Deletion of 3 residues from the C-terminus of MCFD2 affects binding to ERGIC-53 and causes combined factor V and factor VIII deficiency

Combined deficiency of coagulation factors V and VIII (F5F8D, Online Mendelian inheritance in Man no. 227300) is a rare autosomal recessive bleeding disorder first described by Oeri et al in 1954.¹  Individuals affected with this disorder present with a moderate bleeding tendency and plasma levels of factors V and VIII in the range of 5% to 30% of normal. Extensive genetic analysis of F5F8D patients linked the disorder to causative mutations in the 2 genes LMAN1 and MCFD2.^2-7 LMAN1 encodes the transmembrane mannose-lectin ERGIC-53 (ER-Golgi intermediate compartment protein of 53 kDa),⁸  whereas MCFD2 is a soluble luminal protein with 2 calmodulin-like EF hand motifs.⁷  MCFD2 interacts with ERGIC-53 in a calcium dependent manner and the complex recycles between the endoplasmic reticulum (ER) and the ER-Golgi intermediate compartment (ERGIC).⁹  The ERGIC-53/MCFD2 complex is believed to capture factor V and factor VIII in the ER and to package the 2 coagulation factors into transport vesicles that mediate transport to the ERGIC. Indeed, chemical crosslinking identified a triple complex of ERGIC-53, MCFD2 and factor VIII.¹⁰  In the absence of the ERGIC-53/MCFD2 complex, the secretion of factor V and factor VIII is inefficient resulting in low plasma levels and bleeding

The study was approved by the Ethical Committee of the University Hospitals, Geneva, Switzerland.

The patient, a 10-year-old girl with factor V levels of 10% to 13% and factor VIII levels of 5% to 19%, is a single child of nonconsanguineous parents of South American origin. The patient suffers from easy bruising, epistaxis, and gingival bleeding. Her mother did not report any bleeding problem, nor did any other member of the maternal family. Her father had several episodes of epistaxis during childhood but apparently no major bleeding manifestation.

Blood samples were obtained from the affected child and her mother after informed consent was obtained in accordance with the Declaration of Helsinki. The father was unavailable for analysis. Primers used for polymerase chain reaction (PCR) amplification and sequencing of LMAN1 have been previously described.⁶  The MCFD2 gene was analyzed by PCR amplification of the 4 exons and intron-exon junctions followed by sequencing. Primer sequences are available on request.

Southern blot analysis was performed according to standard procedures. Genomic DNA (4 μg) from the patient, her mother, and 2 controls was digested with Sac I or HindIII. A 510-bp probe containing MCFD2 exon 1 and part of intron 1 was obtained by PCR using oligonucleotide primers: MCFD2×1F' 5′GGGGCGAAGCCGAGGAAGA3′ and MCFD2×1R: 5′CAGAGAGGGAATACAACAGG3′ and cloned using pDRIVE (Qiagen, Basel, Switzerland). The deleted allele was confirmed and sequenced by PCR using 2 forward primers: MCFD2delF2: 5′ATTGCTTGAGCCCAGAGCTCGAG3′ and MCFD2delF3: 5′CCAATGAATAGTAAGGTACTC3′ and the reverse primer MCFD2×1R.

Cloning of pcDNA3[HA-MCFD2 WT] was described previously.⁹  pcDNA3[HA-MCFD2 ΔSLQ] was generated by changing Ser144 into a stop codon using PCR-mutagenesis. Constructs were transfected into HeLa cells (CCL-2, American Type Culture Collection, Manassas, VA) using FuGENE6 (Roche Applied Science, Rotkreuz, Switzerland). Forty-eight hours after transfection, cells were harvested, lysed, and subjected to immunoprecipitation using mouse monoclonal antibody G1/93 against human ERGIC-53 (ALX-804-602; Alexis, Lausen, Switzerland), covalently coupled to protein A Sepharose.^9,11  Immunoblotting was performed with G1/93 and antihemagglutinin (HA; Covance, Princeton, NJ).

DNA fragments encoding MCFD2 WT and ΔSLQ were cloned into the pET-16b plasmid (N-terminal polyhistidine tag) and expressed in the Escherichia coli BL21(DE3) codonplus strain induced with 0.5 mM isopropyl β-D-thiogalactopyranoside. Polyhistidine-tagged proteins were purified from lysates with Ni²⁺-nitriltriacetic acid columns (GE Healthcare, Uppsala, Sweden), cleaved by incubation with factor Xa (Novagen, Madison, WI) and further purified using a Benzamidine Sepharose column (Amersham Biosciences, Uppsala, Sweden). Circular dichroism (CD) measurements were conducted on a Jasco J-725 spectropolarimeter (Jasco, Tokyo, Japan) at room temperature using protein samples of 0.15 mg/mL in 150 mM NaCl, 5 mM CaCl₂, and 10 mM Tris-HCl, at pH 7.5.

To identify the cause of F5F8D in a young patient, we screened the LMAN1 and MCFD2 genes for mutations. None were identified in LMAN1. Instead, the patient was found to be heterozygous for a novel nonsense mutation in MCFD2 exon 4 (c.431C > G; p.Ser144X), transmitted by the mother. Because no other mutation was identified by sequencing all MCFD2 exons (including the noncoding exon 1) and intron-exon junctions, we performed Southern blot analysis of genomic DNA digested with Sac I using a probe containing MCFD2 exon 1 and part of intron 1 (Figure 1A). For the mother and 2 controls, an expected band of 25.5 kb was detected (Figure 1B left). In contrast, for the patient, a smaller band of approximately 17 kb was detected in addition to the normal band (Figure 1B left, the patient's DNA sample (P) migrates more slowly than the samples from the controls (C) and from the patient's mother (M) so that the patient's normal band appears to be of greater size). Southern blot analysis with HindIII was also performed using the same probe. Again, a smaller band was detected for the patient, compatible with a deletion of approximately 8.5 kb from the 5′ of MCFD2 (Figure 1B right panel, one HindIII site is included in the deletion). This was confirmed by PCR amplification of the deleted allele followed by sequencing. The deletion eliminates 8.4 kb of DNA 5′ of MCFD2 including exon 1 which contains nearly the entire 5′UTR and is unlikely to result from a single recombination event because 2 small sequences of 55 bp (in the 5′ upstream region) and 96 bp (in intron 1) are retained in the deleted allele in inverted orientation. There is no doubt that this is a null mutation because it knocks out the MCFD2 promoter sequence. Notably, this is the first large deletion identified in a patient with F5F8D.

The remaining MCFD2 allele contains the novel nonsense mutation p.Ser144X. The effect of this mutation on the MCFD2 protein is comparatively mild because only the last 3 C-terminal amino acids are truncated. To determine the underlying molecular mechanism responsible for F5F8D, we analyzed if MCFD2 ΔSLQ can still bind ERGIC-53. To this end, HA-tagged versions of wildtype (WT) and ΔSLQ MCFD2 (Figure 2B) were transiently expressed in HeLa cells and coimmunoprecipitated with ERGIC-53. Although equally expressed as MCFD2 WT, MCFD2 ΔSLQ coprecipitated only minimally with ERGIC-53 (Figure 2C). Quantification revealed that the ΔSLQ mutation in MCFD2 reduced binding to ERGIC-53 by 80% compared with MCFD2 WT (Figure 2D). To examine if impaired binding to ERGIC-53 may be due to a structural defect, recombinant MCFD2 WT and ΔSLQ were expressed in E coli, purified, and analyzed by circular dichroism (CD). The CD spectrum of MCFD2 WT in the presence of Ca²⁺ is typical for an α-helical conformation with minima at around 207 and 222 nm. However, the CD spectrum of the MCFD2 ΔSLQ mutant showed a significantly lower α-helical content (Figure 2E). We conclude that the C-terminal ΔSLQ deletion perturbs the 3-dimensional structure of MCFD2 and as a consequence impairs binding to ERGIC-53 and causes F5F8D.

To date, 30 distinct mutations^4-6,12-15  have been identified in LMAN1, accounting for approximately 75% of F5F8D patients, compared with 11 mutations identified in the MCFD2 gene.^7,14-16  In LMAN1, no missense mutations have been identified which affect binding of either MCFD2 or factors V and VIII. In contrast, 2 missense mutations (Asp129Glu and Ile136Thr) in the second EF-hand domain of MCFD2 were shown to abolish binding to ERGIC-53 and cause F5F8D.⁷  No structural analysis was performed with these mutants. Here we show that the ΔSLQ missense mutation modifies the α-helical content of MCFD2 and impedes binding to ERGIC-53. Thus, our structural observations explain why ΔSLQ leads to a bleeding phenotype similar to a MCFD2 null mutation although ΔSLQ truncates only 3 residues near the putative calcium-binding domain of the second EF-hand. Collectively, our data clearly emphasize the importance of the second EF-hand of MCFD2 for protein conformation, interaction with ERGIC-53, and factor V and factor VIII transport.

We thank Dr Michael Morris for helpful discussions. We are grateful to Luciana Palumbo for expert technical assistance.

This study was supported by a Swiss National Science Foundation Professorship to M.N.A. B.N. is supported by the Roche Research Foundation. Y.K. is a recipient of a Japan Society for the Promotion of Science Research Fellowship for Young Scientists. K.K. was supported in part by CREST project from the Japan Science and Technology Agency and by Grants-in-Aid from the Ministry of Education, Culture, Sports, Science, and Technology of Japan. H-P.H. is supported by the Swiss National Science Foundation and the University of Basel, Switzerland.

National Institutes of Health

Contribution: M.N.A. directed the study and designed and performed the genetic analysis. B.N. and H-P.H designed and performed the study of the mutant MCFD2 in HeLa cells. Y.K., K.Y., and K.K. designed and performed the circular dichroism measurements of the recombinant MCFD2. F.B. and P. d.M. performed the clinical study. All authors contributed to writing the paper.

Conflict-of-interest disclosure: The authors declare no competing financial interests.

Correspondence: Marguerite Neerman-Arbez, Department of Genetic Medicine and Development, 1 Rue Michel Servet, 1211 Geneva, Switzerland; e-mail: Marguerite.Arbez@medecine.unige.ch.