Identification of Novel Alternative SplicingEvents in the Huntingtin Gene andAssessment of the FunctionalConsequences Using Structural ProteinHomology Modelling
1, Matthew Mort 2, Lyn Ell
iston1, Rhian M. Thomas1,Simon P. Brooks3, Stephen B. Dunnett 3 and Lesley Jones1
1 - Institute of Psychological Medicine and Clinical Neuroscience, MRC Centre for Neuropsychiatric Genetics and Genomics,Hadyn Ellis Building, Cardiff University, Cardiff CF24 4HQ, UK2 - Institute of Medical Genetics, Cardiff University, Heath Park, Cardiff CF14 4XN, UK3 - Brain Repair Group, School of Biosciences, Cardiff University, Museum Avenue, Cardiff CF10 3AX, UK
Correspondence to Lesley Jones: [email protected]://dx.doi.org/10.1016/j.jmb.2013.12.028Edited by S. A. Teichmann
Abstract
Huntington's disease (HD) is an inherited progressive neurodegenerative disorder caused by a pathologicalCAG trinucleotide repeat expansion in the large multi-exon gene, huntingtin (HTT). Although multiplepathogenic mechanisms have been proposed for HD, there is increasing interest in the RNA processing of theHTT gene. In mammals, most multi-exon genes are alternatively spliced; however, few alternative transcriptshave been described for HTT. Given the numerous protein bands detected in mouse and human brain tissueby Western blotting using anti-huntingtin antibodies, we examined whether alternative splicing of HTT mayaccount for some of these fragments. Using RT-PCR in mouse brain, we detected two novel splice variants ofHtt that lacked the 111-bp exon 29 (HttΔex29) or retained a 57-bp portion of intron 28 (Htt+57in28) via use of acryptic splice site. The alternative transcripts were present in wild-type and homozygous Hdh(Q150/Q150)mouse brain at all ages and in all brain regions and peripheral tissues studied. Differential splicing of HttΔex29was found in the cerebellum of Hdh(Q150/Q150) mice with a significant reduction in transcript levels in mutantanimals. In human brain, we detected similar splice variants lacking exons 28 and 29. The ability ofalternatively spliced transcripts to encode different protein isoforms with individual functions in the cell,combined with the known role of splicing in disease, renders these novel transcripts of interest in the context ofHD pathogenesis.
Introduction
Huntington's disease (HD) is an autosomal dom-inant progressive neurodegenerative disorder withonset in midlife that proceeds inexorably to death [1].The triad of motor, cognitive and behaviouralsymptoms can be partially managed; however, nodisease-modifying therapy is available. Multiplemechanisms have been implicated in the pathwayfrom the causative CAG expansion in HTT todisease [2], and it remains unclear which of theseare critical to disease aetiology. Neither has thenormal function of the huntingtin (HTT) protein beenfully elucidated. There is evidence to suggest thatHTT exists in shorter forms in addition to its
full-length state [3–6]. Cleavage of HTT can accountfor some of these fragments [6]; however, theidentity of others remains unknown.We hypothesised that alternative splicing of the
huntingtin (HTT) gene might account for some ofthese smaller products. Alternative splicing is theprocess whereby DNA encoding a single gene canbe processed into more than one mRNA transcriptthrough the inclusion or exclusion of specific exons[7]. In humans, most multi-exon genes are alterna-tively spliced [8] although little has been reportedabout HTT splicing [9–11]. The process of alterna-tive splicing has been shown to be associated withdisease; for instance, transcript variants of MAPTare implicated in disorders such as progressive
1429Alternative Splicing Events in the Huntingtin RNA
supranuclear palsy, corticobasal degeneration andParkinson's disease [12]. In HD, an HTT exon 1transcript produced only from the expanded CAGallele has been detected and it is postulated that thistranscript, which translates to give an exon 1 protein,is the pathogenic moiety in HD [11].HTT comprises 67 exons giving two major
transcripts of either ~10 kb or 13 kb dependingupon the 3′ untranslated region (UTR) [9]. Murine Httexons range from 48 bp to 341 bp in size except forexons 1 and 67, which are 366 bp and 3903 bp,respectively, but contain either 5′ UTR or 3′ UTR(Table S1). Therefore, any splicing event resulting ina noticeable change in protein size could involveeither multiple exons being skipped or the introduc-tion of a frame shift and premature termination codon(PTC). Here, we describe two novel transcripts inaddition to the standard Htt mRNA transcript inmouse. One lacks exon 29 (HttΔex29) and the otherretains 57 bp of intron 28 (Htt+57in28) due to thepresence of a cryptic splice site. We have shown bystructural protein homology modelling that thisregion of HTT contains a predicted binding sitewith implications for HTT's function in nucleocyto-plasmic protein transport and nuclear RNA export.Reduction of HTT expression has pathogenic effectsin mice and cells [13–20], and it is not clear whichfunctions are critical for normal development andneurogenesis.
Results
Alternative transcripts in the Hdh(Q150) mousemodel of HD
Htt was amplified in 28 overlapping segmentsspanning all Htt exons (P1–P28, Table S2a) using5-, 12- and 24-month reverse-transcribedHdh(Q150) mouse brain RNA. One segment span-ning exons 28–32 yielded three bands on gelelectrophoresis (Fig. 1b). The bands were band-stabbed, reamplified and sequenced to confirm themas the constitutive Htt transcript (band A; Fig. 1a andb) and two novel splice variants of Htt, one lackingexon 29 (HttΔex29) and the other retaining 57 bp ofthe 3′ end of intron 28 (Htt+57in28) (bands B and C,respectively; Fig. 1a and b). Bioinformatic analysis ofthe region upstream of the natural exon 29 acceptorsplice site revealed that the cryptic splice site used inthe Htt+57in28 transcript had an NNSplice score of0.95, stronger than the natural splice site used inthe wild-type constitutively spliced Htt transcript,which had an NNSplice score of 0.55 (Fig. 1c). Thepresence of the three transcripts was demonstratedin all subdissected brain regions studied (Fig. 1b) inaddition to all peripheral tissues studied of bothmutant and wild-type Hdh (Q150) mice (Fig. 1b). The
three bands corresponding to the three different Htttranscripts were also detected in the Hdh(Q92)mouse model of HD across all brain regions studied(data not shown).
Determining expression levels of Htt alternativetranscripts in the Hdh(Q150) knock-in HDmouse model
Quantitative PCR (qPCR) was performed toexplore differential splicing of the Htt transcripts inthe motor cortex, caudate nucleus and cerebellum of5-month-old Hdh(Q150) HD mice (WT, n = 5; MT,n = 4). Primers specific for the constitutive transcript(P30; Table S2a), HttΔex29 (P29; Table S2a) andHtt+57in28 (P31; Table S2a) were used to selectivelyamplify these splice variants for qPCR. No signifi-cant difference was found in expression levels of theconstitutive transcript (Fig. 2a) orHtt+57in28 (Fig. 2c)in any brain region studied, when compared be-tween Hdh(Q150) and wild-type mice. By contrast,the HttΔex29 splice variant was expressed atsignificantly lower levels in the cerebellum ofHdh(Q150) mice (Fig. 2b) (p = 0.0118).
Three-dimensional homology modelling ofHTTΔex29
The Htt transcript lacking exon 29 (HttΔex29)maintains the reading frame on translation and leadsto a protein lacking the 37 amino acids encoded byexon 29. Investigation of evolutionary sequenceconservation using orthologous proteins (Fig. 3)showed that HTT exon 29 is highly conserved with81% of residues showing 100% conservation across11 different species, a significantly higher level ofconserved residues (Fisher's exact test, p = 0.0025)compared with the entire HTT protein where 57% ofresidues are conserved. The evolutionary conser-vation of exon 29 indicates a potentially functionalrole for this region of the protein. To assess thepotential functional consequences of HTTΔex29, weconstructed a three-dimensional (3D) structure of theconstitutive wild-type and HTTΔex29 proteins usingprotein homology modelling encompassing residues503–1999 of HTT. The I-TASSER server is restrictedto protein sequences less than 1500 amino acidresidues to reduce the time taken to generate theprotein model and to increase the quality of thequery-template alignment that will result in a high-er-quality 3D model [21]. In this study, the functionalrole of exon 29 was assessed by this method; thus, awindow of amino acids 503–1999 was selected inorder to centre on exon 29, which is encoded byamino acids 1233–1269 in murine HTT. The top fivestructural analogs found in the Protein Data Bank(PDB) are shown in Table 1. Predicted GeneOntology (GO) terms are listed in Table 2, showingpredicted functional roles for this region including the
Fig. 1. Alternative splicing in mouse Htt. (a) Schematic of the Htt gene (mouse), showing exons 27–32 (boxed regions).Exons are not drawn to scale. Transcripts B and C denote the novel alternative splice variants identified in this study.Sequencing confirmed the identity of these products. (b) Amplified products using primer sets P13, P29 generating theamplicon specifically for HttΔex29 and P31 generating the amplicon specifically for Htt+57in28. TheHttΔex29 (band B) andHtt+57in28 (band C) were detected in brain from wild-type and Hdh(Q150) mice at 5, 12 and 24 months of age. They werepresent in all brain regions (CN, caudate nucleus; MC, motor cortex; CB, cerebellum; HC, hippocampus) of 8-month-oldwild-type and Hdh(Q150) mouse brain studied. They were also detected in all peripheral tissues studied including thekidney, muscle, lung, heart and liver of both Hdh(Q150) and wild-type mice at 13 months of age. (c) The genomic DNAsequence of the Htt gene. Exon 29 is highlighted in grey. The first coding base (+1) of the predicted acceptor splice site isshown; in parentheses, the predicted splice site score (between 0 and 1) is shown; a score ≥0.4 is predicted to be anacceptor splice site. The novel alternatively spliced transcript that utilises the upstream cryptic splice site, leading to theretention of 57 bp of intronic sequence, is highlighted in yellow. The introduced PTC is boxed in red.
1430 Alternative Splicing Events in the Huntingtin RNA
binding and transport of proteins and gene regula-tion. The structures for constitutive HTT are shown inFig. 4 and Supplementary Fig. 1. The modelledregion is largely made of α-helices joined bycoiled regions, highlighted in Supplementary Fig. 1.
α-Helical regions commonly have a functional role inDNA binding or spanning a membrane [22]. In theconstitutive HTT protein, residues 1233 to 1269 areencoded by exon 29 (shown in red in Fig. 4) andexon 29 is an α-helical region predicted to be directly
Fig. 2. qPCR of alternative transcripts of murine Htt. A box plot showing the distribution of delta CT (DCT) values forqPCR of constitutive Htt (a), the HttΔex29 transcript (b) and the Htt+57in28 transcript (c) in the caudate nucleus, motorcortex and cerebellum of 5-month-old Hdh(Q150) mice. There was a significantly lower expression of the HttΔex29transcript in the cerebellum of Hdh(Q150/Q150) mice compared to wild-type animals (b) (*t test and statistical significancewere determined using a Bonferroni-corrected threshold of 0.05). No significant difference was detected between HD andwild-type mice for either of the other twoHtt transcripts (a and c). Themedian DCT value (horizontal line) and outliers (opencircles) are depicted.
1431Alternative Splicing Events in the Huntingtin RNA
involved in protein binding. A predicted proteinbinding site that involved 11 contact residuesshown in green in Fig. 5 (p.Q1121, p.H1125,
Fig. 3. Multiple protein sequence alignment of HTT exon 29for HTT exon 29 encompassing residues 1233 to 1269 (a). Anthe alignment, with ~81% of residues in the HTT exon 29 fallinbetween groups of strongly similar properties, whilst a period (properties. The residues in the MSA are coloured to indicatehydrophobic amino acids, blue indicates acidic residues, mageamine residues. (b) shows the same MSA for HTT exon 29protein–protein interaction sites (PPI), identified in Fig. 4.
p.M1179, p.G1246, p.R1249, p.D1253, p.S1256,p.E1260, p.M1288, p.V1291 and p.Q1295) wasfound. Of the 11 contact residues, 5 are located in
. Multiple protein sequence alignment (MSA) of 11 speciesasterisk (*) indicates positions that are 100% conserved ing into this category. (a) A colon (:) indicates conservation.) indicates conservation between groups of weakly similartheir physicochemical properties. Red indicates small andnta indicates basic and green denotes hydroxyl/sulfhydryl/as a sequence logo and is annotated with the predicted
Table 1. Ranking of the top five structural analogs in the PDB for the 3D modelled region of mouse wild-type HTT protein(residues 503–1999, encompassing 1497 residues in total)
PDB identifier Description TM-score
1u6g Ring-box 1, E3 ubiquitin protein ligase (RBX1; human), which encodes a HEAT repeat protein 0.7123ea5 GTP-binding nuclear protein Ran (RAN; human) 0.4071wa5 GTP-binding nuclear protein Ran (Ran; dog) 0.4033icq GTP-binding nuclear protein GSP1/CNR1 (GSP1; yeast) 0.3872qna Karyopherin (importin) beta 1 (KPNB1; human) 0.379
Ranking is based on the TM-score of the structural alignment between the query structure and known structures in the PDB library.TM-scores are in the range of 0–1. The higher the TM-score, the greater the structural similarity between two model proteins.
1432 Alternative Splicing Events in the Huntingtin RNA
exon 29. This binding site was similar in bothsequence and structure to a binding site in Trans-portin 1 (TNPO1), which binds to the Tap protein.
The HTTΔex29 transcript in human cellsand tissue
Two methods were used to determine whether thesame splice variants found in mouse Htt were alsopresent for HTT derived from human brain tissue.First, an amplicon was generated between exons 27and 31 (primer set P32; Table S2b) using HeLa cellRNA and in control and HD patient prefrontal corticalRNA. No product additional to the expected bandwas obvious and so the gel was band-stabbed at thepredicted location of the splice variant lacking exon
Table 2. Consensus of predicted GO terms for the 3Dmodelled region of mouse wild-type HTT protein (residues503–1999, encompassing 1497 residues in total), whichincludes exon 29 (residues 1233 to 1269)
GO term name Domain GO identifier
Protein binding Molecular function GO:0005515Protein transporter
activityMolecular function GO:0008565
Regulation of geneexpression
Biological process GO:0010468
Nuclear export Biological process GO:0051168Regulation of catalytic
activityBiological process GO:0050790
Regulation ofmacromoleculebiosynthetic process
Biological process GO:2000112
Protein modification bysmall protein
Biological process GO:0032446
Cellular developmentalprocess
Biological process GO:0048869
Negative regulation ofmolecular function
Biological process GO:0044092
Regulation of RNAmetabolic process
Biological process GO:0051252
Protein targeting Biological process GO:0006605RNA transport Biological process GO:0050658
The GO consensus is restricted here to two domains: molecularfunction, the elemental activities of a gene product at themolecular level, such as binding or catalysis, and biologicalprocess, operations or sets of molecular events with a definedbeginning and end.
29. HTTΔex28 was amplified and its identity wasconfirmed by sequencing (Fig. 5a and b).Second, primers specific to the human splice
variants HTTΔex28 and HTTΔex29 were generated(primer sets P33 and P34, respectively; Table S2b).These primers only generated a PCR product whenused in a nested PCR with the amplicon betweenexons 27 and 31 as template using primer set P32 togenerate this initial amplicon (Fig. 5b). The specificityof these splice variant-specific primers was confirmedby the absence of any product by this nested PCRmethod using a full-length HTT plasmid as template(data not shown). Data are shown here for thepresence or absence of the amplicons in one patientand one control for each brain region (Fig. 5c). Thiswas repeated for 12 HD patients and 12 controls, andthis confirmed that there was no apparent associationof the presence or absence of a band with HD status.
Discussion
The HTT gene harbours the expanded andunstable CAG repeat, the cause of HD [23]. A clearunderstanding of the normal function of HTT, itssubcellular localisation and the form in which it existswithin the cell are important for elucidation of thepathogenic mechanisms in the disease. The mech-anisms by which this mutation culminates in cellulardysfunction still remain largely unresolved althoughmultiple pathways and processes are implicated:RNA toxicity [24], transcriptional dysregulation [25],DNA repair [26], mitochondrial dysfunction andenergy deficits [27], aberrant protein cleavage[6,28], protein mislocalisation [29] and cellulartransport defects [30].It has been well established that HTT exists in
truncated forms in addition to the full-length protein[6,31], although the identity of many of thesefragments remains to be determined. Here, weprobed for the existence of alternative transcripts ofthe Htt gene that could potentially account for anytruncated HTT protein product. Three alternativetranscripts were found in the region of exon 28/29(Fig. 1), which were most abundant in the cerebellumand least abundant in the caudate nucleus ofHdh(Q150) mutant and wild-type mice. One variant
Fig. 4. 3D structural model of murine HTT centred around exon 29. 3D structural model of themouse constitutive wild-typeHTT protein (residues 503–1999). Exon 29 is located at residues 1233 to 1269. Highlighted in red is an α-helical region that isencoded by exon 29. The 11 contact residues predicted to form a protein binding site are highlighted in green, with 5 of theinterface residues being located in the α-helical region that is encoded by exon 29. Therefore, in the alternatively splicedvariant lacking exon 29 (HTTΔex29), it is likely that the predicted protein binding site would be abolished.
1433Alternative Splicing Events in the Huntingtin RNA
demonstrated an in-frame retention of 57 bp of intron28 upstream of exon 29 (Htt+57in28), which utilised acryptic splice site, predicted to be stronger than thedownstream natural splice site. It is most likely thatthe favourable recruitment of the natural splice site isdue to the presence of auxiliary cis-acting splicingregulatory elements, which serve to strengthen exondefinition of exon 29. Additionally, however, the useof the cryptic splice site introduced a PTC at 25nucleotides downstream of this sequence (Fig. 1c).Although many such products undergo nonsense-mediated decay [32], notably this would result inan HTT protein product of 1240 amino acids(~136 kDa) when translated, which coincides withan immunoaffinity-purified protein product of HTT [6].The other alternative transcript demonstrated exon29 skipping (HttΔex29), and as this was an in-frameevent, it was predicted to remove 37 amino acids(~4 kDa) from HTT. We have been unable to resolvefull-length HTT at ~348 kDa from this putative novelHTT isoform given the small size difference of thetwo potential products, even when performed usinga 6% Tris–glycine gel by SDS-PAGE run for severalhours. Neither have we found any evidence in theliterature that alternative versions of full-length HTTexist other than those with different polyglutaminelength stretches. We attribute this to the limitations ofresolution at such high molecular weights: othertechnical approaches are likely to be necessary toconfirm the existence of these novel splice variantsat the protein level.
To investigate the potential functional conse-quences of exon 29 skipping in HTTΔex29, weperformed 3D modelling using amino acids 503–1999 of murine HTT. The coding sequence of exon29 was found to contribute to a predicted proteinbinding site. This binding site is similar in bothsequence and structure to a binding site found inTNPO1, a transport receptor that transports sub-strates from the cytoplasm to the nucleus throughnuclear pore complexes by recognising nuclearlocalisation signals [33,34]. In particular, it transportsRNA-binding proteins [35]. The comparable site inTNPO1 binds to the Tap protein, thereby mediatingthe nuclear export of RNAs containing the constitu-tive transport element [36] and general mRNAtransport [37]. The putative binding site in HTTcould therefore bind to Tap, suggesting a role forHTT in the transport of substrates from the cyto-plasm to the nucleus [38] and nuclear export ofRNAs. Another polyglutamine-containing protein,Ataxin-1, binds to the Tap protein and may play arole in RNA processing and transport [39]. Indeed,aberrant nuclear export of RNAs acts as a patho-genic mechanism in several diseases includingmyotonic dystrophy, osteogenesis imperfecta type Iand certain motor neuron diseases [40]. Despitebeing implicated in RNA transport [41] and transla-tion [42], it remains to be known whether HTT canalter RNA export.This predicted protein binding site in HTT was
contributed to by residues from exons 28 and 29;
28 29 30 31 3227
(A) Constitutively spliced HTT transcript
30 31 3227
(B) HTTΔex28 transcript
29
P32 P32
(a)
(b)
(c)
Fig. 5. Alternative splicing in human HTT. (a) Model of the HTT gene (human), showing exons 27–32 (boxed regions).Note that exons are not drawn to scale. (b) HTT amplified with primer set P32 and band-stabbed blindly at the potentiallocation of the HTTΔex29 variant using RNA from HeLa cells and control (cont) and HD patient (pt) frontal cortex. This wasshown to beHTTΔex28 (band B) underneath the constitutiveHTT transcript (band A). (c) A nested PCR using the amplicongenerated by primer set P32 as template and splice variant-specific primers (primer sets P33 and P34) demonstrated thepresence of HTTΔex28 and HTTΔex29, respectively, in the human brain (CN, caudate nucleus; CB, cerebellum; BA4,Brodmann area 4; BA9, Brodmann area 9).
1434 Alternative Splicing Events in the Huntingtin RNA
therefore, it is probable that skipping of either exonwould affect the structure of this region. The murineprotein model lacking exon 29 still contained thebinding site, but it involved different contact residues.Therefore, it is likely that this would result in analtered binding affinity for any binding partner ratherthan a complete loss of binding.Levels of the HttΔex29 variant were significantly
reduced in the cerebellum of Hdh(Q150/Q150) micecompared to wild type and relative to the constitutivetranscript, suggesting that the splicing of Htt isperturbed by the HD mutation in cerebellum. Thecerebellum is known to be subject to higher levels ofalternative splicing than other brain regions [43], andit has been shown to degenerate early in HD [44]. A
direct association between the mis-splicing of Httand tissue-specific disease pathogenesis is yet to beshown in HD, although the exon 1–intron 1 transcriptrecently shown exclusively in Hdh(Q150/Q150)diseased mice [11] supports the notion that mis-spli-cing of HTT itself may be an important event in HD. Itis therefore possible that the HttΔex29 variant foundhere could contribute to disease pathogenesis andfurther work is required to understand its functionalsignificance and any effect it could have on potentialgene silencing therapies in HD.With this in mind, the study was translated to
human subjects using postmortem control and HDpatient microdissected brain regions. This verifiedthe presence of the HTTΔex29 variant and also
1435Alternative Splicing Events in the Huntingtin RNA
revealed exon 28 skipping (HTTΔex28) in humanbrain. These were both present at low levels as theycould only be detected using second round PCR byband-stabbing and reamplification or by a nestedPCR method. As a result, qPCR to assess expres-sion levels and a link to pathogenicity was notpossible. As was the case for the murine model ofHD, in human brain, exon 29 skipping was anin-frame event whilst exon 28 skipping was anout-of-frame event resulting in a PTC within exon 29(10th codon). This transcript may undergo nonsen-se-mediated decay; however, the generation of atruncated HTT protein product cannot be ruled out,and if it were to be translated, the protein would be1197 amino acids in length (~133 kDa). Proteinproducts smaller than full-length HTT have beendetected in human brain material by Western blotting[4], and it remains to be determined whether theseproducts correspond to protein isoforms encoded bythe novel HTT RNA species.Mis-splicing and disease is a well-reported phe-
nomenon, and various databases that documentthese associations exist.†,‡, § In the present study,we found that both the murine and human HTTtranscripts showed alternative splicing. It is wellestablished that alternative splicing events at themRNA level can result in distinct protein isoformswith unique functions [45]. This work therefore raisesseveral important points. Firstly, that HTT may existas a more diverse pool of alternative transcripts thanoriginally thought. Secondly, that the full-length HTTprotein may exist as multiple protein isoforms withspecific cellular functions. Characterisation of anynovel protein isoform could provide more informationabout the normal function of the HTT protein, whichis still not fully understood [46]. Thirdly, this workcorroborates the recent work at the Bates laboratory,which suggests that mis-splicing of HTT occurs inHD [11]. Mis-splicing per se could be a potentialtherapeutic target [47] and could open a new avenuefor HD treatment. Finally, a comprehensive knowl-edge of any novel HTT transcripts could impact thedesign of RNA silencing therapies, currently of greatinterest in HD research [48,49].
Materials and Methods
Samples
Mouse
Hdh(Q150) mice [50] were bred on 129/Ola × C57BL6/Jbackground and were maintained as described elsewhere[51]. HdhQ92 mice were originally bred and maintained ona mixed 129SvEv × CD1 background [52] but werebackcrossed onto a mixed 129SvEv × C57BL6/J back-ground and subsequently maintained over N6 generationson a C57BL/6J background as described in Brooks et al.[53]. The brain was removed and regions were microdis-
sected into motor cortex, striatum, prefrontal cortex,cerebellum, hippocampus and the remaining tissue or“rest of brain”. The age of these animals ranged from 5 to24 months. Peripheral tissues (lung, liver, muscle, kidneyand heart) were dissected from 13-month-old Hdh(Q150)wild-type and homozygous mutant animals bred on aC57BL/6J background. All experiments were carried outin accordance with the United Kingdom Animals (ScientificProcedures) Act of 1986 and subject to local ethicalreview.
Human
HD patient and control cerebellum, caudate nucleus,prefrontal cortex (Brodmann area 9) and motor cortex(Brodmann area 4) samples were obtained from the NewZealand Neurological Brain Bank in accordance with theethical principles stated in the Declaration of Helsinki.
RNA extraction and reverse transcription
RNA was extracted from tissue in lysing matrix D tubes(MP Biomedicals) using a Fastprep machine at a speed of4.0 for 20 s (brain); at speeds of 4.0 for 20 s, 4.5 for 25 s,and 5.0 for 30 s (liver); at speeds of 5.5 for 30 s and 5.5 for20 s (heart); and at a speed of 5.5 for 30 s twice (lung,muscle and kidney) in Trizol (Life Technologies) accordingto standard protocol. RNA cleanup was performed usingthe RNeasy MinElute kit (Qiagen) according to standardprotocol. DNase treatment of the RNA was carried outusing Turbo DNA-free (Ambion). Extracted RNA wasquantified using a Nanodrop ND-1000 spectrophotometer(Thermoscientific), and its quality was verified in randomlyselected samples on an Agilent 2100 Bioanalyzer (AgilentTechnologies). cDNA was produced from RNA usingSuperscript III (Life Technologies) and either random oroligo(dT)12–18 primers (Life Technologies) according tostandard protocol.
Primer design
Murine Htt mRNA sequence (gi: 315221149) andgenomic DNA sequence (gi: 149354224) were alignedusing NCBI-BLAST [54] in order to identify intron/exonboundaries. mRNA fragments of approximately 1000 bpwere inserted into primer 3 [55], and overlapping primer setswere designed to amplify the wholeHtt gene in amplicons ofapproximately 500–700 bp. Twenty-eight sets of overlap-ping primers were selected as shown in Table S2.In order to discriminate the three Htt transcripts found,
we selectively designed primers to amplify the novel HttΔex29 variant, the novel Htt+57in28 variant and thefull-length Htt transcript (Table S2a).
PCR
PCR was carried out with BioTaq DNA polymeraseaccording to standard protocol (Bioline) and thermocyclingconsisting of 2 min at 95 °C, followed by 30 cycles of95 °C for 30 s; 58 °C, 59 °C, 60 °C or 62 °C for 30 s (seeTable S2 for annealing temperatures); 72 °C for 30 s witha final extension at 72 °C for 10 min. PCR products were
1436 Alternative Splicing Events in the Huntingtin RNA
visualised using gel electrophoresis (1–2% agarose) withethidium bromide and a UVP gel documentation system.Any bands in addition to the predicted amplicon on the gelwere isolated using the band stab method of Bjourson andCooper [56]. Reamplification of these isolated sequenceswas performed as described above.
Quantitative PCR
qPCR was performed using Power SYBR Green PCRMaster Mix (Applied Biosciences) and an HT5700machine(ABI) with thermocycling conditions of 50 °C for 2 minand 95 °C for 10 min followed by 40 cycles of 95 °C for15 s and 60 °C for 1 min. For each transcript (HttΔex29,Htt+57in28 and constitutive Htt), expression levels fromthree different brain regions (motor cortex, caudatenucleus and cerebellum) were compared between wild-type (n = 5) and Hdh(Q150/Q150) (n = 4) mice by meansof a t test using a Bonferroni corrected significance level of0.05. Htt expression levels were normalised to that ofubiquitin C using forward primer (5′ GAGTTCCGTCTGCTGTGTGA 3′) and reverse primer (5′ CCTCCAGGGTGATGGTCTTA 3′).
Sequencing
DNA sequencing of isolated PCR products was per-formed by Sanger sequencing using Big Dye 3.1 (LifeTechnologies) with an ABI3730 DNA analyser followed bysequencing product purification with Dynabeads (LifeTechnologies). Sequencing traces were analysed usingSequence Scanner software v1.0 (Applied Biosystems).
3D homology modelling
The 3D structure of the mouse wild-type HTT protein(NP_034544.1; p. 503–1999) was computed using ahomology modelling technique employing the I-TASSERserver [57]. I-TASSER was also employed to predictbinding sites residing within the modelled structures andany associated GO terms. Several templates were used toderive the predicted 3D structure; the top ranking structuralanalogs are listed in Table 1. The model used in thisanalysis had a C-score of −2.27. The C-score provides ameasure of confidence in the quality of a model predictedby I-TASSER. C-scores are typically in the range of −5 to2, where a higher C-score indicates greater confidence inthe accuracy of a given model. Chimera [58] was used tovisualise and annotate the 3D structure in Figs. 4 and S1.The numbering of the amino acids in HTT refers toreference sequence RefSeq: NM_010414.2 and RefProt:NP_034544.1.
Bioinformatics
Further identification of a possible functional role for theregion encoded by Htt exon 29 (which is not present in HttΔex29) was assessed using a multiple sequence align-ment (MSA) derived from 11 orthologous proteins thatincluded mouse, human, chimp, gorilla, rhesus macaque,rat, cow, cat, dog, chicken and zebrafish. Clustal Omegawas used to perform the MSA [59]. The sequence logo of
theMSA (Fig. 3) was constructed withWebLogo [60]. Splicesite prediction (Fig. 1c) was made with NNSplice [61].Supplementary data to this article can be found online at
http://dx.doi.org/10.1016/j.jmb.2013.12.028.
Acknowledgements
We acknowledge the New Zealand NeurologicalFoundation Human Brain Bank for provision ofhuman brain material. We thank Zubeyde Bayram-Weston for her help with mouse tissue dissection.The technical assistance of Cardiff University'sCentral Biotechnology Services is gratefully ac-knowledged. CHDI funded the maintenance of theanimal colonies. We also acknowledge the MRCCentre grant (G0801418) (L.J.) and support from theE. J. Norman fund.Conflict of Interest Statement: The authors
declare that there is no conflict of interest.
Received 16 October 2013;Received in revised form 23 December 2013;
Accepted 25 December 2013Available online 3 January 2014
[1] Ross CA, Tabrizi SJ. Huntington's disease: from molecularpathogenesis to clinical treatment. Lancet Neurol2011;10:83–98.
[2] Jones L, Hughes A. Pathogenic mechanisms in Huntington'sdisease. Int Rev Neurobiol 2011;98:373–418.
[3] Kim YJ, Yi Y, Sapp E, Wang Y, Cuiffo B, Kegel KB, et al.Caspase 3-cleaved N-terminal fragments of wild-type andmutant huntingtin are present in normal and Huntington'sdisease brains, associate with membranes, and undergocalpain-dependent proteolysis. Proc Natl Acad Sci U S A2001;98:12784–9.
1437Alternative Splicing Events in the Huntingtin RNA
[4] Mende-Mueller LM, Toneff T, Hwang SR, Chesselet MF,Hook VY. Tissue-specific proteolysis of Huntingtin (htt) inhuman brain: evidence of enhanced levels of N- and C-terminal htt fragments in Huntington's disease striatum. JNeurosci 2001;21:1830–7.
[5] Toneff T, Mende-Mueller L, Wu Y, Hwang SR, Bundey R,Thompson LM, et al. Comparison of huntingtin proteolyticfragments in human lymphoblast cell lines and human brain.J Neurochem 2002;82:84–92.
[6] Landles C, Sathasivam K, Weiss A, Woodman B, Moffitt H,Finkbeiner S, et al. Proteolysis of mutant huntingtin producesan exon 1 fragment that accumulates as an aggregatedprotein in neuronal nuclei in Huntington disease. J Biol Chem2010;285:8808–23.
[7] Kim E, Goren A, Ast G. Alternative splicing: currentperspectives. Bioessays 2008;30:38–47.
[8] PanQ,ShaiO, LeeLJ,FreyBJ,BlencoweBJ.Deepsurveyingofalternative splicing complexity in the human transcriptome byhigh-throughput sequencing. Nat Genet 2008;40:1413–5.
[9] Lin B, Rommens JM, Graham RK, Kalchman M, MacDonaldH, Nasir J, et al. Differential 3' polyadenylation of theHuntington disease gene results in two mRNA species withvariable tissue expression. Hum Mol Genet 1993;2:1541–5.
[10] Casanova E, Alonso-Llamazares A, Zamanillo D, Garate C,Calvo P, Chinchetru MA. Identification of a long huntingtinmRNA transcript in mouse brain. Brain Res 1996;743:320–3.
[11] Sathasivam K, Neueder A, Gipson TA, Landles C, BenjaminAC, Bondulich MK, et al. Aberrant splicing of HTT generatesthe pathogenic exon 1 protein in Huntington disease. ProcNatl Acad Sci U S A 2013;110:2366–70.
[12] Trabzuni D, Wray S, Vandrovcova J, Ramasamy A, Walker R,Smith C, et al. MAPT expression and splicing is differentiallyregulated by brain region: relation to genotype and implication fortauopathies. Hum Mol Genet 2012;21:4094–103.
[13] Duyao MP, Auerbach AB, Ryan A, Persichetti F, Barnes GT,McNeil SM, et al. Inactivation of the mouse Huntington'sdisease gene homolog Hdh. Science 1995;269:407–10.
[14] Nasir J, Floresco SB, O'Kusky JR, Diewert VM, Richman JM,Zeisler J, et al. Targeted disruption of the Huntington's diseasegene results in embryonic lethality and behavioral and morpho-logical changes in heterozygotes. Cell 1995;81:811–23.
[15] Zeitlin S, Liu JP, Chapman DL, Papaioannou VE, EfstratiadisA. Increased apoptosis and early embryonic lethality in micenullizygous for the Huntington's disease gene homologue.Nat Genet 1995;11:155–63.
[16] White JK, AuerbachW, Duyao MP, Vonsattel JP, Gusella JF,Joyner AL, et al. Huntingtin is required for neurogenesis andis not impaired by the Huntington's disease CAG expansion.Nat Genet 1997;17:404–10.
[17] O'Kusky JR, Nasir J, Cicchetti F, Parent A, Hayden MR.Neuronal degeneration in the basal ganglia and loss of pallido-subthalamic synapses in mice with targeted disruption of theHuntington's disease gene. Brain Res 1999;818:468–79.
[18] Dragatsis I, Levine MS, Zeitlin S. Inactivation of Hdh in thebrain and testis results in progressive neurodegeneration andsterility in mice. Nat Genet 2000;26:300–6.
[19] Auerbach W, Hurlbert MS, Hilditch-Maguire P, Wadghiri YZ,Wheeler VC, Cohen SI, et al. The HD mutation causesprogressive lethal neurological disease in mice expressingreduced levels of huntingtin. HumMol Genet 2001;10:2515–23.
[20] Trushina E, Dyer RB, Badger 2nd JD, Ure D, Eide L, TranDD, et al. Mutant huntingtin impairs axonal trafficking inmammalian neurons in vivo and in vitro. Mol Cell Biol2004;24:8195–209.
[21] Roy A, Xu D, Poisson J, Zhang Y. A protocol for computer-based protein structure and function prediction. J Vis Exp2011:e3259.
[22] Branden T, Tooze J. Introduction to protein structure. 2nd ed.Garland Publishing; 1998.
[23] A novel gene containing a trinucleotide repeat that isexpanded and unstable on Huntington's disease chromo-somes. The Huntington's Disease Collaborative ResearchGroup. Cell 1993;72:971–83.
[24] Mykowska A, Sobczak K, Wojciechowska M, Kozlowski P,Krzyzosiak WJ. CAG repeats mimic CUG repeats in themisregulation of alternative splicing. Nucleic Acids Res2011;39:8938–51.
[25] Thomas EA, Coppola G, Tang B, Kuhn A, Kim S, GeschwindDH, et al. In vivo cell-autonomous transcriptional abnormal-ities revealed in mice expressing mutant huntingtin in striatalbut not cortical neurons. Hum Mol Genet 2011;20:1049–60.
[26] Goula AV, Pearson CE, Della Maria J, Trottier Y, TomkinsonAE, Wilson III DM, et al. The nucleotide sequence, DNAdamage location, and protein stoichiometry influence thebase excision repair outcome at CAG/CTG repeats. Bio-chemistry 2012;51:3919–32.
[27] Costa V, Scorrano L. Shaping the role of mitochondria in thepathogenesis of Huntington's disease. EMBO J2012;31:1853–64.
[28] Juenemann K, Weisse C, Reichmann D, Kaether C,Calkhoven CF, Schilling G. Modulation of mutant huntingtinN-terminal cleavage and its effect on aggregation and celldeath. Neurotox Res 2011;20:120–33.
[29] Havel LS, Wang CE, Wade B, Huang B, Li S, Li XJ.Preferential accumulation of N-terminal mutant huntingtin inthe nuclei of striatal neurons is regulated by phosphorylation.Hum Mol Genet 2011;20:1424–37.
[30] Pardo R, Molina-Calavita M, Poizat G, Keryer G, Humbert S,Saudou F. pARIS-htt: an optimised expression platform tostudy huntingtin reveals functional domains required forvesicular trafficking. Mol Brain 2010;3:17.
[31] Kegel KB, Meloni AR, Yi Y, Kim YJ, Doyle E, Cuiffo BG, et al.Huntingtin is present in the nucleus, interacts with thetranscriptional corepressor C-terminal binding protein, andrepresses transcription. J Biol Chem 2002;277:7466–76.
[32] Maquat LE. Nonsense-mediated mRNA decay: splicing,translation and mRNP dynamics. Nat Rev Mol Cell Biol2004;5:89–99.
[33] Sorokin AV, Kim ER, Ovchinnikov LP. Nucleocytoplasmictransport of proteins. Biochemistry (Mosc) 2007;72:1439–57.
[34] Troakes C, Hortobagyi T, Vance C, Al-Sarraj S, Rogelj B,Shaw CE. Transportin 1 colocalization with Fused inSarcoma (FUS) inclusions is not characteristic for amyotro-phic lateral sclerosis-FUS confirming disrupted nuclearimport of mutant FUS and distinguishing it from frontotem-poral lobar degeneration with FUS inclusions. NeuropatholAppl Neurobiol 2013;39:553–61.
[35] Lee BJ, Cansizoglu AE, Suel KE, Louis TH, Zhang Z, ChookYM. Rules for nuclear localization sequence recognition bykaryopherin beta 2. Cell 2006;126:543–58.
[36] Gruter P, Tabernero C, von Kobbe C, Schmitt C, Saavedra C,Bachi A, et al. TAP, the human homolog of Mex67p, mediatesCTE-dependent RNA export from the nucleus. Mol Cell1998;1:649–59.
[37] Truant R, Kang Y, Cullen BR. The human tap nuclear RNAexport factor contains a novel transportin-dependent nuclearlocalization signal that lacks nuclear export signal function. JBiol Chem 1999;274:32167–71.
1438 Alternative Splicing Events in the Huntingtin RNA
[38] Hilditch-Maguire P, Trettel F, Passani LA, Auerbach A,Persichetti F, MacDonald ME. Huntingtin: an iron-regulatedprotein essential for normal nuclear and perinuclear organ-elles. Hum Mol Genet 2000;9:2789–97.
[39] Irwin S, Vandelft M, Pinchev D, Howell JL, Graczyk J, Orr HT,et al. RNA association and nucleocytoplasmic shuttling byataxin-1. J Cell Sci 2005;118:233–42.
[40] Mastroyiannopoulos NP, Shammas C, Phylactou LA.Tackling the pathogenesis of RNA nuclear retention inmyotonic dystrophy. Biol Cell 2010;102:515–23.
[41] Savas JN, Ma B, Deinhardt K, Culver BP, Restituito S, WuL, et al. A role for huntington disease protein in dendriticRNA granules. J Biol Chem 2010;285:13142–53.
[42] Culver BP, Savas JN, Park SK, Choi JH, Zheng S, Zeitlin SO,et al. Proteomic analysis of wild-type and mutant huntingtin-associated proteins in mouse brains identifies unique interac-tions and involvement in protein synthesis. J Biol Chem2012;287:21599–614.
[43] de la Grange P, Gratadou L, Delord M, Dutertre M, AuboeufD. Splicing factor and exon profiling across human tissues.Nucleic Acids Res 2010;38:2825–38.
[44] Rub U, Hoche F, Brunt ER, Heinsen H, Seidel K, Del TurcoD, et al. Degeneration of the cerebellum in Huntington'sdisease (HD): possible relevance for the clinical pictureand potential gateway to pathological mechanisms of thedisease process. Brain Pathol 2013;23:165–77.
[45] Nilsen TW, Graveley BR. Expansion of the eukaryoticproteome by alternative splicing. Nature 2010;463:457–63.
[46] Zheng Q, Joinnides M. Hunting for the function of Huntingtin.Dis Model Mech 2009;2:199–200.
[47] Singh RK, Cooper TA. Pre-mRNA splicing in disease andtherapeutics. Trends Mol Med 2012;18:472–82.
[48] Carroll JB, Warby SC, Southwell AL, Doty CN, Greenlee S,Skotte N, et al. Potent and selective antisense oligonucleo-tides targeting single-nucleotide polymorphisms in theHuntington disease gene / allele-specific silencing of mutanthuntingtin. Mol Ther 2011;19:2178–85.
[50] Lin CH, Tallaksen-Greene S, Chien WM, Cearley JA,Jackson WS, Crouse AB, et al. Neurological abnormalitiesin a knock-in mouse model of Huntington's disease. HumMolGenet 2001;10:137–44.
[51] Brooks S, Higgs G, Jones L, Dunnett SB. Longitudinalanalysis of the behavioural phenotype in Hdh(CAG)150Huntington's disease knock-in mice. Brain Res Bull2012;88:182–8.
[52] Wheeler VC, Auerbach W, White JK, Srinidhi J, Auerbach A,Ryan A, et al. Length-dependent gametic CAG repeat instabilityin the Huntington's disease knock-in mouse. Hum Mol Genet1999;8:115–22.
[53] Brooks S, Higgs G, Jones L, Dunnett SB. Longitudinalanalysis of the behavioural phenotype in HdhQ92Huntington's disease knock-in mice. Brain Res Bull2012;88:148–55.
[55] Rozen S, Skaletsky H. Primer3 on the WWW for generalusers and for biologist programmers. Methods Mol Biol2000;132:365–86.
[56] Bjourson AJ, Cooper JE. Band-stab PCR: a simple techniquefor the purification of individual PCR products. Nucleic AcidsRes 1992;20:4675.
[57] Roy A, Kucukural A, Zhang Y. I-TASSER: a unified platformfor automated protein structure and function prediction. NatProtoc 2010;5:725–38.
[58] Pettersen EF, Goddard TD, Huang CC, Couch GS, GreenblattDM, Meng EC, et al. UCSF Chimera–a visualization system forexploratory research and analysis. J Comput Chem2004;25:1605–12.
[59] Goujon M, McWilliam H, Li W, Valentin F, Squizzato S, PaernJ, et al. A new bioinformatics analysis tools framework atEMBL-EBI. Nucleic Acids Res 2010;38:W695–9.
[60] Crooks GE, Hon G, Chandonia JM, Brenner SE.WebLogo: a sequence logo generator. Genome Res2004;14:1188–90.
[61] Reese MG, Eeckman FH, Kulp D, Haussler D. Improvedsplice site detection in Genie. J Comput Biol 1997;4:311–23.
