document

letter

nature genetics • volume 21 • february 1999 195

Hephaestin, a ceruloplasmin homologue implicated inintestinal iron transport, is defective in the sla mouse

Christopher D. Vulpe1,2, Yien-Ming Kuo2, Therese L. Murphy3, Lex Cowley3, Candice Askwith4,Natasha Libina2, Jane Gitschier1,2 & Gregory J. Anderson3

1Howard Hughes Medical Institute and the 2Department of Medicine, University of California at San Francisco, San Francisco, California, 94143, USA.3Clinical Sciences Unit, Queensland Institute of Medical Research and Joint Clinical Sciences Program, University of Queensland, Brisbane, Queensland,Australia. 4Department of Pathology, University of Utah School of Medicine, Salt Lake City, Utah, USA. Correspondence should be addressed to J.G. (e-mail: [email protected]).

Iron is essential for many cellular functions; consequently, dis-turbances of iron homeostasis, leading to either iron deficiencyor iron overload, can have significant clinical consequences.Despite the clinical prevalence of these disorders, the mecha-nism by which dietary iron is absorbed into the body is poorlyunderstood. We have identified a key component in intestinaliron transport by study of the sex-linked anaemia (sla) mouse,which has a block in intestinal iron transport1. Mice carrying thesla mutation develop moderate to severe microcytic hypo-chromic anaemia1. Although these mice take up iron from theintestinal lumen into mature epithelial cells normally2, the sub-sequent exit of iron into the circulation is diminished3. As aresult, iron accumulates in enterocytes and is lost duringturnover of the intestinal epithelium4. Biochemical studies havefailed to identify the underlying difference between sla andnormal mice, therefore, we used a genetic approach to identifythe gene mutant in sla mice. We describe here a novel gene,Heph, encoding a transmembrane-bound ceruloplasmin homo-logue that is mutant in the sla mouse and highly expressed inintestine. We suggest that the hephaestin protein is a multi-copper ferroxidase necessary for iron egress from intestinalenterocytes into the circulation and that it is an important linkbetween copper and iron metabolism in mammals.Building on prior genetic mapping studies, we narrowed theregion likely to contain the sla gene through exclusion mappingand ancestral chromosome mapping (Fig. 1). Previously, the slalocus was mapped to the central span of the mouse X chromo-some, 3 cM centromeric of tabby (Ta) and 10 cM telomeric ofbent tail (Bn; ref. 5). A recent intra-specific cross identified a criti-cal region between the markers DXMit45 and DXMit16 likely tocontain the gene6. A portion of this 5-cM region was excluded byanalysis of a large deletion mutant (Ta25H) that does not sufferfrom anaemia7. By testing markers in the DXMit45−DXMit16region for their presence or absence in Ta25H mice, we found thatthe sla locus must lie proximal to DXMit16, DXMit96 andDXMit114 (data not shown). Ancestral chromosome mappingfurther narrowed the sla mutation, which arose in a different (andunknown) genetic background from the C57BL/6J strain in whichit is currently maintained. We identified four dinucleotide repeatmarkers (DXMit93, DXMit63, DXMit113 and DXMit114) in theDXMit45−DXMit16 region in which alleles in C57BL/6J-sla differfrom those in C57BL/6J. As one of these markers, DXMit114, isalso deleted in Ta25H mice, the remaining three markers define aregion of approximately 2 cM in which the sla gene must reside.

A coincident project led to the discovery of a sla candidate gene,termed hephaestin (Heph) after the Greek god of metal-working.We noted several mouse and human ESTs with homology to ceru-loplasmin (encoded by the Cp gene), a serum multi-copper ferrox-

idase, and we assembled a complete mouse cDNA that encoded aprotein 50% identical to mouse ceruloplasmin (Fig. 2). All type I,II and III copper-binding sites in ceruloplasmin are conserved inthe predicted hephaestin protein, as are all cysteine residuesinvolved in disulfide bond formation. In contrast to ceruloplas-min, hephaestin contains a predicted carboxy-terminal transmem-brane domain, which suggests a membrane-bound protein with aceruloplasmin-like domain located in an extra-cytosolic compart-ment or in the extracellular space. Radiation-hybrid mapping indi-cated that human HEPH maps within 14.55 cR of DXS1194 inXq11−12 (lod score, 7.81), a region with homology by synteny tothe mouse sla region (data not shown). This localization was con-firmed in both human and mouse by PCR analysis of somatic cellhybrids and of several human YACs (665A3 and 626G11, MIT YAClibrary) that contained DXS1194 (data not shown). Furthermore,we identified a Heph-containing mouse BAC (285L22) and foundit to overlap with a YAC containing DXMit113 (Fig. 1), thereforedemonstrating that Heph maps to the sla interval.

Fig. 1 Mapping of the sla locus. The region on the X chromosome containingthe sla gene was defined by genetic mapping, exclusion mapping with theTa25H deletion mouse and ancestral chromosome mapping. The central regionof the mouse X chromosome is expanded and the marker and mutant loci areindicated. The region excluded by analysis of the Ta25H deletion is indicated bya hatched bar. Alleles found during ancestral chromosome mapping inC57BL/6J or C57BL/6J-sla animals for the polymorphic markers studied are indi-cated by C or S, respectively. Lines indicate the locations of the BAC (285L22)containing Heph and the DXMIT113-containing YAC (356H8). The location andrelative order of markers in this region are based on the 1998 chromosomecommittee consensus report (http://www.informatics.jax.org), MIT YAC contigs(http://carbon.wi.mit.edu:8000/cgi-bin/mouse/index) and MRC physical map ofthe mouse X chromosome as of July 1998 (http://www.mgc.har.mrc.ac.uk/xmap/xmap.html).

DXMit45

DXMit60, 61

DXMit210DXMit8

DXMit147

DXMit93DXMit63

DXMit113

DXMit114

DXMit16, 96Ar Ta

Genetic mapping

Heph

Ta25HExclusion mapping

Ancestral chromosome mapping

© 1999 Nature America Inc. • http://genetics.nature.com©

199

9 N

atu

re A

mer

ica

Inc.

• h

ttp

://g

enet

ics.

nat

ure

.co

m

letter

196 nature genetics • volume 21 • february 1999

Heph expression is consistent with the primary intestinal phe-notype of decreased iron transport in affected sla mice. Expres-sion studies with northern blots prepared from adult mousetissues revealed high Heph expression in small intestine andcolon, and low levels of expression in other tissues (Fig. 3). Hephexpression contrasts that of Cp, which is highly expressed in liver

and expressed to a lesser extent in other tissues including brain8

and lung, but is not expressed in intestine. In situ hybridizationstudies indicated that intestinal expression of Heph is limited tovilli (Fig. 3). Almost no signal was observed in crypt cells. A rolefor hephaestin in the transport of iron by enterocytes is consis-tent with this finding, as iron absorption occurs in villi9.

Fig. 2 Comparison of mouseceruloplasmin- and hephaestin-derived amino acid sequences.Similarity between the derivedamino acid sequence of hep-haestin and ceruloplasmin wasdetermined with the GCG8(Genetics Computer Group)PILEUP program. Conservedresidues that constitute type I, IIand III copper-binding sites inceruloplasmin are highlightedin light blue, dark blue andgreen, respectively. Conserveddisulfide-bonding cysteines areshown in red. Hephaestin con-tains an additional 86 residuesat the C terminus, which con-tains a predicted transmem-brane domain, highlighted inlight green. Brackets indicatethe region of hephaestindeleted in the sla mouse.

Fig. 3 Expression of Heph in wild-type mice. a, Northern-blot analysis of HephmRNA. Two northern blots containing poly(A)+ RNA (from a Swiss-Webstermouse) were hybridized with a Heph cDNA probe. A single, approximately 4.5-kb band is detected in multiple tissues, suggesting widespread distribution, buthigher expression was observed in small intestine. A 2-h exposure of the sameblot for small intestine is also shown. A control hybridization of an Actb cDNAprobe (β-actin) with the same blots is shown. b–e, In situ hybridization studies.Adjacent sagittal cross-sections of proximal small intestine from wild-type CD1mice were hybridized in situ with radiolabelled antisense (b) and sense (c)Heph RNA probes and visualized as darkfield images. A ×10 magnification ofantisense hybridization (d) indicates predominant expression of Heph in thecells of the mid to upper villus (V) versus the cells of the crypt (C) of the smallintestine. A ×40 magnification (e) demonstrates continued expression of Hephin the distal villus tip.

a b c d e


199

9 N

atu

re A

mer

ica

Inc.

• h

ttp

://g

enet

ics.

nat

ure

.co

m

letter


Examination of intestinal mRNA in sla mice revealed a partialdeletion of Heph. RT-PCR amplification of the central portion ofHeph resulted in a smaller product in sla mice compared withnormal controls (Fig. 4). Sequence analysis indicated a deletionof 582 nt (2,094−2,676), predicting an in-frame omission of 194aa in hephaestin (Fig. 2). Northern-blot analysis of mRNA pre-pared from sla and normal mice confirmed a smaller message insla mice (Fig. 4).

On the basis of the above findings, and by comparison with thestructure of Cp, we predicted that three Heph exons would bedeleted in sla mice; however, analysis of genomic DNA showed thatonly two of three exons are deleted (Fig. 5). Using human CP exon-intron boundaries as a guide, we determined Heph exon-intronboundaries for the three exons contained in the 582-nt regiondeleted in sla mice. Only two of three exons failed to amplify fromgenomic DNA of sla mice by PCR (data not shown). Hybridizationwith exon-specific probes to HindIII-digested DNA confirmed thepresence of exon 9 and the absence of exons 10 and 11 in genomicDNA of affected mice (data not shown). A genomic deletion ofapproximately 3.5 kb with breakpoints distal to exon 9 and proxi-mal to exon 12 explains these results, although a more complexrearrangement is possible. Thus, deletion of two exons of Heph isassociated with a sla transcript lacking three exons, possibly due toa disruption of splicing signals in the ‘first’ intron.

We tested for segregation of the genomic deletion and the slaphenotype in the intra-specific cross performed previously,which yielded 23 recombinations between DXMit45 and the slalocus and three recombinations between DXMit16 and sla.Southern blots prepared from genomic DNA from the 26 recom-binants were hybridized with a probe spanning the deletion (datanot shown). In all cases, the altered and missing bands indicativeof the deletion cosegregated with the affected sla phenotype. Weconclude that this mutation underlies the phenotype in sla miceand designate the allele as Hephsla.

Hephaestin provides the first molecular insight into the path-way of iron export from intestinal epithelial cells. The diminishedtransport of iron from the mucosal cell to the circulation in sla

mice provides evidence that hephaestin has a role in this impor-tant physiological pathway. On the basis of its homology withceruloplasmin, we propose that hephaestin is a ferroxidase neces-sary for iron release from intestinal epithelial cells. As hephaestincontains only one putative membrane-spanning domain, it isunlikely to be a transmembrane iron carrier itself, and we pro-pose that hephaestin interacts with an iron-transport protein tofacilitate the movement of iron across the membrane.

High-affinity iron uptake by Saccharomyces cerevisiae uses ananalogous system10. The yeast protein Fet3, a multi-copper oxi-dase with homology to ceruloplasmin (and with ferroxidase activ-ity), is essential for high-affinity iron import11. Similar to thepredicted hephaestin protein, Fet3 is tethered to the cell surfacevia a C-terminal transmembrane domain, and its ferroxidasedomain is extracellular12. Fet3 forms a complex on the plasmamembrane with Ftr1, the actual transmembrane iron transporter,and both proteins are essential for iron uptake13. In contrast to therole we propose for hephaestin, the ferroxidase activity of Fet3 isnecessary for iron import rather than iron export.

Hephaestin, like Fet3 in yeast, represents a link between copperand iron metabolism in mammals and offers a basis for the iron-deficiency anaemia associated with copper deficiency. Copper defi-ciency results in the decreased absorption of dietary iron, whichenters intestinal epithelium normally but cannot exit into the circu-lation14. Indeed, intestinal iron accumulation in copper-deficientswine is similar to the iron accumulation seen in sla mice14. Theadministration of copper, but not iron, to copper-deficient pigs alle-viates the anaemia and facilitates the egress of iron from tissuesincluding intestine. If hephaestin requires copper to function nor-mally, as its homology predicts, then this provides an explanation forthe copper-dependent release of iron from intestinal epithelial cells.

Copper and iron are similarly linked in systemic iron metabo-lism. The congenital absence of ceruloplasmin in humans leadsto iron accumulation in many tissues15. Iron, however, does notaccumulate in the intestine and only a mild anaemia has beennoted in some (6/9) patients16, suggesting that hephaestin andceruloplasmin perform similar but distinguishable functions.

Fig. 4 Analysis of Heph in sla mice. a, RT-PCR of Heph mRNA. Poly(A)+ mRNAfrom small intestine of two control C57BL/6J and two mutant C57BL/6-sla ani-mals was prepared. RT-PCR with primers spanning the central region of cDNA(nt 2,067−3,510) amplified a smaller size product from sla mice compared withthat from control mice. Sequence analysis of the RT-PCR products revealed a582-nt deletion (2,094–2,676) in sla RT-PCR products. b, Northern analysis ofHeph mRNA. Poly(A)+ mRNA was prepared from small intestine and colon ofC57BL/6J and C57BL/6J-sla animals. The blot was hybridized with Heph cDNAprobe from outside the deletion (nt 1,610−2,226). A 4-h exposure is shown. Atranscript of approximately 4.5 kb is evident in C57BL/6J mice, and a shortertranscript of approximately 4 kb is present in C57BL/6J-sla mice.

Fig. 5 Schematic depiction of the partial genomic deletionof Heph in sla animals. In C57BL/6J DNA, exons 9−10 areboth present on a 3-kb HindIII fragment, whereas exon 11is present on a 6.5-kb fragment. The cDNA sequence indi-cates a HindIII site in exon 12. In DNA from sla mice, anexon-9 probe hybridizes with an approximately 6-kb frag-ment, whereas exons 10 and 11 do not hybridize.

a b


199

9 N

atu

re A

mer

ica

Inc.

• h

ttp

://g

enet

ics.

nat

ure

.co

m

letter

198 nature genetics • volume 21 • february 1999

Hephaestin also provides at least a partial explanation for theparadoxical observation that whereas nutritional copper defi-ciency results in anaemia, anaemia is not manifest in two inher-ited disorders of copper deficiency: Menkes disease in humansand the mottled mouse. In these disorders, defects in a copper-transporting ATPase result in a cellular phenotype of defectivecopper export in most cells (except hepatocytes) and the entrap-ment of copper in some tissues, such as intestine17. We suggestthat in these conditions, copper is available for incorporationinto hephaestin and would therefore be able to function in theexport of iron from the gut; thus, no anaemia is manifested.

Humans maintain iron homeostasis by regulating intestinaliron absorption because the capacity to excrete iron is very lim-ited18. Iron absorption is decreased in iron-replete mammals andis increased when iron stores are low. In iron excess, a decrease inbasolateral iron export results in the accumulation of iron inabsorptive enterocytes, which are lost during the normal recy-cling of intestinal mucosa19. Regulation of iron export may there-fore have a role in the regulation of iron absorption. In sla mice,iron export is blocked despite overall iron deficiency, and slaenterocytes resemble cells from an iron-replete or overloadedanimal. In contrast, haemochromatosis patients inappropriatelyabsorb iron into an ever-more iron-overloaded body20. Dysregu-lation of iron absorption21 due to a mutation in the haemochro-matosis gene product (HFE) results in elevated intestinal ironexport. The target of HFE regulation is not known, but we sug-gest that regulation of the activity of hephaestin could providethe means to regulate iron export from the intestine.

MethodsDNA extraction and microsatellite analysis. DNA was extracted from livertissue of C57BL/6J and C57BL/6J-sla mice as described22. Genomic DNAfrom Ta25H/Y mice was kindly provided by N. Brockdorff. Microsatellitemarkers were amplified by PCR with DXMit primers (Research Genetics)as described6.

Identification of Heph cDNA. Initial Heph cDNAs were identified by BLASTof dbEST with Cp cDNA at the NCBI server. A mouse EST correspondingwith 3´ UTR and C-terminus and a human EST corresponding with the cen-tral region of the cDNA were initially identified. Iterative searching of dbESTidentified additional mouse and human ESTs. Mouse cDNA was initiallyextended by hemi-nested PCR of a mouse brain cDNA library (2.5 µl of Strat-agene library 937914) with one primer in the vector (T7) and nested primerscorresponding with Heph cDNA (primer 1, 5´−CCCTGTAACTTTGCC-CATTC−3´; primer 2, 5´−CCCCTGAGGTTTGCATAGAG−3´). PCR amplifi-cation reaction conditions were a 16:1 mix (0.5 µl) of Klentaq (Ab peptides)and PFU (Stratagene) in 10×PC2 buffer (Ab peptides), primers (10 pms each)and dNTPs (0.25 mM; Pharmacia), with one cycle of 95 °C for 2 min, fol-lowed by 35 cycles of 95 °C, 10 s; 55 °C, 30 s−1 min; 68 °C, 4−8 min and a finalextension of 10 min at 68 °C. Additional sequence was obtained by RT-PCR ofC57BL/6J intestinal mRNA with a primer in the mouse EST and a primer cor-responding to a region of hephaestin conserved between human ceruloplas-min and human hephaestin EST. 5´ sequence was obtained using 5´-RACEwith mouse embryo cDNA (Clontech Marathon cDNA) by PCR nesting fol-lowing the manufacturer’s instructions (AP1 primer (Clontech) and primer3, 5´−TGGTCGACTGGCCTTATTCT−3´, followed by AP2 primer (Clon-tech) and primer 4, 5´−CCAGACCTTGGCCGAGGTATTCTG−3´). cDNAclones were obtained from Research Genetics and/or Genome Systems. Semi-automated sequencing of cDNA clones and PCR was performed by theUCSF-HHMI sequencing core and UCSF Human Genetics core facility usingan ABI 373A automated sequencer (Applied Biosystems).

Mapping of Heph. Primers corresponding with the 3´ UTR of human Hephwere used on the Stanford Radiation Hybrid panel G3 DNA for localization.

The forward primer was 5´−ACTGAGGCCAAGTGAGCTG−3´ and thereverse primer was 5´−CAACATTCCTTTCAGTGCCA−3´. PCR data weresubmitted to the radiation hybrid e-mail server (http://www-shgc.stan-ford.edu/RH/index.html), which indicated an X-chromosome locationwithin 14.55 cR of DXS1194 in Xq11−12 (lod score, 7.81). PCR amplifica-tion of DNA from an X-chromosome regional somatic cell hybrid mappingpanel obtained from Coriell Cell Repository confirmed the Xq11−12 loca-tion. We obtained MIT YACs containing DXS1194 (Research Genetics) andperformed PCR analysis with 3´ UTR primers on genomic DNA preparedusing Qiagen columns following the manufacturer’s instructions. We car-ried out hybridization screening with a probe corresponding to the 3´ UTRof mouse Heph of a gridded mouse C57BL/6J BAC library (Genome Sys-tems; FBAC-4471) in CHURCH buffer at 65 °C and identified a single BACclone (285L22). PCR amplification with the primers corresponding to 5´UTR, intron sequence between exons 9 and 10 and 3´ UTR confirmed thepresence of a complete gene in the BAC. BAC ends were obtained by directsequencing of BAC DNA (5 µl). PCR primers were designed to the BACends and were tested on mouse YACs in the DXMit45−DXMit16 interval. Asingle mouse YAC (356H8) containing DXMIT113 also contained sequencecorresponding with a 285L22 BAC end.

RNA preparation and RT-PCR. Tissues from C57BL/6J-sla and C57BL/6Jwild-type mice were snap frozen in liquid nitrogen. We isolated total RNAby homogenizing each sample in Trizol reagent (Life Technologies), follow-ing the manufacturer’s instructions. For northern blots, we isolatedpoly(A)+ RNA from total RNA using Poly(A)+ Tract mRNA Isolation Sys-tem IV (Promega). For RT-PCR, poly(A)+ RNA was prepared with a Phar-macia Quick Prep mRNA preparation kit following the manufacturer’sinstructions. Reverse transcription of mRNA was performed with Super-script reverse transcriptase (Bethesda Research Laboratories) following themanufacturer’s instructions. PCR products were separated on 1% Trisacetate EDTA agarose gels and visualized with ethidium bromide.

Northern blots and in situ hybridization. Multiple tissue mouse poly(A)+

mRNA blots were obtained from Origene. C57BL/6J and C57BL/6J-slablots were prepared by electrophoresis of sample (1 µg) on a 1.2% agarose-formaldehyde gel. The gel was washed in 20 symbol 180×SSC for 30 minand capillary transferred onto a Hybond N (Amersham-Pharmacia) nylonmembrane. The blot was prehybridized in a solution consisting ofNa2HPO4 (0.5 M), EDTA (1 mM), 7% SDS and 1% BSA for 4 h at 65 °C,then hybridized overnight at the same temperature with the relevant Hephprobe. The filter was washed as follows: 2×SSC, 0.1% SDS for 3×20 min atRT; then 0.2×SSC, 0.1% SDS for 2×30 min at 65 °C and exposed to Fuji RXX-ray film at –70 °C. A probe for hybridization of the multiple-tissuenorthern blots was prepared by RT-PCR of C57BL/6J intestinal mRNAusing primers corresponding with nt 2,068−2,810 in the cDNA. A 300-bpprobe to the Heph-deleted region was generated by PCR using primers cor-responding with nt 2,369−2,668. A 637-bp probe outside the deleted regionwas generated using the following primers: forward, 5´−TGTGACTGCT-GAGATGGTGC−3´; reverse, 5´−TGCAGCAGAGAAGTACATCC−3´. Insitu hybridization was performed as described23 with antisense and senseradiolabelled RNA probes corresponding to nt 1,122−1,809 of Heph onduodenal sections from CD-1 mice (Charles River Laboratory).

GenBank accession numbers. Mouse EST, AA269874; human EST,W46354; full-length Heph cDNA, AF082567.

AcknowledgementsWe thank N. Brockdorff and E. Tuddenham for assaying the sla animals; M.Gunthorpe and J. DeYoung for DNA synthesis and sequencing; M. Schuelerand H. Willard for communicating unpublished results; and M. Fleming andN. Andrews for unpublished sla mapping data. J.G. is an associate investigatorwith the Howard Hughes Medical Institute. This work was supported in part bya grant from the National Health and Medical Research Council of Australia toG.J.A. and by a grant from the National Institutes of Health to J.G.

Received 1 September; accepted 21 December 1998.


199

9 N

atu

re A

mer

ica

Inc.

• h

ttp

://g

enet

ics.

nat

ure

.co

m

letter


1. Grewal, M.S. A sex-linked anemia in the mouse. Genet. Res. 3, 238–247 (1962).2. Manis, J. Active transport of iron by intestine: selective genetic defect in the

mouse. Nature 227, 385–386 (1970).3. Bannerman, R.M. Genetic defects of iron transport. Fed. Proc. 35, 2281–2285 (1976).4. Edwards, J.A., Hoke, J.E., Mattioli, M. & Reichlin, M. Ferritin distribution and

synthesis in sex-linked anemia. J. Lab. Clin. Med. 90, 68–76 (1977).5. Falconer, D.S. & Isaacson, J.H. The genetics of sex-linked anaemia in the mouse.

Genet. Res. 3, 248–250 (1962).6. Anderson, G.J. et al. Mapping the gene for sex-linked anemia: an inherited defect

of intestinal iron absorption in the mouse. Genomics 48, 34–39 (1998).7. Brockdorff, N., Kay, G., Cattanach, B.M. & Rastan, S. Molecular genetic analysis of

the Ta25H deletion: evidence for additional deleted loci. Mamm. Genome 1,152–157 (1991).

8. Klomp, L.W., Farhangrazi, Z.S., Dugan, L.L. & Gitlin, J.D. Ceruloplasmin geneexpression in the murine central nervous system [see comments]. J. Clin. Invest.98, 207–215 (1996).

9. Skikne, B.S. & Baynes, R.D. in Iron Metabolism in Health and Disease (eds Brock,J.H., Halliday, J.W., Pippard, M.J. & Powell, L.W.) 151–187 (W.B. Saunders, London,1994).

10. Askwith, C.C., de Silva, D. & Kaplan, J. Molecular biology of iron acquisition inSaccharomyces cerevisiae. Mol. Microbiol. 20, 27–34 (1996).

11. Askwith, C. et al. The FET3 gene of S. cerevisiae encodes a multicopper oxidaserequired for ferrous iron uptake. Cell 76, 403–410 (1994).

12. De Silva, D.M., Askwith, C.C., Eide, D. & Kaplan, J. The FET3 gene product requiredfor high affinity iron transport in yeast is a cell surface ferroxidase. J. Biol. Chem.

270, 1098–1101 (1995).13. Stearman, R., Yuan, D.S., Yamaguchi-Iwai, Y., Klausner, R.D. & Dancis, A. A

permease-oxidase complex involved in high-affinity iron uptake in yeast. Science271, 1552–1557 (1996).

14. Lee, G.R., Nacht, S., Lukens, J.N. & Cartwright, G.E. Iron metabolism in copper-deficient swine. J. Clin. Invest. 47, 2058–2069 (1968).

15. Harris, Z.L. et al. Aceruloplasminemia: molecular characterization of this disorderof iron metabolism. Proc. Natl Acad. Sci. USA 92, 2539–2543 (1995).

16. Logan, J.I. Hereditary deficiency of ferroxidase (aka caeruloplasmin) [editorial]. J.Neurol. Neurosurg. Psychiatry 61, 431–432 (1996).

17. Danks, D.M. in The Metabolic Basis of Inherited Disease (eds Scriver, C., Beaudet,A., Sly, W. & Valle, D.) 1411–1432 (McGraw-Hill, New York, New York, 1989).

18. Bothwell, T.H., Charlton, R.W., Cook, J.D. & Finch, C.A. in Iron Metabolism in Man256–283 (Blackwell Scientific Publications, Oxford, 1979).

19. Lombard, M., Chua, E. & O’Toole, P. Regulation of intestinal non-haem ironabsorption. Gut 40, 435–439 (1997).

20. Halliday, J.W. & Powell, L.W. in Iron and Human Disease (eds Lauffer, R.B. et al.)131–160 (CRC Press, Boca Raton, 1992).

21. Bothwell, T.H. & MacPhail, A.P. Hereditary hemochromatosis: etiologic,pathologic, and clinical aspects. Semin. Hematol. 35, 55–71 (1998).

22. Davis, L.G., Dibner, M.D. & Battey, J.F. in Basic Methods in Molecular Biology 42–43(Elsevier, New York, 1986).

23. Kuo, Y.M., Gitschier, J. & Packman, S. Developmental expression of the mousemottled and toxic milk genes suggests distinct functions for the Menkes andWilson disease copper transporters. Hum. Mol. Genet. 6, 1043–1049 (1997).


199

9 N

atu

re A

mer

ica

Inc.

• h

ttp

://g

enet

ics.

nat

ure

.co

m

Documents

document