HINT2

Summary: IBA annotation for cytoplasm based on phylogenetic inference. While the IBA includes cytoplasmic HINT family members (e.g., HINT1 is cytoplasmic), human HINT2 is specifically characterized as a mitochondrial protein. PMID:16762638 demonstrates clear mitochondrial localization via immunofluorescence, fractionation, and GFP-fusion. The cytoplasmic annotation may reflect family-level inference that does not apply to HINT2 specifically.

Reason: Human HINT2 has an N-terminal mitochondrial targeting sequence and experimental evidence demonstrates exclusive mitochondrial localization. Unlike the cytoplasmic HINT1, HINT2 functions within mitochondria. The IBA inference from the broader HINT family does not apply to this paralog.

Summary: IBA annotation for mitochondrion supported by phylogenetic inference including mouse Hint2 ortholog. This is strongly supported by experimental evidence from PMID:16762638 demonstrating mitochondrial localization by immunofluorescence co-localization with MitoTracker, mitochondrial fractionation, and GFP-fusion constructs.

Reason: IBA annotation is well-supported by experimental literature. HINT2 has a validated N-terminal mitochondrial targeting sequence (~35 aa) and multiple lines of experimental evidence confirm mitochondrial localization.

Summary: IEA annotation based on automated transfer from mouse ortholog (UniProtKB:Q9D0S9). This is correct and consistent with experimental evidence from PMID:16762638.

Reason: Automated annotation is correct. The orthologous mouse Hint2 and human HINT2 both localize to mitochondria, as established by direct experimental evidence.

Summary: IDA annotation from HPA (Human Protein Atlas) based on immunofluorescence data. This is consistent with the primary literature (PMID:16762638) showing mitochondrial localization.

Reason: HPA immunofluorescence data confirms mitochondrial localization, consistent with the original characterization study.

Summary: HTP annotation from high-throughput mitochondrial proteome study (Morgenstern et al., 2021). This large-scale quantitative proteomics study provides high-confidence identification of HINT2 in the human mitochondrial proteome.

Reason: High-throughput proteomics confirms mitochondrial localization, consistent with low-throughput experimental studies.

Summary: Primary experimental evidence from Martin et al. (2006) demonstrating HINT2 mitochondrial localization via immunofluorescence with MitoTracker co-localization, mitochondrial fractionation, and GFP-fusion constructs. This is the foundational characterization of HINT2 subcellular localization.

Reason: Gold standard experimental evidence establishing HINT2 as a mitochondrial protein. Multiple independent methods confirm localization.

Summary: ISS annotation based on sequence similarity to mouse Hint2 (UniProtKB:Q9D0S9). This is valid as both orthologs localize to mitochondria.

Reason: Sequence similarity-based transfer is appropriate given the conserved N-terminal mitochondrial targeting sequence and experimental validation of localization.

Summary: IEA annotation for mitochondrial matrix based on Ensembl Compara transfer from mouse ortholog. This is consistent with the N-terminal mitochondrial targeting sequence (aa 1-17 per UniProt) which would direct the mature protein to the matrix after import and cleavage. The Reactome annotation suggesting outer membrane localization appears to be erroneous (see review of GO:0005741).

Reason: The mitochondrial matrix localization is well-supported by the protein's N-terminal mitochondrial targeting sequence. UniProt annotates residues 1-17 as a transit peptide for mitochondrial import. After cleavage, the mature protein (aa 18-163) would be expected to reside in the matrix. This is consistent with HINT2's role in regulating mitochondrial NAD+ homeostasis and SIRT3-dependent deacetylation, as SIRT3 is a matrix-localized sirtuin. The Reactome outer membrane annotation appears to be erroneous and should not be used to question the matrix localization. The PMID:18653718 study also noted that HINT2 appeared "associated with mitochondrial membranes, probably facing the interior of the organelle" - consistent with matrix or inner membrane localization.

Summary: TAS annotation based on Reactome pathway R-HSA-9693214 claiming RHOD binds HINT2 at the mitochondrial outer membrane to trigger Ca2+ influx. The Reactome entry cites "Chen et al. 2017" and "Bagci et al. 2020". However, this annotation conflicts with multiple lines of evidence showing HINT2 has an N-terminal mitochondrial targeting sequence (aa 1-17 per UniProt) that directs the mature protein to the matrix. The Reactome interpretation may be erroneous or conflating HINT2 with a different protein.

Reason: This annotation should be removed for several reasons. (1) HINT2 has a well-characterized N-terminal mitochondrial targeting sequence (aa 1-17 per UniProt FT TRANSIT) that directs the protein to the mitochondrial matrix, not the outer membrane. (2) The primary characterization study (PMID:16762638) localized HINT2 to mitochondria but did not demonstrate outer membrane localization specifically. (3) The Reactome citation to "Chen et al. 2017" for RHOD-HINT2 interaction triggering Ca2+ influx is questionable - more recent literature (e.g., PMC8677646) shows HINT2 interacts with MCU (mitochondrial calcium uniporter) which is in the inner membrane/matrix, not the outer membrane. The outer membrane localization may be a curation error conflating the Ca2+ signaling function with outer membrane localization. HINT2's N-terminal targeting sequence is incompatible with outer membrane localization of the mature protein.

Summary: IBA annotation for hydrolase activity based on phylogenetic inference across the HIT family. This is correct but overly general. HINT2 has specifically characterized adenosine 5'-monophosphoramidase activity (GO:0043530) which is a more informative term.

Reason: While hydrolase activity is not incorrect, the more specific term GO:0043530 (adenosine 5'-monophosphoramidase activity) is experimentally established with detailed kinetics (kcat = 0.0223 s-1, Km = 128 uM) and should be preferred.

Proposed replacements: adenosine 5'-monophosphoramidase activity

Summary: IEA annotation for nucleotide binding based on UniProt keyword mapping. This is true but uninformative for a nucleotide-processing enzyme. The nucleotide (AMP) is a product of the enzymatic reaction, not a regulatory binding site.

Reason: While HINT2 does bind nucleotides as part of its catalytic mechanism, this is inherent to its enzymatic function. The term "nucleotide binding" alone is uninformative and the more specific catalytic function (GO:0043530) better captures the biological role.

Summary: IEA annotation for generic catalytic activity based on InterPro domain mapping (HIT-like domain IPR011146). This is correct but extremely vague. The specific enzymatic activity (adenosine 5'-monophosphoramidase activity) is well-characterized.

Reason: Generic "catalytic activity" should be replaced with the specific enzymatic function GO:0043530 (adenosine 5'-monophosphoramidase activity) which is experimentally validated.

Proposed replacements: adenosine 5'-monophosphoramidase activity

Summary: IEA annotation for hydrolase activity based on UniProt keyword mapping. This is a duplicate of the concept captured by the IBA annotation. The term is correct but too general.

Reason: Should use the more specific GO:0043530 (adenosine 5'-monophosphoramidase activity) which is experimentally established.

Proposed replacements: adenosine 5'-monophosphoramidase activity

Summary: IEA annotation for the specific adenosine 5'-monophosphoramidase activity based on automated annotation (ARBA rule and Rhea reaction). This correctly captures the primary enzymatic function of HINT2.

Reason: Correct specific annotation for HINT2's primary molecular function. This is the core enzymatic activity with well-characterized kinetics (kcat = 0.0223 s-1, Km = 128 uM for AMP-pNA substrate).

Summary: IPI annotation for protein binding with CFTR based on high-throughput mammalian membrane two-hybrid (MaMTH-HTS) screening. HINT2 is a mitochondrial matrix/inner membrane protein, while CFTR is a plasma membrane chloride channel. The physiological relevance of this interaction is highly questionable given the distinct subcellular compartments.

Reason: This annotation should be removed because (1) "protein binding" is an uninformative term that does not specify the nature or consequence of the interaction, (2) the interaction was detected in a high-throughput artificial system (membrane two-hybrid) that may not reflect physiological conditions, and critically (3) HINT2 and CFTR localize to completely different subcellular compartments (mitochondria vs. plasma membrane), making physiological interaction extremely unlikely. HINT2 has an N-terminal mitochondrial targeting sequence and multiple studies confirm its mitochondrial localization (PMID:16762638). CFTR is a plasma membrane chloride channel. There is no plausible biological context where these proteins would interact in vivo. High-throughput protein-protein interaction screens frequently produce false positives, and this appears to be one.

Summary: Primary experimental evidence from Martin et al. (2006) establishing HINT2 as an adenosine 5'-monophosphoramidase. Using the model substrate AMP-pNA, they determined kcat = 0.0223 s-1 and Km = 128 uM. The active site histidine (H149) was confirmed by mutagenesis - H149A mutation abolished enzymatic activity.

Reason: Gold standard experimental evidence defining HINT2's core enzymatic function. This is the primary molecular function of HINT2 with detailed kinetic characterization and active site validation.

Summary: Experimental evidence from Strom et al. (2020) comparing substrate specificities between HINT1 and HINT2. They confirmed similar phosphoramidate hydrolysis activity for both enzymes, with HINT2 showing higher kcat and Km values compared to HINT1. Also demonstrated a broader pH range for maximum catalytic activity (pK1 = 6.76, pK2 = 8.41).

Reason: Confirmatory experimental evidence for HINT2's adenosine 5'-monophosphoramidase activity with additional characterization of pH dependence and substrate specificity.

Summary: IEA annotation for lipid metabolic process based on UniProt keyword mapping (KW-0443 Lipid metabolism). This is derived from correlative evidence in PMID:18653718 where HINT2 knockdown affected steroidogenic response in H295R adrenocortical cells. The connection to lipid metabolism is indirect and mechanistically unclear - the authors explicitly state the mechanism "still remains to determine at the molecular level."

Reason: The evidence linking HINT2 to lipid metabolism is phenotypic and correlative, not mechanistic. PMID:18653718 shows that HINT2 knockdown reduced steroidogenic response, but the authors acknowledge the mechanism is unknown. More recent work (Wang et al. 2025) shows HINT2 deficiency exacerbates hepatic steatosis through effects on mitochondrial NAD+/SIRT3 pathways, but this represents an indirect consequence of disrupted mitochondrial function rather than direct involvement in lipid metabolic enzymes or pathways. HINT2's primary function is adenosine 5'-monophosphoramidase activity - there is no evidence it directly participates in lipid metabolic reactions. The term "lipid metabolic process" is too broad and implies direct involvement that is not supported.

Summary: IEA annotation for steroid biosynthesis based on UniProt keyword mapping (KW-0752). This is derived from PMID:18653718 which showed that HINT2 knockdown in adrenocortical H295R cells reduced pregnenolone secretion in response to angiotensin II, K+, forskolin, or hydroxycholesterol stimulation. However, the mechanism is explicitly stated as unknown, and the effect appears to be through mitochondrial calcium signaling and membrane potential rather than direct participation in steroidogenic enzymes.

Reason: The evidence from Lenglet et al. (2008) is correlative and the mechanism is explicitly unknown. The abstract states that HINT2's effects "could be related to its ability to maintain a favorable mitochondrial potential" and that the mechanism "still remains to determine at the molecular level." Importantly, HINT2 overexpression had "no effect" on steroidogenesis - only knockdown showed effects, suggesting HINT2 is not rate-limiting but rather maintains basal mitochondrial function required for optimal steroidogenic capacity. HINT2 is NOT a steroidogenic enzyme (like CYP11A1, HSD3B, etc.) - it is an adenosine 5'-monophosphoramidase. The steroidogenic effects are likely secondary to disruption of mitochondrial calcium handling or membrane potential rather than direct involvement in steroid biosynthetic pathways. This represents classic over-annotation from phenotypic observations.

Summary: IEA annotation for apoptotic process based on UniProt keyword mapping (KW-0053). The original evidence from PMID:16762638 shows HINT2 overexpression sensitizes hepatoma cells to Fas-mediated apoptosis (reduced viability, enhanced mitochondrial depolarization, increased caspase cleavage). However, this is a PHENOTYPIC observation in a specific cell model, not evidence of direct mechanistic involvement in the apoptotic machinery. The authors describe HINT2 as an "apoptotic sensitizer" - meaning it modulates sensitivity to apoptotic stimuli, likely through effects on mitochondrial function (membrane potential, NAD+ homeostasis), rather than being a core component of apoptotic signaling.

Reason: The evidence from Martin et al. (2006) demonstrates that HINT2 modulates cellular sensitivity to apoptotic stress, but does not establish HINT2 as a direct participant in apoptotic signaling pathways. The observed effects (mitochondrial depolarization, caspase activation) are downstream consequences of HINT2's role in mitochondrial function, not evidence of direct apoptotic function. The abstract explicitly states the "biological function" of HINT2 is "uncertain" and classifies it as an "apoptotic sensitizer" rather than an apoptotic regulator. More recent studies (Wang et al. 2025, Fan et al. 2020) link HINT2 to mitochondrial NAD+ homeostasis and SIRT3-dependent deacetylation, suggesting the apoptotic phenotype is secondary to bioenergetic roles. GO annotation for "apoptotic process" implies direct involvement, which is not supported by the mechanistic evidence.