Splice-site mutation of exon 3 deletion in the gamma-glutamyl carboxylase gene causes inactivation of the enzyme

TO THE EDITOR

Gamma-glutamyl carboxylase (GGCX) is an integral membrane protein that catalyzes the post-translational modification of certain glutamates to gamma-carboxyglutamates in vitamin K-dependent (VKD) proteins. Carboxylation is required for the biological activity of numerous VKD proteins involved in a broad range of physiological functions. Mutations in GGCX have been mainly associated with bleeding disorders because these mutations cause undercarboxylation of VKD coagulation factors and of the anticoagulant proteins (Napolitano et al., 2010). GGCX mutations have also been linked to pseudoxanthoma elasticum (PXE)-like syndrome, a non-bleeding disorder caused by functional defects in the matrix Gla protein (MGP) (Vanakker et al., 2007). Patients with PXE-like syndrome have been reported to have comorbid bleeding disorder of vitamin K-dependent coagulation factors deficiency (VKCFD), which is characterized by the simultaneous functional defects of multiple VKD coagulation factors (Li et al., 2009; Rongioletti et al., 1989; Vanakker et al., 2007).

Recently, Kariminejad et al. (2014) reported thirteen patients with phenotypes typical of PXE-like syndrome, but with no coagulation abnormalities (Kariminejad et al., 2014). Genetic analysis of these patients’ ATP-binding cassette subfamily C member 6 (ABCC6) gene (the causative gene for the classical PXE syndrome) excluded the ABCC6 mutations. However, all affected members were found to be homozygous for a splice-site mutation (c.373+3G>T) in the GGCX gene, which causes the deletion of exon 3 in the GGCX mRNA (GGCX-Δex3). It has been suggested that the phenotypes displayed by the affected patients were associated with the GGCX-Δex3 mutation (Kariminejad et al., 2014). However, the reason for the absence of bleeding diathesis in these patients remained unclear, and no functional study on the GGCX-Δex3 mutation was available.

To examine the functional consequences of the GGCX-Δex3 mutation, we determined the carboxylation activity of this mutant using our recently established GGCX-deficient cell-based assay with two reporter proteins (a chimeric coagulation factor, FIXgla-PT and MGP) (Supplementary Methods). This approach, unlike the traditional in vitro GGCX activity assay, allows us to assess the functionality of GGCX using its natural protein substrates in a cellular milieu that requires the enzyme to interact with its physiologic components for VKD carboxylation (Tie et al., 2016). Our result shows that the GGCX-Δex3 mutant abolished carboxylation activity for both reporter proteins (Figure 1a). These results agree with the findings of Kariminejad et al. (2014), who observed uncarboxylated MGP in the affected patients. However, the inability of the GGCX-Δex3 mutant to carboxylate the coagulation factor reporter-protein disagrees with their observation that all the affected patients had normal coagulation factor activities.

Figure 1. Characterization of GGCX mutations identified in PXE-like and VKCFD patients.

(a) Cell-based activity assay of GGCX and its mutant identified from PXE-like patients. Wild-type GGCX and the GGCX-Δex3 mutant were transiently expressed in GGCX-deficient HEK293 reporter cells. Transfected cells were cultured in complete medium containing 5 µg/ml vitamin K. The concentrations of the carboxylated reporter proteins (FIXgla-PT, black bars and MGP, gray bars) in the cell culture medium were measured by ELISA. Wild-type GGCX activity was normalized to 100%. Data are presented as mean±SD (n=3). (b) Carboxylation activity of GGCX and the Arg485Pro mutant in response to increasing concentrations of vitamin K. Wild-type GGCX (filled circles) and the Arg485Pro mutant (filled squares) were transiently expressed in GGCX-deficient HEK293 reporter cells. The enzymatic activity for FIXgla-PT carboxylation was determined as described above. (c) Schematic representation of the proposed membrane topology of GGCX. The exon 3 encoded region is shown in red. (d) Subcellular localization of GGCX and the GGCX-Δex3 mutant in HEK293 cells. EGFP-tagged GGCX or GGCX-Δex3 was transiently co-expressed with the ER marker mCherry-ER-3 in HEK293 cells. Forty-eight hours post-transfection, cell nucleus was stained by Hoechst 33342 and directly used for fluorescence confocal image collection. GGCX fusions were visualized by the green fluorescence of EGFP, ER marker was visualized by the red fluorescence of mCherry, and cell nuclei were visualized by the blue fluorescence of Hoechst. Scale bar = 10 µm. (e) Immunoblotting analysis of GGCX (lane 1) and the GGCX-Δex3 mutant (lane 2) proteins. Full-length GGCX band is indicated by arrow.

To further clarify the effect of GGCX-Δex3 mutation on the carboxylation of coagulation factors, we characterized GGCX mutants of a VKCFD patient with compound heterozygous mutations of GGCX-Δex3 and Arg485Pro (Rost et al., 2004). If the GGCX-Δex3 mutation located on one allele expresses an inactive enzyme, as we observed (Figure 1a), the Arg485Pro mutation on the other allele should play a major role in the clinical phenotype of this VKCFD patient, whose coagulation disorder has been partially corrected by vitamin K administration (Rost et al., 2004). To test this hypothesis, we titrated the carboxylation activity of the Arg485Pro mutant in response to increasing concentrations of vitamin K. Compared with the wild-type enzyme, the Arg485Pro mutant requires a ~5-fold higher vitamin K concentration to reach half-maximal carboxylation of FIXgla-PT (Figure 1b). At the optimum vitamin K concentration, the Arg485Pro mutant has ameliorated coagulation factor carboxylation up to ~50%. This agrees with the patient’s clinical result, in which vitamin K administration partially restored coagulation factor activities to ~60% (Rost et al., 2004). These results support our hypothesis that, when the GGCX-Δex3 mutation on one allele encodes an inactive enzyme, the Arg485Pro mutation on the other allele plays a key role in the clinical phenotype of this VKCFD patient.

Deletion of exon 3 results in an in-frame deletion of 53 amino acid residues (between Phe73 to Gly125) in the GGCX molecule, which includes part of the first and second transmembrane domains (TMDs) and the lumenal loop between TMD1 and TMD2 (Figure 1c) (Tie et al., 2000). Therefore, exon 3 deletion could disrupt the proper integration of TMD1 and TMD2 into the endoplasmic reticulum (ER) membrane, which would result in the misfolding of the GGCX protein. Exon 3 also encodes residue Cys99, which forms the only disulfide bond with Cys450 in the GGCX molecule that is essential for GGCX folding and maturation (Tie et al., 2003). To test the effect of exon 3 deletion on GGCX maturation, we fused an enhanced green fluorescence protein (EGFP) at the C-terminus of the wild-type GGCX and the GGCX-Δex3 mutant, co-expressed these fusion proteins with the ER marker protein mCherry-ER-3 in HEK293 cells, and examined their expression and localization by fluorescence confocal microscopy and immunoblotting analysis (Supplementary Methods). Our results show that, while the wild-type GGCX appears to be localized to the ER (Figure 1d), the GGCX-Δex3 mutant shows a ubiquitous diffusion throughout the entire cell, which suggests the mis-localization and probable misfolding of the mutant protein. The immunoblotting result is consistent with this interpretation, revealing that the GGCX-Δex3 mutant shows more protein bands than the wild-type GGCX, suggesting both degradation (the protein bands are smaller than the wild-type GGCX) and aggregation (the protein bands are larger than the wild-type GGCX) of the mutant protein (Figure 1e). Together, these results suggest that the GGCX-Δex3 mutant protein is misfolded and mis-localized in the cell, and is therefore unable to carry out its biological function in the ER.

Our results suggest that exon 3 deletion in the GGCX gene results in an inactive enzyme. This result cannot explain the normal coagulation activity of the patients who had a homozygous splice-site mutation of c.373+3G>T, causing exon 3 to be skipped in the GGCX mRNA (Kariminejad et al., 2014). One possible explanation for this discrepancy is that our result is based on cell-based characterization of the GGCX-Δex3 mutant. However the patients’ splice-site mutation (c.373+3G>T) may not completely abolish normal processing of the GGCX pre-mRNA, and this could produce a portion of correctly processed mRNA for translating the functional protein, as has been reported in other diseases (Mikkola et al., 1997; Schernthaner-Reiter et al., 2016). It is also possible that this splice-site mutation could result in several variant mRNA transcripts of GGCX, as has previously described for other genes (Gallinaro et al., 2006; Kallabi et al., 2015). The affected patients reported by Kariminejad et al. (2014) have normal coagulation factor carboxylation, but defects in their MGP carboxylation; it is therefore possible that some of the mRNA transcript may produce a GGCX mutant protein that causes MGP-carboxylation defects but has a smaller effect on coagulation factor carboxylation, as we have previously observed in a VKCFD patient (Tie et al., 2016).

Splice-site mutations make a significant contribution to human genetic disease, but the effects of these mutations on pre-mRNA splicing remain to be elucidated (Krawczak et al., 2007). Mutations in exon-intron junctions could lead to mis-splicing and could result in exon skipping, activation of a cryptic splice-site, or intron retention. Importantly, a single splice-site mutation can result in multiple mRNA transcripts and can affect the protein function (Gallinaro et al., 2006; Kallabi et al., 2015; Mikkola et al., 1997; Schernthaner-Reiter et al., 2016). Therefore, it is important to clarify the effect of the GGCX splicing-site mutation (c.373+3G>T) on the splicing mechanism and the possible GGCX mRNA transcripts in these affected patients, especially when mutations in the causative gene for the classical PXE syndrome have been excluded. It would also be beneficial to identify possible variations on the vitamin K epoxide reductase (VKOR) gene in the affected patients, as VKOR is the essential enzyme that provides the reduced form of vitamin K needed for GGCX to function.

In conclusion, we have characterized the exon 3 deletion mutation of GGCX which was identified from patients with a homozygous splice-site mutation (c.373+3G>T) in the GGCX gene. Our results suggest that clarifying the correlation between the GGCX genotypes and their clinical phenotypes will remain challenging, especially when these mutations are located in splice junctions that could affect pre-mRNA processing.

Abbreviations

ABCC6

ATP-binding cassette subfamily C member 6

ER

endoplasmic reticulum

FIXgla-PT

prothrombin with its Gla domain replaced by that of FIX

EGFP

enhanced green fluorescence protein

GGCX

gamma-glutamyl carboxylase

MGP

matrix Gla protein

PXE

pseudoxanthoma elasticum

TMD

transmembrane domain

VKCFD

vitamin K-dependent coagulation factors deficiency

VKD

vitamin K-dependent

VKOR

vitamin K epoxide reductase