Aims Acute intermittent porphyria (AIP) is a disorder of the haem biosynthetic pathway caused by mutations in the hydroxymethylbilane synthase (HMBS) gene. Knowledge of the spectrum of mutations present in South Africa is limited. This study presents the molecular profile of 20 South African patients with AIP, and the kinetic analysis of one novel expressed mutated HMBS enzyme and a previously identified mutation at the same position.
Methods Genomic DNA was isolated from affected probands and selected family members, the HMBS gene amplified and mutations characterised by direct sequencing and restriction enzyme analysis. One of the novel mutations (p.Lys98Glu), a previously characterised mutation at the same position (p.Lys98Arg), and the wild-type enzyme were expressed, purified and subjected to partial kinetic characterisation.
Results Four new mutations, p.Lys98Glu, p.Asp230Aspfs*20, c.161-1G>A and c.422+3_6delAAGT, are described. Seven previously described mutations were found, while four patients revealed no mutations. Mutation analysis of five offspring of one of the probands carrying the p.Trp283X mutation revealed two asymptomatic carriers. Kinetic analysis showed that the p.Lys98Glu mutation results in loss of substrate affinity, whereas the previously described p.Lys98Arg mutation causes the loss of binding between the enzyme and its dipyrromethane cofactor, rendering the enzyme inactive.
Conclusions This study comprises the most comprehensive characterisation of HMBS gene mutations in patients with AIP in South Africa. The biochemical characterisation of expressed HMBS mutants reveals insight into the mechanism of catalytic activity loss, which may inspire investigation into individualised therapy based on the molecular lesion identified.
- MOLECULAR BIOLOGY
- CHEMICAL PATHOLOGY
Statistics from Altmetric.com
If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.
Acute intermittent porphyria (AIP) (MIM no. 176000) is an autosomal dominant disorder of haem biosynthesis caused by a partial deficiency of hydroxymethylbilane synthase (HMBS), also known as porphobilinogen (PBG) deaminase (EC 22.214.171.124). This enzyme catalyses the head-to-tail condensation of four molecules of PBG to form hydroxymethylbilane. This is facilitated by the dipyrromethane (DPM) cofactor, which is covalently linked to the enzyme in its active site cleft.1 In AIP, porphyrins and their precursors PBG and 5-aminolevulinic acid may be produced in excess and excreted in urine during and for many years after an acute attack.2 Biochemical diagnosis of AIP is based on the measurement of these urinary precursors, the presence of a plasma porphyrin fluorescence peak at 619 nm when excited at ±400 nm and the determination of erythrocyte HMBS activity. Molecular analysis of the HMBS gene is not required to make a diagnosis of AIP, but is helpful in confirming a biochemical diagnosis and is essential for presymptomatic diagnosis. A large proportion of patients remain clinically asymptomatic, but are at risk of an acute attack. There is, thus, a need for genetic screening of family members at risk.
Clinically, AIP manifests as intermittent attacks of neurovisceral dysfunction, which can be precipitated by various factors such as drugs, hormones and alcohol.1 ,3 Indeed, 80%–90% of patients who inherit a defective AIP gene remain clinically latent for life,4 reflecting the low penetrance of this disease. A recent study revealed an incidence of symptomatic AIP of 0.13 per year per million population in Europe.5
The human HMBS gene consists of 15 exons spanning over 10 kb of genomic DNA, and is located at chromosome 11q23.3. Located in the 5′ flanking region and in intron 1 are two distinct promoters, which generate the housekeeping (containing exons 1 and 3–15) and erythroid-specific (containing exons 2–15) transcripts by alternative splicing of exons 1 and 2.6 Over 400 mutations in the HMBS gene have been reported.7 To date, 10 patients in South Africa have had a diagnosis of AIP confirmed at the gene level,8–10 of whom 4 carry the p.Arg116Trp mutation, common in the Netherlands due to a founder effect.11 The ancestry of three of the patients carrying the p.Arg116Trp mutation suggests a mixed genetic make-up of Malay, European and Khoi origins. The European component may well be Dutch, thereby explaining the presence of this mutation in these apparently unrelated patients.
This study expands the cohort of South African patients with a molecular diagnosis of AIP, revealing four novel mutations and opens the way for future family studies. Furthermore, one of the new mutations (p.Lys98Glu) and a mutation previously described in a Finnish family (p.Lys98Arg) at the same position12 were expressed and purified to investigate the effect of these mutations on kinetic behaviour of the encoded enzyme. It is possible that the elucidation of mechanisms of enzyme dysfunction associated with specific molecular mutations may reveal appreciation of the disease process.
Materials and methods
Selection of study cohort
Thirty-nine patients with a clinical and biochemical diagnosis of AIP were designated as potential participants in this study, and 20 patients were ultimately recruited. The remainder of the patients were not traceable. Five of the offspring of one of the recruited patients were also investigated. HMBS enzyme activities are not routinely measured as part of the investigation of AIP and, as a consequence, enzyme activities were not performed, except in one case. Ethical approval was obtained from the Health Sciences Faculty Research Ethics Committee (HREC/REF: 324/2010), University of Cape Town, South Africa. No members of this cohort were, to our knowledge, related to South African patients previously investigated for HMBS mutations.8–10
PCR amplification of all 15 exons and flanking intronic sequences was performed using primers designed with Primer Designer V.2.0 (primer details available on request). PCR products spanning all exons of the human HMBS gene were directly sequenced by Sanger sequencing using Big Dye Terminator chemistry to determine the precise sequence variations. Where possible, sequencing was performed in both forward and reverse directions. Mutation nomenclature was according to den Dunnen and Antonarakis.13
Restriction enzyme digestion
Sequence analysis revealed four new mutations. The absence of these mutations was confirmed in at least 50 race-matched control DNA specimens by restriction enzyme digest analysis (c.292A>G using StyI; c.689_690delAC using BstY1; c.422+3_6delAAGT using XcmI; c.161-1G>A using DdeI). Genomic DNA of all five children of one of the probands was also interrogated with restriction analysis (c.848G>A using XbaI). Restriction enzyme maps for wild-type and polymorphic DNA fragments were constructed using Webcutter2.0.
Red cell HMBS assay
HMBS enzyme activity was determined by fluorimetry as previously described.14 The assay is based on a coupled-enzyme procedure in which added 5-aminolevulinic acid and its dehydratase present in erythrocytes are used to generate PBG as substrate for the HMBS. Coproporphyrin was used as a standard. HMBS enzyme activity was expressed as nmol/hour/mL red blood cells.
Selected HMBS mutants were engineered using QuikChange Site-Directed Mutagenesis kit. The mutated plasmid DNA was transformed into supercompetent JM109 Escherichia coli cells (Promega, Madison, Wisconsin).
Screening clones and confirmation of mutated sequences
Plasmid DNA from overnight cultures of the engineered mutants was extracted using the Wizard Plus SV Miniprep DNA Purification system (Promega). PCR products were purified by the Illustra GFX PCR DNA and Gel Band Purification Kit (AEC-Amersham Biosciences). The entire HMBS gene was screened, confirming the presence of desired and the absence of unwanted mutations. The mutated sequences were aligned to the known wild-type fragment(s) using GenBank data.
Expression and purification of wild-type and mutant HMBS protein
The HMBS-containing pTrcHis-A vector plasmid was a gift from Professor H. Dailey (University of Georgia, Athens, Georgia). Expression and purification of human wild-type and mutant HMBS was as described for human protoporphyrinogen oxidase15 with the following modifications: incubation was at 37°C for 22 hours for mutant expression, and Tween20 (Sigma-Aldrich, South Africa) replaced n-octylglucopyranoside during purification. Purity was confirmed by 7.5%–17.5% gradient sodium dodecyl sulfate polyacrylamide gel electrophoresis. Protein quantification was determined using the Bio-Rad Protein microassay (Bio-Rad Laboratories, Germany).
HMBS assay for enzyme kinetics
The activity of purified recombinant HMBS was determined16 with the following modifications: 50 µL of appropriately diluted enzyme was added to decreasing concentrations of PBG (90 µM starting concentration) in assay buffer (0.1 M Tris-HCl, pH 8.0, containing 55 µM dithiothreitol) to a final volume of 1 mL. Bovine serum albumin of equivalent concentration to HMBS was used as a non-enzymatic control. Uroporphyrin I (Frontier Scientific, Logan, Utah) was used as a standard. HMBS enzyme activity was expressed as pmol/hour/ng protein. Where recombinant HMBS activity was minimal, dilution of the sample decreased until meaningful results were obtained (wild-type HMBS typically diluted 1:2500 and the K98R mutant 1:5).
Kinetic characterisation of wild-type and mutant HMBS
KM and Vmax were determined by an iterative curve-fitting procedure using GraphPad Prism 6.01 (GraphPad Software, USA) and a least squares fit of the Michaelis-Menten equation to experimental kinetic data with a maximum of 1000 iterations. Increasing concentrations of PBG were used to a level of substrate excess. Thereafter, the enzyme turnover rate kcat and efficiency of the enzyme expressed as kcat/KM were determined. The data represent the mean from three independent experiments performed on different days.
Sequence analysis of the HMBS gene in 20 index patients with biochemical or clinical evidence of AIP revealed 11 mutations, 4 of which have not been previously described (table 1). The p.Arg195Cys and p.Gly236Ser mutations were found in Black patients, but have been previously described in a cohort of Finnish12 and French19 patients with AIP, respectively. The p.Arg149X mutation in exon 9, which creates a premature stop codon, was identified in a Black patient whose clinical phenotype has previously been reported.20 This mutation has also been found in Finnish families.12 Similarly, the p.Arg116Trp mutation in exon 8 identified in two patients of mixed race was shown in the same study but has also been commonly found in the Dutch AIP population.11 Two Caucasian siblings were identified with the known p.Trp283X mutation21 in exon 14, and restriction analysis of DNA obtained from the offspring of the male sibling showed that two of five children had inherited this mutation, but neither child has developed symptoms consistent with AIP. Two Black patients and a patient of mixed ancestry revealed the p.Leu258Leufs*33 mutation, and a Black patient the p.Asp61Thrfs*37 mutation. Both mutations have previously been described in patients of unknown ethnicity,8 ,22 the former in South Africa.
Four patients with a biochemical diagnosis of AIP were found not to harbour HMBS mutations in any exons or flanking regions interrogated. Furthermore, the South African R59W protoporphyrinogen oxidase mutation was also not detected, making a diagnosis of variegate porphyria unlikely, particularly with the reported absence of skin lesions and, in patients 14 and 15, an absence of elevated stool porphyrins (results not shown). Red blood cell HMBS enzyme analysis of patient 15 revealed activity of 13.8 nmol/hour/mL red blood cells, 49% of normal control activity (average activity of 3 controls: 28.0 nmol/hour/mL; HMBS enzyme activity range: 24.0–54.0 nmol/hour/mL14.) This is consistent with the expected 50% decreased in enzyme activity in heterozygous patients with AIP.
The four novel mutations (table 1) included a missense mutation (p.Lys98Glu), two splice site mutations (c.422+3_6delAAGT and c.161-1G>A) and a frameshift mutation (p.Asp230Aspfs*20). The p.Lys98Glu mutation resulted from an A-to-G transition in exon 7 predicting the substitution of positively charged lysine for a negatively charged glutamic acid. The c.422+3_6delAAGT mutation in intron 8 and c.161-1G>A mutation in intron 4 generate a putative donor and acceptor splice site mutation, respectively. In the p.Asp230Aspfs*20 mutation, a deletion in exon 12 of the second and third bases of codon GAC (aspartic acid) generates a new codon GAT, also representing aspartic acid. Further, a frameshift occurs which predicts substitution of 19 amino acids in codons 230–248 followed by chain termination.
The novel p.Lys98Glu and previously described p.Lys98Arg HMBS mutations were expressed in JM109 supercompetent cells. Screening of the entire HMBS gene did not reveal the introduction of additional mutations and direct sequencing showed successful engineering of both mutants (results not shown). The p.Lys98Glu and p.Lys98Arg mutants showed a relative activity of 53.5% and 0.1%, respectively (table 2).
Relative to the wild-type enzyme, the p.Lys98Glu mutant revealed an approximately fourfold increase in KM and almost halved Vmax, while the p.Lys98Arg mutant revealed a slightly lower KM but a negligible Vmax (table 3). The catalytic efficiency of the mutants, reflected by the kcat/KM, was reduced, profoundly so for the p.Lys98Arg mutant.
In keeping with the presence of the DPM cofactor, the p.Lys98Glu mutant HMBS showed the characteristic spectral shift from 564 to 495 nm, thus confirming its ability to bind the cofactor.18 In contrast, the absence of a spectral shift by the p.Lys98Arg mutant indicated a loss of cofactor-binding ability (figure 1).
This report reflects the most current interrogation of HMBS mutations in patients with AIP in South Africa. Regrettably this cohort is incomplete, the task being complicated by difficulties with tracing and following up patients in a fragmented healthcare system. Nevertheless, the molecular investigation of patients with AIP is an important step in the management of families, where screening is essential to identify those with latent disease so as to prevent acute attacks.3 ,23 The sensitivity for mutation detection determined by sequencing is 95%,24 and increases to 98.1% if gene dosage analysis is included.25 However, mutations were not detected after sequencing in 4 out of 20 index cases in the small cohort of patients in this study, but this still achieves greater diagnostic accuracy than erythrocyte HMBS activity analysis.3
This study identified a total of 11 different mutations in the 20 index cases screened. It is well known that most AIP mutations are private and limited to the families in which they arise. This is supported by the fact that, to date, more than 400 mutations have been identified in the HMBS gene in AIP. Seven of the mutations found have been previously described. Patient 1 in our cohort was HIV positive, presented with tetraparesis26 and revealed a single base-pair deletion resulting in the previously described p.Asp61Thrfs*37 mutation.22 The p.Arg149X mutation creates a premature stop codon and was first identified in two Finnish families,12 but in the current study it was identified in a Black male. Indeed, this mutation has also been detected in Spanish probands27 and occurs at CpG dinucleotides, known mutation hotspots. A previous study of South African patients with AIP described the p.Leu258Leufs*33 mutation8 in a patient of unknown race and has now been found in three apparently unrelated mixed ancestry and Black patients. If the patient in the original study was Black or of mixed ancestry, it is tempting to speculate the presence of a founder effect, but further investigation of South African patients with AIP is required to confirm this.
The p.Arg116Trp mutation was found in a patient of mixed ancestry, and this mutation was revealed in the same study of three South African mixed ancestry patients.8 A relatively high frequency of this mutation (19/80 families) has been found in the Dutch AIP population, and a distinctive haplotype was found to segregate with the p.Arg116Trp mutation in this population.11 This suggests a founder effect, and as the ancestry of all South African patients with this mutation is mixed, it is likely that the mutation is of Dutch origin, considering the Dutch colonial history in South Africa. A further founder mutation is p.Trp283X, found in 56% of patients with AIP in Switzerland and has the highest prevalence of any single HMBS gene mutation in an AIP population.28 The two siblings found to harbour this mutation are Caucasian, but are not known to have Swiss ancestry, and haplotype studies would be required to establish if their mutation is linked to the founder mutation. The utility of a molecular diagnosis of asymptomatic family members is highlighted by the revelation that two of five offspring of one of the siblings were also found to carry the mutation. This allows for appropriate counselling of the parents and children, both those carrying the mutation and those found to be mutation free.
Four patients with clinical and biochemical evidence of AIP were shown not to have coding or flanking region mutations. Three of these patients demonstrated a clinical history consistent with an acute attack, but none had skin lesions. Two of these patients (14 and 15), where stool porphyrins were measured, revealed results incompatible with variegate porphyria. Taken together with the absence of the R59W protoporphyrinogen oxidase mutation in all four patients, a diagnosis of AIP is more likely than variegate porphyria. One of these patients (patient 14) presented peripubertally, at the age of 11 years, but demised 3 years later. Little is known about the prevalence of AIP in the paediatric population, but a prospective study involving 61 children in northern Sweden found that 10% of participants from 5 to 13 years of age presented evidence of an acute attack.29 A recent study from Murcia, Spain, has identified nine children aged 0.5–16 years presenting with AIP.30 Both regions demonstrate a high prevalence of AIP due to founder mutations.31 ,32 Another patient where no mutation could be found demonstrated an approximately 50% decrease in erythrocyte HMBS activity, consistent with AIP. In such patients, analysis of the promoter region is required to exclude mutations.33 Quantitative PCR for gene dosage analysis is also useful to exclude large deletions and insertions or functional intronic mutations caused by nonsense-mediated decay.3 A recent report further suggests mRNA analysis to exclude pseudoexon inclusion as a cause of AIP in mutation negative patients.34 Further investigation is required to fully characterise the molecular defect in our patients where no exonic mutation was found.
This study revealed four new mutations. Lysine at position 98 has previously been demonstrated to mutate to arginine,12 but we show a change to glutamic acid (p.Lys98Glu) in two Caucasian patients. This invariant lysine moiety is located within the active site cleft and in close proximity to the DPM cofactor, and forms salt bridges with the acetate cofactor side chains.35 Kinetic studies using wild-type HMBS reveal a KM (8.49 μM) similar to that previously described (8.7 μM).36 The newly described p.Lys98Glu mutant demonstrates decreased substrate affinity, as the KM is approximately fourfold greater than that of the wild type. The Vmax is 62% of the wild-type enzyme producing a sixfold decreased catalytic efficiency, expressed as the ratio kcat/KM. This mutant replaces a positively charged amino acid with the negative one, which is insufficient to prevent cofactor binding to apoenzyme, but likely disrupting interaction with the DPM cofactor and decreasing substrate affinity and catalytic rate. The decreased substrate affinity is not enough to totally negate enzyme activity, as reflected by a specific enzyme activity half of the wild-type enzyme, but is likely severe enough to be disease causing, as reflected by the significantly decreased catalytic efficiency.
The KM of the p.Lys98Arg mutant was similar to that of the wild type, indicating that the affinity of the enzyme for the PBG substrate is unaffected and yet, intriguingly, it maintains <1% of residual enzyme activity. During purification of this mutant, it was noted that the eluted protein lacked the characteristic pink colour (most likely DPM),37 which led us to speculate that there could be loss of the DPM cofactor, resulting in apoenzyme expression. Spectral shift analysis showed that the mutant had indeed lost its binding to the cofactor, while p.Lys98Glu maintained cofactor binding. The cofactor acts as an initiation point for the sequential head-to-tail addition of PBG to ultimately form the product hydroxymethylbilane. The absence of the cofactor negates this process and accounts for the almost complete loss of enzyme activity of the p.Lys98Arg mutant. Unexpectedly, despite the loss of cofactor, this mutant is able to maintain some binding affinity of substrate. The bulky arginine moiety in the mutant may disrupt the docking or covalent binding of the cofactor to the enzyme, while a charge switch in the p.Lys98Glu mutant has no such effect. Several other mutants (Arg149Gln, Arg173Gln38 and Arg116Trp)39 are also reported to result in the expression of apoenzyme. It is suggested that the disrupted interaction of these residues with the cofactor likely prevents successful docking of hydroxymethylbilane, the DPM precursor, in the active site of the apoenzyme.
The p.Asp230Aspfs*20 mutation in exon 12 causes a frameshift, resulting in a truncated protein most likely with compromised enzyme activity. Two novel intronic lesions were detected, and are hypothesised to adversely influence gene splicing. A mutation in the same position as c.161-1G>A in intron 4 has been described, except that a G-to-C transition occurs,40 while the c.422+3_6delAAGT mutation likely creates a splicing aberration in intron 8. The impact of these lesions should be confirmed with reverse transcriptase PCR, which generates copy DNA from transcribed mRNA and allows the direct detection of spicing anomalies.
There is a dearth of information on AIP in Black and mixed ancestry patients and, in particular, the scope of HMBS mutations is almost unknown in the African context. A study of seven West and North African and Afro-Caribbean patients with AIP characterised six different mutations, four of which were novel.41 In this study, three of the four novel mutations were found in Black or mixed ancestry patients. Although studies are limited, it appears that AIP is not uncommon in these patients, and the disease-causing mutations are heterogeneous except for, perhaps, p.Arg116Trp (c.346C>T) which, in mixed ancestry patients, may be of Dutch origin. From the data in this investigation, it is not possible to make inferences about the prevalence of HMBS mutations in South Africa and would require a comprehensive study. Further studies are also required to delineate the scope of mutations in other regions of Africa, and the continuing emergence of novel mutations and the heterogeneity of mutations demonstrates the usefulness of prior characterisation of HMBS mutations in newly diagnosed patients to allow detection of relatives with latent AIP.
Take home messages
The overall prevalence of acute intermittent porphyria in Black South African patients is unknown, but current evidence suggests that the disease is not uncommon.
Several novel mutations have been uncovered in the South African population, in Black, mixed ancestry and Caucasian groups.
The biochemical characterisation of expressed HMBS mutants contributes to the mechanism of loss of catalytic activity.
The authors acknowledge the care provided to two patients with AIP and the permission granted to use their mutation data in this manuscript (Dr Veronica Ueckermann, Dr Lize-Marie Wium and Dr Darren Joseph, Department Internal Medicine, University of Pretoria). Sequencing was performed by the DNA Sequencing Unit, Central Analytical Facilities, Stellenbosch University, South Africa.
Handling editor Tahir Pillay
Contributors PM: conceived and oversaw the project. PF and AC: collection and interpretation of patient mutation data. EP and AC: enzyme expression; collection and interpretation of enzyme kinetic data. MS and CWS: clinical management of AIP patients at Groote Schuur Hospital. The first draft of the manuscript was written by PF and critically reviewed by all authors.
Funding Project funding was provided by the Medical Research Council (self-initiated research grant awarded to PM), the National Research Foundation Incentive Award for rated researchers (PM) and the University of Cape Town Research Committee.
Competing interests None.
Ethics approval Health Sciences Faculty Research Ethics Committee (HREC/REF: 324/2010), University of Cape Town, South Africa.
Provenance and peer review Not commissioned; externally peer reviewed.