Although many mutations in the ornithine transcarbamylase gene have been correlated with `late onset' of hyperammonemia in patients, the effects of these mutations on enzyme function are largely unknown. Three recurrent mutations (R40H, R277W and R277Q) found in patients with `late onset' disease were incorporated into `mature' human ornithine transcarbamylase cDNA and overexpressed in Escherichia coli. The three recombinant mutant enzymes were purified to homogeneity on an affinity column and their biochemical characteristics were compared to the wild type enzyme. The R277W and R277Q mutants display markedly reduced affinity for l-ornithine, loss of substrate inhibition, alkaline shift of pH optimum, and reduced thermal stability compared to the wild type enzyme. These differences, particularly the reduced affinity for l-ornithine, are sufficient to account for their biochemical effects. In contrast, the `mature' R40H mutant was biochemically indistinguishable from the wild type enzyme in vitro.
`Late onset' hyperammonemia often results from `leaky' mutations in the ornithine transcarbamylase enzyme (OTCase, EC 2.1.3.3) which catalyzes the formation of citrulline and inorganic phosphate from carbamyl phosphate and l-ornithine in the urea cycle (1 ). Clinical onset of the disease in patients with partial ornithine transcarbamylase deficiency (OTCD) can occur at any age after the first month of life (2 ,3 ). Patients with late onset OTCD typically show residual enzyme activity in liver tissue. In contrast, complete enzyme deficiency results in acute neonatal hyperammonemia within the first few days of life, and if the patient survives the neonatal period, a severe clinical course invariably follows.
Human OTCase is a homotrimeric protein with 322 amino acid residues in each polypeptide chain. The `pre-protein' has a 32 amino acid leader peptide at the N-terminus that is required for directing the protein into the mitochondria, where it is cleaved. We have previously used the Escherichia coli aspartate transcarbamylase (ATCase; PDB: 8AT1) structure (5 ) to develop a homology based model of human OTCase (6 ) that closely resembles the recently reported crystal structure of Pseudomonas aeruginosa OTCase (7 ). Each polypeptide chain has two domains, termed polar and equatorial. When missense mutations that result in clinical symptoms were mapped on the model, mutations causing neonatal hyperammonemia tend to affect `buried' residues, while most `late onset' mutations were on the surface of the protein (6 ).
R40H, R277W and R277Q OTCase are three recurrent mutations at CpG dinucleotide `hot spots' that result in a `late onset' phenotype. Arginine 40 and arginine 277 are conserved among many species, but their location within the structural model is remote from the active site (6 ). Arginine 40 is located in the polar domain at the start of a helix near the N-terminus of the mature enzyme and appears to interact with a conserved glutamate at position 52. Arginine 277 is located in an equatorial domain loop that is homologous to the 240's loop in the catalytic subunit of E.coli aspartate transcarbamylase, and appears to form an ionic interaction with an aspartate at position 196.
While most patients with `late onset' OTCD have varying levels of residual enzyme activity in liver homogenates, the mutant enzymes have not been purified and studied in detail. As many more `private mutations' are identified in the OTC gene, expression studies are necessary in order to distinguish deleterious effects from polymorphism. These studies may also help to clarify why some patients can be asymptomatic for years and then suffer an apparently sudden, and sometimes fatal, onset of hyperammonemia. A better understanding of how the mutant enzymes function may also help in the development of new and better therapies for OTCD and in the selection of better candidates for gene therapy. This work reports on the clinical phenotype and enzymatic function of three OTCase mutants.
R40H. Information is available on 13 patients. Age of presentation ranged from 3 to 56 years. Ammonia levels as high as 2400 [mu]M (normal <35 [mu]M) were reported in one patient (8 ). Residual liver OTCase activities for the three patient samples measured in our laboratory were 5%, 7.5% and 30% of normal. Nine patients died, two are alive with normal intelligence and the outcome for two patients is unknown. R277Q. The age of presentation was 10-15 months, and the ammonia level in one of our patients was 140 [mu]M. Residual OTCase activities in liver samples were 50-60% of normal in the two patients for whom it was measured and the Km of OTCase for l-ornithine in liver homogenates from these two patients was 10-fold higher than normal. Two patients died and one is alive and shows normal intelligence.R277W. Age of first clinical presentation ranged from the neonatal period (one patient) to 8 years. Nine of the 11 patients presented at or after one year of age and the mean age of presentation was 2.4 years. Plasma ammonia levels at presentation ranging from 129 to 406 [mu]M were reported for four patients. Residual OTCase activity in liver homogenates from nine patients ranged from 2% to 22% of normal and for those in whom the Km for l-ornithine was determined (three patients) it was found to be ~10-fold higher than normal. In the one patient in which it was examined, the pH optimum was shifted toward alkalinity. Of the 10 patients for whom information was available, six have normal intelligence after variable periods of follow up, two patients died and two have severe mental retardation.
Results of studies of purified wild type and mutant OTCases are summarized in Table 1 . The wild type enzyme and R40H show similar properties, whereas the two mutants at residue 277 show noticeably abnormal characteristics. Both the wild type enzyme and R40H show substrate inhibition when titrated with l-ornithine in buffers above pH 8 (Fig. 1 ), while R277Q and R277W do not.
As the clinical features illustrate, the severity of clinical disease associated with late onset OTCD is highly variable. Some patients are completely asymptomatic, whereas others can die from their disease. The variability in clinical outcome probably reflects many factors, including differences in diet, the effects of physical or mental stress and yet unknown genetic or environmental factors. Altered regulation of gene expression, polymorphisms of other enzymes involved in nitrogen metabolism, even differences in hepatocellular organization may all play a role. Perturbations in mitochondrial targeting or enzymatic cleavage of the leader peptide sequence during or after unfolding and translocation into the mitochondrial matrix may also affect activity in vivo. Although making accurate correlation of genotype and phenotype is challenging, examining the properties of mature OTCase in vitro shows some of these variables to be eliminated or verified.
Results suggest that wild type OTCase is optimized for its environment. The maximal velocity of the wild type enzyme is reached at 1 mM l-ornithine, a concentration thought to be within the mitochondrial physiological range (10 ). In addition, the optimal pH for the enzyme is similar to that of the mitochondrial matrix.
The finding that all of the properties of R40H examined in vitro (catalytic activity, substrate affinity, pH dependence of activity and stability) were very similar to the wild type enzyme was surprising, since assays of liver homogenates from patients with the R40H mutation who died showed only 1-30% of normal OTCase activity (11 -13 ). The specific activity of purified R40H is 75-90% of wild type at physiological concentrations of l-ornithine. R40H was purified several times from different stocks and plasmids from these stocks were resequenced to confirm the presence of the R40H mutation and the absence of any other mutations.
Both the human OTCase homology model (6 ) and the P.aeruginosa catabolic OTCase crystal structure (7 ) show arginine 40 interacting with glutamate 52. These residues are strongly conserved in all OTCases (6 ), suggesting they have an important role.
Replacing arginine 40 with histidine would not necessarily disrupt an ionic interaction with glutamate 52 and all of the experimental evidence indicates that the structure and stability of the enzyme are not perturbed significantly by the mutation. Both wild type OTCase and R40H have similar pH dependent velocity profiles suggesting that, even if the histidine titrates, it does not affect the function of the enzyme significantly. Increasing the ionic strength of the buffer by the addition of 100 mM KCl increases the activities of both R40H and the wild type enzyme by, approximately, a factor of two (data not shown). Thermal inactivation experiments provide further evidence against structural destabilization in R40H. Both the wild type enzyme and R40H have very similar thermal inactivation profiles, with a midpoint around 58oC. The thermal stability and the pH dependence data for R40H suggest that it would have a half-life in vivo similar to the wild type enzyme.
All of these results indicate that, in contrast to R277W and R277Q, the low activity of R40H in vivo is not intrinsic to the protein, but the result of other factors. The mutation may reduce rates of synthesis, or interfere with translocation to the mitochondrion, removal of the leader sequence or the interactions of the protein with other cellular components (10 ,14 ). Several factors may be involved in combination, since the phenotype of the R40H mutation is quite variable, and at least one adult with this mutation was asymptomatic (13 ). The amount of cross-reactive material was reported in two patients with the R40H mutation, as well as in transfection experiments with COS 1 cells. In all of these cases, the amount of cross-reactive material was markedly reduced, and the patient with less cross-reactive material showed more severe symptoms (8 ,13 ).
Thus, defects in mitochondrial targeting, translocation and post-translational processing are likely to contribute to the low in vivo activity of R40H. The signal sequence in human OTCase is 32 residues long, placing arginine 40 close to the N-terminus of the mature protein. Arginine 40 is in fact the first strongly conserved residue in the mature OTCase sequence (6 ). Since arginine 40 is well conserved in bacteria in which mitochondrial translocation does not occur, it is unlikely that it plays a role as part of an uncleaved mitochondrial signal sequence. In addition, the leader peptide appears sufficient to direct chimeric proteins to the mitochondrion (15 ). However, mutations near the N-terminus may interfere with recognition and cleavage of the leader peptide, resulting in an accumulation of the OTCase precursor protein. In rat OTCase which is >90% identical to human, the precursor form of OTCase shows greater instability and lower activity than the mature form (16 ) and similar behavior would be expected for the human protein.
These mutants have similar characteristics, despite the difference in the chemical properties of their amino acid side chains. Although binding of carbamyl phosphate is slightly weaker than for wild type, the much weaker affinity for l-ornithine appears to be the primary cause of their deleterious effects in vivo. At 1 mM l-ornithine and pH 8, conditions thought to be similar to the mitochondrial matrix, R277Q has 7% and R277W has 3% of the activity of the wild type enzyme. This reduced activity is sufficient to explain why mutations at position 277 result in hyperammonemia.
R277W has slightly lower substrate affinity and specific activity and a slightly higher susceptibility to thermal inactivation than R277Q. This is consistent with tryptophan being a more radical substitute than glutamine. The combination of these effects may explain why patients with R277W have more severe symptoms than those with R277Q.
Arginine 277 is located in a flexible loop in E.coli aspartate transcarbamylase (ATCase) catalytic subunit, a protein with which OTCase shares a similar fold (7 ). Motion of this loop in ATCase plays a key role in catalysis (17 ), and may have a similar function in OTCase. Arginine 277 also appears to contribute to the stability of the enzyme, since both R277Q and R277W are inactivated at lower temperatures than wild type.
The altered pH optimum for activity of R277Q and R277W suggest that these mutations have a global effect on the ionizable groups that are critical to function. If a conformational change is occurring in the arginine 277 mutants due to a loss of an ionic interaction, aspartate 196 would be a likely candidate, since the side chains of arginine 277 and aspartate 196 form an ionic bond in the Pseudomonas catabolic OTCase crystal structure (7 ). Substitution of aspartate 196 by valine causes neonatal OTCD with almost complete loss of enzymatic activity (18 ). Abolishment of this interaction would be sufficient to produce the reduction in inactivation temperature relative to wild type found in R277Q and R277W. The interaction of arginine 277 with aspartate 196 may also be responsible for the altered pH optimum for activity of R277Q and R277W. Although the side chain of aspartate normally ionizes with a pK of 4, pK's as high as 9.9 have been reported for aspartate residues buried in the interior of the protein (19 ). The pK of aspartate 196 would be expected to be higher in the R277Q and R277W mutants than in the wild type enzyme, as a result of the elimination of its interaction with arginine 277. If the residues involved in the catalytic mechanism require that aspartate 196 be ionized, this could account for the increase in the pH optimum for activity in the two 277 mutants.
This study has identified two classes of mutants, both of which produce `late onset' hyperammonemia, but which have very different biochemical bases. OTCD produced by R277Q and R277W almost certainly results from the effects of the mutations on l-ornithine affinity and protein stability. OTCD produced by R40H appears to be the result of factors extrinsic to the protein, such as disruption of mitochondrial targeting and translocation or post-translational processing. The ability to identify mutations that reduce activity in vivo by reducing the intrinsic catalytic activity of the enzyme may have implications for gene therapy. Reports in which the presence of even a single defective chain in a trimer of OTCase can abolish activity (20 ,21 ) suggest that some of these missense mutations may interfere with attempts to use wild type OTCase in gene therapy. In other cases, hybridization of two mutant OTCases, each with minimal activity partially restores activity (22 ). A better understanding of structure- function relationships, and particularly the role of subunit interactions in the function of this important enzyme, will be required to develop effective strategies for gene therapy.
Clinical information for 13 patients with the R40H mutation, 11 with the R277W mutation and three with the R277Q mutation was either available to our laboratory from referral of samples for mutation identification (seven patients) or was obtained from the medical literature (20 patients) (9 ,11 -13 ,18 ,23 ,24 ). Of the patients seen in our laboratory , three were positive for R277W, one for R277Q and three for R40H. Two of the patients with the R40H mutation were siblings; an additional two pairs of patients had a familial relationship. All other patients were from different families.
A detailed description of the construction of this plasmid has been published elsewhere (25 ). The nucleotide sequence encoding `mature' human OTCase was cloned into a pET21a+ based expression vector with a methionine initiation codon added at position 1. Transcription in a T7 RNA polymerase expressing host is driven by a strong T7 promoter that is under the regulation of the lactose (lac) operator. In this expression system, ~25% of the soluble protein is human OTCase. Co-transformation with a plasmid (pACYC184) containing bacterial chaperonin genes (GroES and GroEL) markedly reduced the precipitation of recombinant protein in inclusion bodies, and enhanced the yield of active enzyme to ~90% of the soluble protein (25 ).
Expression vectors for R40H (CGT to CAT), R277W (CGG to TGG) and R277Q (CGG to CAG) were constructed by PCR site-directed mutagenesis (26 ). Overlapping sense and anti-sense 19mer oligonucleotides containing the mutation at the 10th position were synthesized. These primers and two additional flanking primers were used to amplify by PCR two fragments of the OTC gene containing unique restriction sites. The resulting two fragments were combined in one PCR reaction amplifying a common DNA fragment containing the mutation and the restriction sites at the 5' and 3' ends, respectively. This fragment `cassette' was cloned into the expression vector. The cloned fragment was sequenced to confirm the presence of the mutation and to exclude PCR errors.
Conditions used for culturing bacteria, induction of OTCase, expression and purification of the protein were as described before (25 ).
Enzyme activity was measured by a colorimetric assay which detects the formation of l-citrulline (27 ). Unless otherwise noted, the buffer was 50 mM Tris acetate, 2 mM EDTA, pH 8.3. Standard curves for l-citrulline were generated for each experiment since this assay is sensitive to ambient conditions (27 ,28 ). Product formation was linear with time even at the lowest carbamyl phosphate concentrations used. One unit is the amount of enzyme required to catalyze the formation of 1 [mu]mol of l-citrulline in 1 min at 25oC.
A borate buffer was used to minimize changes in pH with temperature. Samples of enzyme were diluted to 1.25 [mu]g/ml in 50 mM borate, 50 mM KCl, 2 mM EDTA, pH 8.3, incubated at temperatures between 25 and 70oC for 10 min, then quickly chilled on ice and assayed for activity within 15 min.
A tripartite buffer containing 51 mM 2-(N-morpholino)ethanesulfonic acid, 1 mM N-ethylmorpholine, and 51 mM diethanolamine, designed to maintain constant ionic strength between pH 5 and 11 (29 ), was used to determine the pH dependence of Vmax and Km for l-ornithine. Full titration curves in which l-ornithine was varied were collected between pH 7 and 9.5, and were fit using the software package NONLIN (30 ) to either the Hill equation or to an equation describing cooperative substrate binding with partial uncompetitive substrate inhibition (28 ).
Supported by Public Heath Service grants DK-47870 (to MT) and DK-17335 (to NMA) from the National Institute of Diabetes, Digestive and Kidney Diseases.
ATCase, aspartate transcarbamylase; OTCase, ornithine transcarbamylase (enzyme); OTC, ornithine transcarbamylase (gene); OTCD, ornithine transcarbamylase deficiency.
Human Molecular Genetics
Pages
Introduction
Results
Clinical features (see Materials and Methods for sources and references to clinical data)
Studies of purified enzyme
Discussion
R40H
R277Q and R277W
Materials And Methods
Clinical information
OTCase expression in E.coli
Construction of R40H, R277W, R277Q mutant OTCase expressing plasmids
Purification of `wild type' and mutant OTCases
Activity assay
Thermal inactivation
pH dependence
Acknowledgements
Abbreviations
References
REFERENCES
This page is maintained by OUP admin. Last updated Mon May 12 18:10:04 BST 1997. Part of the OUP Journals World Wide Web service. Copyright Oxford University Press, 1996