Keywords

1 Introduction

Amino acid catabolism provides nitrogen for the synthesis of biologically important compounds like hormones and neurotransmitters as well as energy for the cell. This process occurs mainly in the liver and kidneys and the enzymatic pathway involved depends on the nature of the amino acid.

Phenylalanine and tyrosine are important for protein biosynthesis and intermediates in the biosynthesis of catecholamines. They are both catabolized through the tyrosine degradation pathway, which converts tyrosine into fumarate and acetoacetate, two substrates of the mitochondrial tricarboxilic cycle (TCA cycle) (see Chap. 2).

Fumarylacetoacetate hydrolase (FAH, E.C. 3.7.1.2) is the last enzyme of the tyrosine degradation pathway containing 420 amino acids in a homodimeric form (Mahuran et al. 1977; Phaneuf et al. 1991). Up to now, close to 100 mutations in the FAH gene have been associated with hereditary tyrosinemia type 1 (HT1; HGMD® Professional 2016.1, accessed on April 2016) (Fig. 3.1) (Angileri et al. 2015). Despite multiple efforts, no clear link between mutation (genotype) and HT1 phenotype has been found. This chapter will focus on molecular aspects of the FAH gene and its corresponding protein in addition to give a complete listing of all the mutations identified to date.

Fig. 3.1
figure 1

Schematic representation of the FAH gene region and the corresponding protein. The location of the 98 mutations identified on the FAH gene is presented. E exon, I intron

2 Architecture of the fah Gene

The FAH gene is localized on chromosome 15 (15q23–25) and consists of 14 exons spanning over 35 kb of DNA (Awata et al. 1994; Labelle et al. 1993; Phaneuf et al. 1991) (Fig. 3.1). All exon-intron junctions possess the 5′ splice donor (gt) and 3′ splice acceptor (ag) consensus sequence and the major splicing product of the FAH gene results in an mRNA of 1260 coding nucleotides (Labelle et al. 1993). Table 3.1 summarizes the length of exons and introns and the number of HT1 mutations found in each of these elements while Fig. 3.1 is a schematic representation of the FAH gene region and protein. Exons 9 and 12 have the largest clusters of HT1 disease-causing FAH mutations. Interestingly both of these exons contain metal and substrate binding sites.

Table 3.1 Characteristics of FAH gene and its corresponding protein

2.1 Fah Alternative Transcripts

Two minor alternative splicing products of the FAH gene have also been found in normal fibroblasts, namely del100 and del231 (Dreumont et al. 2005).

The del100 transcript lacks exon 8 and, as a consequence, the reading frame is shifted and a premature termination codon (PTC) appears in 3′ end of exon 10. While this transcript is subjected to nonsense-mediated mRNA decay (NMD), a small part of it is transcribed into a protein that shares the first 202 amino acids with FAH and as a stretch of 67 amino acids completely different in the C-terminal. The pattern of DEL100 expression differs from the one of FAH and its function remains unknown (Dreumont et al. 2005).

The del231 transcript is less abundant than del100 and lacks exons 8 and 9. The corresponding protein should be similar to FAH except for the 77 amino acids encoded by both exons. While this transcript is not subjected to NMD, the corresponding protein has never been observed (Dreumont et al. 2005). The identification of this transcript has led to the hypothesis that intron 8 would be removed before introns 7 and 9 during normal FAH splicing (Dreumont et al. 2005).

The biological relevance of these two minor transcripts has not been demonstrated further. Interestingly, the abundance of del100 and del231 transcripts changes in presence of mutations affecting splicing donors/acceptors sites (Q279R, c.707-1G>A, c.707-1G>C) or enhancer elements (N232N, V259L) and in presence of the nonsense mutation W262X (Table 3.2) (Dreumont et al. 2001, 2004; Perez-Carro et al. 2014; Morrow et al. submitted).

Table 3.2 Splicing mutations involved in HT1 disease

2.2 Splicing Mutations

Up to now, 25 FAH mutations associated with HT1 phenotype have been reported to affect splicing (Table 3.2). Among these, four are located at the exon side of the exon/intron junction (p.Q64H, p.Q279R. p.P281P and p.G305R) and three others are located within exons 8, 9 and 12 (p.N232N, p.V259L and p.G337S). While these later mutations do not alter core sequence elements of splicing, they are probably modifying exonic splicing enhancers (ESE) or silencers (ESS) sites. The importance of these sites in splicing efficiency is gaining increasing support (reviewed in Ward and Cooper 2010) and it was recently shown that 20–45% of pathogenic single nucleotide polymorphisms (SNPs) affect splicing (Wu and Hurst 2016). It is therefore likely that other HT1 causing mutations may affect splicing.

3 FAH Protein

FAH forms a homodimer that catalyzes the hydrolytic cleavage of a carbon-carbon bound in fumarylacetoacetate (FAA) to yield fumarate and acetoacetate. It is the first member of an expanding family of metalloenzymes characterized by a unique α/β fold and involving a Glu-His-water catalytic triad (Timm et al. 1999). Orthologs of FAH are found in different species and share a high degree of homology (Fig. 3.2). The protein can be separated in two distinct domains; the FAH N-terminal domain and the FAH C-terminal domain.

Fig. 3.2
figure 2

Alignment of FAH orthologs. The protein sequence were retrieved from the NCBI website; Homo sapiens (NP_000128.1), Mus musculus (NP_034306.2), Rattus Norvegicus (NP_058877.1), Drosophila melanogaster (NP_524830.2), Arabidopsis thaliana (NP_172669.2), Danio rerio (NP_955895.1 with N-terminal of Fisher et al. 2008), Caenorhabditis elegans (NP_509083.1), Xenopus tropicalis (NP_001107523.1), Macaca mulatta (NP_001244458.1). The multiple sequence alignment was performed with MUSCLE tool (http://www.ebi.ac.uk/Tools/msa/muscle/) from the EMBL-EBI website. Numbers at the bottom of the figure correspond to H. sapiens amino acids sequence. N-terminal and C-terminal domains of FAH are presented, as well as region of interest for FAH activity

3.1 FAH N-Terminal Domain

The N-terminal domain of FAH consists of ~100 residues that are encoded by exons 1–4 (Timm et al. 1999) (Pfam: PF09298, InterPro: IPR015377) (Fig. 3.3a). Little is known about this structural domain except that it forms a structure consisting of an SH3-like barrel. This domain is not involved in dimerization or in active site formation, but it could have a regulatory function given its contacts with the C-terminal domain (Bateman et al. 2001, 2007; Timm et al. 1999).

Fig. 3.3
figure 3

FAH region of interest and effect of nonsense mutations as well as deletion mutations causing frameshift. (a) Scaled schematic representation of FAH exons. FAH N-terminal domain (Pfam: PF09298), FAH C-terminal doamin (PF01557), Hxx… and …xxE: conserved motif of the active site (Ran et al. 2013). P post-translational modifications (Bian et al. 2014; Huttlin et al. 2010; Vaca Jacome et al. 2015), M metal binding sites, S substrate binding sites, A active site (Bateman et al. 2001, 2007; Timm et al. 1999). Most of these features can also be found on the Homo sapiens FAH UniProt entry (P16930). (b) Schematic representation of the nonsense mutations causing HT1. The corresponding mutations can be found in Tables 3.3 and 3.4. (c) Schematic representation of the FAH deletion mutations causing frameshift. The corresponding mutations can be found in Tables 3.3 and 3.5

The FAH N-terminal domain contains two identified post-transcriptionally modified amino acids (N-acetyl-S2 and phospho-S92), but the reasons/effects of these modifications have not been investigated (UniProt: P16930) (Huttlin et al. 2010; Vaca Jacome et al. 2015).

Among the nine exonic FAH mutations found in this domain (Fig. 3.1), five can be linked to aberrant mRNA processing and two results in the p.W78X nonsense mutation yielding a protein lacking the entire FAH catalytic C-terminal domain (Figs. 3.1 and 3.3b) (Table 3.3). Among the five mutations that can be linked to aberrant mRNA processing, the c.1A>G (p.M1?/exon 1) mutation results in the start codon loss, while p.Q64H (exon 2) has been shown to affect splicing by promoting the retention of 94 nucleotides from intron 2 and resulting in the apparition of a PTC after nine missense amino acids. In addition to these two documented mutations, p.S23P (exon 1), p.F62C (exon 2) and p.Q64fs (exon 2) are likely to affect splicing due to the predicted alteration of an ESE site (S23P), activation of an exonic cryptic donor site (F62C) and direct alteration of the wild-type cryptic donor site (p.Q64fs) (Desmet et al. 2009). However, no experimental data on mRNA and protein are available for the later three mutations preventing a conclusion on their real effect on FAH mRNA, protein stability or activity. Of note, S23P was proposed to possibly affect FAH dimerization (Heath et al. 2002) and recombinant F62C was found to be an insoluble/inactive protein (Bergeron et al. 2001). The two remaining disease-causing mutations of the N-terminal domain give rise to normal FAH mRNA and are therefore likely to have a structural effect. Indeed, p.N16I (exon 1) produces an inactive and insoluble protein as shown from expression analysis of FAH in patient liver extract and from recombinant expression in cultured cells (Bergeron et al. 2001; Phaneuf et al. 1992) while p.A35T (exon 2) was shown to be expressed at low level and to have a decreased activity both in cultured fibroblasts and liver extracts (Cassiman et al. 2009).

Table 3.3 HT1 causing mutations in the FAH N-terminal domain

3.2 FAH C-Terminal Domain

The FAH C-terminal domain is composed of ~300 residues that are encoded by exons 5–14 (Fig. 3.3a) (Pfam: PF01557, InterPro: IPR011234). It is shared between members of the FAH family of metalloenzymes and characterized by a ß-sandwich fold forming a deep pocket in the catalytic domain and containing a metal ion at its base (Ran et al. 2013; Timm et al. 1999). All enzymes of the FAH family share the ability to cleave C-C bond of their substrate through a Glu-His-water triad involving either a HxxE or Hxx…xxE motif. For extensive alignments between members of the FAH family please refer to (Ran et al. 2013).

The FAH C-terminal domain has functional roles in metal-ion binding, catalysis and dimerization. The dimer formation is needed to form the pocket of the active site, while multiple residues have been shown to be involved in FAH dimerization, the longest stretch of amino acids located at the dimer interface is spanning from the end of exon 5 through exon 6 and the first half of exon 7 (Timm et al. 1999). From the crystal structure of mouse FAH, it was shown that the metal ion is coordinated by residues present in exons 5, 7, 8 and 9 (Bateman et al. 2001, 2007; Timm et al. 1999). Moreover, based on the study of Ran and co-workers, the FAH active site corresponds to the Hxx…xxE motif (Ran et al. 2013) (Figs. 3.2 and 3.3a). The first part of this motif (Hxx…), which would correspond to the lid domain of the active site, is mainly located in exon 5 and the last part of the motif (…xxE) in exon 13. Also based on mouse FAH crystal structure, substrate binding sites are located in exon 5, 9 and 12 (Bateman et al. 2001, 2007; Timm et al. 1999). In addition to these features, phosphorylation of S309 (exon 11) and Y395 (exon 14) have been observed but the role of these post-translational modifications has not been investigated further (Bian et al. 2014). As can be seen on Fig. 3.3a, all these functional elements are spread all over the FAH C-terminal domain, which explains why even nonsense and deletions mutations located in FAH last exons are causing HT1. The listing of nonsense mutations causing HT1 is presented in Table 3.4 and the corresponding proteins are depicted in Fig. 3.3b, while the listing of deletion mutations causing frameshift is presented in Table 3.5 and proteins are depicted in Fig. 3.3c (see also Fig. 3.1 for localization of mutations in exons).

Table 3.4 Nonsense mutations found in FAH C-terminal domain
Table 3.5 Deletion mutations causing frameshift in FAH C-terminal domain

3.2.1 Missense HT1 Mutations Located in the FAH C-Terminal Domain

While the FAH C-terminal domain is ~3 times larger then the FAH N-terminal domain, it contains nearly eight times more disease causing mutations (69 versus 9) due to its importance for FAH function. As mentioned above, the two exons containing the most disease-causing mutations are exon 9 and 12 (Table 3.1 and Fig. 3.1). Based on the fact that it was recently shown that 20–45% of pathogenic SNPs affect splicing (Wu and Hurst 2016), FAH missense mutations of the C-terminal domain were separated according to experimental data and to their potential effect on splicing as determined by the Human Splicing Finder website (Desmet et al. 2009). Nineteen mutations were found to potentially affect splicing (Table 3.6), while 29 mutations were not (Table 3.7). Not surprisingly all mutations affecting important residues for FAH activity are found in Table 3.7.

Table 3.6 FAH C-terminal domain missense mutations that could affect splicing
Table 3.7 Missense mutations located in the FAH C-terminal domain and affecting protein function

4 Most Frequent FAH Mutations and Their Geographical Localization

In total, more than 98 FAH mutations have been reported at this time to cause HT1. This number will likely increase, since SNPs causing the disease are found in SNPs database such as the ones from the Exome Aggregation Consortium (ExAC) (URL: http://exac.broadinstitute.org) and NHLBI Exome Sequencing project (http://evs.gs.washington.edu/EVS/). For example, a W234X SNP is reported on the ExAC website (last access: April 2016). This mutation is located at the end of exon 8 and based on experimental data from Table 3.4 is likely to cause HT1.

The worldwide incidence of HT1 is relatively low, with 1/100,000 affected individual (Hutchesson et al. 1996). The population that possesses the highest incidence of HT1 is the French Canadian population of the Saguenay-Lac-Saint-Jean (SLSJ) region in the province of Quebec (Canada). Not surprisingly, the most frequent FAH mutation in SLSJ region (~90% of all the disease causing alleles; c.1062 + 5G>A (IVS12 + 5G>A)) is also the most frequent worldwide (32.3% of the reported alleles) (Table 3.8) (Angileri et al. 2015). Since c.1062 + 5G>A accounts for the third of the HT1 reported allele and due to the fact that it is reported in a wide range of ethnic groups, it is likely to be a very old mutation (Angileri et al. 2015). The second most frequent HT1 mutation encountered worldwide is c.554-1G>T (IVS6-1G>T) with a frequency of 16.4% (Table 3.8). While this mutation is not associated to a specific cluster, it is more prevalent in the Mediterranean region and in southern Europe (Angileri et al. 2015).

Table 3.8 HT1 most frequent mutations

Two other clusters of HT1 are found in the world. The first one is in the Finnish population of Pohjanmaa where the c.786G>A (p.W262X) represents ~88% of disease causing alleles (Angileri et al. 2015; Kvittingen 1991; Mustonen et al. 1997). The other cluster is in an immigrant population from Pakistan living in the United Kingdom (predominantly in Birmingham), and for which the c.192G>T (p.Q64H) mutation accounts for 42% of the alleles reported (Angileri et al. 2015; Hutchesson et al. 1998). These two mutations are also among the most frequent mutations worldwide, with frequencies of 5.6% and 4.3% (Table 3.8; third and fourth rank respectively). The geographical distribution of almost all of the FAH mutations has been the subject of a recent review (Angileri et al. 2015).

5 Correlation Between FAH Mutations and HT1 Phenotype

HT1 is classified in three different forms depending on the clinical phenotype of patients and the age of onset. The acute form presents before 2 months of age with acute liver failure, while the subacute form presents between 2 and 6 months of age with liver disease and the chronic form presents after 6 months of age with slowly progressive liver cirrhosis and hypophosphatemic rickets (Bergman et al. 1998; Mitchell et al. 2001; van Spronsen et al. 1994) (See Morrow and Tanguay, Chap. 2). However, despite multiple efforts, no clear genotype-phenotype relationships have been unveiled (Arranz et al. 2002; Bergman et al. 1998; Dursun et al. 2011; Rootwelt et al. 1996).

5.1 HT1 Pseudodeficiency

To date, one missense mutation (c.1021C>T, p.R341W) has been described as a pseudodeficiency variant since individuals homozygous for this mutation are healthy due to residual FAH activity while compound heterozygotes with another FAH mutation develop HT1 (Rootwelt et al. 1994b). This mutation does not change the mRNA level nor its size but it results in decreased amount of FAH protein with less activity than the wild-type protein (Bergeron et al. 2001; Rootwelt et al. 1994b), suggesting that a minimal requirement of FAH activity is needed to prevent HT1 disease.

5.2 Reversion of FAH Mutation

A mosaic pattern of FAH expression in liver of HT1 patients has been reported for four splicing mutations; c.192G>T (p.Q64H), c.836A>G (p.Q279R), c.1009G>A (p.G337S) and c.1062 + 5G>A (IVS12 + 5G>A) (Demers et al. 2003; Dreumont et al. 2001; Kvittingen et al. 1993, 1994; Poudrier et al. 1998). The presence of FAH positive nodules was shown to be due to the reversion of the primary point mutation (Demers et al. 2003; Kvittingen et al. 1993, 1994) and favored by the selective advantage that the reversion would provide (Demers et al. 2003). Interestingly, it was also shown that the severity of the disease is directly correlated with the extent of HT1 mutation reversion in the liver of HT1 patients (Demers et al. 2003).

6 Concluding Remarks

This report summarizes the available information for each of the FAH mutations reported in the literature and places them back in their molecular context. While it does not provide explanations for the effect of all mutations on FAH mRNA and protein, it does suggest new ways to look at them in addition to highlight the importance of splicing mutations in HT1.