INVESTIGATION Saccharomyces cerevisiae Sen1 as a Model for the Study of Mutations in Human Senataxin That Elicit Cerebellar Ataxia Xin Chen, Ulrika Müller, Kaitlin E. Sundling, and David A. Brow1 Department of Biomolecular Chemistry, University of Wisconsin School of Medicine and Public Health, Madison, Wisconsin 53706-1532 ORCID ID: 0000-0003-1871-4477 (X.C.) ABSTRACT The nuclear RNA and DNA helicase Sen1 is essential in the yeast Saccharomyces cerevisiae and is required for efficient termination of RNA polymerase II transcription of many short noncoding RNA genes. However, the mechanism of Sen1 function is not understood. We created a plasmid-based genetic system to study yeast Sen1 in vivo. Using this system, we show that (1) the minimal essential region of Sen1 corresponds to the helicase domain and one of two flanking nuclear localization sequences; (2) a previously isolated terminator readthrough mutation in the Sen1 helicase domain, E1597K, is rescued by a second mutation designed to restore a salt bridge within the first RecA domain; and (3) the human ortholog of yeast Sen1, Senataxin, cannot functionally replace Sen1 in yeast. Guided by sequence homology between the conserved helicase domains of Sen1 and Senataxin, we tested the effects of 13 missense mutations that cosegregate with the inherited disorder ataxia with oculomotor apraxia type 2 on Sen1 function. Ten of the disease mutations resulted in transcription readthrough of at least one of three Sen1-dependent termination elements tested. Our genetic system will facilitate the further investigation of structure–function relationships in yeast Sen1 and its orthologs. RANSCRIPTION termination by eukaryotic RNA polymer- 2008). A set of core factors distinguish the Sen1-dependent Tase II (Pol II) uses at least two pathways, one that is cou- pathway from the poly(A)-dependent pathway, including Sen1 pled to cleavage and polyadenylation of the nascent transcript and the RNA-binding proteins Nrd1 and Nab3 (Steinmetz and [the poly(A)-dependent pathway] and one that involves the Brow 1996, 1998; Conrad et al. 2000; Steinmetz et al. 2001; activity of the RNA/DNA helicase Sen1 (the Sen1-dependent Carroll et al. 2007). However, some short messenger RNA pathway) (Kuehner et al. 2011). The Sen1-dependent path- (mRNA) genes, such as CYC1, may have hybrid terminators that way was first identified in the budding yeast Saccharomyces require factors from both pathways (Steinmetz et al. 2006b). cerevisiae and is responsible for transcription termination of S. cerevisiae Sen1 is a 252-kDa superfamily 1 helicase many short, noncoding RNA genes, including small nuclear encoded by the essential SEN1 gene. Its helicase domain is (sn) and small nucleolar (sno) RNAs (Winey and Culbertson located in the C-terminal half of the protein (Figure 1A), and 1988; Steinmetz and Brow 1996; Rasmussen and Culbertson the N-terminal 975 amino acids are dispensable for viability 1998; Steinmetz et al. 2001). It also restricts the elongation (DeMarini et al. 1992).TheorthologofSen1 in the fission and accumulation of pervasive cryptic unstable transcripts yeast Schizosaccharomyces pombe has ATP-dependent, 59-to-39 (Arigo et al. 2006b; Thiebaut et al. 2006) and regulates tran- DNA and RNA unwinding activities in vitro (Kim et al. 1999). scription of some protein-coding genes by premature termina- Two hypomorphic alleles of S. cerevisiae SEN1, sen1-1 and tion, i.e., attenuation (Steinmetz et al. 2001; Arigo et al. 2006a; nrd2-1, have G1747D and E1597K substitutions in the heli- Steinmetz et al. 2006b; Jenks et al. 2008; Kuehner and Brow case domain, respectively, and exhibit terminator read- through on a subset of Pol II-transcribed genes (DeMarini Copyright © 2014 by the Genetics Society of America et al. 1992; Steinmetz and Brow 1996; Steinmetz et al. doi: 10.1534/genetics.114.167585 2001, 2006a; Kuehner and Brow 2008; Mischo et al. 2011; Manuscript received June 20, 2014; accepted for publication August 11, 2014; fi published Early Online August 12, 2014. Hazelbaker et al. 2013). These ndings indicate that the Sen1 Supporting information is available online at http://www.genetics.org/lookup/suppl/ helicase domain plays an important role in Pol II termination. doi:10.1534/genetics.114.167585/-/DC1. 1Corresponding author: University of Wisconsin-Madison, Department of Biomolecular Two main models have been proposed for the mechanism Chemistry, 420 Henry Mall, Madison, WI 53706-1532. E-mail: [email protected] of Sen1-dependent termination. In one model, Sen1 is Genetics, Vol. 198, 577–590 October 2014 577 Figure 1 A plasmid-based system for testing mutations in yeast Sen1. (A) Comparison of S. cerevisiae Sen1 with human Senataxin. The conserved heli- case domains and their subdomains are highlighted by colored boxes. Numbers indicate amino acid residues. The loca- tion of missense disease mutations in Sentaxin are shown by colored dots. (B) Anti-Sen1(HD) immunoblotting of cell extracts from strains with either chromo- somal (46a)orplasmid-borne(pRS313- SEN1) SEN1. Nrd1 antiserum was used as a loading control (bottom). (C) Schematic genotype of yeast strain DAB206 after transformation with a mutant SEN1 allele and a terminator-reporter construct. Its chromosomal SEN1 and CUP1 genes are disrupted, and the SEN1 deletion is complemented by pRS316-SEN1. Spon- taneous loss of pRS316-SEN1 is selected for by plating on medium containing 5-FOA. Terminator readthrough can be assayed by plating on medium contain- ing copper sulfate, which is detoxified by the CUP1 gene product. (D) Schematic of the ACT1-CUP1 construct for measuring terminator activity. (E) Schematic of the CUP1 reporter for measuring attenuator activity. proposed to be analogous to bacterial Rho helicase, trans- result from defects in expression of genes essential for long- locating 59 to 39 along the nascent transcript until colliding term neuron survival. with paused Pol II and terminating its elongation (Steinmetz To better understand the role of Sen1 in the transcription and Brow 1996; Porrua and Libri 2013). In the other model, termination process, we created a plasmid-based assay to Sen1 unwinds RNA/DNA hybrids (R loops) formed between characterize Sen1 function in vivo. We found that the essen- the nascent transcript and the template strand, thereby tial region of Sen1 corresponds to the helicase domain and allowing recognition of termination elements in the nascent a flanking nuclear localization signal, but that efficient tran- transcript by RNA-binding proteins (Mischo et al. 2011). scription termination by Sen1 requires its N-terminal domain Both models can account for an essential role of the Sen1 as well. Using the crystal structure of the related helicase helicase domain in termination. Upf1 as a guide, we determined that the growth and termi- The human ortholog of yeast Sen1, called Senataxin, is nation defects caused by nrd2-1 (sen1-E1597K) result from encoded by the gene SETX. Mutations in SETX cosegregate disruption of an intradomain salt bridge in the first RecA re- with two progressive neurodegenerative disorders with ju- peat. Examining the utility of Sen1 as a surrogate for Sena- venile onset: autosomal recessive ataxia oculomotor apraxia taxin, we analyzed the effects of 13 human AOA2 mutations type 2 (AOA2) and autosomal dominant amyotrophic lateral and showed that six are recessive lethal and that an addi- sclerosis type 4 (ALS4) (Chen et al. 2004; Moreira et al. tional five result in recessive termination defects. Our studies 2004; Lemmens et al. 2010). The mechanism by which the establish a facile genetic system for studying structure–func- SETX mutations cause AOA2 and ALS4 has not been deter- tion relationships in Sen1 and underscore the importance of mined, but many of the missense disease mutations are the helicase domain in Sen1-dependent termination of Pol II located in the conserved helicase domain (Figure 1A and transcription. Supporting Information, Supporting References, and Table S1), suggesting that the diseases are associated with dys- function of the helicase activity. Senataxin has been impli- Materials and Methods cated in Pol II transcription termination on protein-coding Plasmid construction genes (Skourti-Stathaki et al. 2011). Gene expression pro- filing showed that AOA2 mutations in SETX cause changes The SEN1 gene, including 434 bp upstream of the start co- in gene expression profiles in patient fibroblasts, including don and 212 bp downstream of the stop codon, was ampli- genes involved in neurogenesis and neuronal functions fied from genomic DNA of the strain 46a by two PCR (Fogel et al. 2014). Thus, it is possible that AOA2 and ALS4 reactions. One created an upstream SalI site and an internal 578 X. Chen et al. PstI site via a silent mutation in codon 965. The other cre- codons 1175–1890 and replace them with SETX codons ated the same internal PstI site and a downstream SpeI site. 1769–2484 to generate pRS313-SEN1/SETXhelicase. (The sequences of all primers used in this study are available pGAC24-CYC1 (Steinmetz and Brow 2003) and the on request.) The PCR products were digested with the ap- pRS316-NRD1-CUP1 reporter plasmids (Kuehner and Brow propriate restriction enzymes and ligated into SalI/SpeI-cut 2008) have been described elsewhere. The pGAC24-SNR47 pRS425 to create pRS425-SEN1(PstI). The SalI/SpeI insert reporter plasmid was generated by amplifying the SNR47 from pRS425-SEN1(PstI) was subcloned into pRS313, pRS315, terminator (Carroll et al. 2004) (positions +100 to +215, and pRS316. Point mutations in SEN1 were created by the relative to +1 transcription start site) from DAB206 geno- QuikChange procedure (Stratagene) and subcloned into mic DNA with an XhoI site introduced at each end. The PCR pRS313. product was digested with XhoI and ligated to XhoI-cut Deletions of SEN1 were created using pRS313-SEN1(PstI). pGAC24 (Lesser and Guthrie 1993). The C-terminal sequences were deleted by the QuikChange Yeast strains procedure, leaving the stop codon intact.