Lamin C conserves DNA replication factors via phase separation during oxidative stress for DNA replication recovery

Graphical abstract

Dear Editor,

Excessive reactive oxygen species (ROS) cause damage to biomolecules and lead to DNA replication fork slowdown and even stalling (Sies and Jones, 2020; Somyajit et al., 2017; Wilhelm et al., 2016); this state is referred to as oxidative stress. Eukaryotic cells employ diverse strategies to maintain redox homeostasis, including the formation of biomolecular condensates through phase separation. In mussels, for example, Catecholic 3,4-dihydroxyphenyl-l-alanine (Dopa)-containing mussel foot protein 3 and mussel foot protein 6 form redox insulators via phase separation to protect DOPA from oxidation (Valois et al., 2020). In mammalian cells, oxidative stress induces the assembly of stress granules through the phase separation of mRNAs and mRNA-associated proteins to improve cell survival (Guillén-Boixet et al., 2020; Wang et al., 2021).

Lamins are the main components of the nuclear lamina (NL) beneath the nuclear envelope in metazoan cells. In mammalian somatic cells, lamins are categorized into two types: the A-type, which comprises lamin A and lamin C, and the B-type, which includes lamin B1 and lamin B2 (Burke and Stewart, 2013). Several studies have reported that the overexpression of wild-type (WT) lamin C or mutants of lamin A, lamin C, or lamin B1 results in the formation of lamin protein condensates (Izumi et al., 2000; Sylvius et al., 2005, 2008). These condensates are regarded as nonfunctional protein aggregates that may lead to disease.

In this work, we found that endogenous lamin A/C could form nuclear protein condensates in response to H₂O₂-induced oxidative stress (Fig. 1A). We also observed that, in cells stably expressing low levels of green fluorescent protein (GFP)-lamin A (GLA) or GFP-lamin C (GLC) under oxidative stress, both the GLC and GLA proteins could form nuclear protein condensates, whereas GFP alone could not (Figs. 1B and S1). By quantifying the number of cells exhibiting lamin condensates under oxidative stress, we revealed that a significantly greater proportion of the cells formed GLC condensates than formed GLA condensates (Fig. 1C). Hence, we focused on the lamin C condensates afterward.

We first established a lamin C-mClover (LCmC) knock-in HeLa cell line, named lamin A knockout (LAO)-LCmC, to elucidate the properties and functions of the lamin C condensates (Fig. S2). LAO-LCmC cells express endogenous lamin C, which is tagged with a GFP variant mClover, but it lacks the expression of lamin A, potentially due to altered alternative splicing of the LMNA gene following gene editing (Fig. S2C). Upon exposure to oxidative stress, the LCmC proteins in LAO-LCmC cells were also induced to assemble into lamin C condensates (Fig. 1D and 1E). In addition to exogenous H₂O₂, the intracellular ROS generated by treating cells with menadione, a compound that generates oxidants through redox cycling in cells (Thor et al., 1982), also induced the assembly of lamin C condensates (Fig. S3A). We determined whether oxidative stress was unique in inducing the formation of lamin C condensates by testing several other stresses/conditions. We found that all these stresses/conditions failed to induce the formation of lamin C condensates (Fig. S3A). We further revealed that the oxidative stress-induced lamin C condensates did not contain any markers of the nuclear bodies that we tested (Fig. S3B). Overall, we conclude that lamin C condensates are newly recognized nuclear bodies and that their formation is a specific response to oxidative stress. Hence, we designate the lamin C condensate as the “lamin C body.”

We speculated that the lamin C bodies were formed by phase separation. We found that GLA or GLC bodies formed by transient overexpression and LCmC bodies induced by oxidative stress were both dissolved by the treatment of cells with 1,6-hexanediol (Figs. 1F and S4A), a compound known to disrupt liquid-like bodies formed by phase separation. Through a fluorescence recovery after photobleaching (FRAP) assay, we revealed that, after photobleaching, the fluorescence intensity of the GLA and GLC bodies recovered by approximately 42% and 52%, respectively, indicating that they possess liquid-like properties (Fig. S4B and S4C), whereas the fluorescence intensity of the LCmC bodies induced by oxidative stress recovered by approximately 18%, indicating that the oxidative stress-induced endogenous lamin C bodies possess gel-like properties (Fig. 1G and 1H). We subsequently performed an in vitro phase separation assay. Because full-length lamin C proteins easily precipitate in vitro, likely because of their coiled-coil domains, we overexpressed a series of lamin C truncates lacking a partial coiled-coil domain in LMNA-knockout (KO) HeLa cells to identify a truncated protein suitable for in vitro phase separation assays (Figs. S5 and S6A). As a result, the ∆1B∆2B59 truncated protein could assemble into lamin condensates in cells (Fig. S6A–C). The purified ∆1B∆2B59 protein could form droplet-like condensates at various concentrations, including at micromolar physiological concentrations, in physiological saline buffer with 10% polyethylene glycol (PEG) (Figs. 1I and S6D). Collectively, these data indicate that, under oxidative stress, lamin C assembles condensates through phase separation.

We transiently overexpressed several lamin truncations in cells to identify the essential domains within the lamin C molecule that facilitate phase separation (Fig. S7A). The results revealed that the truncations lacking the rod domain (Head-nuclear localization signal (NLS), NLS-Ig-like, GLC-NLS-C, and GLA-NLS-C) and the truncations lacking the head domain (GLA-ΔHead and GLC-ΔHead) failed to undergo phase separation, whereas the C-terminus deletion truncate (ΔC) retained the phase separation ability, indicating that both the rod domain and the head domain are essential for phase separation of the lamin proteins (Fig. S7). We noted that five amino acids with positive charges within the head domain are highly conserved (Fig. S8A). Eliminating these positive charges through the mutation of these arginines to alanines in GLC (GLC-5A) largely reduced its ability to form condensates (Fig. S8B and S8C). We also established several cell lines (mCL cells) stably expressing mCherry-lamin A (mC-LA), mCherry-lamin C-WT (mC-LC-WT) or mCherry-lamin C-5A (mC-LC-5A) by infecting LMNA-KO HeLa cells with lentiviruses and observed that mC-LC-5A completely lost its ability to assemble the lamin C bodies under oxidative stress (Fig. 1J and 1K). The 5A mutant of ∆1B∆2B59 (∆1B∆2B59-5A) also failed to form droplet-like condensates in vitro (Figs. S6D and S8D). Collectively, these data indicate that the positive charges provided by the arginine residues within the head domain promote the phase separation of lamin C proteins.

When we explored the regulatory mechanisms for the dynamics of lamin C bodies, we found that lamin C bodies formed by overexpression disassembled during mitosis, akin to the disassembly of the NL induced by the phosphorylation of lamin proteins, and lamin C bodies gradually disassembled upon removal of oxidative stress (Fig. S9A–C). The mass spectrometry analysis revealed that the level of S22 phosphorylation (pS22) in lamin A/C, a canonical site for lamin A/C phosphorylation that regulates NL disassembly, was lower in cells treated with (Murray-Nerger and Cristea, 2021) H₂O₂ than in control cells (Fig. S9D). Furthermore, compared with those of the control cells (HBSS), the levels of pS22-lamin C decreased significantly in cells under oxidative stress (H₂O₂), but when the cells were transferred to fresh H₂O₂-free medium, the level of pS22-lamin C recovered rapidly (Fig. S9E). These findings suggest that the pS22 level of lamin C fluctuates with changes in the level of cellular oxidative stress and that phase separation of the lamin C protein may be regulated by S22 phosphorylation. We verified this result by mutating S22 to aspartic acid (S22D) to mimic phosphorylation at this site or to alanine (S22A) to mimic its nonphosphorylated status and overexpressed these proteins in cells. We found that S22 phosphorylation diminished the phase separation ability of lamin C, whereas S22 nonphosphorylation preserved the phase separation ability and that the nonphosphorylation of lamin C also prevented disassembly of the lamin C bodies during mitosis (Fig. S9F–H). Collectively, these results indicate that a reduced level of pS22-lamin C promotes lamin C body assembly under oxidative stress, whereas an increased level of pS22-lamin C promotes disassembly of lamin C bodies upon the removal of oxidative stress.

Next, we investigated the biological significance of the lamin C bodies. Through an EdU incorporation assay, we found that most of the WT and LMNA-KO HeLa cells treated with H₂O₂ presented a very weak EdU intensity (Fig. S10A–D), suggesting that DNA replication in both types of cells under this stress was almost completely stalled. When the cells were moved into H₂O₂-free medium, DNA replication gradually recovered (Fig. S10B–D). However, compared with WT HeLa cells, LMNA-KO HeLa cells presented slower DNA replication recovery, as indicated by lower ratios of cell numbers with a high EdU intensity during the indicated recovery processes (Fig. S10B–D). The results of the DNA fiber assay revealed that the DNA replication fork velocity in LMNA-KO HeLa cells during recovery was much slower than that in control cells (Fig. S10E and S10F). Collectively, these data indicate that lamin A/C enhances DNA replication recovery upon removal of oxidative stress.

We assessed the roles of the lamin C bodies in DNA replication recovery by performing rescue experiments for both lamin A and lamin C in LMNA-KO HeLa cells through the transient expression of GFP, GLC-WT, GLC-5A, or GLC-5A-FUSN (GFP-lamin C-5A fused with the N-terminus of FUS protein (FUSN)) in mC-LA cells. FUSN, the intrinsically disordered region of FUS, has been used to confirm the role of phase separation in protein function (Sun et al., 2021), and here, we confirmed that FUSN rescued the phase separation ability of GLC-5A (Fig. S11). Compared with GFP and GLC-5A expression, the expression of GLC-WT and GLC-5A-FUSN, both of which can form lamin C bodies, rescued the velocity of the DNA replication fork and promoted the recovery of DNA replication (Fig. 2A–D). These results indicate that the lamin C bodies retain the ability to restore DNA replication during oxidative stress.

Finally, we investigated the mechanisms by which lamin C bodies orchestrate the DNA replication recovery process after oxidative stress. Through rapid immunoprecipitation mass spectrometry and immunofluorescence labeling, we found that the DNA replication factors proliferating cell nuclear antigen (PCNA), RPA1, RPA2, DNA polymerase delta (POLD) catalytic subunit (POLD1) and DNA ligase 1 (LIG1); the antioxidants peroxiredoxin 1 (PRDX1), PRDX2 and PRDX6; and the classic lamin A/C binding proteins barrier-to-autointegration factor (BAF) and lamina-associated polypeptide 2 (LAP2) were enriched in the lamin C bodies (Figs. 2E, 2F and S12A–D). Through a coimmunoprecipitation assay, we confirmed that lamin A/C interacted with these proteins under both control and oxidative stress conditions (Fig. S13), indicating that these proteins were recruited into the lamin C bodies via interactions with lamin A/C. Based on these results, we speculated that lamin C bodies might protect DNA replication factors from the impairment of oxidative stress by clustering them together with antioxidant proteins, and that the disassembly of lamin C bodies upon the removal of oxidative stress might promote the recovery of DNA replication through releasing DNA replication factors. To verify this, we assessed the interaction between PCNA and POLD1, a crucial interaction for the facilitation of DNA replication (Punchihewa et al., 2012), in S-phase cells after oxidative stress using a proximity ligation assay (PLA) (Fig. 2G). We observed that, compared with that in the control condition (HBSS-R2h), the PCNA‒POLD1 interaction in the DNA replication recovery process (H₂O₂-R2h) was reduced in both the WT and LMNA-KO HeLa cells and that the PCNA‒POLD1 interaction in LMNA-KO HeLa cells was weaker than that in WT HeLa cells under both HBSS-R2h and H₂O₂-R2h conditions (Fig. 2G and 2H). More importantly, compared with that under HBSS-R2h conditions, the degree of the decrease in the PCNA‒POLD1 interaction under H₂O₂-R2h conditions was significantly greater in LMNA-KO HeLa cells than in WT HeLa cells (Fig. 2H). These results indicate that oxidative stress impairs the PCNA‒POLD1 interaction and that lamin A/C may weaken this impairment. Through the PLA assay with mC-LA cells transiently expressing different lamin C proteins, we found that the expression of both FLAG-TagBFP-lamin C-WT (FBLC-WT) and FLAG-TagBFP-lamin C-5A-FUSN (FBLC-5A-FUSN) reversed the decrease in the PCNA‒POLD1 interaction during the DNA replication recovery process, whereas the expression of both FLAG-TagBFP and FLAG-TagBFP-lamin C-5A (FBLC-5A) did not (Fig. 2I and 2J). These results indicate that lamin C may weaken the impairment of the PCNA‒POLD1 interaction induced by oxidative stress by forming lamin C bodies.

Based on these results, we propose a working model to elucidate the roles of lamin C during oxidative stress (Fig. S14). Under oxidative stress, lamin C in the nucleus undergoes phase separation to assemble the lamin C bodies in response to stress. During this process, lamin C binds to DNA replication factors and antioxidants to promote their accumulation in lamin C bodies to protect DNA replication factors during oxidative stress. Upon the removal of oxidative stress, the lamin C bodies gradually disassemble, releasing DNA replication factors for the recovery of DNA replication. DNA replication errors, stalls, and even damage frequently occur in cells under oxidative stress, and in response, these cells may quickly inhibit DNA replication initiation (Davalli et al., 2018; Fragkos et al., 2015), leading to the entry of DNA replication factors into an idle state, during which these factors need to be properly managed for later rapid use upon the removal of oxidative stress. We found that the lamin C bodies induced by oxidative stress provide temporary storage sites for DNA replication factors and protect them by concentrating them and antioxidants within gel-like condensates. Antioxidants within lamin C bodies may be able to remove ROS more efficiently and weaken the ability of ROS to impair DNA replication factors. More importantly, the assembly/disassembly cycle of lamin C bodies in response to the cellular redox status fits the dynamic regulatory needs of DNA replication factors well, temporarily and protectively storing these factors during oxidative stress, or releasing them in a timely manner to participate in DNA replication recovery once stress is relieved.

In summary, for the first time, in this work, we report that endogenous lamin C proteins are able to assemble condensates via phase separation and reveal a crucial function of lamin C condensates in the protection of DNA replication factors during oxidative stress. In addition to storing DNA replication factors, lamin C bodies may also enrich and safeguard other factors and regulators of relevant metabolic pathways during oxidative stress for use in the cell recovery process upon the removal of oxidative stress. Overall, this work provides significant implications for understanding the cellular responses to both intrinsic and extrinsic stressors, although a more complete physiological regulatory framework underlying the phase separation of lamin C to form lamin C bodies remains to be elucidated.

References

[1] Burke B, Stewart CL. The nuclear lamins: flexibility in function. Nat Rev Mol Cell Biol 2013;14:13–24.

[2] Davalli P, Marverti G, Lauriola A et al Targeting oxidatively induced DNA damage response in cancer: opportunities for novel cancer therapies. Oxid Med Cell Longev 2018;2018:2389523.

[3] Fragkos M, Ganier O, Coulombe P et al DNA replication origin activation in space and time. Nat Rev Mol Cell Biol 2015;16:360–374.

[4] Guillén-Boixet J, Kopach A, Holehouse AS et al RNA-induced conformational switching and clustering of G3BP drive stress granule assembly by condensation. Cell 2020;181:346–361.

[5] Izumi M, Vaughan OA, Hutchison CJ et al Head and/or CaaX domain deletions of lamin proteins disrupt preformed lamin A and C but not lamin B structure in mammalian cells. Mol Biol Cell 2000;11:4323–4337.

[6] Murray-Nerger LA, Cristea IM. Lamin post-translational modifications: emerging toggles of nuclear organization and function. Trends Biochem Sci 2021;46:832–847.

[7] Punchihewa C, Inoue A, Hishiki A et al Identification of small molecule proliferating cell nuclear antigen (PCNA) inhibitor that disrupts interactions with PIP-box proteins and inhibits DNA replication. J Biol Chem 2012;287:14289–14300.

[8] Sies H, Jones DP. Reactive oxygen species (ROS) as pleiotropic physiological signalling agents. Nat Rev Mol Cell Biol 2020;21:363–383.

[9] Somyajit K, Gupta R, Sedlackova H et al Redox-sensitive alteration of replisome architecture safeguards genome integrity. Science 2017;358:797–802.

[10] Sun M, Jia M, Ren H et al NuMA regulates mitotic spindle assembly, structural dynamics, and function via phase separation. Nat Commun 2021;12:7157.

[11] Sylvius N, Bilinska ZT, Veinot JP et al In vivo and in vitro examination of the functional significances of novel lamin gene mutations in heart failure patients. J Med Genet 2005;42:639–647.

[12] Sylvius N, Hathaway A, Boudreau E et al Specific contribution of lamin A and lamin C in the development of laminopathies. Exp Cell Res 2008;314:2362–2375.

[13] Thor H, Smith MT, Hartzell P et al The metabolism of menadione (2-methyl-1,4-naphthoquinone) by isolated hepatocytes. A study of the implications of oxidative stress in intact cells. J Biol Chem 1982;257:12419–12425.

[14] Valois E, Mirshafian R, Waite JH. Phase-dependent redox insulation in mussel adhesion. Sci Adv 2020;6:eaaz6486.

[15] Wang L, Yang W, Li B et al Response to stress in biological disorders: Implications of stress granule assembly and function. Cell Prolif 2021;54:e13086.

[16] Wilhelm T, Ragu S, Magdalou I et al Slow replication fork velocity of homologous recombination-defective cells results from endogenous oxidative stress. PLoS Genet 2016;12:e1006007.

RIGHTS & PERMISSIONS

The Author(s) 2025. Published by Oxford University Press on behalf of Higher Education Press.