevated ROS.**A, confocal imaging of mitochondrial health visualized with CMXRos (the scale represents 10 μm). B, quantitative analysis of CMXRos fluorescence intensity from confocal images. C, quantified mitochondrial length from CMXRos-labeled mitochondria. D, quantified ATP concentration of cells expressing lamin A variants. E, schematic representation of the ATP dependency assay using 2-deoxyglucose (2-DG) and oligomycin (OA). F, quantified mitochondrial versus glycolytic dependency. G, quantified glucose versus beta oxidation, that is, fatty acid oxidation (FAO) and amino acid oxidation (AAO) dependency. H, unsupervised principal component analysis (PCA) of the K97E and WT metabolome. I–K, relative abundance of cholesterol (I), phosphorylethanolamine (J), and inosine (K) obtained from metabolome analysis through GC–MS. L, metabolic pathway enrichment conducted using significantly differentially abundant metabolites. M, quantified ratio of selected metabolites from central carbon and fatty acid metabolism pathways. N, quantification of mitochondrial superoxide using MitoSOX (n = 10). O, quantification of cellular ROS using CellROX (n = 10). ROS, reactive oxygen species.

To check the bioenergetic footprint of the K97E-containing cells, we measured the amount of ATP in the cells. A drastic reduction in ATP content was observed in cells expressing the lamin A K97E mutant (Fig. 4D). At this point, it is important to differentiate between the contribution of mitochondrial and nonmitochondrial biochemical pathways leading to ATP depletion. To that end, we used selective pathway inhibitors like oligomycin (OA) and 2-deoxyglucose (2-DG) as described in the protocol and depicted in Figures 4E and S5 (69). Results showed that concomitant with an overall decrease in ATP, there was a significant (∼3-fold) decrease in mitochondrial contribution to ATP production upon K97E mutation (Fig. 4F). Further, while dependency on glucose oxidation decreased significantly, the normalized contribution of FAO and amino acid oxidation (AAO) was increased in the mutant (Fig. 4G). In order to gain further insight into the metabolic alterations associated with expression of the K97E mutant, untargeted analysis of the cellular metabolome was performed. The unsupervised principal components analysis showed an inherent tendency of segregation of the K97E and WT metabolome (Fig. 4H). The supervised orthogonal projection to latent structure-discriminant analysis (orthogonal partial least squares-discriminant analysis) of the metabolomic signature showed clear separation of the WT and K97E metabolome (Fig. S6A). The permutation test (100 permutations) showed a p value <0.05, indicating goodness of model fit and predictive accuracy (Fig. S6B). Among metabolites that contributed to the separation of the metabolome, cholesterol (Fig. 4I) and phosphorylethanolamine (Fig. 4J) were found to be notably depleted, whereas inosine (Fig. 4K) was found to be elevated in the K97E mutant even after false discovery rate (FDR) correction. Metabolic pathway analysis was conducted using differentially abundant metabolites identified by volcano plot analysis (Fig. 4L). It indicated a significant alteration in multiple pathways as listed in Table S1. Particularly, alterations in the TCA cycle, valine, leucine, and isoleucine metabolism, and alanine, aspartate, and glutamate metabolism were significant even after FDR correction. In addition, the pantothenate and CoA biosynthesis pathway, which is essential for the entry of glucose-derived pyruvate into the TCA cycle as well as the metabolism of fatty acids and branched-chain amino acids, was also altered. Mitochondria play an important role in the metabolism of these amino acids and fatty acids via the TCA cycle. Taken together, these pointed to a coordinate derangement of mitochondrial metabolism, as was indicated by the energetics assay. In a biochemical pathway, the ratio of two interconnected metabolites has been shown to offer a surrogate signature for biochemical flux and the relative enzymatic activities through the steps in between (70, 71, 72, 73, 74). Hence, we explored the ratio of metabolites involved in energy-producing metabolic pathways to get a more detailed picture of the effect on the energy metabolism. As shown in Figure 4M, there was a decrease in citrate/pyruvate, succinate/pyruvate, malate/pyruvate, succinate/2-ketoglutarate, and succinate/citrate ratios in the mutant, indicating a lower TCA cycle activity. This was also associated with an increase in the pyruvate–glucose ratio in the mutant, plausibly because of a lack of glucose-derived pyruvate utilization along with an increase in glycolytic dependence (Fig. 4F). Therefore, an impaired glucose metabolism toward ATP production was also indicated by a decrease in glucose dependency in the mutant. On the other hand, there was an increase in the gluconic acid–glucose ratio (Fig. S6C) in the mutant, indicating oxidative stress. Interestingly, the ratio of palmitate to citrate, as well as myristate or stearate to citrate, was significantly lower in the mutant, which may be a result of elevated FAO, as was indicated by the ATP assay, and/or a decrease in de novo fatty acid synthesis because of low ATP availability. This was also reflected in the lower AMP/adenine ratio along with a lower cholesterol-to-citrate ratio (Fig. S6C), further indicating a shift toward the FAO over de novo fatty acid synthesis. To validate this further, we measured cellular and mitochondrial ROS levels and observed an increase in both mitochondrial superoxide (Fig. 4N) and total cellular ROS (Fig. 4O) to be associated with the metabolic deviation in the K97E-expressing cells.

Discussion

Laminocardiomyopathy, driven by mutations in LMNA, presents clinically with left ventricular dilation, systolic dysfunction, conduction block, and life-threatening arrhythmias, often culminating in myocardial infarction (1, 33). Mutant lamin A/C compromises nuclear envelope integrity and mechanotransduction, perturbing gene expression programs vital for cardiomyocyte homeostasis (1, 34, 35). Concurrent shifts toward excessive FAO exacerbate oxidative stress management and metabolic efficiency (75, 76). Furthermore, our previous study about downregulated interleukin-17–NF-κB signaling points toward reduced inflammation compromising antioxidant defenses while perturbing various metabolic processes (1). Although the mechanistic pathways by which lamin A orchestrates mitochondrial dynamics, OXPHOS flux, and cellular bioenergetic circuitry toward pathogenesis of DCM remain largely unexplored, hence their elucidation is critical.

Our proximity-based BioID2 identified that lamin A interacts with major enzymes like PKM, ENO1, PGK1, TPI1, ALDOA, TKT, TALDO1, MDH2, ATP5F1A, ATP5F1B, SLC25A5, and others, enriching toward ATP synthesis, the pentose phosphate pathway, glycolysis, and other carbon metabolism pathways (Fig. S7, A and B), whose presence in the nucleus is now well reported and is essential for nuclear bioenergetics and epigenetic remodeling (77, 78, 79, 80). These interactions highlight a previously underappreciated role of lamin A in nuclear metabolic regulation, beyond its established structural functions, although the functional relevance of these interactions remains to be elucidated. Our recent study demonstrated the phenomenon of liquid–liquid phase separation in the condensates formed by K97E inside the nucleus (81). These result from increased self-association, which in turn could reduce the interaction with other nuclear factors emphatically, as observed in our BioID2-MS dataset, enriching toward the PHB2 complex. We validated the results using molecular docking, followed by blot overlay, far Western and immunoprecipitation and established, for the first time, the direct interaction between lamin A and PHB2. Furthermore, we surmised that the K97E–PHB2 interaction has been weakened, which was again validated by orthogonal methods, thereby depleting the lamin A–PHB2 complex in the nucleus. PHB2 is not only limited to its role in mitochondrial homeostasis but also plays a key role in several cellular processes that are often overseen (66, 82). Interestingly, PHB2 was reported in the pathogenesis of DCM, where its depletion led to a distinct shift in the OXPHOS pathway and reduced ATP levels, leading to heart failure (23). This is in association with PHB2's interaction with DNAJC19, a mitochondrial chaperone, whose alteration is directly correlated to the pathogenesis of lamin A–independent DCM with ataxia (83). In this study, mitochondria in K97E-expressing cells displayed significant fragmentation as seen through shortened branch lengths and increased motility arising from an increase in fission events accompanied by a decrease in fusion frequency. Furthermore, we elucidated that PHB2-mediated downregulation of Opa1 led to decreased mitochondrial fusion. It has also been reported that mitochondrial fragmentation because of improper Opa1 processing, in which PHB2 also plays a role (22), might lead to heart failure in mice (84). In addition, elevated E2F/TP53 DNA damage response activity during lamin A mutation has also been reported to destabilize Opa1, leading to the development of DCM (44, 85). On the other hand, γH2AX levels in K97E-expressing nuclei revealed a marked increase in nuclear pH2AX foci (Fig. S8, A–C), indicative of elevated DNA damage. This genomic instability may arise from increased ROS levels or from impaired DNA repair capacity because of the loss of functional lamin A (86). Such DNA damage may further contribute to the destabilization of Opa1, as previously implicated through E2F/TP53-mediated pathways. Furthermore, increased mitochondrial fission axis could also be explained through profound actin cytoskeletal disruption in K97E-expressing cells, leading to elevated MAA levels and elevated phospho-Drp1. We observed disorganized actin structures, consistent with impaired Rho–ERK–FAK signaling and reduced formin-mediated polymerization. Instead, the surplus G-actin plausibly fuels the mitochondrial “actin cloud,” which enhances ER–mitochondrial contacts and recruits Drp1 to promote mitochondrial fission (48). While global actin organization is disrupted, there is no direct evidence that short, highly mobile MAA filaments, which are required for mitochondrial fission, are abolished. Taken together, it can be conjectured that local actin clouds can be a principle but not the sole-determining factor for elevated mitochondrial fission. In addition, actin-independent or minimally actin-dependent mechanisms, including Drp1 sensitization driven by altered bioenergetics, reduced ATP availability, elevated ROS, and changes in mitochondrial membrane properties, may further contribute to enhanced fission under these conditions (87). Here, we can also appreciate that the nuclear exclusion of PHB2 could also impair transcriptional programs vital for cardiomyocyte survival, precipitating the disease. As PHB2 is a transcription factor for both ERK1 and ERK2 (88), its nuclear exclusion might trigger the downregulation of the mitogen-activated protein kinase pathway in K97E lamin A mutation. Along with lowered ATP availability and elevated ROS in tandem, further strengthened the appearance of severed actins, leading to a destabilized actin fraction down the cascade. In either case, the associated actins lose their integrity and mediate abnormalities in mechanotransduction signals. In our case, we observed altered traction forces exerted by cells expressing K97E mutant. As muscle myoblasts heavily rely on ATP and actin stress fibers for optimal cell differentiation (89, 90), one can expect an alteration in this vital process. It must be emphasized that studies conducted over the past decade already showed that pathogenesis in cardiomyopathy is frequently linked to underlying metabolic abnormalities (91, 92). Our untargeted metabolomics uncovered derangement of several metabolic pathways, including glucose, fatty acid, and amino acid metabolism, in the K97E-expressing cells. Bioenergetic assay indicated a relative elevation in FAO and/or AAO-derived ATP, possibly to compensate for the decrease in glucose-derived ATP. Taken together with reduced fatty acid–citrate ratios, these indicated an elevation in FAO, which was found to be positively associated with ROS level in earlier studies (93, 94, 95). Several other metabolites that were found to change included inosine, which was elevated in K97E-expressing cells. Inosine was shown to increase mitochondrial biogenesis and mitophagy as well as mitochondrial respiration through OXPHOS by eliciting TCA cycle enzymes, which could be an adaptive response for survival (96, 97). Similar to our study, LMNA knockout mice mimicking DCM are also reported to show impaired beta-oxidation through accumulation of medium- and long-chain acylcarnitines along with depleted high-energy nucleotides, such as ATP, ADP, and AMP (76). We conjecture that the reduced ATP along with increased ROS could possibly arise from reduced mitochondrial membrane potential and/or elevated FAO, which needs further detailed investigations.

Lamin A-mediated familial partial lipodystrophy, metabolic-associated fatty liver disease, lipid metabolism abnormalities in mandibuloacral dysplasia (75), and regulation of the mechanistic target of rapamycin signaling cascade (98) have been repeatedly linked with metabolism defects. The most notable phenotype suggestive of our mechanistic framework came from a very recent patient study who showed lamin A variant G105L, causing similar trends of reduced mitochondrial health, biomass, oxygen consumption rate, reduced ATP production along with elevated ROS (99). Our work uncovers a previously unrecognized pathway by which lamin A mutations perturb PHB2 and cytoskeletal remodeling–mediated mitochondrial dysfunction that converge on bioenergetic collapse and genomic stress. By integrating nuclear architecture, cytoskeletal dynamics, and mitochondrial physiology, our study provides a framework for a better understanding of laminopathies and subsequent development of novel therapeutic strategies.

Experimental procedures

Subcloning in BioID2 plasmid

Lamin A WT and lamin A K97E were cloned as reported earlier (34) in the pEGFP vector (Clontech). EGFP–lamin A variants were used as templates for PCR. The purified PCR was digested with EcoRI and BamHI and subcloned into the BioID2-myc plasmid, received as a gift from Dr Kyle Roux, in which we subcloned WT and K97E lamin A using the cloning primers:

Sense: CCGGAATTCCGGATGGAGACCCCGTCCCAG.

Antisense: CCGGGATCCCGGTTACATGATGCTGCAGTTCT.

The subcloned construct was sequenced and validated.

Cell culture and transfection

C2C12 cells (American Type Culture Collection [ATCC]) were cultured in ATCC-formulated Dulbecco's modified Eagle's medium supplemented with penicillin–streptomycin and Glutamax (Gibco) as previously described (13, 34, 100). HEK293T (ATCC) cells were maintained using the same parameters as C2C12. The cells were maintained according to standard protocols at 37 °C and 5% CO_2_. Cells at passages 2 to 3 were seeded to a confluency of 50% to 60% for transfection-induced expression. For each transfection, 1:1.5 of DNA with Lipofectamine 3000 (Invitrogen) was utilized, following the manufacturer's instructions. Cells were processed for subsequent experiments 24 to 36 h post-transfection.

Staining for immunofluorescence

Cells were fixed with 4% paraformaldehyde for 20 min at room temperature. Subsequently, the fixed cells were permeabilized with 0.5% Triton X-100 for 6 min and then incubated with primary antibodies diluted in 5% goat serum in PBS for 2 h at room temperature or overnight at 4 °C (according to the manufacturer’s protocol). Following which, cells were washed thrice with 0.05% Tween-20 in PBS and thrice with PBS before incubating with fluorochrome-conjugated secondary antibodies for 1 h at 37 °C. It was followed by washing steps. Finally, the cells were mounted with VectaShield (Fisher Scientific).

Fluorescence microscopy and image analysis

For confocal imaging, the slides were visualized using a 60X water or 100X oil immersion objective with a numerical aperture of 1.4 and a refractive index of 1.515 on a NIKON TiE inverted research microscope (Nikon) equipped with a digital zoom, as described in previous literatures (1). For XY and Z-stack imaging, a resonant scanner with a line averaging of 4.0 was utilized. Z-stack images were acquired with a step size of 0.15 μm. Live cell imaging was performed with a culture dish under a thermal incubator (Tokai Hit) maintained at 37 °C with cells in the CO_2_-independent media.

For super-resolution structured illumination microscopy, a super-resolution Plan Apochromat TIRF 100×/1.5 numerical aperture/working distance 0.13 mm objective (Nikon) was employed to capture the morphology and distribution of RFP-LifeAct-decorated actins or MLS-RFP–marked mitochondria within cells (35, 100).

Images were exported using NIS-Elements Analysis AR (version 4.13), and ImageJ software (version 1.8.0_112, National Institutes of Health) was utilized for all confocal, super resolution, and IB image analysis. Mitochondrial length, fission–fusion rate, and velocity were calculated using Mitometer using default parameters (43).

Proximity biotin labeling and MS analysis

Sample preparation

HEK293T cells were initially cultured up to the third passage, when transfection was carried out with the plasmids in four 100 mm dishes using calcium phosphate. Following 24 h of transfection, the media were replaced with complete media supplemented with 50 μM of biotin. The cells were then cultured for an additional 18 h before harvesting. Before harvesting, cells were washed thrice with ice-cold PBS and lysed according to the protocol outlined by Kim et al. (101, 102). Specifically, 2 ml of BioID lysis buffer containing 50 mM Tris (pH 7.4), 500 mM NaCl, 0.4% SDS, 1 mM DTT, and protease inhibitor cocktail (PIC) was added to the scraped cell pellet, homogenized, and sonicated for 3 min at 50% amplitude. Subsequently, 2% Triton X-100 was added to the lysate, followed by another round of sonication for 5 min. Finally, 2 ml of 50 mM Tris was added to the lysate, and a third round of sonication was performed for 5 min under ice-cold conditions. The lysate was then clarified by centrifugation at 13,000g for 10 min. The cleared lysate was then added to 30 μl of Streptavidin agarose beads (Pierce) and incubated overnight at 4 °C with shaking. The following day, the beads were washed sequentially with wash buffer 1 containing 2% SDS in 50 mM Tris, wash buffer 2 containing 0.1% (w/v) deoxycholic acid, 1% (w/v) Triton X-100, 1 mM EDTA, 500 mM NaCl, and 50 mM Hepes (pH 7.5), wash buffer 3 containing 0.5% (w/v) deoxycholic acid, 0.5% (w/v) NP-40, 1 mM EDTA, 250 mM LiCl, and 10 mM Tris–Cl (pH 7.4), and finally with wash buffer containing only 50 mM Tris. The beads were then collected by centrifugation, and 100 μl of protein loading buffer was added. The mixture was heated at 95 °C for 15 min before analysis.

MS sample preparation and analysis

Protein samples were in-gel digested as described by Rawat et al. (103). Digested peptides were analyzed on a Q-Exactive mass spectrometer coupled with a Nano flow HPLC system (Easy-nLC 1200; Thermo Scientific). The sample was loaded onto an Easy Spray Column PepMap RSLC C18 (3 μm, 100 A0, 75 μm × 15 cm). Samples were eluted with a 60 min gradient between solution A (5% acetonitrile, water containing 0.2% formic acid) and solution B (95% acetonitrile in water containing 0.2% formic acid) at a flow rate of 300 nl/min. Mass spectra were acquired with an Orbitrap analyzer at the resolving power of 70,000 at m/z 200. The scan range selected was 400 to 1650 m/z. MS/MS was carried out in higher-energy collisional dissociation (normalized collision energy: 30%) mode with a resolving power of 17,500 at m/z 200. TopN 10 ions were taken for MS/MS. The lock mass option was enabled for accurate mass measurements.

The raw data files were searched against the Protein Database: Human FASTA database (January 5, 2023): HUMAN-uniprot-download_true_format_fasta_query__28homo_20sapiens_29_20AND_-023.01.05-06.28.17.74.fasta (20,330 sequences) and contaminant FASTA database containing 298 sequences using the Sequest HT search engine on Thermo Proteome Discoverer (version 2.2.0388). N-terminal acetylation and methionine oxidation were included as dynamic modifications; carbamidomethyl of cysteine as a static modification. A precursor mass tolerance of 10 ppm and a fragment mass tolerance of 0.05 Da were used. FDR setting was 0.01 and 0.05 for both peptide and protein levels; Percolator q value validation; delta Cn was 0.05. The following filters were applied to the Proteome Discoverer–analyzed data: 1) master is equal to master, 2) minimum 1 unique peptide and 1 peptide/protein, and 3) high peptide confidence.

Analysis of interactome

For differential analysis of the interactome, Venny 2.0 was utilized after filtering the raw data for the proteins, which are homogeneously highly abundant in the sample replicates. The loss of interaction (WT ∩ K97E′) was enriched using ShinyGO (version 0.77) using the Comprehensive Resource of Mammalian Protein Complex Protein–Protein Interaction database. Further, the interaction with PHB2 was confirmed using orthogonal techniques.

Immunoprecipitation

Cells were washed with ice-cold PBS, scraped, and pelleted in 1.5 ml microfuge tubes. The pellets were stored at −80 °C until further processing. For lysis, 1 ml of radioimmunoprecipitation assay buffer (50 mM Tris–Cl [pH 7.5], 150 mM NaCl, 1% NP-40, 0.1% SDS, 0.5% sodium deoxycholate, 1 mM EDTA, and 1× protease inhibitor cocktail) was added per 10^7^ cells (∼750 μl for a 90 mm dish), followed by sonication. The lysate was precleared by incubating with 25 μl of equilibrated Dynabead Protein-G (Invitrogen) per mL of lysate at 4 °C for 1 h with rotation. Beads were then pelleted using DynaMag. The cleared lysate was divided into two equal portions: one incubated with the anti-myc antibody and the other with normal rabbit IgG (negative control). Both were incubated overnight at 4 °C with 50 μl of Dynabead Protein-G per mL of lysate. Beads were washed thrice with lysis buffer. Bound proteins were eluted with 30 μl Laemmli sample buffer (5% β-mercaptoethanol) and heated at 95 °C for 15 min. Eluates were analyzed by IB.

Immunoblot

Postlysis protein concentration was assessed using the Bradford assay, following previously established protocols (1, 34, 100). Equal volumes of protein samples and 2X Laemmli buffer were combined and boiled at 95 °C for 10 min and cooled before gel loading. Subsequent to gel loading, proteins were separated via 10% SDS-PAGE and transferred onto a nitrocellulose membrane (Merck Millipore). The iIB procedures were conducted as described in prior literature (86, 104) by blocking the membrane in 5% blocking buffer (nonfat dry milk [NFDM] or bovine serum albumin) and incubating with primary and secondary antibodies according to the manufacturer’s protocol with antibodies mentioned in Methods Table S1. Blot reuse has been indicated in the “uncropped immunoblots.” Band intensities were quantified using ImageJ.

Blot overlay

Blot overlay was performed using the modified protocol used by Bhattacharjee et al. (35). Purified, heterologously expressed lamin A proteins (7 μg; WT and K97E) were resolved in 10% SDS-polyacrylamide gels and transferred onto nitrocellulose membranes. The membranes were incubated overnight with HEK293T cell lysate at 4 °C. The membranes were probed with lamin A and PHB2 antibodies to confirm input and interaction.

Far Western blot

Far Western blot was performed using the modified protocol standardized by Wu et al. (105). After blotting, the denatured proteins were renatured in situ by sequential incubation in renaturation buffer containing decreasing concentrations of guanidine hydrochloride (6 M, 3 M, 1 M, 0.1 M, and 0 M) in renaturation buffer (10% glycerol, 0.1 M NaCl, 20 mM Tris [pH 7.5], 1 mM EDTA, 0.1% Tween-20, guanidine hydrochloride, 1% NFDM, and 1 mM DTT) each for 30 min to 1 h, as described by Wu et al. The membrane was subsequently blocked with 5% NFDM and incubated overnight at 4 °C with HEK293T cell lysate. Following incubation, the membrane was probed with lamin A and PHB2 antibodies to confirm protein input and assess the interaction.

Quantitative PCR

qPCR was performed as per the previous standardized protocol by Sengupta et al. (1) with primers mentioned in Methods Table S2.

Nuclear fractionation

A subcellular fractionation buffer was produced by dissolving 20 mM Hepes (pH 7.4), 10 mM KCl, 2 mM MgCl_2_, 1 mM EDTA, 1 mM DTT, 1 mM EGTA, and PIC. Cells grown on two 60 mm culture dishes were washed thrice with ice-cold PBS and scraped into 400 μl of the cold fractionation buffer, followed by incubation on ice for 15 min. The cell suspension was then mechanically ruptured using a 27-gauge syringe, followed by another 20-min incubation on ice. The lysate was then centrifuged at 720g for 5 min to separate the nuclear pellet from the cytoplasmic supernatant, including mitochondria and membranes. The nuclear pellet was washed with 500 μl subcellular fractionation buffer, gently resuspended by pipetting, and homogenized further by passing through a 25-gauge needle to shave off impurities. The sample was centrifuged again at 700g for 10 min, after which the supernatant was discarded, and the nuclear pellet was then lysed using radioimmunoprecipitation assay lysis buffer and processed for IB.

Homology modeling and molecular docking

The 3D structure of the target protein was generated using homology modeling via SWISS-MODEL (https://swissmodel.expasy.org/), using 8J4I and 6JLB as templates and validated. For molecular docking, the modeled proteins were submitted to the HADDOCK web server (https://wenmr.science.uu.nl/haddock2.4), and rigid-body docking, followed by energy minimization and water refinement, was performed to generate multiple complex conformations (106). The top-ranked cluster, based on HADDOCK score and binding energies, was selected and analyzed for binding efficiencies.

Actin fractionation

To separate G-actin and F-actin, cells are first grown in a 60 mm dish until they reach approximately 60% confluency. Following this, the cells are washed twice with ice-cold PBS. Subsequently, actin-stabilizing lysis buffer, composed of 0.1 M Pipes (pH 6.9), 30% glycerol, 5% dimethyl sulfoxide (DMSO), 1 mM MgSO_4_, 1 mM EGTA, 1% Triton X-100, 1 mM ATP, and 1X PIC, is added, and the cells are incubated for 10 min. After scraping the cells and transferring them to a 1.5 ml Eppendorf tube, centrifugation is performed at 25,000g for 75 min at 4 °C. The resulting supernatant, containing G-actin, is collected in a separate Eppendorf tube and mixed with an equal volume of 2X loading buffer. Meanwhile, the pellet, enriched with F-actin, is dissolved in 400 μl of 1X loading buffer. Following this, both fractions are heated for 10 min at 100 °C to denature proteins. Finally, equal amounts of each fraction are loaded onto a gel, and separation is confirmed using IB.

Traction force microscopy

Preparation of TFM substrate

TFM utilizing the bead tracking method is conducted on polyacrylamide substrates with a stiffness of 10 KPa. The preparation of these substrates follows the procedures outlined by Kulkarni et al. (107), with minor modifications. A 27 mm glass bottom dish is initially treated with 10% 3-aminopropyltriethoxysilane (Sigma) for 15 min, followed by washing and subsequent treatment with a 0.5% glutaraldehyde solution (Sigma) for 45 min. The dish is thoroughly washed, air dried, and immediately utilized. Concurrently, a polyacrylamide solution with a stiffness of ∼10 KPa is prepared according to the methodology described by Tse and Engler (108). This involves the addition of 10% acrylamide, 0.1% bisacrylamide in water, along with 0.01% 0.2 μm FluoSpheres (Invitrogen), followed by thorough homogenization. Subsequently, the solution is polymerized using 10% ammonium persulfate and N,N,N',N'-tetramethylethylenediamine. Upon addition of the crosslinker and catalyst, 35 μl of the mixture is poured into the silanated glass bottom dish. A poly-l-lysine–coated 22 mm coverslip is then inverted onto the mixture to create a thin layer of polyacrylamide substrate. After 30 min, the coverslip is carefully removed, and the gel is washed to remove any excess acrylamide. Subsequently, 50 mg/ml Sulpho-SANPAH is applied to cover the polymerized gel, followed by exposure to UV light for 30 min. After aspiration of the solution, 1 ml of 100 μg/ml fibronectin is pipetted onto the activated gel and incubated overnight at 4 °C. Following the incubation period, the gel is washed thrice with PBS and sterilized using UV sterilization before cell seeding.

Image acquisition for TFM

Cells were trypsinized after 24 h of transfection and seeded into the prepared gels at a very low confluency (1 cell/100 μm) to ensure visibility of a single cell per frame. The cells are allowed to grow for at least 18 h before image acquisition. CO_2_-independent media (1 ml) are added to the dish to maintain uniform conditions throughout the experiment. Imaging is performed using a 40X objective on the Nikon inverted fluorescent microscope. Initially, a differential interference contrast image is captured along with the initial bead positions using a 561 nm laser. Subsequently, 1 ml of 10X (2.5%) trypsin is used to detach the cells for 30 min without disturbing the dish. After 30 min, another image is taken to capture localized bead movement. The images are then converted to TIFF format for subsequent traction force calculation.

Traction force measurement

The MATLAB algorithm developed by Kulkarni et al. is utilized for traction force measurement. The exact protocol outlined in their work is followed, employing the acquired images to obtain the constrained traction forces exerted by the cells.

ATP dependency determination

ATP dependency was carried out by protocols followed by Islam et al. (69), with slight modifications. Cells were seeded in replicates in a white-walled and clear 96-well plate with a seeding density of 10^4^ cells per well. Twenty-four hours post-transfection, the cells were supplemented with growth media containing 10 mM 2-DG, 1 μM OA, and a third set with a combination of both. DMSO-treated cells (0.01%) were taken as a control. After an hour of incubation with the drugs, the cells from the white-walled plate were lysed for 15 min with constant shaking, followed by ATP determination using the ATP determination kit (Abcam) by quantifying the luminescence emitted. The data were normalized with the viability assay performed in the replicate clear plate using the 3-(4,5-dimethylthiazol-2-yl)-2,5-diphenyltetrazolium bromide (MTT) assay.

After which, the dependencies were calculated using the formulas.

• I.Glucose dependence (%) = ([DMSO control – 2-DG]/[DMSO control – DGOA]) × 100.
• II.FAO and AAO capacity (%) = 100 – glucose dependence.
• III.Mitochondrial dependence (%) = ([DMSO control – OA]/[DMSO control – DGOA]) × 100.
• IV.Glycolytic capacity (%) = 100 – mitochondrial dependence.

Cellular and mitochondrial ROS determination

Cells were seeded in replicates in black-walled and clear 96-well plates with a seeding density of 10^4^ cells per well. Twenty-four hours post-transfection, the cells were supplemented with 500 nM MitoSOX or 5 μM CellROX Deep Red (Invitrogen) and incubated at 37 °C for 30 min. After which the cells were washed with PBS, and the fluorescent emission was measured to obtain the relative superoxide levels.

Metabolome analysis using MS

Sample preparation

Following 24 h of transfection, cells were washed twice with 150 mM NaCl at the indicated time points before metabolite extraction. Metabolite extraction and analysis were performed following a modified version of the protocol described earlier (109). Briefly, 500 μl of extraction solvent (water:acetonitrile:isopropanol = 2:3:3) containing 10 μM of 3-amino-4-methoxybenzoic acid as internal standard was added to the cell culture plate. Then, cells were scraped and collected in 1.5 ml Eppendorf tubes and stored in −80 °C freezer until further processing. Cell lysis was achieved through three cycles of rapid freeze–thawing using liquid nitrogen and a 37 °C incubator, followed by centrifugation at 13,000g for 30 min at 4 °C. A 100 μl aliquot of the supernatant was transferred into a 0.6 ml glass crimp-top microvial (27312; Supelco) and dried in a vacuum concentrator for approximately 1.5 h. Each sample was then treated with 30 μl of 2% methoxamine reagent (TS-45950; ThermoScientific) and incubated at 50 °C for 1 h with the lid closed to enable methyloxime derivatization. After cooling to room temperature, the samples were silylated by adding 50 μl of MSTFA reagent (M-132; Supelco) and heating at 65 °C for 1 hour in sealed tubes. Blank samples were prepared using only the extraction solvent and following the same procedure. For quality control (QC), pooled samples were created by combining equal volumes of metabolite extract from all samples.

GC–MS analysis: MS protocol

Samples were analyzed with a 7890B GC coupled to a 5977B single-quadrupole mass spectrometer (Agilent). The chromatographic separation of analytes was achieved on an HP-5MS column (30 m × 0.25 mm × 0.25 μm) with helium as the carrier gas. The front inlet was used in splitless mode at a temperature of 300 °C. The oven temperature was maintained at 70 °C for 5 min, followed by a ramp to 280 °C at 5 °C /min. The temperature was then increased to 295 °C at 10 °C /min and held at 295 °C for 4 min. MS source and MS quad temperatures were set to 230 °C and 150 °C, respectively. The electron ionization–MS spectra were acquired in full scan mode in the m/z range of 45 to 500. Samples were run in randomized order with intermittent injections of blank and pooled QC samples.

Data analysis

Column performance and consistency of instrument response were checked by manual inspection of chromatograms using MassHunter qualitative analysis software (Agilent). Data were deconvoluted to extract features using MassHunter quantitative analysis software. Compound identity was based on the National Institute of Standards and Technology library score as well as matching with authentic standards wherever available. Peak integration was performed with at least two unique ions for that feature. Features showing significant peaks in blank samples were removed. The blank areas were subtracted from the rest of the features. All features were normalized with respect to the internal standard as well as the protein concentration of each sample measured postcentrifugation during sample preparation. Features showing CV >30% in the pooled QC samples were removed from further analysis. The data were then sum-normalized, log-transformed, and Pareto scaled for multivariate analysis using MetaboAnalyst 5.0 (110) (https://www.metaboanalyst.ca/). Unsupervised principal component analysis and supervised orthogonal partial least squares-discriminant analysis were employed to identify group segregation patterns as well as to find features contributing to the segregation. Volcano plot analysis of the protein concentration-normalized data was used to screen for features showing fold change in relative abundance >1.25 (with raw p value <0.05) between the WT and the mutant, which were used for analysis of metabolic pathways using MetaboAnalyst 5.0. FDR correction (adjusted p < 0.05) was applied to select specifically altered metabolites of interest.

Statistical analysis

All the statistical analyses were done using GraphPad Prism (version 10, GraphPad Software, Inc). Unpaired Student's t test was used to determine the significance of the results with a 5% tolerance, unless mentioned. IB and qPCR experiments were each performed in three biological replicates (n = 3) for validation. GC–MS analysis included six replicates (n = 6). All the error bars in the graph represent SEM, with individual data points as biological replicates. Statistical significance marks indicate p values: ∗p ≤ 0.05, ∗∗p ≤ 0.01, ∗∗∗p ≤ 0.001, and ∗∗∗∗p ≤ 0.0001.

Data availability

All data are subject to availability on reasonable request.

Supporting information

This article contains supporting information.

Conflict of interest

The authors declare that they have no conflicts of interest with the contents of this article.