CHCHD10 has a defined human heterodimerization activity with CHCHD2 and a secondary, experimentally demonstrated nuclear corepressor activity. Its principal mitochondrial membrane-homeostasis role is supported, but the immediate molecular mechanism is not equivalent to an intrinsic protease, chaperone or constitutive MICOS scaffold. This review preserves all 22 original source objects, their qualifiers and partner identifiers; the normal seed has no alternative-product objects. One molecular-function proposal is added for transcription corepression. No new biological process is proposed.
Human HEK293, HeLa and primary-fibroblast experiments combine DSG/DSS crosslinking, separate gene knockouts, reciprocal protein detection and co-IP to resolve the CHCHD2–CHCHD10 heterodimer. This is more specific than a generic screen edge [PMID:30084972, "CHCHD2 and CHCHD10 form heterodimers"]. The original ND and three CHCHD2-binding assertions are refined using this separately cited evidence; their original codes and source papers remain unchanged. The older screen/co-IP papers are not redescribed as having measured dimer stoichiometry.
The inference is not that every tissue contains only a dimer. In 2025, native mouse-tissue complex analysis resolves a larger approximately 100–140 kDa species, and CHCHD10 still oligomerizes after Chchd2 loss. Human endogenous co-IP in that study corroborates the pair. Its lipid/stress perturbation concerns Chchd2 knockout and does not establish CHCHD10 lipid transfer PMID:41053020(https://pmc.ncbi.nlm.nih.gov/articles/PMC12501252/).
In human cells, loss of CHCHD2 and CHCHD10 changes OMA1-dependent OPA1 processing and cristae, and CHCHD10 re-expression improves the processing phenotype. Single and combined loss differ, and some disease variants retain activity. This supports the existing organization annotations without making CHCHD10 the OPA1-cleaving enzyme or a proven direct OMA1 inhibitor. The source itself leaves the immediate mechanism unresolved PMID:32338760(https://pmc.ncbi.nlm.nih.gov/articles/PMC7268789/). SLP2 association and PHB instability add a disease-context route, not proof that CHCHD10 itself is a PHB-like structural scaffold PMID:35656794(https://academic.oup.com/brain/article/145/10/3415/6599799).
The second core is contextual nuclear transcription corepression. Full human HEK293 experiments show repression with an intact but not mutated COX4I2 oxygen-responsive element, derepression after CHCHD10 knockdown, reciprocal CXXC5 association and CXXC5-dependent recruitment to the endogenous promoter. GO:0003714 permits activity through a DNA-bound transcription-regulatory complex; it does not require CHCHD10 to bind DNA directly. The nuclear IBA is retained as core-supported activity, while the historical PAINT placement remains unexamined. Wider tissue relevance is a biological question, not a reason to erase the measured activity PMID:29540477(https://pmc.ncbi.nlm.nih.gov/articles/PMC5925800/).
The same study also performs a positive purified-component experiment: adding recombinant CHCHD10 stimulates purified bovine COX, and human-cell depletion/re-expression and CHCHD2/MNRR1 dependence support respiratory regulation [PMID:29540477, "stimulation of COX activity"]. This positive result is included in the biological summary. GO:0008047's usage comment, however, requires direct enzyme interaction without covalent modification. The COX preparation contains submolar associated MNRR1/CHCHD10, human association is MNRR1-dependent, and the inspected main Methods do not resolve the recombinant CHCHD10 species or production recipe. The exact direct-enzyme-contact MF is therefore left as a question rather than asserted from an assumed binary mechanism. The proposed ABL2 recruitment model is not an experimentally demonstrated ternary bridge; no intrinsic kinase or COX catalysis is assigned to CHCHD10.
The original MICOS study is genuinely positive: native anti-mitofilin and reverse CHCHD10 co-IP, proximity ligation, and human/mouse BN-PAGE support association [PMID:26666268, "mitofilin was immunoprecipitated"]. That evidence is preserved in non-core binding. Later controlled reciprocal co-IP and detergent-dependent gradients separate most CHCHD10/CHCHD2 from mitofilin, while leaving possible transient association. Human native-complex experiments and the newer mouse complexomics also resolve distinct species PMID:29112723(https://pmc.ncbi.nlm.nih.gov/articles/PMC5886281/), PMID:29121267(https://pmc.ncbi.nlm.nih.gov/articles/PMC5886208/), PMID:41053020(https://pmc.ncbi.nlm.nih.gov/articles/PMC12501252/).
The MICOS part_of assertion is UNDECIDED because these actual complex assignments conflict. This is not a claim that components must be essential, permanently bound or purified to qualify. A narrow disputed finding records the membership issue without invalidating the older paper's association or morphology data.
The G66V source is particularly easy to misread. Its patient cells preserve cristae ultrastructure and MICOS assembly despite mitochondrial fragmentation, contrasting with S59L. The retained inner-membrane-organization role uses independent target evidence; G66V's negative ultrastructure result is never converted into a positive PMID:30092269(https://pmc.ncbi.nlm.nih.gov/articles/PMC7015038/). Human full-length S59L experiments and localization are separately established [PMID:24934289, "CHCHD10 is an intermembrane space protein"]. The disease models do not prove that all variants operate by simple loss of normal function.
Mouse Q7TNL9 is Chchd10, with current experimental membrane-organization support from PMID:28585542. That source separates endogenous mouse HT22 depletion from human tagged constructs expressed in mouse NIH3T3 hosts and human-cell experiments. The exact historical Ensembl/ISS transfer is not claimed to have been reconstructed. PAINT PTN000338319 remains a distinct unresolved historical-node provenance question; neither a short donor list nor target self-reference is criticized.
The source evidence extract preserves literal original assertions, the nine positive current IntAct records across two PMIDs, official term/identity fields, the mouse donor records, HPA location and source URLs/hashes. The two other binding papers have direct inspected primary assays even though the current whole-target IntAct query returned no exact matching record. Absence from that query is not an experimental negative. The 2018 T61I record is explicitly mutant-context evidence, not an untagged wild-type confirmation.
The selected PMID:20888800 cache is authentic abstract-only; full original location/organization experiments were not recovered. These correct existing functions are retained using independently inspected human evidence, not rejected from an incomplete source. The normal PMID:27499296 cache is only partial despite its flag, while separate genuine full PMC text was accessed. The exact-main PMID:30092269 cache and normal PMID:32338760/29112723 caches remain abstract-only; separately retrieved full primary text supports the assessments. For PMID:35656794, selected indexed original Methods and Results/Fig. 1 were read, not a successful local full-article download or all supplementary figures. Reference availability flags describe actual external access as well as cache access.
The isolated normal intake produced the immutable seed, UniProt and GOA. Genuine default Falcon research timed out; the explicit Perplexity-lite fallback failed with a quota/authentication response. No provider report was fabricated. This journal and the source ledger record manual primary research. Existing main caches were reused byte-for-byte; fresh variants remained separate. No source field, product scope or machine cache was edited to fit the review.
This supersedes the earlier species contrast: larger CHCHD10-containing complexes are observed in human cells as well as mouse tissues. Actual human control-fibroblast 2D-PAGE resolves approximately 170 and 220 kDa populations in glucose medium [PMID:29121267, "higher molecular weight complexes"; Results/Fig. 6A]. Culture in galactose changes additional complex populations. Crosslinking/knockout experiments independently establish human heterodimers, while their geometry does not exclude larger assemblies. The 2025 mouse study adds the differentiated-tissue 100–140 kDa distribution and persistence of CHCHD10 oligomerization after CHCHD2 loss, rather than establishing larger complexes for the first time. Electrophoretic size is not an exact subunit count.
The molecular-function ND root assertion is now REMOVE because positive molecular activities supersede the no-information statement. The original source fields remain exact. This is not an obsolete-ID workaround or a negative-function claim. The three binding refinements and core already cite the actual heterodimerization study, so no redundant NEW MF is introduced. The existing nuclear corepressor NEW remains unchanged.
Positive COX stimulation remains explicit. The heterodimer paper also discusses the [PMID:30084972, "indirect interaction between CHCHD10 and COX"; Discussion]. This corroborates the specific unresolved direct-contact question; it does not negate the measured stimulation. No additional direct enzyme-activator activity or respiratory process term is asserted from an assumed mechanism.
Thirteen authored-file self-quotation fields (nine unique strings) are removed while their bare provenance pointers remain. Primary excerpts and literal source records remain accessible; authored interpretations are not independent evidence. The four abstract-only selected caches remain unchanged, with actual external full or selected-text access described by their locators. Their optional full_text_unavailable flags are omitted to avoid implying that the cache itself contains those external passages. The full preceding notes remain byte-exact; this appendix supersedes earlier misleading complex-species and self-quotation wording.