Report

Yeast Genes That Enhance the Toxicity of a Mutant Huntingtin Fragment or α-Synuclein

See allHide authors and affiliations

Science  05 Dec 2003:
Vol. 302, Issue 5651, pp. 1769-1772
DOI: 10.1126/science.1090389

Abstract

Genome-wide screens were performed in yeast to identify genes that enhance the toxicity of a mutant huntingtin fragment or of α-synuclein. Of 4850 haploid mutants containing deletions of nonessential genes, 52 were identified that were sensitive to a mutant huntingtin fragment, 86 that were sensitive to α-synuclein, and only one mutant that was sensitive to both. Genes that enhanced toxicity of the mutant huntingtin fragment clustered in the functionally related cellular processes of response to stress, protein folding, and ubiquitin-dependent protein catabolism, whereas genes that modified α-synuclein toxicity clustered in the processes of lipid metabolism and vesicle-mediated transport. Genes with human orthologs were overrepresented in our screens, suggesting that we may have discovered conserved and nonoverlapping sets of cell-autonomous genes and pathways that are relevant to Huntington's disease and Parkinson's disease.

Huntington's disease (HD) is a fatal, inherited neurodegenerative disorder that is characterized by disturbances in movement, cognition, and personality. The mutation that causes HD is an expansion of CAG repeats [encoding polyglutamine (polyQ)] in the gene IT-15 (which encodes the huntingtin protein) (1). Parkinson's disease (PD) is a major neurodegenerative disorder characterized by muscle rigidity, bradykinesia, resting tremor, and postural instability (2). Although the vast majority of cases of PD are idiopathic, a small percentage are caused by missense mutations of the α-synuclein gene (3, 4). One neuropathological feature shared by HD and PD is the occurrence of ubiquitinated intraneuronal inclusion bodies in diseased brains. Huntingtin, and/or truncation products of huntingtin, are the major components of cytoplasmic and nuclear inclusion bodies observed in HD, and α-synuclein is the major component of inclusion bodies (Lewy bodies) in PD. Huntingtin (5) and α-synuclein (6) assemble into fibrillar protein aggregates that display many properties of amyloid in vitro and in vivo. The precise roles of protein aggregation, amyloid formation, and inclusion bodies in HD, PD, and other amyloid diseases remain controversial, and it is not clear whether common pathogenic mechanisms occur in these disorders.

Here, we have used the baker's yeast Saccharomyces cerevisiae as a model eukaryotic organism to test the hypothesis that the downstream targets and molecular mechanisms by which a mutant huntingtin fragment and α-synuclein mediate toxicity are distinct. Similar to neurons, yeast transformed with mutant huntingtin fragments form inclusion bodies in a process regulated by yeast homologs of Hsp40 and Hsp70 (7, 8). As in many types of mammalian cells, overexpression of mutant huntingtin fragments in yeast has no effect on cell viability. This feature allows genetic screens to be performed to identify genes that unmask, or enhance, toxicity. Here, we have used a yeast gene deletion set (YGDS) of 4850 viable mutant haploid strains (9, 10) to identify genes that enhance toxicity of a mutant huntingtin fragment or α-synuclein.

In our model of huntingtin inclusion body formation in S. cerevisiae (8), yeast are transformed with constructs that express exon 1 of the huntingtin gene with a normal (HD20Q) and expanded (HD53Q) polyQ repeat under control of the CUP1 promoter. The aggregation and inclusion body–forming properties of huntingtin fragments with expanded polyQ tracts can be reproduced faithfully in S. cerevisiae (8). Similarly, overexpression of wild-type or mutant (A53T) human α-synuclein in yeast results in the formation of cytoplasmic inclusion bodies that, at the level of light microscopy, are similar to those formed by mutant huntingtin fragments in yeast (11).

A collection of 4850 yeast strains was transformed with constructs that express HD53Q or α-synuclein under the control of inducible promoters and was plated onto selective media in the absence of induction (12). Mutants that were sensitive to a HD53Q or α-synuclein were isolated by replica plating onto media that contained the appropriate inducer of protein expression. We confirmed putative HD53Q- or α-synuclein–sensitive mutants by retesting isolated colonies in spotting assays that measure cell viability (Fig. 1). Although positive colonies were selected originally because of their complete lack of growth (synthetic lethality) after induction, the retests indicated that a sublethal effect on toxicity (synthetic sickness) occurred in many of the deletion strains (Fig. 1). Of 4850 mutants, 52 (∼1%) were identified with enhanced toxicity of HD53Q, and 86 (∼2%) with enhanced toxicity of wild-type α-synuclein (Tables 1 and 2). Although overexpression of α-synuclein caused a modest decrease in cell viability in wild-type yeast (BY4741), all mutant strains that enhanced α-synuclein toxicity had a phenotype that was more severe than that observed in wild-type yeast (Fig. 1B).

Fig. 1.

Expression of a mutant huntingtin fragment (HD53Q) or α-synuclein causes synthetic sickness or lethality in yeast gene deletion strains. (A) Yeast cells transformed with empty vector (pSAL4), HD20Q, or HD53Q were grown in liquid synthetic complete medium lacking uracil (SC-URA) to log phase and then induced for 24 hours in SC-URA + copper. Samples of cells were removed from liquid cultures before (T = 0) and after (T = 24 hours) copper induction, were spotted on plates containing SC-URA ± 400 μM copper, and were incubated at 30°C for 3 days. Shown are fivefold serial dilutions starting with equal numbers of cells. Spotting assays derived from single transformants for 10 unique deletion strains and the parental control strain (BY4741) are shown. (B) Yeast cells transformed with empty vector (p426GAL) or α-synuclein (α-syn) were grown in liquid synthetic complete medium lacking uracil (SC-URA + glucose) to log phase and then induced for 6 hours in SC-URA + galactose. Samples of cells were removed from liquid cultures before (T = 0) and after (T = 6 hours) galactose induction, were spotted on plates containing SC-URA ± galactose, and were incubated at 30°C for 3 days. Shown are fivefold serial dilutions starting with equal numbers of cells. Spotting assays derived from single transformants for nine unique deletion strains and the parental control strain (BY4741) are shown.

Table 1.

Yeast strains synthetically sick or lethal with HD53Q. The ortholog category indicates yeast genes with human orthologs.

Strain Ortholog Function
1. apj1Δ Yes Hsp40 chaperone
2. apm2Δ Yes Nonselective vesicle transport
3. ayr1Δ Yes Ketoreductase; acylglycerone-phosphate reductase
4. cit2Δ Yes Citrate synthase, peroxisomal
5. cmk1Δ Yes Protein histidine kinase
6. cos111Δ No Possibly involved in ubiquitin pathway
7. cps1Δ Yes Gly-X carboxypeptidase
8. dcg1Δ No Possibly involved in cell wall biosynthesis
9. fil1Δ No Translation factor
10. fpr2Δ Yes Peptidyl-prolyl cis-trans isomerase
11. gda1Δ Yes Guanosine diphosphatase
12. glo2Δ Yes Hydroxyacylglutathione hydrolase
13. gre2Δ Yes Alpha-acetoxy ketone reductase
14. gsh2Δ Yes Glutathione synthase
15. hlj1Δ Yes Hsp40 chaperone in ER
16. hlr1Δ No Unknown, similar to Lre1 (Pkc1p-MAPK pathway)
17. hms1Δ Yes Transcription factor
18. ipk1Δ No Phosphatidylinositol phosphate kinase
19. kgd1Δ Yes Alpha-ketoglutarate dehydrogenase
20. msb1Δ Yes Activates Pkc1p-MAPK pathway
21. mrp11Δ No Protein of the mitochondrial large ribosomal subunit
22. mup1Δ Yes Methionine permease
23. pc16Δ No Cyclin-dependent protein kinase
24. phm8Δ No Possibly involved in phosphate metabolism
25. prm5Δ No Possibly involved in cell stress
26. psp1Δ No Possibly involved in DNA replication
27. rim4Δ Yes RNA binding
28. sam2Δ Yes S-adenosylmethionine synthetase 2
29. sas3Δ Yes Histone acetyltransferase
30. sdt1Δ No 5′-Nucleotidase
31. sip18Δ No Binds phospholipids
32. sng1Δ No Probable transport protein
33. stp2Δ Yes Transcription factor
34. tea1Δ No Transcriptional activator
35. tvp15Δ No Possibly involved in vesicular transport
36. ubp13Δ Yes Ubiquitin C-terminal hydrolase
37. vps70Δ Yes Possibly involved in vacuolar trafficking
38. yhb1Δ Yes Nitric oxide dioxygenase; oxygen transporter
39. yrb30Δ No Unknown
40. ybr100wΔ No Possibly involved in DNA damage repair
41. ybr255wΔ No Unknown
42. ydr215cΔ No Unknown
43. ygr015cΔ Yes Alpha or beta hydrolase fold family
44. yjr107wΔ No Has similarity to acylglycerol lipase
45. ykr017cΔ Yes Has a TRIAD composite zinc finger domain
46. ykr064wΔ No Transcription factor
47. ylr128wΔ Yes Basic helix-loop-helix leucine zipper protein
48. ymr160wΔ No Unknown
49. ynl296wΔ No Possibly involved in vacuolar trafficking
50. yor292cΔ Yes Peroxisomal protein
51. yor300wΔ No Bipolar budding and bud site selection
52. ypl067cΔ No Unknown
Table 2.

Yeast strains synthetically sick or lethal with α-synuclein. The ortholog category indicates yeast genes with human orthologs.

Strain Ortholog Function
1. ape2Δ Yes Aminopeptidase
2. arl3Δ Yes ARF small monomeric GTPase activity
3. aro1Δ No Arom pentafunctional enzyme
4. cog6Δ Yes Involved in vesicular transport to the Golgi
5. crh1Δ Yes Cell wall protein
6. cvt17Δ No Lipase
7. dpp1Δ Yes Diacylglycerol pyrophosphate phosphatase
8. fun26Δ Yes Nucleoside transporter
9. gip2Δ Yes Regulatory subunit for PP1 phosphatase
10. glo4Δ Yes Hydroxyacylglutathione hydrolase
11. gtt1Δ No Glutathione transferase
12. hbs1Δ Yes ∼To translation elongation factor EF-1alpha
13. hsp30Δ No Heat shock protein for pH homeostasis
14. ino4Δ No Transcript. factor (phospholipid syn. genes)
15. mad1Δ No Involved in spindle-assembly checkpoint
16. mal31Δ Yes Maltose transporter
17. mei4Δ No Required for meiotic recombination
18. met17Δ Yes O-acetylhomoserine (thiol)-lyase
19. met32Δ Yes Transcription factor
20. msb3Δ Yes RAB GTPase activator
21. nbp2Δ Yes Poss. involved in cytoskeletal organization
22. nit2Δ Yes Nitrilase
23. nup53Δ Yes Component of nuclear pore complex
24. opi3Δ Yes Phosphatidylethanolamine N-methyltransf.
25. pca1Δ Yes P-type copper-transporting ATPase
26. pex2Δ Yes Peroxisomal biogenesis protein
27. pex8Δ No Peroxisomal biogenesis protein
28. pho13Δ Yes 4-Nitrophenylphosphatase
29. pox1Δ Yes Acyl-CoA oxidase
30. ptk2Δ Yes Serine/threonine protein kinase
31. rpl41aΔ No Structural constituent of ribosome
32. rny1Δ Yes Endoribonuclease
33. sac2Δ Yes Involved in protein sorting in the late Golgi
34. sap4Δ No Serine/threonine phosphatase
35. sod2Δ Yes Manganese superoxide dismutase
36. stf1Δ No ATPase inhibitor
37. stp2Δ Yes Transcription factor
38. suv3Δ Yes Mitochondrial RNA helicase (DEAD box)
39. swr1Δ Yes Member of Snf2p DNA helicase family
40. thi7Δ No Thiamin transporter
41. tlg2Δ Yes Syntaxin homolog (t-SNARE)
42. thr1Δ No Homoserine kinase
43. tna1Δ Yes Nicotinamide mononucleotide permease
44. tsl1Δ No Alpha, alpha-trehalose-phosphate synthase
45. ubc8Δ Yes Ubiquitin-conjugating enzyme
46. vps24Δ Yes Sorts proteins in the prevacuolar endosome
47. vps28Δ Yes Required for traffic to vacuole
48. vps60Δ No Vacuolar protein sorting
49. war1Δ No Transcription factor
50. yat1Δ Yes Outer carnitine acetyltransferase, mitochondrial
51. ybr013cΔ No Unknown
52. ybr284wΔ Yes AMP deaminase
53. ybr300cΔ No Unknown
54. ycl042wΔ No Unknown
55. ycr026cΔ Yes Contains type I phosphodiesterase domain
56. ycr050cΔ No Unknown
57. ycr051wΔ Yes Contains ankyrin (Ank) repeats
58. ycr085wΔ No Unknown
59. ydl118wΔ No Possibly involved in meiotic nuclear division
60. ydr154cΔ No Unknown
61. ydr220cΔ No Unknown
62. yfr035cΔ No Unknown
63. ygl109wΔ No Unknown
64. ygl165cΔ No Unknown
65. ygl226wΔ No Unknown
66. ygl231cΔ Yes Unknown
67. ygl262wΔ No Unknown
68. ygr130cΔ Yes Unknown
69. ygr154cΔ No Unknown
70. ygr201cΔ Yes Translation elongation factor
71. ygr290wΔ No Unknown
72. yhr199cΔ No Unknown
73. yjl118wΔ No Unknown
74. yjl122wΔ No Unknown
75. yjl135wΔ No Unknown
76. yjr154wΔ No Unknown
77. ykl098wΔ No Unknown
78. ykl100cΔ Yes Unknown
79. ykr023wΔ Yes Unknown
80. ykr035cΔ No Unknown
81. ylr365wΔ No Unknown
82. ylr376cΔ No Possibly involved in DNA repair
83. ymr226cΔ Yes Oxidoreductase
84. yml089cΔ No Unknown
85. ymr289wΔ No Unknown
86. ypl136wΔ No Unknown

Of the HD53Q-sensitive mutants, 77% (40/52) corresponded to genes for which a function or genetic role has been determined experimentally or can be predicted (Table 1) (13). Thirty-five percent (14/40) of these genes clustered in the functionally related categories of response to stress, protein folding, and ubiquitin-dependent protein catabolism, based on annotations in the Yeast Proteome Database (Fig. 2) (14). The remaining genes were dispersed among numerous and diverse functional categories (Fig. 2). Fiftytwo percent (27/52) of the genes we identified are annotated as having human orthologs (Table 1), a value that is significantly higher than the percentage of genes in the yeast genome with mammalian orthologs (∼31%, P ≤ 1 × 10–10) (15).

Fig. 2.

Comparison of the relative percentage of genes in functional categories for the yeast gene deletion set (YGDS) and the huntingtin/α-synuclein synthetic lethal screens. Thirty-five percent (14/40) of genes that enhanced HD53Q toxicity clustered in the functionally related categories of response to stress, protein folding, and ubiquitin-dependent protein catabolism (*), whereas 32% (18/57) of genes that modified α-synuclein toxicity clustered in the functionally related categories of lipid metabolism and vesicle-mediated transport (#).

We next characterized the effects of HD20Q overexpression in yeast gene deletion strains sensitive to HD53Q. About 77% (40/52) of the deletion strains transformed with HD20Q displayed wild-type viability or exhibited a slight decrease in cell viability (table S1). Although 23% (12/52) of strains transformed with HD20Q displayed a noticeable loss of viability, in each case the phenotype was equivalent to or less severe than that observed with HD53Q (table S1). HD53Q-induced synthetic sickness or lethality observed in yeast gene deletion strains was partially rescued by overexpressing human orthologs (FKBP2, GSS, and DNAJA2) of several yeast genes identified in the screen (FPR2, GSH2, and HLJ1, respectively) (fig. S1). To determine whether a correlation exists between HD53Q aggregation and toxicity in the yeast gene deletion strains, filter-trap assays were conducted on protein lysates from 10 strains (fig. S2). No correlation between levels of aggregation and viability was observed in these strains (fig. S2).

Of the α-synuclein–sensitive mutants that were identified in the yeast screen, 66% (57/86) corresponded to genes for which a function or genetic role has been determined experimentally or can be predicted (Table 2). Thirty-two percent (18/57) of these genes clustered in the functionally related categories of lipid metabolism and vesicle-mediated transport (Fig. 2). As with the HD53Q screen, the remaining genes in the α-synuclein screen were distributed among numerous functional categories (Fig. 2), and a high percentage of the genes (50% or 43/86) have human orthologs (Table 2).

Evidence increasingly suggests that genes involved in response to stress, protein folding, and the ubiquitin-mediated protein catabolism play important roles in HD and the polyQ disorders (16, 17). Overexpression of the chaperones Hsp40 and/or Hsp70 in fruit fly and mouse models of polyQ disorders suppresses neurodegeneration, whereas mutation of genes involved in ubiquitin-mediated protein catabolism enhances neurodegeneration (1822). Our yeast screen identified two Hsp40 homologs (Apj1 and Hlj1) that when deleted enhance toxicity of HD53Q. These results indicate that in wild-type yeast cells Hsp40 chaperones are necessary to suppress polyQ toxicity. In addition to chaperones, eight genes (FPR2, GRE2, GSH2, HLR1, PRM5, SIP18, YHB1, and YJR107W) involved in various forms of cellular stress (osmotic, oxidative, and nitrosative) and three genes involved in ubiquitin-mediated protein catabolism (UBP13, YBR203W, and YKR017C) were identified as enhancers of HD53Q toxicity in yeast.

Several genes in the functional categories of response to stress and ubiquitin-mediated protein catabolism were also isolated in the α-synuclein screen (SOD2, GTT1, HSP30, TSL1, and UBC8) (Table 2). However, in contrast to the HD53Q screen, the levels of genes in these categories were not increased above background levels (Fig. 2). A recent study demonstrated that Hsp70 overexpression rescues α-synuclein–induced neurode-generation in a Drosophila model for PD (23). Although considerable genetic, cell biological, and biochemical evidence suggests that genes involved in response to stress, protein folding, and ubiquitin-mediated protein catabolism play important roles in the pathobiology of PD, the results from the α-synuclein yeast screen indicate that genes involved in lipid metabolism and vesicle-mediated transport may also be primary pathways that regulate toxicity of α-synuclein.

α-Synuclein localizes to nerve terminals and may be associated with synaptic vesicles, based on immunohistochemistry and ultrastructural analyses (24). α-Synuclein binds lipid membranes (2527) and can inhibit phospholipase D2 in vitro (28). α-Synuclein interacts with synphilin-1 (29), which has been proposed to function as an adaptor protein linking α-synuclein to proteins involved in vesicular transport. Although the precise function of α-synuclein is still not clear, this protein has been linked to learning, development, and plasticity (30) and most likely plays a role in synaptic vesicle recycling. Recent in vitro studies suggest that prefibrillar intermediates called protofibrils formed by α-synuclein can bind and permeabilize acidic phospholipid vesicles (31), which has been proposed to lead to defective sequestration of dopamine into vesicles and subsequent generation of reactive oxygen species in the cytoplasm that contribute to neuronal dysfunction and cell death (32). Taken together, these results are consistent with those from our α-synuclein genetic screen and with studies examining the biological and pathobiological effects of α-synuclein in yeast. α-Synuclein, and not a mutant huntingtin fragment, localized to membranes, caused the accumulation of lipid droplets and inhibited phospholipase D and vesicular trafficking (11). The results from the yeast screen are also consistent with recent expression-profiling studies in Drosophila that overexpress α-synuclein, which showed that lipid and membrane transport mRNAs were tightly associated with α-synuclein expression (33).

The results from yeast screens clearly indicate that toxicity mediated by α-synuclein and a mutant huntingtin fragment is regulated by nonoverlapping sets of conserved genes and pathways. The major functional categories enriched in the α-synuclein genetic screen did not overlap with any of the major categories observed in the HD53Q screen, and only 1 out of 138 genes that enhanced toxicity was found in common to both screens (STP2). Collectively, these results suggest that distinct pathogenic mechanisms may underlie HD and PD.

Supporting Online Material

www.sciencemag.org/cgi/content/full/302/5651/1769/DC1

Materials and Methods

Figs. S1 and S2

Table S1

References

References and Notes

View Abstract

Stay Connected to Science

Navigate This Article