Characterization based on genotype–biochemical phenotype association in fructose-1,6-bisphosphatase deficiency

doi:10.21203/rs.3.rs-2185039/v1

Purpose

Fructose-1,6-bisphosphatase (FBPase) deficiency, caused by an FBP1 mutation, is an autosomal recessive disorder characterized by hypoglycemic lactic acidosis. The mechanism by which the mutations cause enzyme activity loss is uncertain.

Methods

We performed whole-exome sequencing in an adult patient with severe hypoglycemic lactic acidosis and identified that the patient carried compound heterozygous missense mutations of FBP1 with c.491G > A (p.G164D) and c.581T > C (p.F194S).

Results

Biochemical analysis revealed that FBP1 mutant (G164D or F194S) decreased protein expression and enzyme activity loss. The interactome analysis for binding partners demonstrated that G164D and F194S mutants interact with the proteins involved in unfolded protein response. Additionally, G164D and F194S mutants aggregated in the endoplasmic reticulum, suggesting the involvement of protein misfolding in its pathogenesis. All FBP1 missense mutations previously reported were classified into three functional categories: Type 1 mutations, located at pivotal residues in enzyme activity motifs with no effects on protein expression; Type 2 mutations, which mediate changes in amino acid hydrophobicity and structurally cluster around the substrate-binding pocket, are associated with aggregation in the endoplasmic reticulum, and decreased protein expression; and Type 3 mutations, which are likely non-pathogenic mutations.

Conclusion

Protein misfolding contributes to FBPase deficiency pathogenesis, particularly in Type 2 mutations.

Biological sciences/Genetics/Mutation

Biological sciences/Cell biology/Protein folding/Chaperones

fructose-1

6-bisphosphatase deficiency

FBP1

hypoglycemia

protein misfolding

heat shock protein

Fructose-1,6-bisphosphatase (FBPase) is a key regulatory enzyme in gluconeogenesis that catalyzes the hydrolysis of fructose 1,6-bisphosphate to fructose 6-phosphate and inorganic phosphate ¹. Baker and Winegrad first reported FBPase deficiency in 1970 ², and Kikawa et al. subsequently identified three FBP gene (FBP1) mutations associated with FBPase deficiency in 1997 ³. FBPase deficiency is an autosomal recessive disorder that generally courses with hypoglycemia and metabolic acidosis ⁴. Although fasting and febrile infections are known to trigger life-threatening episodes of hypoglycemia and lactic acidosis in infancy, these episodes rarely occur in adults due to increased glycogen storage ⁴. Several harmful mutations in the coding region of FBP1 have been reported ⁵. Nevertheless, with incidence rates estimated to range between 1/350,000 and 1/900,000, FBPase deficiency remains a very rare inherited disease 5^,6. Due to its low incidence, elucidating the molecular mechanism by which the identified mutations cause the loss of enzyme activity is still challenging, particularly in terms of genotype phenotype correlation and adult-onset cases. Here, we present the case of a 22-year-old female patient who was diagnosed with FBPase deficiency and exhibited a novel compound heterozygous mutation of the FBP1 gene. Our results demonstrate the underlying mechanism by which the identified mutations caused FBPase deficiency involves protein misfolding, and we examine and categorize the biochemical phenotypes of all previously reported FBP1 missense mutations into three functional groups.

Mutation analysis

DNeasy Blood and Tissue Kits (QIAGEN, Hilden, Germany) were used to extract the genomic DNA from the patient’s blood. Then, all 7 exons of the FBP1 gene were amplified by polymerase chain reaction (PCR) and sequenced using a 3130 Genetic Analyzer (Applied Biosystems, Massachusetts, USA). Table S1 (Additional file 1) describes the primer information and PCR conditions.

Whole-exome Sequencing

The whole-exome sequencing involved the targeted capture of all the exon sequences using SureSelect Human All Exon v6 (Agilent Technologies, California, USA), followed by massive parallel sequencing of the enriched exon fragments on the HiSeq 2500 platform (Illumina) using the 125-bp paired-end mode as per manufacturer’s protocol ⁷. The sequenced reads were aligned to the human genome reference (GRCh37) using the default parameter settings in Burrows-Wheeler Aligner version 0.7.10, while the PCR duplicates were eliminated using Picard-tools version 1.39 (http://picard.sourceforge.net/). Candidate mutations with (i) depths ≥ 8; (ii) Number of variant reads ≥ 4; and (iii) variant allele frequency (VAF) 0.4–0.6, 1 were adopted, as VAFs of pathogenic germline mutations are estimated to be approximately 0.5 in heterozygous state and 1 in homozygous state. Mutations were further filtered by excluding (i) variants presenting only in unidirectional reads; (ii) insertions and deletions in simple repeat regions; (iii) synonymous SNVs; and (iv) known variants listed in the 1000 Genomes Project (Nov 2010 release), Exome Sequencing Project (ESP) 6500, Human Genome Variation Database (HGVD; October 2013 release), and ExAC database with frequencies > 0.001 in order to exclude non-pathogenic variants.

Construction Of Fbp1 Expression Vectors

The total RNA was extracted from the patient’s blood using RNeasy Kits (QIAGEN, Hilden, Germany) and reverse-transcribed using SuperScript 2 reverse transcriptase (Thermo Scientific, Massachusetts, USA) and oligo (dT) primers. The amplification of the coding region of the FBP1 gene was carried out using the forward primer 5’-CACCATGGCTGACCAGGCGCCCTTCG-3’ and the reverse primer 5’-TCACTGGGCAGAGTGCTTCTCATAC-3’. The PCR procedure consisted of the following steps: a) denaturing at 98°C for 2 min, followed by 94°C for 15 secs, and b) annealing for 30 secs at 58°C and extension at 72°C for 1 min for 30 cycles. The PCR products were subcloned into pGEM-T Easy Vector (Promega, Wisconsin, USA). Then, the FBP1 fragments (WT and 2 mutants) were cut from the pGEM-T Easy Vector using EcoRI and subcloned into a pCMV-Myc-N Vector (Clontech, California, USA). A corresponding pCMV-Myc-N Vector was used as a negative control.

Mutagenesis Of Fbp1 Plasmids

D119N, P120L, R158W, G164S, A177D, G207R, N213K, G260R, E281K, P284R, G294E, G294V, and V325A mutations were introduced by site-directed mutagenesis (Quick Change Lightning Site-Directed Mutagenesis KIT, Agilent Technologies) using the primers listed in Table S2 (Additional file 1). XL10-Gold ultracompetent cell DNA was isolated from cultured single clones and by sequencing to confirm successful mutagenesis.

Generation Of Fbp1 Knockout Hepg2 Cells

The protocol used for the CRISPR/Cas9 system was consistently based on that reported by Cong et al., Science 2013 ⁸. The backbone vectors pX459 pSpCas9(BB)-2A-Puro and pX462 pSpCas9n(BB)-2A-Puro were obtained from Addgene (Massachusetts, USA). The target guide RNA sequences were designed at exon 5 including c.491G, exon 6 including c.581T, and exon 8 including the Japanese common mutation site (c.960-961insG) of the FBP1 genome (5’-GCAGCCGGCTACGCACTGTA-3’, 5’-GCACCAAAATGAACTCCCCGA-3’ and 5’-GTCGGGGGATCCCAAGATCAC-3’) ³. To clone the exon 5 and exon 6 target sequences into pX462 and the exon 8 target sequence into the pX459 backbone, the oligos were synthesized using Eurofins genomics (Tokyo, Japan) (Additional file 1: Table S3). Then, these oligos were submitted to annealing and phosphorylation by means of a T4 DNA Ligase Reaction Buffer and a T4 Polynucleotide Kinase (New England Biolabs, Massachusetts, USA) used at 37℃ for 30 min and 95℃ for 5 min. pX459 and pX462 were digested using BbsI (Thermo Scientific, Massachusetts, USA) at 37℃ for 30 min, with gel purification being performed by use of a QIAquick Gel Extraction Kit (QIAGEN, Hilden, Germany). Ligation reactions of pX459, pX462 and the annealed oligos were performed for 10 min at room temperature using a Quick Ligation Kit (New England Biolabs, Massachusetts, USA). Then, the ligated oligos were purified using PlasmidSafe exonuclease (Cambio, Cambridge, UK) at 37℃ for 30 min. The plasmids were transfected into Stbl3, and the appropriate transfectants were amplified and collected using NucleoBond Xtra Midi (Takara, Kusatsu, Japan). Three plasmids (pX462-exon5 gRNA, pX462-exon6 gRNA and pX459-exon8 gRNA) were cotransfected into HepG2 cells using lipofectamine 3000 (Thermo Scientific, Massachusetts, USA) as per manufacturer’s protocol. The plasmid-expressing HepG2 cells were selected by puromycin (Wako, Osaka, Japan), and the limiting dilution method was used to establish the monoclonal cell line.

Cell Culture And Transient Transfection

The human hepatocarcinoma cell line FBP1-KO HepG2 culture was conducted using Dulbecco’s modified Eagle’s medium containing antibiotics and 10% fetal bovine serum. The FPB1-KO HepG2 cells were plated on 6-well plates and then transfected with plasmid DNA (2.5 µg) complexed with Lipofectamine 3000 reagent (7.5 µl) and P3000 reagent (5 µl) in 250 µl of Opti-MEM (Thermo Scientific, Massachusetts, USA). Subsequently, the FBPase activity of HepG2 cells was measured 48 hours after transfection using a nicotinamide adenine dinucleotide phosphate (NADP)-coupled spectrophotometric assay.

Fbpase Activity Assay

The FBPase activity was calculated from an NADP-coupled spectrophotometric assay as described by Kikawa et al ⁹. The assay mixture (300 µl) was composed of 40 µg protein of cell lysate, 50 mmol/L Tris-HCl buffer (pH 7.5), 2.0 mmol/L MgCl₂, 1.0 mmol/L EDTA, 0.2 mmol/L NADP, 3.5 U/mL glucose-6-phosphate dehydrogenase, 1.5 U/mL glucose-6-phosphate isomerase, and 100 µmol/L FBP, which served as the substrate. A 96-well plate reader was used to record the rate of NADPH formation.

Real-time Rt-pcr Analysis

Real-time RT-PCR analysis

RT-qPCR experiments were performed as previously described ^10–12. The FBP1 gene-specific mRNA expression values were determined and normalized to those of β-actin as an internal control. Briefly, the total RNA (4 µg) was extracted using an RNeasy kit (Qiagen, Valencia, CA, USA) and reverse-transcribed using a ReverTra Ace qPCR RT Kit (Toyobo, Tokyo, Japan). The cDNA products were subjected to RT-PCR using a Step One Plus Real-Time PCR system (Applied Biosystems, Massachusetts, USA). All the primer information is provided in Table S4 (Additional file 1).

Immunoblot Analysis

Immunoblot analyses were performed as previously described ¹³. The antibodies included FBP1 (SIGMA rabbit polyclonal, clone: HPA005857), c-Myc (Santa Cruz mouse monoclonal, clone: 9E10), actin (SIGMA rabbit polyclonal, clone: A2066), ATF6 (Novus Biologicals mouse monoclonal, clone: 70B1413.1), HSP70 (StressMarq mouse monoclonal, clone: N27F3-4), and HSP90 (Santa Cruz mouse monoclonal, clone: sc-13119), HSP60 (Abcam rabbit polyclonal, clone: ab46798) and TCP1 (Bethyl laboratories rabbit polyclonal, clone: A303-444A).

Er Fractionation

The extraction of the ER was performed using an Endoplasmic Reticulum Enrichment Kit (Novus Biologicals, Colorado, USA), as per manufacturer’s instructions.

Immunofluorescence Analysis

Immunofluorescence analysis was performed to examine the cellular expression of the WT and mutant FBP1. The FBP1-KO HepG2 cells were cultured in 4-well chamber slides (2 × 10⁴ cells/well) and then transfected with plasmids (0.5 µg). Twenty-four hours after the transfection, the cells were treated with CellLight® ER-GFP (12 µl/well; Thermo Scientific, Massachusetts, USA), which is a marker of ER, followed by fixation in 100% ethanol at − 20℃ for 10 min and incubation with blocking solution and primary antibodies against FBP1 (SIGMA rabbit polyclonal, clone: HPA005857), c-Myc (Santa Cruz mouse monoclonal, clone: 9E10) and GFP (MBL rabbit polyclonal, clone: 598) 48 hours after the transfection. A confocal laser microscope (LSM710, Carl Zeiss, Germany) was used to obtain fluorescence images.

Immunoprecipitation Assay

The commercially available kit c-Myc-tagged Protein Mild Purification Kit (Medical & Biological Laboratories, Aichi, Japan) was used to perform the immunoprecipitation assay. HepG2 cell extracts containing different Myc-tagged FBP1 variants were incubated with anti-Myc beads at 4℃ for 1 hour. The beads were then rinsed and eluted using a wash solution and elution peptides.

Mass Spectrometry Sample Preparation

The FPB1-KO HepG2 cells were transfected with Myc-tagged FBP1-WT, Myc-tagged FBP1-G164D, Myc-tagged FBP1-F194S containing plasmid DNA. These cells were treated with the mannosidase inhibitor kifunensine (200 µM) for 48 hours. Myc-tagged FBP1 HepG2 cells solubilization was obtained with the following buffer: 50 mM Tris-HCl (pH 7.5), 1.0 mM MgCl₂, 0.1 mM EDTA, 0.5 mM phenylmethannesulphonyl fluoride, 2.5 µg/mL leupeptin and 1.0 µg/mL antipain. Immunoprecipitation with anti-Myc beads (Medical & Biological Laboratories, Aichi, Japan) was performed at 4℃ for 1 hour. The beads were then rinsed and eluted using a wash solution and elution peptides. Immune complexes were separated by SDS-PAGE. Bands were exsected from the gel and examined by mass spectrometry to find corresponding proteins. Respective gel pieces were rinsed two times using 100 mM bicarbonate in acetonitrile with subsequent protein digestion by trypsin. 0.1% formic acid was then added to the supernatant, and the peptides were subjected to liquid chromatography-tandem mass spectrometry (LC-MS/MS) using a LTQ Mass Spectrometer (Thermo Scientific). Analysis of MS/MS data set results was conducted using the Mascot software program (Matrix Science).

Proteomic Analysis And Database Search

Proteomic analysis and database search were performed as previously described ¹⁴. Kyoto Encyclopedia of Genes and Genomes (KEGG) pathway annotation and Gene Ontology (GO) were performed using the STRING interaction database.

Protein Structure Based On Protein Data Bank

Ke et al. revealed the detailed crystal structure of Sus scrofa FBP1 complexed with fructose 6-phosphate, AMP, and magnesium ¹⁵; Fig. 1g shows the structure of FBP1 dimer based on their report (DOI: 10.2210/pdb1FBP/pdb). The FBP1 amino acid sequences between Homo sapiens and Sus scrofa show approximately 90% similarity. Particularly, the functional motifs directly associated with FBPase activity are highly conserved.

Statistics

The results are presented as mean ± SD. Student’s t-test was used to analyze the continuous variables, and P-values < 0.05 were considered statistically significant.

Study Approval

In accordance with the protocols issued by the Chiba University Hospital, written informed consent for the genetic studies was obtained from the patient and her family.

Case presentation and whole-exome sequencing

We accompanied the case of a 22-year-old Japanese female patient who presented with a severe hypoglycemic attack and acidosis induced by prolonged fasting. Blood exams, including a blood gas analysis, showed very low levels of plasma glucose (12 mg/dL) with high levels of lactate (110 mg/dL) and severe metabolic acidosis (pH 6.85, PaCO₂ 19 mm Hg, HCO₃⁻ 2.5 mmol/L). The patient reported experiencing similar episodes triggered by fasting or febrile illness during her early childhood (Additional file 1: Table S5). We suspected the case to be an inherited metabolic issue, such as fatty acid oxidation or gluconeogenesis disorders causing hypoglycemic attacks with lactic acidosis, and further examinations were performed in order to establish a definitive diagnosis. Radiological examination of the abdomen by computerized tomography (CT) revealed a fatty liver, and the liver biopsy showed hepatic steatosis without other alterations (Additional file 3: Figure S1b). The acylcarnitine profile exhibited no specific pattern, suggesting a normal β-oxidation pathway. However, the urinary organic acid profile exhibited significantly elevated glycerol, lactate, and pyruvate levels, indicating a possible FBPase deficiency. The oral fructose tolerance test demonstrated a rapid decrease in blood glucose levels and an increase in lactate concentrations (Fig. 1a). Additionally, fructose loading increased glycerol levels in the urinary organic acid profile (Fig. 1b). Although clinical findings up to this point strongly indicated a case of FBPase deficiency, a differential diagnosis from other metabolic disorders that course with hypoglycemic attacks and lactic acidosis was difficult due to overlapping clinical features ⁶.

As previous studies have shown that genetic analyses using next-generation sequencing is useful for the molecular diagnosis of complex metabolic diseases including FBPase deficiency ¹⁶, we performed the whole-exome sequencing in samples from the proband and her family in order to identify potential underlying genetic defects. This approach revealed the 111 filtered variants in the proband (Additional file 2: Table S6). Two heterozygous variants were identified in a region of FBP1 among them, whereas there was no variants associated to mitochondrial fatty acid oxidation disorders and gluconeogenesis such as OCTN2, CACT, CPT1, CPT2, LCHAD, MCAD, SCAD, MTP, VLCAD, ACAD9, ETFDH, ETFA, ETFB, HMGCS2, HMGCL, PC, PCK1, PCK2, G6PC and PGM1. The identified variants in FBP1 were missense mutations (II-1 in Fig. 1c and 1e) with c.491G > A p.G164D, which is a novel mutation, and c.581T > C p.F194S, which has been previously reported ¹⁷. According to the exome sequencing results of family members, her parents carried single mutation respectively: G164D mutation from father (I-1 in Fig. 1c and 1e) and F194S mutation from her mother (I-2 in Fig. 1c and 1e). Then we confirmed that both of FBP1 variants with G164D and F194S in the proband were heterozygous missense mutations (Fig. 1d) using Sanger sequencing, suggesting that it could be in a compound heterozygous state. Based on these findings, we made a definitive diagnosis of FBPase deficiency and advised the patient to avoid both prolonged fasting and intake of fructose-rich foods under a fasted state.

The G164 and F194 are located in the β-strand structure (Fig. 1g, 3a), and both of identified mutations (G164D and F194S) could lead conformational changes due to the substitution of hydrophobic with hydrophilic amino acids (Table 1). In general, FBPase catalyzes the hydrolysis of fructose 1,6-bisphosphate to fructose 6-phosphate in the presence of divalent cations, such as magnesium, manganese, or zinc ¹⁸, whereas AMP acts as an allosteric inhibitor. The human liver FBPase comprises four identical polypeptide chains each containing 338 amino acid residues, which are assembled as relatively flat tetramers with subunits conventionally labeled C1–C4 ¹⁸. These homotetramers consist of two intimate dimers. Figure 1g shows the structure of FBP1 dimer. We found that neither G164D nor F194S corresponded to the pivotal amino acid residues within functional motifs such as AMP binding sites, metal binding sites, or substrate binding sites (Fig. 1g, 3a). Thus, we decided to examine if and how these mutations affect FBPase enzymatic activity.

Protein expression and enzymatic activity of the FBP1 mutations were lower than those in the wild-type

We established FBP1 mutants in hepatocytes to analyze molecular function. In order to eliminate the influence of endogenous FBPase, we generated FBP1-KO HepG2 cells using the CRISPR/Cas9 system (Additional file 3: Figure S1a). Next, Myc-tagged FBP1 cDNA constructs of wild-type (WT), mutant clones (G164D or F194S), or cotransfects (G164D and F194S) were transduced in the FBP1-KO HepG2 cells. NADP-coupled spectrophotometric assays were used to examine the enzymatic activity of mutant FBPase. As a result, the FBPase activities of all of mutants (G164D; 0.73 ± 0.34, F194S; 0.80 ± 0.56, and cotransfected clone; 0.40 ± 0.41 mmol/min/mg protein) were significantly lower than WT clone (4.82 ± 1.61 mmol/min/mg protein) (Fig. 1f). We validated the transcription level of FBP1 was not greater in WT clone, but immunoblot analysis revealed that the protein levels of FBP1 mutants were markedly reduced (Fig. 2a). To evaluate the enzymatic activity in FBP1 protein, we additionally performed an in vitro NADP-coupled spectrophotometric assay with precipitated FBP1 protein by Myc-tag. The result represented that enzyme activity has completely diminished in FBP1 protein with both G164D and F194S (Additional file 3: Figure S2a and S2b).

Subsequently, immunofluorescence analysis was used to evaluate the intracellular localization of the FBP1 gene product. We found that the mutant FBP1 with G164D or F194S aggregated in the cytoplasm, whereas FBP1 protein in WT clone was diffusely localized in the cytoplasm (Fig. 2b). Importantly, this aggregated distribution of FBP1 protein was also found in the liver biopsy from the presented patient (Fig. 2c), therefore we concluded that FBP1 mutations with G164D and F194S cause protein aggregation and the pathogenic loss in enzymatic activity.

FBP1 mutations with G164D and F194S aggregated in the endoplasmic reticulum due to protein misfolding

To address underlying mechanisms of aggregation and reduction in protein expression, we surveyed the interaction partners of FBP1 using Liquid chromatography-tandem MS (LC-MS/MS) analysis for the proteins pulled down by immunoprecipitation with anti-Myc tag. In total, 408 proteins were detected with LC-MS/MS analysis for WT, G164D mutant and F194S mutant (Fig. 2d). Next, we evaluated the molecular functionalities of those proteins interacting with FBP1 and its G164D or F194S mutated variants, including common binding partner among WT and mutants, by Ingenuity Pathway Analysis (IPA) (Fig. 2e). The analysis indicated that FBP1-interacting proteins can be mainly involved in protein production and degradation such as translation, post-transcriptional regulation, ubiquitination pathway and unfolded protein response. To test whether FBP1, especially its mutants, has such a connection to protein production and degradation as predicted, we additionally investigated the FBP1-interactome based on the STRING interaction database ¹⁹. Notably, the interactome analysis revealed that the pathways and gene ontology terms connected to unfolded protein binding, heat shock protein binding, protein processing in endoplasmic reticulum and proteasome were highly enriched (Fig. 2f). It is important that the FBP1 mutants (G164D and F194S) more strongly interacted with almost all these proteins compared to FBP1 WT (Additional file 1: Table S7), suggesting that these molecular pathways are involved in the pathogenesis of FBPase deficiency.

Inherited mutations can break native protein folding, resulting in the formation of misfolded proteins that are consequently retained in the endoplasmic reticulum (ER). These unfolded proteins undergo mannose trimming by ER associated mannosidase and degradation by proteasome (endoplasmic reticulum-associated degradation; ERAD), which can be suppressed via the inhibition of ER associated mannosidase activity using kifunensine ²⁰. According to previous studies, protein expression of FBP1 is regulated by the ubiquitin proteasome and autophagy pathways ²¹. Thus, to address the involvement of ERAD or protein degradation pathways in FBP1 mutant protein expression, we examined the effects of various inhibitors, such as a proteasome inhibitor (MG-132), an autophagy inhibitor (3-methyladenine), and a potent inhibitor of the ER mannosidase I (kifunensine), on the protein expression of the FBP1 mutants with G164D and F194S. Immunoblot analysis revealed that MG-132 and 3-methyladenine had no impact on the protein expression of these mutants (Additional file 3: Figure S2c). By contrast, kifunensine upregulated the protein expression of these FBP1 mutants with G164D and F194S, suggesting that these mutants underwent ERAD (Fig. 2i). For further confirmation this notion, we performed confocal microscopy analysis to determine the intracellular localization of the FBP1 mutants with G164D and F194S, particularly in relation to ER markers. These mutants were primarily, but not completely, found to be colocalized in the ER (Fig. 2g). In addition, immunoblot analyses using the ER fraction revealed that the protein expression of FBP1 mutants with G164D and F194S compared with WT was markedly high in the ER compared to that in the whole cell lysate (Fig. 2h).

The heat-shock proteins (HSPs) are a family of molecular chaperones, which collectively form a network that is critical for protein folding ²². Moreover, missense mutations were shown to shift the folding equilibrium toward a partially folded state, thus increasing the cellular fraction of HSPs relative to the WT protein ²³. Based on these findings, we evaluated the binding of HSP70, HSP90, HSP60 and TCP1 to the FBP1 proteins using anti-Myc immunoprecipitation with subsequent immunoblot analysis (Fig. 2j). The FBP1 mutants with G164D and F194S had significantly increased interactions with HSP70, HSP90, HSP60 and TCP1, to a greater extent than the WT (Fig. 2k), indicating that these mutants affect the folding equilibrium. Taken together, the FBP1 variants with G164D and F194S mutations caused protein misfolding, which led to the reduction of protein expression and aggregation via the ERAD system.

Examination of mutation genotype and functional phenotype in all FBP1 missense mutants

To date, various FBP1 mutations (14 missense mutations, 12 deletion mutations, four nonsense mutations, four insertions/duplications, two splices, and one indel) have been reported ⁵. The reported missense mutations are D119N ¹⁶, P120L ⁵, R158W ⁶, G164S ³, A177D ³, F194S ¹⁷, G207R ⁵, N213K ²⁴, G260R ²⁵, E281K ²⁶, P284R ¹⁷, G294E ²⁷, G294V ²⁴, and V325A ³. We examined the relationship between the mutated site and the substitution of hydrophobicity in these 14 types of FBP1 missense mutations and G164D mutation we newly identified in this patient (Fig. 3a). Eight types of mutation (G164D, G164S, A177D, F194S, G207R, G260R, P284R, and G294E) exhibited substitutions of hydrophobic amino acids with hydrophilic amino acids, and one mutation (R158W) exhibited a substitution of a hydrophilic amino acid with a hydrophobic amino acid. These nine mutations, defined as mutations with changing hydrophobicity, were not located at key amino acid residues (Fig. 3a, indicated by arrowheads) within functional motifs of the substrate binding site, metal binding site, or AMP binding site. Six mutations (D119N, P120L, N213K, E281K, G294V, and V325A) did not generate any change to amino acid hydrophobicity, and four of these mutations (D119N, P120L, N213K, and E281K) were directly located at important amino acid residues of the enzyme activity [D119N and E281K were located at key residues in the metal binding site; P120L was located in the linker lesion between the metal binding site (D119 and L121) and the substrate binding site (D122); and N213K was located in the substrate binding region (N213-Y216)] (Fig. 3a, indicated by arrowheads).

Based on this information, we assessed how that the alteration of hydrophobicity affects the enzyme activitiy and the protein aggregation. We generated the mutant HepG2 cell line on FBP1-KO background as used previously. Then, these clones were separated into two groups: 6 mutants without change in hydrophobicity (indicated in blue) and 9 mutants with change in hydrophobicity (indicated in red) (Fig. 3). We firstly validated the similar gene transduction in each clone by qPCR (Additional file 3: Figure S3). Although two of mutants, G207R and V325A, kept the enzyme activity compared to WT clone, most of clones demonstrated a loss of FBPase activity (Fig. 3b).

In terms of protein level, there were several exceptions, but FBP1 expression tended to be different depending on the substitution of hydrophobicity. The mutant group without changes in hydrophobicity (indicated in blue), except G294V, maintained a similar level of protein expression as WT, whereas the mutants group with changes in hydrophobicity (indicated in red), except G207R, exhibited decreased protein expression (Fig. 3c, 3d). Consistent with the protein expression, immunofluorescence staining demonstrated that all mutants with protein expression similar to WT were diffusely localized in the cytoplasm, whereas mutants with decreased protein expression aggregated in the cytoplasm (Fig. 4a, 4b). Subsequently, we observed a strong negative correlation between the number of cells with FBP1 aggregates and its protein expression among the FBP1 missense mutants (Fig. 4c). These data suggested that the protein aggregation could be linked to the substitution of hydrophobicity, and FBPase activities has more broadly affected by the missense mutations independent of hydrophobicity status. As there were several exceptions including G207R and V325A, the other constructive feature could be associated to enzymatic function.

FBP1 missense mutations were categorized into three functional phenotypes

Based on these results, we categorized the FBP1 missense mutations into three functional groups (Table 1, Fig. 5). The Type 1 mutations (D119N, P120L, N213K, and E281K) are direct substitution of pivotal amino acid residues inside enzyme activity site (Fig. 5a, 5b). These mutations did not change their amino acid hydrophobicity and protein expression and were diffusely localized in the cytoplasm similar to WT (Table 1). Therefore, these mutations cause a primary loss of FBPase enzymatic activity through mutations of key amino acid residues in the functional motif of enzymatic activity (indicated by arrowheads) without affecting protein expression and cytoplasmic localization. Indeed, Type 1 mutations are characterized by no change in hydrophobicity.

Type 2 mutations (R158W, G164D, G164S, A177D, F194S, G260R, P284R, G294E, and G294V) are likely to be out of the important amino acid residues in the functional motif (Fig. 5a) and appear to cluster around the substrate binding pocket (Fig. 5c). In line with the changes in amino acid hydrophobicity (except for G294V), Type 2 mutations decreased protein expression and caused aggregation in the cytoplasm, possibly due to protein misfolding.

Type 3 mutations (G207R and V325A) were structurally distant from sites associated with the enzyme activity motif and substrate binding pocket (Fig. 5d). These mutations exhibited normal FBPase enzyme activity and their protein expression and cytoplasmic localization were similar to those of WT regardless of amino acid hydrophobicity, indicating that Type 3 mutations are likely non-pathogenic in terms of biochemical phenotype of FBPase deficiency. In fact, the previous reports suggested that the cases with V325A had no functional defect ³.

As expected, when we examined the binding ability of all mutants to HSP, only Type 2 mutants increased the interaction with HSP70 and HSP90 to a greater extent than those seen with WT, Type 1, and Type 3 mutants (Fig. 5e and Additional file 3: Figure S4). Furthermore, the binding ability of FBP1 WT and mutants to either HSP70 or HSP90 was significantly correlated with the number of cells with FBP1 aggregates (%) (Fig. 5f). Thus, we postulate that decreased protein expression due to protein misfolding in association with HSP recognition and protein aggregation via ERAD system is involved in the pathogenesis of FBPase deficiency, particularly in Type 2 mutations.

FBPase deficiency is sometimes misdiagnosed or diagnosed in delay following initial suspicion of other energy-related deficiencies. Our patient’s case and the observations we made in the present study illustrate how FBPase deficiency can be confused with disorders of fatty acid oxidation. While acylcarnitine profiles are useful for the diagnosis of the latter, we advocate that fructose tolerance test and genetic analyses would be beneficial and provide informative confirmation to the diagnosis of patients with FBPase deficiency.

In general, fructose loads can lead to large, rapid expansions in the hexose- and triose-phosphate pools, potentially providing increased substrate for all central carbon metabolic pathways, including glycolysis, glycogenesis, gluconeogenesis, lipogenesis, and oxidative phosphorylation. In FBPase deficiency, marked decrease in intracellular free phosphate due to hepatic accumulation of fructose 1,6-bisphosphate can inhibit glycogenolysis, leading to fructose induced hypoglycemia.

To date, several harmful mutations in the coding region of FBP1 have been reported. Santer et al. reviewed 35 different mutations, including 14 missense mutations, 12 deletion mutations, 4 nonsense mutations, 4 insertions/duplications, two splices, and one indel ⁵. Although the FBP1 c.581T > C (which results in the missense mutation F194S) and c.490G > A (which affects the neighboring nucleotide) mutations have been previously demonstrated 3^,17, the c.491G > A mutation (responsible for the missense mutation G164D) demonstrated in this study is novel. The G164 and F194 mutations, which are located in the β-strand (Fig. 1g and 3a), demonstrated the substitution of hydrophobic amino acids with hydrophilic amino acids (Table 1), suggesting certain conformational changes. The examinations carried out in this study confirmed that the FBP1 mutants with G164D and F194S decreased protein expression and resulted in a loss of FBPase enzyme activity.

The interactome analysis based on Liquid chromatography-tandem MS data for binding partners demonstrated that FBP1, particularly in its mutant forms, interacts with the proteins involved in the molecular chaperone related to unfolded protein response including heat shock protein (HSP). Protein misfolding has been recognized as an important pathophysiological cause of protein deficiency in some genetic disorders, such as Fabry disease, Pompe disease, and Gaucher disease ²⁸. Inherited mutations can break native protein folding, resulting in the formation of misfolded proteins that are consequently retained in the endoplasmic reticulum (ER). The heat-shock proteins (HSPs) are a family of molecular chaperones, which collectively form a network that is critical for protein folding ²². It has been suggested that HSP70 recognizes unfolded proteins, and HSP90 recognizes partially folded proteins, but fully folded proteins do not bind with either HSP70 or HSP90 ²³. Additionally, missense mutations have been shown to shift protein folding equilibrium toward a partially folded state, thus increasing the cellular fraction of HSP70 and HSP90 relative to WT proteins ²³. Some proteins can only be partially folded by HSP70 and therefore require additional assistance from HSP60 (chaperonin) in order to acquire a folded functional conformation ²². Aberrant protein aggregation was found to be controlled by chaperonins containing TCP-1. Based on the findings that FBP1 variants with G164D and F194S mutations have increased interactions with HSP70, HSP90, HSP60 and TCP1 and are partially aggregated and trapped in the ER, we conclude that the decrement in protein expression detected in FBPase deficiency is, at least in part, a result of protein misfolding.

Finally, 15 FBP1 missense mutations (Figs. 3, 4 and 5) were reviewed and classified into three categories. Type 1 causes loss of enzyme activity due to mutations in functional domain with unchanged protein expression and diffuse cytoplasmic localization. Type 1 mutations do not change their amino acid hydrophobicity. Type 2 causes loss of enzyme activity with reduction in protein expression and ER aggregation due to protein misfolding. Type 2 mutations change amino acid hydrophobicity, except for G294V. It is noted that G294V from the Type 2 category exhibits unchanges in its amino acid hydrophobicity, suggesting that G294 plays a role in FBP1 protein folding. Type 3 mutations are likely non-pathogenic mutations in regard to obvious phenotype of FBPase deficiency, whereas the relationship between these mutations and disease onset are obscure. This finding is consistent with clinical characteristics, but the further investigation could be needed to uncover the role of our defined sites.

In the context of genotype biochemical phenotype associations, previous in vitro studies evaluated only two missense mutations (D119N and G164S) and, consistent with our findings, D119N decreased enzymatic activity with no impact on protein expression ¹⁶, whereas G164S decreased protein expression ²⁹. Thus, G207R and V325A were defined as Type 3 mutations. The G207R mutation, wherein the second allele exhibited a deletion of exon 8, was present in FBPase deficiency ⁵. This variant has been previously reported in the heterozygous state with an allele frequency of 0.0001498 in 10 European (non-Finnish) individuals ⁵, although its pathological effects have not yet been elucidated. By contrast, the V325A mutation is not recorded in the dbSNP database. Although this mutation was observed in patients with FBPase deficiency, it was found to harbor the same mutant alleles as G164S. Based on an examination of the chimeric V325A mutation, Kikawa et al. suggested that this mutation does not play a pathogenic role in FBPase deficiency ³, which is consistent with the findings of our study. Besides missense mutations, several types of FBP1 mutations (12 deletion mutations, four nonsense mutations, four insertions/duplications, two splices, and one indel) have been reported ⁵. Previous in vitro studies reported that c.704delC, c.838delT, and c.960dupG all decreased truncated FBP1 protein expression ^16,29,30. However, the mechanism that leads to a decrease in truncated FBP1 protein expression has not been sufficiently elucidated. Considering our findings, protein misfolding may be involved in the underlying disease pathophysiology in these truncated types of FBP1 mutations, similar to Type 2 mutations.

The clinical applications of pharmacological chaperone therapy for Fabry disease have become increasingly popular. The chaperone molecules support the folding of mutated enzymes and increase their stability and activity, which may be broadly applicable to other protein deficiencies. Life-threatening episodes of hypoglycemia and lactic acidosis are sometimes triggered by fasting and febrile infections during childhood in cases of FBPase deficiency. Yet, to date, no preventive agent has been developed for these patients and their symptoms. Our findings indicate the possibility that certain patients with Type 2 mutation may respond to such kind of chaperone molecules, although Type 1 mutations are likely to be untreatable cases of FBPase deficiency. At any rate, further studies are warranted in order to fully understand the pathophysiology of this rare disease as well as to clarify the usefulness and efficacy of pharmacological chaperone therapy.

ER

endoplasmic reticulum

FBPase

Fructose-1,6-bisphosphatase

FBP1

FBP gene

HSP

heat-shock proteins

PCR

polymerase chain reaction

VAF

variant allele frequency

WT

wild-type

Competing interests

The authors declare no competing financial interests.

Rahil JF, de Maine MM, Benkovic SJ. Rapid-quench and isotope-trapping studies on fructose-1,6-bisphosphatase. Biochemistry. 1982;21(14):3358-3363.
Baker L, Winegrad AI. Fasting hypoglycaemia and metabolic acidosis associated with deficiency of hepatic fructose-1,6-diphosphatase activity. Lancet. 1970;2(7662):13-16.
Kikawa Y, Inuzuka M, Jin BY, et al. Identification of genetic mutations in Japanese patients with fructose-1,6-bisphosphatase deficiency. American journal of human genetics. 1997;61(4):852-861.
Nyhan W BB, Al-Aqeel A. Fructose-1,6-diphosphatase deficiency. In: Atlas of Inherited Metabolic Diseases. London: Hodder Arnold; 2012:354-358.
Santer R, du Moulin M, Shahinyan T, et al. A summary of molecular genetic findings in fructose-1,6-bisphosphatase deficiency with a focus on a common long-range deletion and the role of MLPA analysis. Orphanet journal of rare diseases. 2016;11:44.
Lebigot E, Brassier A, Zater M, et al. Fructose 1,6-bisphosphatase deficiency: clinical, biochemical and genetic features in French patients. Journal of inherited metabolic disease. 2015;38(5):881-887.
Yoshida K, Sanada M, Shiraishi Y, et al. Frequent pathway mutations of splicing machinery in myelodysplasia. Nature. 2011;478(7367):64-69.
Cong L, Ran FA, Cox D, et al. Multiplex genome engineering using CRISPR/Cas systems. Science. 2013;339(6121):819-823.
Kikawa Y, Shin YS, Inuzuka M, Zammarchi E, Mayumi M. Diagnosis of fructose-1,6-bisphosphatase deficiency using cultured lymphocyte fraction: a secure and noninvasive alternative to liver biopsy. Journal of inherited metabolic disease. 2002;25(1):41-46.
Sakuma I, Higuchi S, Fujimoto M, et al. Cushing Syndrome Due to ACTH-Secreting Pheochromocytoma, Aggravated by Glucocorticoid-Driven Positive-Feedback Loop. The Journal of clinical endocrinology and metabolism. 2016;101(3):841-846.
Tamura A, Ogasawara T, Fujii Y, et al. Glucagonoma With Necrolytic Migratory Erythema: Metabolic Profile and Detection of Biallelic Inactivation of DAXX Gene. The Journal of clinical endocrinology and metabolism. 2018;103(7):2417-2423.
Tanaka T, Ohkubo S, Tatsuno I, Prives C. hCAS/CSE1L associates with chromatin and regulates expression of select p53 target genes. Cell. 2007;130(4):638-650.
Suzuki S, Tanaka T, Poyurovsky MV, et al. Phosphate-activated glutaminase (GLS2), a p53-inducible regulator of glutamine metabolism and reactive oxygen species. Proceedings of the National Academy of Sciences of the United States of America. 2010;107(16):7461-7466.
Hosokawa H, Romero-Wolf M, Yui MA, et al. Bcl11b sets pro-T cell fate by site-specific cofactor recruitment and by repressing Id2 and Zbtb16. Nat Immunol. 2018;19(12):1427-1440.
Ke HM, Zhang YP, Lipscomb WN. Crystal structure of fructose-1,6-bisphosphatase complexed with fructose 6-phosphate, AMP, and magnesium. Proceedings of the National Academy of Sciences of the United States of America. 1990;87(14):5243-5247.
Li N, Chang G, Xu Y, et al. Clinical and Molecular Characterization of Patients with Fructose 1,6-Bisphosphatase Deficiency. International journal of molecular sciences. 2017;18(4).
Matsuura T, Chinen Y, Arashiro R, et al. Two newly identified genomic mutations in a Japanese female patient with fructose-1,6-bisphosphatase (FBPase) deficiency. Molecular genetics and metabolism. 2002;76(3):207-210.
Kaur R, Dahiya L, Kumar M. Fructose-1,6-bisphosphatase inhibitors: A new valid approach for management of type 2 diabetes mellitus. European journal of medicinal chemistry. 2017;141:473-505.
Szklarczyk D, Franceschini A, Kuhn M, et al. The STRING database in 2011: functional interaction networks of proteins, globally integrated and scored. Nucleic acids research. 2011;39(Database issue):D561-568.
Wang F, Song W, Brancati G, Segatori L. Inhibition of endoplasmic reticulum-associated degradation rescues native folding in loss of function protein misfolding diseases. The Journal of biological chemistry. 2011;286(50):43454-43464.
Brown CR, Chiang HL. A selective autophagy pathway that degrades gluconeogenic enzymes during catabolite inactivation. Communicative & integrative biology. 2009;2(2):177-183.
Johnston CL, Marzano NR, van Oijen AM, Ecroyd H. Using Single-Molecule Approaches to Understand the Molecular Mechanisms of Heat-Shock Protein Chaperone Function. Journal of molecular biology. 2018;430(22):4525-4546.
Karras GI, Yi S, Sahni N, et al. HSP90 Shapes the Consequences of Human Genetic Variation. Cell. 2017;168(5):856-866 e812.
Herzog B, Morris AA, Saunders C, Eschrich K. Mutation spectrum in patients with fructose-1,6-bisphosphatase deficiency. Journal of inherited metabolic disease. 2001;24(1):87-88.
Herzog B, Wendel U, Morris AA, Eschrich K. Novel mutations in patients with fructose-1,6-bisphosphatase deficiency. Journal of inherited metabolic disease. 1999;22(2):132-138.
Afroze B, Yunus Z, Steinmann B, Santer R. Transient pseudo-hypertriglyceridemia: a useful biochemical marker of fructose-1,6-bisphosphatase deficiency. European journal of pediatrics. 2013;172(9):1249-1253.
Asberg C, Hjalmarson O, Alm J, Martinsson T, Waldenstrom J, Hellerud C. Fructose 1,6-bisphosphatase deficiency: enzyme and mutation analysis performed on calcitriol-stimulated monocytes with a note on long-term prognosis. Journal of inherited metabolic disease. 2010.
Parenti G. Treating lysosomal storage diseases with pharmacological chaperones: from concept to clinics. EMBO molecular medicine. 2009;1(5):268-279.
Moon S, Kim JH, Han JH, et al. Novel compound heterozygous mutations in the fructose-1,6-bisphosphatase gene cause hypoglycemia and lactic acidosis. Metabolism: clinical and experimental. 2011;60(1):107-113.
Kikawa Y, Inuzuka M, Jin BY, et al. Identification of a genetic mutation in a family with fructose-1,6- bisphosphatase deficiency. Biochemical and biophysical research communications. 1995;210(3):797-804.

Table 1. Categories of FBP1 missense mutations

Mutation

Location

Enzymatic Activity

Protein Expression

Intracellular

Aggregation

Chaperone Binding Ability

Type of Amino Acid

Hydropathy Index

Structure

Amino Acid Change

Compared to Wild-type

Localization

(%)

HSP70

HSP90

Wild-type

Mutant

Wild-type

Mutant

Type 1: Loss of enzyme activity due to mutations in functional domain with unchanged protein expression and diffuse cytoplasmic localization

D119N

Metal binding site

decrease

no change

1.4

Diffuse

(cytoplasm)

9.4

0.1

0.3

hydrophilic-

acidic

hydrophilic-

neutral

-3.5

β-strand

P120L

Linker lesion at metal and substrate binding site

decrease

no change

1.3

Diffuse

(cytoplasm)

8.5

1.1

0.5

hydrophobic-

aliphatic

hydrophobic-

aliphatic

-1.6

3.8

β-strand

N213K

Substrate binding site

decrease

no change

1.1

Diffuse

(cytoplasm)

13.3

1.5

0.5

hydrophilic-

neutral

hydrophilic-

basic

-3.5

-3.9

ND

E281K

Metal binding site

decrease

no change

0.9

Diffuse

(cytoplasm)

33.6

2.4

0.5

hydrophilic-

acidic

hydrophilic-

basic

-3.5

-3.9

turn

Type 2: Loss of enzyme activity with decreased protein expression and ER aggregation in association with hydrophobicity change

R158W

decrease

0.5

Aggregated

(in ER)

69.6

7.7

7.8

hydrophilic-

basic

hydrophobic-

aromatic

-4.5

-0.9

α-helix

G164D

decrease

0.4

Aggregated

(in ER)

71.4

13.7

15.8

hydrophobic-

aliphatic

hydrophilic-

acidic

-0.4

-3.5

β-strand

G164S

decrease

0.6

Aggregated

(in ER)

42.7

6.4

12.2

hydrophobic-

aliphatic

hydrophilic-

neutral

-0.4

-0.8

β-strand

A177D

decrease

0.2

Aggregated

(in ER)

73.9

24.0

37.5

hydrophobic-

aliphatic

hydrophilic-

acidic

1.8

-3.5

β-strand

F194S

decrease

0.4

Aggregated

(in ER)

62.6

15.4

19.8

hydrophobic-

aromatic

hydrophilic-

neutral

2.8

-0.8

β-strand

G260R

decrease

0.4

Aggregated

(in ER)

73.5

18.9

28.2

hydrophobic-

aliphatic

hydrophilic-

basic

-0.4

-4.5

ND

P284R

decrease

0.3

Aggregated

(in ER)

57.7

41.2

47.2

hydrophobic-

aliphatic

hydrophilic-

basic

-1.6

-4.5

α-helix

G294E

decrease

0.4

Aggregated

(in ER)

64.1

64.9

89.9

hydrophobic-

aliphatic

hydrophilic-

acidic

-0.4

-3.5

ND

G294V

decrease

0.6

Aggregated

(in ER)

65.9

6.1

3.5

hydrophobic-

aliphatic

hydrophobic-

aliphatic

-0.4

4.2

ND

Type 3: Likely non-pathogenic mutations in regards to biochemical phenotype

G207R

no change

1.0

Diffuse

(cytoplasm)

12.7

3.1

3.9

hydrophobic-

aliphatic

hydrophilic-

basic

-0.4

-4.5

ND

V325A

no change

1.0

Diffuse

(cytoplasm)

25.4

2.3

1.3

hydrophobic-

aliphatic

hydrophobic-

aliphatic

4.2

1.8

α-helix

Abbreviations: ND, no data; ER, endoplasmic reticulum

There is NO Competing Interest.

FBPaseDSupFiglegend.pdf
SupplementalTable.docx
SupplementalTable6.xlsx
Supplemental Table 6

Characterization based on genotype–biochemical phenotype association in fructose-1,6-bisphosphatase deficiency

Status:

Journal Publication

Version 1

Abstract

Purpose

Methods

Results

Conclusion

Figures

Introduction

Material And Methods

Mutation analysis

Whole-exome Sequencing

Construction Of Fbp1 Expression Vectors

Mutagenesis Of Fbp1 Plasmids

Generation Of Fbp1 Knockout Hepg2 Cells

Cell Culture And Transient Transfection

Fbpase Activity Assay

Real-time Rt-pcr Analysis

Immunoblot Analysis

Er Fractionation

Immunofluorescence Analysis

Immunoprecipitation Assay

Mass Spectrometry Sample Preparation

Proteomic Analysis And Database Search

Protein Structure Based On Protein Data Bank

Statistics

Study Approval

Results

Case presentation and whole-exome sequencing

FBP1 mutations with G164D and F194S aggregated in the endoplasmic reticulum due to protein misfolding

Examination of mutation genotype and functional phenotype in all FBP1 missense mutants

FBP1 missense mutations were categorized into three functional phenotypes

Discussion

Conclusions

Abbreviations

Declarations

References

Tables

Additional Declarations

Supplementary Files

Status:

Journal Publication

Version 1