Tumor suppressor candidate 3 (TUSC3) is a protein that in humans is encoded by the TUSC3 gene on chromosome 8p22.
Structurally, TUSC3 is an integral membrane protein localized to the endoplasmic reticulum (ER) that contains an N-terminal thioredoxin-like domain.
Functionally, TUSC3 has roles in N-linked glycosylation and ER quality control. It acts as an accessory subunit of the oligosaccharyltransferase (OST) complex, modulating N-linked glycosylation of a subset of nascent proteins. TUSC3 also functions at a later stage of N-glycan processing. A July 2026 study found that it also regulates G2-to-G1 glucose trimming on BMP/Dpp in mouse embryonic fibroblasts and Drosophila, acting as a substrate-specific regulator of ER quality control.
In November 2025, research established that TUSC3 functions as a key regulator of ER magnesium homeostasis, assembling into a functional complex with the ERMA (TMEM94) magnesium transporter, which is required for efficient magnesium uptake into the ER lumen. Loss-of-function variants in TUSC3 lead to ER magnesium depletion, which initiates a cellular stress response now understood to be the primary driver of TUSC3-related intellectual disability.
TUSC3 has been studied as a candidate tumor suppressor gene in human cancers. As of 2026, cancer risk has not been reported in individuals with germline biallelic TUSC3 variants.
Structure TUSC3 is a multi-pass transmembrane protein of the endoplasmic reticulum (ER). It consists of a large N-terminal domain that faces the ER lumen and a C-terminal region containing four transmembrane helices that anchor the protein in the membrane. The luminal domain adopts a thioredoxin-like fold containing a conserved redox-active Cys-X-X-Cys motif and a substrate-binding groove, which mediates transient interactions with target proteins during both N-glycosylation and subsequent G2-to-G1 glucose trimming. The C-terminal transmembrane bundle allows TUSC3 to be incorporated into the STT3B-containing oligosaccharyltransferase (OST) complex as one of two alternative accessory subunits (the other being MAGT1). The same transmembrane region also enables incorporation of TUSC3 into a complex with the ER magnesium transporter ERMA (TMEM94).
Function
N-linked glycosylation N-linked glycosylation is the enzymatic process in which a preassembled oligosaccharide (sugar chain) is transferred to specific asparagine residues within Asn-X-Ser/Thr consensus sequences of nascent polypeptides. This process occurs within the endoplasmic reticulum (ER) lumen and is catalyzed by the oligosaccharyltransferase (OST) complex. Historically, primary disruptions to this pathway have been classified as congenital disorders of glycosylation (CDG).
Mechanism Mammalian cells express two functional variants of the OST complex: STT3A, which operates co-translationally at the translocon, and STT3B, which operates post-translationally to glycosylate acceptor sites skipped by STT3A. TUSC3, together with its paralog MAGT1, selectively associates with the STT3B complex.
TUSC3 and MAGT1 share high sequence identity and display functional redundancy within the STT3B complex. Simultaneous loss of both TUSC3 and MAGT1 severely impairs STT3B-dependent glycosylation, causing accumulated misfolded proteins in the ER, activation of the unfolded protein response, and ER stress-induced apoptosis. However, knockout of TUSC3 alone does not cause systemic N-glycan loss or alter serum glycoprotein patterns in most tissues, as MAGT1 compensates for its absence. Thus, loss of TUSC3 does not produce a systemic glycosylation defect.
Although TUSC3 and MAGT1 show functional redundancy in vitro, their endogenous expression patterns differ significantly in vivo: MAGT1 is predominantly expressed in immune cells, whereas TUSC3 expression is highly enriched in the central nervous system. As a result, the loss of TUSC3 cannot be adequately compensated for in brain tissue due to insufficient local MAGT1 expression, leading to tissue-specific neurological pathology.
ER quality control and glucose trimming
Beyond its role in initial glycan transfer, TUSC3 functions at a post-translational stage of N-glycan processing to regulate ER quality control (ERQC). TUSC3 acts as a substrate-specific regulator that promotes G2-to-G1 glucose trimming on target glycoproteins, such as bone morphogenetic protein 4 (BMP4) and its Drosophila ortholog Decapentaplegic (Dpp). By facilitating this trimming step, TUSC3 serves as a gatekeeper for substrate entry into the ERQC cycle, influencing whether target proteins successfully fold and enter the secretory pathway or are directed toward ER-associated degradation.
Magnesium homeostasis Beyond its association with the OST complex, TUSC3 serves an essential regulatory role in maintaining endoplasmic reticulum (ER) magnesium homeostasis. TUSC3 lacks an intrinsic ion-channel or transport pore. Instead, it forms an ER-localized protein complex with ERMA (ER Mg2+ ATPase, encoded by the TMEM94 gene), a P-type ATPase responsible for driving magnesium uptake into the ER lumen. Within this transport complex, TUSC3 regulates and enhances ERMA transport kinetics, acting as a required accessory subunit for efficient luminal magnesium reuptake.
Loss-of-function variants in TUSC3 result in a selective depletion of luminal ER magnesium without altering cytosolic magnesium levels. Because divalent cations are required for the proper activity of ER-resident chaperones, this localized deficit triggers chronic ER stress. Luminal magnesium depletion selectively activates the PERK-eIF2α pathway of the unfolded protein response, while leaving the IRE1-XBP1 and ATF6 pathways largely unaffected.
In neural tissues, where TUSC3 expression is highly enriched compared to peripheral tissues, sustained PERK activation selectively suppresses the translation of key postsynaptic proteins—including PSD-93, PSD-95, and GluA1—and impairs CREB phosphorylation. TUSC3 knockout mouse models display neurodevelopmental deficits characteristic of human autosomal recessive intellectual disability, including impaired spatial learning, memory, and social behavior. Fibroblasts derived from patients with TUSC3 mutations similarly exhibit ER magnesium deficiency and heightened baseline unfolded protein response activation.
Restoration of ER magnesium levels through targeted magnesium supplementation (such as magnesium L-threonate) mitigates PERK-mediated ER stress, restores synaptic protein translation, and reduces cognitive deficits in model systems. This evidence establishes ER magnesium dysregulation, rather than a primary glycosylation failure, as the primary pathogenic driver of TUSC3-associated neurodevelopmental pathology. Intellectual Disability Clinical presentation and genetics Biallelic loss-of-function variants in TUSC3 cause a form of non-syndromic autosomal recessive intellectual disability, historically designated as MRT7. Reported genetic alterations include loss of function point mutations (such as missense, nonsense, frameshift, and splice-site variants) as well as genomic microdeletions encompassing the TUSC3 locus on chromosome 8p22. Affected individuals typically present in early childhood with variable developmental delay, speech impairment, behavioral features (such as hyperactivity, autistic traits, or emotional dysregulation), and moderate-to-severe cognitive deficits. Because TUSC3 expression is highly enriched in neural tissues relative to peripheral tissues, systemic organ involvement, dysmorphic features, and metabolic abnormalities are generally absent.
At the cellular level, patient-derived fibroblasts and neuronal models exhibit reduced luminal endoplasmic reticulum (ER) magnesium concentrations and elevated baseline unfolded protein response signaling via the PERK pathway. Because this disease mechanism stems from ER magnesium transporter dysfunction rather than a primary disruption of protein glycosylation, standard biomarker screens for N-linked glycosylation defects—such as serum transferrin isoelectric focusing—are typically normal in affected individuals. Diagnosis is primarily established through genetic testing, such as whole exome sequencing or targeted gene panels, identifying biallelic variants or genomic deletions involving TUSC3. Historical classification as a congenital disorder of glycosylation Congenital disorder of glycosylation (CDG) was the first proposed disease association for TUSC3. Because TUSC3 encodes a non-catalytic accessory subunit of the endoplasmic reticulum (ER)-bound oligosaccharyltransferase (OST) complex—the research groups that independently identified TUSC3 mutations in patients with non-syndromic autosomal recessive intellectual disability in 2008 proposed that the condition represented a CDG. This classification, sometimes referred to as TUSC3-CDG, was based on TUSC3’s position within the glycosylation machinery rather than direct biochemical evidence of a glycosylation defect in patients.
The inference appeared reasonable at the time: most CDGs are identified through abnormal glycosylation patterns detectable on routine clinical testing, and several other CDGs cause neurodevelopmental impairment through disrupted glycosylation of brain proteins. However, both founding studies reported normal results on standard glycosylation screens in affected patients’ fibroblasts—an atypical finding for a CDG that left the actual disease mechanism unresolved for more than a decade.
Reassessment: from glycosylation to magnesium homeostasis A critical turning point came in 2009, when TUSC3–along with its paralog MAGT1–was identified as a major mediator of cellular magnesium influx, demonstrating that knockdown of TUSC3 significantly lowers intracellular magnesium concentrations. This discovery prompted a re-examination of the glycosylation hypothesis. A 2025 study examined N-glycosylation of several candidate substrates—including the magnesium transporter CNNM3 and the neuronal proteins NCAM1, N-cadherin, and GluA2—in TUSC3 knockout mice and found no detectable glycosylation defect in any of them, consistent with the normal clinical glycosylation screens reported in patients since 2008.
The same 2025 study provided the primary mechanistic link: TUSC3 acts through ERMA (TMEM94) to maintain ER magnesium levels. Loss of TUSC3 produces a luminal magnesium deficit selectively activates the PERK-eIF2α branch of the unfolded protein response, while leaving the IRE1-XBP1 and ATF6 pathways largely unaffected. This selective, chronic ER stress is especially consequential in neural tissues, where TUSC3 expression is highly enriched and local MAGT1 expression is insufficient to compensate. Sustained PERK activation selectively impairs synaptic translation. This suppresses the expression of key synaptic proteins—including PSD-93, PSD-95, and GluA1–and impairs CREB phosphorylation, driving the synaptic and cognitive deficits characteristic of TUSC3-related intellectual disability.
Building on these findings, 2026 research further clarified TUSC3’s role in ER quality control (ERQC). While TUSC3 is not required for global N-glycosylation, it functions downstream during N-glycan processing as a substrate-specific regulator that controls G2-to-G1 glucose trimming on specific targets such as BMP4/Dpp. These two distinct roles—magnesium homeostasis and limited, substrate-specific glycan processing—explain why loss of TUSC3 does not produce the systemic N-glycosylation failure seen in classic CDGs, but instead causes discrete signaling and ER stress dysregulation.