ANALYSIS AND -MEDIATED DECAY OF

RNAS WITH 5´-HYDROXYL TERMINI

by

SALLY E. PEACH

B.S., Massachusetts Institute of Technology, 2009

A thesis submitted to the

Faculty of the Graduate School of the

University of Colorado in partial fulfillment

of the requirements for the degree of

Doctor of Philosophy

Molecular Biology Program

2016

© 2016

SALLY PEACH

ALL RIGHTS RESERVED

ii This thesis for the Doctor of Philosophy degree by

Sally E. Peach

has been approved for the

Molecular Biology Program

by

Sandra L. Martin, Chair

David L. Bentley

Richard E. Davis

Katerina J. Kechris

Jeffrey S. Kieft

Jay R. Hesselberth, Advisor

Date: 05/20/2016

iii Peach, Sally E. (Ph.D., Molecular Biology Program)

Analysis and Kinase-mediated Decay of with 5´-hydroxyl Termini

Thesis directed by Assistant Professor Jay R. Hesselberth

ABSTRACT

RNA cleavage by metal ion-independent and self-cleaving produces RNA fragments with 5´-hydroxyl (5´-OH) and 2´,3´-cyclic phosphate termini. However, RNAs with 5´-OH termini (5´-OH RNAs) are not detected by conventional sequencing methods, and 5´-OH RNAs cannot be directly degraded by canonical 5´-phosphate-dependent 5´→3´ .

To identify new 5´-OH RNA fragments, I exploited the unique ligation mechanism of E. coli RtcB RNA to develop 5OH-seq, a novel method to capture and sequence

5´-OH RNAs. I applied 5OH-seq to budding yeast and captured known 5´-OH fragments produced by tRNA Splicing (SEN) during processing of -containing pre-tRNAs and by Rny1 during stress-induced cleavage of tRNA anti-codon loops. I identified numerous novel 5´-OH fragments derived from mRNAs: some 5´-OH mRNA fragments were derived from single, localized cleavages, while others were likely produced by multiple, distributed cleavages. Many 5´-OH mRNA fragments were produced upstream of codons for highly electrostatic peptides, suggesting that the fragments may be generated by co-translational mRNA decay, such as no-go decay (NGD).

5´-OH RNAs must be phosphorylated to become substrates for 5´-phosphate- dependent 5´→3´ exoribonucleases, a process I term “kinase-mediated decay” (KMD).

During the unfolded response (UPR), the ER-localized Ire1 cleaves HAC1 mRNA, yielding 2´,3´-cyclic phosphate and 5´-OH RNA fragments that are

iv ligated by Trl1 tRNA ligase to produce mature HAC1 mRNA and activate downstream target . The HAC1 fragments cleaved by Ire1 have previously been identified with

5´-phosphate (5-PO4) termini in vivo in xrn1∆ yeast, suggesting the HAC1 fragments are

KMD substrates. I found that HAC1 3´- accumulate with 5´-PO4 termini in xrn1∆ yeast when Trl1 is present, but not kinase-dead Trl1-D425N or RtcB ligase, suggesting that

Trl1 phosphorylates the HAC1 3´- for decay by Xrn1. Failure to degrade the HAC1

3´-exon promotes aberrant splicing and UPR activation without stimulus.

In summary, this thesis describes a method to directly capture and sequence 5´-OH

RNAs (5OH-seq), a comprehensive analysis of 5´-OH RNAs in budding yeast, and evidence for a unified theory of 5´-OH RNA degradation by kinase-mediated decay

(KMD). I also present preliminary data on the abundance of 5´-OH RNAs in mammalian tissues that lack the RNA 5´-kinase CLP1.

The form and content of this thesis are approved. I recommend its publication.

Approved: Jay Hesselberth

v DEDICATION

To Mom, for her boundless love and support throughout all of my crazy journeys,

and for her timeless feminist wisdom, which helped shape my path:

“Don’t marry the doctor, BE the doctor.”

vi ACKNOWLEDGEMENTS

First and foremost, I would like to acknowledge and thank my advisor, Jay

Hesselberth. He approaches science with an excitement and curiosity that is contagious, and he has been beyond essential in helping to push my research forward. As I’ve said on numerous occasions, I definitely won the PI lottery.

The rest of the Hesselberth lab has been equally important for my success. Monica

Ransom is an incredibly thoughtful person, and keeps up the spirits with scientific advice, presentation feedback, and cupcakes. Patrick Cherry provided important data for the kinase-mediated decay model (Figure 3.4), and is faster than Googling a biochemistry question. Erich Chapman has been a great role model and science brother, and he has made things far more interesting for us all. Big thanks to Austin Gillen(s) for secret -80ºC communications, and Kerri York for teaching me so much during that first year in the lab.

Finally, to *Dr. Suzi*: it just wasn’t the same without you around. It has been a pleasure completing my graduate studies in this lab, and I will miss everyone greatly.

I am extremely thankful to my thesis committee. My chair Sandy Martin has been a fantastic role model. Dick Davis has been the most thorough and thoughtful critic of manuscripts. Jeff Kieft and David Bentley and their labs have been exceptionally generous with ideas, reagents, and machinery. Katerina Kechris has patiently answered my naïve statistics questions and helped me develop more rigorous data analysis approaches.

I have been fortunate to be a part of two fantastic graduate programs and one amazing department. Thanks to Arthur Gutierrez-Hartmann, director of the Medical

Scientist Training Program (MSTP), for years of solid leadership and advice, and to Angie

Ribera, former associate director of MSTP, for incredible support during one of the hardest

vii moments of my life. Bob Sclafani, director of the Molecular Biology Program, and Mark

Johnston, chair of the Biochemistry and Medical Department, have been extremely encouraging and have taken an active role in my development as a scientist. The administrative staff behind the scenes made my life ultra easy, including: Annie Vazquez,

Sabrena Heilman, Sue Brozowski, Lindsay Quandt, Karen Vockrodt, Michele Parsons,

Emily Dailey, and Jodi Cropper. I have also appreciated the financial support of the

Molecular Biology NIH training grant T32 GM008730.

Science has been mailed to me from all over the country and the world. Many thanks to Marc Hammarlund from Yale University; Stewart Shuman and Beate Schwer from Sloan Kettering; Roy Parker and Yuriko Harigaya from University of Colorado

Boulder; Jingyan Wu and Anita Hopper from the Ohio State University; Ender Karaca and

James Lupski from the Baylor College of Medicine; and Javier Martinez and Stefan

Weitzer from the Institute of Molecular Biotechnology in Vienna. All of these scientists have shared samples and data with me during my efforts to find novel 5´-OH RNAs.

Several previous advisors have helped me on my way here. Ten years ago, Roger

Summons accepted me into his lab at MIT with no previous experience, and D’Arcy

Meyer-Dombard put up with teenaged-Sally and taught me how to wield a pipette. Dianne

Newman supervised my first significant research project as a senior at MIT, which certainly changed my life’s direction. Jake Jaffe was a truly phenomenal mentor at the

Broad Institute without whom I would have never applied to MD/PhD programs.

My friends and colleagues in the medical and graduate school have been critical to keeping me afloat. The list is large and incomplete, but here goes… From MSTP: Eric

Nguyen for Battlestar Galactica, hosting all of my parties, and explaining happiness; Chris

viii Knoeckel for being my free Udi’s psychiatrist and teaching me the secrets to graduate school success; Mira Estin, Sarah Nelson, Hannah Scarborough, Veronica Searles Quick, and Genevieve Park for forming a core group of MD/PhD women that I could both look up to and befriend; Jing Zhang for that, plus an ever-growing collection of strange banana experiences; and the rest of my surviving MSTP “2011” cohort, especially Tamara

Burleson Garcia for bringing some small-town Southern sense to this crazy world and just enough pink. From graduate school: my MolBio cohort, especially Ryan Sheridan, Alexis

Zukowski, and Maggie Balas; Becky Fusby for keeping it real in the truest sense of that phrase; Seena Shah for one all-too-short year and “a proper English garden”; and Charlotte

Siska for cows, tutus, and rolling in the dust. From medical school: Garland Castañeda for just enough sarcasm, spreadsheets, and summits to help me survive the first two years; and, of course, Stevie Kohler for continuously teaching me how to be human.

I’m also extremely thankful for those outside my Anschutz academic bubble, especially: Christie Ciarlo, who has been an essential adventuring partner, scientific comrade, and dear friend; MinWah Leung for setting an extremely high bar for both exploration and cheerfulness; and Camilo Guáqueta, for being my rock during this last grad school crunch time, BBB.

Finally, I would like to thank my family, who have been there for me since the beginnings of Little Woo: my sister Katie, who has taught me so much about overcoming adversity; my brother James, for his unfailing kindness and strength; my dad, for introducing me to the world of a physician; and, most importantly, to my amazing mom for her endless love, which will follow me always.

for life’s not a paragraph

ix TABLE OF CONTENTS

CHAPTER

I. INTRODUCTION ...... 1

Overview of RNA Modifications ...... 2

Generation of RNAs with 5´-hydroxyl Termini ...... 4

5´-OH RNAs in Biological Systems ...... 10

RNA Decay Pathways ...... 30

Kinase-Mediated Decay ...... 40

Methods for Global Mapping of RNA Modifications ...... 48

II. GLOBAL ANALYSIS OF RNA CLEAVAGE BY 5´-HYDROXYL RNA SEQUENCING ...... 53

Introduction ...... 53

Materials and Methods ...... 55

Results ...... 60

Discussion ...... 80

Summary ...... 89

III. KINASE-MEDIATED DECAY OF EUKARYOTIC MRNA CLEAVAGE PRODUCTS ...... 91

Introduction ...... 91

Materials and Methods ...... 93

Results ...... 99

Discussion ...... 110

Summary ...... 115

x IV. ANALYSIS OF 5´-HYDROXYL RNAS IN MAMMALS ...... 117

Introduction ...... 117

Materials and Methods ...... 118

Results ...... 122

Discussion ...... 128

Summary ...... 134

V. SUMMARY AND FUTURE DIRECTIONS ...... 135

Applications and Improvements of 5OH-Seq ...... 136

Future Studies and Considerations of Kinase-Mediated Decay ...... 142

REFERENCES ...... 149

xi LIST OF TABLES

TABLE

1.1 Functions of several eukaryotic metal ion-independent endoribonucleases...... 11

2.1 Oligonucleotides ...... 57

2.2 Yeast strains ...... 57

3.1 Plasmids...... 94

3.2 Yeast strains...... 95

3.3 Oligonucleotides ...... 98

4.1 Oligonucleotides ...... 119

xii LIST OF FIGURES

FIGURE

1.1 Two distinct mechanisms of endoribonuclease cleavage...... 9

1.2 Yeast and human pathways for processing of intron-containing tRNAs...... 15

1.3 Rny1 (yeast) and (humans) cleave the anticodon loops of tRNAs

under stress conditions...... 19

1.4 The unfolded protein response (UPR) in yeast is primarily regulated by the non-

canonical cytosolic splicing of HAC1 mRNA...... 25

1.5 Human IRE1α cleaves XBP1 and ER-proximal mRNAs...... 28

1.6 Canonical cytosolic mRNA decay pathways: Xrn1 and the exosome...... 34

1.7 Decay pathways for defective mRNAs: nonsense-mediated decay (NMD), non-stop decay (NSD), and no-go decay (NGD)...... 37

1.8 Examples of kinase-mediated decay (KMD)...... 46

2.1 Purified E. coli RtcB ligates RNA substrates with 3´-PO4 and 5´-OH termini .... 62

2.2. Schematic of method to clone and sequence RNAs with 5´-OH termini (5OH-seq)...... 63

2.3 5OH-seq captures RNAs with 5´-PO4 termini after in vitro dephosphorylation...... 67

2.4 Cleavages in tRNA in , anticodon loops, and D arm and loop ...... 68

2.5 Analysis of mRNAs captured by 5OH-seq...... 71

2.6 5´-OH RNAs exist in regions of clustered or discrete cleavage...... 72

2.7 Single cleavage in RPS31 mRNA is validated by link-RT-PCR...... 74

xiii 2.8 5´-OH RNAs accumulate upstream of regions encoding consecutive polycharged amino acids...... 77

2.9 HAC1 mRNA cleavage fragments accumulate with

5´-OH and 5´-PO4 termini...... 81

2.10 Some 5´-OH RNAs increase with tunicamycin treatment and occur in a common motif...... 83

3.1 Xrn1 degrades the phosphorylated HAC1 3´-exon cleavage product...... 102

3.2 The HAC1 3´-exon mRNA cleavage product requires the kinase activity of Trl1 to promote degradation by Xrn1...... 105

3.3 Xrn1 suppresses HAC1 mRNA splicing and UPR activation...... 108

3.4 Trl1 phosphorylates NGD 3´ fragments prior to 5´→3´ degradation by Xrn1...... 111

3.5 Models of kinase-mediated decay of mRNA cleavage fragments...... 113

4.1 CLP1 shRNA knockdowns validated by qRT-PCR and Western blot...... 123

4.2 CLP1-deficient cell lines have increased 5OH-signal surrounding PolyA sites...... 124

4.3 CLP1 kinase-dead mutants differentially accumulate 5OH-RNAs in mouse spinal cords...... 126

4.4 SeC 3´-tiRNAs accumulate in the spinal cords of mice...... 129

4.5 SeC tRNA Northern blot requires optimization...... 132

xiv CHAPTER I

INTRODUCTION

RNA is a diverse molecule comprising the majority of nucleic acid content in a given cell. It can act as the messenger encoding protein sequence (mRNA), the molecule decoding the sequence and carrying amino acids during protein translation (tRNA), and a structural and catalytic component of the ribosome (rRNA). The constant production of this essential molecule necessitates robust decay mechanisms to break down RNA and recycle . For instance, the canonical Xrn1 and exosome decay pathways in yeast maintain cytosolic homeostatsis by processive mRNA degradation. However, little is known about RNAs which exist outside of the known decay pathways. What happens when an RNA is cleaved and the resulting RNA terminus is immune to degradation by known exoribonucleases?

This dissertation seeks to address this question by employing both broad, transcriptome-wide analyses of RNAs with 5´-hydroxyl termini (5´-OH RNAs) as well as a focused study of the molecular mechanism of kinase-mediated decay (KMD) for specific

5´-OH RNAs within the unfolded protein response (UPR) and the no-go decay (NGD) pathway. This chapter provides an overview of RNA modifications, the generation and regulation of 5´-OH RNAs, the known RNA decay mechanisms, and the current methods of terminal RNA modification detection by next-generation sequencing. In Chapter II, I discuss the development and validation of the novel 5OH-seq method, which combines the unique ligation specificity of E. coli RtcB with massively parallel next-generation sequencing to directly capture and sequence 5´-OH RNAs. I describe the findings of the first transcriptome-wide analysis of 5´-OH RNAs in Saccharomyces cerevisiae, including

1 the enrichment of 5´-OH RNAs upstream of regions encoding polycharged amino acids and 5´-OH RNAs cleaved during UPR induction. In Chapter III, I further explore the

5OH-seq results in yeast. I demonstrate the molecular mechanism of the decay of two 5´-

OH RNAs: the 3´-exon fragment of HAC1 mRNA generated by Ire1 cleavage and the 3´- exon fragment of a no-go decay (NGD) reporter construct. I term these and other related processes “kinase-mediated decay” (KMD). In Chapter IV, I apply the 5OH-seq method to mammalian RNA samples and describe results of a preliminary analysis of 5´-OH RNAs in mouse spinal cords, human cell lines, and human patient-derived fibroblasts with compromised 5´-OH kinase activity. In Chapter V, I summarize the collective findings and describe future directions and applications to RNA decay, including extensions to human disease. Together, these chapters provide a comprehensive overview of 5´-OH

RNAs in eukaryotes and define the molecular mechanism of their degradation by KMD.

Overview of RNA Modifications

Modification of nucleobases and ribose sugars drastically increases the information coding potential of RNA, creating a “combinatorial chemistry playground” (Jackman and

Alfonzo 2013). The first glimpse of RNA modification arose when the hydrolyses of yeast

RNA with a and 5´ yielded inorganic phosphate, suggesting that some phosphoryl groups were attached to C5 of the ribose (Gulland and

Jackson 1938). Since this early discovery of what is now known as the 5´-phosphate terminal modification (5´-PO4), over 100 RNA modifications, both terminal and internal, have been identified on all major RNA types (tRNA, rRNA, and mRNA) and several minor

RNA subspecies (snRNA and miRNA) in all domains of life (Machnicka et al. 2012).

2 tRNA provides a stunning example of internal RNA modifications. In 1957, Davis and Allen described an abundant “fifth ” that comprised 4% of yeast tRNA nucleotides (Davis and Allen 1957). The chemical structure and properties of this pseudouridine (Ψ) nucleotide, the C5-glycoside isomer of uridine, were revealed shortly thereafter (Cohn 1960). Over the next half century, dozens of tRNA modifications were discovered (Reviews: Hopper 2013; Jackman and Alfonzo 2013; Helm and Alfonzo 2014).

An analysis of 561 sequenced tRNA molecules from a variety of species found 11.9% of the all nucleotides differed from the canonical bases and 16.4% of nucleotides in 34 sequenced yeast cytoplasmic tRNAs are modified (Sprinzl and Vassilenko 2005; Phizicky and Alfonzo 2010). Though the role of most tRNA modifications is not well-understood, it is abundantly clear that many RNA modifications are required for normal cellular functioning. Over 200 diseases have been linked to tRNA hypomodification, including cancers, asthma, diabetes, and neurological diseases (Torres et al. 2014; Abbott et al. 2014).

In addition to modifications of the internal nucleotides, the 5´ and 3´ termini of RNA can be covalently modified. The nature of these termini can inhibit or promote recognition by that modify the end chemistry by phosphorylation, cyclic phosphate diesterase activity, adenylation, guanylation, ligation, and so on. Perhaps the most well-studied example of a regulated 5´-terminal RNA modification is the 7-methylguanosine (m7G) cap on the 5´ terminus of mRNA. In the mid 1970s, a series of publications demonstrated the presence of the m7G cap on the 5´ terminus of mRNAs from viruses, slime mold, yeast, and mammals (Shatkin 1976). During , nascent pre-mRNA is dephosphorylated by RNA triphosphatase Cet1 (Tsukamoto et al. 1997). A guanine nucleotide is transferred to the resulting 5´-bisphosphate mRNA by mRNA

3 Ceg1 creating a 5´-to-5´ triphosphate linkage. Methylation of the guanine by mRNA (guanine-N7-)-methyltransferase Abd1 yields a m7G-mRNA (Mao et al. 1995). The m7G cap recruits eukaryotic initiation factor 4E (eIF4E) to promote translation (Sonenberg et al. 1979), and it also promotes 5´ proximal intron excision (Ohno et al. 1987). Importantly, the m7G cap prevents recognition by 5´→3´

Xrn1, and must first be removed by Dcp1/Dcp2 to promote decay of the mRNA (discussed further in the “RNA Decay” section).

In this dissertation, I seek to characterize another terminal RNA modification: the 5´- hydroxyl terminus. I consider the 5´-OH terminus a “hidden” terminus for two reasons.

First, 5´-OH RNAs have evaded extensive characterization due to their incompatibility with sequencing methods that rely on 5´-phosphate-dependent chemistries. Therefore, in order to comprehensively identify 5´-OH RNAs with precise resolution, I needed to develop a new method to specifically capture and map this previously “hidden” RNA species (Chapter II and Chapter IV). The second reason is that, similar to the m7G capped mRNAs, 5´-OH RNAs cannot be recognized or degraded by 5´-phosphate-dependent

5´→3´ exoribonucleases. In Chapter III, I describe how the “hidden” 5´-OH terminus can be phosphorylated, creating a competent substrate for 5´→3´ RNA decay.

Generation of RNAs with 5´-hydroxyl Termini

Within this section, I introduce three key processes generating 5´-OH RNAs: intrinsic cleavage, self-cleaving ribozymes, and metal-ion independent endoribonucleases.

Cleavages by several metal-ion independent endoribonucleases are of particular interest in

4 regulated pathways, which are further described in the subsequent section, “5´-OH RNAs in biological systems”.

Intrinsic RNA cleavage

RNA can assume a conformation promoting sequence-independent cleavage. This intrinsic RNA cleavage likely proceeds via an in-line SN2-like reaction that requires an initial deprotonation of the 2´-OH of the upstream ribose. Following deprotonation, the

2´-O functions as the nucleophile and attacks the neighboring 3´-5´ bridging phosphate prompting the 5´-O of the downstream ribose to act as a leaving group. Intrinsic RNA cleavage yields a 5´ RNA fragment with a 2´,3´-cyclic phosphate terminus and a 3´ RNA fragment with a 5´-OH terminus. The cleavage is relatively non-specific and likely accounts for a significant portion of in vitro RNA fragmentation.

Certain metal ions can enhance intrinsic RNA cleavage. Metal ions promote RNA cleavage by acting as Brönsted bases and abstracting a proton from the 2´-OH (Brown et al. 1985). Indeed, acceleration of RNA cleavage by divalent lead cations was first shown over 65 years ago (Dimroth et al. 1950). Other divalent cations Mg2+ and and Zn2+ also increase the rate of cleavage compared to uncatalyzed reactions, and lanthanide ions Eu3+,

Tb3+, and Yb3+ are even more efficient catalysts, accelerating the rate of dinucleotide cleavage 3-4 orders of magnitude (Breslow and Huang 1991).

Intrinsic RNA cleavage is important for many RNA sequencing methods. Single- stranded RNA cleaves at a rate ~100 times higher than double-stranded RNA (Soukoup and Beaker, 1999), enabling partial mapping of RNA structure by differential cleavage in the presence of metal cations (Forconi and Herschlag 2009). Intentional RNA

5 fragmentation is a critical step in the optimization of various sequencing library preparations, including the 5OH-seq method described within this thesis.

Self-cleaving ribozymes

Though the potential for enzymatic activity of RNA had been suggested by Carl

Woese, Francis Crick, and Leslie Orgel in 1967, the experimental evidence for the necessity of RNA in catalysis came a decade later. While investigating the mechanism of (RNase P) in the sequence-independent generation of 5´ termini of E. coli tRNAs, Sidney Altman’s group showed that purified RNase P contained a single protein and a single nucleic acid component. Treatment of purified RNase P with micrococcal (MN) or RNase A inhibited cleavage of a tRNATyr precursor, and addition of total cellular RNA after MN treatment partially restored the enzymatic activity, suggesting that the RNA component was required for enzymatic function (Stark et al. 1978). Separately,

Tom Cech’s group was investigating the removal mechanism of a 413 bp intervening sequence (IVS) from the 26S pre-rRNA in Tetrahymena thermophila. They initially purified an active form of 26S pre-rRNA that completed IVS excision and circularization even following extensive measures to inactivate potential protein contaminants including

SDS-phenol extraction, boiling in SDS, and treatment with proteases (Cech et al. 1981).

These results suggested, but did not definitely prove, that excision was not performed by a protein. In a follow up study, in vitro transcription of purified plasmid rDNA with a minimal E. coli RNA system yielded a pre-rRNA independently capable of autoexcision and autocyclization of the IVS (Kruger et al. 1982). This seminal publication also coined the term “” for RNA with enzymatic properties. Shortly thereafter,

6 Altman’s group definitively showed that RNA was the catalyitic subunit of RNase P

(Guerrier-Takada et al. 1983). For these discoveries, Thomas Cech and Sidney Altman were awarded the Nobel Prize in Chemistry in 1989.

Self-cleaving ribozymes are a subset of short ribozymes that achieve RNA cleavage and ligation at specific sites. Examples of such ribozymes include: hammerhead (Prody et al. 1986), hairpin (Buzayan et al. 1986), hepatitis delta virus (HDV) (Sharmeen et al. 1988),

Varkud satellite (VS) (Saville and Collins 1990), glmS (Winkler et al. 2004), and twister

(Roth et al. 2014) ribozymes. Though it was initially thought that all ribozymes were metalloenzymes that required coordination of divalent metal ions to achieve catalysis, analysis of the hammerhead, hairpin, and VS ribozymes in the absence of metal ions revealed that RNA cleavage rates were unaffected (Murray et al. 1998). Instead, self- cleaving ribozymes employ a site-specific internal transesterfication reaction (Das and

Piccirilli 2005), which is similar to the mechanism of intrinsic RNA cleavage described above and yields final products with 2´,3´-cyclic phosphate and 5´-OH termini.

Metal ion-independent endoribonucleases

Endoribonucleases can be divided into two distinct categories based on their reliance on divalent cations, their mechanism of action, and the terminal chemistry of the resulting

RNA cleavage fragments (Figure 1.1) (Yang 2011). In general, metal ion-dependent endoribonucleases use a conserved aspartic acid residue to coordinate two divalent metal cations, usually Mg2+, and the non-bridging oxygens of the scissile phosphate. One of these divalent metal cations facilitates the deprotonation of a water molecule, creating a hydroxide that acts as nucleophile to attack the phosphate. The 3´-O of the 5´-ribose acts

7 as the leaving group and is deprotonated by a general acid. The cleavage occurs 5´ of the scissile phosphate and yields a 5´ RNA fragment with a 3´-OH terminus and a 3´ RNA fragment with a 5´-PO4 terminus (Figure 1.1A) (Steitz and Steitz 1993). Because generation and attack of the nucleophile is relatively independent of nucleic acid conformation, metal ion-dependent endoribonucleases are able to cleave ssRNA, dsRNA, and DNA:RNA hybrids. Endoribonucleases in this category include RNases H, P, and Z, as well as Dicer and Drosha, and are relegated to a minor role in this document (Yang

2011).

In contrast, metal ion-independent endoribonucleases coordinate amino acids to promote acid-base catalysis. For instance, residue His12 of RNase A deprotonates the

2´-O creating the attacking nucleophile. His119 serves as a general acid catalyst and protonates the 5´-O leaving group. The chemistry of metal ion-independent endoribonucleases ensures a break 3´ of the scissile phosphate and generates a 5´ RNA fragment with a 2´,3´-cyclic phosphate terminus and a 3´ RNA fragment with a 5´-OH terminus (Figure 1.1B) (Raines 1998; Fedor and Williamson 2005). Metal ion-independent endoribonucleases require accessibility of the 2´-OH, and therefore almost exclusively cleave ssRNA.

Though mammals have evolved over a dozen metal ion-dependent endoribonucleases, I focus on several well-characterized endoribonucleases that are conserved from yeast through humans (Table 1.1). In the next section, I describe regulatory pathways involving four metal ion-independent (Sen2, Sen34,

Rny1/Angiogenin, and Ire1). The role of Las1 in rRNA biogenesis is explored in the

“Kinase-Mediated Decay” section.

8 5’-RNA 5’-RNA A Cytosine Cytosine O O

5’ Fragment O O OH HO OH Mg2+ P -O O HO - Adenine O :- O HO P Mg2+ Asp O O - H O OH Adenine :B O 3’-RNA 3’ Fragment O OH

3’-RNA 5’-RNA Cytosine O B 5’-RNA 5’-RNA O OH Cytosine Cytosine 5’-RNA O O Cytosine O 5’ Fragment O P O-

O O O O OH O O OH 5’-RNA P P Cytosine P :B- O O O- :OH- O O- -O O Adenine O HO HO O + Adenine A--H O - 3’ Fragment O P O O OH OH 3’-RNA O OH 3’-RNA

Figure 1.1. Two distinct mechanisms of endoribonuclease cleavage. A. Metal ion-dependent endoribonucleases (e.g., RNase H, RNase P, RNase Z) coordinate a metal ion to initiate cleavage of the scissile phosphate in ssRNA, dsRNA, and DNA:RNA hybrids. Cleavage generates a 5´-fragment with a 3´-OH terminus and a 3´-fragment with a 5´-PO4 terminus. B. Metal ion-independent endoribonucleases (e.g., Sen2, Sen34, Ire1, Rny1) use amino acid residues to promote catalysis in the absence of a metal ion. Cleavage generates a 5´- fragment with a 2´,3´-cyclic phosphate terminus and a 3´-fragment with a 5´-OH terminus. Figure used with permission from (Cooper, 2015b).

9 5´-OH RNAs in Biological Systems

Although a transcriptome-wide identification of 5´-OH RNAs via a direct sequencing method had never been performed prior to my work, there are numerous well- studied examples of 5´-OH RNAs in regulated pathways of RNA cleavage. In this section,

I discuss the generation of 5´-OH RNAs in three pertinent examples: processing of intron- containing tRNAs by the tRNA splicing endonuclease (TSEN), cleavage of tRNAs by

Rny1/Angiogenin to yield stress-induced tRNA fragments (tiRNAs), and the cleavage of

HAC1/XBP1 mRNA by inositol-requiring 1 (IRE1). The regulated decay of these fragments are further described in the “Kinase-Mediated Decay” section.

TSEN: Processing of intron-containing tRNAs

Of the predicted 295 tRNAs encoded in the yeast genome, 59 tRNAs representing

10 codons contain an intron that must be removed following transcription (Lowe and Eddy

1997). The first intron-containing tRNA in yeast was discovered in 1977 when DNA sequencing of SUP4 revealed a 14 bp intervening sequence adjacent to the 3´ end of the anticodon triplet in tRNATyr (Goodman et al. 1977). An intron of ~18 bp was found 3´ of the anticodon loop in tRNAPhe the following year (Valenzuela et al. 1978). Though the functional purpose of tRNA introns remains mysterious, the processing of intron- containing tRNAs is now well-understood (Figure 1.2A).

The first step in maturation of intron-containing tRNAs is processing by the tRNA

Splicing Endonuclease (TSEN). Intron-containing tRNAs are transcribed in the nucleus, then exported to the cytoplasm where they are cleaved by the TSEN complex to remove the intron. Sen2 cleaves on the 5´ end of the intron and Sen34 cleaves on the 3´ end of the

10 Table 1.1 Functions of several eukaryotic metal ion-independent endoribonucleases.

Protein Name Function References Yeast Human Sen2 TSEN2 Processing of intron- (Trotta et al. 1997) Sen34 TSEN34 containing tRNAs Stress-induced tRNA (Yamasaki et al. 2009; Rny1* Angiogenin* anticodon loop cleavage Thompson and Parker 2009b) UPR (yeast/humans); (Cox and Walter 1996; Ire1 IRE1α RIDD (humans) Hollien et al. 2009) ITS2 cleavage in rRNA (Schillewaert et al. 2012; Las1 LAS1L processing Butterfield et al. 2014)

UPR: Unfolded Protein Response; RIDD: Regulated IRE1-Dependent Decay of mRNA; ITS2: Internal Transcribed Spacer 2 * Rny1 and Angiogenin are not homologous but perform an analogous function. See text in the “5´-OH RNAs in Biological Systems” section for additional explanation.

11 intron (Ho et al. 1990). The enzymes generate a 5´ fragment with a 2´,3´-cyclic phosphate terminus and a 3´ fragment with 5´-OH terminus (Peebles et al. 1983; Knapp et al. 1979).

T-RNA Ligase 1 (Trl1) – also known as RNA Ligase 1 (Rlg1) – resolves the termini and ligates the two tRNA exons. Trl1 was first purified and shown to perform tRNA splicing in 1986 (Phizicky et al. 1986). The specific molecular mechanism was worked out in the late 1980s and early 1990s via a series of experiments primarily from John Abelson’s and Craig Greer’s groups. In brief, Trl1 ligates the 5´ and 3´ tRNA halves via key three enzymatic activities (Greer et al. 1983; Phizicky et al. 1986), each performed by a separate modular domain: (i) hydrolysis of the 2´,3´-cyclic phosphate terminus of the 5´ tRNA half to a 2´-phosphate terminus by the 2´,3´-cyclic phosphodiesterase (CPDase) domain; (ii) phosphorylation of the 5´-OH terminus of the 3´ tRNA half by the NTP-dependent polyribonucleotide kinase domain; and (iii) adenylation of the 5´-phosphate terminus to form an RNA-adenylate intermediate which is attacked by the 3´-OH terminus to ligate the ends via a phosphodiester bond (Xu et al. 1990; Apostol et al. 1991). The three domains

(CPDase, kinase, and adenylyltransferase/ligase) are all essential for tRNA splicing and yeast viability, but can function independently and perhaps arose from fusion events

(Sawaya 2003). The ligation produces a 2´-phosphate at the splice junction. For the tRNA to be functional, this phosphate must be removed by the essential 2´-

Tpt1, which transfers the 2´-phosphate to NAD+, forming ADP-ribose 1´,2´-cyclic phosphate and nicotinamide (Culver et al. 1993; 1997; Spinelli et al. 1997) (Figure 1.2A).

Following splicing, some tRNAs also undergo retrograde transport to the nucleus where they are further modified (Shaheen and Hopper 2005; Takano 2005). Decay of the tRNA

12 intron, which also carries 5´-OH and 2´,3´-cyclic phosphate termini is further described in the “Kinase-Mediated Decay” section (Wu and Hopper 2014).

In humans, intron-containing tRNAs are cleaved by conserved TSEN complex members TSEN2 and TSEN34, but the resulting 3´ and 5´ tRNA exons are ligated via a different mechanism than yeast. It had been speculated since the 1970s that the tRNA exons were joined by a direct ligation pathway that differed from the yeast system described above; however, the responsible enzyme was not identified until 2011. Popow et. al followed the ligation of a 3´-PO4 dsRNA over three chromatographic purification steps and identified 91 potential by MS/MS (Popow et al. 2011). One of the proteins,

HSPC117, is the human homolog of E. coli RtcB. Due to its genetic proximity to RNA 3´- terminal cyclase RtcA, the E. coli RtcB was thought to be an RNA repair enzyme

(Genschik et al. 1998; Galperin and Koonin 2004). Depletion of human RTCB by RNAi reduced ligation of tRNA halves in vivo, and mutation of a conserved cytosine residue for metal ion coordination (Cys122→Ala122) abolished tRNA ligation in vitro, thus demonstrating that RTCB is the mammalian tRNA ligase (Popow et al. 2011). (Additional functions of RTCB and consequences of mutation are discussed in the “IRE1α: non- canonical splicing of XBP1 mRNA and RIDD” section.)

Almost simultaneously, two groups showed that archaeal (Pyrobaculum aerophilum) RtcB and bacterial (E. coli) RtcB also perform ligation of 2´,3´-cyclic phosphate RNAs and 5´-OH RNAs (Englert et al. 2011; Tanaka and Shuman 2011). Further biochemical studies by Shuman’s group demonstrated RtcB joins the ends by a multi-step mechanism: i) resolution of the 2´,3´-cyclic phosphate using cyclic phosphate diesterase

(CPDase) activity, ii) formation of a covalent RtcB-GMP intermediate on His337, iii)

13 guanylation of the 3´-PO4 RNA terminus to form a (3′)pp(5′)G intermediate, and iv) attack of the 5´-OH terminus on the (3′)pp(5′)G to form the ligation junction (Chakravarty et al.

2012; Chakravarty and Shuman 2012) (Figure 1.2B). The mechanism of RtcB represents a novel method of RNA ligation. I take advantage of the unique enzyme specificity of E. coli RtcB to ligate 5´-OH RNAs to a 3´-PO4 RNA/DNA oligonucleotide in the initial step of the 5OH-seq method (Chapter II).

Rny1/Angiogenin: Stress-induced tRNA fragments

A growing body of literature is dedicated to understanding the generation and consequences of tRNA cleavage (Review: Saikia and Hatzoglou 2015). Presently, two major categories of tRNA fragments have been demonstrated in eukaryotic cells: stress- induced tRNA fragments (tiRNAs) and tRNA-derived fragments (tRFs). This discussion will focus on tiRNAs, as they are the most well-studied tRNA fragments and have 5´-OH termini.

In both prokaryotes and eukaryotes, tRNAs are cleaved in the anticodon loop during specific cellular stress conditions to yield 3´ and 5´ halves termed tiRNAs. The 3´-tiRNAs are produced with 5´-OH termini, and the 5´-tiRNAs have 2´,3´-cyclic phosphate termini.

Studies in E. coli were the first to identify a mechanism for in vivo tiRNA generation. In response to bacteriophage T4 infection, the PrrC nuclease cleaves tRNALys in the anticodon loop (Levitz et al. 1990). Another study showed that colicin E5 cleaved 3´ of a queuine wobble base in tRNAs for Tyr, His, Asn, and Asp (Ogawa et al. 1999). In contrast to eukaryotic tRNA anticodon loop cleavages, which affect a minority of the tRNA pool, the

14

tron-containing tron-containing tRNAs. Figure 1.2. Yeast andof1.2. human Yeast forin pathways Figure processing

15 Figure 1.2. Yeast and human pathways for processing of intron-containing tRNAs. A. Yeast pathway for cleavage and ligation. (1) The TSEN complex cleaves intron- containing tRNAs, yielding a 5´-half with a 2´,3´-cyclic phosphate terminus and a 3´-half with a 5´-OH terminus. (2) Trl1 resolves the 2´,3´-cyclic phosphate terminus to a 2´-PO4, (3) phosphorylates the 5´-OH terminus, and (4) adenylates the new 5´-PO4 terminus. (5) The 3´-OH attacks the adenylated intermediate. (6) Tpt1 removes the 2´-PO4 at the splicing junction. B. Human pathway for cleavage and ligation. (1) The TSEN complex cleaves intron- containing tRNAs as in A. (2) RTCB resolves the 2´,3´-cyclic phosphate terminus to a 3´- PO4, and (3) guanylates the new 3´-PO4 terminus. (4) The 5´-OH terminus attacks the guanylated intermediate.

16 prokaryotic stress response shuts down translation by cleaving a large percentage of the pool (Thompson and Parker 2009a).

Although tRNA fragments had been identified in the the urine and sera of cancer patients in the late 1970s, evidence of in vivo eukaryotic tRNA cleavage came much later.

The first study demonstrating in vivo eukaryotic tRNA cleavage showed that starvation by amino acid deprivation of Tetrahymena induced anticodon loop cleavage in asparagine, glycine, and threonine tRNAs (Lee and Collins 2005). In 2009, RNase T2 family member

Rny1 was identified as the endoribonuclease responsible for tiRNA generation in yeast under oxidative stress conditions (Thompson and Parker 2009b). Yeast tiRNAs are also produced in numerous other stress conditions, including: nitrogen starvation, methionine starvation, heat shock, and entrance into stationary phase (Thompson et al. 2008). Rny1 is normally sequestered in the vacuole or secreted, and is only released into the cytoplasm in the presence of cellular stressors (Thompson and Parker 2009b) (Figure 1.3A).

In mammals, Angiogenin is an endoribonuclease that cleaves tRNA anticodon loops under stress conditions. Though Angiogenin shares this common function with Rny1, it is not a homolog, and the mammalian Rny1 homolog RNASET2 does not cleave tRNAs

(Yamasaki et al. 2009). Angiogenin is a secreted ribonuclease of the RNase A superfamily that was initially discovered as an angiogenic factor in conditioned medium of HT-29 colon adenocarcinoma cells (Fett et al. 1985). While it was known that Angiogenin hydrolyzes tRNAs when injected into Xenopus oocytes (Saxena et al. 1992), it was not until 2009 that

Angiogenin was shown to generate tiRNAs in human fetus hepatic tissue and human osteosarcoma cells (U2OS) (Fu et al. 2009; Yamasaki et al. 2009). To become activated,

Angiogenenin must translocate from the nucleus to the cytoplasm and dissociate from

17 Ribonuclease/Angiogenin Inhibitor 1 (RNH1) (Thompson and Parker 2009a). Once active in the cytoplasm, Angiogenin cleaves tRNAs near the anti-codon loop to yield 3´ and 5´ tiRNAs (Figure 1.3B). Angiogenin is downregulated in several neurodegenerative diseases, and loss of function mutations have been identified in patients with amyotrophic lateral sclerosis (ALS) (Greenway et al. 2004), and Parkinson’s disease (PD) (van Es et al.

2011). However, Angiogenin can also promote cellular growth and is upregulated in many solid tumors including the colorectal, prostate, and breast cancers (Shimoyama et al. 1999;

Katona et al. 2005; Montero et al. 1998).

Several studies have implicated tiRNAs in inhibition of both translation and apoptosis. Transfection of 5´-tiRNAs, but not 3´-tiRNAs, promotes stress granule formation and moderately inhibits protein synthesis, independent of eIF2α phosphorylation state (Emara et al. 2010). Synthetic 5´-tiRNAs can also displace eIF4G/A from mRNAs to inhibit translation in rabbit reticulocyte lysate (Ivanov et al. 2011). tiRNAs can saturate the Dicer binding pocket, decreasing cleavage of dsRNA substrates (Durdevic et al. 2013). Finally, both 5´-tiRNAs and 3´-tiRNAs can bind cytochrome c, competitively inhibiting the binding of apoptotic protease-activating factor 1 (APAF1), thus suppressing apoptosis (Saikia et al. 2014).

Ire1: non-canonical splicing of HAC1 mRNA

The unfolded protein response (UPR) is a conserved eukaryotic intracellular signaling pathway activated in response to endoplasmic reticulum (ER) stress (Walter and

Ron 2011; Moore and Hollien 2012). When the protein folding load exceeds ER capacity,

18

Figure 1.3. Rny1 (yeast) and Angiogenin (humans) cleave the anticodon loops of tRNAs under stress conditions. A. In yeast, Rny1 is normally sequestered in the vacuole or secreted. Following certain stress conditions, Rny1 is released to the cytoplasm. Rny1 cleaves tRNAs in the anticodon loop, generating a 5´-tiRNA with a 2´,3´-cyclic phosphate terminus and 3´-tiRNA with a 5´-OH terminus. Does Rny1 target other RNAs? B. In humans, Angiogenin (Ang) is usually localized to the nucleus or secreted. During cellular stress, Ang enters the cytoplasm and dissociates from Ribonuclease/Angiogenin Inhibitor 1 (RNH1). Ang generates the same tiRNAs as Rny1 in A. These tiRNAs may interact with several downstream pathways. Does Angiogenin target other RNAs?

19 the UPR is triggered, resulting in upregulation of protein-folding machinery (Kozutsumi et al. 1988). The net result is an increase in the size and folding-capacity of the ER and a resolution of the unfolded protein burden (Schuck et al. 2009). The UPR is naturally triggered by a multitude of environmental stimuli, diseases, and developmental processes including: oxidative stress; viral infection (Bechill et al. 2006; 2008); inherited diseases of protein misfolding, such as cystic fibrosis and retinitis pigmentosa; and the development of antibody-secreting plasma cells (Gass et al. 2002; Iwakoshi et al. 2003) or pancreatic ß- cells (Harding et al. 2001; Zhang et al. 2002). The UPR is commonly chemically induced by tunicamycin, an inhibitor of N-linked glycosylation of proteins, which decreases protein stability. Dithiothreitol (DTT), which reduces disulfide bonds and thus results in decreased protein stability, can also be used to induce the UPR.

In budding yeast, the UPR is mediated by a single pathway controlling the non- canonical cytosolic splicing of HAC1 mRNA, which encodes a transcription factor that resolves ER stress by upregulating key protein folding machinery and chaperones. This unexpected and unprecedented means of RNA regulation was revealed by a series of elegant experiments conducted independently by Peter Walter at UCSF and Kazutoshi

Mori at UT Southwestern and Kyoto University in the early 1990s, for which they were jointly awarded the Lasker Award in 2012.

Several early studies suggested the existence of a regulated system for increasing protein expression under specific cellular stress conditions. The first of these studies showed that the mRNA levels of two glucose regulated proteins (GRPs) p4A3 and p3C5 increased under conditions of glucose starvation in the hamster K12 cell line (Lee et al.

1983). While it was initially thought that induction of GRPs was due to defective

20 glycosylation caused by glucose depletion, induction of GRPs by amino acid analogues argued against this theory (Kelley and Schlesinger 1978). Impaired protein folding was also shown to regulate the expression of GRPs. In simian CV-1 cells, expression of a misfolded influenza virus haemagglutinin (HA) but not wildtype HA induced GRP78 and

GRP94/HSP90 (Kozutsumi et al. 1988). GRP78 was later renamed BiP and is encoded by the KAR2 gene in yeast (Normington et al. 1989).

The discovery of Ire1 and Hac1 in the molecular regulation of the UPR was spurred by the identification of a 22 nt conserved sequence in the KAR2 promoter. This conserved

UPR element (UPRE) is required for induction of yeast BiP in response to tunicamycin

(Mori et al. 1992; Kohno et al. 1993). Independently, Walter and Mori expressed the UPRE upstream of the E. coli ß-galactosidase gene (lacZ) creating a colorimetric system for screening yeast genomic libraries and ultimately identifying Ire1/Ern1, a putative (Nikawa and Yamashita 1992), as a required gene for upregulation of KAR2 expression following treatment with tunicamycin (Cox et al. 1993; Mori et al. 1993).

Further studies revealed Ire1 spans the ER membrane due to an N-terminal ER signal adjacent to a hydrophobic region and contains a C-terminal cytosolic kinase domain whose enzymatic activity is essential for complementation of temperature sensitive ern1-1 mutants (Mori et al. 1993). The initial presumption was that Ire1 would phosphorylate and activate a transcription factor that would subsequently recognize the UPREs upstream of target genes. Indeed, using an adapted yeast in vivo one-hybrid screening system (Mori et al. 1996) or an overexpression system (Cox and Walter 1996), Mori and Walter each identified Hac1/Ern4 as the transcription factor responsible for binding the UPRE and upregulating expression of target genes.

21 Peter Walter’s group then proceeded to make the revolutionary discovery that Hac1 signaling was not regulated by phosphorylation, but by non-canonical cytosolic mRNA splicing, publishing two essential studies in the same issue of Cell. In one publication, they showed that HA-Hac1 was undetectable by immunofluorescence with an α–HA antibody under normal growth conditions, but was detected in the nucleus when cells were induced with tunicamycin. Northern analysis demonstrated that the HAC1 mRNA shortened in an

Ire1-dependent manner following treatment with tunicamycin, and sequencing revealed the excision of a 252 nt intervening sequence (IVS) and splice junctions of the mRNA that differed from consensus recognition sequences (Cox and Walter 1996). The second study used a sectoring screen in a kar2-∆HDEL yeast stain to identify Trl1/Rlg1 as the HAC1 mRNA ligase (Sidrauski et al. 1996). The following year, Walter’s group demonstrated that Ire1 cleaves at a conserved 5´-CN|CNNG-3´ motif in stem loops at the

3´ and 5´ splice junctions (Sidrauski and Walter 1997).

The series of experiments described above unveiled the first and only known example of a signaling pathway that is regulated by the cytosolic splicing of an mRNA.

Ire1 oligomerizes upon sensing unfolded proteins via its luminal domain (Gardner and

Walter 2011). This oligomerization promotes the activation of the Ire1 cytosolic endoribonuclease domain. When the Ire1 endoribonuclease domain is activated, it cleaves

HAC1 mRNA at two locations on either side of the excised intron at sequence-specific recognition motifs (Kawahara et al. 1998), generating fragments with 2´,3´-cyclic phosphate and 5´-OH termini (Gonzalez et al. 1999). Non-canonical cytosolic splicing of the exons is then carried out by tRNA ligase Trl1 via the three distinct enzymatic activities previously described in “TSEN: Processing of intron-containing tRNAs.” Following

22 removal of this phosphate by Tpt1 (Spinelli et al. 1997), spliced HAC1 mRNA is translated, yielding a functional transcription factor that upregulates approximately 7% of yeast genes to modulate the ER stress response (Travers et al. 2000), including key UPR markers

KAR2, EUG1 (Tachibana and Stevens 1992), and PDI1 (Cox et al. 1993). While the UPR is non-essential for growth in the absence of cellular stress, HAC1 and IRE1 null mutants grow slowly compared to wild type yeast and both are required under UPR-inducing conditions (Figure 1.4).

Key publications in the 2000s further extended this complex model of HAC1 mRNA splicing and resolved two unanswered questions by attributing regulatory significance to both the intron and the 5´- and 3´-UTRs. Previous experiments had revealed curious findings of unspliced HAC1 mRNA (HAC1U): In situ hybridization experiments localized HAC1U to the cytoplasm (Chapman and Walter 1997), and large portions of

HAC1U comigrated with polyribosomes on sucrose gradients (Cox and Walter 1996). This suggested that Hac1U is partially translated by ribosomes that subsequently stall on the

HAC1U mRNA. These findings were resolved by demonstrating that translational attenuation is mediated by a long-range 16 nucleotide base-pairing interaction between the intron and the 5’ UTR (Rüegsegger et al. 2001). There still remained the question of what might promote the colocalization of HAC1 mRNA and Ire1 prior to cleavage. Mutational probing of the HAC1 3´-UTR revealed a prominent, extended stem-loop with a conserved bipartite element (3´ BE) which targets HAC1 mRNA to Ire1 foci to promote cleavage during ER stress (Aragón et al. 2008) (Figure 1.4).

Given these latest examples of UPR regulation by the mRNA transcript itself, I was particularly curious about the findings of a recent study that captured HAC1 3´-exons with

23 5´-PO4 termini in ∆ xrn1∆ yeast (these genes are discussed in the “RNA Decay

Pathways” section) (Harigaya and Parker 2012). The existence of HAC1 3´-exons with

5´-PO4 termini is perplexing because HAC1 mRNA is cleaved by metal ion-independent endoribonuclease Ire1, and the 3´-exon should therefore be generated with a 5´-OH terminus (Figure 1.4). If the HAC1 3´-exon can exist with a 5´-PO4 terminus, it must be phosphorylated after Ire1 cleavage. Furthermore, because the HAC1 3´-exon with a 5´-PO4 terminus accumulates when Xrn1 is deleted, Xrn1 must be required for the degradation of this exon. In Chapter II, I use 5OH-seq to replicate the capture of HAC1 3´-exons with

5´-PO4 termini in xrn1∆ yeast. This finding sets up the experimentation in Chapter III, where I investigate the phosphorylation and decay of the HAC1 3´-exon intermediate.

IREα: non-canonical splicing of XBP1 mRNA and RIDD

In contrast to the single UPR pathway found in yeast, the human UPR is mediated via three branches: PERK, ATF6, and IREα (Walter and Ron 2011). (This discussion will be limited to the IRE1α pathway, as it is the only pathway conserved with yeast and the only pathway regulated by mRNA splicing.) Similar to yeast, the IREα branch signals via the non-canonical cytosolic splicing of XBP1 mRNA, the human ortholog of yeast HAC1.

In contrast to yeast HAC1 mRNA, XBP1 mRNA is targeted to the ER membrane by the nascent peptide of the partially translated Xbp1 protein (Yanagitani et al. 2009; 2011), and the two conserved recognition motifs that are cleaved by IRE1α excise a 26 nt intron following UPR activation (Figure 1.5).

Until 2014, there was debate as to the primary mammalian pathways for XBP1 mRNA ligation following IRE1α-mediated cleavage: a yeast-like pathway and a direct-

24

Figure 1.4. The unfolded protein response (UPR) in yeast is primarily regulated by the non-canonical cytosolic splicing of HAC1 mRNA. (1) HAC1 mRNA is targeted to the ER membrane by a bipartite element (BE) in the 3´- UTR. (2) Translation of HAC1 is attenuated by long-range base-pairing of the intron and the 5´-UTR. (3) During ER stress, unfolded proteins are sensed by Ire1, which promotes oligomerization, autophosphorylation, and activation of the endoribonuclease domain. (4) Active Ire1 cleaves HAC1 mRNA in two conserved stem loops, generating a 5´-exon with a 2´,3´-cyclic phosphate terminus and 3´-exon with a 5´-OH terminus. (5) Trl1 ligates the exons, and (6) Tpt1 removes a 2´-PO4 at the splicing junction. (7) Spliced HAC1 mRNA is translated to Hac1 protein, (8) which translocates to the nucleus, binds UPR elements (UPREs), and upregulates expression of ~7% of the yeast genome. Inset: (?) Ire1 cleavage produces HAC1 3´-exon mRNA intermediates with 5´-OH termini. In dcp2∆ xrn1∆ yeast strains, these intermediates accumulate with 5´-PO4 termini. Does phosphorylation of the HAC1 3´-exon by Trl1 promote decay by Xrn1?

25 ligation pathway. In the hypothetical yeast-like pathway, the three enzymatic roles of yeast

Trl1 would be carried out by analogous but non-homologous human counterparts: CNP, a cyclic phosphodiesterase; CLP1, a polynucleotide kinase (discussed in further detail in the

“Kinase-Mediated Decay” section); and an unknown ligase. TRPT1, a phosphodiesterase homologous to the yeast Tpt1, would be a likely candidate for 2´-PO4 removal. However,

CNP and TRPT1 are non-essential in mice, and TRPT1-null mice have intact XBP1 splicing (Harding et al. 2007). Rather than resolve the 2´,3´-cyclic phosphate and 5´-OH termini, ligate, and remove the 2´-PO4 via four enzymes, the direct ligation pathway would join these ends with a single enzyme (Schwer et al. 2008). As previously discussed in the

“5´-OH RNAs in Biological Systems” section, RTCB performs such a direct ligation during processing of intron-containing tRNAs in humans (Popow et al. 2011). RTCB can also ligate Ire1-cleaved HAC1 mRNA in a yeast trl1∆ strain (Tanaka et al. 2011). However, a report demonstrating that XBP1 mRNA ligation was not impaired when RTCB was depleted by RNAi in HeLa cells (Iwawaki and Tokuda 2011) argued against RTCB as the ligase for XBP1 mRNA.

In late 2014, a series of three publications revealed that RTCB is the mammalian

UPR ligase (Lu et al. 2014; Kosmaczewski et al. 2014; Jurkin et al. 2014) (Figure 1.5).

The first created a synthetic circuit wherein an XBP1 splicing-dependent 4-OHT-regulated

Cre sensor activates Bim expression to generate a robust apoptotic response in SNL cells

(Lu et al. 2014). This clever biological engineering approach enabled an unbiased screen that identified RTCB as the requisite XBP1 mRNA ligase. Kosmaczewski et al. bypassed the essential requirement of RtcB in tRNA splicing in C. elegans by expressing pre-spliced tRNAs for Leu(CAA) and Tyr(GUA) in vivo. They were able to demonstrate impaired XBP1

26 mRNA splicing of endogenous intron-containing tRNAs in RtcB null worms

(Kosmaczewski et al. 2014). (We employ a similar strategy in Chapter III to evaluate the requirement of Trl1 in the phosphorylation of the 3´ cleavage product of HAC1 mRNA in the yeast UPR.) The final publication showed that simultaneous depletion of RTCB and its Archease by Dox-inducible shRNAs abolishes XBP1 mRNA splicing in HeLa cells (Jurkin et al. 2014).

Recent studies have linked RTCB ligase activity to several essential processes. The

RTCB-mediated splicing of XBP1 is required for secretion of antibodies in human plasma cells (Jurkin et al. 2014). In the orthologous C. elegans system, RTCB-1 splicing of xbp-1 protects neurons from degeneration induced by human α–synuclein, which is a hallmark of Parkinson’s disease (Ray et al. 2014). RtcB also localizes to the site of neuronal injuries in C. elegans and inhibits axon regeneration independent of xbp-1 ligation, tRNA ligation, and cofactor Archease (Kosmaczewski et al. 2015).

In many eukaryotes, activated IRE1 broadly cleaves mRNA in a process termed

Regulated Ire1-Dependent Decay (RIDD) (Hollien and Weissman 2006) (Figure 1.5).

During RIDD, Ire1 cleaves mRNAs that are colocalized to the ER, generating RNA fragments with 2´,3´-cyclic phosphate and 5´-OH termini. RIDD was first identified by microarray studies of Drosophila S2 cells depleted of either IRE1 or XBP-1 by RNAi and induced with DTT. A population of mRNAs were differentially repressed in IRE1-depleted cells but not XBP-1-depleted cells, and this was shown to be the result of mRNA cleavage performed by IRE1 that was dependent on co-translational ER-localization of the mRNA

(Hollien and Weissman 2006). Subsequent studies have extended RIDD to plants (Hayashi et al. 2012), fission yeast (Kimmig et al. 2012), and mammals (Hollien et al. 2009). While

27

Figure 1.5. Human IRE1α cleaves XBP1 and ER-proximal mRNAs. (1) The nascent peptide of partially translated XBP1 targets the mRNA to the ER. (2) During ER stress, IRE1α oligomerizes, promoting autophosphorylation and activation of the endonuclease domain. (3) Active IRE1α cleaves XBP1, generating a 5´-exon with a 2´,3´-cyclic phosphate terminus and a 3´-exon with a 5´-OH terminus. (4) RTCB ligates the exons. (5) Spliced XBP1 mRNA is translated to XBP1S protein and upregulates expression of protein chaperones and folding machinery. (6) Active IRE1α also cleaves ER-proximal mRNAs in a process termed Regulated IRE1-Dependent Decay (RIDD).

28 RIDD has not been demonstrated in budding yeast, the conservation of Ire1 and the

HAC1/XBP1 pathway in eukaryotes suggests a similar pathway may be present, and I investigate this further in Chapter II. Importantly, although the 3´ mRNA fragments have

5´-OH termini following cleavage by IRE1, the initial RIDD paper demonstrated that these species accumulate in XRN1-depleted cells (Hollien and Weissman 2006). This discrepancy is not addressed and RIDD-substrates are prime candidates for “kinase- mediated decay”, which is explained later in this chapter and discussed in Chapter III.

At the time of this dissertation, hundreds of publications have purported a UPR phenotype in human disease, comprising an extensive breadth of syndromes. In 1999, a group in Osaka, Japan was the first to implicate the UPR in a specific disease, early-onset familial Alzheimer’s disease (FAD). Their study demonstrated that missense-mutations in the human presenilin-1 (PS1) gene – the most common autosomal dominant cause of early- onset FAD – increased sensitivity of neuroblastoma cells to tunicamycin and inhibited induction of GRP78/BIP mRNA (Katayama et al. 1999). Since this initial study, aberrant regulation of the UPR has been linked to multiple disease processes, including: i) protein misfolding (e.g., proinsulin 2 in ß-cell death and procollagen in osteogenesis imperfecta); ii) polyglutamate expansions, such as Machado-Joseph disease; iii) cancers, including myeloma, breast cancer, many others; and iv) neurodegenerative diseases, specifically

Parkinson’s disease, Alzheimer’s disease, and amyotrophic lateral sclerosis (Reviews:

(Kaufman 2002; Lee and Ozcan 2014)). While description of the causes and implications of UPR misregulation in each of these diseases is beyond the scope of this thesis, such extensive connections between UPR and disease demonstrates the paramount importance of understanding the molecular mechanisms modulating this ubiquitous stress response.

29 RNA Decay Pathways

Within this section, I provide an overview of RNA decay. I outline the known primary cytosolic mRNA degradation pathways of Xrn1 and the exosome. I also describe several degradation pathways for defective mRNAs: nonsense-mediated decay (NMD), non-stop decay (NSD), and no-go decay (NGD). Finally, I introduce previously unaddressed questions of 5´-OH RNA decay.

5´→3´ degradation by 5´-phosphate-dependent exoribonucleases

The constant production of mRNAs necessitates robust mechanisms of regulated mRNA turnover. In the late 1970s, studies of the 5´ mRNA cap structure in Xenopus, wheat germ, and mouse L cells suggested that a 5´→3´ decay process might hydrolyze uncapped mRNAs (Furuichi et al. 1977; Shimotohno et al. 1977). Shortly thereafter, Audrey Stevens purified an exoribonuclease from yeast and demonstrated that it performed 5´ hydrolysis of radiolabelled RNA with a 5´-PO4 termini (Stevens 1980). This protein was named eXoRiboNuclease 1 (Xrn1) (Larimer and Stevens 1990), and subsequently shown to degrade mRNA by processive 5´-phosphate–dependent 5´→3´ hydrolysis (Hsu and

Stevens 1993; Muhlrad et al. 1994; Stevens 2001).

Most 5´-capped mRNAs are deadenylated and decapped prior to degradation. The first step of decay is deadenylation by the Ccr4/Pop2/Not complex (Brown and Sachs 1998;

Tucker et al. 2001) or a secondary Pan2/Pan3 complex (Boeck et al. 1996). Deadenylation is regulated by several RNA binding proteins (RBPs), particularly the poly-A binding protein (Pab1), which binds the polyadenylated mRNA and inhibits the Ccr4 deadenylase

(Caponigro and Parker 1995). Decapping proteins 1 and 2 (Dcp1/Dcp2) subsequently

30 remove the 7-methylguanylate caps of the deadenylated transcripts (van Dijk et al. 2002); however, decapping proceeds independently of deadenylation for a subset of mRNA transcripts (Muhlrad and Parker 2005). As with deadenylation, decapping is a highly regulated process, and likely predominantly regulated by Pat1/Lsm1-7 mediated recruitment of Dcp1/Dcp2 (Review: Parker 2012).

The final product of most mRNA decapping events is a deadenylated mRNA with a 5´-PO4 terminus that Xrn1 recognizes and degrades (Figure 1.6). Xrn1 is non-essential in yeast, presumably due to secondary pathways of degradation; however, xrn1∆ strains have a decreased rate of mRNA decay, which varies between a 1.5 to 6-fold reduction depending on the mRNA transcript (Cao and Parker 2001). In contrast, Xrn1/Pacman knockdowns in higher-order eukaryotes C. elegans and Drosophila confer either developmental failures or lethality, confirming that 3´→5´ decay pathways cannot compensate for defective 5´→3´ decay (Newbury and Woollard 2004; Jones et al. 2012).

In addition to Xrn1, there are two other known 5´-phosphate-dependent 5´→3´ exoribonucleases in yeast. The most well-studied of these is Rat1, known as Xrn2 in higher-order eukaryotes. The essential RAT1 gene was initially identified in a screen for defective poly(A) mRNA transport from the nucleus (thus its name: Ribonucleic Acid

Trafficking 1), and encodes an exoribonuclease similar to Xrn1 (Amberg et al. 1992;

Kenna et al. 1993). Rat1 is predominantly expressed in the nucleus, and overexpression of wild-type RAT1 can only partially restore growth defects in xrn1∆ yeast (Poole and Stevens

1995). However, expression of RAT1 with mutations that increase cytosolic localization can fully complement xrn1∆ yeast. Similarly, Xrn1 is primarily located in the cytoplasm, but XRN1 tagged with a nuclear localization signal (NLS) can rescue temperature-sensitive

31 rat1-1 yeast (Johnson 1997). Thus, Xrn1 and Rat1 are functionally interchangeable exoribonucleases, but are restricted to the cytoplasm and nucleus, respectively. Rat1 is required for pre-rRNA trimming (Henry et al. 1994) (described in the “Kinase-Mediated

Decay” section), snoRNA processing (Petfalski et al. 1998), and mRNA transcription termination (Kim et al. 2004; Ghazal et al. 2009).

Functional studies of a Rai1 homolog Ydr3706 recently revealed a third 5´- phosphate-dependent 5´→3´ exoribonuclease in yeast. Though initially evaluating the effects of Ydr3706 on decapping, a group at Columbia discovered the protein had distributed 5´→3´ exoribonuclease activity and proposed the new name Decapping 1 (Dxo1). The study showed that Dxo1 can degrade 5´-PO4 RNA, but cannot proceed through a RNA-DNA heteroduplex, suggesting that Dxo1 is weaker than Xrn1 and more prone to stalling (Chang et al. 2012). The mammalian homolog DXO is restricted to the nucleus and acts as a 5´ end quality control mechanism to detect and degrade incompletely 5´ end-capped pre-mRNAs (Jiao et al. 2013); however, it is not known what role cytosolic yeast Dxo1 might play in cytosolic mRNA decay.

3´→5´ degradation by the exosome

A secondary pathway for RNA degradation following deadenylation is 3´→5´ degradation by the exosome (Figure 1.6). The exosome was first discovered when immunoprecipitation of Ribosomal RNA Processing 4 (Rrp4), a protein required for

3´→5´ exonuclease activity on 5.8S rRNA, revealed that it was part of a much larger complex (Mitchell et al. 1997). Mass spectrometry analysis enabled identification of four associated proteins – thereafter named Rrp41, Rrp42, Rrp43, and Rrp44 – each of which is

32 required for 5.8S rRNA 3´ end formation (Mitchell et al. 1997). Studies over the subsequent decade demonstrated that the exosome is a complex comprised of 10 key proteins: the exonuclease and endonuclease domain-containing Rrp44/Dis3 protein and a nine-member ring structure formed by six RNase PH proteins (Rrp41, Rrp42, Mtr3, Rrp43, Rrp46,

Rrp45) and three small RNA-binding proteins (Rrp4, Rrp40, and Csl4) analogous to bacterial PNPase (Allmang et al. 1999; Liu et al. 2006). While the core is present in both the nucleus and the cytoplasm, the exosome requires four Ski proteins

(Ski2, Ski3, Ski7, and Ski8) for cytoplasmic mRNA decay (Anderson and Parker 1998; van Hoof et al. 2000b). The Ski2 protein is an ATPase that may act as an RNA to unwind structured RNAs and remove bound proteins; a related protein, Mtr4, may perform the same activity when associated with the nuclear exosome (Lykke-Andersen et al. 2009).

Deadenylation occurs prior to 3´→5´ exosomal decay of mRNAs, possibly due to removal of factors bound to the poly(A) tail that would otherwise inhibit the exosome (Parker 2012).

In contrast to the 5´-PO4 terminus requirement of Xrn1, the effect of RNA 3´ terminal modifications on recognition and degradation by the exosome is unknown.

The components of the exosome are well-conserved in humans, and mutations can greatly compromise neurodevelopment. For example, a recent study that sequenced the exomes of patients with pontocerebellar hypoplasia type 1 (PCH1) revealed recessive mutations in EXOSC3, the human homolog of yeast RRP40 (Wan et al. 2012). A morpholino knockdown of exosc3 mRNA in zebrafish conferred maldevelopment consistent with the human phenotype, and expression of wildtype exosc3 mRNA restored normal development.

33

Figure 1.6. Canonical cytosolic mRNA decay pathways: Xrn1 and the exosome. (1) The first step in cytosolic mRNA decay is deadenylation by the Ccr4/Pop2/Not complex or the Pan2/Pan3 complex. (2) Dcp1 and Dcp2 are recruited to the 5´ terminus and Dcp2 7 removes the m G cap, leaving a 5´-PO4 terminus. (3) Xrn1-mediated degradation of decapped 5´-PO4 mRNA proceeds 5´→3´. The cytosolic exosome is comprised of a 9-member ring structure and the Rrp44 nuclease. (4) Degradation of deadenylated mRNA by the exosome proceeds 3´→5´. Adapted from (Parker 2012).

34 Decay pathways for defective mRNAs

In addition to the major Xrn1 and exosome decay pathways, three other minor pathways for mRNA quality control have been described in yeast: nonsense-mediated decay (NMD), non-stop decay (NSD), and no-go decay (NGD) (Figure 1.7). The most well-studied of these mRNA quality control mechanism, NMD, was first suggested in 1979 when amber nonsense mutations introduced into the yeast URA3 gene decreased mRNA stability (Losson and Lacroute 1979). Other than mRNAs with the nominal nonsense mutations, NMD acts on a broad array of substrates with anomalous translation termination, as well as a subset of “normal” mRNAs (Parker 2012). These defective mRNAs are marked by Upf1, Upf2, and Upf3 interaction and are subject to rapid deadenylation-independent decapping followed by 5´-phosphate-dependent degradation by Xrn1 (Baker and Parker 2004) (Figure 1.7A). NMD is conserved in Drosophila

(Gatfield et al. 2003) and mammals (Lejeune et al. 2003); however, metazoans have additional mechanisms for mediating NMD, including SMG6, an endonucleolytic component that cleaves mRNA at premature translation termination codons (PTCs)

(Huntzinger et al. 2008). SMG6-mediated cleavage allows cytosolic decay by XRN1 and the exosome to proceed from the site of PTC in the 5´→3´ and 3´→5´ directions, respectively (Stevens et al. 2002).

Nonstop decay (NSD) occurs when a translating ribosome reaches the 3´ end of an mRNA without reaching a stop codon (Figure 1.7B). The stalled ribosome is recycled by

Dom34/Hbs1 (discussed in more detail below) and triggers rapid 3´→5´ Ski-dependent degradation of the mRNA by the exosome (Frischmeyer 2002; van Hoof et al. 2002; Tsuboi et al. 2012). NSD of the mRNA occurs concomitantly with rapid proteasome-mediated

35 decay of the nascent peptide promoted by Ltn1 (Wilson et al. 2007) or Not4 (Dimitrova et al. 2009). NSD can be differentiated from canonical exosomal decay by its requirement of the GTPase domain of Ski7 (van Hoof et al. 2002). Though NSD is primarily a 3´→5´ decay process, decapping by Dcp1 and 5´→3´ decay by Xrn1 does play a minor role in degrading mRNAs without stop codons (Frischmeyer 2002).

No-go decay (NGD) refers to endonucleolytic cleavage of mRNA that occurs when a translating ribosome stalls, followed by degradation of the cleavage fragments (Figure

1.7C). NGD was first demonstrated in yeast expressing a PGK1 reporter with a stable stem- loop insertion in the coding region (PGK1-sL). Deletion of XRN1 leads to accumulation of fragments 3´ of the stem loop, and cytosolic exosome mutant strains ski2∆, ski3∆, and ski7∆ accrue fragments 5´ of the stem loop (Doma and Parker 2006). Rare codons, runs of polylysine and polyarginine, depurination sites, and possibly frameshift sites can also trigger NGD (Parker 2012).

Dom34 is required for the accumulation of NGD fragments for certain NGD triggers (Doma and Parker 2006), but is dispensable for others (Kuroha et al. 2010; Passos et al. 2009; Chen et al. 2010). Dom34 is structurally similar to translation termination factor eRF1 and interacts with Hbs1, a protein related to eRF3. A cryo-EM structure of the

Dom34-Hbs1 complex bound to a stalled ribosome showed that the N-terminal domain

(NTD) of Dom34 reached deep into the A-site (Becker et al. 2011). This association may disrupt the mRNA stabilizing network, leading to dissociation of the nascent peptide and rescue of the ribosome. Dom34 is thought to recycle ribosomes at the 3´-ends of truncated messages in both NSD and NGD (Tsuboi et al. 2012).

36

Figure 1.7. Decay pathways for defective mRNAs: nonsense-mediated decay (NMD), non-stop decay (NSD), and no-go decay (NGD).

37 Figure 1.7. Decay pathways for defective mRNAs: nonsense-mediated decay (NMD), non-stop decay (NSD), and no-go decay (NGD). A. Mechanism of nonsense-mediated decay (NMD). (1) A translating ribosome encounters a premature stop codon. (2) The mRNA is rapidly decapped by Dcp1/Dcp2. (3) Xrn1 performs rapid 5´→3´ degradation of the 5-PO4 mRNA. B. Mechanism of nonstop decay (NSD). (1) A translating ribosome reaches the end of an mRNA without encountering a stop codon. (2) Dom34/Hbs1 rescue the stalled ribosome, and (3) the truncated mRNA is rapidly degraded 3´→5´ by the exosome. C. Mechanism of no-go decay (NGD). (1) A translating ribosome stalls due to a strong secondary structure, polycharged bases, etc. (2) An unidentified endoribonuclease cleaves upstream of the stall, and (3) Dom34/Hbs1 rescue the stalled ribosome. (4) The exosome degrades the 5´-NGD fragment, and (5) Xrn1 degrades the 3´-NGD fragment. Inset: (?) Does the NGD cleavage produce a 3´-fragment with a 5´-PO4 or a 5´-OH that must be phosphorylated prior to decay? Adapted from (Parker 2012; Shoemaker and Green 2012)

38 Although targets of NGD have not been specifically identified in higher-order eukaryotes, the Dom34 homolog Pelota is highly conserved. Pelota is required for mitotic and meiotic cell division in Drosophila (Eberhart and Wasserman 1995), and homozygous

Pelota deficient mice fail to develop past embryogenesis day 7.5 (E7.5) (Adham et al.

Arg 2003). In mice with a CNS-specific tRNA UCU SNP, loss of Pelota binding partner

GTPBP2 induces ribosome stalling and neurodegeneration, indicating the potential importance of Pelota in mammalian NSD/NGD (Ishimura et al. 2014).

The endonuclease responsible for cleavage in NGD decay remains unknown, and this is a key gap in understanding NGD. Early reports suggested that Dom34 and the

Dom34 ortholog had RNA nuclease activity (Lee et al. 2007), but subsequent studies could not reproduce these results (Passos et al. 2009). Importantly, as the nuclease has still not been identified, the ends generated after cleavage remain ambiguous. If the cleavage is performed by a metal ion-dependent endonuclease, the 3´ fragment would have a 5´-PO4 terminus. However, if the cleavage is performed by a metal ion-independent endonuclease or in-line cleavage, the 3´ fragment would have a 5´-OH terminus. In this latter case of cleavage chemistry, the 3´ fragment with a 5´-OH terminus would require phosphorylation prior to decay by Xrn1.

5´→3´ decay pathways do not account for 5´-OH RNAs

The RNA decay field does not currently address the 5´→3´ degradation of 5´-OH

RNAs. Three known 5´→3´ exoribonucleases Xrn1, Rat1, and Dxo1 (XRN1, XRN2, and

DXO in mammals) require an RNA substrate with a 5´-PO4 terminus (Stevens 2001). The

NSD and NGD pathways for cytosolic decay of defective mRNAs both propose a final step

39 in 5´ end degradation that requires 5´-phosphate-dependent degradation by Xrn1. The

5´-OH termini generated by metal-ion independent endoribonucleases, such as the HAC1 mRNA intermediates produced by yeast Ire1 cleavage during the UPR, are immune to decay by these known 5´→3´ exoribonucleases. Therefore, 5´-OH RNAs cannot be substrates for degradation by Xrn1 or Rat1 without phosphorylation of the 5´ terminus.

Similarly, if the 3´ cleavage products of NGD are generated with a 5´-OH rather than a 5´-

PO4, they would also require phosphorylation to become competent for 5´→3´ decay.

Kinase-Mediated Decay

Throughout this thesis, I use the term “kinase-mediated decay” (KMD) to refer to any decay process where a 5´-OH RNA is phosphorylated prior to decay by a 5´-phosphate- dependent 5´→3´ ribonuclease. On the following pages, I describe KMD in three previously published examples: T4 phage infection of E. coli, tRNA intron processing in yeast, and pre-rRNA processing in yeast. Prior to this dissertation, these three disparate examples have not been united by a common theory of regulated RNA decay.

Kinase-mediated decay of T4 bacteriophage mRNA

E. coli lack the 5´→3´ exoribonucleases found in eukaryotes, and instead accomplish mRNA degradation primarily via multiple endoribonucleolytic cleavages promoting 3´→5´ decay. The initial step in mRNA decay is conversion of the 5´- triphosphate mRNA terminus to a 5´-PO4 terminus by RNA pyrophosphohydrolase (RppH)

(Celesnik et al. 2007). RNase E and its homolog RNase G recognize the 5´-PO4 terminus of mRNA and cleave downstream, creating a 3´-fragment with a 5´-PO4 terminus that itself

40 can be recognized for additional cleavages (Mackie 1998; Spickler et al. 2001). The result of these cleavages is several smaller mRNA fragments, which are degraded by one of three

3´→5´exoribonucleases – polynucleotide (PNPase), RNase R, and RNase

II – and oligoribonuclease (Review: Andrade et al. 2009)).

Bacteriophage T4 infects E. coli and must coordinate expression of “early” and

“late” genes during the infection. RegB is a unique sequence-specific endoribonuclease encoded by the T4 phage genome that cleaves primarily at tetranucleotide GGAG in the

Shine-Dalgarno sequence of “early” genes (Uzan et al. 1988; Sanson et al. 2000; Saïda et al. 2003). As a metal ion-independent endoribonuclease, RegB leaves a 5´-OH terminus on the 3´ cleavage fragment (Uzan et al. 1988; Saïda et al. 2003). This 5´-OH terminus was initially confusing, because E. coli RNases E and G require a 5´-PO4 terminus to promote rapid RNA turnover by creating entry sites for PNPase, RNase R, and RNase II. Durand et al. solved the quandary when they showed that T4 polynucleotide kinase (PNK) phosphorylates the 5´terminus, creating a substrate competent for decay by the endogenous

RNases E and G (Durand et al. 2012) (Figure 1.8A).

In this system of KMD, the T4 phage uses its own kinase (PNK) to phosphorylate

5´-OH RNAs for decay by the E. coli host (RNases E and G) (Durand et al.

2012). Because RegB specifically cleaves “early” phage mRNAs, the T4 phage can turn off expression of “early” genes by promoting their turnover. This joint viral-host KMD is the first known example of phosphorylation of 5´-OH RNAs by an RNA kinase prior to decay by a 5´-phosphate-dependent nuclease, and it is the only example thus far occurring in prokaryotes.

41 Kinase-mediated decay of yeast tRNA introns

In addition to cleavage of pre-tRNA by TSEN and ligation of tRNA exons by

TRL1, intron-containing tRNAs are also processed by KMD. As described in the “5´-OH

RNAs in Biological Systems” section, cleavage of the pre-tRNA by TSEN liberates the tRNA intron with 5´-OH and 2´,3´-cyclic phosphate termini (Knapp et al. 1979). The fate of this tRNA intron was recently uncovered following the identification of Xrn1 in a genome-wide screen for tRNA processing defects (Wu et al. 2015). Deletion of the XRN1

Leu Lys Trp gene resulted in the accumulation of tRNA CAA, tRNA UUU, tRNA CCA, and

Pro tRNA UGG introns with 5´-PO4 termini, which suggested that tRNA introns are phosphorylated prior to 5´-phosphate-dependent decay by Xrn1 (Wu and Hopper 2014).

Using temperature-sensitive Trl1 mutants rlg1-4 and rlg1-10 (Phizicky et al. 1992), Wu and Hopper demonstrated that the phosphorylation of the 5´-OH terminus of tRNA introns is greatly diminished when Trl1 function is impaired (Wu and Hopper 2014). Thus in this example of KMD, the 5´-OH tRNA intron is phosphorylated by Trl1, which promotes

5´→3´ degradation by Xrn1 (Figure 1.8B).

While kinase-mediated decay of tRNA introns has not yet been demonstrated in mammals, Cleavage and factor 1 subunit 1 (CLP1) is likely the mammalian kinase of tRNA introns. CLP1 is one of two known polyribonucleotide 5´-OH in humans. The activity was first identified by Weitzer and Martinez by following siRNA-kinase activity over eight biochemical purification steps and analyzing the final fraction by mass spectrometry (Weitzer and Martinez 2007b). While endogenously- produced miRNAs have 5´-PO4 termini (Nykänen et al. 2001), siRNAs introduced for experimental knockdown of mRNAs have 5´-OH termini and are phosphorylated by CLP1

42 in the cytoplasm to become RISC substrates (Weitzer and Martinez 2007b). In the same study, Weitzer and Martinez demonstrated that both tRNA introns and tRNA 3´-exons are phosphorylated on their 5´-termini in vitro in the presence of purified or immunoprecipitated CLP1-TSEN complex. This activity is abolished by single-site mutations (either S128A or K127A) that impair the catalytic activity of CLP1 (Weitzer and

Martinez 2007a).

Three recent studies have demonstrated the importance of CLP1 in maintaining specific neuronal populations and have linked CLP1 to tRNA processing in vivo. Mice homozygous for a kinase-dead CLP1K127A mutation develop progressive spinal motor neuron loss, similar to SOD1 mouse models of acute lateral sclerosis (Hanada et al. 2013).

Humans with a rare recessive homozygous kinase-defective CLP1R140H mutation display pontocerebellar hypoplasia and progressive neurodegenerative features (Schaffer et al.

2014; Karaca et al. 2014). In all three studies, aberrantly processed tRNA fragments from intron-containing tRNAs accumulate in cells with CLP1 mutations. In the human patient- derived fibroblasts, these include tRNA introns (Karaca et al. 2014) and tRNA 3´-exons

(Schaffer et al. 2014), which accumulate with 5´-OH termini. It is not known whether these fragments are causative of disease, though transfection of the tRNA 3´-exon with a 5´-OH terminus increased cellular death of CLP1 mutant cells during oxidative stress (Schaffer et al. 2014). Taken together, these results suggest that CLP1 could potentially function in a role analogous to Trl1 in the decay of tRNA intermediates, and that the kinase activity of

CLP1 may be required for normal neural development.

43 Kinase-mediated decay of yeast pre-rRNA

A second example of kinase-mediated decay in yeast was recently demonstrated in rRNA processing (Gasse et al. 2015). Eukaryotic rRNA processing was first described in the early 1960s. In Jim Darnell’s lab at MIT, Klaus Scherrer showed that “giant” RNA accumulated in the rapidly-sedimenting 45S fraction of a sucrose gradient when HeLa cells were labelled for a short time with uridine-2-C14, and that increasing the labelling time resulted in RNA accumulating in the 35S, and then the 28S and 16S fractions (Scherrer and Darnell 1962; Scherrer et al. 1963; Scherrer 2003). At the Fox Chase Cancer Center,

Bob Perry published similar results (Perry 1962; Perry et al. 1964). Thus, it was proposed that 45S “giant” RNA was the precursor to ribosomal RNA.

Surprisingly, the molecular details of the complex processing pathway of ribosomal

RNA have still not been completely solved over half a century later. In yeast, rRNA is encoded by 150 genomic copies of rDNA repeat units in a tandem array on

XII (Fernández-Pevida et al. 2015). The 5S pre-rRNA is transcribed as a single RNA pol

III transcript and Rex1 performs a rapid, non-essential maturation step, trimming 12 nt from the 3´ end (van Hoof et al. 2000a). In contrast to this simple maturation pathway, the maturation of the 25S, 18S, and 5.8SS/L rRNAs from the RNA Pol I 35S pre-rRNA transcript requires 8 endonucleolytic cleavages (A0, A1, A2, A3, B0, B1L, C2, and D) and 8

“trimming” steps (B1S, B2, C1´, C1, and 4 steps between 7SS/L to 5.8SS/L) (Reviews:

Woolford and Baserga 2013; Fernández-Pevida et al. 2015). While the enzymes and complexes responsible for the trimming steps have been identified, the source of four of the endonucleolytic cleavages (A0, A1, B1L, and C2) have remained elusive.

44 In 2015, the long-sought-after endonuclease performing the C2 cleavage was identified. It was already known that internal transcribed spacer 2 (ITS2) between the 5.8S and 25S rRNAs is cleaved from 27SB pre-rRNA, and these rRNAs are further matured by several exoribonucleases (Veldman et al. 1981). Importantly, 25S rRNA was known to be primarily trimmed by Rat1 in 5´-phosphate-dependent 5´→3´ decay (Geerlings et al. 2000).

Gasse et al. demonstrated that Las1, a metal ion-independent endonuclease, cleaves the

ITS2 spacer, generating a 26S pre-rRNA with a 5´-OH terminus. They further show that phosphorylation of the 26S pre-rRNA by the RNA/DNA 5´-kinase Grc3 initiates decay by the nuclear 5´→3´ exonuclease Rat1 to produce 25S´ rRNA (Gasse et al. 2015) (Figure

1.8C).

The components of the yeast KMD ribosomal rRNA processing system are conserved through humans. LAS1L performs the 3´ and 5´ cleavage of ITS2 in human 45S pre-rRNA (Schillewaert et al. 2012). Then Nol9, the human nucleolar RNA 5´-kinase, likely phosphorylates the 5´ end of the 26S pre-rRNA followed by trimming by the nuclear

5´-phosphate-dependent 5´→3´ exoribonuclease XRN2, the Rat1 homolog (Heindl and

Martinez 2010). KMD of 26S pre-rRNA may be required for normal mammalian development as a LAS1L mutation was recently found in patients with spinal muscle atrophy with respiratory distress (SMARD) (Butterfield et al. 2014).

Extensions of kinase-mediated decay

The preceding examples of eukaryotic KMD support two key conclusions. First, given that grc3∆ yeast are inviable and single amino acid substitutions impairing Gcr3 catalytic activity abolish 26S pre-rRNA phosphorylation in vitro (Gasse et al. 2015), KMD

45

Examples of kinase-mediated Examples decay (KMD). .

Figure 1.8

46 Figure 1.8. Examples of kinase-mediated decay (KMD). A. KMD of T4 phage mRNA. (1) Bacteriophage T4 infects E. coli. (2) T4-encoded RegB, a sequence-specific endonuclease, cleaves “early” T4 genes in the Shine Delgarno (SD) sequence, generating 3´-fragments with 5´-OH termini. (3) T4 PNK phosphorylates the 5´-OH termini. (4) E. coli RNases E and G cleave the 5´-PO4 RNAs to create entry points for 3´→5´ decay (Durand et al. 2012). T4 factors are displayed in red, and E. coli factors in blue. B. KMD of yeast tRNA introns. (1) Sen2 and Sen34 cleave an intron-containing pre- tRNA, leaving a 5´-OH terminus and a 2´,3´-cyclic phosphate terminus. (2) Trl1 phosphorylates the 5´-OH terminus, creating a 5´-PO4 substrate for degradation by Xrn1 (3) (Wu and Hopper 2014). C. KMD of yeast pre-rRNA. (1) Las1 cleaves the ITS2 spacer in 27SB pre-rRNA, generating 26S pre-rRNA with a 5´-OH terminus. (2) Grc3 phosphorylates this terminus, (3) promoting decay by Rat1 to produce 25S´ rRNA (Gasse et al. 2015).

47 may be an essential process in eukaryotes. Second, KMD is probably a wide-spread phenomenon. It has already been shown to occur in both the cytoplasm and the nucleus, with each of the known 5´-OH kinases (Trl1 and Grc3) and two 5´→3´ exoribonucleases

(Xrn1 and Rat1). In this dissertation, particularly in Chapter III, I show that KMD is required for regulation of the yeast UPR, and I present evidence that KMD may function broadly in mRNA surveillance to degrade 5´-OH cleavage products.

Methods for Global Mapping of RNA Modifications

Next-generation sequencing allows for massively parallel sequencing of DNA using clustered amplification and iterative addition of reversibly terminated, fluorescently labeled nucleotides. Though initially developed as a tool for genomic sequencing, next- generation sequencing technology was creatively harnessed to map RNA by using degenerate or oligo-dT primers to reverse transcribe cDNA. In 2008, a series of high- profile papers applied this new “RNA-seq” method to analyze the transcriptomes of

Saccharomyces cerevisiae, Schizosaccharomyces pombe, Prunus persica, Arabidopsis thaliana, mouse and human cells (Wang et al. 2009). RNA-seq was rapidly combined with ingenious modified linkers and unique ligation chemistries to specifically capture and sequence a variety of RNA modifications. Below, I describe several such methods: Cap

Analysis Gene Expression (CAGE) (Shiraki et al. 2003; Kodzius et al. 2006), 5´ phosphate sequencing (German et al. 2008; Harigaya and Parker 2012; Pelechano et al. 2016), and

2´,3´-cyclic phosphate sequencing (Schutz et al. 2010; Honda et al. 2015). Despite these and other methods to capture and sequence RNA terminal chemistries, 5´-OH RNAs had never been directly mapped prior to my work.

48 Cap Analysis Gene Expression (CAGE)

Two of the key limitations of RNA-seq are the resolution of 5´ and 3´ termini of

RNA transcripts and the difficulties in isolating specific populations of interest, such as mRNA. The Cap Analysis Gene Expression (CAGE) method was pioneered to specifically capture m7G capped RNAs, allowing the precise mapping of the 5´ end as well as a quantitative measure for gene expression comparison (Shiraki et al. 2003; Kodzius et al.

2006). In brief, the diol residue of the mRNA cap is chemically biotinylated, allowing for specific purification of capped RNA hybridized to a single strand of reverse-transcribed

DNA using streptavidin. T4 RNA Ligase is then used to attach a linker with an Mme1 recognition motif, and subsequent treatment with Mme1 results in cleavage of 20 or 21 nucleotides downstream of the original capsite. The resulting fragments can then be concatenated and sequenced. The initial study using the CAGE method revealed tissue-specific alternative transcription starting points (TSPs) for a subset of genes (Shiraki et al. 2003). Application of a related cap-based sequencing method,

RAMPAGE (RNA Annotation and Mapping of Promoters for the Analysis of Gene

Expression), to 36 stages of Drosophila development found at least two promoters in more than 40% of developmentally expressed genes (Batut et al. 2013).

5´-phosphate sequencing

miRNA-directed cleavage by Argonaute yields an mRNA fragment with a 5´- phosphate terminus. To identify miRNA targets on a massive scale, Pam Green’s lab pioneered a method to specifically capture and sequence RNAs with 5´-PO4 termini

(German et al. 2008). An RNA oligonucleotide with a Mme1 recognition motif is directly

49 ligated to RNAs with 5´-PO4 termini. The RNA is reverse transcribed with an oligo-dT primer, followed by second strand synthesis. Digestion with Mme1 shortens the 3´ terminus to 20 or 21 nts from the original 5´-PO4. Finally, a DNA adapter is ligated to the

3´ terminus, and the library is PCR amplified and sequenced. Once sequencing reads are trimmed of the initial 5´ RNA oligonucleotide, the 5´ ends of reads map to 5´ phosphate cleavage fragments. This method in conjunction with Paired Analysis of RNA Ends

(PARE) successfully identified hundreds of miRNA-target RNA pairs in Arabidopsis, including nearly 100 previously validated miRNAs (German et al. 2008). The 5´-phosphate sequencing method has also been applied to yeast to demonstrate a novel discard pathway for debranched lariat-intermediates and identify introns in genes previously thought to be intron-less (Harigaya and Parker 2012). Importantly, the surprising capture of HAC1 3´- exon mRNA cleavage fragments by 5´-phosphate cloning in dcp2∆ xrn1∆ yeast cells suggests a stable phosphorylated intermediate in the HAC1 mRNA ligation pathway

(Harigaya and Parker 2012). This finding provides essential motivation for the studies described in Chapter III. Another group recently developed a nearly identical method to sequence 5´-PO4 RNA, which they termed “5P-seq” (Pelechano et al. 2016).

Sequencing of RNAs with 2´,3´-cyclic phosphate termini

Our lab has extensive experience in the preparation of 2´,3´-cyclic phosphate libraries (Schutz et al. 2010). In this method, A. thaliana tRNA ligase is used to attach a linker to RNAs terminating in 2´,3´-cyclic phosphates. Yeast Tpt1 is then used to dephosphorylate the 2´ phosphate at the ligation junction. Following reverse transcription with a primer complementary to the linker, cDNA is purified and sequenced. The first

50 2´,3´-cyclic phosphate sequencing study captured the 3´ terminus of the U6 snRNA as well as cleavages in the anti-codon loop of 5 tRNA species (Glu-CTC, Glu-TTC, Lys-CTT,

Lys-TTT, His-GTG) consistent with Rny1 cleavage (Schutz et al. 2010). This method has subsequently been used to identify sites of RNase L cleavage in poliovirus (Cooper et al.

2014) and influenza A infected cells (Cooper et al. 2015), and to identify targets of Ire1 cleavage in fission yeast (Kimmig et al. 2012). A similar protocol to specifically sequence cyclic phosphate RNAs – “cP-RNA-seq” – identified a novel category of angiogenin cleavage products termed 5´ Sex HOrmone-dependent TRNA-derived RNAs (SHOT-

RNAs) in breast and prostate cancer cell lines (Honda et al. 2015).

Towards 5´-OH sequencing

Although other methods can partially infer 5´-OH RNA mappings, there are key limitations. For instance, although in vitro phosphorylation prior to 5´-PO4 sequencing would in principle capture 5´-OH RNAs, this population would be a mixed population with in vivo 5´- PO4 RNAs. A figure in the most recent 5PSeq paper suggested 5´-OH RNAs might be captured following 5´-P RNA removal by Terminator followed by phosphorylation with PNK; however, this approach was not performed and was not discussed within the text (Pelechano et al. 2016). Cyclic phosphate cloning theoretically captures all 5´ fragments at 5´-OH RNA cleavage sites generated by a single endoribonucleolytic cleavage; however, in practice, the signal generated by this method is overwhelmingly derived from ncRNAs including rRNAs, tRNAs, and snRNAs. In contrast, a method specifically capturing 5´-OH RNA allows for advantageous use of the polyA tail to easily enrich for polyadenylated species, including mRNA populations.

51 In Chapter II, I describe a novel method to directly capture and sequence 5´-OH

RNA. In my initial analysis, I show how the method enables new inquiries into this unexplored RNA modification, and in Chapter IV, I apply 5OH-seq to mammalian samples. Given the adoption of other RNA modification sequencing protocols and their widespread applications, I anticipate that the 5OH-Seq method will be used by other labs to pursue questions such as those outlined in Chapter V.

52 CHAPTER II

GLOBAL ANALYSIS OF RNA CLEAVAGE BY

5´-HYDROXYL RNA SEQUENCING1

Introduction

Metal ion-independent endoribonucleases including the tRNA splicing endonuclease (Trotta et al. 1997), Ire1 (Sidrauski and Walter 1997) and the ribonucleases

T2 (Luhtala and Parker 2010) and L (Cooper et al. 2014) create cleavage products with 5´- hydroxyl (5´-OH) and 2´,3´-cyclic phosphate termini, as do self-cleaving ribozymes like the hammerhead, HDV, VS, hairpin and twister ribozymes (Ferré-D'Amaré and Scott

2010). In addition, accelerated RNA cleavage by intramolecular phosphoester transfer at structurally distorted phosphodiester bonds generates 5´-OH and 2´,3´-cyclic phosphate products (Soukup and Breaker 1999). However, products of RNA cleavage with 5´-OH and 2´,3´-cyclic phosphate termini are not typically identified in global surveys of RNA populations, and thus their frequency and role in RNA metabolism remain unknown.

Methods for studying RNAs with specific termini can shed light on RNA transcription initiation, processing, and decay pathways. Several methods have been developed for the global study of mRNA 5´-termini, but most of these do not capture 5´-

OH termini, or do so indirectly via a 5´-PO4 intermediate. In a recent example, 5´-OH termini were indirectly inferred from cloning of 5´-PO4 termini created upon phosphorylation of 5´-OH termini (Schifano et al. 2014), but this method requires several enzymatic manipulations prior to the ligation step. Similarly, the Parallel Analysis of RNA

1 Portions of this chapter were reused with permission from Peach et al., 2015. 53 Ends (PARE) method (German et al. 2009) begins with ligation of an RNA adapter to the

5´-PO4 terminus of an RNA substrate with T4 RNA ligase, and thus this method does not capture 5´-OH termini. Other methods compare signals among different enzymatic manipulations of 5´ mRNA cap structures to enable specific detection of mRNA 5´-termini

(Gu et al. 2012), but these methods are unable to capture 5´-OH termini.

A method was previously developed to identify RNA fragments with 2´,3´-cyclic phosphate and 2´-PO4/3´-OH termini using the A. thaliana tRNA ligase (Schutz et al.

2010). Recent studies used this method to discover virus and host targets of Ribonuclease

L cleavage (Cooper et al. 2014), and to identify the Gas2 mRNA as a novel target of the

Ire1 endonuclease in the S. pombe unfolded protein response (UPR) (Kimmig et al. 2012).

Fungal and plant tRNA join RNAs with 5´-OH and 2´,3´-cyclic phosphate termini via 5´-PO4 and 5´-adenylate intermediates (Greer et al. 1983), and thus the method employing A. thaliana tRNA ligase does not distinguish between 5´-OH and 5´-PO4 RNA termini. In addition, libraries produced by this method typically contain high levels of ribosomal RNA fragments with cyclic-phosphate termini, which may mask signals from lower abundance fragments from mRNA. Therefore I was motivated to develop a new method to capture the other products of RNA cleavage – 5´-OH RNA fragments – that also allows mRNAs to be enriched via oligo-dT selection of their polyadenylate tails.

E. coli RtcB RNA ligase catalyzes a unique chemistry involving the attack of a 3´- guanylylate intermediate by a 5´-OH nucleophile, yielding a 3´-5´ phosphodiester bond and

GMP (Tanaka et al. 2011). As noted previously (Desai and Raines 2012; Tanaka et al.

2011), RtcB RNA ligase therefore represents a unique reagent for direct capture of 5´-OH

RNAs. In this chapter, I describe the development of the “5OH-seq” method using the E.

54 coli RtcB RNA ligase to globally identify 5´-OH RNA fragments in complex RNA populations. I apply the 5OH-seq method to study 5´-OH RNA fragments in budding yeast.

Materials and Methods

Expression and purification of E. coli RtcB RNA ligase

The E. coli rtcB gene was PCR amplified, recombined into pDONR221 (Gateway

BP, Invitrogen) and Sanger sequenced to confirm lack of mutations in the open reading frame. The rtcB ORF was recombined into pET-53-DEST (creating an N-terminal 6xHis tag; Gateway LR, Invitrogen) and transformed into BL21 (DE3) RIPL cells (Stratagene).

The pET-53-RtcB expression construct is available from Addgene (Plasmid 51282). For

RtcB protein expression, cells were grown from a single colony in 1 L terrific broth with selective antibiotics (50 µg/mL ampicillin and 34 µg/mL chloramphenicol) at 37˚C to an

OD600 of 0.6 and then chilled on ice for 30 min. Ethanol was added to 2%, expression was induced with 0.1 mM IPTG and cultures were grown 16 hours at 17˚ C. Cells were harvested by centrifugation and lysed for 1 hour at 4˚C in Buffer A (50 mM Tris pH 7.4,

250 mM NaCl, 10% sucrose) with 0.2 mg/mL lysozyme. Triton X-100 was added to a final concentration of 0.1%, and the cells were sonicated briefly to reduce viscosity. Following clarification by centrifugation at 10,000 x g, the supernatant was incubated for 1 hour at

4˚C with 3 mL of pre-washed Nickel-NTA beads (Qiagen). The beads were washed with

25 mL of Buffer A, then loaded onto a column and washed twice with 25 mL Buffer E (50 mM Tris-HCl (pH 7.4), 10% glycerol, 250 mM NaCl) containing 15 mM imidazole. The column was eluted with 3 mL each of Buffer E with 50, 150 and 300 mM imidazole. Peak fractions were identified by SDS-PAGE and were pooled and diluted to 50 mM NaCl with

55 Buffer D (50 mM Tris-HCl (pH 7.4), 10% glycerol, 5 mM DTT. 1 mM EDTA). RtcB protein was further purified on a 1 mL HiTrap Heparin HP column (GE Healthcare Life

Sciences), washed with Buffer D containing 50 mM NaCl, and eluted with 3 mL each of

Buffer D containing 150 mM, 300 mM and 500 mM NaCl. RtcB protein eluted predominantly in the 150-300 mM NaCl fractions, with a typical yield of 2 mg / L culture.

Peak fractions were pooled, quantitated by Bradford assay, snap frozen in liquid nitrogen and stored at -80˚C.

RtcB ligation assay

Oligonucleotides used in this study are listed in Table 2.1. Ligation assays were used to confirm recombinant RtcB function. A linker oligonucleotide with a 3´-phosphate was incubated at equimolar concentration (200 pM) with another oligonucleotide with a

5´-OH terminus in the presence of 1.25 µM RtcB in buffer (1 mM GTP, 50 mM MOPS pH

7.4, 2 mM MnCl2) for 1 hour at 37˚C. The reaction was quenched with stop dye (95% formamide, 5 mM EDTA pH 8.0, 0.01% xylene cyanol and bromophenol blue), denatured at 70˚C for 3 minutes, and electrophoresed on 10% TBE-Urea polyacrylamide gel and stained with SYBR Gold (Invitrogen) prior to visualization.

Yeast strains and culturing

Yeast strains used in this study are listed in Table 2.2. Single colonies were inoculated in YEPD and incubated at 30˚C overnight with rotation. Cultures were diluted to OD600 of 0.1 in YEPD. In untreated conditions, yeast cells were harvested in mid-log phase. To induce the UPR, yeast were grown to mid-log phase, then treated for 2 hours

56 Table 2.1. Oligonucleotides

Ligation Assay RNA Oligonucleotides Sequence 5′ Amino – rGrUrUrUrUrCrCrCrArGrUrCrA M13-Linker rCrGrArC – 3′ phosphate 5′ OH – rArArGrArArGrArGrArArArGrArArG RPS31-5OH rArArGrGrUrCrUrArCrArCrCrArCrC – 3′ Amino

5OH Library Oligonucleotides 5′ desthiobiotin – ACACTCTTTCCCTACACGACGC 5OH-Linker TCTTCCGATCTrNrNrNrNrNrNrNrN – 3′ phosphate 5´ Amino – GTGACTGGAGTTCAGACGTGTG 5OH-RT-amino CTCTTCCGATCTNNNNNNNNN – 3´

Validation Primers M13-F 5′ GTTTTCCCAGTCACGAC RPS31-F 5′ CCCTGGAAGTTGAATCTTCTG RPS31-R 5′ GCGTTAACCTTGTAGACGGAAT

Table 2.2. Yeast strains

Strain Genotype XRN1 (BY4742) MATα his3Δ1 leu2Δ0 lys2Δ0 ura3Δ0 xrn1∆ MATα his3Δ1 leu2Δ0 lys2Δ0 ura3Δ0 xrn1Δ::hygMX

57 with tunicamycin (final concentration of 2.5 µg/mL, Sigma-Aldrich). Cells were harvested by centrifugation and total RNA was isolated by hot acid phenol extraction and treated with TURBO DNase (Ambion) to remove genomic DNA.

5´-hydroxyl RNA library construction and sequencing

Total RNA from budding yeast (15 µg) was enriched for a polyadenylated mRNA fraction using oligo-dT magnetic beads (Ambion). PolyA-enriched RNA was incubated with 1.1 µM RtcB in buffer (1 mM GTP, 50 mM MOPS pH 7.4, 2 mM MnCl2), 20 units

RNase Inhibitor (Enzymatics) and 10 µM 5OH-Biotin-Linker at 37 ˚C for 1 hour.

Following phenol/chloroform extraction and ethanol precipitation, RNA was fragmented with Ambion Fragmentation Reagent for 15 minutes at 70 ˚C, and the reaction was quenched with the addition of EDTA. Samples were denatured in stop dye (95% formamide, 0.01% xylene cyanol / bromophenol blue), heated to 65 ˚C for 5 minutes and separated on a 10% acrylamide TBE-Urea gel. Following staining with SYBR Gold

(Invitrogen), the gel was visualized by blue light transillumination and a gel slice containing RNA above 80 nt was excised and pulverized. RNA was eluted from the gel by

2-hour incubation at 40˚C in 0.3 M sodium acetate, pH 5.2, 1 mM EDTA, pH 8.0 and precipitated with 2.5 volumes of ethanol. cDNA was prepared using an Illumina- compatible primer with a 9 nt degenerate region (oligonucleotide 5OH-RT-amino) and

Protoscript II RT (NEB), and was purified using Agencourt RNAClean XP Beads

(Beckman Coulter). Second strand synthesis was performed with RNase H, E. coli DNA

Polymerase, and E. coli Ligase (Enzymatics). Double-stranded DNA products were purified using magnetic Streptavidin beads (Invitrogen), eluted with 25 mM biotin in

58 elution buffer (Omega Bio-Tek) and PCR amplified with Illumina TruSeq primers and

Phusion DNA polymerase. PCR reactions were purified with SPRI Size Select Beads

(Beckman Coulter). Indexed libraries were quantified by Qubit (Invitrogen), mixed to a final concentration of 1-10 nM and sequenced on an Illumina MiSeq in a 50 cycle run. Raw and processed sequencing data are available at NCBI GEO under accession number

GSE61527.

Analysis of DNA sequencing data

Sequences in FASTQ format were preprocessed to remove molecular indexes using umitools (https://github.com/brwnj/umitools). Processed reads were aligned to the S. cerevisiae genome (sacCer1) with Bowtie (Langmead et al. 2009) and visualized in the

UCSC Genome Browser (Karolchik et al. 2014). Coverage signals in bedGraph format were generated, normalized to UMI-corrected Counts Per Million reads mapped (CPM), and analyzed with BEDTools (Quinlan and Hall 2010) and custom scripts. Studies of mRNA cleavage focused on the collection of 5´-OH fragments in the upper decile of the

CPM distribution (i.e., the most abundant 10%) that were reproducible in 5OH-seq libraries prepared from two independent cultures. Significant discrete 5´-OH signals within mRNAs were identified as previously described (Harigaya and Parker 2012). Briefly, the CPM values for each nucleotide of a given mRNA were fit to a transcript-specific negative binomial distribution, and P values were computed for each site against the background distribution. P values for all sites in all mRNAs were combined and adjusted using the

Benjamini-Hochberg false discovery rate (FDR) (<0.05). Clustered hits were determined by identifying regions of high 5´-OH signal density using the Model-based Analysis of

59 ChIP-Seq (MACS) algorithm (Zhang et al. 2008), selecting regions longer than 100 bp, and discarding regions included in the list of mRNAs with single fragments. Additional workflow and analysis scripts are available at https://github.com/hesselberthlab/5OH.

Processed data can be visualized on the UCSC Genome Browser in the track hub at https://github.com/hesselberthlab/trackhub.

Validation of 5´-hydroxyl RNA fragments by RtcB ligation and RT-PCR

5´-OH RNA fragments were verified by RtcB-mediated ligation of an oligonucleotide linker, followed by reverse transcription and gene-specific PCR. Total

RNA (5 µg) was ligated to an RNA oligonucleotide with a 5´-amino group and a 3´- phosphate (M13-Linker-RNA) with RtcB. Reactions were phenol-chloroform extracted, ethanol precipitated, and reverse transcribed with Superscript III RT (Invitrogen) and gene- specific primers. cDNA was PCR amplified using primer pairs that hybridize to full length mRNA or the linker-ligated 5´-OH RNA fragment. PCR products were resolved on 10%

TBE non-denaturing 29:1 acrylamide gels and stained with SYBR Gold.

Results

Development of a method for capture of 5´-hydroxyl RNAs

To demonstrate the utility of RtcB RNA ligase as a reagent for global capture of

5´-OH RNA, I first asked whether RtcB could ligate unrelated oligonucleotide substrates

– a prerequisite for global capture of 5´-OH molecules. I overexpressed and purified E. coli

RtcB ligase (Figure 2.1A) and used the enzyme to perform ligations with

60 oligoribonucleotide substrates with 2´-OH/3´-PO4 and 5´-OH termini with minimal predicted base pairing (Figure 2.1B). I found that RtcB could ligate these substrates in a bimolecular ligation with ~25% efficiency, confirming previous studies of RtcB intermolecular ligation (Das et al. 2013) and suggesting that RtcB could be used in global profiling of cellular RNAs with 5´-OH termini.

I next established a general protocol to capture 5´-OH RNA from complex RNA populations using the RtcB ligase (Figure 2.2). I began by chemically fragmenting total

RNA from budding yeast, producing many RNA fragments with 5´-OH and 2´,3´-cyclic phosphate termini. RNA fragments with 5´-OH termini were ligated to an oligonucleotide with 5´-desthiobiotin and 3´-PO4 groups using E. coli RtcB ligase. This oligonucleotide linker contains an 8-base random sequence at its 3´ end that serves as a molecular index during sequence analysis, enabling quantitation of unique ligation events and estimation of relative amounts of 5´-OH fragments in the initial RNA population (Kivioja et al. 2012).

After gel purification to remove excess unligated linker, RNA was reverse transcribed with a second oligonucleotide primer containing an Illumina primer and nine randomized bases at the 3´ end to facilitate random priming to RNA templates. Double- stranded RNA:DNA hybrids were converted to double stranded DNA by second-strand synthesis, and double-stranded DNA fragments containing the desthiobition linker were purified on immobilized streptavidin and eluted with free biotin. Purified double-stranded

DNA fragments were PCR amplified with primers that incorporate library-specific indexes and flow cell sequences compatible with Illumina sequencing.

61

Figure 2.1. Purified E. coli RtcB ligates RNA substrates with 3´-PO4 and 5´-OH termini A. SDS-PAGE analysis of E. coli RtcB protein. Recombinant RtcB protein (10 µg) was electrophoresed through a 10% SDS-PAGE gel and stained with Coomassie blue. B. RtcB-mediated bimolecular ligation of RNA substrates with 3´-PO4 and 5´-OH termini was analyzed by polyacrylamide gel electrophoresis. The ligation product at 50 nt was dependent on RtcB, with an efficiency of ~25%.

62

Figure 2.2. Schematic of method to clone and sequence RNAs with 5´-OH termini (5OH-seq). RNA cleavage generates a 5´-fragment with a 2´,3´-cyclic phosphate terminus and a 5´- fragment with a 5´-OH terminus. An adaptor with a 5´-desthiobiotin and 3´-phosphate is ligated to the 5´-OH RNA fragment with E. coli RtcB ligase. This adaptor has 8 randomized positions at its 3´ end which serve as a molecular index to uniquely identify individual ligation events upon sequencing (Kivioja et al. 2012). Following ligation, excess adaptor is removed by gel purification. A primer with a 3´ random hexamer region is used for reverse transcription to create cDNA, followed by second-strand synthesis to generated double-stranded DNA products. After streptavidin isolation, DNA fragments are PCR amplified and analyzed by Illumina sequencing.

63 Global analysis of 5´-hydroxyl RNAs in budding yeast

I used this library construction protocol to identify 5´-OH RNA fragments in the budding yeast S. cerevisiae. I prepared libraries from oligo-dT enriched yeast total RNA, in order to increase signals that might be present on polyadenylated mRNA fragments.

These libraries were sequenced on the Illumina MiSeq platform with 50 bp single reads to identify 5´-OH RNA fragments. The 5OH-seq libraries had an average of 2.7 million reads per sample, with an average of 23% of the reads aligning uniquely to the yeast genome and an additional 30% of reads aligning multiple times. The majority of unaligned reads were due to Illumina primer sequence amplification. I examined the distribution of uniquely aligning reads within existing RNA annotations, and found that in polyA-enriched samples,

21% mapped to mRNA regions, 0.4% mapped to abundant non-coding RNA species

(tRNA, snoRNA, and snRNA), and 65% of the reads were derived from ribosomal RNA.

Using these sequences, I also assessed whether RtcB exhibited a bias for specific sequences at the ligation junction. I found that the base composition of each position downstream of the ligation site (i.e., the 5´-OH terminus) mirrored budding yeast A/T content (62%), suggesting that RtcB does not have strong preference for 5´-OH RNAs with specific sequences. The 8 nt molecular index in ligated adaptors (Figure 2.2 and Table 2.1) allowed me to examine base composition upstream of the ligation site, and I found that RtcB preferred adaptors with an adenosine at their 3´ terminus, which was present in 42% of sequences, while the remaining 7 positions had more random composition. Because the molecular index is trimmed from reads prior to alignment, this ligation bias has no effect on sequence alignment, but does limit the diversity of possible molecular indexes.

64 To validate the protocol, I analyzed the capture of 5´-OH termini generated in vitro and in vivo. Ribosomal RNAs have 5´-PO4 termini produced by endonucleolytic (Venema et al. 1995) and exonucleolytic processing (Geerlings et al. 2000). These ends should be underrepresented in 5´-OH libraries, but I would expect to recover these termini in 5´-OH libraries prepared from RNA samples that had been dephosphorylated in vitro (Figure

2.3A). Therefore, I compared libraries prepared from total RNA samples that were treated or not with recombinant shrimp alkaline (rSAP) to convert 5´-PO4 RNAs in the samples to 5´-OH groups. An average of 2574 reads aligned to the 5´ end of 5.8S rRNA in the dephosphorylated sample compared to an average of 358 5´-OH fragments aligned to the same position in the untreated samples, (Figure 2.3B), representing a 7-fold enrichment of reads at the 5´ end of this RNA upon phosphatase treatment and confirming that the method captures 5´-OH RNAs and discriminates against 5´-PO4 RNAs.

The levels of ribosomal RNA fragments found in the poly(A)-enriched library suggested that oligo-dT capture of polyadenylated mRNA is not sufficient to deplete ribosomal RNA signals from libraries, and that other strategies for ribosomal RNA depletion might be used to further reduce signals from ribosomal RNA. I prepared libraries from an RNA sample following treatment with a commercial strategy to remove ribosomal

RNA by hybridization to antisense probes (Sigurgeirsson et al. 2014). I found that signals from ribosomal RNAs were ~10-fold less abundant in these libraries compared to libraries prepared from oligo-dT selected RNA, but also unexpectedly found that the commercial kit contained a phosphatase activity that converted the 5´-PO4 termini of ribosomal RNAs to 5´-OH groups, leading to their capture by 5OH-seq independent of additional phosphatase treatment (data not shown).

65 Similar to rRNA, some small nucleolar RNAs (snoRNAs) are produced from endonucleolytic and exonucleolytic processing of precursor RNA molecules and have 5´- monophosphate termini (Petfalski et al. 1998). Dephosphorylation of the 5´-PO4 termini of these snoRNAs would produce 5´-OH termini, and I found abundant signals for these in

5´-OH libraries following rSAP treatment. For example, the 5´-end of the box C/D snoRNA

U18, which is produced by Rnt1 and Rat1 processing of the EFB1 pre-mRNA (Giorgi et al. 2001), was 6-fold more abundant in the dephosphorylated sample compared to the untreated sample (Figure 2.3C). Overall I identified six snoRNAs (snR18, 24, 55, 57, 61,

64) with 5´-PO4 termini that were captured by 5OH-seq after in vitro phosphatase treatment.

Cleavage of pre-tRNA molecules by the tRNA splicing endonuclease (SEN) creates intron and 3´-exon products with 5´-OH groups (Trotta et al. 1997) (Figure 2.4A). Yeast pre-tRNA molecules range from 71 to 133 nt and processed tRNAs from 71 to 87 nt.

Sequences in my libraries have an effective read length of 42 nt, and thus my libraries can be used to quantify cleavages within ~30-40 nt of the 5´-end of processed tRNA molecules, depending on variable loop length. I analyzed signals from tRNAs in the 5´-OH library prepared from rRNA-depleted total RNA. The presence of 5´-OH fragments corresponding

Leu to the 5´-terminus of the 3´-exon of the intron-containing tRNA CAA (Figure 2.4B, top panel) confirms that 5OH-seq can capture known 5´-OH termini generated in vivo. I also

Ser identified products of cleavage in the anticodon loop of tRNA AGA (Figure 2.4B, middle panel), consistent with 2´,3´-cyclic phosphate capture of anticodon loop cleavage (Schutz et al. 2010). Finally, I identified high levels of 5´-OH fragments that mapped to the D arm and loop of 28 tRNA species (Figure 2.4B, bottom panel). I compared these 5´-OH

66

Figure 2.3. 5OH-seq captures RNAs with 5´-PO4 termini after in vitro dephosphorylation. A. Schematic of RtcB-mediated ligation to a 5´-OH RNA substrate produced by phosphatase-mediated removal of a 5´-PO4 group. B. Abundance of 5´-OH fragments (per-site, UMI-corrected Counts Per Million reads mapped, CPM) mapped to 5.8S rRNA (157 nt). In an untreated sample (top), 5´-OH signal was uniformly low across the body of the 5.8S rRNA. 5´-OH RNA fragments cloned from total RNA treated with shrimp (bottom) mapped to the 5´-end of 5.8S rRNA (grey arrows), indicating conversion of the 5´-PO4 to a 5´-OH group C. Abundance of 5´-OH fragments (CPM) in the snR18 show that the 5´-PO4 of snR18, produced by Rnt1-mediated processing of the EFB1 pre-mRNA (Petfalski et al. 1998), is specifically captured upon phosphatase treatment and 5OH-seq (grey arrows).

67

Figure 2.4. Cleavages in tRNA in introns, anticodon loops, and D arm and loop A. Schematics of tRNAs. On the left, a mature tRNA with the CCA terminator and the D, anticodon, variable, and TΨC loops. On the right, a portion of an intron-containing pre- tRNA with cleavage sites recognized by tRNA Splicing Endonuclease (SEN) represented by black lines. Sites within tRNAs that were captured abundantly by 5OH-seq are shaded. Arrows indicate the most abundant fragment in each tRNA shown in B. B. Examples of tRNA fragments captured by 5OH-seq in rRNA-depleted total RNA. Leu Fragments from tRNA CAA mapped to the 5´-terminus of the 3´ exon (top panel). Anti- Ser Asp codon loop cleavages were found in tRNA AGA (middle panel). tRNA GUC was the most abundant species captured by 5OH-seq and primarily showed D-arm and loop cleavage (bottom panel). Arrows indicate the most abundant fragment in each tRNA and correspond to the arrows in A. Key regions of the tRNA are indicated by solid bars labeled as Intron, Anticodon loop, and D arm & loop.

68 fragments in tRNAs to a previous study of 5´-PO4 RNA fragments (Harigaya and Parker

2012), and found that 5´-PO4 signals were localized mainly to the 5´ ends of tRNA, indicating that the 5´-OH fragments represent a distinct pattern of tRNA cleavage fragments (data not shown). I did not identify the 5´-OH intron products of SEN cleavage, presumably because they are phosphorylated by Trl1 and degraded by Xrn1

(Wu and Hopper 2014).

mRNA decay events identified with 5´-hydroxyl RNA capture

In order to focus on 5´-OH fragments present in mRNA, I collected sequence reads from libraries prepared from poly(A)-enriched RNA that mapped uniquely within annotated mRNA transcripts (Nagalakshmi et al. 2008), and found 710 mRNAs that had significant signals in two replicate samples (35% overlap among mRNAs). I examined the distribution of 5´-OH fragments within these mRNAs (Figure 2.5A), and found that 76.7% of 5´-OH fragments mapped within open reading frames, 9.9% were in 5´ untranslated regions (UTRs) and 13.4% were in 3´ UTRs. I found 81 mRNAs with 5´-OH fragments that mapped exclusively to the 5´ and 3´ UTRs, and 551 mRNAs with 5´-OH fragments exclusively in their open reading frames. The remaining 78 mRNAs contained fragments distributed throughout the mRNA. Many 5´-OH fragments were biased toward the 3´-end of mRNAs, possibly reflecting a bias imposed by poly(A) purification prior to 5OH-seq

(Figure 2.5A).

I identified a subset of mRNAs with 5´-OH signals that also were expressed under similar conditions as measured by a previous RNA-seq experiment (Levin et al. 2010). For each of these 629 mRNAs, I calculated a summary statistic by dividing the sum of its CPM

69 values by the length of the mRNA in kilobases (Counts Per Kilobase per Million reads mapped; CPKM), and compared mRNA CPKM values to their expression levels in Reads

Per Kilobase per Million mapped (RPKM) (Levin et al. 2010). There was a modest but significant correlation between levels of 5´-OH fragments and mRNA abundance (Pearson

R = 0.31; p < 10-16), indicating that 5´-OH fragment abundance is partly a reflection of mRNA abundance (Figure 2.5B). I also evaluated whether these fragments were associated with intrinsic structural properties of RNAs. I compared 5´-OH fragment levels to previous measures of global RNA structures using Ribonuclease S1 and V1 mapping (Kertesz et al.

2010), and found minimal correlation between these data sets (Pearson R = -0.04; p <

0.048). This suggests that 5´-OH sites may be uniformly distributed among RNA secondary structure elements (i.e., single and double-strand regions).

There were unique patterns of 5´-OH fragment distribution within mRNAs. I identified 187 mRNAs with multiple fragments mapping to localized regions within the gene body, and these 5´-OH fragments clustered within a mean distance of 191 nt. For example, the IMH1, RPN2, and YEF3 mRNAs had clustered 5´-OH fragments within their coding regions, and the SAC1 mRNA had clustered fragments within its 3´ UTR (Figure

2.6A). The pattern of clustered 5´-OH fragments suggests that a localized process cleaves these specific mRNAs in multiple locations within a defined region.

I found 17 mRNAs that each had a single predominant 5´-OH fragment (FDR <

0.05). For example, the ADH1 and YNK1 mRNAs had single sites of cleavage in their open reading frames, CYS4 had a predominant cleavage in its 3´ UTR, and a predominant cleavage in the MDH1 mRNA is 8 nt downstream of its stop codon (Figure 2.6B). I also identified a prominent 5´-OH fragment in the RPS31 mRNA, which encodes an in-frame

70

Figure 2.5. Analysis of mRNAs captured by 5OH-seq. A. 5´-OH fragments are found in numerous mRNAs. Each of 710 mRNAs was divided into proportionally sized bins for the 5´ and 3´ UTR regions (2 bins each, left and right of the solid vertical lines), and the open reading frame (ORF; 20 bins). 5´-OH fragment abundance (CPM) in each bin was summed, and mRNAs were sorted by their maximum bin value from 5´ to 3´. Each mRNA is displayed in a row. A total of 11 mRNAs have signal exclusively within their 5´ UTRs (top) and 69 mRNAs have signal exclusively in their 3´ UTRs (bottom). The grey center line marks the middle of the ORF region; 5´-OH signal was predominantly in the 5´ portion of 143 mRNAs (above horizontal dashed line), while remaining mRNAs had signals mainly localized toward their 3´ ends. B. Comparison of signals in mRNAs in 5´-OH and RNA-seq libraries. There was a significant correlation (Pearson R = 0.31; p < 10-16) between 5´-OH signals (counts per million reads per kilobase; CPKM) and RNA-seq (FPKM) (Levin et al. 2010) for 629 mRNAs (a subset of the mRNAs in A. Dotted lines indicate median RNA-seq FPKM and 5´-OH CPKM values, and the linear regression is plotted as a grey line.

71

Figure 2.6. 5´-OH RNAs exist in regions of clustered or discrete cleavage. A. Plots of 5´-OH abundance (CPM) versus position (bp) for four representative mRNAs (of 187 total) with multiple, distributed 5´-OH fragments. All mRNAs are plotted from 5´ to 3´, with UTR regions in grey, and ORF regions in black. 5´-OH fragments in the IMH1, RPN2, and YEF3 mRNAs were predominantly localized to their coding regions, and 5´- OH fragments in the SAC1 mRNA were restricted to its 3´ UTR. B. Three representative mRNAs (of 17 total) with single, predominant 5´-OH fragments. The ADH1 and YNK1 mRNAs each contained a single predominant 5´-OH fragment in their ORF regions, and the CYS4 mRNA contained a single 5´-OH fragment in its 3´ UTR. The MDH1 mRNA had a predominant 5´-OH fragment 8 nt downstream of its stop codon.

72 fusion of ubiquitin to the S31 protein of the ribosomal small subunit (Finley et al. 1989).

The 5´-OH fragment mapped 3 nt downstream of the codons for the Arg Gly-Gly (RGG) terminus of ubiquitin, suggesting that this fragment might be generated during translation and ubiquitin processing (Lacombe et al. 2009) (Figure 2.7A). However, another mRNA encoding a ubiquitin-ribosomal protein, RPL40A, had significant levels of 5´-OH fragments located farther downstream of the codons encoding the ubiquitin C-terminus, suggesting that ubiquitin processing is not responsible for RPS31 mRNA cleavage (Figure

2.7B).

In the RPS31 mRNA, the 7 codons following the Arg-Gly-Gly residues – and at the site of the 5´-OH fragment – are AAA and AAG codons that encode Lys and Arg residues, which are known to trigger co-translational mRNA cleavage (Tsuboi et al. 2012). In addition, a similar fragment from the RPS31 mRNA was previously identified in ribosomal profiling experiments (Guydosh and Green 2014) (Figure 2.7C), suggesting that this fragment is associated with ribosomes. I validated the RPS31 mRNA fragment by RtcB- mediated linker ligation followed by reverse transcription and PCR, yielding product sizes consistent with 5OH-seq (Figure 2.7D). Based on this amplification, I estimate that these fragments are <1% of the abundance of full length RPS31 mRNA.

Given that the RPS31 5´-OH fragment mapped upstream of a region of six consecutive basic residues, I analyzed the global relationship between 5´-OH fragments and nearby codon content. I first examined annotated open reading frames to identify sequences encoding amino acid sequences of at least five consecutive identical or related residues (e.g., five or more consecutive lysine residues, or a mix of five consecutive lysine and arginine residues), and determined the average coverage of 5´-OH fragments 75 nt

73

Figure 2.7. Single cleavage in RPS31 mRNA is validated by link-RT-PCR.

74 Figure 2.7. Single cleavage in RPS31 mRNA is validated by link-RT-PCR. A. A single predominant 5´-OH RNA fragment (grey arrow) maps to the RPS31 mRNA, encoding the ubiquitin-Rps31 fusion protein. The fragment maps to a position 3 nt downstream of the codons for Arg-Gly-Gly (RGG) residues (vertical black arrow), which signal ubiquitin removal by deubiquitinases (Lacombe et al. 2009). B. Multiple, distributed 5´-OH signals are further downstream of RGG codons (vertical black arrow) in the RPL40A mRNA, which encodes the ubiquitin-Rpl40a fusion protein. C. An abundant RNA fragment recovered in ribosome profiling experiments (Guydosh and Green 2014) maps to the same position in the RPS31 mRNA as the 5´-OH fragment in A. This fragment was present in all samples, but was most abundant in ribosome profiling done in cells treated with the non-hydrolysable GTP analog GMP-PNP (shown). D. The 5´-OH fragment in RPS31 was validated by RtcB ligation-mediated RT-PCR (“link- RT-PCR”). 5´-OH fragments in total RNA were incubated with an RtcB substrate (a synthetic oligonucleotide with a 3´-phosphate) in the presence or absence of RtcB ligase. These reactions were divided and reverse transcribed with a primer specific for RPS31 that hybridizes downstream of the 5´-OH terminus. cDNA products of reverse transcription were PCR amplified with primers that hybridize to RPS31 up- and downstream of the 5´- OH terminus and analyzed by gel electrophoresis, yielding abundant fragments at the appropriate size (415 bp), independent of RtcB ligation (left panel). A second PCR was performed with a 5´ primer that hybridizes to the RtcB substrate, and analysis of these products showed ligation-dependent PCR products at the expected sizes (major 5´-OH fragment at 215 bp; minor products at ~400 bp). A portion of A is displayed on the right for reference.

75 upstream and 75 nt downstream from the start of these regions (Figure 2.8A). I analyzed

5´-OH fragments cloned from rRNA-depleted RNA, and found that 5´-OH fragments accumulated upstream of the 240 instances of consecutive basic residues composed of both lysine and arginine codons (Figure 2.8B). This relationship was more pronounced than the accumulation of 5´-OH fragments upstream of regions with consecutive lysine or arginine residues alone (Figure 2.8B). I also found high levels of 5´-OH fragments upstream of the

105 instances of consecutive glutamate residues, but this signal was diminished when I considered consecutive acidic residues composed of aspartate alone (n=102) or regions composed of both glutamate and aspartate (n=642) (Figure 2.8B). The highest levels of

5´-OH fragments were upstream of regions composed of a mixture of acidic and basic residues, particularly the combination of Lys and Glu (Figure 2.8C). The levels of 5´-OH fragments were higher when longer lengths of consecutive amino acids were considered

(Figure 2.8C). I identified specific examples of mRNAs with 5´-OH fragments upstream of regions encoding consecutive acidic residues including CBF5 and HSC82 (Figure

2.8D). Analysis of other consecutive codons showed that they are not associated with 5´-

OH fragment accumulation (data not shown). These findings suggest a relationship between the electrostatic properties of a nascent peptide and mRNA cleavage events that yield 5´-OH fragments, possibly via a co-translational process such as no-go mRNA decay

(Shoemaker and Green 2012).

The eukaryotic 5´→3´ exonuclease Xrn1 specifically recognizes 5´-PO4 RNA substrates, but has greatly reduced activity on 5´-OH RNA in vitro (Stevens 2001).

Consistent with this activity, xrn1∆ budding yeast accumulate several types of 5-PO4 RNAs

(Harigaya and Parker 2012), and phosphorylation of the 5´-OH RNA products of tRNA

76

Figure 2.8. 5´-OH RNAs accumulate upstream of regions encoding consecutive polycharged amino acids.

77 Figure 2.8. 5´-OH RNAs accumulate upstream of regions encoding consecutive polycharged amino acids. A. Schematic of 5´-OH signals with respect to regions encoding specific peptide regions in B. B. The mean 5´-OH fragment coverage in rRNA-depleted total RNA was calculated 75 nt upstream (negative) and downstream (positive) of the start position of all mRNA sequences encoding at least 5 consecutive basic or acidic residues, including Arg (n=12), Lys (n=59), a mixture of Lys and Arg (n=280), Glu (n=105), Asp (n=102), and a mixture of Glu and Asp (n=642). A peak of 5´-OH fragments is present ~30 nt upstream of regions composed of Lys and Arg, but this signal is lower in regions encoding Lys or Arg alone (top panels). Coverage of 5´-OH fragments is elevated higher in regions encoding Glu than regions encoding Asp, or Glu and Asp (bottom panels). C. The mean 5´-OH fragment coverage was calculated 75 nt upstream (negative) and downstream (positive) of the start position of all mRNA sequences encoding varying lengths of consecutive basic and acidic residues. 5´-OH fragments were enriched upstream of regions encoding polyelectrostatic (both basic and acidic) resides (Arg, Lys, Glu, Asp; left panel) and were highest upstream of regions encoding only Lys and Glu (right panel). In both panels, regions encoding longer consecutive stretches (≥20) of polyelectrostatic amino acids had higher 5´-OH signals than regions encoding shorter stretches (≥5). D. Representative examples of mRNAs (CBF5 and HSC82) with 5´-OH fragments that map to positions upstream of sequences encoding polyelectrostatic amino acid motifs. Regions encoding polyelectrostatic amino acid stretches are indicated by grey boxes above the gene models.

78 splicing is required for their turnover by Xrn1 (Wu and Hopper 2014). Given the high specificity of Xrn1 for 5´-PO4 RNA substrates in vitro, it is possible that other 5´-OH RNA fragments could be subject to a similar turnover pathway with a requisite phosphorylation event prior to Xrn1-mediated turnover.

Identification of endonucleolytic cleavages produced during the unfolded protein response

During the unfolded protein response (UPR), misfolded or aggregated proteins accumulate in the lumen of the endoplasmic reticulum and activate the kinase- endoribonuclease Ire1 (Ron and Walter 2007). Activated Ire1 endonuclease cleaves an intron from the HAC1 pre-mRNA (Gonzalez et al. 1999), and the cleaved exons are subsequently ligated by yeast Trl1 tRNA ligase (Sidrauski et al. 1996). After ligation, the mature HAC1 mRNA is translated, and Hac1 protein traffics to the nucleus where it drives expression of hundreds of stress response genes (Sidrauski et al. 1996). I mapped 5´-OH

RNA fragments in cells treated with the N-linked glycosylation inhibitor tunicamycin

(Tm), which causes protein misfolding and UPR activation. The primary 5´-OH RNA fragments captured from the HAC1 mRNA mapped upstream of a sequence encoding consecutive electrostatic residues (Figure 2.9A), consistent with the transcriptome-wide enrichment of 5´-OH signal upstream of mRNA sequences encoding polyelectrostatic residues (Figure 2.8B). The canonical Ire1 cleavage sites in the HAC1 pre-mRNA were identified at the 5´ end of the intron and the 5´ end of the 3´-exon (Figure 2.9B, all panels).

In ∆xrn1 yeast, the HAC1 5´-OH RNA products of Ire1 cleavage increased with in vitro shrimp alkaline phosphatase (SAP) treatment, consistent with their phosphorylation in vivo

– presumably by Trl1 5´-kinase (Sidrauski et al. 1996) – followed by Xrn1-mediated

79 degradation (Figure 2.9B, top panel). The increase in 5´-OH signal at the the 5´ end of the

HAC1 3´-exon in the xrn1∆ strain treated with tunicamycin (Figure 2.9B, second panel) compared to the XRN1 strain treated with tunicamycin (Figure 2.9B, fourth panel) may be due to an increase in stability of the primary HAC1 transcript prior to cleavage.

I compared 5´-OH mRNA signals in untreated cells and cells treated with tunicamycin, and found a 5´-OH fragment from the KAR2 mRNA (the yeast ortholog of

BiP/HSP70) that was 17-fold more abundant in the presence of tunicamycin (Figure

2.10A). Initially I attributed this signal to upregulation of KAR2 mRNA expression by the

UPR transcriptional response (Normington et al. 1989), but I also identified six other ER- related mRNAs with 5´-OH fragments that increased 5-fold or more upon tunicamycin treatment and were present in a sequence motif similar to the KAR2 cleavage site:

5´-CCAUU|A-3´ (Figure 2.10A, lower panel; Figure 2.10B). One of these mRNAs is the

UPR-induced thioredoxin peroxidase TSA1 mRNA (Kimata et al. 2005), where 5´-OH fragments accumulated at two such motifs upon tunicamycin treatment (Figure 2.10C). A shared motif within these mRNAs suggests that a site-specific endoribonuclease specifically cleaves certain mRNAs during UPR induction.

Discussion

I developed a method to directly capture RNAs with 5´-OH termini using the unique substrate specificity of E. coli RtcB RNA ligase (Tanaka et al. 2011) (Figures 2.1, 2.2).

This method is highly selective for the capture of 5´-OH products (Figure 2.3), and yields reproducible signals among several classes of cellular RNA. RNA fragments with 5´-OH termini have not been directly detected in other global studies, and the collection of 5´-OH

80

Figure 2.9. HAC1 mRNA cleavage fragments accumulate with 5´-OH and 5´-PO4 termini.

81 Figure 2.9. HAC1 mRNA cleavage fragments accumulate with 5´-OH and 5´-PO4 termini. A. The majority of 5´-OH fragments from the HAC1 mRNA map upstream of codons for consecutive electrostatic amino acid residues. Their abundance was similar in DMSO- treated cells (DMSO, top panel) and cells treated with 2.5ug/mL tunicamycin for 2 hours (+Tm, bottom panel) cells. B. Ire1 cleavage fragments from HAC1 mRNA are captured in 5OH-Seq libraries. Low levels of fragments resulting from intron cleavage were found in all conditions at the 5´end of the intron and the 5´ end of the 3´-exon (all panels). The fragments were most abundant in the xrn1∆ strain treated with tunicamycin and subjected to in vitro dephosphorylation with shrimp alkaline phosphatase (+Tm, +SAP, top panel).

82

Figure 2.10. Some 5´-OH RNAs increase with tunicamycin treatment and occur in a common motif. A. 5´-OH fragment abundance (CPM) in KAR2 mRNA was examined in cells treated with vehicle (DMSO, top) and tunicamycin (middle). A 5´-OH fragment increased 17-fold upon tunicamycin treatment (middle) and mapped to the sequence 5´-CCAUU|A-3´ (bottom). B. Consensus sequence motif in logo format derived from sequences flanking (10 nt up- and downstream) 5´-OH fragments that increased in abundance 5-fold or more upon tunicamycin treatment. A total of 10 instances of the motif were found among 23 mRNAs. C. Two 5´-OH fragments in the TSA1 mRNA increased 11.4-fold and 8.7-fold upon tunicamycin treatment (top and middle panel) and mapped to the consensus motif in B (bottom panel, arrows).

83 RNA fragments reported here collectively comprise a previously uninterrogated RNA class.

I identified three categories of tRNA fragments with 5´-OH termini: those derived from cleavage at the boundary of the intron and 3´ exon, cleavage in the anticodon loop, and cleavage in the D-arm/loop (Figure 2.4). The first of these validates the method by demonstrating capture of a known 5´-OH terminus generated in vitro during the well- characterized processing of an intron-containing pre-tRNA by the tRNA Splicing

Endonuclease (SEN) (Trotta et al. 1997). A mechanism of anticodon loop cleavage is also known: Rny1 can cleave the anticodon loop during stress conditions (Thompson and Parker

2009b), and related fragments have been previously captured by 2´,3´-cyclic phosphate cloning (Schutz et al. 2010). However, the most abundant tRNA cleavages observed in

5OH-seq were in the D-arm and loop, which may represent a novel tRNA stress fragment generated simultaneously with the 5´-tRF fragment (Saikia and Hatzoglou 2015).

I identified a relationship between 5´-OH mRNA fragments and mRNA sequences that encoded consecutive basic peptide sequences (Figures 2.7, 2.8). In the RPS31 mRNA, a 5´-OH fragment accumulates upstream of codons encoding a polybasic peptide (Figure

2.7A). Production of the fragment does not appear to be related to ubiquitin processing, as another ubiquitin fusion RPL40A does not have 5´-OH fragments near the site of ubiquitin cleavage (Figure 2.7B). A similar fragment in the RPS31 mRNA was also identified in ribosomal profiling experiments done in budding yeast (Guydosh and Green 2014).

(Figure 2.7C). These experiments predominantly recover RNA fragments that are protected from ribonuclease digestion by ribosomes, but do so without relying on a specific

5´ end chemistry (Ingolia et al. 2012). Additionally, several previous studies showed that

84 nascent peptides with specific electrostatic properties cause ribosome pausing and trigger co-translational mRNA decay. In non-stop decay (NSD), translation of the polyadenylate tails of mRNAs that lack stop codons leads to a nascent polylysine peptide (Ito-Harashima et al. 2007), which is proposed to interact with the negatively charged exit tunnel of the ribosome (Lu and Deutsch 2008) to trigger ribosome stalling and mRNA decay. Similarly, internal polybasic peptides trigger co-translational mRNA decay, and this process is dependent on the non-canonical release factors Dom34 and Hbs1 (Tsuboi et al. 2012). In addition, global measures of translation indicated that polybasic sequences cause ribosomal stalling, suggesting they interfere with ribosome elongation (Brandman et al. 2012). The presence of this RPS31 5´-OH fragment in ribosome profiling data suggests that it may be associated with ribosomes and may be a product of a ribosome-directed decay process driven by the polybasic coding region.

Many studies have used synthetic reporter mRNAs to study the role of nascent peptides in determining decay of a translating mRNA (Tsuboi et al. 2012). However, there are no known physiological mRNA substrates of no-go decay (NGD), and the endonuclease activity responsible for mRNA cleavage during NGD is not known

(Shoemaker and Green 2012). Detailed mapping of mRNA cleavage events during NGD showed that mRNA cleavage occurs around 20 nt upstream of a polybasic sequence

(Tsuboi et al. 2012), which is consistent with the distance between 5´-OH fragments and other polyelectrostatic sequences (Figure 2.8). The 3´-fragments generated in NGD are degraded by Xrn1 (Doma and Parker 2006), which requires 5´- PO4 substrates (Stevens

2001). Perhaps the 3´-fragments of NGD are generated by an endonuclease that creates

5´-OH RNA products that are phosphorylated by a cellular 5´-kinase activity (e.g., the

85 tRNA ligase Trl1), promoting their decay by Xrn1. This scenario would be similar to 5´- kinase-mediated turnover of spliced tRNA introns by Xrn1 (Wu and Hopper 2014). Thus, it is possible that many of the 5´-OH fragments I identified are products of NGD. Future studies will determine whether the production of 5´-OH fragments in mRNA sequences encoding highly electrostatic peptides (Figure 2.8) is dependent on NGD factors such as

Dom34 and Hbs1 (Tsuboi et al. 2012).

I also identified 5´-OH fragments upstream of polyacidic and other polyelectrostatic sequences composed of a mixture of basic and acidic residues (Figure

2.8). These signals were more prominent than any regions composed of any other single amino acid. Polyacidic stretches in the C-terminus of several human proteins are associated with ribosome frameshifting in these regions (Gould et al. 2014). However, it is not clear whether frameshifting is a consequence of the polyacidic nascent peptide, or whether it is a consequence of the propensity out-of-frame polyglutamate codons to encode new Met initiation codons. Notably, the 5´-OH fragment I identified 8 nt downstream of the MDH1 stop codon (Figure 2.6B) is near the site of ribosomal stop codon read through in other fungi, which appends a peroxisomal targeting sequence (PTS) to the Mdh1 protein for its localization (Stiebler et al. 2014). If a similar phenomenon accounts for the MDH1 mRNA

5´-OH fragment in S. cerevisiae, it suggests that production of 5´-OH fragments is a general feature of mRNA decay at sites of non-canonical ribosome elongation (i.e., frameshifting or read through).

A potential artifact of mRNA purification with oligo-dT beads is the enrichment of

A-rich 5´-OH RNAs independent of their non-templated 3´-polyadenylate tails. In particular, some codons for electrostatic residues (Arg, Lys, Glu) are A-rich, and might be

86 preferentially isolated by oligo-dT beads. I found that 5´-OH fragments upstream of polylysine stretches (encoded by AAA and AAG codons) were indeed increased in polyA libraries relative to rRNA-depleted libraries, but 5´-OH fragment levels upstream of other mixed polyelectrostatic sequences were present at similar levels in polyA selected and rRNA-depleted libraries. Furthermore, in polyA selected samples and rRNA-depleted samples, I detected no enrichment of 5´-OH signal upstream of polyasparagine residues, which are encoded by AAU and AAC codons and are well-represented in the yeast transcriptome (n=138). Therefore, I conclude that oligo-dT purification introduces minimal codon bias in the 5OH-seq signal.

During activation of the unfolded protein response, the HAC1 pre-mRNA is cleaved by the kinase-endoribonuclease Ire1, producing 2´,3´-cyclic phosphate and 5´-OH RNA fragments (Gonzalez et al. 1999) that are substrates for ligation by tRNA ligase (Sidrauski et al. 1996). However, previous studies that identified 5´-PO4 mRNA fragments from dcp2∆ xrn1∆ yeast recovered HAC1 cleavage products at the sites of Ire1 cleavage

(Harigaya and Parker 2012). Using 5OH-seq, I recovered HAC1 mRNA cleavage products at their highest levels in xrn1∆ cells after in vitro dephosphorylation (Figure 2.9B, top panel). The presence of these 5´-PO4 RNA fragments possibly indicates that the Trl1 5´- phosphorylation step is rapid, but further HAC1 mRNA ligation by Trl1 is slow or often not completed. Corroborating this idea, fungal ligases have rapid 5´-kinase kinetics, but a slower ligation step (Remus and Shuman 2014). Similarly, the products of metazoan

Regulated IRE1-Dependent Decay (RIDD) are degraded by Xrn1 (Hollien and Weissman

2006), implying that a 5´-kinase phosphorylates 5´-OH products of Ire1 cleavage prior to their decay by Xrn1.

87 Previous studies indicated that HAC1 is the sole target of Ire1 cleavage in budding yeast (Niwa et al. 2005); however, a more recent study provided evidence that yeast Ire1 can cleave known human RIDD substrates in vitro, and suggested that DAP2 and MFα1 mRNAs may be endogenous targets of yeast RIDD (Tam et al. 2014). I identified 23 mRNAs with 5´-OH fragments that increased more than 5-fold with tunicamycin treatment.

Ten of these fragments were present in the consensus motif 5´-CAUU|A-3´, suggesting that the mRNAs are targeted by a site-specific endonuclease following UPR induction (Figure

2.10B). An obvious candidate for this endonuclease is Ire1, but the 5´-CAUU|A-3 cleavage motif identified by 5OH-seq differs from the canonical Ire1 recognition motif for HAC1 mRNA in budding yeast (Sidrauski and Walter 1997), as well as motifs in other RIDD targets in fission yeast and metazoans (Maurel et al. 2014). However, Ire1 cleavage specificity can vary with oligomerization state, and Ire1 may not necessarily cleave at the canonical motif during RIDD (Gaddam et al. 2013). Further study will determine whether these mRNAs are directly cleaved by Ire1, or are an indirect consequence of UPR activation.

The ability to globally map 5´-OH RNA fragments will enable new studies of 5´- kinase enzymes that phosphorylate 5´-OH substrates. For example, the human RNA 5´- kinase CLP1 was first identified as a component of the CF IIm sub-complex of cleavage and polyadenylation specificity factor (de Vries et al. 2000), and has recently been implicated in quality control of RNA 5´-termini during transfer RNA processing (Weitzer and Martinez 2007b). Mutations in CLP1 cause severe neurodevelopmental phenotypes, and reduce its 5´-kinase activity in vitro (Schaffer et al. 2014; Karaca et al. 2014). Previous studies proposed that reduction in 5´-kinase activity causes accumulation of aberrant RNA

88 fragments (Hanada et al. 2013), but it is currently not known how accumulation of these molecules might lead to human neurodevelopmental defects or the spinal motor neuron deficits that develop in mice expressing kinase-dead Clp1 (Hanada et al. 2013). Turnover of tRNA fragments may be analogous to the turnover of tRNA introns in yeast, where phosphorylation of 5´-OH tRNA introns leads to their rapid 5´→3´ decay by Xrn1 (Wu and

Hopper 2014). Application of the 5OH-seq method in cells harboring mutant CLP1 with reduced 5´-kinase activity might identify new 5´-OH RNA substrates of CLP1, providing insight into this class of neurodevelopmental disorders. In Chapter IV, I present the preliminary results of such an experiment.

Summary

In this chapter, I described 5OH-seq, a novel method that uses the unique ligation chemistry of E. coli RtcB to capture 5´-OH RNAs prior to library preparation and high- throughput sequencing. 5OH-seq is specific for 5´-OH RNAs, and I validated this by showing enrichment for RNAs with 5´-PO4 termini only after in vitro dephosphorylation.

By performing 5OH-seq on polyA+ RNA from budding yeast, I identified several known and previously unknown 5´-OH RNAs. I discovered 5´-OH tRNA fragments from the D- arm/loop, anticodon loop, and introns of tRNA. I demonstrated that 5´-OH RNAs are enriched upstream of polycharged regions, potentially identifying the first in vivo targets of no-go decay. I also showed that a subset of 5´-OH mRNAs increased with tunicamycin treatment, which may signify that regulated Ire1-dependent decay (RIDD) occurs in yeast.

Of all the findings in this initial 5OH-seq study, the confirmation of HAC1 mRNA fragments with 5´-PO4 termini rather than 5´-OH termini in xrn1∆ strains is the most

89 intriguing for follow-up studies. It suggests that: i) the kinase activity of Trl1 can be separated from the ligation activity, and ii) Xrn1 is responsible for the degradation of HAC1 mRNA intermediates. In the next chapter, I directly test the hypothesis that the HAC1 mRNA 3´-exon intermediate with a 5´-OH terminus is phosphorylated by Trl1 prior to degradation by Xrn1, a process I term “kinase-mediated decay” (KMD).

90 CHAPTER III

KINASE-MEDIATED DECAY OF

EUKARYOTIC MRNA CLEAVAGE PRODUCTS

Introduction

Many steps in RNA processing are initiated by RNA cleavage events that produce

RNAs with 2´,3´-cyclic phosphate and 5´-OH termini. During the unfolded protein response (UPR), unfolded proteins in the ER lumen promote oligomerization of the transmembrane kinase-endoribonuclease Ire1, which cuts a cytoplasmic mRNA to create

2´,3´-cyclic phosphate and 5´-OH products (Gonzalez et al. 1999). In budding yeast, Ire1 cleaves an intron from the HAC1 mRNA and the exons are ligated by the multifunctional

Trl1 RNA ligase (Sidrauski et al. 1996), whereas in metazoans the products of XBP1 cleavage by IRE1 are ligated by the RTCB RNA ligase (Lu et al. 2014; Kosmaczewski et al. 2014). Although the HAC1 3´-exon is produced with a 5´-OH terminus following Ire1- mediated cleavage, two transcriptome-wide sequencing studies captured the HAC1 3´-exon with a 5´-PO4 terminus in xrn1∆ yeast (Harigaya and Parker 2012, Peach et al. 2015)

(Chapter II), indicating that the HAC1 3´-exon is normally phosphorylated after Ire1 cleavage and then degraded by Xrn1.

RNAs with 5´-OH termini must be phosphorylated to become competent for decay by 5´-phosphate-dependent exoribonucleases, and I have termed this process “kinase- mediated decay” (KMD). Three previous examples of KMD suggest that it is a common mode of RNA turnover. In the first example of KMD, bacteriophage T4 cleaves “early genes” with T4 endoribonuclease RegB shortly after infection of E. coli, generating 3´

91 RNA cleavage fragments with 5´-OH termini. Phosphorylation of the 5´-OH terminus by

T4 polynucleotide kinase (PNK) yields a 5´-PO4 RNA intermediate, which is degraded by the host endonucleases RNase E and G (Durand et al. 2012). In the second example, the yeast SEN complex cleaves intron-containing pre-tRNAs in yeast, releasing an intron with a 2´,3´-cyclic phosphate terminus and a 5´-OH terminus (Knapp et al. 1979). Trl1 phosphorylates the 5´-OH terminus enabling degradation of the intron by the 5´-phosphate- dependent 5´→3´ exoribonuclease Xrn1 (Wu and Hopper 2014). In the most recent example of KMD, the Las1 endonuclease cleaves the ITS2 spacer of 27SB pre-rRNA in yeast, generating 26S pre-mRNA with a 5´-OH terminus. RNA/DNA 5´-kinase Grc3 phosphorylates the terminus of 26S pre-rRNA, and the 5´-PO4 terminus promotes recognition and degradation by nuclear 5´→3´ exoribonuclease Rat1 to produce mature

25S´ rRNA (Gasse et al. 2015).

Insight into unknown endonucleolytic factors that cleave mRNA substrates may be gained by examining the termini of RNA cleavage products. In no-go mRNA decay

(NGD), mRNA features that slow or stall ribosome movement are sensed by the non- canonical release factors Dom34/Hbs1, which couple ribosome slowing to mRNA cleavage (Doma and Parker 2006; Parker 2012). However, the factors that catalyze NGD endonucleolytic cleavage are not known. Thus, while Xrn1 degrades the 3´ product of cleavage (Doma and Parker 2006), it is not known whether NGD products are created with

5´-PO4 termini, or with 5´-OH termini that are subsequently phosphorylated.

In this chapter, I establish the role of 5´-OH RNA kinase Trl1 and 5´-phosphate- dependent 5´→3´ exoribonuclease Xrn1 in two KMD pathways for mRNA cleavage products in budding yeast. In the first example, I use a series of experiments to demonstrate

92 that Trl1 phosphorylates the 5´-OH terminus of the HAC1 mRNA 3´-exon cleavage fragment. The 5´-PO4 terminus promotes degradation by Xrn1, which in turn suppresses

HAC1 mRNA splicing and UPR activation. In the second example, I present an experiment performed by Patrick Cherry, which shows that NGD fragments are generated with 5´-OH termini, and require Trl1 phosphorylation prior to their decay by Xrn1. These two pathways suggest that KMD may function broadly to degrade 5´-OH mRNA cleavage products.

Materials and Methods

Yeast strains and plasmids

Plasmids used in this chapter were obtained from previous studies or created as indicated in Table 3.1 with further details of the (10x-tRNA) plasmid below. Yeast strains used in this chapter are described in Table 3.2. Plasmids were introduced into yeast by standard lithium acetate transformation. For plasmid shuffles, yeast were plated on 5- fluoroorotic acid (5-FOA). The following yeast strains were a gift from S. Shuman: SPT05,

SPT06, SPT07, SPT08 (Tanaka et al. 2011).

Generation of (10x-tRNA) plasmid

A sequence was designed that encodes the 10 intron-containing yeast tRNAs without introns (i.e, “pre-spliced”). The expression of this “tRNA block” is driven by the

SUP4 promoter and terminated by RPR1. This sequence was ordered from IDT as a gBlock and recombined into pDONR221 in a Gateway BP reaction (Thermo Fisher). The product plasmid (pDONR221-10X-tRNA) was recombined by a Gateway LR reaction into

93

Source 2014)Hopper and (Wu et2011)al. (Tanaka et2011)al. (Tanaka et2011)al. (Tanaka text in Described ofplasmidchange(TRL1) Quick the 2006) and Parker (Doma TRP1), promoter SUP4 TRP1), � HIS3), TPI promoter TPI HIS3), � -D425N CEN HIS3), native-D425N HIS3),promoter CEN tRNA 2 LEU2),YEpXRN1 � RTCB2 TRL1 Genotype 2 (pGP564 promoter URA3), CEN TRL1 (pRS416-TRL1 promoter HIS3), CEN TRL1 (pRS413-TRL1 (pRS423- (pAG424-10X- (pRS413- UAS LEU2),GAL1 CEN (pRP1251

Plasmids.

sL)

- -D425N) tRNA) Name (XRN1) (TRL1 URA3) (TRL1) (RTCB) (10x- (TRL1 (PGK1 Table 3.1 Table

94

FOA FOA FOA FOA FOA FOA ::KanMx ::KanMx ∆ ∆ dxo1 xrn1 Δ Δ Δ 0 trp1 0 trp1 0 trp1 Δ Δ Δ

sL) sL) sL) sL) 0ura3 0ura3 0ura3 Δ Δ Δ 0 met15 0 met15 0 met15 Δ Δ Δ ::HygMx ::HygMx 1 leu2 1 leu2 1 leu2

∆ Δ Δ Δ ::kanMx Δ Description his3 MATa his3 MATa with (XRN1) SPT02 transformed his3 MATa his3-11,15can1-100 ade2-1 ura3-1 trp1-1 MATa leu2-3,112 trl1 plasmidshufflewith 5- (TRL1), on SPT05 transformed with (PGK- SPT06 transformed plasmidshufflewith 5- on (RTCB), SPT05 transformed with (PGK- SPT08 transformed plasmid5- withon shuffle (10x-tRNA), SPT05 transformed with (TRL1-D425N) SPT10 transformed xrn1 SPT05 plasmidshufflewith 5- (TRL1), on SPT12 transformed with (PGK- SPT14 transformed plasmidshufflewith 5- on (RTCB), SPT12 transformed with (PGK- SPT16 transformed plasmid5- withon shuffle (10x-tRNA), SPT12 transformed with (TRL1-D425N) SPT18 transformed

-D425N)

sL) sL)

-D425N)

sL) sL) - (TRL1

tRNA) (TRL1 tRNA)

(TRL1 URA3) (TRL1 (TRL1) (PGK1-(TRL1) (RTCB) (PGK1- (RTCB) (10x- (10x- tRNA) tRNA) ∆ ∆ ∆ ∆ ∆ ∆ ∆ strains.

(XRN1) ∆ (TRL1 URA3) (TRL1 (TRL1) (PGK1-(TRL1) (RTCB) (PGK1 (RTCB) (10x- (10x- xrn1 xrn1 xrn1 xrn1 xrn1 xrn1 xrn1 ∆ ∆ ∆ ∆ ∆ ∆ ∆ ∆ ∆ ∆ ∆ ∆ ∆ ∆ ∆ ∆ Name XRN1 xrn1 xrn1 dxo1 trl1 trl1 trl1 trl1 trl1 trl1 trl1 trl1 trl1 trl1 trl1 trl1 trl1 trl1

Index SPT01 SPT02 SPT03 SPT04 SPT05 SPT06 SPT07 SPT08 SPT09 SPT10 SPT11 SPT12 SPT14 SPT15 SPT16 SPT17 SPT18 SPT19 Table 3.2Yeast Table

95 pAG424-ccdB (created by removing the GPD promoter from pAG424-GPD-ccdB (Alberti et al. 2007), which has the TRP1 gene and a 2µ origin). The pDONR221-10X-tRNA, pAG424-ccdB (empty vector) and pAG424-10X-tRNA (10x-tRNA) plasmids will be made available from Addgene.

Yeast spot assays

Yeast strains were spotted on –Trp, -Ura, and 5-Fluoroorotic Acid (5-FOA) SC plates in serial dilutions. Yeast were grown at 30ºC and imaged at 5-7 days.

Yeast culture conditions and RNA preparation

Single colonies were inoculated in selective drop-out media supplemented with relevant amino acids and incubated at 30˚C overnight with rotation. For UPR induction, overnight cultures were diluted to OD600 of 0.2, grown to mid-log phase, and treated for 2 hours with tunicamycin (final concentration of 2.5 µg/mL, Sigma-Aldrich) or DMSO. For stem-loop reporter experiments, overnight cultures were pelleted, washed, resuspended in

YEP with 2% raffinose and 2% galactose, and grown to mid-log phase. In both UPR and stem-loop experiments, cells were harvested by centrifugation and total RNA was isolated by hot acid phenol extraction. Total RNA was treated with TURBO DNase (Ambion) to remove genomic DNA.

RT-PCR and qRT-PCR

DNase-treated RNA was reverse transcribed with Protoscript II RT (New England Biolabs) using degenerate primers. HAC1 and PGK1 were PCR amplified from cDNA using

96 oligonucleotides listed in Table 3.3. Products were run on a 1.5% agarose TAE gel, stained with CYBR-Gold (Life Technologies) and imaged with AlphaImager HP (ProteinSimple).

Densitometry was performed with ImageJ. cDNA was also subjected to quantitative real- time PCR (qRT-PCR) using the LightCycler 480 SBYR Green I kit (Fischer Scientific) with the KAR2 F/R and PGK1 F/R primers. qRT-PCR was perfomed on a LightCycler 480

(Roche).

Primer extension

Primers against the 3´-UTR of HAC1 mRNA and snRNA U6 were end-labeled with

PNK (New England Biolabs) and γ32P-ATP (Perkin Elmer). Labeled primers were purified by Illustra MicroSpin G-25 columns (GE Healthcare), precipitated with 100% ethanol and washed with 75% ethanol. Radiolabelled primers and DNase-treated RNA

(5 µg) were denatured in TRIS, pH 8.0, 31 mM NaCl, 6.21 mM DTT at 90ºC for 3 minutes, folded at 55 ºC for 3 minutes, and then immediately cooled in liquid nitrogen. GoScript (1 U, Promega) was added in a final concentration of 533uM dNTPs,

3.23 mM MgCl2, 6mM DTT, 80mM KCl, and 55 mM Tris pH 8.3. Primers were extended for 45 minutes at 42ºC and 15 minutes at 55ºC. GoScript RTase was inactivated by heating for 15 minutes at 65ºC. RNA was hydrolyzed with 10mM NaOH at 90ºC for 3 minutes.

Formamide loading dye was added and ssDNA products were run on a 6% TBE 7M Urea gel. The gel was exposed on a phosphor-imager screen and imaged on a Typhoon 9400

(GE Healthcare).

97

Table 3.3. Oligonucleotides

Name Sequence HAC1 3´-UTR ACCCTCTTGCGATTGTCTTC U6-67 GAACTGCTGATCATCTCTG HAC1 F CATGGGAGCTGCAGATGTT HAC1 R AAATGAATTCCAACCTGACTGC PGK1 F TCTTAGGTGGTGCCAAAGGTT PGK1 R GCCTTGTCGAAGATGGAGTC KAR2 F AAGACAAGCCACCAAGGATG KAR2 R AGTGGCTTGGACTTCGAAAA PGK1 3´-NGD CGGATAAGAAAGCAACACCTGG

98 Northern blotting

For stem-loop reporter northern blots, 5µg of total RNA was run on a 6% TBE 7M

Urea gel. RNA was transferred to nylon membrane (Hybond N+, GE) and UV-crosslinked.

The PGK1 3´-NGD probe was end-labeled with PNK (New England Biolabs) and γ32P-

ATP (Perkin Elmer) and incubated with the membrane in ULTRAhyb-Oligo Buffer

(Thermo Fisher Scientific) at 42ºC overnight. The membrane was exposed on a phosphor- imager screen and imaged on a Typhoon 9400 (GE Healthcare).

Results

Xrn1 and Trl1 are required for degradation of the HAC1 3´ exon.

Previous studies showed that HAC1 3´-exon cleavage products can accumulate with

5´-PO4 termini in dcp2∆ xrn1∆ (Harigaya and Parker 2012) and xrn1∆ yeast cells (Peach et al. 2015) (Chapter II). This suggests that a 5´-OH RNA kinase must phosphorylate the

HAC1 3´-exon prior to decay by Xrn1. Using a radiolabelled primer complementary to the

3´-UTR of HAC1 mRNA, I performed primer extensions on several samples and analyzed the 74 nt predicted HAC1 3´-exon product (Figure 3.1A). The bands representing the

HAC1 3´-exon occur in one of three configurations in this assay. If the HAC1 3´-exon is present and has a 5´-OH terminus, it cannot be degraded by 5´-phosphate-dependent 5´→3´ and thus appears as the full 74 nt product (Figure 3.1B, lane 1). If the HAC1

3´-exon exists with a 5´-PO4 terminus and Xrn1 is present, it is degraded, and does not appear in the assay. Finally, if the HAC1 3´-exon exists with a 5´-PO4 terminus and Xrn1

99 is deleted, it is shortened 1 to 3 nt by 5´-phosphate-dependent 5´→3´ exonucleolytic

“nibbling”, which has been previously reported for 5´-PO4 RNAs in xrn1∆ yeast (Muhlrad et al. 1995; Hsu and Stevens 1993). Thus, a HAC1 3´-exon with a 5´-PO4 terminus appears as a doublet slightly below the expected product (Figure 3.1B, lane 2). In the xrn1∆ strain

(BY4741 background), the HAC1 3´-exon accumulates with a 5´-PO4 terminus in the absence of UPR induction (Figure 3.1C, lane 3), and the accumulation of the HAC1 3´- exon with a 5´-PO4 terminus increases upon UPR induction with N-linked glycosylation inhibitor tunicamycin (Tm) (Figure 3.1C, lane 4). Normal degradation of the HAC1 3´- exon is restored when XRN1 is expressed from the genome (Figure 3.1C, lanes 1, 2). or from a plasmid (Wu and Hopper 2014) (Figure 3.1C, lanes 5, 6). Rai homolog Dxo1 was recently shown to have distributed 5´→3´ exoribonuclease activity (Chang et al. 2012), but the dxo1∆ strain does not accumulate the HAC1 3´-exon, suggesting it is nonessential for degradation of the HAC1 3´-exon (Figure 3.1C, lanes 7, 8). Taken together, these findings indicate that the HAC1 3´-exon is phosphorylated, and that the degradation of the phosphorylated HAC1 3´-exon fragment is primarily accomplished by Xrn1-mediated

5´→3´ decay (Figure 3.1C).

E. coli RtcB does not have 5´-OH kinase activity, but can ligate HAC1 mRNA and intron-containing tRNAs in yeast via an orthogonal chemistry, catalyzing ligation of 2´,3´- cyclic phosphate and 5´-OH termini without a 5´-PO4 intermediate (Tanaka et al. 2011).

Given that RtcB can rescue Trl1 deletions, but does not replace its 5´-OH kinase activity,

I compared HAC1 3´-exon accumulation in the presence (TRL1) or absence (RTCB) of cytosolic 5´-OH kinase activity using W303 plasmid shuffle strains (Tanaka et al. 2011).

When Trl1 is present and Xrn1 is deleted, the HAC1 3´-exon fragment accumulates with a

100 5´-PO4 terminus (indicated by the lower doublet of nibbled 5´-PO4 fragments), and this accumulation increases upon Tm treatment (Figure 3.1D, lanes 3, 4), consistent with the

BY4741 xrn1∆ strain. However, when E. coli RtcB replaces Trl1, and no cytosolic 5´-OH kinase activity is present, the HAC1 3´-exon never accumulates with a 5´-PO4 terminus, even following Tm treatment in the absence of Xrn1 (Figure 3.1D, lanes 7, 8). This indicates that Trl1 phosphorylates the 5´ terminus of the HAC1 3´-exon and that Xrn1 degrades this fragment. In comparison to the BY4741 strains, all trl1∆ W303 strains accumulate a HAC1 3´-exon fragment with a 5´-OH terminus (indicated by the full length fragment which is immune to decay by Xrn1) in the absence of induction (Figure 3.1D, lanes 1, 3, 5, 7), and this fragment increases with Tm treatment (Figure 3.1D, lanes 2, 4,

6, 8). This may indicate inefficient ligation of the exogenously expressed ligases Trl1 and

RtcB.

The 5´-OH kinase activity of Trl1 is required for degradation of the HAC1 3´-exon.

I next sought to specifically test the role of Trl1 5´-kinase activity in the decay of the HAC1 3´-exon. In yeast, Trl1 is essential because it is required for the ligation of tRNA halves following cleavage by the TSEN complex (Greer et al. 1983). Similar to previous studies in C. elegans (Kosmaczewski et al. 2014), we constructed a plasmid that expresses all 10 intron-containing yeast tRNAs in their “pre-spliced” forms. This (10x-tRNA) plasmid allows us to bypass the essential tRNA splicing activity of Trl1 (Figure 3.2A). To validate the rescue of trl1∆ cells with “pre-spliced” tRNAs, we performed a plasmid shuffle and found that trl1∆ could be complemented by the (10x-tRNA) plasmid (Figure 3.2B).

101

. product -exon cleavage-exon 3´

HAC1 phosphorylated he . Xrn1 degrades .Xrn1 t degrades Figure 3.1

102 Figure 3.1. Xrn1 degrades the phosphorylated HAC1 3´-exon cleavage product. A. Diagram of HAC1 mRNA and the HAC1 3´UTR primer, which is extended with reverse transcriptase. The expected primer extension product length of the HAC1 3´exon cleavage product is 74 nt. B. Schematic of HAC1 3´-exon bands in the primer extension assay. A HAC1 3´-exon with a 5´-OH terminus escapes decay by 5´-phosphate-dependent exoribonucleases and appears as a single 74 nt band on the gel (Lane 1). A HAC1 3´-exon with a 5´-PO4 terminus is rapidly decayed by Xrn1 and does not appear on the gel (no band). A HAC1 3´-exon with a 5´-PO4 terminus is stabilized in the absence of Xrn1, but shortened by 5´-phosphate- dependent exonucleolytic “nibbling” (Muhlrad et al. 1995; Hsu and Stevens 1993), and appears as a 73 nt and 71 nt doublet slightly below the expected length (Lane 2). U6 primer extension serves as a positive control (67 nt). C. XRN1, xrn1∆, xrn1∆ (XRN1), and dxo1∆ BY4741 strains in log phase were treated with 2.5µg/mL tunicamycin (Tm) or DMSO for 2 hours. Radiolabelled HAC1 3´UTR primer was extended on DNase-treated total RNA isolated from these strains. Primer extension of U6 RNA (67 nt) was used as a loading control. Band intensity was quantified with ImageJ, and is displayed in the bar graph immediately below the gel image as percentage of HAC1 primer extension signal with a 5´-PO4 or 5´-OH terminus. The HAC1 3´-exon accumulates with a 5´-PO4 terminus in the xrn1∆ strain (indicated by lower doublet bands derived from “nibbled” HAC1 3´-exon fragments). This accumulation increases with Tm treatment (Lanes 3, 4). The (XRN1) plasmid restores degradation of the intermediate in the xrn1∆ strain (Lanes 5, 6), and dxo1∆ has no effect on HAC1 3´-exon accumulation (Lanes 7, 8). D. trl1∆ and trl1∆ xrn1∆ W303 strains with expression of either Trl1 or RtcB ligase were treated and evaluated as in C. The HAC1 3´-exon accumulates primarily with a 5´-PO4 terminus (indicated by the lower doublet bands derived from the “nibbled” HAC1 3´-exon) in the trl1∆ xrn1∆ (TRL1) strain and this accumulation increases with Tm treatment (Lanes 3, 4). The trl1∆ xrn1∆ (RTCB) strain does not accumulate the HAC1 3´-exon with a 5´-PO4 terminus (Lanes 7, 8). All strains accumulate the HAC1 3´-exon with a 5´-OH terminus (indicated by the single upper band) independent of Xrn1 presence or Trl1 or RtcB expression, and this accumulation increases with Tm treatment.

103 I measured the HAC1 3´-exon mRNA intermediate by primer extension in trl1∆ and trl1∆ xrn1∆ W303 yeast strains expressing the (10x-tRNA) plasmid alone or the (10x- tRNA) plasmid and a plasmid encoding kinase-dead TRL1-D425N (Sawaya 2003) (Figure

3.2C). In the absence of Trl1 5´-kinase activity, the HAC1 3´exon accumulates with a

5´-OH terminus (indicated by the single upper band) (Lanes 1, 3, 5, 7), and the accumulation increases when the cells are treated with tunicamycin (Lanes 2, 4, 6, 8). There is an increase in the 5´-OH HAC1 3´-exon fragment in Lane 4 compared to Lane 2; this may indicate increased stabilization of the HAC1 pre-mRNA in the absence of Xrn1 in the

W303 strain, which in turn results in increased Ire1 cleavage.

Taken together, my results thus far indicate that Trl1 phosphorylates the 5´-OH terminus of the HAC1 3´-exon, and the phosphorylated HAC1 3´-exon is then degraded by

Xrn1. This represents the third example of KMD in eukaryotes, and the first shown to occur on eukaryotic mRNA cleavage fragments.

Xrn1 5´→3´ decay of the phosphorylated HAC1 3´-exon is required to suppress aberrant

HAC1 mRNA splicing and UPR activation

Given the regulated decay of the HAC1 3´-exon intermediate, I next asked whether

Xrn1 might also affect HAC1 splicing and UPR activation (Figure 3.3). In the absence of

UPR induction, HAC1 mRNA splicing is significantly elevated in the xrn1∆ strain compared to the XNR1 replete strain (Figure 3.3A, B, Lanes 1, 3). Furthermore, HAC1 mRNA is nearly completely spliced in the xrn1∆ strain after treatment with Tm (Figure

3.3A, B, Lane 4). Expression of Xrn1 in the xrn1∆ strain restores HAC1 splicing to wildtype levels (Figure 3.3A, B, Lanes 5, 6), and deletion of Dxo1, the minor 5´→3´

104

Xrn1. to degradation promote by Trl1

-exon cleavageactivityrequires of mRNA -exon kinase the product 3´ HAC1 HAC1 The The

Figure 3.2. Figure 105

Figure 3.2. The HAC1 3´-exon mRNA cleavage product requires the kinase activity of Trl1 to promote degradation by Xrn1. A. Diagram of the 10x-tRNA plasmid used to rescue trl1∆ yeast strains. “Pre-spliced” forms of the 10 intron-containing tRNAs in yeast are encoded in tandem on a high-copy TRP1 plasmid. Expression of a single transcript is driven by a SUP4 promoter and terminated by the RPR1 terminator. B. Spot assays demonstrate rescue of trl1∆ with the (10x-tRNA) plasmid. Serial dilutions of trl1∆ (TRL1 URA3) (10x-tRNA) or trl1∆ (10x-tRNA) were spotted onto –trp, –ura, and 5-FOA plates. Plates were photographed after 5-7 days of incubation at 30˚C. The trl1∆ strain complemented with the (10x-tRNA) plasmid is viable on –trp plates, but is slow growing compared to strains with Trl1 expression. The trl1∆ (10x-tRNA) strain does not grow on -ura plates but is viable on 5-fluoroorotic acid (5-FOA) plates, demonstrating loss of the (TRL1 URA3) plasmid and rescue of trl1∆ with the (10x-tRNA) plasmid. Spot assay data provided by Laura White. C. Primer extension with radiolabelled HAC1 3´UTR and U6 primers was performed on DNase-treated total RNA isolated from yeast strains where essential tRNA ligation is bypassed by the (10x-tRNA) plasmid. Band intensity of the HAC1 3´-exon fragment with a 5´-OH terminus was quantified with ImageJ, normalized to trl1∆ (10x-tRNA), and is displayed in the bar graph immediately below the gel image. Kinase-dead Trl1 (TRL1- D425N) cannot restore normal Xrn1-mediated degradation of the HAC1 3´-exon.

106 cytosolic exoribonuclease, has little effect on HAC1 mRNA splicing (Figure 3.3, 4, Lanes

7, 8). These results indicate that Xrn1 normally suppresses baseline HAC1 mRNA splicing.

If the increase in HAC1 mRNA splicing leads to a rise in Hac1 protein levels, I would expect elevated transcription of Hac1 target genes indicating activation of the UPR.

Indeed, in the absence of Xrn1, expression of the Hac1 transcriptional target KAR2 (Kohno et al. 1993) was elevated 4-fold above XRN1 cells without chemical UPR induction (Figure

3.3C, Lane 3). This modest but significant elevation was reduced in the other strains tested.

Following UPR induction, all strains induce KAR2 to similar levels, perhaps reaching a level of maximal upregulation.

These results collectively demonstrate the multi-step process of kinase-mediated decay. HAC1 mRNA is initially cleaved by Ire1, generating a 3´-exon fragment with a

5´-OH terminus that cannot be degraded by Xrn1. Phosphorylation by Trl1 5´-kinase activity creates a 5´-PO4 terminus, which allows Xrn1 to degrade the HAC1 3´exon intermediate. The decay of this intermediate prevents ligation by Trl1, which inhibits

HAC1 splicing and UPR activation. Thus, KMD is required to control UPR activation in yeast.

Trl1 phosphorylates no-go decay 3´-fragments prior to degradation by Xrn1.

We asked whether KMD might play a broader role in 5´-OH RNA surveillance.

RNAs with 5´-OH termini accumulate upstream of consecutive basic residues (Peach et al.

2015) (Chapter II), which are known to trigger no-go mRNA decay (Parker 2012), suggesting that NGD 3´ fragments may be generated with 5´-OH termini. However, because Xrn1 is required for the decay of NGD 3´ cleavage products (Doma and Parker

107

*

Figure 3.3. Xrn1 suppresses HAC1 mRNA splicing and UPR activation.

108 Figure 3.3. Xrn1 suppresses HAC1 mRNA splicing and UPR activation. A. The indicated strains were grown to log phase and treated for 2 hours with 2.5µg/mL tunicamycin (Tm) or DMSO. RT-PCR across the HAC1 mRNA splicing junction was used to assess HAC1 splicing. Unspliced HAC1 is the upper band at 461 bp, and spliced HAC1 is lower band at 209 bp. A minor intermediate band of an unknown species is indicated with an asterisk. Gel is representative of three biological replicates. B. Quantification of HAC1 splicing in A for three biological replicates. The calculation for HAC1 mRNA splicing percentage is: (spliced HAC1) / (spliced HAC1 + unspliced HAC1) × 100. White bars indicate DMSO (control) treatment and black bars indicate Tm treatment. C. KAR2 expression from the samples in A was evaluated by qRT-PCR. KAR2 expression is normalized to PGK1 expression for each sample, then compared to the normalized KAR2 expression in the XRN1 strain for each of three replicates. B, C: **p<0.01, ***p<0.001, ***p<0.0001. Student’s t-test (2-tailed). Error bars are SEM.

109 2006) we asked whether Trl1 5´-kinase phosphorylates the 5´-OH terminus of NGD products prior to their degradation by Xrn1.

To evaluate the decay of NGD fragments, we used a NGD reporter (Doma and

Parker 2006) consisting of a long stem-loop inserted in the PGK1 gene under the control of the GAL1 promoter (PGK1-sL) (Figure 3.4A). This stem-loop causes elongating ribosomes to stall, leading to mRNA cleavage upstream of the stem-loop. Patrick Cherry expressed this reporter in four yeast strains and evaluated NGD 3´-fragment accumulation following induction (Figure 3.4B). PGK1-sL 3´ NGD cleavage fragments are degraded in strains expressing 5´-OH RNA kinase Trl1, but accumulate in the absence of Xrn1, confirming previous studies (Doma and Parker 2006). However, when Trl1 is replaced by the orthogonal E. coli RNA ligase RtcB, PGK1-sL 3´ NGD cleavage fragments accumulate at elevated levels in the presence or absence of Xrn1, indicating that Trl1 acts upstream of

Xrn1.

Thus, a PGK1-sL 3´ NGD cleavage fragment is generated with a 5´-OH terminus that must be phosphorylated by Trl1 prior to 5´-phosphate-dependent 5´→3´ degradation by Xrn1. KMD of NGD 3´-fragments is the second example of eukaryotic mRNA cleavage fragment degradation by KMD, and indicates that KMD may play a broader role in 5´-OH mRNA surveillance.

Discussion

I have described two examples of kinase-mediated decay (KMD) in which RNA

5´-OH kinase Trl1 phosphorylates a cleaved 5´-OH mRNA and promotes 5´-phosphate- dependent 5´→3´ decay by Xrn1 (Figure 3.5). In the first example, the 3´-exon of HAC1

110

Figure 3.4. Trl1 phosphorylates NGD 3´ fragments prior to 5´→3´ degradation by Xrn1. A. Diagram of no-go decay reporter PGK1-sL. A strong stem-loop inserted in PGK1 mRNA causes ribosome stalling and triggers NGD. PGK1-sL fragments detected by Northern blot with a radiolabelled 3´-PGK1 probe are indicated below the reporter cartoon. B. Northern blot analysis with the 3´-PGK1 probe was performed on RNA isolated from yeast strains grown in YEP-galactose-raffinose to induce expression of PGK1-sL. PGK-sL 3´ NGD fragments accumulate when XRN1 is deleted, or when Trl1 is replaced by RtcB, which does not have 5´-OH kinase activity. Experimental data provided by Patrick Cherry.

111 mRNA is phosphorylated by Trl1 prior to 5´→3´ degradation by Xrn1. KMD of the HAC1

3´-exon suppresses HAC1 mRNA splicing and UPR activation in the absence of true stimulus (Figure 3.5A). In the second example, NGD 3´ cleavage fragments are generated with 5´-OH termini and must be phosphorylated by Trl1 to become competent substrates for 5´-phosphate-dependent Xrn1 degradation (Figure 3.5B). These two examples, combined with the two previous examples of eukaryotic KMD for processing tRNA introns

(Wu and Hopper 2014) and 25S rRNA maturation (Gasse et al. 2015), help make the case that KMD is a fundamental, multi-purpose pathway of regulated RNA decay in yeast.

The mechanism of NGD mRNA cleavage remains mysterious. The proposal that

Dom34 is the NGD endonuclease is controversial (Lee et al. 2007; Passos et al. 2009), and the reduction in NGD cleavage in the absence of ribosomal small subunit proteins is likely due to reductions in overall levels of translating ribosomes (Bhattacharya et al. 2010).

Other characteristics of NGD cleavage suggest that it may not be catalyzed by an endonuclease: NGD cleavage is inefficient, with only a small percentage of mRNA substrate cleaved (Doma and Parker 2006), and cleavage can occur in a variety of sequence contexts (Tsuboi et al. 2012). Patrick Cherry’s data show that NGD mRNA cleavage products have 5´-OH termini, suggesting that the NGD cleavage might be catalyzed by intrinsic phosphoester transfer promoted by stalled ribosomes.

KMD is likely required for the decay of 5´-OH RNA cleavage products generated by many metal-ion independent endoribonucleases. For instance, in higher-order eukaryotes, Ire1 homologs cleave ER-proximal mRNAs during ER stress in a process termed Regulated IRE1-Dependent Decay (RIDD) (Hollien and Weissman 2006). The cleavage chemistry of Ire1 dictates that the 3´ mRNA fragments must have 5´-OH termini.

112

Figure 3.5. Models of kinase-mediated decay of mRNA cleavage fragments. A. (1) Upon sensing unfolded proteins in the ER, Ire1 cleaves HAC1 mRNA, leaving a 5´- OH terminus on the 3´ exon. (2) The terminus is phosphorylated by Trl1. (3) Normally, the 3´-exon is degraded by Xrn1. (4) However, in the absence of Xrn1, the RNA is available for ligation resulting in splicing of HAC1 mRNA and aberrant activation of the UPR. B. (1) When a translating ribosome stalls at a strong secondary structure in mRNA, endonucleolytic cleavage leaves a 5´-OH terminus on the 3´ NGD fragment. (2) Phosphorylation by Trl1 creates a substrate for degradation by Xrn1 (3).

113 However, these fragments are normally degraded by 5´-phosphate-dependent 5´→3´ decay by Xrn1, which requires that the 3´-fragments are first phosphorylated (Hollien and

Weissman 2006).

Under certain stress conditions including oxidative stress and nutrient starvation, yeast Rny1 and mammalian Angiogenin cleave the anticodon loop of some tRNAs

(Thompson and Parker 2009b; Yamasaki et al. 2009; Fu et al. 2009). These stress-induced tRNA fragments (tiRNAs) can then repress translation and apoptosis (Ivanov et al. 2011;

Saikia et al. 2014). The 3´-tiRNAs are produced with 5´-OH termini, but it is not known how they are degraded. If 3´-tiRNAs are decayed by nuclear 5´-phosphate-dependent exoribonuclease Rat1/XRN2, they must first be phosphorylated.

A third example of potential KMD substrates are the targets of RNase L (Cooper et al. 2014). During viral infection of human cells, RNase L cleaves both viral and host RNAs, which function as ligands for RIG-I and MDA5 (Malathi et al. 2007; 2010). Although it is known that deletion of cytosolic exosome component SKIV2L limits the activation of RIG-

I-like receptors (RLRs) and the production of type I interferons (Eckard et al. 2014), it is not known whether cytosolic decay factors directly affect the RNase L cleavage fragments.

The 3´-fragments of RNase L are produced with 5´-OH termini, and if they are degraded by Xrn1, would also require phosphorylation of the 5´-terminus.

In humans, the UPR is a tripartite signaling system regulated by XBP1, ATF4, and

ATF6 proteins. XBP1 mRNA, the human homolog of HAC1, is cleaved in the cytoplasm by activated IRE1α, which generates a 5´-exon fragment with a 2´,3´-cyclic phosphate terminus and a 3´-exon fragment with a 5´-OH terminus. The two exons are joined by

RTCB, which first resolves the cyclic phosphate to a 3´-PO4 prior to ligation (Lu et al.

114 2014; Chakravarty and Shuman 2012). The RNA terminal phosphate cyclase RTCA may directly antagonize RTCB by converting the 3´-PO4 terminus back to a 2´,3´-cyclic phosphate. In Drosophila, loss of function mutations in Rtca enhances Xbp1 splicing and axonal growth (Song et al. 2015). Similarly, decreased Rtca expression in the adult mouse retinal ganglion cells promotes axon regeneration after optic nerve crush(Song et al. 2015).

An exome screen of consanguineous families with children afflicted by severe neurodevelopmental deficits identified a single site homozygous mutation in CLP1 that reduces RNA 5´OH-kinase activity in vitro (Hanada et al. 2013; Karaca et al. 2014;

Schaffer et al. 2014). Although patient-derived fibroblasts with the homozygous CLP1R140H mutation accumulate aberrant tRNA fragments, the affect on other RNAs was not assessed.

Given that RTCA may directly oppose RTCB ligation of XBP1 by cyclizing the 3´-PO4 terminus of the 5´-exon, perhaps CLP1 can oppose RTCB ligation by phosphorylating the

5´-OH terminus of the 3´-exon. Phosphorylation of the XBP1 3´-exon by CLP1 would also promote 5´-phosphate-dependent decay by Xrn1, which would be yet another example of eukaryotic KMD.

Summary

In this chapter, I followed up on a mysterious finding: HAC1 3´-exons are captured with 5´-PO4 termini in xrn1∆ yeast (Harigaya and Parker 2012; Peach et al. 2015) (Chapter

II), despite being generated with 5´-OH termini following Ire1-mediated cleavage. By analyzing HAC1 3´-exon accumulation and HAC1 mRNA splicing in yeast strains with

Xrn1 and Trl1 deletions, I demonstrated the first example of kinase-mediated decay (KMD) of a eukaryotic mRNA. Even in the absence of ER stress, Ire1 cleaves HAC1 pre-mRNA

115 at low levels, generating a 3´-exon mRNA fragment with a 5´-OH terminus. 5´-OH RNA kinase Trl1 phosphorylates the 5´-OH terminus, yielding a 5´-PO4 RNA that is then degraded by 5´-phosphate-depedent 5´→3´ exoribonuclease Xrn1. KMD of the cleaved

3´-exon of HAC1 is necessary to prevent HAC1 ligation and suppress activation of the UPR in the absence of true ER stress.

KMD is also responsible for the turnover of no-go decay (NGD) 3´-fragments. Prior to this study, the nature of the 5´ terminus of the 3´-fragment was unknown. In this chapter,

I shared Patrick Cherry’s data, which show that the 3´ NGD fragment from the PGK-sL reporter is generated with a 5´-OH terminus. Trl1 phosphorylates the 5´-terminus of this 3´

NGD fragment, enabling recognition and decay by Xrn1. Given that NGD occurs on a variety of mRNA substrates, this suggests that KMD functions as a broad surveillance system for 5´-OH mRNAs.

116 CHAPTER IV

ANALYSIS OF 5´-HYDROXYL RNAS IN MAMMALS

Introduction

Cleavage and polyadenylation factor 1 (CLP1) is an essential mammalian polyribonucleotide 5´-OH kinase that is a component of both the pre-mRNA cleavage complex II (de Vries et al. 2000) and the tRNA splicing endonuclease (TSEN) complex

(Paushkin et al. 2004). CLP1 phosphorylates synthetic siRNAs in vivo prior to incorporation into the RNA-induced silencing complex (RISC) (Weitzer and Martinez

2007b). CLP1 also phosphorylates the 3´-exon and intron generated from cleavage of intron-containing tRNAs by TSEN (Weitzer and Martinez 2007b).

Studies of consanguineous families with childhood neurological disease identified nine pedigrees with a single mutation in CLP1 (Schaffer et al. 2014; Karaca et al. 2014).

Affected children with a homozygous CLP1R140H mutation develop motor-sensory defects, cortical dysgenesis, and microcephaly. They also demonstrate structural defects of the pons and cerebellum consistent with pontocerebellar hypoplasia (PCH). Studies of patient- derived fibroblasts with the kinase-impaired CLP1R140H mutation and a mouse model with a kinase-dead CLP1K127A mutation showed differential accumulation of tRNA fragments including intron-containing pre-tRNAs (Schaffer et al. 2014), introns processed from tRNAs (Karaca et al. 2014), and 5´ exons (Hanada et al. 2013) when CLP1 kinase activity is compromised, though it is not known whether these fragments are causative of disease.

In Chapter II, I described the development of the 5OH-Seq method and its application to transcriptome-wide analysis of 5´-OH RNA in yeast. Given the severe

117 neurodevelopmental phenotypes causes by single site mutations in the CLP1 mammalian kinase, I sought to use 5OH-seq to identify additional targets of CLP1 in mouse and humans. Here, I demonstrate the utility of the 5OH-Seq method on three sets of mammalian samples with impaired CLP1 kinase function: human 293FT cells with CLP1 shRNA knockdown; patient-derived fibroblasts with homozygous and heterozygous CLP1R140H mutations; and mouse embryonic fibroblasts (MEFs) and spinal cords from mice with kinase-dead CLP1K127A mutations. I show enhanced 5´-OH RNA accumulation around

Poly-A sites in human cell lines with decreased CLP1 expression and identify a novel SeC tRNA anti-codon loop cleavage in spinal cords of mice but not in MEFs. These studies demonstrate the utility of the 5OH-seq in identification of new 5´-OH RNA fragments in mammalian studies.

Materials and Methods

Generation and validation of human cell lines

Following standard protocols, 293FT cell lines were transduced with one of three vectors targeting CLP1 mRNA via shRNA (Table 4.1, shRNA #1-3) or a non-targeting control vector, along with packaging plasmids VSV-G and ∆8.9. Transduced cells were selected with puromycin. CLP1 shRNA knockdown cell lines were validated by qRT-PCR with CLP1 qPCR F/R and Actin qPCR F/R primers. CLP1 shRNA knockdowns were further validated by western blots with primary antibodies against CLP1 (ab133669,

Abcam) and Actin (15939, Pierce) and fluorescent secondary antibodies (DyLight488, ab96875, Abcam). Patient-derived fibroblasts were a kind gift from the Lupski lab

118 Table 4.1. Oligonucleotides

shRNA targeting sequences shRNA #1 CCGGCCTGGCACTAACATCAAGCTTCTCGAGA (TRCN0000015041) AGCTTGATGTTAGTGCCAGGTTTTT shRNA #2 CCGGGCTGACCCGAAACAAGAAATTCTCGAGA (TRCN0000329954) ATTTCTTGTTTCGGGTCAGCTTTTTG shRNA #3 CCGGTGGTAACCTGACTCTTCTAATCTCGAGAT (TRCN0000329955) TAGAAGAGTCAGGTTACCATTTTTG

Validation Primers Actin qPCR F GGATGCAGAAGGAGATCACTG Actin qPCR R CGATCCACACGGAGTACTTG Clp1 qPCR F AAATTCTGCACGGGAGGACC Clp1 qPCR R AGATGCCTCCACCTCAAAGC Clp1 F TGATGACAAGAAGCCAACCA Clp1 R TGTTAGTGCCAGGAGTGGTG

Northern Probes SeC tRNA 5´ AGCCTGCACCCCAGACCACTGAGGAT SeC tRNA 3´ GGTGGAATTGAACCACTCTGTCGCTAAACAGC

119 (Karaca et al. 2014). Genomic DNA was amplified by PCR with CLP1 F/R primers and the CLP1 R140H as validated by Sanger sequencing with the CLP1 qF primer.

RNA isolation and enrichment

293FT shRNA cell lines and patient-derived fibroblasts were grown in high- glucose DMEM (Thermo Fisher Scientific) supplemented with 10% FBS (Life

Technologies) and 1% penicillin/streptomycin (Life Technologies). RNA and DNA were isolated from human cell lines using the standard procedures for TRIzol reagent (Ambion).

From 293FT shRNA cell lines, PolyA+ RNA was enriched from total RNA using oligo- dT(25) dynabeads (Ambion). Patient-derived fibroblast samples were split in half. One half was enriched for PolyA+ RNA with oligo-dT beads, the other half was depleted of rRNA using Ribo-Zero Gold rRNA Removal Kit (Illumina). Total RNA from MEFs and mouse spinal cords was received as a gift from the Martinez lab (Hanada et al. 2013). This total

RNA was also depleted of rRNA using the Ribo-Zero kit.

Library preparation

5OH-Seq libraries from all samples were prepared essentially as described in

Chapter II (Peach et al. 2015). PolyA-seq libraries were prepared from the 293FT shRNA cell lines according to the published protocol (Derti et al. 2012). All libraries were sequenced by Illumina HiSEQ 2000/2500 with single 50bp reads. Raw data from all libraries are deposited in GEO, accession number pending.

120 Data analysis

5OH-Seq data was preprocessed as described in Chapter II with additional downstream analysis. For analysis of mouse tRNAs, I aligned 5OH-seq libraries to a custom tRNA database created using annotated tRNAs and including 15 nucleotides upstream and downstream to capture leader sequences and potential processing defects.

To compare 5OH-signal density surrounding mRNA features, I analyzed 40 bins of 25 bases to capture 500 nucleotides upstream and downstream of a given feature. I calculated the average 5OH-seq signal within each bin for each mRNA and summed these means. All

Python, R, and bash scripts are available at https://github.com/hesselberthlab/5OH.

Northern blot

DNA oligonucleotides complementary to the 3´ and 5´ halves of SeC tRNA were end-labeled with PNK (New England Biolabs) and γ32P-ATP (Perkin Elmer). Total RNA

(5µg) from mouse spinal cords was run on a denaturing 10% TBE-Urea gel (Thermo

Fisher). RNA was transferred to positively-charged nylon membrane (Hybond-N+

Membrane, GE Healthcare Amersham) and fixed with UV light. Blocking and hybridization was performed in ULTRAhyb-oligo buffer (Thermo Fisher Scientific) at

42ºC. Membranes were exposed on a phosphor-imager screen and imaged on a Typhoon

9400 (GE Healthcare).

121 Results

Increased 5OH-signal surrounding PolyA sites in CLP1 knockdown cell lines

CLP1 is a component of the pre-mRNA cleavage complex II, which processes 3´ ends of pre-mRNA to form the mature polyadenylated mRNA (de Vries et al. 2000). To assess whether decreased CLP1 expression would affect 5´-OH RNA formation near PolyA sites, our lab generated stable shRNA knockdowns of CLP1 mRNA in 293FT cell lines. I validated CLP1 knockdown by qRT-PCR (Figure 4.1A) and Western blot (Figure 4.1B).

I performed 5OH-seq on RNA isolated from validated CLP1 knockdown cell lines and analyzed 5OH-seq signals 500bp upstream and downstream of PolyA sites (Figure 4.2, top). All three CLP1 shRNA knockdown lines had enhanced 5OH-seq signal density surrounding PolyA sites compared to the scrambled control. As an additional control, I analyzed PolyA-seq signal around the same sites (Figure 4.2, bottom). Depletion of CLP1 had no effect on PolyA signal.

No significant differences between patient and parent CLP1R140H samples

CLP1 kinase activity is decreased more than 50% in patients with a homozygous

CLP1R140H mutation compared to wildtype CLP1 (Schaffer et al. 2014). To test whether

CLP1 might broadly phosphorylate 5´-OH RNA prior to 5´-phophate-dependent decay

(discussed further in Chapter III), I applied 5OH-seq to patient-derived fibroblasts from individuals with heterozygous or homozygous CLP1R140H mutations. I used the DESeq package (Anders and Huber 2010) to compare 5OH-seq signals of heterozygotes and

122

A

B

CLP1

Figure 4.1. CLP1 shRNA knockdowns validated by qRT-PCR and Western blot. A. CLP1 shRNA cell lines were validated by qRT-PCR. CLP1 signal is normalized to ACTIN signal, and compared to the scrambled control cell line. All CLP1 shRNA cell lines show at least 67% reduction in CLP1 mRNA levels compared to the scrambled control. B. CLP1 shRNA cell lines were validated by western blot with a primary antibody against CLP1. All CLP1 shRNA cell lines (Lanes 2-4) show at least a 50% reduction in CLP1 protein levels compared to the scrambled control (Lane 1).

123

Figure 4.2. CLP1-deficient cell lines have increased 5OH-signal surrounding PolyA sites. 5OH-seq signal (top) and PolyA-seq signal (bottom) was assessed within 500 nucleotides upstream and downstream of poly(A) sites. 40 bins of 25 bases were created upstream and downstream of known poly(A) sites. The average 5OH-seq signal within each bin for each mRNA was calculated and the sum of the means is plotted on the y axis. Position relative to the poly(A) sites is plotted on the x axis. CLP1 shRNA knockdown cell lines and the scrambled control have nearly identical PolyA-seq signals. CLP1 deficient cell lines (blue, green, purple) have an elevated signal surrounding poly(A) sites compared to the scrambled control (red). This difference is particularly pronounced upstream of PolyA sites.

124 homozygotes. Using an FDR of 0.1, I was unable to identify significantly different sites of

5´-OH RNA.

5´-OH signal increases in kinase-dead Clp1 samples

In the previous two experiments, CLP1 levels were decreased by shRNA knockdown or CLP1 functioning was decreased by a single amino acid change outside the

ATP-binding pocket that decreases but does not abolish kinase activity. In contrast, a single amino acid change in the Walker A motif of CLP1 completely abolishes CLP1 kinase activity (Hanada et al. 2013). To assess the affect of a kinase-dead CLP1 mutation on the accumulation of 5´-OH RNAs, I performed 5OH-seq on RNA from mice homozygous for a CLP1K127A mutation.

Prior to sequencing, I run libraries on a TBE gel to visualize library content.

Libraries from samples with kinase-dead CLP1 displayed more intensity compared to samples with wildtype CLP1, and this difference was particularly pronounced in samples from mouse spinal cords (Figure 4.3A).

I assessed the 5OH-seq signal 500bp upstream and downstream of several key genomic features: Poly-A signals, start codons, stop codons, and transcription start sites

(TSS) (Figure 4.3B). 5OH-seq libraries from MEFs had similar traces for all genomic features independent of CLP1 status (left panels). The 5OH-seq signals from CLP1+/+ spinal cord RNA were similar to MEFs, with the 4-month time point having a slightly lower signal than the 20-day time point. RNA from the spinal cords of mice with kinase- dead CLP1K127A/K127A had greatly diminished 5OH-seq signal, and this change is independent of time. Thus, the CLP1 mutation has a clear effect on 5OH-seq signal in the

125

Figure 4.3. CLP1 kinase-dead mutants differentially accumulate 5OH-RNAs in mouse spinal cords.

126 Figure 4.3. CLP1 kinase-dead mutants differentially accumulate 5OH-RNAs in mouse spinal cords. A. 5OH-seq libraries were prepared from RNA isolated from MEFs or spinal cords of mice with wildtype CLP1 or kinase-dead CLP1. For quality control, 5OH-seq libraries were run on a TBE gel after PCR amplification and before size-selection and sequencing. Lanes 1- 8 are the libraries sequenced and analyzed in B. Lanes 9-12 are mock libraries prepared from pre-fragmented yeast RNA in the presence or absence of RtcB ligase or Reverse Transcriptase (RT). Lane 13 is a no template (nt) PCR control. 5OH-seq libraries from kinase-dead CLP1 samples (lanes 2, 4, 6, 8) showed enhanced intensity compared to wildtype CLP1 samples (lanes 1, 3, 5, 7). B. 5OH-seq signal was assessed within 500 nucleotides upstream and downstream of Poly- A signals, start codons, stop codons, and transcription start sites (TSS). 5OH-seq libraries from MEF RNA have similar traces indicating no changes with CLP1 mutation or in different mouse backgrounds. Libraries from spinal cord RNA differ based on animal age and on CLP1 mutation status. CLP1 +/+: Wild-type CLIP1. CLP1 K/K: kinase dead CLP1K127A/K127A. SC: spinal cord.

127 spinal cords of affected mice. However, this is a general effect on mRNA and is present at all genomic features, rather than a feature-specific effect. Furthermore, these results do not support the hypothesis that CLP1 broadly phosphorylates mRNA prior to decay by Xrn1.

Novel SeC anti-codon loop tRNA cleavage identified by 5OH-seq

CLP1 phosphorylates the 3´-exon and intron of cleaved intron-containing tRNAs in vitro (Weitzer and Martinez 2007b), and mammalian cells with mutations in CLP1 accumulate several types of tRNA fragments (Hanada et al. 2013; Schaffer et al. 2014;

Karaca et al. 2014). I asked whether 5OH-seq could capture known or novel tRNA fragments in cells with kinase-dead CLP1. I aligned the 5OH-seq reads to a tRNA-specific genome and analyzed the summed signals for each tRNA species (Figure 4.4).

Surprisingly, the strongest signal was found in the anticodon loop of selenocysteine (SeC) tRNA in spinal cord RNA, though this signal was independent of CLP1 mutation (Figure

4.4B). During certain stress conditions, angiogenin can cleave tRNAs in the anticodon loop generating a 3´ stress-induced tRNA fragment (3´-tiRNA) with a 5´-OH terminus (Fu et al.

2009; Yamasaki et al. 2009); however, such a cleavage in SeC tRNA has not been described (Figure 4.5A). To validate the fragment by Northern blot, I designed probes against the 5´ and 3´ halves of the SeC tRNA (Figure 4.5B). I captured the full-length SeC tRNA with both probes but was unable to validate either tiRNA half (Figure 4.5C).

Discussion

Patients with homozygous CLP1R140H mutation suffer from pontocerebellar hypoplasia and severe neurological deficits (Schaffer et al. 2014; Karaca et al. 2014).

128

Figure 4.4. SeC 3´-tiRNAs accumulate in the spinal cords of mice.

129 Figure 4.4. SeC 3´-tiRNAs accumulate in the spinal cords of mice. A. 5OH-seq reads from MEF RNA (first four columns) and spinal cord RNA (last four columns) were uniquely aligned to a tRNA genome and normalized to counts per million aligned reads (CPMs). 5OH-seq CPMs for each tRNA species was summed and is displayed on the y axis. Samples with wildtype CLP1 (CLP1 +/+) are colored in black; samples with kinase-dead CLP1 (CLP1 K/K) are displayed in red. The predominant peak is a 5OH signal consistent with cleavage in the anticodon loop of SeC tRNA generating a 3´-tiRNA. This signal is present in spinal cord samples for both time points and is essentially absent in both sets of MEF samples. B. Zoomed view of the boxed SeC tRNA plot from A.

130 Though CLP1 is a ubiquitously expressed gene, most tissues are unaffected by this single site mutation. It is therefore perhaps unsurprising that I did not identify significant changes in 5OH-Seq signal when comparing RNA isolated from patient-derived fibroblasts of

CLP1+ / R140H parents to CLP1R140H / R140H patients. I conducted my experiments in a rich media and did not treat the cells with any cellular stressors. In one of the publications that first describes the CLP1R140H mutation, the patient-derived fibroblasts are cultured in N3 media supplemented with BDNF, GDNF, NT3 and CNTF to induce neuronal properties, termed “iNeurons” (Xue et al. 2013; Schaffer et al. 2014). The iNeurons with a CLP1R140H homozygous genotype show a stronger translocation of CLP1 to the cytoplasm and an even more pronounced tRNA accumulation phenotype when compared to the uninduced fibroblasts. It is therefore possible that a differential 5OH-seq signal would be generated upon iNeuron induction of these cells or upon cellular stressors that endogenous neuronal populations are selectively exposed to, such as oxidation.

When CLP1 kinase activity is completely abolished by the K127A mutation in the

Walker A motif, there is an increased signal intensity on the gel of 5OH-seq libraries prior to sequencing; however, the normalized 5OH-seq signal does not corroborate this effect or correlate with the anticipated biology of KMD (Figure 4.4). One of the current limitations of the 5OH-seq method is the absence of an internal standard. I normalize the 5OH-seq signal by counts of UMI-corrected reads per million reads aligned (CPMs), a normalization method that is analogous to standards of other sequencing methods. However, a widespread signal could be erased by this normalization procedure. I anticipate that including a sequencing standard prior to library preparation (further described in Chapter V) could improve the identification of 5´-OH RNAs.

131 A SeC tRNA SeC 3´ tiRNA 3´ 3´ 5´

HO-5´

angiogenin? B C

3´ 3´probe

5´ 5´probe

CLP1 K/K - 4 months 4 - K/K CLP1 days 20 - K/K CLP1

CLP1 +/+ - 4 months 4 - +/+ CLP1 days 20 - +/+ CLP1

full tRNA 3’ probe

3´ tiRNA

full tRNA 5’ probe

Figure 4.5. SeC tRNA Northern blot requires optimization. A. A cartoon of cleavage generating a 5´-OH tRNA fragment which is competent for RtcB ligation and 5OH-seq. Angiogenin is known to cleave tRNAs in the anticodon loop, and could potentially be involved in SeC tRNA cleavage. B. Location of the 3´ (green) and 5´ (red) northern probes in the SeC tRNA molecule. C. Northern blot with the probes shown in (C). Both the 3´ and 5´ probes anneal to the full- length SeC tRNA. No smaller cleavages fragments are identified.

132 Analysis of 5OH-seq reads aligning uniquely to tRNA sequences revealed a cleavage in the anticodon loop of selenocysteine (SeC) tRNA (Figure 4.4). This signal was found reproducibly in RNA from mouse spinal cord samples and was independent of CLP1 mutation. I was unable to validate this fragment by northern blot with probes designed against the 3´ or 5´ half of SeC tRNA (Figure 4.5). However, given the strength of the signal in the 5OH-seq data, I anticipate that the SeC 3´-tiRNA fragment may be validated by: optimizing hybridization conditions of the probes, greatly extending phosphor screen exposure time, changing the probe design, or using a more sensitive method for RNA detection, such as double DIG-labelled LNA probes.

Although anticodon-loop cleavage has been demonstrated in Met, Gly, Pro, Val,

Arg, and Tyr tRNAs (Fu et al. 2009; Yamasaki et al. 2009), this study is the first to identify anticodon-loop cleavage in SeC tRNA. The SeC tRNA can only pair with the UGA codon when a selenocysteine insertion sequence (SECIS) is present in mRNA (Berry et al. 1993).

When cells are grown in the presence of selenium, the SeC tRNA overrides the normal

UGA stop codon and incorporates selenocysteine into the nascent peptide. Twenty-five genes encode selenocysteine-containing proteins (selenoproteins) in humans, with most selenoproteins performing catalytic redox reactions (Labunskyy et al. 2014). Cleavage of

SeC tRNA may enable a cell to rapidly and specifically cease production of selenoproteins.

It is possible that this regulatory mechanism controlling selenoprotein production is ongoing in differentiated cells, and disrupting CLP1 kinase activity either directly prevents the degradation of these 3´ tiRNAs by inhibiting 5´-phosphate-dependent decay, or increases their production via an unknown mechanism.

133 Summary

In the chapter, I investigated the role of CLP1 on the accumulation of 5´-OH RNAs by performing 5OH-seq on several mammalian samples with mutations in CLP1. In 293FT cells with stable expression of shRNAs directed against CLP1 mRNA, I showed a CLP1- depedent divergence in 5OH-seq signal surrounding PolyA sites, confirming the importance of CLP1 in the pre-mRNA cleavage complex II for appropriate 3´-end formation. In RNA from spinal cords of mice, I identified a novel 5´-OH RNA generated by cleavage in the anticodon loop of SeC tRNA, which may link neural development to precise temporal regulation of selenoprotein expression. Further analysis of these 5OH-seq libraries may uncover additional 5´-OH RNAs that are differentially present in CLP1 kinase-impaired or CLP1-deficient cells.

134 CHAPTER V

SUMMARY AND FUTURE DIRECTIONS

This dissertation describes my two main contributions to the field of RNA decay: a method to directly capture and sequence 5´-OH RNAs (5OH-seq) and a unified theory of

5´-OH RNA degradation by kinase-mediated decay (KMD). I developed the 5OH-seq method to enable transcriptome-wide analysis of the uncharacterized 5´-OH RNA end modification. In Chapter II, I applied the 5OH-seq method to Saccharomyces cerevisiae, a quintessential eukaryotic model organism. Surprisingly, 5´-OH RNAs are found throughout the yeast transcriptome with areas of distinct and dispersed cleavages. I identified several hotspots of 5´-OH RNA accumulation, including upstream of mRNA sequences encoding polycharged amino acids and within a specific 5´-CCAUU|A-3´ motif following tuncamycin treatment. In Chapter IV, I extended the 5OH-seq method to mammalian samples, and showed that 5OH-seq can identify novel 5´-OH RNA fragments in mammals, such as the selenocysteine (SeC) 3´-tiRNA in mouse spinal cord RNA.

The widespread accumulation of 5´-OH RNAs led me to ask how this subset of

RNAs might be degraded. Given the perplexing accumulation of 3´-exon HAC1 mRNA fragments with 5´-PO4 termini (Harigaya and Parker 2012; Peach et al. 2015), I hypothesized that 5´-OH RNAs could be phosphorylated by RNA kinases to enter the canonical Xrn1 5´→3´degradation pathway. In Chapter III, I showed that this process, which I termed “kinase-mediated decay” (KMD), is required for the degradation of 3´-exon

HAC1 mRNA fragments and a NGD stem-loop reporter. I determined that KMD provides regulatory control of UPR induction in yeast by determining whether the HAC1 3´-exon

135 mRNA fragment is ligated and translated (UPR activation) or decayed by Xrn1 (UPR suppression).

In this final chapter, I reflect on the applications of 5OH-seq and the implications of KMD. I propose the use of 5OH-seq for identification of endonuclease targets and rapid clinical diagnostic tools. I also suggest strategic improvements to the 5OH-seq method. In the second half of the chapter, I outline future directions for the study of KMD, including a discussion of the potential for KMD in regulating the mammalian UPR, and the possibility of KMD inhibition by an RNA phosphatase.

Applications and Improvements of 5OH-Seq

Targets of metal ion-independent endoribonucleases

The most pressing application of 5OH-seq is the identification of novel targets of metal ion-independent endoribonucleases. In Chapter I, I discussed the mechanism of metal ion-independent endoribonucleases, which cleave ssRNA to yield a 5´-fragment with a 2´,3´-cyclic phosphate terminus and a 3´-fragment with a 5´-OH terminus. Given the successful capture of tRNA halves by 5OH-seq in both yeast (Chapter II) and mammals

(Chapter IV), I am particularly interested in finding novel targets of the endoribonucleases known to generate these fragments.

Under certain stress conditions, Rny1 in yeast and Angiogenin in humans cleave subclasses of tRNAs in the anticodon loop to generate 5´-tRNA halves with 2´,3´-cyclic phosphate termini and 3´-tRNA halves with 5´-OH termini (Yamasaki et al. 2009; Fu et al.

2009; Thompson and Parker 2009b). These stress-induced tRNA fragments (tiRNAs) can

136 inhibit translation and apoptosis in mammalian cells (Emara et al. 2010; Ivanov et al. 2011;

Saikia et al. 2014). In Chapter II, 5OH-seq enabled identification of an anticodon loop

Ser cleavage in tRNA AGA in non stressed conditions in yeast, which likely represents a low level constitutive cleavage by Rny1, as has been previously suggested (Thompson and

Parker 2009b). What is the nature of the tRNAs cleavage by Rny1? Does Rny1 cleave in this region merely based on a non-specific cleavage in an accessible, hypomodified region of an abundant RNA species, or are there other structural components driving recognition?

Other than the anticodon loops of tRNA, what are the other targets of Rny1, and are these cleavages functionally significant in mediating the effects of Rny1 by altering specific gene expression?

I discovered an anticodon loop cleavage in SeC tRNA in RNA isolated from mouse spinal cords (Chapter IV). The tRNASeC cleavage fragment is a novel 3´ tiRNA that has not been previously described in tiRNA literature, which prompts several questions. Is the

SeC 3´-tiRNA a neural tissue-specific tiRNA, and if so, is it a functional tiRNA or merely a marker of normal development? Is the SeC tiRNA produced by angiogenin? Furthermore, angiogenin plays a critical role in angiogenesis (Fett et al. 1985) and is particularly important in promoting the growth of solid tumors (Shimoyama et al. 1999; Katona et al.

2005; Montero et al. 1998). Does angiogenin target mRNAs to specifically decrease expression of tumor suppressors? Is the ribonuclease activity important for the function of secreted angiogenin?

In addition to answering the above questions, 5OH-seq can be used to find novel substrates of the other metal ion-independent endoribonucleases. Metazoans have evolved dozens of endoribonucleases (Yang 2011). While most have known substrates, many

137 uncharacterized endoribonucleases remain. It would be of particular interest to study these

“orphan endos” – endoribonucleases without known substrates – to determine their targets, which may in turn lead to additional insight into their function.

RIDD substrates in S. cerevisiae and mammals

In plants, S. pombe, Drosophila, and mammals, activated IRE1α cleaves ER- proximal mRNAs in a process termed Regulated IRE1-Dependent Decay (RIDD) (Hollien and Weissman 2006; Hollien et al. 2009; Kimmig et al. 2012; Hayashi et al. 2012). Most studies identifying RIDD substrates propose that an ER-localized mRNA must carry a consensus cleavage motif for IRE1α recognition (Maurel et al. 2014). This motif, 5´-

CN|GNNG-3´ and most often 5´-CU|GCAG-3´, is similar to the HAC1/XBP1 cleavage sites. In Chapter II, I showed that the 5OH-seq signals within 23 mRNAs increased at least 5-fold upon tunicamycin treatment compared to the DMSO control. Seven mRNAs had sites of 5´-OH cleavage that occurred in a common sequence motif 5´-CAUU|A-3´; however, this sequence does not match any known endoribonuclease motifs, and is highly divergent from the common RIDD motifs. The increase in cleavage of specific mRNAs in a tunicamycin-dependent manner certainly argues for the existence of RIDD in S. cerevisiae, though the divergence from a consensus motif necessitates less specificity in yeast Ire1 targeting. A recent study argued that RIDD substrates may not necessarily require the 5´-CN|GNNG-3´ motif (Gaddam et al. 2013). If the increased 5´-OH RNAs I identified following UPR induction with tunicamycin are indeed fragments of mRNAs cleaved by yeast Ire1, this not only demonstrates RIDD in S. cerevisiae, but also suggests

138 that 5OH-seq can be used to identify additional targets of IRE1α in human cells and other model organisms.

Additional analyses of 5OH-seq data in yeast

In Chapter II, I describe a series of computational analyses of 5OH-seq data from key yeast libraries including: comparisons to RNA-seq (Levin et al. 2010) and S1/V1 structure mapping (Kertesz et al. 2010) datasets; identification of discrete and clustered cleavages; enrichment analysis upstream and downstream of polycharged residues; and motif discovery. While this is a comprehensive series of analyses, it is certainly not exhaustive. Ribosome profiling data has a known triplet periodicity, and a recent publication demonstrated the same frame-dependent signals in 5´-phosphate sequencing data (Pelechano et al. 2016). Preliminary analysis demonstrated a frame-dependent signal for wild-type 5OH-seq libraries, but a more detailed examination of this periodicity may reveal a link with co-translational decay processes, such as NGD.

An in-depth assessment of the effect of specific codons on 5OH-seq signal should also be performed. Given that polybasic stretches are known to trigger NGD, I limited the analysis in Chapter II to runs of specific groups of amino acids (i.e., polyK, polyE, acidic, basic, etc.). However, lysine is encoded by both AAA and AAG codons, but only stretches of AAA trigger ribosomal sliding (Koutmou et al. 2015). Thus, specific codons rather than the biochemical properties of the amino acid encoded may promote cleavage by altering translation speed. Though I saw enhancement of signal upstream of polybasic stretches and polyacidic stretches with runs of at least 5 codons, I may find that individual codons have a greater contribution on 5´-OH RNA generation.

139 Rapid microarray-based screening of 5´-OH RNAs

While the plummeting costs of DNA sequencing have enabled a shift to whole exome and whole genome sequencing, there is still a need for validated tools for rapid clinical diagnostics. Dozens of patent applications for microarray-based screening tools have been filed in the United States, and such screening tools are used to detect monogenic disease-related mutations as well as bacterial, viral, and fungal infections (Yoo et al. 2009).

Future 5OH-seq experiments might reveal differential 5´-OH RNA signals between affected and unaffected populations for a given disease state. It may therefore be useful to develop a tool for quicker, less labor-intensive quantitation of select markers of RNA decay defects.

Assessing differential expression of subset of 5´-OH RNA targets could be accomplished by combining RtcB 5´-OH RNA ligation with microarray technology. A

DNA/RNA adapter with a 3´-phosphate and a recognition site for MmeI could be ligated to 5´-OH RNAs by RtcB. The ligated RNA could be reverse transcribed by using an adaptor with an oligo(dT) sequence followed by 5 cycles of PCR with primers complimentary to the adapters. The DNA could be cleaved with MmeI to yield fragments 20 or 21 nucleotides

3´ of the initial ligation site. This DNA could then be labelled with differential fluorescent dyes and applied to a chip with probes against candidate 5´-OH RNA sites. While this provides less information than the 5OH-seq method, this 5OH-Chip method would be compatible with high-throughput preparation methods and would allow rapid analysis of clinically relevant fragments on a much shorter time scale (~2 days vs. 2+ weeks).

140 Improvement of the 5OH-seq method

There are key improvements that should be made to the 5OH-seq method to overcome two principal limitations. One of the limitations of the current 5OH-seq protocol is purely technical in nature. The linker oligo contains eight degenerate bases to account for PCR amplification bias, and the reverse transcription primer contains nine degenerate bases for random priming. To avoid priming of the linker oligo with the degenerate reverse transcription primer, I purify the linker by running the ligated sample on a TBE-Urea gel, manually excising and crushing a gel band, and eluting the RNA. This step prevents the method from being compatible with high throughput robotics and also adds several additional hours and reagents to the protocol. To decrease reverse priming, the linker could be added at a lower concentration or the UMI of the linker could be shortened.

Alternatively, RNA could be reverse transcribed with an oligo-dT primer, the resulting cDNA could be fragmented, and an Illumina-compatible primer could be directly ligated to the 3´ end. This would prevent background amplification, and would limit the libraries to poly-adenylated species, eliminating the need for polyA enrichment or rRNA depletion.

The second limitation of the current 5OH-seq protocol is the lack of robust metrics for quantitative comparisons. The 5OH-seq libraries are normalized by UMI-corrected reads per million aligned sequences, referred to as counts per million (CPMs). While this aligns related samples on a similar scale, it may prevent analyses across key datasets. If samples maintain similar relative levels of 5´-OH RNA cleavages, but there is a gross change in processing by genome-wide modification or deletion of decay factors, this change may not be detected. It would therefore be useful to normalize libraries to a “spike- in” control of several 5´-OH RNAs with known concentrations. The External RNA

141 Controls Consortium (ERCC) recently developed a set of universal synthetic RNA spike- in standards for RNA-seq (Jiang et al. 2011; Zook et al. 2012), and these standards are commercially available. Use of such standards enables absolute quantification of 5´-OH

RNAs, a fundamental question that remains unanswered. These standards also enable quantification of experimental variability in the preparation and sequencing of the libraries, providing a much needed quality assurance metric in the 5OH-seq protocol.

Future Studies and Considerations of Kinase-Mediated Decay

Within this dissertation, I coined the term “kinase-mediated decay” (KMD), which encompasses any pathway where phosphorylation of 5´-OH RNA precedes 5´-phosphate- dependent degradation. Prior to this dissertation, there were two examples of KMD in eukaryotes: the decay of excised tRNA introns and the processing of the 5´ end of 28S rRNA. In Chapter III, I describe two new substrates that require phosphorylation of 5´-

OH RNA prior to 5´-phosphate dependent degradation: the 3´-exon of HAC1 after cleavage by Ire1 and the 3´ cleavage fragment during NGD. These are the first examples of KMD that is required for the degradation of mRNA fragments. Though these examples solidify

KMD as a bona fide pathway for cytoplasmic decay of mRNA fragments, many additional questions remain. Below I discuss six unexplored aspects of KMD, including implications for mammalian XBP1 mRNA splicing.

Decay kinetics

My initial studies of KMD evaluated accumulation of 5´-OH RNAs at single time points in yeast with stable deletions of decay factors. While this allowed me to demonstrate

142 that KMD is required for 5´-OH RNA degradation, it does not address the kinetics of this reaction. Several questions regarding KMD kinetics should be addressed in future studies including: What is the general turnover rate of 5´-OH RNAs? How quickly are 5´-OH

RNAs phosphorylated after cleavage, and how soon after phosphorylation are these 5´-PO4

RNAs degraded? This is particularly interesting for determining the window of opportunity for cellular action, such as the repression of translation by tiRNAs.

Localization of KMD

Though the two examples of KMD described in Chapter III are performed by enzymes with cytosolic localization (Trl1 and Xrn1), it is not clear precisely where KMD of mRNAs occurs. Processing bodies (P-bodies) are foci containing translationally- repressed messenger ribonucleoproteins (mRNPs) (Parker and Sheth 2007). The complete proteome of P-bodies is not known, but many components of 5´→3´ mRNA decay are present, including Xrn1 and decapping proteins Dcp1/Dcp2. In xrn1∆ yeast, the NGD reporter PGK-sL aggregates in distinct cystoplasmic foci and colocalizes with Dcp2, suggesting that NGD fragments accumulate within P-bodies (Cole et al. 2009). Does KMD of NGD fragments therefore occur within P-bodies? Is Trl1 present in P-bodies or does phosphorylation of the mRNA fragment precede entrance to the P-body? Are other 5´-OH

RNAs fated for degradation in P-bodies? mRNAs stored within P-bodies are not constitutively degraded; some mRNAs exit P-bodies and resume translation (Brengues et al. 2005). This raises a particularly intriguing question: can 5´-OH RNAs be temporarily stored and later religated either in cis or in trans to RNAs with 2´,3´-cyclic phosphate termini?

143 KMD between compartments

In the four examples of eukaryotic KMD so far, there are only two kinase- exoribonuclease pairs – the nuclear Grc3-Rat1 and cytosolic Trl1-Xrn1 – which raises several questions for future studies. These enzymes show little substrate specificity and are agnostic to RNA sequence. Are these kinases and exoribonucleases therefore interchangeable for KMD? For instance, expressing a mutant Rat1 with cytosolic localization restores the growth defects of xrn1∆ strains (Johnson 1997). Would this cytoplasmic Rat1 also rescue KMD of the known substrates in xrn1∆ strains?

Phosphorylation and degradation occur within the same compartment (i.e., both steps are either nuclear or cytosolic) in the known examples. Is it possible for a 5´-OH RNA to be phosphorylated in the cytoplasm, and then transported to the nucleus for decay, or vice versa? Can the human 5´-OH RNA kinases (CLP1 and NOL9) restore KMD in yeast with kinase-dead Trl1 and Grc3?

Decay of trailing mRNA in 3´-end formation

CLP1, one of the human 5´-OH RNA kinases, was first identified as part of the II (CFIIm) complex, and is thought to bridge cleavage and polydenylation specificity factor (CPSF) and CFIm (de Vries et al. 2000). Accordingly, depletion of CLP1 in HeLa cells decreases pre-mRNA 3´ cleavage activity (de Vries et al. 2000). Curiously, although the yeast homolog is also required for 3´ mRNA end formation (Haddad et al.

2012), the yeast Clp1 cannot phosphorylate 5´-OH RNA. Whether the additional kinase activity of human CLP1 affects 3´ cleavage, and in particular, the decay of the fragment downstream of cleavage is not known. In the torpedo model of RNA Polymerase II

144 transcriptional termination, 5´-phosphate dependent 5´→3´ exoribonucleases Rat1/XRN2 rapidly degrade RNA downstream of the poly(A) cleavage site and ultimately dissociate

RNA Pol II from its DNA template (Luo and Bentley 2004). Could CLP1 potentially phosphorylate the 5´ terminus of the 3´ fragment to facilitate rapid degradation by XRN2?

Since this activity is clearly not required in yeast, could another factor in humans alter the

5´-terminus following cleavage?

Regulation of UPR by KMD in mammals

In mammals, the IRE1α arm of the UPR is regulated by the non-canonical cytosolic splicing of XBP1 mRNA, the HAC1 homolog. In Chapter III, I showed that phosphorylation of the 5´ terminus of the HAC1 mRNA 3´ exon by Trl1 poises the exon either for ligation and UPR activation or for degradation by Xrn1 resulting in UPR suppression. While it is possible that XBP1 splicing could also be regulated by KMD, this seems unlikely for several reasons. First, in contrast to the single pathway of UPR control by HAC1 mRNA splicing in yeast, humans have two additional branches controlling UPR induction, PERK and ATF6. Therefore, there has been less selective pressure on an intricate pathway to regulate XBP1 splicing by degradation of intermediates. Second, the intron of XBP1 is only 26 nt in humans, compared to the 252 nt intron of yeast HAC1, which greatly minimizes the distance between the cleaved exons. Indeed, a recent publication from the Walter lab proposed that the intron was expelled by a conformational

RNA zipper that quickly aligns the cleavage sites for splicing by RTCB (Peschek et al.

2015). Third, RTCB directly ligates the two exons rather than the distributed multi-step ligation yeast pathway of Trl1. While I have not yet performed a direct comparison of the

145 ligation efficiency of yeast Trl1 and human RTCB, I and others have shown that E. coli

RtcB quickly and efficiently ligates ends, especially ends in close proximity such as complementary RNAs (Tanaka and Shuman 2011). Finally, though humans have conserved components of KMD, including at least two 5´-OH RNA kinases (CLP1, NOL9) and several 5´-phosphate-dependent exoribonucleases (XRN1, XRN2, DXO, and CPSF3), it is unclear whether cytosolic KMD is truly feasible. XRN1 is primarily cytosolic; however, NOL9 is localized in the nucleolus (Heindl and Martinez 2010) and CLP1 is primarily (~85%) located in the nucleus (Schaffer et al. 2014). There may not be a sufficient concentration of 5´-OH RNA kinase in the cytoplasm to rapidly phosphorylate this specific substrate prior to ligation by RTCB.

Although I believe it is unlikely that XBP1 splicing and the human UPR is regulated by KMD, it remains possible that an unknown cytoplasmic 5´-OH RNA kinase could phosphorylate the XBP1 3´-exon for degradation. A prime target for future experiments testing either UPR-specific KMD or general cytosolic KMD in humans is NEDD4-binding protein 2 (N4BP2), which is located in the cytoplasm and has 5´-polynucleotide kinase activity (although it is not known whether this includes 5´-OH RNA substrates).

Importantly, although XBP1 splicing is unlikely to be regulated by KMD, the possibility for general cytoplasmic degradation of mRNAs by KMD cannot be ruled out.

RNA phosphatase

In signaling pathways, kinases and act as molecular switches to modulate function of protein intermediates. A kinase transfers a phosphate to a serine, tyrosine, or threonine residue to alter confirmation of an affected protein. Phosphorylation

146 can activate or deactivate enzymatic activity, change protein localization, or alter binding sites for cofactors and protein partners. Phosphatases remove phosphates from the same residues and antagonize these changes. Perhaps RNA end modifications can be similarly regulated for precise, reversible control. If an RNA kinase can alter an RNA to enter a regulated decay pathway, is there a RNA phosphatase counteracting this fate and preserving the RNA?

mRNA capping enzymes were the first eukaryotic proteins described with in vivo

RNA triphosphatase activity. The N-terminal domains of these mRNA capping enzymes have sequence similarity to the protein tyrosine phosphatase (PTP) family, and the cytosine of the conserved HCX5R, motif in CEL-1 is required for RNA triphosphatase activity in C. elegans (Takagi et al. 1997). The baculovirus (BVP) from the

Autographa californica nuclear polyhedrosis virus also contains a PTP domain and can further hydrolyze the 5´ terminus of RNAs from diphosphate to monophosphate in vivo

(Takagi et al. 1998). Sequence similarity studies revealed that human PIR1, which contains

41% sequence similarity to BVP, also has RNA 5´-triphosphatase and 5´-bisphosphatase activities (Deshpande et al. 1999). Together, these enzymes form a subgroup of PTP family members that recognize RNA as a substrate. None of these known RNA phosphatases can hydrolyze 5´-monophosphate RNAs to 5´-OH RNAs; however there are phosphatases that act on nucleic acids indiscriminately. Calf intestinal phosphatase is a standard molecular biology reagent that dephosphorylates 5´-PO4 RNA and DNA in vitro. In humans, there are four such alkaline phosphatases: placental (PLAP), placental-like (ALPPL2), intestinal

(ALPI), and the tissue non-specific live/bone/kidney alkaline phosphatase (ALPL).

Importantly, measuring alkaline phosphatase (ALP) levels is part of the standard panel of

147 “liver function tests”; elevated ALP is a marker of liver damage and certain cancers, especially osteosarcoma (Lange et al. 1982; Khan et al. 2015). If a phosphatase acts in direct opposition to the RNA 5´-kinases, it is either: i) not yet known, or ii) a broad-acting

DNA/RNA phosphatase, such as the alkaline phosphatases. Future studies should investigate the potential inhibition of RNA phosphatases on KMD.

148 REFERENCES

Abbott JA, Francklyn CS, Robey-Bond SM. 2014. Transfer RNA and human disease. Front Genet 5: 158.

Adham IM, Sallam MA, Steding G, Korabiowska M, Brinck U, Hoyer-Fender S, Oh C, Engel W. 2003. Disruption of the pelota gene causes early embryonic lethality and defects in cell cycle progression. Mol Cell Biol 23: 1470–1476.

Alberti S, Gitler AD, Lindquist S. 2007. A suite of Gateway cloning vectors for high- throughput genetic analysis in Saccharomyces cerevisiae. Yeast 24: 913–919.

Allmang C, Petfalski E, Podtelejnikov A, Mann M, Tollervey D, Mitchell P. 1999. The yeast exosome and human PM-Scl are related complexes of 3“ --> 5” exonucleases. Genes & Development 13: 2148–2158.

Amberg DC, Goldstein AL, Cole CN. 1992. Isolation and characterization of RAT1: an essential gene of Saccharomyces cerevisiae required for the efficient nucleocytoplasmic trafficking of mRNA. Genes & Development 6: 1173–1189.

Anders S, Huber W. 2010. Differential expression analysis for sequence count data. Genome Biol 11: R106.

Anderson JSJ, Parker R. 1998. The 3′ to 5′ degradation of yeast mRNAs is a general mechanism for mRNA turnover that requires the SKI2 DEVH box protein and 3′ to 5′ exonucleases of the exosome complex. EMBO J 17: 1497–1506.

Andrade JM, Pobre V, Silva IJ, Domingues S, Arraiano CM. 2009. The role of 3“-5” exoribonucleases in RNA degradation. Prog Mol Biol Transl Sci 85: 187–229.

Apostol BL, Westaway SK, Abelson J, Greer CL. 1991. Deletion analysis of a multifunctional yeast tRNA ligase polypeptide. Identification of essential and dispensable functional domains. J Biol Chem 266: 7445–7455.

Aragón T, van Anken E, Pincus D, Serafimova IM, Korennykh AV, Rubio CA, Walter P. 2008. nature07641. Nature 457: 736–740.

Baker KE, Parker R. 2004. Nonsense-mediated mRNA decay: terminating erroneous gene expression. Current Opinion in Cell Biology 16: 293–299.

Batut P, Dobin A, Plessy C, Carninci P, Gingeras TR. 2013. High-fidelity promoter profiling reveals widespread alternative promoter usage and transposon-driven developmental gene expression. Genome Res 23: 169–180.

Bechill J, Chen Z, Brewer JW, Baker SC. 2008. Coronavirus infection modulates the unfolded protein response and mediates sustained translational repression. J Virol 82: 4492–4501.

149 Bechill J, Chen Z, Brewer JW, Baker SC. 2006. Mouse hepatitis virus infection activates the Ire1/XBP1 pathway of the unfolded protein response. Adv Exp Med Biol 581: 139–144.

Becker T, Armache J-P, Jarasch A, Anger AM, Villa E, Sieber H, Motaal BA, Mielke T, Berninghausen O, Beckmann R. 2011. Structure of the no-go mRNA decay complex Dom34-Hbs1 bound to a stalled 80S ribosome. Nat Struct Mol Biol 18: 715–720.

Berry MJ, Banu L, Harney JW, Larsen PR. 1993. Functional characterization of the eukaryotic SECIS elements which direct selenocysteine insertion at UGA codons. EMBO J 12: 3315–3322.

Bhattacharya A, McIntosh KB, Willis IM, Warner JR. 2010. Why Dom34 stimulates growth of cells with defects of 40S ribosomal subunit biosynthesis. Mol Cell Biol 30: 5562–5571.

Boeck R, Tarun S, Rieger M, Deardorff JA, Müller-Auer S, Sachs AB. 1996. The yeast Pan2 protein is required for poly(A)-binding protein-stimulated poly(A)-nuclease activity. J Biol Chem 271: 432–438.

Brandman O, Stewart-Ornstein J, Wong D, Larson A, Williams CC, Li G-W, Zhou S, King D, Shen PS, Weibezahn J, et al. 2012. A ribosome-bound quality control complex triggers degradation of nascent peptides and signals translation stress. Cell 151: 1042–1054.

Brengues M, Teixeira D, Parker R. 2005. Movement of eukaryotic mRNAs between polysomes and cytoplasmic processing bodies. Science 310: 486–489.

Breslow R, Huang DL. 1991. Effects of metal ions, including Mg2+ and lanthanides, on the cleavage of ribonucleotides and RNA model compounds. Proc Natl Acad Sci USA 88: 4080–4083.

Brown CE, Sachs AB. 1998. Poly(A) tail length control in Saccharomyces cerevisiae occurs by message-specific deadenylation. Mol Cell Biol 18: 6548–6559.

Brown RS, Dewan JC, Klug A. 1985. Crystallographic and biochemical investigation of the lead(II)-catalyzed hydrolysis of yeast phenylalanine tRNA. Biochemistry 24: 4785–4801.

Butterfield RJ, Stevenson TJ, Xing L, Newcomb TM, Nelson B, Zeng W, Li X, Lu H-M, Lu H, Farwell Gonzalez KD, et al. 2014. Congenital lethal motor neuron disease with a novel defect in ribosome biogenesis. Neurology 82: 1322–1330.

Buzayan JM, Gerlach WL, Bruening G. 1986. Non-enzymatic cleavage and ligation of RNAs complementary to a plant virus satellite RNA. Nature 323: 349–353.

Cao D, Parker R. 2001. Computational modeling of eukaryotic mRNA turnover. RNA 7: 1192–1212.

150 Caponigro G, Parker R. 1995. Multiple functions for the poly(A)-binding protein in mRNA decapping and deadenylation in yeast. Genes & Development 9: 2421–2432.

Cech TR, Zaug AJ, Grabowski PJ. 1981. In vitro splicing of the ribosomal RNA precursor of Tetrahymena: involvement of a guanosine nucleotide in the excision of the intervening sequence. Cell 27: 487–496.

Celesnik H, Deana A, Belasco JG. 2007. Initiation of RNA decay in by 5' pyrophosphate removal. Molecular Cell 27: 79–90.

Chakravarty AK, Shuman S. 2012. The sequential 2',3'-cyclic phosphodiesterase and 3“- phosphate/5-”OH ligation steps of the RtcB RNA splicing pathway are GTP- dependent. Nucleic Acids Res 40: 8558–8567.

Chakravarty AK, Subbotin R, Chait BT, Shuman S. 2012. RNA ligase RtcB splices 3'- phosphate and 5'-OH ends via covalent RtcB-(histidinyl)-GMP and polynucleotide- (3“)pp(5”)G intermediates. Proc Natl Acad Sci USA 109: 6072–6077.

Chang JH, Jiao X, Chiba K, Oh C, Martin CE, Kiledjian M, Tong L. 2012. Dxo1 is a new type of eukaryotic enzyme with both decapping and 5“-3” exoribonuclease activity. Nat Struct Mol Biol 19: 1011–1017.

Chapman RE, Walter P. 1997. Translational attenuation mediated by an mRNA intron. Curr Biol 7: 850–859.

Chen L, Muhlrad D, Hauryliuk V, Cheng Z, Lim MK, Shyp V, Parker R, Song H. 2010. Structure of the Dom34-Hbs1 complex and implications for no-go decay. Nat Struct Mol Biol 17: 1233–1240.

Cohn WE. 1960. Pseudouridine, a carbon-carbon linked ribonucleoside in ribonucleic acids: isolation, structure, and chemical characteristics. J Biol Chem 235: 1488–1498.

Cole SE, LaRiviere FJ, Merrikh CN, Moore MJ. 2009. A convergence of rRNA and mRNA quality control pathways revealed by mechanistic analysis of nonfunctional rRNA decay. Mol Cell 34: 440–450.

Cooper DA, Banerjee S, Chakrabarti A, García-Sastre A, Hesselberth JR, Silverman RH, Barton DJ. 2015. RNase L targets distinct sites in influenza A virus RNAs. ed. S. Perlman. J Virol 89: 2764–2776.

Cooper DA. 2015. The detection of ribonuclease cleavage sites in host and viral RNAs during viral infections. (PhD Thesis).

Cooper DA, Jha BK, Silverman RH, Hesselberth JR, Barton DJ. 2014. Ribonuclease L and metal-ion-independent endoribonuclease cleavage sites in host and viral RNAs. Nucleic Acids Res 42: 5202–5216.

151 Cox JS, Shamu CE, Walter P. 1993. Transcriptional induction of genes encoding endoplasmic reticulum resident proteins requires a transmembrane protein kinase. Cell 73: 1197–1206.

Cox JS, Walter P. 1996. A novel mechanism for regulating activity of a transcription factor that controls the unfolded protein response. Cell 87: 391–404.

Culver GM, McCraith SM, Consaul SA, Stanford DR, Phizicky EM. 1997. A 2'- phosphotransferase implicated in tRNA splicing is essential in Saccharomyces cerevisiae. J Biol Chem 272: 13203–13210.

Culver GM, McCraith SM, Zillmann M, Kierzek R, Michaud N, LaReau RD, Turner DH, Phizicky EM. 1993. An NAD derivative produced during transfer RNA splicing: ADP-ribose 1“-2” cyclic phosphate. Science 261: 206–208.

Das SR, Piccirilli JA. 2005. General acid catalysis by the hepatitis delta virus ribozyme. Nat Chem Biol 1: 45–52.

Das U, Chakravarty AK, Remus BS, Shuman S. 2013. Rewriting the rules for end joining via enzymatic splicing of DNA 3“-PO4 and 5-”OH ends. Proc Natl Acad Sci USA 110: 20437–20442.

Davis FF, Allen FW. 1957. Ribonucleic acids from yeast which contain a fifth nucleotide. J Biol Chem 227: 907–915. de Vries H, Rüegsegger U, Hübner W, Friedlein A, Langen H, Keller W. 2000. Human pre-mRNA cleavage factor II(m) contains homologs of yeast proteins and bridges two other cleavage factors. EMBO J 19: 5895–5904.

Derti A, Garrett-Engele P, Macisaac KD, Stevens RC, Sriram S, Chen R, Rohl CA, Johnson JM, Babak T. 2012. A quantitative atlas of polyadenylation in five mammals. Genome Res 22: 1173–1183.

Desai KK, Raines RT. 2012. tRNA ligase catalyzes the GTP-dependent ligation of RNA with 3“-phosphate and 5-”hydroxyl termini. Biochemistry 51: 1333–1335.

Deshpande T, Takagi T, Hao L, Buratowski S, Charbonneau H. 1999. Human PIR1 of the protein-tyrosine phosphatase superfamily has RNA 5'-triphosphatase and diphosphatase activities. J Biol Chem 274: 16590–16594.

Dimitrova L, Kuroha K, Tatematsu T, Inada T. 2009. Nascent peptide-dependent translation arrest leads to Not4p-mediated protein degradation by the proteasome. Journal of Biological Chemistry 284: 10343–10352.

Dimroth K, Jaenicke L, Heinzel D. 1950. Die Spaltung der Pentose-nucleinsäure der Hefe mit Bleihydroxyd. Justus Liebigs Annalen der Chemie 566: 206–210.

152 Doma MK, Parker R. 2006. Endonucleolytic cleavage of eukaryotic mRNAs with stalls in translation elongation. Nature 440: 561–564.

Durand S, Richard G, Bontems F, Uzan M. 2012. Bacteriophage T4 polynucleotide kinase triggers degradation of mRNAs. Proc Natl Acad Sci USA 109: 7073–7078.

Durdevic Z, Mobin MB, Hanna K, Lyko F, Schaefer M. 2013. The RNA methyltransferase Dnmt2 is required for efficient Dicer-2-dependent siRNA pathway activity in Drosophila. Cell Rep 4: 931–937.

Eberhart CG, Wasserman SA. 1995. The pelota locus encodes a protein required for meiotic cell division: an analysis of G2/M arrest in Drosophila spermatogenesis. Development 121: 3477–3486.

Eckard SC, Rice GI, Fabre A, Badens C, Gray EE, Hartley JL, Crow YJ, Stetson DB. 2014. The SKIV2L RNA exosome limits activation of the RIG-I-like receptors. Nat Immunol 15: 839–845.

Emara MM, Ivanov P, Hickman T, Dawra N, Tisdale S, Kedersha N, Hu G-F, Anderson P. 2010. Angiogenin-induced tRNA-derived stress-induced RNAs promote stress- induced stress granule assembly. Journal of Biological Chemistry 285: 10959–10968.

Englert M, Sheppard K, Aslanian A, Yates JR, Söll D. 2011. Archaeal 3'-phosphate RNA splicing ligase characterization identifies the missing component in tRNA maturation. Proc Natl Acad Sci USA 108: 1290–1295.

Fedor MJ, Williamson JR. 2005. The catalytic diversity of RNAs. Nat Rev Mol Cell Biol 6: 399–412.

Fernández-Pevida A, Kressler D, La Cruz De J. 2015. Processing of preribosomal RNA in Saccharomyces cerevisiae. Wiley Interdiscip Rev RNA 6: 191–209.

Ferré-D'Amaré AR, Scott WG. 2010. Small self-cleaving ribozymes. Cold Spring Harb Perspect Biol 2: a003574–a003574. http://cshperspectives.cshlp.org/content/2/10/a003574.full.

Fett JW, Strydom DJ, Lobb RR, Alderman EM, Bethune JL, Riordan JF, Vallee BL. 1985. Isolation and characterization of angiogenin, an angiogenic protein from human carcinoma cells. Biochemistry 24: 5480–5486.

Finley D, Bartel B, Varshavsky A. 1989. The tails of ubiquitin precursors are ribosomal proteins whose fusion to ubiquitin facilitates ribosome biogenesis. Nature 338: 394– 401.

Forconi M, Herschlag D. 2009. Metal ion-based RNA cleavage as a structural probe. Meth Enzymol 468: 91–106.

153 Frischmeyer PA. 2002. An mRNA Surveillance Mechanism That Eliminates Transcripts Lacking Termination Codons. Science 295: 2258–2261.

Fu H, Feng J, Liu Q, Sun F, Tie Y, Zhu J, Xing R, Sun Z, Zheng X. 2009. Stress induces tRNA cleavage by angiogenin in mammalian cells. FEBS Lett 583: 437–442.

Furuichi Y, LaFiandra A, Shatkin AJ. 1977. 5'-Terminal structure and mRNA stability. Nature 266: 235–239.

Gaddam D, Stevens N, Hollien J. 2013. Comparison of mRNA localization and regulation during endoplasmic reticulum stress in Drosophila cells. Mol Biol Cell 24: 14–20.

Galperin MY, Koonin EV. 2004. “Conserved hypothetical” proteins: prioritization of targets for experimental study. Nucleic Acids Res 32: 5452–5463.

Gardner BM, Walter P. 2011. Unfolded Proteins Are Ire1-Activating Ligands That Directly Induce the Unfolded Protein Response. Science 333: 1891–1894.

Gass JN, Gifford NM, Brewer JW. 2002. Activation of an unfolded protein response during differentiation of antibody-secreting B cells. J Biol Chem 277: 49047–49054.

Gasse L, Flemming D, Hurt E. 2015. Coordinated Ribosomal ITS2 RNA Processing by the Las1 Complex Integrating Endonuclease, Polynucleotide Kinase, and Exonuclease Activities. Mol Cell 60: 808–815.

Gatfield D, Unterholzner L, Ciccarelli FD, Bork P, Izaurralde E. 2003. Nonsense- mediated mRNA decay in Drosophila: at the intersection of the yeast and mammalian pathways. EMBO J 22: 3960–3970.

Geerlings TH, Vos JC, Raué HA. 2000. The final step in the formation of 25S rRNA in Saccharomyces cerevisiae is performed by 5“-->3” exonucleases. RNA 6: 1698– 1703.

Genschik P, Drabikowski K, Filipowicz W. 1998. Characterization of the Escherichia coli RNA 3'-terminal phosphate cyclase and its sigma54-regulated operon. J Biol Chem 273: 25516–25526.

German MA, Pillay M, Jeong D-H, Hetawal A, Luo S, Janardhanan P, Kannan V, Rymarquis LA, Nobuta K, German R, et al. 2008. Global identification of microRNA–target RNA pairs by parallel analysis of RNA ends. Nature Biotechnology 26: 941–946.

Ghazal G, Gagnon J, Jacques P-E, Landry J-R, Robert F, Elela SA. 2009. Yeast RNase III triggers polyadenylation-independent transcription termination. Mol Cell 36: 99– 109.

154 Giorgi C, Fatica A, Nagel R, Bozzoni I. 2001. Release of U18 snoRNA from its host intron requires interaction of Nop1p with the Rnt1p endonuclease. EMBO J 20: 6856–6865.

Gonzalez TN, Sidrauski C, Dörfler S, Walter P. 1999. Mechanism of non-spliceosomal mRNA splicing in the unfolded protein response pathway. EMBO J 18: 3119–3132.

Goodman HM, Olson MV, Hall BD. 1977. Nucleotide sequence of a mutant eukaryotic gene: the yeast tyrosine-inserting ochre suppressor SUP4-o. Proc Natl Acad Sci USA 74: 5453–5457.

Greenway MJ, Alexander MD, Ennis S, Traynor BJ, Corr B, Frost E, Green A, Hardiman O. 2004. A novel candidate region for ALS on chromosome 14q11.2. Neurology 63: 1936–1938.

Greer CL, Peebles CL, Gegenheimer P, Abelson J. 1983. Mechanism of action of a yeast RNA ligase in tRNA splicing. Cell 32: 537–546.

Guerrier-Takada C, Gardiner K, Marsh T, Pace N, Altman S. 1983. The RNA moiety of ribonuclease P is the catalytic subunit of the enzyme. Cell 35: 849–857.

Gulland JM, Jackson EM. 1938. 5-Nucleotidase. Biochem J 32: 597–601.

Guydosh NR, Green R. 2014. Dom34 Rescues Ribosomes in 3′ Untranslated Regions. Cell 156: 950–962.

Haddad R, Maurice F, Viphakone N, Voisinet-Hakil F, Fribourg S, Minvielle-Sébastia L. 2012. An essential role for Clp1 in assembly of polyadenylation complex CF IA and Pol II transcription termination. Nucleic Acids Res 40: 1226–1239.

Hanada T, Weitzer S, Mair B, Bernreuther C, Wainger BJ, Ichida J, Hanada R, Orthofer M, Cronin SJ, Komnenovic V, et al. 2013. nature11923. Nature 495: 474–480.

Harding HP, Lackey JG, Hsu HC, Zhang Y, Deng J, Xu RM, Damha MJ, Ron D. 2007. An intact unfolded protein response in Trpt1 knockout mice reveals phylogenic divergence in pathways for RNA ligation. RNA 14: 225–232.

Harding HP, Zeng H, Zhang Y, Jungries R, Chung P, Plesken H, Sabatini DD, Ron D. 2001. Diabetes mellitus and exocrine pancreatic dysfunction in perk-/- mice reveals a role for translational control in secretory cell survival. Molecular Cell 7: 1153–1163.

Harigaya Y, Parker R. 2012. Global analysis of mRNA decay intermediates in Saccharomyces cerevisiae. Proc Natl Acad Sci USA 109: 11764–11769.

Hayashi S, Wakasa Y, Takaiwa F. 2012. Functional integration between defence and IRE1-mediated ER stress response in rice. Sci Rep 2: 670.

155 Heindl K, Martinez J. 2010. Nol9 is a novel polynucleotide 5'-kinase involved in ribosomal RNA processing. EMBO J 29: 4161–4171.

Helm M, Alfonzo JD. 2014. Posttranscriptional RNA Modifications: playing metabolic games in a cell's chemical Legoland. Chem Biol 21: 174–185.

Henry Y, Wood H, Morrissey JP, Petfalski E, Kearsey S, Tollervey D. 1994. The 5' end of yeast 5.8S rRNA is generated by exonucleases from an upstream cleavage site. EMBO J 13: 2452–2463.

Ho CK, Rauhut R, Vijayraghavan U, Abelson J. 1990. Accumulation of pre-tRNA splicing“2/3”intermediates in a Saccharomyces cerevisiae mutant. EMBO J 9: 1245.

Hollien J, Lin JH, Li H, Stevens N, Walter P, Weissman JS. 2009. Regulated Ire1- dependent decay of messenger RNAs in mammalian cells. J Cell Biol 186: 323–331.

Hollien J, Weissman JS. 2006. Decay of endoplasmic reticulum-localized mRNAs during the unfolded protein response. Science 313: 104–107.

Honda S, Loher P, Shigematsu M, Palazzo JP, Suzuki R, Imoto I, Rigoutsos I, Kirino Y. 2015. Sex hormone-dependent tRNA halves enhance cell proliferation in breast and prostate cancers. Proc Natl Acad Sci USA 112: E3816–25.

Hopper AK. 2013. Transfer RNA post-transcriptional processing, turnover, and subcellular dynamics in the yeast Saccharomyces cerevisiae. Genetics 194: 43–67.

Hsu CL, Stevens A. 1993. Yeast cells lacking 5'-->3‘ exoribonuclease 1 contain mRNA species that are poly(A) deficient and partially lack the 5’ cap structure. Mol Cell Biol 13: 4826–4835.

Huntzinger E, Kashima I, Fauser M, Saulière J, Izaurralde E. 2008. SMG6 is the catalytic endonuclease that cleaves mRNAs containing nonsense codons in metazoan. RNA 14: 2609–2617.

Ishimura R, Nagy G, Dotu I, Zhou H, Yang X-L, Schimmel P, Senju S, Nishimura Y, Chuang JH, Ackerman SL. 2014. RNA function. Ribosome stalling induced by mutation of a CNS-specific tRNA causes neurodegeneration. Science 345: 455–459.

Ito-Harashima S, Kuroha K, Tatematsu T, Inada T. 2007. Translation of the poly(A) tail plays crucial roles in nonstop mRNA surveillance via translation repression and protein destabilization by proteasome in yeast. Genes & Development 21: 519–524.

Ivanov P, Emara MM, Villen J, Gygi SP, Anderson P. 2011. Angiogenin-Induced tRNA Fragments Inhibit Translation Initiation. Mol Cell 43: 613–623.

Iwakoshi NN, Lee A-H, Vallabhajosyula P, Otipoby KL, Rajewsky K, Glimcher LH. 2003. Plasma cell differentiation and the unfolded protein response intersect at the transcription factor XBP-1. Nat Immunol 4: 321–329.

156 Iwawaki T, Tokuda M. 2011. Function of yeast and amphioxus tRNA ligase in IRE1alpha-dependent XBP1 mRNA splicing. Biochem Biophys Res Commun 413: 527–531.

Jackman JE, Alfonzo JD. 2013. Transfer RNA modifications: nature's combinatorial chemistry playground. Wiley Interdiscip Rev RNA 4: 35–48.

Jiang L, Schlesinger F, Davis CA, Zhang Y, Li R, Salit M, Gingeras TR, Oliver B. 2011. Synthetic spike-in standards for RNA-seq experiments. Genome Res 21: 1543–1551.

Jiao X, Chang JH, Kilic T, Tong L, Kiledjian M. 2013. A mammalian pre-mRNA 5' end capping quality control mechanism and an unexpected link of capping to pre-mRNA processing. Mol Cell 50: 104–115.

Johnson AW. 1997. Rat1p and Xrn1p are functionally interchangeable exoribonucleases that are restricted to and required in the nucleus and cytoplasm, respectively. Mol Cell Biol 17: 6122–6130.

Jones CI, Zabolotskaya MV, Newbury SF. 2012. The 5“ → 3” exoribonuclease XRN1/Pacman and its functions in cellular processes and development. Wiley Interdiscip Rev RNA 3: 455–468.

Jurkin J, Henkel T, Nielsen AF, Minnich M, Popow J, Kaufmann T, Heindl K, Hoffmann T, Busslinger M, Martinez J. 2014. The mammalian tRNA ligase complex mediates splicing of XBP1 mRNA and controls antibody secretion in plasma cells. EMBO J 33: 2922–2936.

Karaca E, Weitzer S, Pehlivan D, Shiraishi H, Gogakos T, Hanada T, Jhangiani SN, Wiszniewski W, Withers M, Campbell IM, et al. 2014. Human CLP1 Mutations Alter tRNA Biogenesis, Affecting Both Peripheral and Central Nervous System Function. Cell 157: 636–650.

Karolchik D, Barber GP, Casper J, Clawson H, Cline MS, Diekhans M, Dreszer TR, Fujita PA, Guruvadoo L, Haeussler M, et al. 2014. The UCSC Genome Browser database: 2014 update. Nucleic Acids Res 42: D764–70.

Katayama T, Imaizumi K, Sato N, Miyoshi K, Kudo T, Hitomi J, Morihara T, Yoneda T, Gomi F, Mori Y, et al. 1999. Presenilin-1 mutations downregulate the signalling pathway of the unfolded-protein response. Nat Cell Biol 1: 479–485.

Katona TM, Neubauer BL, Iversen PW, Zhang S, Baldridge LA, Cheng L. 2005. Elevated expression of angiogenin in prostate cancer and its precursors. Clin Cancer Res 11: 8358–8363.

Kaufman RJ. 2002. Orchestrating the unfolded protein response in health and disease. J Clin Invest 110: 1389–1398.

157 Kawahara T, Yanagi H, Yura T, Mori K. 1998. Unconventional Splicing of HAC1/ERN4 mRNA Required for the Unfolded Protein Response: Sequence-Specific And Non- Sequential Cleavage Of The Splice Sites. Journal of Biological Chemistry 273: 1802–1807.

Kelley PM, Schlesinger MJ. 1978. The effect of amino acid analogues and heat shock on gene expression in chicken embryo fibroblasts. Cell 15: 1277–1286.

Kenna M, Stevens A, McCammon M, Douglas MG. 1993. An essential yeast gene with homology to the exonuclease-encoding XRN1/KEM1 gene also encodes a protein with exoribonuclease activity. Mol Cell Biol 13: 341–350.

Kertesz M, Wan Y, Mazor E, Rinn JL, Nutter RC, Chang HY, Segal E. 2010. Genome- wide measurement of RNA secondary structure in yeast. Nature 467: 103–107.

Khan Y, Mansour I, Ong E, Shrestha M. 2015. Obstructive Jaundice as Initial Presentation of Multiple Myeloma: Case Presentation and Literature Review. Case Rep Med 2015: 686210–4.

Kim M, Krogan NJ, Vasiljeva L, Rando OJ, Nedea E, Greenblatt JF, Buratowski S. 2004. The yeast Rat1 exonuclease promotes transcription termination by RNA polymerase II. Nature 432: 517–522.

Kimata Y, Ishiwata-Kimata Y, Yamada S, Kohno K. 2005. Yeast unfolded protein response pathway regulates expression of genes for anti-oxidative stress and for cell surface proteins. Genes to Cells 11: 59–69.

Kimmig P, Diaz M, Zheng J, Williams CC, Lang A, Aragón T, Li H, Walter P. 2012. The unfolded protein response in fission yeast modulates stability of select mRNAs to maintain protein homeostasis. ed. R. Parker. Elife 1: e00048–e00048.

Kivioja T, Vähärautio A, Karlsson K, Bonke M, Enge M, Linnarsson S, Taipale J. 2012. Counting absolute numbers of molecules using unique molecular identifiers. Nature Publishing Group 9: 72–74.

Knapp G, Ogden RC, Peebles CL, Abelson J. 1979. Splicing of yeast tRNA precursors: structure of the reaction intermediates. Cell 18: 37–45.

Kodzius R, Kojima M, Nishiyori H, Nakamura M, Fukuda S, Tagami M, Sasaki D, Imamura K, Kai C, Harbers M. 2006. CAGE: cap analysis of gene expression. Nature methods 3: 211–222.

Kohno K, Normington K, Sambrook J, Gething MJ, Mori K. 1993. The promoter region of the yeast KAR2 (BiP) gene contains a regulatory domain that responds to the presence of unfolded proteins in the endoplasmic reticulum. Mol Cell Biol 13: 877– 890.

158 Kosmaczewski SG, Edwards TJ, Han SM, Eckwahl MJ, Meyer BI, Peach S, Hesselberth JR, Wolin SL, Hammarlund M. 2014. The RtcB RNA ligase is an essential component of the metazoan unfolded protein response. EMBO Rep 15: 1278–1285.

Kosmaczewski SG, Han SM, Han B, Irving Meyer B, Baig HS, Athar W, Lin-Moore AT, Koelle MR, Hammarlund M. 2015. RNA ligation in neurons by RtcB inhibits axon regeneration. Proc Natl Acad Sci USA 112: 8451–8456.

Koutmou KS, Schuller AP, Brunelle JL, Radhakrishnan A, Djuranovic S, Green R. 2015. Ribosomes slide on lysine-encoding homopolymeric A stretches. Elife 4: 446.

Kozutsumi Y, Segal M, Normington K, Gething MJ, Sambrook J. 1988. The presence of malfolded proteins in the endoplasmic reticulum signals the induction of glucose- regulated proteins. Nature 332: 462–464.

Kruger K, Grabowski PJ, Zaug AJ, Sands J, Gottschling DE, Cech TR. 1982. Self- splicing RNA: autoexcision and autocyclization of the ribosomal RNA intervening sequence of Tetrahymena. Cell 31: 147–157.

Kuroha K, Akamatsu M, Dimitrova L, Ito T, Kato Y, Shirahige K, Inada T. 2010. Receptor for activated C kinase 1 stimulates nascent polypeptide-dependent translation arrest. EMBO Rep 11: 956–961.

Labunskyy VM, Hatfield DL, Gladyshev VN. 2014. Selenoproteins: molecular pathways and physiological roles. Physiol Rev 94: 739–777.

Lacombe T, García-Gómez JJ, La Cruz De J, Roser D, Hurt E, Linder P, Kressler D. 2009. Linear ubiquitin fusion to Rps31 and its subsequent cleavage are required for the efficient production and functional integrity of 40S ribosomal subunits. Molecular Microbiology 72: 69–84.

Lange PH, Millan JL, Stigbrand T, Vessella RL, Ruoslahti E, Fishman WH. 1982. Placental alkaline phosphatase as a tumor marker for seminoma. Cancer Res 42: 3244–3247.

Langmead B, Trapnell C, Pop M, Salzberg SL. 2009. Ultrafast and memory-efficient alignment of short DNA sequences to the . Genome Biol 10: R25.

Larimer FW, Stevens A. 1990. Disruption of the gene XRN1, coding for a 5“----3” exoribonuclease, restricts yeast cell growth. Gene 95: 85–90.

Lee AS, Delegeane AM, Baker V, Chow PC. 1983. Transcriptional regulation of two genes specifically induced by glucose starvation in a hamster mutant fibroblast cell line. J Biol Chem 258: 597–603.

Lee HH, Kim Y-S, Kim KH, Heo I, Kim SK, Kim O, Kim HK, Yoon JY, Kim HS, Kim DJ, et al. 2007. Structural and functional insights into Dom34, a key component of no-go mRNA decay. Molecular Cell 27: 938–950.

159 Lee J, Ozcan U. 2014. Unfolded Protein Response Signaling and Metabolic Diseases. Journal of Biological Chemistry 289: 1203–1211.

Lee SR, Collins K. 2005. Starvation-induced cleavage of the tRNA anticodon loop in Tetrahymena thermophila. J Biol Chem 280: 42744–42749.

Lejeune F, Li X, Maquat LE. 2003. Nonsense-mediated mRNA decay in mammalian cells involves decapping, deadenylating, and exonucleolytic activities. Molecular Cell 12: 675–687.

Levin JZ, Yassour M, Adiconis X, Nusbaum C, Thompson DA, Friedman N, Gnirke A, Regev A. 2010. Comprehensive comparative analysis of strand-specific RNA sequencing methods. Nature Publishing Group 7: 709–715.

Levitz R, Chapman D, Amitsur M, Green R, Snyder L, Kaufmann G. 1990. The optional E. coli prr locus encodes a latent form of phage T4-induced anticodon nuclease. EMBO J 9: 1383–1389.

Liu Q, Greimann JC, Lima CD. 2006. Reconstitution, Activities, and Structure of the Eukaryotic RNA Exosome. Cell 127: 1223–1237.

Losson R, Lacroute F. 1979. Interference of nonsense mutations with eukaryotic messenger RNA stability. Proc Natl Acad Sci USA 76: 5134–5137.

Lowe TM, Eddy SR. 1997. tRNAscan-SE: a program for improved detection of transfer RNA genes in genomic sequence. Nucleic Acids Res 25: 955–964.

Lu Y, Liang F-X, Wang X. 2014. A Synthetic Biology Approach Identifies the Mammalian UPR RNA Ligase RtcB. Molecular Cell 1–13.

Luhtala N, Parker R. 2010. T2 Family ribonucleases: ancient enzymes with diverse roles. Trends Biochem Sci 35: 253–259.

Luo W, Bentley D. 2004. A ribonucleolytic rat torpedoes RNA polymerase II. Cell 119: 911–914.

Lykke-Andersen S, Brodersen DE, Jensen TH. 2009. Origins and activities of the eukaryotic exosome. J Cell Sci 122: 1487–1494.

Machnicka MA, Milanowska K, Osman Oglou O, Purta E, Kurkowska M, Olchowik A, Januszewski W, Kalinowski S, Dunin-Horkawicz S, Rother KM, et al. 2012. MODOMICS: a database of RNA modification pathways--2013 update. Nucleic Acids Res 41: D262–D267.

Mackie GA. 1998. Ribonuclease E is a 5'-end-dependent endonuclease. Nature 395: 720– 723.

160 Malathi K, Dong B, Gale M, Silverman RH. 2007. Small self-RNA generated by RNase L amplifies antiviral innate immunity. Nature 448: 816–819.

Malathi K, Saito T, Crochet N, Barton DJ, Gale M, Silverman RH. 2010. RNase L releases a small RNA from HCV RNA that refolds into a potent PAMP. RNA 16: 2108–2119.

Mao X, Schwer B, Shuman S. 1995. Yeast mRNA cap methyltransferase is a 50- kilodalton protein encoded by an essential gene. Mol Cell Biol 15: 4167–4174.

Maurel M, Chevet E, Tavernier J, Gerlo S. 2014. Getting RIDD of RNA: IRE1 in cell fate regulation. Trends Biochem Sci 39: 245–254.

Mitchell P, Petfalski E, Shevchenko A, Mann M, Tollervey D. 1997. The exosome: a conserved eukaryotic RNA processing complex containing multiple 3“-->5” exoribonucleases. Cell 91: 457–466.

Montero S, Guzmán C, Cortés-Funes H, Colomer R. 1998. Angiogenin expression and prognosis in primary breast carcinoma. Clin Cancer Res 4: 2161–2168.

Moore KA, Hollien J. 2012. The unfolded protein response in secretory cell function. Annu Rev Genet 46: 165–183.

Mori K, Kawahara T, Yoshida H, Yanagi H, Yura T. 1996. Signalling from endoplasmic reticulum to nucleus: transcription factor with a basic-leucine zipper motif is required for the unfolded protein-response pathway. Genes Cells 1: 803–817.

Mori K, Ma W, Gething MJ, Sambrook J. 1993. A transmembrane protein with a cdc2+/CDC28-related kinase activity is required for signaling from the ER to the nucleus. Cell 74: 743–756.

Mori K, Sant A, Kohno K, Normington K, Gething MJ, Sambrook JF. 1992. A 22 bp cis- acting element is necessary and sufficient for the induction of the yeast KAR2 (BiP) gene by unfolded proteins. EMBO J 11: 2583–2593.

Muhlrad D, Decker CJ, Parker R. 1994. Deadenylation of the unstable mRNA encoded by the yeast MFA2 gene leads to decapping followed by 5“-->3” digestion of the transcript. Genes & Development 8: 855–866.

Muhlrad D, Decker CJ, Parker R. 1995. Turnover mechanisms of the stable yeast PGK1 mRNA. Mol Cell Biol 15: 2145–2156.

Muhlrad D, Parker R. 2005. The yeast EDC1 mRNA undergoes deadenylation- independent decapping stimulated by Not2p, Not4p, and Not5p. EMBO J 24: 1033– 1045.

161 Murray JB, Seyhan AA, Walter NG, Burke JM, Scott WG. 1998. The hammerhead, hairpin and VS ribozymes are catalytically proficient in monovalent cations alone. Chem Biol 5: 587–595.

Newbury S, Woollard A. 2004. The 5“-3” exoribonuclease xrn-1 is essential for ventral epithelial enclosure during C. elegans embryogenesis. RNA 10: 59–65.

Nikawa J, Yamashita S. 1992. IRE1 encodes a putative protein kinase containing a membrane-spanning domain and is required for inositol phototrophy in Saccharomyces cerevisiae. Molecular Microbiology 6: 1441–1446.

Niwa M, Patil CK, DeRisi J, Walter P. 2005. Genome-scale approaches for discovering novel nonconventional splicing substrates of the Ire1 nuclease. Genome Biol 6: R3.

Normington K, Kohno K, Kozutsumi Y, Gething MJ, Sambrook J. 1989. S. cerevisiae encodes an essential protein homologous in sequence and function to mammalian BiP. Cell 57: 1223–1236.

Nykänen A, Haley B, Zamore PD. 2001. ATP requirements and small interfering RNA structure in the RNA interference pathway. Cell 107: 309–321.

Ogawa T, Tomita K, Ueda T, Watanabe K, Uozumi T, Masaki H. 1999. A cytotoxic ribonuclease targeting specific transfer RNA anticodons. Science 283: 2097–2100.

Ohno M, Sakamoto H, Shimura Y. 1987. Preferential excision of the 5' proximal intron from mRNA precursors with two introns as mediated by the cap structure. Proc Natl Acad Sci USA 84: 5187–5191.

Parker R. 2012. RNA degradation in Saccharomyces cerevisae. Genetics 191: 671–702.

Parker R, Sheth U. 2007. P bodies and the control of mRNA translation and degradation. Molecular Cell 25: 635–646.

Passos DO, Doma MK, Shoemaker CJ, Muhlrad D, Green R, Weissman J, Hollien J, Parker R. 2009. Analysis of Dom34 and its function in no-go decay. Mol Biol Cell 20: 3025–3032. http://www.molbiolcell.org/content/20/13/3025.full.

Paushkin SV, Patel M, Furia BS, Peltz SW, Trotta CR. 2004. Identification of a Human Endonuclease Complex Reveals a Link between tRNA Splicing and Pre-mRNA 3′ End Formation. Cell 117: 311–321.

Peach SE, York K, Hesselberth JR. 2015. Global analysis of RNA cleavage by 5'- hydroxyl RNA sequencing. Nucleic Acids Res 43: e108–e108.

Peebles CL, Gegenheimer P, Abelson J. 1983. Precise excision of intervening sequences from precursor tRNAs by a membrane-associated yeast endonuclease. Cell 32: 525– 536.

162 Pelechano V, Wei W, Steinmetz LM. 2016. Genome-wide quantification of 5'- phosphorylated mRNA degradation intermediates for analysis of ribosome dynamics. Nat Protoc 11: 359–376.

Perry RP. 1962. The cellular sites of synthesis of ribosomal and 4S RNA. Proc Natl Acad Sci USA 48: 2179–2186.

Perry RP, Srinivasan PR, Kelley DE. 1964. Hybridization of rapidly labeled nuclear ribonucleic acids. Science 145: 504–507.

Peschek J, Acosta-Alvear D, Mendez AS, Walter P. 2015. A conformational RNA zipper promotes intron ejection during non-conventional XBP1 mRNA splicing. EMBO Rep 16: 1688–1698.

Petfalski E, Dandekar T, Henry Y, Tollervey D. 1998. Processing of the precursors to small nucleolar RNAs and rRNAs requires common components. Mol Cell Biol 18: 1181–1189.

Phizicky EM, Alfonzo JD. 2010. Do all modifications benefit all tRNAs? FEBS Lett 584: 265–271.

Phizicky EM, Consaul SA, Nehrke KW, Abelson J. 1992. Yeast tRNA ligase mutants are nonviable and accumulate tRNA splicing intermediates. J Biol Chem 267: 4577– 4582.

Phizicky EM, Schwartz RC, Abelson J. 1986. Saccharomyces cerevisiae tRNA ligase. Purification of the protein and isolation of the structural gene. J Biol Chem 261: 2978–2986.

Poole TL, Stevens A. 1995. Comparison of features of the RNase activity of 5“- exonuclease-1 and 5-”exonuclease-2 of Saccharomyces cerevisiae. Nucleic Acids Symp Ser 79–81.

Popow J, Englert M, Weitzer S, Schleiffer A, Mierzwa B, Mechtler K, Trowitzsch S, Will CL, Luhrmann R, Soll D, et al. 2011. HSPC117 Is the Essential Subunit of a Human tRNA Splicing Ligase Complex. Science 331: 760–764.

Prody GA, Bakos JT, Buzayan JM, Schneider IR, Bruening G. 1986. Autolytic processing of dimeric plant virus satellite RNA. Science 231: 1577–1580.

Quinlan AR, Hall IM. 2010. BEDTools: a flexible suite of utilities for comparing genomic features. Bioinformatics 26: 841–842.

Raines RT. 1998. Ribonuclease A. Chem Rev 98: 1045–1066.

Ray A, Zhang S, Rentas C, Caldwell KA, Caldwell GA. 2014. RTCB-1 mediates neuroprotection via XBP-1 mRNA splicing in the unfolded protein response pathway. J Neurosci 34: 16076–16085.

163 Remus BS, Shuman S. 2014. Distinctive kinetics and substrate specificities of plant and fungal tRNA ligases. RNA 20: 462–473.

Ron D, Walter P. 2007. Signal integration in the endoplasmic reticulum unfolded protein response. Nat Rev Mol Cell Biol 8: 519–529.

Roth A, Weinberg Z, Chen AGY, Kim PB, Ames TD, Breaker RR. 2014. A widespread self-cleaving ribozyme class is revealed by bioinformatics. Nat Chem Biol 10: 56–60.

Rüegsegger U, Leber JH, Walter P. 2001. Block of HAC1 mRNA translation by long- range base pairing is released by cytoplasmic splicing upon induction of the unfolded protein response. Cell 107: 103–114.

Saikia M, Hatzoglou M. 2015. The Many Virtues of tRNA-derived Stress-induced RNAs (tiRNAs): Discovering Novel Mechanisms of Stress Response and Effect on Human Health. Journal of Biological Chemistry 290: 29761–29768.

Saikia M, Jobava R, Parisien M, Putnam A, Krokowski D, Gao X-H, Guan B-J, Yuan Y, Jankowsky E, Feng Z, et al. 2014. Angiogenin-cleaved tRNA halves interact with cytochrome c, protecting cells from apoptosis during osmotic stress. Mol Cell Biol 34: 2450–2463.

Saïda F, Uzan M, Bontems F. 2003. The phage T4 restriction endoribonuclease RegB: a cyclizing enzyme that requires two histidines to be fully active. Nucleic Acids Res 31: 2751–2758.

Sanson B, Hu RM, Troitskayadagger E, Mathy N, Uzan M. 2000. Endoribonuclease RegB from bacteriophage T4 is necessary for the degradation of early but not middle or late mRNAs. J Mol Biol 297: 1063–1074.

Saville BJ, Collins RA. 1990. A site-specific self-cleavage reaction performed by a novel RNA in Neurospora mitochondria. Cell 61: 685–696.

Sawaya R. 2003. Genetic and Biochemical Analysis of the Functional Domains of Yeast tRNA Ligase. Journal of Biological Chemistry 278: 43928–43938.

Saxena SK, Rybak SM, Davey RT, Youle RJ, Ackerman EJ. 1992. Angiogenin is a cytotoxic, tRNA-specific ribonuclease in the RNase A superfamily. J Biol Chem 267: 21982–21986.

Schaffer AE, Eggens VRC, Caglayan AO, Reuter MS, Scott E, Coufal NG, Silhavy JL, Xue Y, Kayserili H, Yasuno K, et al. 2014. CLP1 Founder Mutation Links tRNA Splicing and Maturation to Cerebellar Development and Neurodegeneration. Cell 157: 651–663.

Scherrer K. 2003. Historical review: the discovery of “giant” RNA and RNA processing: 40 years of enigma. Trends Biochem Sci 28: 566–571.

164 Scherrer K, Darnell JE. 1962. Sedimentation characteristics of rapidly labelled RNA from HeLa cells. Biochem Biophys Res Commun 7: 486–490.

Scherrer K, Latham H, Darnell JE. 1963. Demonstration of an unstable RNA and of a precursor to ribosomal RNA in HeLa cells. Proc Natl Acad Sci USA 49: 240–248.

Schifano JM, Vvedenskaya IO, Knoblauch JG, Ouyang M, Nickels BE, Woychik NA. 2014. An RNA-seq method for defining endoribonuclease cleavage specificity identifies dual rRNA substrates for toxin MazF-mt3. Nature Communications 5: 3538.

Schillewaert S, Wacheul L, Lhomme F, Lafontaine DLJ. 2012. The evolutionarily conserved protein Las1 is required for pre-rRNA processing at both ends of ITS2. Mol Cell Biol 32: 430–444.

Schuck S, Prinz WA, Thorn KS, Voss C, Walter P. 2009. Membrane expansion alleviates endoplasmic reticulum stress independently of the unfolded protein response. The Journal of Cell Biology 187: 525–536.

Schutz K, Hesselberth JR, Fields S. 2010. Capture and sequence analysis of RNAs with terminal 2“,3-”cyclic phosphates. RNA 16: 621–631.

Schwer B, Aronova A, Ramirez A, Braun P, Shuman S. 2008. Mammalian 2“,3” cyclic nucleotide phosphodiesterase (CNP) can function as a tRNA splicing enzyme in vivo. RNA 14: 204–210.

Shaheen HH, Hopper AK. 2005. Retrograde movement of tRNAs from the cytoplasm to the nucleus in Saccharomyces cerevisiae. Proc Natl Acad Sci USA 102: 11290– 11295.

Sharmeen L, Kuo MY, Dinter-Gottlieb G, Taylor J. 1988. Antigenomic RNA of human hepatitis delta virus can undergo self-cleavage. J Virol 62: 2674–2679.

Shatkin AJ. 1976. Capping of eucaryotic mRNAs. Cell 9: 645–653.

Shimotohno K, Kodama Y, Hashimoto J, Miura KI. 1977. Importance of 5'-terminal blocking structure to stabilize mRNA in eukaryotic protein synthesis. Proc Natl Acad Sci USA 74: 2734–2738.

Shimoyama S, Yamasaki K, Kawahara M, Kaminishi M. 1999. Increased serum angiogenin concentration in colorectal cancer is correlated with cancer progression. Clin Cancer Res 5: 1125–1130.

Shiraki T, Kondo S, Katayama S, Waki K, Kasukawa T, Kawaji H, Kodzius R, Watahiki A, Nakamura M, Arakawa T. 2003. Cap analysis gene expression for high- throughput analysis of transcriptional starting point and identification of promoter usage. Proc Natl Acad Sci USA 100: 15776–15781.

165 Shoemaker CJ, Green R. 2012. Translation drives mRNA quality control. Nat Struct Mol Biol 19: 594–601.

Sidrauski C, Cox JS, Walter P. 1996. tRNA ligase is required for regulated mRNA splicing in the unfolded protein response. Cell 87: 405–413.

Sidrauski C, Walter P. 1997. The transmembrane kinase Ire1p is a site-specific endonuclease that initiates mRNA splicing in the unfolded protein response. Cell 90: 1031–1039.

Sigurgeirsson B, Emanuelsson O, Lundeberg J. 2014. Analysis of stranded information using an automated procedure for strand specific RNA sequencing. BMC Genomics 15: 631.

Sonenberg N, Rupprecht KM, Hecht SM, Shatkin AJ. 1979. Eukaryotic mRNA cap binding protein: purification by affinity chromatography on sepharose-coupled m7GDP. Proc Natl Acad Sci USA 76: 4345–4349.

Song Y, Sretavan D, Salegio EA, Berg J, Huang X, Cheng T, Xiong X, Meltzer S, Han C, Nguyen T-T, et al. 2015. Regulation of axon regeneration by the RNA repair and splicing pathway. Nat Neurosci 18: 817–825.

Soukup GA, Breaker RR. 1999. Relationship between internucleotide linkage geometry and the stability of RNA. RNA 5: 1308–1325.

Spickler C, Stronge V, Mackie GA. 2001. Preferential cleavage of degradative intermediates of rpsT mRNA by the Escherichia coli RNA . J Bacteriol 183: 1106–1109.

Spinelli SL, Consaul SA, Phizicky EM. 1997. A conditional lethal yeast phosphotransferase (tpt1) mutant accumulates tRNAs with a 2'-phosphate and an undermodified base at the splice junction. RNA 3: 1388.

Sprinzl M, Vassilenko KS. 2005. Compilation of tRNA sequences and sequences of tRNA genes. Nucleic Acids Res 33: D139–40.

Stark BC, Kole R, Bowman EJ, Altman S. 1978. Ribonuclease P: an enzyme with an essential RNA component. Proc Natl Acad Sci USA 75: 3717–3721.

Steitz TA, Steitz JA. 1993. A general two-metal-ion mechanism for catalytic RNA. Proc Natl Acad Sci USA 90: 6498–6502.

Stevens A. 2001. 5′-Exoribonuclease 1: Xrn1. In Ribonucleases - Part B, Vol. 342 of Methods in Enzymology, pp. 251–259, Elsevier.

Stevens A. 1980. Purification and characterization of a Saccharomyces cerevisiae exoribonuclease which yields 5'-mononucleotides by a 5‘ leads to 3’ mode of hydrolysis. J Biol Chem 255: 3080–3085.

166 Stevens A, Wang Y, Bremer K, Zhang J, Hoepfner R, Antoniou M, Schoenberg DR, Maquat LE. 2002. Beta -Globin mRNA decay in erythroid cells: UG site-preferred endonucleolytic cleavage that is augmented by a premature termination codon. Proc Natl Acad Sci USA 99: 12741–12746.

Tachibana C, Stevens TH. 1992. The yeast EUG1 gene encodes an endoplasmic reticulum protein that is functionally related to protein disulfide . Mol Cell Biol 12: 4601–4611.

Takagi T, Moore CR, Diehn F, Buratowski S. 1997. An RNA 5'-triphosphatase related to the protein tyrosine phosphatases. Cell 89: 867–873.

Takagi T, Taylor GS, Kusakabe T, Charbonneau H, Buratowski S. 1998. A protein tyrosine phosphatase-like protein from baculovirus has RNA 5'-triphosphatase and diphosphatase activities. Proc Natl Acad Sci USA 95: 9808–9812.

Takano A. 2005. tRNA Actively Shuttles Between the Nucleus and in Yeast. Science 309: 140–142.

Tam AB, Koong AC, Niwa M. 2014. Ire1 has distinct catalytic mechanisms for XBP1/HAC1 splicing and RIDD. Cell Rep 9: 850–858.

Tanaka N, Meineke B, Shuman S. 2011. RtcB, a Novel RNA Ligase, Can Catalyze tRNA Splicing and HAC1 mRNA Splicing in Vivo. Journal of Biological Chemistry 286: 30253–30257.

Tanaka N, Shuman S. 2011. RtcB is the RNA ligase component of an Escherichia coli RNA repair operon. J Biol Chem 286: 7727–7731.

Thompson DM, Lu C, Green PJ, Parker R. 2008. tRNA cleavage is a conserved response to oxidative stress in eukaryotes. RNA 14: 2095–2103.

Thompson DM, Parker R. 2009a. Stressing out over tRNA cleavage. Cell 138: 215–219.

Thompson DM, Parker R. 2009b. The RNase Rny1p cleaves tRNAs and promotes cell death during oxidative stress in Saccharomyces cerevisiae. J Cell Biol 185: 43–50.

Torres AG, Batlle E, Ribas de Pouplana L. 2014. Role of tRNA modifications in human diseases. Trends Mol Med 20: 306–314.

Travers KJ, Patil CK, Wodicka L, Lockhart DJ, Weissman JS, Walter P. 2000. Functional and genomic analyses reveal an essential coordination between the unfolded protein response and ER-associated degradation. Cell 101: 249–258.

Trotta CR, Miao F, Arn EA, Stevens SW, Ho CK, Rauhut R, Abelson JN. 1997. The yeast tRNA splicing endonuclease: a tetrameric enzyme with two subunits homologous to the archaeal tRNA endonucleases. Cell 89: 849–858.

167 Tsuboi T, Kuroha K, Kudo K, Makino S, Inoue E, Kashima I, Inada T. 2012. Dom34:hbs1 plays a general role in quality-control systems by dissociation of a stalled ribosome at the 3' end of aberrant mRNA. Mol Cell 46: 518–529.

Tsukamoto T, Shibagaki Y, Imajoh-Ohmi S, Murakoshi T, Suzuki M, Nakamura A, Gotoh H, Mizumoto K. 1997. Isolation and characterization of the yeast mRNA capping enzyme beta subunit gene encoding RNA 5'-triphosphatase, which is essential for cell viability. Biochem Biophys Res Commun 239: 116–122.

Tucker M, Valencia-Sanchez MA, Staples RR, Chen J, Denis CL, Parker R. 2001. The transcription factor associated Ccr4 and Caf1 proteins are components of the major cytoplasmic mRNA deadenylase in Saccharomyces cerevisiae. Cell 104: 377–386.

Uzan M, Favre R, Brody E. 1988. A nuclease that cuts specifically in the ribosome binding site of some T4 mRNAs. Proc Natl Acad Sci USA 85: 8895–8899.

Valenzuela P, Venegas A, Weinberg F, Bishop R, Rutter WJ. 1978. Structure of yeast phenylalanine-tRNA genes: an intervening DNA segment within the region coding for the tRNA. Proc Natl Acad Sci USA 75: 190–194. van Dijk E, Cougot N, Meyer S, Babajko S, Wahle E, Séraphin B. 2002. Human Dcp2: a catalytically active mRNA decapping enzyme located in specific cytoplasmic structures. EMBO J 21: 6915–6924. van Es MA, Schelhaas HJ, van Vught PWJ, Ticozzi N, Andersen PM, Groen EJN, Schulte C, Blauw HM, Koppers M, Diekstra FP, et al. 2011. Angiogenin variants in Parkinson disease and amyotrophic lateral sclerosis. Ann Neurol 70: 964–973. van Hoof A, Frischmeyer PA, Dietz HC, Parker R. 2002. Exosome-mediated recognition and degradation of mRNAs lacking a termination codon. Science 295: 2262–2264. van Hoof A, Lennertz P, Parker R. 2000a. Three conserved members of the RNase D family have unique and overlapping functions in the processing of 5S, 5.8S, U4, U5, RNase MRP and RNase P RNAs in yeast. EMBO J 19: 1357–1365. van Hoof A, Staples RR, Baker RE, Parker R. 2000b. Function of the Ski4p (Csl4p) and Ski7p Proteins in 3“-to-5” Degradation of mRNA. Mol Cell Biol 20: 8230–8243.

Veldman GM, Klootwijk J, van Heerikhuizen H, Planta RJ. 1981. The nucleotide sequence of the intergenic region between the 5.8S and 26S rRNA genes of the yeast ribosomal RNA operon. Possible implications for the interaction between 5.8S and 26S rRNA and the processing of the . Nucleic Acids Res 9: 4847– 4862.

Venema J, Henry Y, Tollervey D. 1995. Two distinct recognition signals define the site of endonucleolytic cleavage at the 5'-end of yeast 18S rRNA. EMBO J 14: 4883– 4892.

168 Walter P, Ron D. 2011. The Unfolded Protein Response: From Stress Pathway to Homeostatic Regulation. Science 334: 1081–1086.

Wan J, Yourshaw M, Mamsa H, Rudnik-Schöneborn S, Menezes MP, Hong JE, Leong DW, Senderek J, Salman MS, Chitayat D, et al. 2012. Mutations in the RNA exosome component gene EXOSC3 cause pontocerebellar hypoplasia and spinal motor neuron degeneration. Nat Genet 44: 704–708.

Wang Z, Gerstein M, Snyder M. 2009. RNA-Seq: a revolutionary tool for transcriptomics. Nat Rev Genet 10: 57–63.

Weitzer S, Martinez J. 2007a. hClp1: a novel kinase revitalizes RNA metabolism. Cell Cycle 6: 2133–2137.

Weitzer S, Martinez J. 2007b. The human RNA kinase hClp1 is active on 3′ transfer RNA exons and short interfering RNAs. Nature 447: 222–226.

Wilson MA, Meaux S, van Hoof A. 2007. A genomic screen in yeast reveals novel aspects of nonstop mRNA metabolism. Genetics 177: 773–784.

Winkler WC, Nahvi A, Roth A, Collins JA, Breaker RR. 2004. Control of gene expression by a natural metabolite-responsive ribozyme. Nature 428: 281–286.

Woolford JL, Baserga SJ. 2013. Ribosome biogenesis in the yeast Saccharomyces cerevisiae. Genetics 195: 643–681.

Wu J, Bao A, Chatterjee K, Wan Y, Hopper AK. 2015. Genome-wide screen uncovers novel pathways for tRNA processing and nuclear-cytoplasmic dynamics. Genes & Development 29: 2633–2644.

Wu J, Hopper AK. 2014. Healing for destruction: tRNA intron degradation in yeast is a two-step cytoplasmic process catalyzed by tRNA ligase Rlg1 and 5“-to-3” exonuclease Xrn1. Genes & Development 28: 1556–1561.

Xu Q, Teplow D, Lee TD, Abelson J. 1990. Domain structure in yeast tRNA ligase. Biochemistry 29: 6132–6138.

Xue Y, Ouyang K, Huang J, Zhou Y, Ouyang H, Li H, Wang G, Wu Q, Wei C, Bi Y, et al. 2013. Direct conversion of fibroblasts to neurons by reprogramming PTB- regulated microRNA circuits. Cell 152: 82–96.

Yamasaki S, Ivanov P, Hu G-F, Anderson P. 2009. Angiogenin cleaves tRNA and promotes stress-induced translational repression. J Cell Biol 185: 35–42.

Yanagitani K, Imagawa Y, Iwawaki T, Hosoda A, Saito M, Kimata Y, Kohno K. 2009. Cotranslational targeting of XBP1 protein to the membrane promotes cytoplasmic splicing of its own mRNA. Mol Cell 34: 191–200.

169 Yanagitani K, Kimata Y, Kadokura H, Kohno K. 2011. Translational pausing ensures membrane targeting and cytoplasmic splicing of XBP1u mRNA. Science 331: 586– 589.

Yang W. 2011. : diversity of structure, function and mechanism. Q Rev Biophys 44: 1–93.

Yoo SM, Choi JH, Lee SY, Yoo NC. 2009. Applications of DNA microarray in disease diagnostics. J Microbiol Biotechnol 19: 635–646.

Zhang P, McGrath B, Li S, Frank A, Zambito F, Reinert J, Gannon M, Ma K, McNaughton K, Cavener DR. 2002. The PERK eukaryotic initiation factor 2 alpha kinase is required for the development of the skeletal system, postnatal growth, and the function and viability of the pancreas. Mol Cell Biol 22: 3864–3874.

Zhang Y, Liu T, Meyer CA, Eeckhoute J, Johnson DS, Bernstein BE, Nusbaum C, Myers RM, Brown M, Li W, et al. 2008. Model-based analysis of ChIP-Seq (MACS). Genome Biol 9: R137.

Zook JM, Samarov D, McDaniel J, Sen SK, Salit M. 2012. Synthetic spike-in standards improve run-specific systematic error analysis for DNA and RNA sequencing. ed. J. Kelso. PLoS ONE 7: e41356.

170