Abstract
Before human genome sequencing, a genome-wide study of sibling centenarian pairs identified a longevity-associated locus on chromosome 4. Here, we mapped the genes in this locus and identified a collagen gene, COL25A1. Introducing an SNP linked to longevity that changes a serine predicted to be phosphorylated to leucine in COL25A1, into col-99, the C. elegans ortholog, extended lifespan. These col-99(gk694263[S106L]) SNP-mutants exhibited enhanced innate immune-related transcriptional responses, and their lifespan extension was abolished by inhibiting the p38 MAPK pathway. YAP-1, a transcriptional co-activator responsive to extracellular matrix changes, was essential for this longevity. Mechanistically, we find that this SNP modifies furin-mediated cleavage of this transmembrane collagen in vitro, and expressing the cleaved extracellular domain of COL-99 alone was sufficient to prolong C. elegans’ lifespan. These findings reveal a potential mechanism by which a human centenarian-associated SNP in COL25A1 influences furin cleavage and shedding of the collagen ectodomain to promote healthy longevity.
Similar content being viewed by others
Introduction
Centenarians not only achieve exceptional longevity but also often exhibit a delayed onset of age-related diseases, providing valuable insights into the biological mechanisms of healthy aging1. While lifestyle factors, such as diet and physical activity, contribute to extended lifespans, as observed in the Okinawan population2, the genetic determinants that interact with these factors remain less understood.
Genome-wide association studies (GWAS) have begun to uncover genetic variations associated with longevity3. Notably, studies such as the New England Centenarian Study and the Long Life Family Study have identified genetic loci that may contribute to extended lifespan and reduced disease incidence among centenarians4. One such study by Puca and colleagues5 identified a significant locus on chromosome 4 (marker D4S1564) associated with exceptional longevity in a cohort of 308 individuals from 137 long-lived families. However, this study was prior to the human genome being sequenced, and thus, the specific genes and variants within this locus were not characterized.
Here, we focus on the chromosome 4 locus identified by Puca et al. and investigate the collagen gene COL25A1 as a candidate for influencing longevity5. We utilized Caenorhabditis elegans as a model organism to assess the role of a centenarian-associated single nucleotide polymorphism (SNP), S106L, introduced into col-99, the ortholog of human COL25A16,7,8. We demonstrate that this SNP promotes lifespan extension in C. elegans. We find that the S106L mutation alters the cleavage rate of COL-99 in vitro. The cleaved extracellular domain of COL-99 is sufficient to extend lifespan of C. elegans. Additionally, our findings indicate that genes involved in the innate immune response are upregulated in col-99(gk694263[S106L]) mutants and that the p38 mitogen-activated protein kinase (MAPK) pathway and the transcriptional co-activator yes-associated protein 1 (YAP-1) are crucial for this lifespan extension.
Our study provides novel insights into the genetic mechanisms regulating aging by identifying a specific genetic variant that influences lifespan through modulation of the innate immune response and MAPK signaling pathways. These findings may contribute to developing targeted interventions for age-related diseases and enhance our understanding of the genetic factors underlying human longevity.
Results
Introduction of centenarian COL25A1 SNP in C. elegans ortholog col-99
In 2001, Puca and colleagues performed a linkage study in centenarian sibling pairs to identify genetic loci associated with exceptional longevity5. Interestingly, they identified a 3 cM locus on chromosome 4 that was strongly associated with longevity5. However, the human genome was not yet sequenced at the time of publication, and this region had not been characterized. Therefore, we investigated this locus and found 108 ensemble features, including 33 protein-coding genes, 14 lincRNA, 2 miRNA, 2 snoRNA, and 9 snRNA (Figs. 1a, S1a, Supplementary Table 1). To identify causally implicated protein-coding genes, we cross-referenced genes to identify those with longevity-associated SNPs in centenarian GWAS or with studies showing that altering gene function would increase lifespan in any organism (Fig. S1a, Supplementary Table 1). Strikingly, we found a collagen gene, COL25A1, which was previously shown to be significantly associated with human healthy aging (“Wellderly phenotype”), but without mechanistic validation9.
a Depiction of the human chromosome IV with the putative longevity region highlighted in red. The overlapping open reading frames located in this region are displayed in separate rows, and COL25A1 is highlighted in red (for details, see Supplementary Table 1). b Lifespan curve showing short-lived col-99(ok1204) mutant compared to wild-type. Automated Lifespan was performed using n > 87. The log-rank test was performed for statistical analysis of mean lifespans. c Schematic showing the parallels drawn between human COL25A1 and C. elegans col-99 gene and the associated SNPs. Grey lines indicate observed SNPs in COL25A1; Red boxes indicate SNPs/mutations associated with diseases, while SNPs marked in green are associated with longevity. Dark blue bars indicated collagen domains. Predicted furin sites are marked by dashed lines; orange clamps indicate cysteine-knots. SNPs in dark grey are potential longevity-associated SNPs matched on col-99. Light blue boxes are known for deletion/stop mutation in C. elegans. The Brown box indicates the transmembrane domain.
In a subsequent literature search, we found that COL25A1 exhibited several SNPs associated with exceptional lifespan and/or healthy aging (Fig. 1b, Supplementary Table 1). Most of the molecular alterations of these longevity-associated SNPs were predicted to result in the loss or reduction of the function of COL25A1. To determine whether the loss of COL25A1 is beneficial, we investigated the effect of a loss-of-function mutation in the COL25A1 orthologue col-99 in C. elegans. However, the col-99(ok1204) deletion mutant was short-lived with a mean lifespan of 18.4 days, approximately 5 days less than the wild type (Fig. 1b; blue line, Supplementary Table 2). Overexpression of full-length COL-99 protein tagged with GFP, on the other hand, did not affect the mean lifespan (Fig. 1b; green line, Supplementary Table 2). Assuming COL25A1 is responsible for the above-mentioned linkage, these findings indicate that the effect of SNPs in COL25A1 on longevity has a more complex mechanism.
S106L SNP in col-99 extends lifespan in C. elegans
To investigate this unknown mechanism, we tried to map all human SNPs in the C. elegans col-99 gene. However, due to the vast evolutionary distance and the repetitive nature of collagen, we could only reliably map three sites with some confidence (human S116L → C. elegans S106, G225C → G215, R402C → R393) (Fig. 1c). We focused on the S116L SNP in COL25A1 (S106L in col-99 in C. elegans) as it is proximal to a conserved furin cleavage site with the potential to impact domain processing (Fig. 1c).
First, to determine if the human S116L SNP introduction into col-99 in C. elegans had any effect on lifespan, we obtained C. elegans harboring this S106L SNP in col-99 (hereafter referred to as col-99(gk694263[S106L]) mutants) and outcrossed ten times (10X) to remove any potential background mutations. We then performed lifespan studies with 4X, 8X, and 10X outcrossed col-99(gk694263[S106L]) mutants and found all outcrossed col-99(gk694263[S106L]) mutants increased mean lifespans (Fig. 2a, S1b, Supplementary Table 2) by approximately 26% compared to wild-type control, indicating not only that S106L in humans might indeed be causative for prolonged life, but also re-assuring our mapping of this SNP between the species.
a Lifespan curve showing an extension of life span in col-99(gk694263[S106L]) SNP mutant. b Lifespan curve showing the extracellular domain (CED06 col-99(syb4332[COL-99EXT])) is sufficient to increase lifespan. The lifespans were performed in at least two biological replicates at 20 °C. Kaplan-Meier Log-Rank test was performed for statistical analysis on mean lifespans using the online software OASIS. A lifespan summary for individual experiments is provided in Supplementary Table 2. c In-vitro furin cleavage assay shows differential cleavage of COL-99 peptide. Representative data from three different assays are shown.
To independently validate that the increased lifespan is due to the single mutation in col-99, we CRISPR-generated the S106L SNP mutation into wild-type animals (hereafter referred to as col-99(syb4350[S106L])). Reassuringly, we found a similar extension in the mean lifespan of col-99(syb4350[S106L]) compared to the wild type (Fig. S1c, Supplementary Table 2). The col-99(gk694263[S106L]) mutants superficially looked wild type but were slightly delayed developmentally by a few hours, with larval L4 stage (Fig. S1d, Supplementary Table 2). Furthermore, the longevity of col-99(gk694263[S106L]) mutants did not correlate with oxidative or heat stress resilience (i.e., col-99(gk694263[S106L]) mutants were more susceptible to 14 mM arsenite or 32 °C treatment; Fig. S1e, f, Supplementary Table 2).
These results confirmed that the S106L mutation in col-99 extends lifespan independently of stress resistance pathways. As the mutation is near a conserved furin cleavage site involved in collagen processing, we next investigated whether the extracellular domain of COL-99 mediates this effect.
The extracellular domain of COL-99 is sufficient for increased longevity
COL25A1 in humans and COL-99 in C. elegans are both single-pass type II transmembrane collagen proteins with predicted furin cleavage sites7,10. For COL25A1, this site is reported to be important for shedding, creating the so-called CLAC (collagen-like Alzheimer's amyloid plaque component) fragment11. Besides its association with Alzheimer’s Disease, the function of this shredded collagen fragment is not understood, although it seems to play a role in muscle development and innervation12,13. Similarly, in C. elegans, col-99 was important for axonal guidance, and the first and third predicted furin cleavage sites seem important for its function6,8.
As S116L (and S106L) are close to the conserved (first) furin cleavage site, we hypothesized that the S106L mutation in C. elegans might affect furin cleavage, leading to a change in the extracellular domain shedding rate. To test this hypothesis, we expressed the extracellular domain of COL-99 under its endogenous promoter in C. elegans. The expression of a shed extracellular domain of COL-99 prolonged the mean lifespan of C. elegans by approximately 11% already (Fig. 2b, Supplementary Table 2), consistent with the longevity effect of the S106L SNP in col-99(gk694263[S106L]) mutant.
Although the furin recognition site is often described as R-X-[K/R]-R⇓, the full recognition sequence is longer and amino acids up to 14 amino acids N-terminally (P14) and six amino acids C-terminally (P6′) from the cleavage site are described to influence furin activity (see Fig. 3a for clarification of P1/P1’ nomenclature and positioning in COL-99 and COL25A1)14. To measure the effect of an S to L mutation at the P2’ site in C. elegans, we synthesized short fluorescently quenched peptides containing the furin cleavage site and a serine or leucine residue, respectively. The quenching was resolved upon cleavage, and an increased fluorescence was observed (Fig. 2c). Contrary to our expectation, the mutant showed a roughly halved Vmax (30 vs 52) compared to the wild-type sequence, indicating cleavage of S106L by furin, but at a slightly decreased rate. Given the abundance of furin, this small difference can hardly explain the observed effect in vivo, especially as it seems to contradict our finding from the expression of the extracellular domain alone, which suggested benefits from increased cleavage.
a Phosphorylation potential of serine residue in the SNP next to the furin cleavage site (highlighted in green) in COL25A1 and COL-99. Bold “Serines” are SNP S106 and S116, respectively. Prediction by NetPhos 3.1. b Lifespan curve showing incapability of phosphomimetic variant of col-99 S106D (CED05 col-99(syb4352[S106D])) to extend lifespan. The lifespan assay was performed in five independent biological replicates at 20 °C. The Kaplan-Meier Log-Rank test was performed for statistical analysis on mean lifespans using the online software OASIS. A lifespan summary for individual experiments is provided in Supplementary Table 2.
Besides its location near the furin cleavage site, both sites are also predicted to be phosphorylated (Fig. 3a). Although extracellular phosphorylation is not completely understood yet, its importance has become more evident over the last few years15,16,17,18. Bioinformatically, several phosphorylation sites in COL25A1 have been predicted19. Furthermore, it was shown at least twice that phosphorylation might affect furin cleavage20. We therefore speculated that phosphorylation of S106 might inhibit COL-99 shedding, thereby reducing the extracellular concentration of the extracellular domain of COL-99. The described SNP might prevent this phospho‑dependent inhibition, thus increasing shedding. Consequently, a CRISPR mutant of col-99 introducing a phosphomimetic S106D exchange (col-99(syb4352[S106D])) did not significantly affect the mean lifespan (3 out of 5 independent trials; Fig. 3b, Supplementary Table 2).
Taken together, this suggests a model wherein phosphorylation of serine 106 influences furin cleavage and release of the extracellular domain of COL-99 to alter downstream processes to increase lifespan.
Innate immune response is upregulated in col-99(gk694263[S106L]) mutants
To identify the downstream mechanisms affected by the S106L mutation in col-99(gk694263[S106L]), we performed transcriptomic profiling of the col-99(gk694263[S106L]) mutant and wild type at the young adult stage. The introduction of col-99 SNP drove limited transcriptomic effects (Fig. 4a, Supplementary Table 3), including a ~ 20% increase in the col-99 transcript level (Fig. 4b), suggesting an autoregulation. Gene Ontology (GO) analysis revealed a downregulation of genes in the broad “oxidation-reduction process” category (Fig. 4c, Supplementary Table 3). On closer inspection, some of these genes are core components of the mitochondrial respiratory chain and inner membrane, directly involved in ATP production. This suggests that col-99(gk694263[S106L]) mutants may have an energy production phenotype; however, we did not functionally assess mitochondrial activity, so this remains a hypothesis.
a Logarithmic scale representing the distribution of differentially expressed genes in col-99(gk694263[S106L]) SNP mutant versus wild-type. b The graph shows an increase in the expression of col-99 transcripts in col-99(gk694263[S106L]) SNP mutant compared to wild-type (in terms of FKPM values). c GO term analysis showing downregulated and upregulated genes in col-99(gk694263[S106L]) SNP mutant to wild-type. RNA sequencing was performed with three biological batches. d Targeted screening to find regulators of T24B8.5 GFP expression. Data is plotted as mean, and error bars represent SEM. Two-tailed Welch’s t-test statistical analysis. The experiment was performed in three biological batches. col-99(SNP) is col-99(gk694263[S106L]).
Meanwhile, innate immune response genes were upregulated (Fig. 4c, Supplementary Table 3). ATF-7 is the master regulator of immune response in C. elegans21, and we found it to be upregulated by col-99(gk694263[S106L]) expression (Fig. S2a, Supplementary Table 3). However, in our epistatic lifespan analysis, we observed that atf-7 was not required for mean lifespan extension in col-99(gk694263[S106L]) mutants (Fig. S2b, Supplementary Table 2). Arginine kinase (creatine kinase) argk-1 was also upregulated in col-99(gk694263[S106L]) mutants (Supplementary Table 3). We verified this increase in col-99(gk694263[S106L]) mutants with a reporter strain argk-1p::GFP using fluorescence microscopy (Fig. S2c). Previously, ARGK-1 overexpression was shown to increase lifespan via AMPK, which triggers the activation of the innate immune response22,23. Therefore, we checked for the requirement of argk-1 in lifespan extension by col-99(gk694263[S106L]) mutants. We assayed the lifespan of argk-1(ok2993); col-99(gk694263[S106L]) double mutants but found that the extended lifespan of col-99(gk694263[S106L]) mutant was independent of argk-1 (Fig. S2d, Supplementary Table 2). Since DAF-16 is also an important modulator of innate immune response and lifespan24, we tested the nuclear localization of DAF-16 in col-99(gk694263[S106L]) mutants. However, the differences in DAF-16 nuclear localization were insignificant (Fig. S2e). Taken together, our approach in testing these genes, differentially expressed in RNA sequencing, did not identify underlying pathways driving longevity.
Next, we employed a candidate RNAi screen approach to identify the pathways mediating the lifespan extension of col-99(gk694263[S106L]). Given the strong innate immune transcriptional signature, we tested two immune-response reporter genes, clec-85 and T24B8.5, that were also differentially regulated by col-99(gk694263[S106L]) (Fig. 4d, S2f, Supplementary Table 3). The transcriptional reporter driving GFP under the T24B8.5 promoter (T24B8.5p::GFP25) showed a two-fold upregulation in the fluorescence intensity in col-99(gk694263[S106L]) mutants compared to wild-type, making it suitable for screening. Given that col-99(gk694263[S106L]) induces T24B8.5p::GFP in vivo, knocking down genes that are required for mediating col-99(gk694263[S106L])-dependent T24B8.5 upregulation may reveal underlying pathways. Therefore, we screened through a targeted candidate library including known immunogenic-, collagen-, and mechano-receptors using the T24B8.5p::GFP reporter strain.
We started with genes in the canonical innate immune response pathway (P38 MAP Kinase pmk-126, Toll and Interleukin 1 Receptor tir-127, ATF cAMP-dependent transcription factor atf-728, and Toll-like receptor tol-129) and found that RNAi targeting tir-1 suppressed the col-99(gk694263[S106L])-mediated T24B8.5p::GFP induction (Fig. 4d), indicating that this immune receptor was required. However, it is unlikely that the extracellular domain of COL-99 would bind the TIR-1 receptor directly for signaling, suggesting additional upstream signaling. To identify the potential direct binding partners/receptors of shed COL-99 extracellular domain, we turned to the COL-99 collagen receptors, the Discoidin Domain Receptors (ddr-1 and ddr-2), and basement membrane component nidogen homolog nid-1, which are putative binding partners of the cleaved COL-99 ectodomain and are involved in col-99-mediated neuronal axon guidance6. However, knockdown of ddr-1, ddr-2, or nid-1 did not suppress col-99(gk694263[S106L])-mediated T24B8.5p::GFP induction (Fig. 4d). COL-99 is expressed in the hypodermis, and the furin cleavage and shedding of the COL-99 ectodomain are required for neuronal axon guidance6. Thus, a neuronal receptor may be required. We tested neuronal and mechanosensitive transient receptor potential (TRP) channel trpa-1/TRPA and osm-9/TRPV, which are shown to be involved in longevity30,31. We found that osm-9 was required for col-99(gk694263[S106L])-mediated T24B8.5p::GFP induction (Fig. 4d). OSM-9 is a mechanosensitive receptor upstream of YAP-1 (yes-associated protein), a mechanoresponsive transcriptional co-activator activated upon stiffness changes in the extracellular collagen network and implicated in the control of lifespan32,33,34. In addition, yap-1 is functionally implicated in Touch Receptor Neuron asymmetric neurite extension35 and pathogen resistance in C. elegans36, consistent with a potential role for neuronal regulation and innate immunity. We found that the knockdown of yap-1 also suppressed col-99(gk694263[S106L])-mediated T24B8.5p::GFP induction (Fig. 4d). Taken together, we identified two receptors, tir-1 and osm-9, and the co-transcriptional activator yap-1 required for col-99(gk694263[S106L]) induction of the transcriptional innate immune response program.
Since SNP could possibly affect the folding or trimerization of COL-99 collagen, we also included the hsp-4::GFP reporter as part of a broader mechanistic screen for stress-responsive pathways, to assess whether col-99 SNP might affect the unfolded protein response (UPR), despite no prior evidence linking COL-99 to ER stress. By contrast to the immune-related reporter screen, yap-1 and tir-1 were not required for col-99(gk694263[S106L]) induction of hsp-4 (Fig. S2g, h), suggesting that our reporter screen findings need to be validated with functional lifespan assays.
The P38 MAPK pathway is required for longevity extension by col-99(gk694263[S106L])
Given the distinct downstream pathways we recovered for col-99(gk694263[S106L]), we aimed to determine the pathway that mediates col-99(gk694263[S106L]) longevity. For instance, our reporter assays revealed that the COL-99 receptors ddr-1 and ddr-2 were not required for immune and UPR responses; this does not rule out their role in lifespan regulation. However, we found that loss of ddr-1, or ddr-2, or the double mutant ddr-1; ddr-2 did not suppress col-99(gk694263[S106L])-induced longevity (Fig. S3a, Supplementary Table 2), suggesting that DDR-1 and DDR-2 are the COL-99 receptor for axonal guidance6 but not for longevity. Another likely candidate for COL-99 ectodomain receptor is TIR-1, and we found that tir-1 was required for col-99(gk694263[S106L])-induced lifespan extension (Fig. 5a, Supplementary Table 2). Since tir-1 is upstream of the p38 MAPK signaling cascade, we tested whether the downstream protein p38 MAPK (pmk-1) was required for lifespan extension. Although RNAi of pmk-1 was not sufficient to block T24B8.5p::GFP induction in col-99(gk694263[S106L]) mutants, knockdown of pmk-1 abolished the longevity of col-99(gk694263[S106L]) mutants (Fig. 5b, Supplementary Table 2). We confirmed these lifespan results of col-99(gk694263[S106L]) with the CRISPR-generated mutant CED04 col-99(syb4350[S106L]), and also showed no lifespan extension with the phosphomimetic CED05 col-99(syb4352[S106D]) mutants in combination with tir-1 or pmk-1 knockdown (Fig. S3b–e, Supplementary Table 2). Consistently, tir-1 and pmk-1 were required for lifespan extension in the CED06 col-99(syb4332[COL-99EXT]) mutant expressing the ectodomain of COL-99 (Fig. 5c, d, Supplementary Table 2).
Survival curve showing the dependence of the increased lifespan of col-99(gk694263[S106L]) SNP mutant on tir-1(a), pmk-1(b). Survival curve showing the dependence of the increased lifespan of CED06 col-99(syb4350[S106L]) SNP mutant on tir-1(c) and pmk-1 (d). e Western Blot showing activation of PMK-1 in col-99(gk694263[S106L]) SNP mutant. f Survival curve showing the dependence of col-99(gk694263[S106L]) SNP on yap-1 for extension of lifespan. g Hypothetical model of signal relay from the extracellular domain of COL-99 to activate the p38 MAPK pathway, which signals to YAP-1 to extend longevity in the col-99(gk694263[S106L]) SNP mutant. Genes displayed are unlikely to be in the same cell. Please, also see Fig. S4. The lifespans were performed in at least two biological replicates at 20 °C. Kaplan-Meier Log-Rank test was performed for statistical analysis on mean lifespans using the online software OASIS. The Western blot was done twice. A lifespan summary and raw western blots for individual experiments are provided in Supplementary Table 2. col-99(SNP) is col-99(gk694263[S106L]).
Furthermore, in line with the requirement of pmk-1 for lifespan extension, col-99(gk694263[S106L]) mutants showed higher p38 MAPK/PMK-1 phosphorylation (Fig. 5e, Supplementary Table 2), an event generally known to drive increased ATF-7 transcription factor levels. Despite this association, ATF-7 was not required for the extended lifespan in col-99(gk694263[S106L]) mutants (Fig. S2b, Supplementary Table 2). This led us to investigate other potential mediators, and we discovered that YAP-1 played a critical role in facilitating the longevity effects of col-99(gk694263[S106L]) mutants (Fig. 5f, Supplementary Table 2). Collectively, our data support a model in which a non-canonical immune response pathway involving osm-9, tir-1, pmk-1, and yap-1 is activated upon blunting the phosphorylation of COL-99 at the S106 site, and thus altered furin-mediated cleavage resulting in the subsequent release of the COL-99 ectodomain (Fig.5g, S4).
Discussion
Previously, in at least three centenarian cohorts and one healthy aging cohort, several SNPs in collagen XXV associated with healthy human aging have been identified9,37,38,39. Despite these associations, causal mechanisms linking specific genetic variants to healthy longevity remain largely unvalidated. In this study, we addressed this gap by demonstrating that a rare coding variant in the collagen gene COL25A1, specifically the S116L SNP in the COL25A1 ortholog, is sufficient to extend the lifespan in C. elegans. By introducing the S106L orthologous mutation into col-99, we observed an approximately 26% increase in lifespan (Figs. 2a, S1b, Supplementary Table 2), suggesting a conserved role of this variant in longevity.
Mechanistically, the S106L mutation influences furin-mediated cleavage of transmembrane COL-99, altering the release of its extracellular domain. This extracellular domain is sufficient to promote lifespan extension, implicating the remodeling of extracellular matrix (ECM) in aging processes. Our results also reveal that the lifespan extension in col-99(gk694263[S106L]) mutants depends on the p38 MAPK pathway and the transcriptional co-activator YAP-1. Interestingly, this effect is independent of ATF-7, a typical downstream target of p38 MAPK involved in immune responses28, indicating alternative signaling pathways are at play.
The involvement of YAP-1, known for its roles in mechanotransduction and cellular proliferation40 and mechanotransduction-mediated longevity33, suggests that ECM changes can influence longevity through mechanoresponsive signaling pathways. The activation of innate immune response genes further supports a model where ECM alterations modulate immune signaling to impact aging.
Our findings suggest that similar mechanisms may operate in humans, given the conservation of these pathways. Proteomic analyses of centenarians have revealed elevated levels of proteins involved in angiogenesis and cell junctions, suggesting enhanced tissue homeostasis41. Additionally, single-cell transcriptomics of peripheral blood mononuclear cells (PBMCs) from centenarians has indicated improved immune function, characterized by increased T cell populations42. Notably, serum proteomic signatures have shown enrichment of collagen proteins, such as COL28A1 and COL6A3, in centenarians43, highlighting the potential role of collagen in longevity. Identifying COL25A1 variants affecting furin cleavage and ECM dynamics opens new avenues for research into the genetic regulation of human aging. Future studies should focus on elucidating where COL-99 serine 106 is phosphorylated, i.e., in the Golgi or the extracellular space. Furthermore, our hypothetical model shows genetic interactions (Fig. 5g) and plausible protein interactions (Fig. S4), but the direct interactions between the COL-99 extracellular domain (EXT) and its signaling partners are needed to fully understand the mechanisms underlying this lifespan extension. We speculate that the interaction of COL-99 EXT with the alleged receptors occurs in a tissue-specific manner, with possibilities including neurons and the intestine, as TIR-1 is expressed in both tissues, and osm-9 is predominantly expressed in neurons. Since col-99 is expressed in both hypodermis and neurons, we hypothesized that cleavage of COL-99 occurs in hypodermis and/or neurons. The EXT domain then signals through TIR-1 in the intestine or neurons to extend lifespan. Given that the p38 MAPK pathway in the intestine plays a crucial role in regulating lifespan, we believe that the interaction of TIR-1 with COL-99 EXT in the intestine regulates lifespan. However, it would be interesting to test the requirement of p38 MAPK in neurons for lifespan regulation.
One limitation of our study is the use of C. elegans as a model organism, which, while highly informative, may not fully recapitulate the complexity of human aging. Additionally, the direct applicability of the S116L variant in humans requires further validation in mammalian systems.
Supporting the relevance of our findings, multiple human genetic studies have associated COL25A1 variants with healthy aging, longevity, and age-related diseases (Fig. 1c, Supplementary Table 1). For instance, SNPs in COL25A1 are associated with Alzheimer’s disease44,45, which maps to the Aβ-binding side in COL25A1, suggesting a failure of the extracellular domain of COL25A1 to scoop up Aβ oligomers44. Similarly, SNPs in COL25A1 are associated with congenital cranial dysinnervation disorder46,47, which can lead to a loss of the most outer collagen domain (COL3) in COL25A1 expressed from the muscle that is required to bind to receptors protein tyrosine phosphatases σ and δ (PTP σ/δ), a process essential for intramuscular motor innervation48. Previous comparisons of centenarians with disease GWAS studies found an overlap of SNPs in genes associated with diseases, such as Alzheimer’s and coronary artery diseases, and other SNPs in the same gene associated with longevity49,50. However, these risk alleles in these genes for diseases are depleted in centenarians and/or compensated by a constellation of multiple beneficial genetic variants in centenarians49,51,52. These observations suggest that centenarians in given disease pathways have resilience variants tipping the balance to health and not down towards disease progression.
In line with this idea, mapping proteins associated with human healthspan, COL25A1 is in the first central node together with LRP1, TOMM40, and CREBBP, which are also implicated in Alzheimer’s disease, and potentially under dietary restrictions might signal through NOTCH to promote human healthspan53. Further GWAS studies revealed SNPs in COL25A1 are associated with resistance to COVID-19 infections54 and higher spermidine levels55, a polyamine compound inducing autophagy and increasing the lifespan of model organisms56, suggesting some implications in resilience.
In summary, our research identifies a causal role for a coding variant in COL25A1 in extending lifespan via modulation of ECM remodeling and immune signaling pathways. These insights advance our understanding of the genetic factors that contribute to longevity and may inform the development of therapeutic strategies targeting age-related diseases.
Methods
Strain maintenance
All strains were maintained at 20 °C, on NGM (nematode growth media) plates seeded with Escherichia coli OP50.
The strains used in this study are: N2 (wild-type Bristol), RB1165 col-99(ok1204) IV, COL-99 overexpression VH2847 hdIs73 [GFP::COL-99, pha-1( + )], AU78 agIs219 [T24B8.5p::GFP::unc-54 3’UTR + ttx-3p::GFP::unc-54 3’UTR] III, yap-1(tm1416), VC1518 atf-7(gk715) III, MAH547 sqEx82 [argk-1p::GFP + rol-6(su1006)], MAH205 argk-1(ok2993) V, TJ356 zIs356 [daf-16p::daf-16a/b::GFP + rol-6(su1006)], SAL146 pha-1(e2123) III; denEx24 [clec-85p::GFP+pha-1( + )], SJ4005 zcIs4 [hsp-4::GFP] V, RB970 ddr-1(ok874), RB788 ddr-2(ok574), VH1681 ddr-1(ok874); ddr-2(ok574) X; evIs111[rgef::GFP] V, VC40561 col-99(gk694263[S106L]) SNP mutant [4 times, 8 times, and 10 times outcrossed]57. col-99(SNP) is interchangeably used in figures for col-99(gk694263[S106L]) mutant, which is originally the VC40561 strain obtained from the Million mutation project. The strain has a missense mutation G → A at position 144813 on chromosome IV, generating col-99(gk694263).
Strains generated in this study
Strains generated by CRISPR gene editing (performed by SUNY Biotech)-F29C4.8j.1 is the isoform of col-99 which was modified to obtain the following transgenic strains: CED04 PHX4350 [col-99(syb4350[S106L])] (col-99 S106L mutant): Precise one base pair missense mutation (TCG → TTG) was introduced in col-99 gene using CRISPR. It resulted in the expression of nonpolar leucine amino acid (L) at the 106th position, instead of polar serine (S) in wild-type COL-99 (termed as S106L).
CED05 PHX4352 [col-99(syb4352[S106D])] (col-99 S106D mutant): The mutation TCG → GAC was induced in col-99 gene using CRISPR gene editing, which resulted in the expression of phosphomimetic aspartic acid residue (D) at 106th position, instead of serine (S) in wild-type COL-99 (termed as S106D).
CED06 PHX4332 [col-99(syb4332[COL-99EXT])] (col-99 extracellular domain): Precisely 3075 bp (95 aa) were deleted after the first methionine coding codon ATG in col-99 gene so that first conserved Furin cleavage site “RRVR” at 104th position8 in the non-collagenous extracellular domain of COL-99 was expressed, leading to expression of only the COL-99 extracellular domain (COL-99 EXT).
Strains generated by genetic cross: col-99(gk694263[S106L]); T24B8.5p::GFP, col-99(gk694263[S106L]); yap-1(tm1416), col-99(gk694263[S106L]); atf-7(gk715), col-99(gk694263[S106L]); argk-1p::GFP, col-99(gk694263[S106L]); argk-1(ok2993), col-99(gk694263[S106L]); DAF-16::GFP, col-99(gk694263[S106L]); clec-85p::GFP, col-99(gk694263[S106L]); hsp-4p::GFP, col-99(gk694263[S106L]); ddr-1(ok874); ddr-2(ok574), col-99(gk694263[S106L]); ddr-2(ok574), col-99(gk694263[S106L]); ddr-1(ok874).
Genotyping col-99(gk694263[S106L]) SNP mutation
A quick, reliable, and inexpensive method of SNP detection by endpoint PCR was followed as described previously58. Briefly, the principle of allele variant detection is based on SuperSelective (SS) primers design, consisting of a 16 bp anchor, a 6 bp optimal bridge, and a 7 bp foot region with terminal mismatch nucleotide G → A for wild-type and col-99(gk694263[S106L]) mutant allele. When primers anneal to the template DNA, any single nucleotide mismatch in the short foot region would inhibit the binding of primers and, hence, prevent primer extension. So, when PCR was run with wild-type gDNA as a template, only primers designed for wild-type allele would bind to wild-type gDNA and yield a specific band (expected size of the amplicon is 348 bp) at an annealing temperature of 62 °C, while a combination of mutant primers with wild-type gDNA should not yield any band.
In-silico prediction of phosphorylation sites in collagen
Full-length sequences of COL-99 and COL25A1 were subjected to the NetPhos 3.1 algorithm (https://services.healthtech.dtu.dk/services/NetPhos-3.1/) and analyzed for serine phosphorylation. Only predictions with a score higher than 0.5 and in the vicinity of the furin cleavage site near the S106 / S116 SNP in question were considered59,60.
Automated lifespan assay
The automated lifespan assay was performed as described by Stroustrup et al.61. Briefly, C. elegans populations were subjected to population lysis, and the eggs were transferred to OP50 plates. At the L4 stage, the animals were transferred to either OP50 or RNAi plates that contained 50 μM FUdR. At day two of adulthood, eggs were removed through repeated washing, and the animals were transferred to lifespan plates of the same composition, featuring a tight-fitting lid (BD Falcon Petri Dishes, 50×9 mm). The plates were placed in the Epson V800 flatbed scanners and continuously imaged for 60 days.
Peptide synthesis
The internally quenched fluorogenic peptides S106S (Abz-VAKSRRVRNS-Lys[DNP]) and S106L (Abz-VAKSRRVRNL-Lys[DNP]) were synthesized by manual and automated solid-phase peptide synthesis (SPPS) (peptide synthesizer from MultiSynTech, Syro I) using the Fmoc (fluorenylmethoxycarbonyl)/t-Bu (tert-butyl) strategy on Rinkamid resin (loading 0.48 mmol, 0.015 mmol scale). First, Fmoc-Lys(DNP)-OH was manually loaded on the resin, and subsequently, the resin was elongated by the peptide synthesizer. All amino acids were coupled using 8 eq each of N,N′-diisopropylcarbodiimide (DIC), and ethyl cyanohydroxyiminoacetate (Oxyma) dissolved in DMF. Fmoc was cleaved off using 20% piperidine in DMF. The final peptides were cleaved by treating the resin for 3 h at room temperature with a mixture of concentrated (conc.) trifluoroacetic acid (TFA)/triisopropylsilane (TIS)/H2O (95:2.5:2.5, v/v/v). Following, the peptides were precipitated in ice-cold diethyl ether, washed five times, and freeze-dried. Purification was achieved by preparative reversed-phase (RP)-HPLC (Hitachi Elite LaChrom) on a VP250/16 NUCLEODUR 100-5 C18ec column (Macherey Nagel) using linear gradients from 20% to 70% B in A [A = 0.1% TFA in water; B = 0.1% TFA in acetonitrile (ACN)] over 60 min. All peptides were identified by RP-HPLC-ESI-MS (column: CC 125/4.6 Nucleodur 100-5 C18e (Macherey-Nagel) using a gradient from 10‑60% of ACN in H2O with 0.1% formic acid (ESI-MS, Thermo Scientific LTQ-XL). Final purity of all compounds was >95%.
Furin cleavage assay using an internally quenched fluorogenic peptides
The activity of furin towards the fluorogenic substrate containing an N-terminal Abz group and a C-terminal Lys-DNP group was measured by the increase in fluorescence at 420 nm that occurs upon cleavage of the peptide bond (|) in the peptide sequence Abz-VAKSRRVR | N(S/L)[-Lys[DNP]. Furin (New England Biolabs, #P8077) concentration was kept constant at a total of 1 unit in the total assay volume of 100 µl. Peptide concentrations were determined based on the absorption of the DNP moieties at 365 nM, calculated based on the extinction coefficient 17’300 M−1 cm−162, and used in concentrations between 1 and 2000 nM. The reaction was performed in 50 mM HEPES (pH 7.5), 10 mM CaCl, 1 mM β-Mercaptoethanol at 25 °C and measured in a Fluoromax 4 P (HORIBA) with the following settings: Excitation 320 nm (5 nm Bandpass); Emission 420 nm (5 nm Bandpass); Integration time 0.5 s; Interval time 1.0 s; total time 300 s. Initial velocity was determined by linear fit in Origin (OriginLabs) and plotted against substrate concentrations. The resulting titration curve was fitted by the Hill equation for evaluation of maximal velocity and approximate Km of the peptide. Representative data from three different assays are shown.
Manual lifespan assay
Gravid adults were treated with hypochlorite solution to obtain eggs. These were grown on OP50-seeded NGM plates until the L4 stage. At the L4 stage, 50 animals were transferred to OP50 NGM plates containing 50 μM FuDR, in four replica plates. For RNAi lifespan experiments, animals were grown on the respective RNAi NGM plates seeded with the corresponding bacteria (L4440/tir-1/pmk-1) until the L4 stage and then transferred to 50 μM FuDR-containing RNAi plates at the L4 stage. The survival of the animals was scored by prodding their tail once with a platinum wire every other day. Unhealthy animals showing vulval bursting or those crawling to the sides/lid of the plates were censored from the population. Lifespan Scoring was started on Day 7 of adulthood, L4 being taken as t = 0 timepoint. The survival curve was plotted using GraphPad Prism. The lifespans were performed in at least two biological replicates at 20 °C. Statistical analysis was performed on mean lifespans using the online software OASIS63. A lifespan summary for individual experiments is provided in Supplementary Table 2.
C. elegans Total RNA sample preparation for Transcriptomics
Synchronized populations of the strains Wild-type (N2), and col-99(gk694263[S106L]) SNP mutant were grown on OP50 until the young adult stage (approximately 10 × 90 mm NGM petri plates with 500–800 animals feeding on OP50 per plate). Young adult animals are referred to as day 1 adults shortly after the L4 stage, without the baggage of embryos. Young adult animals were collected in M9 buffer, washed twice with the buffer, and then flash frozen the pellet for storage at −80 °C until RNA isolation. RNA isolation was done using the Qiagen RNeasy mini kit (#74004) method according to the manufacturer’s protocol. Three biological replicates were collected for the experiment.
RNA sequencing and analysis
RNA concentration and purity were assessed via a Nanodrop spectrophotometer and an Agilent 2100 Bio-Analyzer. cDNA libraries for sequencing were prepared using the Illumina TruSeq RNA Library Prep Kit. Libraries were pooled and sequenced with 150 bp paired-end reads to a target depth of 20 million reads per sample. Reads were aligned to WBcel235 reference genome using the align function and annotated to the corresponding GTF using the featureCounts function from the Rsubread R package64 (v. 2.3.5). Differentially expressed genes were determined using the edgeR65 (v. 3.30.2) and limma66 (v. 3.44.2) R packages. Briefly, genes were filtered based on minimum expression level (>5 counts per million in at least 3 samples). Normalization was carried out using the trimmed mean of M-values method as implemented in calcNormFactors from edgeR. Data were modeled and differentially determined using the limma voom workflow to generate linear models with empirical Bayes moderation. Gene set overrepresentation analysis was performed using the enrichGO function from the clusterProfiler67 (v. 3.15.0) R package, using lists of genes with either significantly increased or decreased expression and compared to a background list of all genes detected in the dataset. Raw reads are available at the NIH Sequence Read Archive (SRA) under accession number PRJNA1198094.
Oxidative stress assay
Young adult wild-type or col-99(gk694263[S106L]) SNP mutant animals grown on OP50 NGM plates were collected in M9 buffer and washed three times with M9 buffer by centrifugation. Approx. 30–40 animals were pipetted into the wells of a 96-well round-bottom plate. Animals were either swimming in 30 μL of M9 buffer (filled in the first two rows of the 96-well round-bottom plate) or 14 mM sodium arsenite solution (in the rest of the 6 rows). Every column of the plate was denoted with one condition or strain. Stress resistance was measured as a function of the motility of the animals scored for 3 days in an automated wormtracker (wMicroTracker by NemaMetrix) platform. The resulting quantified motility output was analyzed and plotted using an R Script generated in our lab (Ewaldlab-LSD @ github). The experiment was done in three biological replicates.
Automated Heat Stress Survival Assay
Synchronized wild-type or col-99(gk694263[S106L]) SNP mutant populations were grown until the young adult stage on the OP50 NGM plate. At the young adult stage, animals were transferred to a 32 °C (or 35 °C) incubator to initiate heat stress. The incubator was equipped with scanners that were set to scan the plates every half an hour and capture the images of the animals for 3 days. Survival data of the animals was recorded and generated automatically after we manually decoded images of the animals from misleading object data points. The survival curve was plotted using an R script developed in our lab (Ewaldlab-LSD @ github). The experiment was done in two biological replicates at 32 °C and once at 35 °C.
Developmental time assay
Ten gravid adult Day 1 animals were allowed to lay eggs on OP50-seeded NGM plates for four hours and then sacrificed. After 48 h, the number of animals was estimated in different life cycle stages. Three plates per condition (technical replicates) were kept to assess the developmental stages. The total number of animals in a particular stage was plotted in a bar graph. Two-way ANOVA statistical analysis was performed in GraphPad Prism.
Fluorescence Microscopy
Gravid adults were treated with sodium hypochlorite solution to obtain eggs. The eggs were grown on OP50 or respective RNAi bacteria seeded on an NGM plate until young adults. Approx. 30 young adult animals were mounted onto a 2% agarose pad and anesthetized with 20 mM levamisole for imaging. For hsp-4p::GFP, young adults were heat-shocked at 33 °C for 2 h and imaged after recovery from shock for 6 h at 20 °C. Animals were stacked together and captured at 10X with one or two fields of view and then stitched together later using ImageJ. Quantifying the total fluorescence intensity per number of pixels was done by running a Python script, GreenIntensityCalculator (publicly available in Github-Ewaldlab), in ImageJ. Data is plotted as mean, and error bars represent SEM. Two-tailed Welch’s t-test statistical analysis of the quantified data was done using GraphPad Prism software (GraphPad Prism 9.0). The experiment was performed in three biological batches, with n > 80 total animals per condition.
DAF-16 nuclear localization assay
The synchronized population was grown until the L4 stage on OP50 NGM plates. At the L4 stage, approximately 30 animals were assessed manually for DAF-16 nuclear localization and categorized as none, low, intermediate, or high based on the intensity of the nuclear localization throughout the body. The percentage of animals showing DAF-16 nuclear localization was plotted in GraphPad Prism 9.0.
Western Blot
Young adult animals were harvested in the M9 buffer and washed three times with the M9 buffer to remove excess bacteria. The pellet was flash-frozen and stored at −80 °C until further use. To isolate protein from the animals, the pellet was resuspended in Buffer ‘C’ solution (50 mM HEPES, 100 mM KCl, 1 mM MgCl2, 1 mM EGTA, 10% glycerol) with protease inhibitor cocktail (#11697498001, Merck) and then subjected to the homogenizer. The tube was then centrifuged to pellet the insoluble fraction and separate the soluble supernatant in another tube. Protein in the sample was estimated using the Bradford assay (#5000006, Bio-Rad). Approx. 30 μg of protein was run in duplicate lanes on SDS-PAGE gel. After blotting the protein on the PVDF membrane (#IPVH00010, Merck-Millipore), it was blocked in a 5% BSA blocking buffer for 1 h at room temperature while shaking. Primary P-PMK-1 antibody (#9215S, Cell Signalling Technology) was used in dilution 1:5000, and alpha-tubulin (#T9026, Sigma) was used in dilution 1:10,000 in the blocking buffer overnight at 4 °C. The next day, the blots were washed three times with 1X TBST buffer and put in secondary anti-rabbit (#7074S, Cell Signalling Technology) (for P-PMK-1) and anti-mouse (#7076S, Cell Signalling Technology) (for alpha-tubulin) antibodies in a dilution 1:10,000 for 1 h at room temperature. Then, the blots were washed three times with TBST buffer and developed in a chemiluminescent chamber. Clarity Western ECL Blotting Substrate (#1705060S, Bio-Rad) was used to develop the blot.
References
Borras, C. et al. Centenarians: An excellent example of resilience for successful ageing. Mech. Ageing Dev. 186, 111199 (2020).
Willcox, B. J., Willcox, D. C. & Suzuki, M. Demographic, phenotypic, and genetic characteristics of centenarians in Okinawa and Japan: Part 1—centenarians in Okinawa. Mech. Ageing Dev. 165, 75–79 (2017).
Deelen, J. et al. Genome-wide association meta-analysis of human longevity identifies a novel locus conferring survival beyond 90 years of age. Hum. Mol. Genet. 23, 4420–4432 (2014).
Sebastiani, P. & Perls, T. T. The genetics of extreme longevity: lessons from the New England Centenarian Study. Front. Genet 3, 277 (2012).
Puca, A. A. et al. A genome-wide scan for linkage to human exceptional longevity identifies a locus on chromosome 4. Proc. Natl. Acad. Sci. 98, 10505–10508 (2001).
Taylor, J., Unsoeld, T. & Hutter, H. The transmembrane collagen COL-99 guides longitudinally extending axons in C. elegans. Mol. Cell. Neurosci. 89, 9–19 (2018).
Teuscher, A. C. et al. The in-silico characterization of the Caenorhabditis elegans matrisome and proposal of a novel collagen classification. Matrix Biol. 1, 100001 (2019).
Tu, H., Huhtala, P., Lee, H.-M., Adams, J. C. & Pihlajaniemi, T. Membrane-associated collagens with interrupted triple-helices (MACITs): evolution from a bilaterian common ancestor and functional conservation in C. elegans. BMC Evol. Biol. 15, 281 (2015).
Erikson, G. A. et al. Whole-genome sequencing of a healthy aging cohort. Cell 165, 1002–1011 (2016).
Tong, Y., Xu, Y., Scearce-Levie, K., Ptáček, L. J. & Fu, Y.-H. COL25A1 triggers and promotes Alzheimer’s disease-like pathology in vivo. Neurogenetics 11, 41–52 (2010).
Hashimoto, T. CLAC: a novel Alzheimer amyloid plaque component derived from a transmembrane precursor, CLAC-P/collagen type XXV. EMBO J. 21, 1524–1534 (2002).
Tanaka, T. et al. CLAC-P/collagen type XXV is required for the intramuscular innervation of motoneurons during neuromuscular development. J. Neurosci.34, 1370–1379 (2014).
Gonçalves, T. J. M. et al. Collagen XXV promotes myoblast fusion during myogenic differentiation and muscle formation. Sci. Rep. 9, 5878 (2019).
Tian, S. A 20 residues Motif Delineates the Furin Cleavage Site and its Physical Properties May Influence Viral Fusion. Biochem. Insights 2, BCI.S2049 (2009).
Harada, H. et al. Extracellular phosphorylation drives the formation of neuronal circuitry. Nat. Chem. Biol. 15, 1035–1042 (2019).
Yalak, G. & Vogel, V. Extracellular phosphorylation and phosphorylated proteins: not just curiosities but physiologically important. Sci. Signal. 5, re7 (2012).
Bordoli, M. R. et al. A secreted tyrosine kinase acts in the extracellular environment. Cell 158, 1033–1044 (2014).
Tagliabracci, V. S. et al. Secreted Kinase phosphorylates extracellular proteins that regulate biomineralization. Science 336, 1150–1153 (2012).
Yalak, G. & Olsen, B. R. Proteomic database mining opens up avenues utilizing extracellular protein phosphorylation for novel therapeutic applications. J. Transl. Med. 13, 125 (2015).
Tagliabracci, V. S. et al. Dynamic regulation of FGF23 by Fam20C phosphorylation, GalNAc-T3 glycosylation, and furin proteolysis. Proc. Natl. Acad. Sci. USA. 111, 5520–5525 (2014).
Fletcher, M., Tillman, E. J., Butty, V. L., Levine, S. S. & Kim, D. H. Global transcriptional regulation of innate immunity by ATF-7 in C. elegans. PLOS Genet. 15, e1007830 (2019).
McQuary, P. R. et al. C. elegans S6K Mutants Require a Creatine-Kinase-like Effector for Lifespan Extension. Cell Rep. 14, 2059–2067 (2016).
Ju, S. et al. C. elegans monitor energy status via the AMPK pathway to trigger innate immune responses against bacterial pathogens. Commun. Biol. 5, 643 (2022).
Kitisin, T., Muangkaew, W. & Sukphopetch, P. Caenorhabditis elegans DAF-16 regulates lifespan and immune responses to Cryptococcus neoformans and Cryptococcus gattii infections. BMC Microbiol. 22, 162 (2022).
Shivers, R. P., Kooistra, T., Chu, S. W., Pagano, D. J. & Kim, D. H. Tissue-specific activities of an immune signaling module regulate physiological responses to pathogenic and nutritional bacteria in C. elegans. Cell Host Microbe 6, 321–330 (2009).
Troemel, E. R. et al. p38 MAPK Regulates Expression of Immune Response Genes and Contributes to Longevity in C. elegans. PLoS Genet. 2, e183 (2006).
Liberati, N. T. et al. Requirement for a conserved Toll/interleukin-1 resistance domain protein in the Caenorhabditis elegans immune response. Proc. Natl. Acad. Sci. 101, 6593–6598 (2004).
Shivers, R. P. et al. Phosphorylation of the conserved transcription factor ATF-7 by PMK-1 p38 MAPK regulates innate immunity in Caenorhabditis elegans. PLoS Genet. 6, e1000892 (2010).
Pujol, N. et al. A reverse genetic analysis of components of the Toll signaling pathway in Caenorhabditis elegans. Curr. Biol. 11, 809–821 (2001).
Xiao, R. et al. A genetic program promotes C. elegans longevity at cold temperatures via a thermosensitive TRP channel. Cell 152, 806–817 (2013).
Riera, C. E. et al. TRPV1 pain receptors regulate longevity and metabolism by neuropeptide signaling. Cell 157, 1023–1036 (2014).
Ewald, C. Y. & Nyström, A. Mechanotransduction through hemidesmosomes during aging and longevity. J. Cell Sci. 136, jcs260987 (2023).
Teuscher, A.C. et al. Longevity interventions modulate mechanotransduction and extracellular matrix homeostasis in C. elegans. Nat. Commun. 15, 276 (2024).
Iwasa, H. et al. Yes-associated protein homolog, YAP-1, is involved in the thermotolerance and aging in the nematode Caenorhabditis elegans. Exp. Cell Res. 319, 931–945 (2013).
Lee, H., Kang, J. & Lee, J. Involvement of YAP-1, the Homolog of Yes-Associated Protein, in the Wnt-Mediated Neuronal Polarization in Caenorhabditis elegans. G3 Genes Genomes Genetics 8, 2595–2602 (2018).
Ma, Y.-C. et al. YAP in epithelium senses gut barrier loss to deploy defenses against pathogens. PLOS Pathog. 16, e1008766 (2020).
Giuliani, C. et al. Impact of demography and population dynamics on the genetic architecture of human longevity. Aging 10, 1947–1963 (2018).
Walter, S. et al. A genome-wide association study of aging. Neurobiol. Aging 32, 2109.e15–28 (2011).
Sebastiani, P. et al. Four genome-wide association studies identify new extreme longevity variants. J. Gerontol. A. Biol. Sci. Med. Sci. 72, 1453–1464 (2017).
Zhang, J. et al. Mechanotransduction and Cytoskeleton Remodeling Shaping YAP1 in Gastric Tumorigenesis. Int. J. Mol. Sci. 20, 1576 (2019).
Santos-Lozano, A. et al. Successful aging: insights from proteome analyses of healthy centenarians. Aging 12, 3502–3515 (2020).
Karagiannis, T. T. et al. Multi-modal profiling of peripheral blood cells across the human lifespan reveals distinct immune cell signatures of aging and longevity. eBioMedicine 90, 104514 (2023).
Sebastiani, P. et al. Protein signatures of centenarians and their offspring suggest centenarians age slower than other humans. Aging Cell 20, e13290 (2021).
Forsell, C. et al. Genetic association to the amyloid plaque associated protein gene COL25A1 in Alzheimer’s disease. Neurobiol. Aging 31, 409–415 (2010).
Wang, Z. et al. Deep post-GWAS analysis identifies potential risk genes and risk variants for Alzheimer’s disease, providing new insights into its disease mechanisms. Sci. Rep. 11, 20511 (2021).
Natera-de Benito, D. et al. Recessive variants in COL25A1 gene as novel cause of arthrogryposis multiplex congenita with ocular congenital cranial dysinnervation disorder. Hum. Mutat. 43, 487–498 (2022).
Shinwari, J. M. A. et al. Recessive mutations in COL25A1 are a cause of congenital cranial dysinnervation disorder. Am. J. Hum. Genet. 96, 147–152 (2015).
Munezane, H. et al. Roles of Collagen XXV and its putative receptors PTPσ/δ in Intramuscular motor innervation and congenital cranial dysinnervation disorder. Cell Rep. 29, 4362–4376.e6 (2019).
Fortney, K. et al. Genome-wide scan informed by age-related disease identifies loci for exceptional human longevity. PLOS Genet. 11, e1005728 (2015).
Giuliani, C. et al. Centenarians as extreme phenotypes: An ecological perspective to get insight into the relationship between the genetics of longevity and age-associated diseases. Mech. Ageing Dev. 165, 195–201 (2017).
Tesi, N. et al. Polygenic risk score of longevity predicts longer survival across an age continuum. J. Gerontol. A. Biol. Sci. Med. Sci. 76, 750–759 (2021).
Ying, K. et al. Depletion of loss-of-function germline mutations in centenarians reveals longevity genes. Nat. Commun. 15, 9030 (2024).
Möller, S. et al. Healthspan pathway maps in C. elegans and humans highlight transcription, proliferation/biosynthesis and lipids. Aging 12, 12534–12581 (2020).
Słomian, D. et al. Better safe than sorry-Whole-genome sequencing indicates that missense variants are significant in susceptibility to COVID-19. PloS One 18, e0279356 (2023).
Moore, A. et al. Genome-wide metabolite quantitative trait loci analysis (mQTL) in red blood cells from volunteer blood donors. J. Biol. Chem. 298, 102706 (2022).
Madeo, F., Carmona-Gutierrez, D., Kepp, O. & Kroemer, G. Spermidine delays aging in humans. Aging 10, 2209–2211 (2018).
Unsoeld, T., Park, J.-O. & Hutter, H. Discoidin domain receptors guide axons along longitudinal tracts in C. elegans. Dev. Biol. 374, 142–152 (2013).
Touroutine, D. & Tanis, J. E. A rapid, superselective method for detection of single nucleotide variants in Caenorhabditis elegans. Genetics 216, 343–352 (2020).
Blom, N., Gammeltoft, S. & Brunak, S. Sequence and structure-based prediction of eukaryotic protein phosphorylation sites. J. Mol. Biol. 294, 1351–1362 (1999).
Blom, N., Sicheritz-Pontén, T., Gupta, R., Gammeltoft, S. & Brunak, S. Prediction of post-translational glycosylation and phosphorylation of proteins from the amino acid sequence. Proteomics 4, 1633–1649 (2004).
Stroustrup, N. et al. The Caenorhabditis elegans Lifespan Machine. Nat. Methods 10, 665–670 (2013).
Karaseva, M. A. et al. An internally quenched fluorescent peptide substrate for Protealysin. Sci. Rep. 9, 14352 (2019).
Han, S. K. et al. OASIS 2: online application for survival analysis 2 with features for the analysis of maximal lifespan and healthspan in aging research. Oncotarget 7, 56147–56152 (2016).
Liao, Y., Smyth, G. K. & Shi, W. The R package Rsubread is easier, faster, cheaper and better for alignment and quantification of RNA sequencing reads. Nucleic Acids Res. 47, e47 (2019).
Chen, Y., Lun, A. T. L. & Smyth, G. K. From reads to genes to pathways: differential expression analysis of RNA-Seq experiments using Rsubread and the edgeR quasi-likelihood pipeline. F1000Research 5, 1438 (2016).
Ritchie, M. E. et al. limma powers differential expression analyses for RNA-sequencing and microarray studies. Nucleic Acids Res. 43, e47 (2015).
Yu, G., Wang, L.-G., Han, Y. & He, Q.-Y. clusterProfiler: an R package for comparing biological themes among gene clusters. Omics J. Integr. Biol. 16, 284–287 (2012).
Acknowledgements
We thank Garif Yalak for help with the collagen phosphorylation predictions, Valerio Izzi for sharing his analysis of the COL25A1 variants in LUSC patients, Harald Hutter for the col-99, ddr-1, and ddr-2 mutants, and WormBase for curated gene and phenotype information. Some strains were provided by the CGC, which is funded by the NIH Office of Research Infrastructure Programs (P40 OD010440). Funding from the Swiss National Science Foundation Funding from the Swiss National Science Foundation SNF P3 Project 190072 to CYE, CS, and AG.
Author information
Authors and Affiliations
Contributions
All authors participated in analyzing and interpreting the data. C.Y.E. and A.G. designed the C. elegans experiments. J.M.G. and I.N. performed in vitro furin cleavage assays. M.M. performed the analysis of RNA sequencing. CS performed the automated lifespan analysis. J.Y.C.P. performed a few col-99(gk694263[S106L]) S.N.P. crosses and conducted some reporter assays. A.G. performed all other lifespan assays and other experiments. A.G., J.Y.C.P., J.M.G. and C.Y.E. wrote the manuscript in consultation with the other authors.
Corresponding author
Ethics declarations
Competing interests
The authors declare the following personal relationship, which may be considered as potential competing interests: co-author Michael R. MacArthur and Sarah Mitchell, Editor-in-Chief of npj Aging, are married. Dr. Sarah Mitchell was not involved in the journal’s review of, or decisions related to, this manuscript. With no relation to the present manuscript, CYE declares to be a co-founder and shareholder of Avea Life AG and Lichi3 GmbH and is employed by Novartis. The other authors declare no competing interests.
Additional information
Publisher’s note Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.
Rights and permissions
Open Access This article is licensed under a Creative Commons Attribution-NonCommercial-NoDerivatives 4.0 International License, which permits any non-commercial use, sharing, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence, and indicate if you modified the licensed material. You do not have permission under this licence to share adapted material derived from this article or parts of it. The images or other third party material in this article are included in the article’s Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the article’s Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this licence, visit http://creativecommons.org/licenses/by-nc-nd/4.0/.
About this article
Cite this article
Goyala, A., Statzer, C., Park, J.Y.C. et al. A centenarian single nucleotide polymorphism in collagen gene COL25A1 promotes longevity in C. elegans. npj Aging 11, 81 (2025). https://doi.org/10.1038/s41514-025-00264-7
Received:
Accepted:
Published:
DOI: https://doi.org/10.1038/s41514-025-00264-7
