Expression of Hailey-Hailey disease mutations in yeast.

The yeast Saccharomyces cerevisiae, with its compact genome and tractable genetics, is a valuable model organism for the heterologous expression of mammalian proteins. In this letter, we show that phenotypic screening in yeast provides insight into the molecular basis of a human disease. Hailey–Hailey disease (HHD), or benign familial pemphigus, is an autosomal dominant intraepidermal blistering skin disorder. HHD has been mapped to ATP2C1 (Hu et al, 2000; Sudbrak et al, 2000), which encodes the Secretory Pathway Ca2þ–ATPase 1 (hSPCA1). This protein shares 49% identity with Pmr1, the founding member of a family of Golgi-localized Ca2þ and Mn2þ -transporting pumps (Rudolph et al, 1989; Sorin et al, 1997). The phenotypes of pmr1 mutants are similar to the ion homeostasis and cell surface abnormalities reported in HHD keratinocytes (Hu et al, 2000; Sudbrak et al, 2000). Defects in protein sorting, processing and glycosylation in pmr1 mutants result from Ca2þ and Mn2þ deficiency within the secretory pathway (Antebi & Fink, 1992; Durr et al, 1998). Such defects can be assessed by hypersensitivity to cation chelators (BAPTA or EGTA) (Rudolph et al, 1989; Wei et al, 2000). In addition, pmr1 mutants fail to clear calcium and manganese from the cytosol, thus their growth is inhibited by excess extracellular levels of these ions (Halachmi and Eilam, 1996; Wei et al, 2000). Quantitative growth assays in BAPTAor Mn2þ supplemented media have been very effective in screening Pmr1 site-directed point mutations. These approaches identified the critical role of membrane helices M4, M5, and M6 in ion binding, transport, and Ca2þ / Mn2þ selectivity (Mandal et al, 2000, 2003; Wei et al, 2000). Strikingly, the distribution of 15 missense mutations identified in HHD patients (Hu et al, 2000; Sudbrak et al, 2000; Chao et al, 2002; Dobson-Stone et al, 2002; Yokota et al, 2002; Li et al, 2003) are clustered in the stretch of polypeptide including M4, M5, M6, and the intervening cytosolic loop that constitutes the nucleotide-binding and phosphorylation domains characteristic of P-type ion pump superfamily (Fig 1). In contrast, chain termination or nonsense mutations scatter along the entire length of hSPCA1 (not shown). Among the point mutations, D742Y and A304S most likely disrupt ion binding, as seen in a homology model of hSPCA1 based on the crystal structure of the related sarco/endoplasmic reticulum Ca2þ -ATPase from rabbit muscle in the high affinity calcium-binding conformation (Toyoshima et al, 2000; Ton and Rao, 2004). The carbonyl oxygen of D742 directly coordinates Ca2þ and has been shown to be indispensable for Pmr1 function, whereas the backbone carbonyl oxygen of A304 binds calcium in SERCA1. L318P, T709M, and P744R are likely to cause structural perturbations, since replacement of the equivalent residues with Ala in yeast Pmr1 could be tolerated without loss of function (Wei et al, 2000, Mandal et al, 2003). Two other substitutions of Ala and Gly residues with bulkier side chains (A746T, G309C) are also likely to destabilize protein structure based on their location in the homology model (Ton and Rao, 2004). A random subset of six HHD missense mutations shown in Fig 1 was introduced into hSPCA1 and expressed in a yeast mutant lacking endogenous Ca2þ -ATPases. We had shown that heterologous expression of ATP2C1 in yeast completely restored mutant phenotypes resulting from disruption of PMR1 (Ton et al, 2002). Here, HHD mutants L341P, C411R, T570I, T709M, and D742Y fail to correct BAPTA or Mn2þ hypersensitivity of pmr1D (Fig 2), indicative of loss of Ca2þ and Mn2þ transport function, respectively. Indeed, Fairclough et al (2003) reported a loss of Ca2þ and Mn2þ -dependent phosphoenzyme intermediate formation, and low protein expression level for four of these mutants (L341P, C411R, T570I, and D742Y) expressed in COS-1 cells, whereas the effects of mutant T709M have not been reported before. The same study, however, failed to discern any defects in enzyme catalysis, expression level or Golgi localization of mutant P201L. Consistent with these results, in the yeast model, P201L, like wild-type hSPCA1, fully restores cellular phenotypes of pmr1D. Could P201L be a benign polymorphism, despite its absence from 100 control sequences (Sudbrak et al, 2000; Fairclough et al, 2003)? If so, the underlying cause of HHD in patients might be undetected mutations in introns or promoter regions. These findings underscore the importance of confirming the effects of amino acid substitutions with functional assays. It has been hypothesized that the autosomal dominant mode of inheritance of HHD is due to haploinsufficiency of SPCA1 rather than dominant negative effects of the mutants (Hu et al, 2000); however, there has not been direct experimental evidence supporting this speculation. We attempted to distinguish these possibilities by expressing the six HHD mutations, shown in Fig 2, in yeast carrying wild-type Pmr1. Given the high level of homology between Pmr1 and hSPCA1, any dominant negative effect of the mutants would have been detectable by changes in phenotype. Yeast expressing both Pmr1 and HHD mutants, however, showed no hypersensitivity to BAPTA or Mn2þ toxicity (Fig S1). Thus, at least in yeast, HHD mutants appear to exert no Abbreviation: HHD, Hailey–Hailey disease