Human Molecular Genetics, 2000, Vol. 9, No. 15 2341-2350
© 2000 Oxford University Press
Structural and functional analysis of mutations in alkaptonuria
1Centro de Investigaciones Biológicas CSIC, Velázquez 144, Madrid 28006, Spain, 2Department of Biochemistry and Molecular Biology, Indiana University School of Medicine, 635 Barnhill Drive, Indianapolis, IN 46202, USA and 3Unidad de Patología Molecular, Fundación Jiménez Díaz, Madrid, Spain
Received 10 July 2000; Accepted 7 August 2000.
 |
ABSTRACT |
|---|
 TOP ABSTRACT INTRODUCTIONRESULTSDISCUSSIONMATERIALS AND METHODSREFERENCES |
|---|
Alkaptonuria (AKU), the prototypic inborn error of metabolism, was the first human disease to be interpreted as a Mendelian trait by Garrod and Bateson at the beginning of last century. AKU results from impaired function of homogentisate dioxygenase (HGO), an enzyme required for the catabolism of phenylalanine and tyrosine. With the novel 7 AKU and 22 fungal mutations reported here, a total of 84 mutations impairing this enzyme have been found in the HGO gene from humans and model organisms. Forty-three of these mutations result in single amino acid substitutions. This mutational information is analysed here in the context of the HGO structure and function using kinetic assays performed using purified AKU mutant enzymes and the crystal structure of human HGO. HGO is a topologically complex structure which assembles as a functional hexamer arranged as a dimer of trimers. We show how the intricate pattern of intra- and inter-subunit interactions and the extensive surfaces required for subunit folding and association of this oligomeric enzyme can be inactivated at multiple levels by single-residue substitutions. This explains, in part, the predominance of missense mutations (67%) in AKU.
|
INTRODUCTION |
|---|
|
TOPABSTRACTINTRODUCTIONRESULTSDISCUSSIONMATERIALS AND METHODSREFERENCES |
|---|
Alkaptonuria (AKU; MIM 203500) has been a prototypic inborn error of metabolism since Garrod (1,2) reported his pioneering studies of the disease, the first to be interpreted as a human autosomal recessive trait. AKU results from a deficiency of homogentisate dioxygenase (HGO) (3), one of the six enzymes required for the catabolism of the aromatic amino acids phenylalanine and tyrosine. This deficiency results in the accumulation of homogentisate in the body fluids leading to the characteristic symptoms of AKU (black urine, ochronosis and arthritis) (4,5). The recent cloning of the HGO gene in humans (6) and other species (Fig. 1, legend) and its full genomic characterization (7) enabled us to show that AKU individuals are homozygotes or compound heterozygotes for loss-of-function mutations in the gene (6).
|
Despite the relatively low incidence of AKU, its marked allelic heterogeneity provides information on functionally important regions of the protein. Forty candidate causative AKU mutations [including seven reported here (Table 1, Fig. 1A, legend for references)] have been identified in the HGO gene on the basis of its absence from the normal population and/or on their co-segregation in homozygosis or compound heterozygosis with the disease phenotype. In addition, a predictably neutral polymorphism resulting in a His80Gln substitution has been found at high frequency in the normal population (8). Notably, 27 candidate mutations (67%) are missense mutations, providing valuable information on functionally relevant individual residues.
|
In addition to human AKU mutations, we use here a positive selection screen to select loss-of-function mutations identifying functionally relevant residues/regions in the Aspergillus nidulans gene encoding HGO. The filamentous fungus A.nidulans has a phenylalanine/tyrosine degradation pathway that is remarkably similar to that in humans. The high degree of amino acid sequence identity (45–50%) between the fungal and human enzymes in this pathway was crucial in identifying human genes for HGO (6,9) and maleylacetoacetate isomerase (10). The relative ease with which classical and reverse genetic analysis can be carried out in A.nidulans facilitated the isolation of a fungal knock-out (
fahA) strain lacking fumarylacetoacetate hydrolase (FAH; the final enzyme in the phenylalanine/tyrosine catabolic pathway). This strain serves as a metabolic model for the severe disease tyrosinaemia type 1 (HT1). The accumulation of toxic metabolites upstream of FAH and downstream of HGO prevents fahA cells from growing in the presence of phenylalanine. The fact that this conditionally lethal phenotype can be rescued by loss-of-function mutations in the gene encoding HGO (hmgA) is the basis of the positive selection screen used here and led us to conclude that alkaptonuria would prevent the fatal consequences of HT1 (11). This prediction has recently been verified by Manning et al. (12) using a mouse model of HT1. Mice that are homozygous for HT1 (Fah–/Fah–) and heterozygous for AKU (Hgo–/Hgo+) were found to develop clonal nodules of functional hepatocytes in which the Hgo+ allele had been inactivated by mutation. This in vivo suppression analysis in hepatocytes enabled an additional 12 deletion or truncation mutations and 10 missense mutations affecting HGO to be identified (12). This remarkable collection of human, fungal and mouse missense HGO mutations can now be understood in the context of the human HGO structure, recently reported by Titus et al. (13). The HGO subunit has a high content of ß-sheet structure and can be described as a compact 280 residue N-terminal domain and an open 140 residue C-terminal domain. Consistent with previous reports of multimeric association in solution (14,15), the HGO crystal structure revealed a hexameric subunit arrangement. The N-terminal domain contains a central ß-sandwich structure flanked by a ß-sheet that mediates interactions with the C-terminal domain of adjacent subunits to form a trimer that stacks base to base as a dimer of trimers. The essential Fe2+ cofactor is co-ordinated by a glutamate and two histidine side-chains in an active site formed by the C-terminal domain and the trimer interface.
Here we report an analysis of structure–function relationships of the enzyme deficient in AKU. In addition to providing a detailed analysis of the structural bases of AKU, this work raises interesting questions for future studies of this prototypic metabolic disease.
|
RESULTS |
|---|
|
TOPABSTRACTINTRODUCTIONRESULTSDISCUSSIONMATERIALS AND METHODSREFERENCES |
|---|
Characterization of novel mutations in AKU patients
We have identified seven novel HGO mutations in our AKU screening program (Table 1). These bring the total number of AKU mutations identified to 40 (Fig. 1, legend for references), which shows the remarkably high allelic heterogeneity found in this rare disease. One of the above mutations, INV13+1G
T, almost certainly results in a splicing defect, as it involves a substitution of the essentially invariant G at position +1 of a donor splice site. Two mutations predictably result in early HGO truncation (Table 1). Finally, four novel missense mutations have been identified, which brings the total number of reported missense mutations (Fig. 1) to 27. Thus, 67% of all AKU mutations cause single amino acid replacements.
Biochemical analysis of human AKU proteins
HGO is highly conserved through evolution, with a remarkable amino acid sequence conservation shared by HGO proteins from bacteria, fungi, plants, invertebrates, fish and man throughout the protein sequence (Fig. 1A). Therefore, residues in many different regions of the protein are likely to be essential for enzyme structure and function. In agreement, single amino acid changes resulting from the 27 human AKU missense mutations are evenly distributed in the HGO sequence (Fig. 1A). In most cases, these single amino acid substitutions affect strictly conserved positions.
To test the effects of a representative sample of these AKU mutations on enzyme function, 19 different His-tagged mutant HGO proteins (18 carrying single residue substitutions predictably resulting in a loss-of-function phenotype and one carrying the His80Gln neutral substitution) were expressed in Escherichia coli, purified by Ni2+ affinity chromatography and assayed for enzyme activity (Table 2). As expected, the H80Q protein showed a specific activity approximating the wild-type, which confirms that this allele is not pathogenic. In contrast, all AKU substitutions resulted in significant loss of activity. Some mutations also caused reduced solubility (suggesting that they interfere with proper protein folding). These data formally establish that the corresponding 18 missense mutations are loss-of-function, causative AKU mutations.
|
Thirteen affinity-purified mutant proteins showed null or very low levels of enzyme activity (Table 2). However, five mutant enzymes (E42A, Y62C, A122D, D153G and M368V) showed specific activities in the range of 22–37% of the wild-type. Steady-state kinetics parameters were determined for the recombinant wild-type enzyme and these five mutants. The Km and Vmax for the wild-type enzyme were
6 µM and 22.3 µmol/min/mg, respectively. Catalytic efficiencies (Vmax/Km) of the five mutant recombinant enzymes were reduced to 7–25% of the wild-type (Table 3). The prevalent Met368Val substitution resulted in a catalytic efficiency that was 14% of that of the wild-type.
|
Mutational analysis of HGO in the fungal model
An A.nidulans wild-type strain grows on either lactose or a mixture of phenylalanine and lactose as carbon source. In contrast, growth of a
fahA strain is normal on lactose, but it is largely prevented on lactose and phenylalanine due to the accumulation of toxic phenylalanine metabolites. Second-site suppressor mutations inactivating HGO [which are relatively frequent due to the mutagenic effect of fumarylacetoacetate (16)] can overcome this toxicity, leading to normal growth. Double mutant colonies can be distinguished among a background of single mutant fahA colonies by their markedly more vigorous growth and their accumulation of ochronotic pigment (Fig. 2). These clones were purified and their HGO-encoding genes were screened for mutations by direct sequencing of PCR-amplified hmgA– alleles, a task facilitated by the haploid nature of A.nidulans and the small size of the three hmgA introns (<100 bp in length). Twenty-two novel mutations were found (Table 4 and Fig. 1, red colour code) with this genetic screen until saturation was indicated by the recurrent finding of several mutations. Seven of these 22 mutations represented missense mutations. The hmgA14 allele truncates the protein beyond residue 14 and almost certainly represents a null allele. The growth phenotype of all mutants was indistinguishable from that of an hmgA14 strain, showing that all the above mutations belong to the loss-of-function class. For all missense mutant strains, close meiotic linkage of the suppressing mutation to the hmgA locus was demonstrated.
|
|
As in humans, most fungal missense mutations affect fully conserved residues. For simplicity, residues affected by fungal mutations will be referred to according to human HGO numbering (Fig. 1A and Table 4 for their precise numbering in the fungal sequence). Of these fungal missense mutations, five affect amino acid positions not altered in AKU patients, whereas hmgA6 affects His292, also altered by the AKU H292R mutation, and hmgA8 involves a substitution in position 189, also affected by the AKU mutation S189I (Fig. 1 and Table 4). In addition to these A.nidulans mutations, a similar in vivo selection procedure carried out by Manning et al. (12) identified nine single residue substitutions inactivating mouse HGO (Fig. 1A).
The structural bases of AKU
A total of 43 single residue substitutions impairing HGO enzyme activity have been identified in AKU patients and model organisms. We next tried to understand the biochemical and genetic data in the context of the recently determined HGO crystal structure (13). The uniform distribution of residues affected by human, mouse and A.nidulans missense mutations throughout the primary structure (Fig. 1A) is maintained in the three-dimensional structure of the HGO subunit (Fig. 1B). The HGO structure revealed a hexameric subunit assembly arranged as a dimer of trimers (Fig. 3A, B and D). Mutations identified in human, fungal and mouse HGO have been grouped according to predicted structural consequences (described below).
|
Mutations directly affecting active site residues
Six mutations affect residues that directly co-ordinate the Fe2+ cofactor or that are predicted to have essential roles in catalysis (13). The iron cofactor, located at the bottom of a V-shaped structure (Fig. 3A), is co-ordinated by the side chains of the fully conserved residues His335, His371 and Glu341. These residues are altered by the human AKU H371R mutation, by the mouse E341D and H371D mutations and by the fungal hmgA10 (H335Y) mutation (Fig. 1A), which strongly supports the crucial role attributed to these amino acids. The essential role of iron agrees with the complete absence of activity of purified H371R HGO (Table 2). The critical role in catalysis attributed to H292 as an active site base (13) is strongly supported by the finding of novel missense mutations affecting this residue in both AKU patients (H292R) and fungi (hmgA6, H292Q) (Fig. 1A, Tables 1 and 4).
Mutations indirectly affecting the active site
Several mutations affecting residues in close intramolecular proximity to the iron cofactor impair enzyme activity by disturbing the active site conformation. Examples are the Finnish AKU mutation R330S (17) and the mouse mutation R330G. The recurrent finding of Arg330 substitutions underscores the predicted structural role of this residue in stabilizing the conformation of the two loops that connect the right and left sides of active site V (Fig. 3A) through potential electrostatic interactions of the Arg330 guanidinium group with the Glu389 carboxyl group and the main chain oxygens of Ala325 and Ala387, so that R330S and R330G mutations are expected to have a strong disruptive effect on the active site by altering the conformations of the His335 iron ligand and the predicted substrate binding site residues Pro332 and Tyr333 (Fig. 3A) (13). The loss of activity caused by the Arg330Ser substitution (Table 1) is also consistent with the central importance attributed to Arg330 in HGO function. The mouse T328I and T328P mutations are likely to have similar effects on Pro332, Tyr333 and His335 by eliminating a hydrogen bond between the Thr328 side chain and the Asp326 main chain that stabilizes a ß-turn between residues Ala325 to Thr328 (Fig. 3A). The His335 and His371 iron ligands are also likely to be perturbed by the mouse G372S mutation, which affects a neighbouring glycine residue having 
and 
values that are disallowed for serine residues. Finally, the Glu341 conformation may also be affected by the AKU W97G mutation. Trp97 is located in the ß-sheet forming one side of the active site (Fig. 3A) on a ß-strand adjacent to that containing His365, a residue that mediates the correct positioning of the Glu341 Fe2+ ligand.
Mutations interfering with the folding of the HGO subunit
Eighteen of 44 (43 if only single amino acid replacements are considered) missense mutations predictably result in subunit misfolding. In agreement, eight of ten mutant AKU proteins in this group show reduced solubility in the E.coli expression system (data not shown). The relatively high abundance of missense mutations in this class possibly reflects the topological complexity of the subunit structure and the intimate association between the N- and C-terminal domains required for the correct folding of the protomer. An example of this intimate association is a knot in which C-terminal domain residues 406–410 pass through a loop between residues 77–110 in the N-terminal domain (Fig. 3A) (13). The Trp97 and Val300 side chains, respectively, affected by the AKU W97G and V300G mutations, are located in the saddle-like hydrophobic core between the N- and the C-terminal domains (Fig. 3A). Although the W97G and V300G mutations would both be expected to cause misfolding, Trp97 is also in close proximity to the active site (see above) and to the structural knot, with a 4.8 Å distance separating the Trp97 and Ser416 
-carbon atoms. Therefore, W97G would be expected to have even greater effects on folding and activity than V300G, in agreement with the markedly reduced activity (Table 2) and solubility (data not shown) of this mutant protein. Other mutations which would disrupt the intra-subunit hydrophobic structure are AKU mutations V181F, F227S, P230S and P230T, fungal mutations hmgA24(L163P) and hmgA22 (Fig. 1A and Table 4) and mouse mutation W268G. Subunit misfolding may also result from mutations disrupting electrostatic interactions. For example, D153G and K248R AKU mutations disrupt side-chain interactions with His207 and Arg184 and with Asp151 and Glu202, respectively. Similarly, the mouse mutation K276E is expected to destabilize the central jelly roll structure by introducing charge repulsion in place of an interaction with Glu178.
Noticeably, the Ser189 residue affected by AKU mutation S189I is a solvent-exposed residue. The essential requirement of a polar amino acid in this position is underscored by the isolation of hmgA18 (R189P), which involves the Arg residue in the fungal protein in equivalent position to human Ser189. Finally, subunit misfolding may also result from the introduction by mutation of unfavourable steric contacts, as predicted for substitutions of glycine to arginine resulting from AKU G161R, G270R, G360R and fungal hmgA17 (G170R) mutations.
Mutations affecting intersubunit interactions
A hexameric association strongly influences the HGO monomer structure (13) and the conformation of the active site. Monomers are organized about three-fold crystallographic axes. In the resulting trimers, the N-terminal ß-sheet of each subunit contacts the C-terminal domain of its adjacent subunit (Fig. 3B). The disk-like trimers stack base-to-base about two-fold axes to form hexamers (Fig. 3D and E).
Nine missense mutations affect residues involved in interactions between subunits in the trimeric assembly, which strongly supports the conclusion (13) that the conformation of the active site is dependent on the HGO quaternary structure. Extensive hydrophobic interactions between neighbouring subunits in the trimeric assembly and the formation of an inter-subunit ß-sheet appear crucial for positioning the His iron ligands and other active site residues (Fig. 3B). Residues Tyr334, His335 and Arg336 of the C-terminal domain (Fig. 3A) contribute a ß-strand to a five-stranded inter-subunit ß-sheet (Fig. 3B) by hydrogen-bonding with the N-terminal residues Gln43 to Ser45 of a three-fold related subunit (i.e. adjacent subunit in the trimeric assembly). An extensive hydrogen bond–salt bridge network involving the Glu42, Arg63 and Glu168 side chains of one subunit and the Arg336 side chain of a three-fold related subunit (Fig. 3C) would be disrupted by the AKU E42A and E168K mutations and the fungal hmgA16 (R336G) mutation (Fig. 1A and Table 4). The importance of these interactions for HGO activity is also demonstrated by the 14-fold decrease in catalytic efficiency resulting from the E42A mutation (Table 3). AKU mutation D291E similarly affects electrostatic contacts (13). Five AKU mutations (L25P, W60G, Y62C, A122D, M368V) and the mouse mutation V266G/A267P predictably affect hydrophobic contacts between subunits in the trimeric assembly. W60G is the most disruptive of these mutations, leading to nearly null enzyme activity. Trp60 is in Van der Waals contact with Tyr333 of a three-fold related molecule that forms one side of the predicted box-shaped substrate-binding pocket (Fig. 3A). The Y62C, A122D and M368V mutations result in 4- to 7-fold reductions in catalytic efficiency (Table 3). Thus, consistent with the importance of intersubunit contacts to the active site conformation, perturbation of the interfaces between subunits in the HGO trimer adversely affects catalysis.
The AKU mutations R225H and I216T affecting residues located in the inter-trimer interface (Fig. 3D and E) strongly supported the conclusion that the functional form of human HGO is the hexamer (13). R225H and I216T mutant HGO proteins show virtually null enzyme activity (Table 2). However, these variants showed increased solubility compared with the wild-type (data not shown) indicating that these substitutions are unlikely to result in general misfolding effects. The novel AKU mutations, R53W and W322R (Table 1), and the mouse D294N mutation (12) also involve residues located in interfaces buried between trimers (Fig. 3E). The Arg53 guanidine donates two direct hydrogen bonds to the two-fold related (i.e. located in an adjacent subunit belonging to the other disk-like trimer in the hexamer) Pro211 and a water-mediated hydrogen bond to the two-fold related residues Gly220 and Ala222. The Trp322 side-chain donates a hydrogen bond to the Glu326 carboxyl group of a two-fold related subunit and is in Van der Waal’s contact with the Pro388 ring. Trp322, as Ile216, is close to the two-fold axes that relate trimers within the hexamer. The two-fold crystallographic axes pass within 1.9 Å of the Pro388 CB atoms. The Asp294 carboxyl is involved in a hydrogen bond network with the Tyr333 hydroxyl group and the main-chain nitrogens of the two-fold related Ala218 and Asn219. D294N would disrupt at least one of the hydrogen bonds, altering the position of Tyr333 and/or destabilizing interactions between trimers. It is also important to bear in mind that several mutations could perturb contacts between both two- and three-fold related subunits. For example, Asp294 is also in Van der Waal’s contact with the three-fold related Phe49 side chain. Thus, the clustering of four AKU mutations and a mouse mutation between trimers and the virtually null enzyme activity resulting from two of these mutations provide strong support for the hexamer as the functional form of HGO. In the absence of hexameric association, the Arg225, Ile216, Arg53 and Trp322 side chains would be solvent-accessible and predictably tolerant of mutation.
C-terminal truncating mutations
Five of the 21 mouse mutations truncate the 445-residue HGO protein downstream of residue 392 and an additional deletion removes eight residues close to the C-terminus (Fig. 1A). Two A.nidulans mutations truncate the protein after positions corresponding to human residues 426 and 430 (Fig. 1A). These mutations suggest a critical role of C-terminal residues in HGO function.
An -helix between residues 409 and 416 occurs beyond the structural knot (described above; Fig. 3A) and the polypeptide becomes disordered between residues 418 and 429. The polypeptide becomes ordered again between residues 430 and 440, protruding away from the HGO subunit to contact an adjacent three-fold related subunit (Fig. 3A and B). Residues 430–435 form an extended section of polypeptide that packs in the vicinity of the three-fold related residues 38, 40, 64, 65, 69, 234 and 266. Residues Leu430, Lys431, His433 and Phe434 are strictly conserved between human and A.nidulans and strongly conserved between all species. The polypeptide undergoes a 90° turn between residues 435 and 438 to bring residues 437–440 in contact with the three-fold related residues 21–23 and 36–37.
Although mouse truncating mutations presumably result in general structural defects, the more carboxyl fungal C-terminal truncations will result in the loss of the interactions of residues Leu430 to Asn440 with residues in the adjacent (three-fold related) subunit. The important role of these residues is underscored by a mouse deletion mutation removing residues 430–438 (Fig. 1A) and the double missense mouse mutation V266G/A267P, affecting the contact of Phe434 with Val266. Loss of these interactions predictably destabilizes the HGO trimer.
|
DISCUSSION |
|---|
|
TOPABSTRACTINTRODUCTIONRESULTSDISCUSSIONMATERIALS AND METHODSREFERENCES |
|---|
The allelic heterogeneity found in AKU, the availability of selective screens for loss-of-function HGO mutations in model organisms and the recent resolution of the crystal structure of the HGO enzyme make AKU ideally placed to understand structure–function relationships in an inherited metabolic disease. HGO almost certainly functions as a hexamer in solution, with a quaternary structure arranged as a dimer of trimers. The relatively high proportion of missense mutations leading to loss of enzyme function, their even distribution throughout the primary and tertiary structures and the high interspecies HGO sequence conservation are all consistent with the location of the HGO active site in the context of the hexameric structure. The extensive surfaces required for folding and association of individual subunits to form an intersubunit active site suggest that the odds of a single substitution causing a loss-of-function are high.
We have used an E.coli overexpression/poly-His affinity tag system to determine the enzymatic activity of the wild-type and 19 purified single-residue substitution variants of human HGO. The apparent Km of purified human HGO is 6.2 µM, which contrasts with the previously reported value of 188 µM for the mouse enzyme (15). The fact that the specific activity of our purified enzyme is several orders of magnitude higher than that reported for the purified mouse enzyme strongly suggests that the latter was assayed in suboptimal conditions. This could possibly be due to the long purification scheme required which might have led to enzyme inactivation or perhaps to selection of a dissociated, largely inactive form of the enzyme, as this mouse enzyme was reported to be trimeric (15).
Of the 19 purified mutant proteins tested for activity, only that carrying a His80Gln substitution showed an activity approximating that of the wild-type. This confirms the prediction that H80Q is a neutral polymorphism (8). All other 18 single-residue substitutions resulted in loss of enzyme activity. Thirteen such proteins showed null or very low activity. Eight of these, including seven (W97G, G161R, S189I, F227S, P230S, P230T and V300G) which are predicted to interfere with proper subunit folding, showed reduced solubility in the E.coli expression system. Therefore, although the null or nearly null enzyme activity (Table 1) would account for the loss-of-function phenotype of these alleles, we cannot rule out that these mutations primarily result in misfolding leading to increased instability and degradation of the mutant polypeptides in the liver. The other five mutant proteins retained substantial activity, with catalytic efficiencies (estimated as Vmax/Km) in the range of 7–25% of the wild-type.
The hexameric enzyme structure raises interesting questions regarding the disease phenotype caused by these apparently partial loss-of-function mutations and the recessive nature of AKU. If only the wild-type hexamer were active and the steady-state level of both allelic variants were similar, the fact that heterozygote carriers are healthy would suggest that only 1/64 activity is required for a normal phenotype (assuming equal synthesis from both alleles). However, this conclusion may be jeopardized by exclusion or degradation of mutated subunits (see Fig. 4 in ref. 18) or by tolerating mutated subunits within hexamers to yield a sufficient number of functional active sites. In any case, it is hardly conceivable that enzyme efficiencies as high as 25% of the wild-type do not suffice for a normal phenotype.
The five alkaptonuria mutations which in our functional assay showed catalytic activities in the range of 7–25% of the wild-type are the prevalent mutation M368V (which has been found both in homozygosity and compound heterozygosity) (8,19,20) and four rare mutations, each detected in just one family. These were A122D (found in compound heterozygosity with M368V), D153G (in compound heterozygosity with an undetected mutation) and E42A and Y62C (both found in homozygosity) (8). These data rule out that a particular combination of alleles is required for these mutations to cause AKU, which a priori might have been a possible pitfall in our functional assay (in which we have not expressed ‘hybrid’ hexamers). One of these five mutations, D153G, is predicted to interfere with the proper folding of the HGO subunit. Here, the primary effect might be accelerated proteolytic turnover of the mutant polypeptide, as demonstrated for certain phenylketonuria alleles (21,22). The other four mutations (M368V, Y62C, E42A and A122D) are predicted to impair trimer assembly. High level bacterial overexpression of these four mutant proteins may artificially drive the formation of less stable HGO hexamers as described by Waters et al. (23) on phenylalanine hydroxylase, that would not normally form at the monomer concentration found in human liver. These unstable hexamers would not be fully dissociated at the enzyme concentrations required for in vitro enzyme assays.
The analysis of HGO presented here represents a significant advance in our understanding of AKU. This work demonstrates that this oligomeric enzyme requires an intricate pattern of intra- and inter-subunit interactions for activity that can be inactivated at multiple levels by single-residue substitutions. Additionally, this work provides a solid framework for directing and interpreting future studies of the concentration and required minimum level of HGO activity and degradation of mutant variants in vivo and the association state and stability of HGO variants in solution.
|
MATERIALS AND METHODS |
|---|
|
TOPABSTRACTINTRODUCTIONRESULTSDISCUSSIONMATERIALS AND METHODSREFERENCES |
|---|
AKU mutation screening
DNA of AKU probands were obtained from peripheral blood lymphocytes after informed consent. Mutation screening is described by Beltràn-Valero de Bernabé et al. (19).
Recombinant plasmids
Plasmid D1:HGO encodes human HGO tagged with the peptide MGH10SSGHIDDDDKHMGS at its N-terminus. The protein is expressed under the control of an IPTG-inducible T7 gp10 promoter. This plasmid was used as template for the introduction of AKU missense mutations by PCR, using the Quick Change (Stratagene, La Jolla, CA) system and appropriate mutagenic oligonucleotides. The HGO gene in mutated expression constructs was fully sequenced to rule out the presence of secondary mutations resulting from PCR amplification.
Expression and purification of recombinant proteins
Expression constructs were constructed in E.coli DH1 and transformed into E.coli BLB21(DE3) pLysS for high level expression. Bacterial clones were inoculated into luria broth (LB) cultures in the presence of 34 µg/ml chloramphenicol and 50 µg/ml ampicillin which were grown at 22°C until OD600 reached 0.6. HGO synthesis was induced on addition of 10 µM IPTG and the cultures were incubated for a further 14 h. Bacteria were collected, washed, resuspended in lysis buffer (50 mM HEPES, 0.5 M NaCl, 5 mM imidazol, 1% Triton X-100, 1 mM benzadimine pH 7.5) and lysed with a French Press. The lysate was cleared by centrifugation for 12 min at 17 000 g and 4°C and passed through a 0.45 µm filter. The soluble, His-tagged protein was purified by Ni2+ affinity chromatography, using 50 mM HEPES, 0.5 M NaCl, 0.5 M imidazol pH 7.5 and dialysed against 20 mM Tris–HCl pH 7.5, 0.5 M NaCl and 0.1% ß-mercaptoethanol. The purity of enzyme preparations was monitored by 10% polyacrylamide–SDS electrophoresis and Coomassie staining.
HGO assays
HGO was freshly prepared and reconstituted with 1 mM FeSO4 and 2 mM ascorbate immediately before enzyme assays, which were made at 37°C. The formation of the reaction product, malelylacetoacetate was monitored every 0.1 s at 330 nm (extinction coefficient 13 500/M/cm), using 1 ml reaction mixtures containing 50 mM potassium phosphate, 0.2 mM FeSO4, 2 mM ascorbate pH 7.4 and 0.2 mM homogentisate (Sigma, St Louis, MO) as a substrate. Steady-state assays to determine kinetic parameters were measured using at least five different substrate concentrations in the range of 0.004–0.4 mM (0.02 for the weight). Maximum velocity (Vmax) and Km values for mutant proteins showing significant activity were deduced from Lineweawer–Burk plots obtained with the Shimazu UVPC Kinetics software v. 2.7 (Shimazu, Germany).
Mutagenesis analysis of A.nidulans HGO
Aspergillus nidulans was grown in appropriately supplemented minimal medium (9,11). Standard genetic markers were used (24). The genotype of the parental strain for the isolation of hmgA mutants was biA1, methG1, argB2, fahA:argB+. Approximately 104 conidiospores of this strain were seeded per plate of medium containing 0.05% (w/v) lactose and 25 mM phenylalanine as carbon sources. After 3–5 days at 37°C, healthy colonies carrying second-site suppressors of the null fahA mutation were observed. Among these clones, those carrying a mutation in the hmgA gene (encoding HGO) were clearly identified by the accumulation of ochronotic pigment around the colonies. Such clones were purified by single spore isolation on glucose minimal medium and replica plating on minimal lactose/phenylalanine medium to confirm their phenotype. Mutant hmgA genes were amplified from genomic DNA by PCR using appropriate external primers and the mutations determined by direct sequencing of the PCR-amplified product. To confirm the genetic linkage of the supressor mutations to hmgA, mutant biA1 methG1 argB2 fahA:argB+ hmgA– (?) strains were crossed with a yA2 inoB2 argB2 strain and the meiotic progeny analysed. Recombinant progeny were in all cases <2% of the parental progeny, indicating tight linkage of the suppressor mutations to the deletion fahA allele (i.e. to hmgA, as the divergent fahA and hmgA open reading frames are separated by only 414 bp). This analysis was carried out with all strains in which missense mutations were detected as well as with hmgA18 and hmgA21 strains.
|
ACKNOWLEDGEMENTS |
|---|
We thank R. Aquaron, A. Cabral, M. Seaman and F. Gayoso-Gómez for their donation of blood (DNA) samples of AKU patients, E. Reoyo for technical assistance and M.C. Estébanez and J.M. Fernández-Cañón for their contribution to the initial stages of this work. This work was supported by grants from the Spanish CICYT (SAF97-1789-E, SAF-99/0013), the Comunidad de Madrid (08.6/0015/1997), the NIH and the Midwest Affiliate of the American Heart Association.
|
FOOTNOTES |
|---|
+ These authors contributed equally to this work
§ To whom correspondence should be addressed. Tel: +34 91 5644562 ext. 4358; Fax: +34 91 5627518; Email: penalva@cib.csic.es
|
REFERENCES |
|---|
|
TOPABSTRACTINTRODUCTIONRESULTSDISCUSSIONMATERIALS AND METHODSREFERENCES |
|---|
1 Garrod, A.E. (1902) The incidence of alkaptonuria: a study in clinical individuality. Lancet, 2, 1616–1620.
2 Garrod, A.E. (1908) The Croonian Lectures on inborn errors of metabolism. Lecture II. Alkaptonuria. Lancet, 2, 73–79.
3 La Du, B.N., Zannoni, V.G., Laster, L. and Seegmiller, J.E. (1958) The nature of the defect in tyrosine metabolism in alkaptonuria. J. Biol. Chem., 230, 251–260.
4 Knox, W.E. (1958) Sir Archibald Garrod’s ‘inborn errors on metabolism’. II, Alkaptonuria. Am. J. Hum. Genet., 10, 95–124.[Web of Science][Medline]
5 La Du, B.N. (1994) Alkaptonuria. In Scriver, C.R., Beaudet, A.L., Sly, W. and Valle, D. (eds), The Metabolic Basis of Inherited Disease. MacGraw-Hill, New York, NY, pp. 1371–1379.
6 Fernández-Cañón, J.M., Granadino, B., Beltrán-Valero de Bernabé, D., Renedo, M., Fernández-Ruiz, E., Peñalva, M.A. and Rodríguez de Cordoba, S. (1996) The molecular basis of alkaptonuria. Nature Genet., 14, 19–24.[Web of Science][Medline]
7 Granadino, B., Beltrán-Valero de Bernabé, D., Fernández-Cañón, J.M., Peñalva, M.A. and Rodríguez de Córdoba, S. (1997) The human homogentisate 1,2-dioxygenase (HGO) gene. Genomics, 43, 115–122.[Web of Science][Medline]
8 Beltrán-Valero de Bernabé, D., Granadino, B., Chiarelli, I., Porfirio, B., Mayatepek, E., Aquaron, R., Moore, M.M., Festen, J.J., Sanmarti, R., Peñalva, M.A. and Rodríguez de Córdoba, S. (1998) Mutation and polymorphism analysis of the human homogentisate 1,2-dioxygenase gene in alkaptonuria patients. Am. J. Hum. Genet., 62, 776–784.[Web of Science][Medline]
9 Fernández-Cañón, J.M. and Peñalva, M.A. (1995) Molecular characterization of a gene encoding a homogentisate dioxygenase from Aspergillus nidulans and identification of its human and plant homologues. J. Biol. Chem., 270, 21199–21205.
10 Fernández-Cañón, J.M. and Peñalva, M.A. (1998) Characterization of a fungal maleylacetoacetate isomerase gene and identification of its human homologue. J. Biol. Chem., 273, 329–337.
11 Fernández-Cañón, J.M. and Peñalva, M.A. (1995) Fungal metabolic model for human type I hereditary tyrosinaemia. Proc. Natl Acad. Sci. USA, 92, 9132–9136.
12 Manning, K., Al-Dhalimy, M., Finegold, M. and Grompe, M. (1999) In vivo suppressor mutations correct a murine model of hereditary tyrosinemia type I. Proc. Natl Acad. Sci. USA, 96, 11928–11933.
13 Titus, G. P, Mueller, H.A., Burgner, J., Rodríguez de Córdoba, S., Peñalva, M.A. and Timm, D.E. (2000) The crystal structure of human homogentisate dioxygenase. Nature Struct. Biol., 7, 542–546.[Web of Science][Medline]
14 Adachi, K., Iwayama, Y., Tanioka, H. and Takeda, Y. (1966) Purification and properties of homogentisate oxygenase from Pseudomonas fluorescens. Biochim. Biophys. Acta, 118, 88–97.[Medline]
15 Schmidt, S.R., Müller, C.R. and Kress, W. (1995) Murine liver homogentisate 1,2-dioxygenase. Purification to homogeneity and novel biochemical properties. Eur. J. Biochem., 228, 425–430.[Web of Science][Medline]
16 Jorquera, R. and Tanguay, R. (1997) The mutagenicity of the tyrosine metabolite, fumarylacetoacetate, is enhanced by glutathione depletion. Biochem. Biophys. Res. Commun., 232, 42–48.[Web of Science][Medline]
17 Beltrán-Valero de Bernabé, D., Peterson, P., Luopajarvi, K., Matintalo, P., Alho, A., Konttinen, Y., Krohn, K., Rodríguez de Córdoba, S. and Ranki, A. (1999) Mutational analysis of the HGO gene in Finnish alkaptonuria patients. J. Med. Genet., 36, 922–923.
18 Scriver, C.R. and Waters, P.J. (1999) Monogenic traits are not simple: lessons from phenylketonuria. Trends Genet., 15, 267–272.[Web of Science][Medline]
19 Beltrán-Valero de Bernabé, D., Jiménez, F.J., Aquaron, R., and Rodríguez de Cordoba (1999) Analysis of alkaptonuria (AKU) mutations and polymorphisms reveals that the CCC sequence motif is a mutational hot spot in the homogentisate 1,2-dioxygenase gene (HGO). Am. J. Hum. Genet., 64, 1316–1322.[Web of Science][Medline]
20 Müller, C.R., Fregin, A., Srsen, S., Srsnova, K., Halliger-Keller, B., Felbor, U., Seemanova, E. and Kress, W. (1999) Allelic heterogeneity of alkaptonuria in Central Europe. Eur. J. Hum. Genet., 7, 645–651.[Web of Science][Medline]
21 Waters, P.J., Parniak, M.A., Hewson, A.S. and Scriver, C.R. (1998) Alterations in protein aggregation and degradation due to mild and severe missense mutations (A104D, R157N) in the human phenylalanine hydroxylase gene (PAH). Human Mutat., 12, 344–354.[Web of Science][Medline]
22 Waters, P.J., Parniak, M.A., Akerman, B.R., Jones, A.O. and Scriver, C.R. (1999) Missense mutations in the phenylalanine hydroxylase gene (PAH) can cause accelerated proteolytic turnover of PAH enzyme: a mechanism underlying phenylketonuria. J. Inherit. Metab. Dis., 22, 208–212.[Web of Science][Medline]
23 Waters, P.J., Parniak, M.A., Nowacki, P. and Scriver, C.R. (1998) In vitro expression analysis of mutations in phenylalanine hydroxylase: linking genotype to phenotype and structure to function. Hum. Mutat., 11, 4–17.[Web of Science][Medline]
24 Clutterbuck, A.J. (1993) Aspergillus nidulans. In O’Brien, S.J. (ed.), Genetic Maps. Locus Maps of Complex Genomes. Cold Spring Harbor Laboratory Press, Cold Spring Harbor, NY, pp. 3.71–3.84.
25 Gehrig, A., Schmidt, S.R., Müller, C.R., Srsen, S., Srsnova, K. and Kress, W. (1997) Molecular defects in alkaptonuria. Cytogenet. Cell Genet., 76, 14–16.[Web of Science][Medline]
26 Felbor, U., Mutsch, Y., Grehn, F., Muller, C.R. and Kress, W. (1999) Ocular ochronosis in alkaptonuria patients carrying mutations in the homogentisate 1,2-dioxygenase gene. Br. J. Ophthalmol., 83, 680–683.
27 Higashino, K., Liu, W., Ohkawa, T., Yamamoto, T., Fukui, K., Ohno, M., Imanishi, H., Iwasaki, A., Amuro, Y. and Hada, T. (1998) A novel point mutation associated with alkaptonuria. Clin. Genet., 53, 228–229.[Web of Science][Medline]
28 Walter, K., Gaa, A. and Schaefer, H.E. (1999) Sequence analysis of the homogentisate 1,2-dioxygenase gene in a family affected by alkaptonuria. J. Med. Genet., 36, 645–646.
29 Porfirio, B., Chiarelli, I., Graziano, C., Mannoni, A., Morrone, A., Zammarchi, E., Beltrán-Valero de Bernabé, D. and Rodríguez de Cordoba, S. (2000) Alkaptonuria in Italy: polymorphic haplotype background, mutational profile and description of four novel mutations in the homogentisate 1,2-dioxygenase gene. J. Med. Genet., 37, 309–312.
CiteULike 
Connotea 
Del.icio.us What's this?
This article has been cited by other articles:
|
|
 |
|
  A. A. Fisher and M. W. Davis Alkaptonuric Ochronosis with Aortic Valve and Joint Replacements and Femoral Fracture: A Case Report and Literature Review Clin. Med. Res., November 1, 2004; 2(4): 209 - 215. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 J. M. Rodriguez, P. Ruiz-Sala, M. Ugarte, and M. A. Penalva Fungal Metabolic Model for Type I 3-Methylglutaconic Aciduria J. Biol. Chem., July 30, 2004; 279(31): 32385 - 32392. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 C. Phornphutkul, W. J. Introne, M. B. Perry, I. Bernardini, M. D. Murphey, D. L. Fitzpatrick, P. D. Anderson, M. Huizing, Y. Anikster, L. H. Gerber, et al. Natural History of Alkaptonuria N. Engl. J. Med., December 26, 2002; 347(26): 2111 - 2121. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||