Glycosylphosphatidylinositol (GPI) serves as a glycolipid anchor for numerous cell surface proteins in eukaryotes (1). In humans, ~150 GPI-anchored proteins (APs) have been identified, displaying roles such as enzymatic activity, antigen presentation, co-receptor engagement, adhesion and involvement in immune responses (1-3).

The core structure of GPI remains highly conserved across diverse organisms, featuring a PI molecule firmly integrated into the cell membrane. This structure is followed by glycans, namely glucosamine (GlcN) and three mannoses (Man) culminating in a terminal ethanolamine phosphate (EtNP) group. The latter binds covalently to the C-terminal end of the target protein (1,4). Variations in GPI backbone, involving modifications in EtNP and various glycan side branches, depend on the organism, cell type and specific protein (1,4). Notably, in mammalian cells, Man1 undergoes modification with an EtNP side chain and Man4 may attach to Man3 via α1,2-linkage (4,5).

To date, 30 genes have been identified encoding proteins implicated in GPI synthesis (4,6). These genes belong to the phosphatidylinositol-glycan biosynthesis class (PIG) and the post-GPI attachment to proteins (PGAP) families (3). The initiation of glycolipid biosynthesis occurs on the cytoplasmic side of the endoplasmic reticulum (ER). GPI N-acetylglucosamine transferase (GPI-GnT) catalyzes transfer of N-acetylglucosamine (GlcNAc) glycan to the PI molecule within the ER membrane, yielding GlcNAc-PI (1). Subsequent steps involve deacetylation, translocation to the ER lumen and acylation of inositol resulting in GlcN-acyl-PI and sequential addition of three Man. The addition of Man1 in ER is orchestrated by GPI-mannosyltransferase (MT)-I, anchored in the ER membrane and composed of the proteins PIGM and PIGX (Fig. 1) (1), where PIGM serves a functional role and PIGX as the stabilizing protein for PIGM enzyme (1).

Following GPI synthesis, GPI-MT-II encoded by PIGV links Man2 to Man-GlcN-acyl-PI, while GPI-MT-III encoded by PIGB adds a third Man yielding the glycolipid Man-Man-Man3-GlcN-acyl-PI (Fig. 1) (1,7). In addition, an EtNP molecule is integrated into the glycan, contributing to formation of the GPI core (1,7). The addition of Man4 as a side chain to Man3 is facilitated by GPI-MT-IV, encoded by PIGZ (1,7). Subsequently, the GPI molecule can bind to proteins featuring a C-terminal GPI attachment hydrophobic signal peptide, mediated by the GPI transamidase enzyme complex (GPI-TA) bound to the EtNP molecule (1). Synthesis proceeds in the Golgi apparatus, where GPI undergoes further lipid remodeling and glycan modification by PGAP enzymes (1,7-9). For example, the enzyme GPI-N-acetylgalactosamine transferase (GalNAc), encoded by the PGAP4 gene, modifies Man1 by adding a GalNAc (10). Ultimately, GPI-APs are transported to the plasma membrane via vesicular transport where they function within lipid rafts or are released into the extracellular space (8). GPI-APs exhibit diverse roles, including enzymatic activity, signaling, cell adhesion, cell wall metabolism, neuritogenesis and immune response (11).

Alterations in genes involved in GPI biosynthesis have been implicated in congenital anomalies such as multiple congenital anomalies-hypotonia-seizures, hyperphosphatasia with mental retardation and anomalies/epilepsy syndrome (3,12,13). In addition, certain types of cancer display altered expression in some PIG genes (14-16). Currently, little is known about the involvement of altered expression of PIGM and PIGX, genes that encode and regulate the GPI-MT-I, respectively, in human disease (17). The characteristics of the PIGM and PIGX genes and their encoded proteins are summarized in the present review, as well as the relevance of both genes in GPI synthesis and certain human health diseases, and their potential role in other biological functions.

PIGM is localized in chromosome 1q23, consists of 7,038 bp and encodes the transcript ENST00000368090.5, resulting in the protein PIGM Q9H3S5(18). The protein consists of 423 amino acid residues (2); the predicted structure by AlphaFold DB indicates that the tertiary structure consists of 10 transmembrane α-helices that alternate with 11 lumenal and cytoplasmic domains (Fig. 2A) (19). At amino acid positions 49-51, the protein harbors a sugar-binding motif, aspartate-any residue-aspartate (DXD) situated within a hydrophilic region flanked by the first and second transmembrane domains. The DXD motif is a prevalent feature in numerous glycosyltransferases and serves a pivotal role in coordinating a manganese ion, crucial for binding to a nucleotide sugar substrate (20). Notably, mutations in the DXD motif, such as the D51A alteration in PIGM, lead to the absence of GPI-APs on the cell surface, indicating the essential role of the DXD motif in expression of GPI-APs (20). Moreover, according to the predicted structure, all lumenal domains are comprised of the amino acid residues 39-79, 162-169, 247-287, 338 and 379-384, where PIGM should exert the catalytic activity (19); to the best of our knowledge however, there is no experimental evidence regarding the functional importance of these regions. There is no predicted site for phosphorylation (2) and thus far, binding to PIGX is the only mechanism proposed to regulate catalytic activity.

PIGX is 23,630-bp long and is located on chromosome 3q29(21). Currently, there is a total of nine known potential mRNA variants of PIGX resulting from alternative splicing (18). However, only two mRNA variants encode two protein isoforms (18). The protein isoform of PIGX Q8TBF5-1 is encoded by mRNA variant ENST00000392391.9, with a size of 258 amino acid residues and a mass of 28,788 Da (2). This isoform has been chosen as the canonical protein since it was first described (2,18). The amino acid residues 1-21 comprise the signal peptide that recognizes the protein as an ER membrane protein (Fig. 2B). The amino acid residues 22-230 form a soluble amino acid chain in the lumen of the ER, while the amino acid residues 231-251 are inserted in the ER membrane and amino acid residues in the C-terminal region, 252-258, are soluble in the cytoplasm (Fig. 2B) (2). PIGX has a non-ATG start codon and instead contains a CTG start codon that is well-conserved in mammals (18); non-ATG start codons are associated with key cellular functions such as development and stress responses (22). Regarding post-translational modifications, PIGX Q8TBF5-1 isoform harbors an N-glycosylation site at asparagine 103, phosphorylation site at serine 136(2) and two ubiquitination sites at lysines 66 and 82(23). By contrast, the isoform Q8TBF5-2 is encoded by the mRNA variant ENST00000296333.10 of PIGX. This isoform has a size of 276 amino acids and a mass of 30,974 Da, distinguishing it from the canonical sequence at positions 177-195, which contain the sequence QAGSRRMIRFRFDSFDKTI (Fig. 3) (2), and comprises the soluble chain in the lumen of the ER. As for the tertiary structure of PIGX, bioinformatic predictions in AlphaFold DB of the canonical protein (19) show that the soluble luminal amino acid chain in the ER consists of a random coil structure and β-sheets, while the transmembrane region is a single α-helix (Fig. 2B). Co-precipitation has demonstrated that PIGX is associated with PIGM (24). However, whether the transmembrane or the large luminal domain is implicated in stabilizing the PIGM protein remains unknown.

According to the Genotype-Tissue Expression project (25,26), there is mRNA expression of PIGM and PIGX in all major tissues (26,27), including the nervous system, heart, digestive system, skin and reproductive system in humans (Fig. 4A). Consistently, Proteomics DB reports the expression of both proteins (PIGX Q8TBF5-1) in the brain and the digestive and reproductive systems, but also in the breast, lung, retina, kidney and thyroid gland (Fig. 4B) (28). Notably, the expression of PIGM and PIGM has been detected in the colon, T lymphocytes, prostate and rectum (Fig. 4B) (28). These data suggest that GPI-MT-I is present in numerous types of tissues regardless of their specialized function in humans.

Mutant mammalian cells with deficiencies in genes implicated in the GPI-anchor biosynthesis have been previously reported, including those encoding GPI-GnT and GPI-TA (29), and the genes PIGV, PIGB, PIGM and PIGX (6,20,29-31). For example, in vitro experiments conducted using human lymphoma cells reveal that lack of PIGM results in elevated GlcN-acyl-PI levels and impaired surface expression of GPI-APs (Fig. 5) (20). Moreover, a recent study using a KO human cell library targeting GPI biosynthetic genes indicated that suppressing expression of specific PIG genes results in the absence of GPI-APs on the cell surface (6). Notably, KO of regulatory subunits of GPI-GnT leads to diminished presence of GPI-APs, while KO of catalytic subunits results in complete absence of these proteins (Fig. 5). Furthermore, elimination of GPI-AP synthesis occurs following KO of genes involved in steps subsequent to GPI-GnT activity. Regarding enzymes catalyzing the transfer of Man1 and 2, GPI-MT-I (PIGM) and GPI-MT-II (PIGV), complete removal of GPI-AP presence is observed upon PIGM KO, whereas residual presence of GPI-APs persists after PIGX KO (6). Experiments conducted in CHO cells derived from hamster adult ovaries demonstrated that defective PIGX leads to the accumulation of GlcN-acyl-PI, imitating the phenotype observed in PIGM-deficient cells (32). Additionally, the aforementioned study revealed diminished expression of protein PIGM in the absence of PIGX, while a 10-fold increase in expression was observed when PIGX was expressed. For these reasons, it is hypothesized that PIGX has a role in stabilizing PIGM (32).

Similar outcomes are associated with mutations affecting other genes involved in GPI synthesis. PIGV KO completely abolishes the surface expression of GPI-APs in human cells (Fig. 5). Conversely, PIGB KO, involved in transferring Man3, allows limited expression of GPI-APs, while PIGZ KO does not impact GPI-AP biosynthesis (6). In summary, the aforementioned studies demonstrate the essential role of enzymes in the initial stages of GPI synthesis, including GPI-MT-I, for expression of GPI-APs on the cell surface. Dysregulation of their expression may lead to alterations in cell surface characteristics.

The role of GPI-MT-I and II in GPI synthesis and secretion has been studied in hyperphosphatasia with mental retardation syndrome (HPMRS) (41-44). This autosomal recessive syndrome is characterized by intellectual disability and elevated serum and remodeling are associated with HPMRS, with the PIGV gene being well-studied (41-44). PIGV mutations include c.53G→A, c.176T→G, c.467G→A, c.905T→C, c.1022C→A and c.1405C→T and in some instances, mutations can be biallelic (42,43). Experiments using PIGV with transmembrane region mutations suggest that these mutations may destabilize the protein (41).

Cells deficient in PIGV and PIGB secrete GPI-APs into the medium, accompanied by the accumulation of incomplete GPI-bearing Man (41). This phenotype is explained by the fact that secretion of alkaline phosphatase (ALP) requires GPI-TA, which, in normal cells, cleaves the C-terminal GPI attachment signal peptide and replaces it with GPI. In PIGX-deficient cells, where incomplete shorter GPIs lacking Man accumulate, ALP is degraded. This suggests that at least one Man residue is required for GPI-TA to cleave the GPI attachment signal. Consequently, it is hypothesized that GPI-TA recognizes incomplete GPI-bearing mannose, cleaving a hydrophobic signal peptide, resulting in secretion of soluble ALP (Fig. 5) (41). Although HPMRS is not observed in patients with a PIGM mutation (41), the aforementioned study suggested a key step during GPI synthesis involving the enzymes GPI-MT-I (PIGM), PIGX and GPI-MT-II (PIGV) for proper recognition of GPI-APs by GPI-TA and subsequent secretion.

Studies indicated a crucial role of GPI-TA in regulating glycolipid biosynthesis, since GPI biosynthesis is suppressed by the ER-associated degradation pathway when cells are defective in transferring the complete GPI core to proteins to prevent GPI accumulation (9,45). Biosynthesis of GPI is upregulated in ER-associated degradation-deficient cells (46). Whether the ER-associated degradation pathway is triggered in PIGM-deficient cells or in cells with decreased expression of PIGM or PIGX that accumulate GlcN-acyl-PI remains unclear.

Several studies have investigated the human interactome and its implication in human diseases (47-49). BioGRID, a public database of genetic and protein interactions (23), indicated that PIGM and PIGX proteins interact physically with other proteins (Fig. 6). Affinity capture-mass spectrometry and two-hybrid assays indicate that PIGM can bind to 39 proteins (Fig. 6A) and PIGX can bind to six (Fig. 6B). Notably, each subunit binds to different proteins. Proteins interacting with PIGM include glycoproteins, G-protein-coupled receptors in humans and the proteins ORF7, ORF14, E and M of severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2) (Fig. 6A) (23). Whether these interactions reveal a specific biological context of ER or a cross-regulation of the proteins remains unknown; G-protein-coupled receptors, glycoproteins and GPI-APs are commonly recruited in lipid rafts (50-52) and recent evidence indicates that GPI-biosynthesis is relevant for the life cycle of SARS-CoV-2(53).

In summary, expression of PIGM and PIGX is reported in most types of cell, suggesting that GPI-APs may be present in most cells. The present review indicated that PIGM and PIGX are key genes for GPI synthesis since their absence may lead to accumulation of GPIs that lack Man in the cell and deficiency of GPI-APs in the cell membrane or their secretion. Absence or altered expression of PIGM gene is associated with PNH and inherited GPI deficiency, which are characterized by thrombosis; whether this phenotype is due to altered expression of specific GPI-APs remains unclear. Altered expression of PIGM and PIGX has been reported in cancer (14,16,37,39); to the best of our knowledge, however, whether these changes may lead to altered expression of GPI-APs that drive malignant phenotype has not been explored. Transcription factor Sp1 may exert a role in the transcription of PIGM; to the best of our knowledge, there are no studies regarding the transcription of PIGX. Regulation of GPI-MT-I enzyme activity depends on the expression of PIGX and, to the best of our knowledge, no other mechanism has been described. However, since PIGM and PIGX may interact with other proteins independently, additional regulation may be involved. Further exploration may enable the development of targeted therapies for cancer.