Wilson disease (WD) is an autosomal recessive defect of copper transport characterized by massive accumulation of copper in the liver, which can lead to liver failure. Mutations in a copper transporting ATPase (WND or ATP7B) have been shown to cause the disease.The toxic milk mouse mutant (tx) accumulates copper in the liver in a manner similar to that observed in patients with WD. However, some physiological differences between tx mice and human WD patients have cast doubts on whether this mutant mouse is a valid model for WD. In this paper we report the isolation of cDNA clones encoding the murine homologue of WND. The predicted amino acid sequence is 1462 amino acids and contains the same functional domains identified in human and rat WND. As in the rat, the fourth metal binding domain is apparently non-functional. Similar levels of a 7.5 kb WND mRNA were detected in liver and kidney from normal and tx mice, indicating that transcription of this gene was unaffected in the mutant mice. The coding sequence of WND cDNA from the tx mouse liver identified a single nucleotide difference between the normal DL mouse and the tx which is predicted to change methionine1356 in the eighth transmembrane domain to valine. This methionine is conserved in all copper ATPases including those from bacteria and yeast. The conclusion that this is the causative mutation is supported by the recent mapping of tx and WND to the same region of mouse chromosome 8. Thus the tx mouse is presented as a valid model for studies of the role of WND in copper transport and for investigation of different treatment strategies for WD.
Copper is an essential element required for a number of important enzymes. Due to the capacity of copper to cause oxidative damage to cells, efficient regulatory mechanisms exist to maintain tissue levels within the critical range. Copper homeostasis in mammals depends on a balance between intestinal absorption and biliary excretion [reviewed in (1 )]. In Wilson disease (WD), an autosomal recessive inherited disorder, an impaired ability to excrete copper into the bile leads to an accumulation in the liver and extensive damage to this organ. WD can present either as a liver disease or a neurological disorder, the latter condition thought to be due to aberrant release of copper from the liver depositing in the central nervous system. As well as the phenotypic variation the age of onset of the disease can be from childhood to adult, but the reason for the variable clinical picture is not well understood.
The defective gene in WD has recently been cloned and encodes a copper transporting P-type ATPase (2 ,3 ). The WD gene (officially designated ATP7B, referred to here as WND for clarity) is expressed primarily in the liver, where it is thought to facilitate removal of excess copper from hepatocytes to the bile for excretion. The mRNA is found in other tissues such as kidney, brain and placenta, but the role of the protein in copper transport in these tissues is not understood. The predicted protein closely resembles the protein product of the Menkes disease gene (officially designated ATP7A, referred to here as MNK) (4 ). The products of MNK and WND, which will be referred to as MNK and WND, respectively, are transmembrane proteins and have the characteristic features of P-type ATPases, such as an ATP binding domain, and a conserved aspartic acid residue, which is phosphorylated during the transduction of copper across the membrane (5 ). An interesting feature of MNK and WND is the presence of six putative metal binding sites in the N-terminal half of the protein; copper ATPases from bacteria (6 ) and yeast (7 ) have only one or two of these putative copper binding sites. Analysis of patients with WD has identified 23 mutations of WND including 11 small insertions and deletions, seven missense mutations, two nonsense mutations and three splice site mutations. The missense mutations have been localized to mainly the transmembrane domain or the ATP binding region (8 ). The correlation of mutation with phenotype is complicated by the fact that most patients are compound heterozygotes.
Toxic milk (tx) is an autosomal recessive mutation in mice which causes hepatic accumulation of copper which commences in the third postnatal week. By 6 months of age the copper concentration can be 100 fold that of the normal adult. This gradual accumulation of copper in the liver resembles that seen in patients with WD. In addition, however, pups are born copper deficient and the milk produced by mutant mothers is low in copper, resulting in death of pups (9 ). Analysis of the morphology of livers from adult tx mice has shown significant differences from the liver damage seen in patients with WD. This fact, together with the absence of reports of copper deficiency in infants of mothers with WD, has raised doubts about whether the tx mouse is a valid model for WD (10 ). Recent studies, however, have mapped the toxic milk mutation to the same region of chromosome 8 as the murine homologue of WND, consistent with a mutation in this gene causing the tx phenotype (11 ,12 ).
Here we report the cloning and sequence of the murine homologue of the WD gene (GenBank submission number U38477). We demonstrate a point mutation in the tx mouse sequence which results in the conversion of a highly conserved methionine to valine in the eighth transmembrane channel, strongly suggesting that the tx mouse is indeed a valid model for WD.
A mouse liver cDNA library from a BALB/c mouse (Clonetech #ML3001b) was initially screened using a rat WND cDNA clone which included nucleotides encoding the C-terminal region of the protein (13 ). This screen identified three positive clones which were plaque purified, subcloned and sequenced as described in the Methods. The sequences of these clones showed a high level of identity to the corresponding rat sequence. The library was screened again with these clones and the procedure repeated until sequence covering the entire coding region was obtained. Translation of the sequence identified an open reading frame with two potential ATG start codons, one which corresponded to the start codon identified in rat and human WND (13 ,14 ), and the other 36 nucleotides upstream which also conformed to the Kozak consensus for initiation codons (15 ). Including this region, the open reading frame for mouse WND mRNA consists of 4386 bases, predicted to encode a protein of 1462 amino acids. Figure 1 shows the alignment of the 5'-regions of the WND cDNAs from mouse, human and rat. The upstream methionine is not present in human (the codon in this position is CCA), but sequence is not available for this region in the rat; it remains to be determined whether the mouse WND has an additional 18 amino acids at the N-terminus.
The sequence of the coding region of the WND homologue in the tx mouse has shown a single base difference between the normal (DL) and tx mutant mouse. We conclude that this is the causative mutation in the tx mouse, making it a true animal model of WD. This conclusion is supported by the following considerations. The tx mutation arose in an inbred strain (DL) and affected mice were subsequently back crossed to produce homozygous tx mutants (9 ). Therefore, any difference in sequence between the two is likely to be the result of a mutation rather than a polymorphism. The mutation results in the alteration of a highly conserved methionine to valine in the eighth transmembrane region of the predicted protein. The conservation of this methionine in all copper ATPases so far sequenced, including those from yeast and bacteria, suggests that it plays an important role in the copper transport process. Strong supporting evidence is also provided by the coincident mapping of both tx (11 ) and WND (12 ) to a site near the centromere of chromosome 8 in mouse. Although collectively these data provide compelling evidence that the methionine to valine change causes the defect in the tx mouse, there is still a remote possibility that the change is a polymorphism. Future work will directly test this possibility by determining whether a normal WND construct will correct the tx defect.
The mutation found in the tx mouse has not been reported in any patients with WD. However, a similar methionine to valine change in transmembrane domain number four (TM4) has been found in one patient with WD (8 ). This change was initially classified as a possible polymorphism. This methionine, however, is also conserved in all copper transporting ATPases and it now appears likely, based on the tx mutation, that the TM4 mutation is the causative mutation in that WD family. In copper transporting proteins the metal ion is often bound to cysteine or methionine residues, thus the methionine in the transmembrane regions may form a transient complex with the copper ion as it passes through the membrane channel.
An interesting feature of the predicted structure of the mouse WND is the presumed loss of the putative copper binding site four in the N-terminal region. The rat and mouse sequences share some of these changes, suggesting that the alterations occurred in a common ancestor. The region of divergence with the human sequence includes a region of about 190 amino acids surrounding the M4 site, so the loss of function of the recognized motif may have been followed by subsequent drift of surrounding sequences. The loss of this site without inactivation of the copper transport function implies that not all copper binding sites are required for function, and raises the question of the role of the six metal binding sites which are conserved between WND and MNK in most species so far examined.
Figure
Animal models of WD are important for both clinical and scientific studies. Clinically the treatment of WD has been generally regarded as very successful with the advent of penicillamine therapy, yet significant complications sometimes develop (18 ). A convenient animal model allows a more complete assessment of various therapies, within the recognized limits of potential biological differences between an experimental animal and humans. Such models are also useful for basic studies of copper transport, which is still not well understood.
The only other confirmed animal model of WND is the LEC rat which has a deletion of the 3' region of the WND homologue (13 ). Many of the physiological features of the LEC rat are also found in the tx mouse. The pattern of hepatic accumulation of copper is very similar: the accumulation of copper commences after weaning and most of the excess copper is bound to metallothionein (9 ,10 ,19 ). The histological appearance of the liver is also similar in both mutants, showing enlarged hepatocytes and abnormal nuclei (10 ,20 -22 ). In both mutants, the amount of ceruloplasmin in plasma is decreased (9 ,23 ). The reduced ceruloplasmin has been shown to be a consequence of lower incorporation of copper into this protein in the LEC rat (23 ,24 ), which is also the case in most Wilson patients [reviewed by Danks (1 )]. Thus the similarities between the rat and mouse models are in accordance with our finding of the mutation in the murine WND homologue.
The differences between the phenotype of patients with WD and the tx mouse have cast doubt on its validity as a model of that disease (10 ). Now it has been established that the tx mouse and the LEC rat have mutations of WND, it is worth considering the possible explanations for their differences. It should be noted that WD itself presents with quite a variable phenotype, and the reasons for this are not understood (1 ). An analysis of the specific phenotypes of animal models may elucidate the factors causing this clinical variability.
The most striking differences between the tx mouse and patients with WD as well as the LEC rat is the production of copper deficient milk by the tx mouse. Although pups of tx dams are always copper deficient, this does not invariably cause death. Indeed in some colonies of tx mice, with the same genetic background, death of pups is rarely observed (J. McC. Howell, pers. comm.), suggesting that environmental factors such as concentration of copper in the mother's diet can modify the phenotype. It is also possible that some degree of copper deficiency may be present in the LEC rat pups, since the copper status of milk from LEC dams has not been reported. Copper deficiency in breast fed babies of mothers with WD has not been reported to the best of our knowledge, but it would be of interest to determine the copper concentration of milk from lactating mothers with WD. In any event the copper deficiency in the suckling mice raises the interesting possibility that WND may be involved in delivery of copper to milk. Another possible explanation for the low copper in milk is that WND is required for the incorporation of copper into ceruloplasmin in the liver, and ceruloplasmin is providing copper to the mammary gland and placenta (25 ).
Another possible reason for the differences in phenotype between the tx mouse and WD is that specific mutations may result in distinct phenotypes. Correlation of phenotype with genotype in WD is complicated by the fact that most patients are compound hetereozygotes, with two different mutations of WND. Interestingly, one patient with a very early onset WD proved to be homozygous for a null allele (8 ). The LEC rat having a partial deletion of WND may be a model of this class of patients. The effect of the tx mutation of the activity of WND has not been established, but it is possible that some residual activity may be maintained with a subsequent modification of the phenotype.
In conclusion, the evidence strongly suggests that the toxic milk mouse is a true model for WD. This is of importance for the study of WD and copper transport in general. The tx mouse may prove to be a better model for WD since the mutation is similar to some mutations found to cause WD, whereas no large deletions, such as found in the LEC rat, have been reported in WD patients. It is possible that the tx mouse retains some residual WND activity which could be of importance in the study of different therapeutic agents used for treatment of WD. A mouse model offers other advantages over the LEC rat, e.g. the possibility of transferring the mutation to many more defined genetic backgrounds, thus exploring the question of the effect of modifying genes on the expression of the mutation, which is of relevance to understanding the phenotypic variation seen in WD (8 ,26 ). The more advanced transgenic possibilities with mice will facilitate gene correction studies.
The tx mutation first arose in 1974 in the DL strain which was at that time at F68 of brother sister mating. The tx mutation has been maintained in the DL strain subsequently and in 1983 was at F88 (9 ). Our colony was started from animals supplied by Dr H. Rauch.
The rat WND clone 8/2, spanning nucleotides 1310-3742 of the rat coding sequence (13 ), was used to probe a BALB/c liver cDNA library in [lambda]gt11 (Clonetech #ML3001b). Inserts from positive [lambda] bacteriophage clones were subcloned into pBluescript KS (Stratagene) for sequence analysis. Clones identified as being representative of the mouse WND homologue were used in subsequent screening of the library to obtain overlapping clones which covered the open reading frame. DNA sequencing was performed on double stranded plasmid DNA according to the protocols provided with the Sequenase sequencing kit (United States Biochemical).
RNA was extracted from various mouse tissues using the RNeasy kit (QIAGEN) according to manufacturers' protocols. Total RNA (10 [mu]g) from each tissue sample was run on a denaturing formaldehyde 1% agarose gel, and blotted on to Hybond N+ nylon membranes as described previously (27 ). RNA was fixed using 0.05 M NaOH, and rinsed in 2 * SSC prior to hybridization. DNA probes were inserts from [lambda] library clones 3 and 4B, which comprised cDNA encoding the region from the second metal binding site to the fifth transmembrane domain. Probes were labeled with [alpha]-32P-dCTP using a random priming kit (Boehringer Mannheim), hybridized overnight at 65oC in a solution containing 500 mM Na2HPO4, 5% SDS, 5 * Denhardt solution, 10 mM Tris pH 7.5, 5 mM EDTA and 100 [mu]g/ml sheared and denatured salmon sperm DNA. Blots were washed to 0.5 * SSC, 0.1% SDS at the same temperature, and exposed at -70oC using intensifying screens.
Using the sequence derived from the mouse library clones, primers were designed which would allow reverse transcription and PCR amplification of the WND homologue from DL and tx liver RNA. Reverse transcription was performed at 42oC with the following reaction mixture: 5 [mu]g total liver RNA, 500 ng primer (oligo dT(30) or internal primer), 300 [mu]M dNTP, 1 unit RNase inhibitor (Promega), 50 mM Tris pH 8.5, 8 mM MgCl2, 30 mM KCl, 1 mM DTT, 24 units AMV reverse transcriptase. A one-tenth volume of this reaction was used as template for PCR amplification, to which was added a reaction mixture containing: 750 ng each of forward and reverse primers, 200 [mu]M dNTP, 50 mM KCl, 10 mM Tris pH 8.3, 1.5 mM MgCl2 and 1.0 units Taq DNA polymerase (Boehringer Manneheim). Amplification was performed using 35 cycles of 95oC for 90 s, 55oC for 90 s, and 72oC for 120 s. PCR products were isolated by electrophoresis on agarose gels, bands were excised and purified using Qiaex DNA extraction kit (QIAGEN) and ligated into T-tailed pBluescript (29 ). Clones were sequenced according to Sequenase protocols for double-stranded plasmids. PCR primer pairs as follows (5' to 3') :
(i)TX011 AACATCCTCAGAAAGC and MW9B.2 GCTTAGGGAGACTTTGATTC; (ii)TX531 CTCTATCGAGGGCAACATCC and MW4B.1 GAGATGGATGTTTGCTGCAC; (iii)MW4B.2 GTGCAGCAAACATCCATCTC and TX2155 TAGAAGTACCACCCACCCAG; (iv)MW3.1 CTCATCTTCTTCATCTTGTG and TX1734 TTGTGTGCCATCTCCAGAGG; (v)MW1.1 CCCTAGCAAGCACATCTCGC and TX2181 ATCTCGTGATCTGTCATGGC; (vi)TX3491 TAACCATCTCCAGTGACATC and TX4202 CCGCTGACTTGGGAGGCAC; (vii)TX4265 CGACCTAGAGAGATGTGAGG and TX4703 TTCTGGAAGAGCAATCCTGC.
We are grateful to Sophie Gazias for maintenance of the mouse colony and assistance with mouse tissue collection. MBT is a Helen M. Schutt Ph.D. Fellow. This work was supported in part by a Block grant from the National Health and Medical Research Council of Australia.
Human Molecular Genetics
Pages
Introduction
Results
Discussion
Materials And Methods
Source of animals
Isolation and sequencing of cDNA clones
Northern blot analysis
RT-PCR and PCR cloning
Acknowledgements
References
REFERENCES
This page is maintained by OUP admin. Last updated Thu Oct 31 15:27:49 GMT 1996. Part of the OUP Journals World Wide Web service.Copyright Oxford University Press, 1996
