| Human Molecular Genetics | Pages |
Spinobulbar muscular atrophy: polyglutamine-expanded androgen receptor is proteolytically resistant in vitro and processed abnormally in transfected cells
Introduction
Results
Conformational analysis of radiolabelled normal (20-Gln) versus expanded (44-Gln) AR
Transfection studies
Discussion
Materials And Methods
In vitro protein synthesis
Proteolytic digestion
Electrophoretic analysis of protein
Western analysis
Transfection and cell toxicity studies
Acknowledgements
References
Spinobulbar muscular atrophy: polyglutamine-expanded androgen receptor is proteolytically resistant in vitro and processed abnormally in transfected cells
The neuronotoxicity of genes with expanded CAG repeats is most likely mediated by their respective polyglutamine (Gln)-expanded gene products. Gln- expanded portions of these products may be sufficient, or necessary, for pathogenesis. We tested whether a Gln-expanded human androgen receptor (AR) is structurally altered, so that it allows for the proteolytic generation of a potentially pathogenic portion that may be resistant to further degradation. We found, in vitro, that a Gln-expanded AR is more proteolytically resistant than normal, and that it yields a distinct set of Gln-expanded fragments even after extended proteolysis in the presence of 2 M urea. Furthermore, COS cells transfected with CAG-expanded AR cDNA generate an aberrant, nuclear-associated 75 kDa derivative containing the Gln-expanded tract. They are also twice as likely to die by 24 h apoptotically than those transfected with normal AR cDNA. Our data support the notion that an unconventional derivative of the Gln- expanded AR is a component of the proximate motor neuronopathic agent in spinobulbar muscular atrophy. They also focus attention on two ways in which neuronotoxic derivatives may originate from various Gln-expanded proteins: (i) generation of an unusual derivative that is pathogenic de novo; and (ii) the toxic accumulation of a normal derivative because of an inability to dispose of it.
INTRODUCTION
Eight hereditary neuronopathies, distinguished by progressive death of a specific group of neurons, are caused by expansion of translated CAG trinucleotide repeats, resulting in abnormally long polyglutamine (polyGln) stretches within their respective mutant gene products (1,2). They are: Huntington's disease (HD), spinocerebellar ataxia type 1 (SCA1), spinocerebellar ataxia type 2 (SCA2), Machado-Joseph disease (SCA3), spinocerebellar ataxia type 6 (SCA6), spinocerebellar ataxia type 7 (SCA7), dentatorubral-pallidoluysian atrophy (DRPLA) and spinobulbar muscular atrophy (SBMA). Only the normal gene product at the X-linked locus for SBMA (3,4), the androgen receptor (AR), has been characterized substantially. Over 140 different `loss-of-function' mutations of the AR gene cause various degrees of androgen resistance (5); yet, none is neuronopathic. Therefore, an AR protein with 40 or more residues in its expanded polyGln tract must become neuronotoxic by a gain of function.
The pathogenesis of the motor neuronopathy in SBMA is unknown, but it is likely to share essential features with that of the seven other disorders in the group. One hypothesis is that Gln-expanded proteins are excessively prone to bind to certain target proteins by polar zippering (6,7) or covalent cross-linking (8). The resulting complexes might be selectively neuronotoxic in themselves, or because their formation reduces the level of the free target protein(s) below a threshold needed for the survival of a specific set of neurons. Several huntingtin-associated proteins, HAP-1 (9), calmodulin (10) and HIP-1 (11), do bind more strongly to the mutant version of huntingtin. This is not true, however, for the binding of ataxin-1 (SCA1) or of AR to glyceraldehyde-3-phosphate dehydrogenase (GAPDH) (12), or for the interaction of huntingtin with a ubiquitin-conjugating enzyme (13). Another hypothesis (14) envisages that `... abnormal post-translational cleavage products ... consisting of abnormally large polyglutamine moieties ...' may render `specific regions of the nervous system vulnerable ...'. Accordingly, certain polyGln-containing huntingtin fragments bind to GAPDH more than does full-length huntingtin (15). Likewise, exon 1 of the HD gene is sufficient to produce an HD-like phenotype in transgenic mice (16), CAG-expanded fragments of several genes have been used to generate effective baits for interacting proteins (11,12), Gln-expanded SCA3 fragments generate insoluble aggregates in vitro (17), and proteolytic susceptibility of huntingtin to apopain, a proapoptotic cysteine protease, increases after substantial expansion of its polyGln tract (18). Finally, Ikeda et al. reported that only a CAG-expanded portion of the Machado-Joseph gene, not the full disease gene, induced death of Purkinje cells in transgenic mice. The CAG-expanded portion was also able to cause apoptosis of COS cells in culture. These observations were postulated to reflect `region-specific proteolysis or processing' that might `expose the expanded polyglutamines', thereby maximizing their pathogenicity (19).
The multiplicity of evidence incriminating Gln-expanded fragments as pathogenetic agents of (CAG)n-expanded neurodegenerative disease stimulated us to ask several questions: does the Gln-expanded AR in vitro have a conformational alteration (gain of structure) that facilitates the creation of potentially pathogenic derivatives?; does the CAG-expanded AR cDNA yield a potentially pathogenic derivative in transfection studies in whole cells?; is this derivative(s) associated with increased cell death?
RESULTS
Conformational analysis of radiolabelled normal (20-Gln) versus expanded (44-Gln) AR
[35S]Methionine (Met)-labelled normal AR and Gln-expanded AR were subjected to partial tryptic digestion (20) for 10 min and then analysed by SDS-PAGE (Fig. 1A). Expectedly, the normal and expanded parental ARs had slightly different mobilities. By densitometry, >80% of the normal full-length protein was digested at 10 min compared with 50% of the Gln-expanded full-length AR. The relative resistance to trypsin of the Gln-expanded AR signifies that it is conformationally different from the normal AR. Nonetheless, most of the fragments in both digests migrated identically. This indicates that a trypsin cleavage site C-terminal to the normal or expanded tract is highly exposed and that an N-terminal portion containing different polyGln tract sizes is cleaved from both the normal and expanded parental ARs soon after the start of tryptic digestion. In contrast, mutant AR and its normal counterpart yield several distinct tryptic fragments: notably the 65, 36 and 20 kDa products in the Gln-expanded AR digest, and those of 55 and 28 kDa in the normal AR digest. The migration differences of the possible pairs of cognate fragments far exceed those expected for the 3.1 kDa contributed by the 24 extra residues in the Gln-expanded tract; this further supports conformational differences between normal and expanded AR.
Figure
To semi-quantitate the relative trypsin resistance of certain Gln-expanded AR fragments, we increased the amount oftrypsin by 100% and extended digestion of [14C]glutamine labelled normal and expanded AR to 1 h in the presence or absence of 2 M urea (Fig. 1B). These harsher conditions highlight the same disparate fragments observed earlier (Fig. 1A) to distinguish the normal from the Gln-expanded AR. The prominence of the 65, 36 and 20 kDa [14C]Gln-labelled fragments in the Gln-expanded, trypsin-treated lanes of Figure 1B is particularly striking. Figure 1C is a Western analysis performed with mAb1C2 (21,22) (a monoclonal antibody that reacts preferentially with long polyGln tracts) after blotting the same [14C]Gln-labelled gel described in Figure 1B. It illustrates that four of the fragments that resist aggressive denaturant proteolysis of the Gln-expanded AR themselves contain the Gln-expanded tract. Appropriately, these four fragments also yield the most intense radioautographic signals in Figure 1B.
Transfection studies
COS cells were transfected with normal or Gln-expanded AR to determine whether they produced unique Gln-expanded derivatives of AR, in association with increased cell death. Western blots using mAb1C2 and F39.4.1 (23) antibodies (epitope amino acids 301-320) were performed on whole-cell extracts obtained 1-4 days later (Fig. 2A and B). Slightly less stringent conditions allowed mAb1C2 to detect both the parental normal AR and the Gln-expanded parent. mAb1C2 clearly exposed a novel band, of ~75 kDa, in whole-cell extracts of COS cells transfected with the (CAG)n-expanded AR cDNA (Fig. 2A). No counterpart was obtained from COS cells transfected with normal AR cDNA, even after prolonged exposure to the chemiluminescence detection system. Likewise, no counterpart was observed in COS cell extracts blotted with F39.4.1 (Fig. 2B). Thus, the ~75 kDa derivative is uniquely associated with expression of the Gln-expanded AR and it contains a polyGln tract, but apparently not the epitope for mAb F39.4.1 (amino acids 301-320).
Figure
A conventional, linear 75 kDa fragment of the AR that includes the polyGln tract should also contain the F39.4.1 epitope. The apparent absence of this epitope implies unconventional proteolysis or some form of misprocessing of the AR. We attempted to assess whether the 75 kDa fragment represents a post-translational event, rather than abnormal behaviour at the DNA or RNA level, by repeating the COS cell transfection experiments using two kinds of AR cDNA: one with a natural, expanded sequence of (CAG)43 trinucleotides followed by a single 3[prime] CAA; another with a random mixture of CAG and CAA trinucleotides yielding an expanded (Gln)41 tract. These experiments also demonstrated the 75 kDa fragment in the cells with the mixed CAG/CAA construct coding for a 41-Gln tract (Fig. 2C). To determine whether the 75 kDa product was associated with the nuclei of transfected COS cells, we prepared extracts of their nuclei (24), and subjected them to western analysis using mAb1C2 and a monoclonal antibody to proliferating cell nuclear antigen (PCNA). Figure 3 shows that the 75 kDa product is nuclear associated.
Figure
COS cells transfected with normal or (CAG)n-expanded AR cDNA were stained with ethidium monoazide bromide and examined by epifluorescence microscopy to assess cell toxicity and the rate of apoptosis. Ethidium monoazide bromide covalently only labels DNA of dead cells and has been used widely to assess cell death (25). Table 1 demonstrates a 2-fold increase in death with distinct apoptotic morphology in cells transfected with Gln-expanded versus normal AR. This effect was observed in the presence or absence of 3 nM mibolerone, a synthetic androgen. Table 1
Exp no.
-Mibolerone
+Mibolerone
Normal AR
Expanded AR
Normal AR
Expanded AR
1
5.4
9.0
5.5
11.0
2
4.2
12.4
8.6
13.2
3
8.1
12.7
8.3
16.8
4
3.6
9.8
5.3
9.6
Average
5.3
11.0
6.9
12.7
DISCUSSION
We have used [14C]Gln or [35S]Met labelling during in vitro synthesis, partial progressive trypsinolysis with and without denaturant, and monoclonal antibodies to two different portions of the N-terminal region of the AR to study the conformational properties of normal and Gln-expanded versions of the AR.
The most striking mobility differences we observed arethe disparate sets of trypsin- and trypsin-2 M urea-resistant fragments derived from the [14C]Gln-labelled mutant compared with the normal AR (Fig. 1B and C). As revealed by their immunoreactivity with mAb1C2, four of the residual fragments in the mutant digest have an expanded polyGln tract that contains 24 more Gln residues than the normal AR. These four fragments do not appear to be the counterparts of otherwise homologous fragments, each with a normal size Gln tract, that are derived from the normal full-length AR. They also display pronounced resistance to proteolysis. Our data suggest that conformational differences are induced by Gln tract expansion, are in the vicinity of the tract itself and can create proteolytically resistant fragments.
Our DNA transfection studies clearly demonstrate a unique ~75 kDa Gln-expanded polypeptide that is generated, accumulated, or both only by cells transfected with the expanded AR cDNAs. We presume that this molecule is derived, at least in part, from the AR cDNA. The fact that it does not appear to contain the F39.4.1 epitope must reflect some form of misprocessing, whether transcriptionally, translationally or post-translationally (Fig. 4). Protein splicing or denaturation-resistant intramolecular post-translational modification of the F39.4.1 epitope are possible. The result of our mixed expanded (CAG/CAA)41 transfection experiments suggests a post-translational event. One particularly provocative possibility is that a proteolytic portion of the Gln-expanded AR, lacking the epitope for F39.4.1, combines with another peptide to form a covalent isodipeptide of ~75 kDa by a transglutaminase-catalysed reaction, as recently described by Kahlem et al. (26).
Figure
In the spongiform encephalopathies (27-29) and Alzheimer's disease (30), presumably neuronotoxic peptide fragments are resistant to proteolysis and denaturation. Likewise, the novel 75 kDa Gln-expanded polypeptide produced in transfected COS cells could represent the accumulation of a conformationally altered, protease-resistant fragment of the AR that lacks immunoreactivity to F39.4.1 because it also resists denaturation. The presence of the 75 kDa fragment in nuclear extracts is in keeping with the recent observation of inclusion bodies containing Gln-expanded protein in brain nuclei of the HD transgenic mouse and SCA3 patients (17,31,32). Cha and Dure (14) postulated that `abnormal post-translational cleavage products' (polyglutamine-expanded) might be neuronotoxic particularly if they accumulate, in part because of reduced susceptibility to endopeptidase. Ikeda et al. (19) predicted the existence of such products to account for the fact that a CAG-expanded portion of the SCA3 gene is cytotoxic, while a full-length version is not. We found that full-length, CAG-expanded AR generates a distinctive set of fragments during denaturant proteolysis in vitro, and an aberrant ~75 kDa derivative that correlates with increased apoptosis in transfected COS cells. The lack of an androgen effect on cell death rate is compatible with the concept that a fragment of the AR, unable to bind androgen, is a prerequisite in the pathogenesis of the neuronopathy in SBMA. Our observations link the postulate and the prediction of previous investigators.
MATERIALS AND METHODS
A coupled rabbit reticulocyte lysate-based transcription-translation system kit was purchased from Promega; [35S]Met and [14C]Gln from Amersham; trypsin from Sigma and urea from Fisher Scientific. The hAR cDNA with an expanded (CAG)44 tract was constructed in the pcDNA3 plasmid, and was a gift from Dr K. Fischbeck, University of Pennsylvania, PA. The monoclonal antibody 1C2 that recognizes proteins with an expanded polyGln tract was a gift from Y. Trottier and J.-L. Mandel, Université Louis Pasteur, Strasbourg. The monoclonal antibody F39.4.1 that recognizes an epitope near the middle of the AR N-terminal domain was a gift from A. O. Brinkmann, Erasmus University, Rotterdam. The monoclonal antibody that recognizes the PCNA (36 kDa) was obtained from Santa Cruz Biotechnology.
In vitro protein synthesis
A portion of the hAR cDNA encoding a 44-Gln tract was replaced by one encoding a 20-Gln tract using the flanking NheI and BfrI restriction sites. The pcDNA3 plasmid (Invitrogen Corporation) contains T7 and Sp6 promoters for in vitro transcription. Transcription-translation was done at 30°C for 2 h in a total volume of 50 µl using 1 µg of plasmid DNA, and amino acid mixtures containing 200 fM l-[35S]Met (1000 Ci/mmol) alone, or 0.34 nM l-[14C]Gln (290 Ci/mmol) with 1 mM radioinert Gln.
Proteolytic digestion
Radiolabelled AR with 20- or 44-Gln was partially digested by adding 40 ng of trypsin in 1 µl of water to 10 µl of in vitro product at room temperature for 0-10 min. The reaction was terminated by adding 1 vol. 2× SDS/sample buffer (3% SDS; 10% glycerol; 5% 2-mercaptoethanol; 0.062 M Tris-HCl, pH 6.8; and 0.01% bromophenol blue). For more vigorous digestion, 15 µl of in vitro product were exposed to 80 ng of trypsin in the presence or absence of 2 M urea for 60 min at room temperature. These reactions were terminated by boiling for 5 min.
Electrophoretic analysis of protein
Proteins were fractionated on 8% SDS-PAGE using a Bio-Rad minigel apparatus or standard 20 × 20 cm gels according to Okajima's method (33), followed by gel drying and autoradiography.
Western analysis
Proteins were electroblotted on nitrocellulose membrane, prepared for incubation with monoclonal antibodies F39.4.1 (23), 1C2 (21,22) and PCNA and developed using the ECL western blotting chemiluminescence detection system (Amersham) as described previously (33).
Transfection and cell toxicity studies
To replace the conventional tract of the (CAG)n trinucleotide repeat in exon 1 by the mixed (CAG/CAA)n trinucleotide repeat, antisense primers (PRMixQ-198) flanked by 24 nucleotides of the original sequence were synthesized. In the primary reaction, two fragments flanking and overlapping with MixQ-198 were amplified. The downstream fragment was amplified by using a sense complementary primer (Primer 20EA) for the flanking 5[prime] end of the MixQ-198 together with a downstream primer (Primer 1.2B°). The upstream fragment was amplified by using an overlapping antisense primer (Primer 5[prime]B flank) for the flanking 3[prime] end of MixQ-198 and an upstream primer designed to contain a NheI restriction siste, (Primer-NheI 1.1A°). The secondary amplification was accomplished by combining the two fragments generated by the primary reaction and the bridging primer MixQ-198. The reaction was done in two stages, primer extension and amplification. In the first stage, the mixture contained the flanking fragments, MixQ-198 and the remaining PCR cocktail. Extension was for two cycles: each consisted of 3 min denaturation at 98°C; 2 min annealing at 57°C (ramping time of 2 min) and 10 min extension at 72°C. After the end of the second cycle, the external flanking primers (NheI 1.1A° and 1.2B°) were added and the reaction was continued for 30 cycles. The final product was gel purified, digested with NheI-AflII and the double-digested fragment generated was ligated into the pcDNA3 AR expression vector. All primer sequences are available upon request.
COS cells were transfected with pcDNA3 plasmids expressing AR with 20- or 44-Gln, and subsequent western blot analysis was performed as described in Beitel et al. (34). Nuclear protein was isolated using Dignam's methodology (24).
Cell toxicity was assessed 24 h after transfection as follows. Monolayers grown in 8-unit multiwell chambers (Lab Tek) were stained with 10 µg/ml ethidium monoazide bromide (EMA; Molecular Probes) in phosphate-buffered saline (PBS) for 10 min at room temperature under UV light exposure. The cells subsequently were washed twice in PBS and fixed in 4% paraformaldehyde for 20 min at room temperature. Unlike standard ethidium bromide, EMA covalently binds to the DNA of dead cells thereby preventing dye leakage post-fixation (25). Total cell numbers were determined in 25× fields under phase microscopy. Dead cells were identified by the presence of bright red nuclear staining under epifluorescence microscopy (Leitz Diaplan Photomicroscope) using a rhodamine filter. The appearance of shrunken, condensed nuclei and the presence of small round apoptotic bodies constituted morphological evidence of apoptosis. Four sister wells were evaluated for each experimental group. A minimum of 140 cells were assessed per well. The extent of cell death was expressed as the mean per cent of EMA-positive cells in each monolayer. Statistical analysis was performed using the Student's unpaired t-test (two-tailed), withP < 0.05 indicating significance.
ACKNOWLEDGEMENTS
We thank L. Beitel for her important contributions and R. Rosenzweig and K. Berckmans for expert preparation of the manuscript. A.A.R.A. wishes to acknowledge the financial support of the University of Bahrain. This work was funded by the Huntington Disease Society of America, the Amyotrophic Lateral Sclerosis Association (USA), and grant MT-13209 from the Medical Research Council of Canada.
REFERENCES
*To whom correspondence should be addressed. Tel: +1 514 340 8222; Fax: +1 514 340 7502; Email: rrosenzw@ldi.jgh.mcgill.ca
This page is run by Oxford University Press, Great Clarendon Street, Oxford OX2 6DP, as part of the OUP Journals
Comments and feedback: www-admin{at}oup.co.uk
Last modification: 14 Feb 1998
Copyright© Oxford University Press, 1998.
CiteULike 
Connotea 
Del.icio.us What's this?
This article has been cited by other articles:
|
|
 |
|
  P. Davies, K. Watt, S. M Kelly, C. Clark, N. C Price, and I. J McEwan Consequences of poly-glutamine repeat length for the conformation and folding of the androgen receptor amino-terminal domain J. Mol. Endocrinol., November 1, 2008; 41(5): 301 - 314. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 R. P. Pelley, K. Chinnakannu, S. Murthy, F. M. Strickland, M. Menon, Q. P. Dou, E. R. Barrack, and G. P.-V. Reddy Calmodulin-Androgen Receptor (AR) Interaction: Calcium-Dependent, Calpain-Mediated Breakdown of AR in LNCaP Prostate Cancer Cells Cancer Res., December 15, 2006; 66(24): 11754 - 11762. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 J. Taylor, S. K. Grote, J. Xia, M. Vandelft, J. Graczyk, L. M. Ellerby, A. R. La Spada, and R. Truant Ataxin-7 Can Export from the Nucleus via a Conserved Exportin-dependent Signal J. Biol. Chem., February 3, 2006; 281(5): 2730 - 2739. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 D.M. Avila, D.R. Allman, J.-M. Gallo, and M.J. McPhaul Androgen Receptors Containing Expanded Polyglutamine Tracts Exhibit Progressive Toxicity when Stably Expressed in the Neuroblastoma Cell Line, SH-SY 5Y Experimental Biology and Medicine, September 1, 2003; 228(8): 982 - 990. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 L. M. Mandrusiak, L. K. Beitel, X. Wang, T. C. Scanlon, E. Chevalier-Larsen, D. E. Merry, and M. A. Trifiro Transglutaminase potentiates ligand-dependent proteasome dysfunction induced by polyglutamine-expanded androgen receptor Hum. Mol. Genet., July 1, 2003; 12(13): 1497 - 1506. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
J. L. Walcott and D. E. Merry Ligand Promotes Intranuclear Inclusions in a Novel Cell Model of Spinal and Bulbar Muscular Atrophy J. Biol. Chem., December 20, 2002; 277(52): 50855 - 50859. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
H.Y. E. Chan, J. M. Warrick, I. Andriola, D. Merry, and N. M. Bonini Genetic modulation of polyglutamine toxicity by protein conjugation pathways in Drosophila Hum. Mol. Genet., November 1, 2002; 11(23): 2895 - 2904. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
N. J. Caplen, J. P. Taylor, V. S. Statham, F. Tanaka, A. Fire, and R. A. Morgan Rescue of polyglutamine-mediated cytotoxicity by double-stranded RNA-mediated RNA interference Hum. Mol. Genet., January 1, 2002; 11(2): 175 - 184. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
W. J. Welch and M. I. Diamond Glucocorticoid modulation of androgen receptor nuclear aggregation and cellular toxicity is associated with distinct forms of soluble expanded polyglutamine protein Hum. Mol. Genet., December 1, 2001; 10(26): 3063 - 3074. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
H.N. Lim, H. Chen, S. McBride, A.M. Dunning, R.M. Nixon, I.A. Hughes, and J.R. Hawkins Longer polyglutamine tracts in the androgen receptor are associated with moderate to severe undermasculinized genitalia in XY males Hum. Mol. Genet., March 22, 2000; 9(5): 829 - 834. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
S. Simeoni, M. A. Mancini, D. L. Stenoien, M. Marcelli, N. L. Weigel, M. Zanisi, L. Martini, and A. Poletti Motoneuronal cell death is not correlated with aggregate formation of androgen receptors containing an elongated polyglutamine tract Hum. Mol. Genet., January 1, 2000; 9(1): 133 - 144. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 K. L. Moulder, O. Onodera, J. R. Burke, W. J. Strittmatter, and E. M. Johnson Jr Generation of Neuronal Intranuclear Inclusions by Polyglutamine-GFP: Analysis of Inclusion Clearance and Toxicity as a Function of Polyglutamine Length J. Neurosci., January 15, 1999; 19(2): 705 - 715. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 J. Hardy and K. Gwinn-Hardy Genetic Classification of Primary Neurodegenerative Disease Science, November 6, 1998; 282(5391): 1075 - 1079. [Abstract] [Full Text]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||