This item is only available as the following downloads:
THE SEARCH FOR MICRORNAS ENCODED BY THE INFLUENZA A VIRUS BY KATHLEEN MAXWELL A THESIS Submitted to the Division of Natural Sciences New College of Florida in partial fulfillment of the requirements for the degree Bachelor of Arts Under the Sponsorship of Dr. Katherine Walstrom Sarasota, Florida May, 2011
Table of Contents Title Page Acknowledgements vii List of Tables viii List of Figures ix Abstract xiii Chapter 1: Introduction 1 1.1 Gene Expression Overview 4 1.1.1 Transcription 4 1.1.2 Translation 8 1.1.3 Regulation of Gene Expression 9 1.2 The Influenza A Virus 11 1.2.1 Basic Information 11 1.2.2 The Infection Process 12 1.2.3 Regulation of Gene Expression 16 1.2.4 Spread of Viral Particles 18 1.2.5 Mutations of the Influenza A Virus 18 1.3 Small Regulatory RNA Molecules 19 1.3.1 MicroRNAs 19 1.3.2 Small Interfering RNAs 26 1.4 MicroRNAs and Viruses 27 1.4.1 MicroRNAs Encoded by DNA Viruses 27 1.4.2 Conservation of MicroRNAs in Viruses 28 1.4.3 Roles of Virally Encoded MicroRNAs 29
1.4.4 Influenza A and MicroRNAs 30 1.5 Methods for Identifying MicroRNAs 30 1.5.1 Computational Methods 30 1.5.2 Experimental Verification of Computationally Predicted 32 MicroRNAs 126.96.36.199 Northern Blotting 32 188.8.131.52 Locked Nucleic Acids to Increase the Sensit ivity of 32 Northern Blotting 184.108.40.206 Microarrays 33 220.127.116.11 High-throughput Sequencing 34 18.104.22.168 Polymerase Chain Reaction 35 22.214.171.124 Quantitative Polymerase Chain Reaction 36 126.96.36.199 Stem-Loop Reverse Transcriptase QPCR 3 7 188.8.131.52 Primer Extension QPCR 38 Chapter 2: Materials and Methods 40 2.1 Overview 40 2.1.1Selection of an Influenza Strain 42 2.1.2 Use of VMir for Pre-miRNA Hairpin Prediction 42 2.1.3 Selection of Hairpins 43 2.2 Preparation of Viral RNA 45 2.2.1 Cell Growth and Infection 45 2.2.2 RNA Isolation 46 2.2.3 Small RNA Isolation 47 2.3 Preparation of QPCR Components 48
2.3.1 Primer Design 48 2.3.2 Reverse Transcription of RNA 49 2.3.3 Creation of a pool of anchor-tailed cDNAs 51 2.3.4 PCR Procedure 52 2.4 Further Analysis of QPCR Products 53 2.4.1 DNA Gel Electrophoresis 53 2.4.2 PCR Product Purification 54 2.4.3 Plasmid Creation 55 2.4.4 Transfection of E. coli 55 2.4.5 Purification of DNA for Sequencing 56 2.5 Procedure to Decrease the Expression Levels of Dicer 57 2.5.1 Procedure for Dicer shRNA 57 2.5.2 Transfection of cells 59 2.6 QPCR Data Analysis 61 Chapter 3: Results and Discussion 63 3.1 Pre-miRNA Hairpins 64 3.1.1 Predicted Hairpins 64 3.1.2 QPCR Results 65 3.1.3 DNA Gel Electrophoresis Results 69 3.2 The Search for MicroRNAs 74 3.2.1 Mature MicroRNAs 75 3.2.2 Other Sequencing Results 84 184.108.40.206 Hairpin 5 Primers 5 3-1 and 5 3-2 84
220.127.116.11 Hairpin 7-1 Primer 7-1 3-2 89 18.104.22.168 Hairpin 7-2 Primers 7-2 3-1, 7-2 3-2, 7 -2 3-4, 93 and 7-2 3-5 3.2.3 Elimination of Possible miRNAs 101 3.3 The Attempt to Decrease Expression Levels of Di cer 103 3.4 Future Work 106 3.5 Improvements 107 3.5.1 Basic Technique 107 3.5.2 Procedure 109 Chapter 4: Conclusions 112 Appendix 113 A.1 Sequence and VMir Diagram of Cal04 Hairpins and Subhairpins 113 A. 2 Sequence of oligoDT Adaptor 114 A.3 Sequence of Universal Reverse Primer 114 A.4 Forward Primers for Predicted Hairpins in Calif ornia 04 114 A.5 Primers for Cal04 Hairpins 116 A.6 Sequence of miR-16 Forward Primer 116 A.7 Vector Information 117 A.8 Primers for Dicer qPCR 119 A.9 Sequencing Results for all small, virally encod ed RNAs 119 A.10 MRNAs with seed sequence match and greater tha n 119 50% complementarity to the possible viral miRNA amplified by primer 7-1 3-1
A.11 Primers that formed primer-dimers 120 A.12 Primers with inconsistent amplification betwee n PCR replicates 120 References 121
Acknowledgements I would like to thank Dr. Kevin Harrod and Dr. Robe rt Rubin of Lovelace Respiratory Research Institute (Albuquerque, NM) fo r giving me the opportunity and funding to conduct this research. I would also like to thank Dr. Harrod for his guidance and positive support that he gave throughout this p rocess, both while I was Albuquerque conducting research and when I returned to New Coll ege to begin the writing process. Also from Lovelace, Jesse VanWestrienen, Dr. Wanli Lei, and Richard Jaramillo were always available to answer questions and help with various procedures. At New College, my sponsor Dr. Katherine Walstrom h as provided an incredible amount of support and feedback throughout the proce ss. I appreciate her availability and the genuine interest she has shown in my future pla ns. That guidance has made thoughts of the upcoming year slightly less daunting. Dr. Ch ris Hart has been immensely helpful, both throughout the summer in Albuquerque, and duri ng this school year. I would also like to thank Dr. Suzanne Sherman for her support t hrough the years and for taking the time to be a part of my committee.
List of Tables 1.1 Documented Influenza pandemics 11 1.2 Major classes of small non-coding regulatory RNA mo lecules 19 2.1 Data including number of cells and virions for each 46 infection study 2.2 Concentration and purity data for cDNAs used for 50 qPCR for hairpins 2.3 Concentration and purity data for cDNAs used for 51 qPCR for miRNAs 2.4 Amounts of reagents used for transfections 60 3.1 Concentrations of cDNA used in each qPCR experiment 67
List of Figures 1.1 Antiparallel DNA strands 5 1.2 Transcription initiation 6 1.3 RNA polymerase II 7 1.4 Replication and transcription of a negative-sense R NA virus 14 1.5 Transcription of viral mRNAs 15 1.6 Transcription of miRNAs 20 1.7 Processing pathways of miRNAs and siRNAs 23 1.8 PAZ domain of Ago2 24 1.9 Ago2 protein 25 1.10 A locked nucleic acid 33 1.11 The PCR process 36 1.12 Stem-loop primer method of preparing RNA for PCR 37 1.13 Primer extension method of preparing RNA for PCR 39 2.1 Flowchart of experimental procedure 41 2.2 VMir output 43 2.3 A pre-miRNA hairpin predicted by VMir 44 3.1 Locations of predicted pre-miRNA hairpins in the 64 California 04 Genome 3.2 QPCR results for Cal04 #1 pre-miRNA hairpins 65 3.3 QPCR results for Cal04 #2 pre-miRNA hairpins 66 3.4 DNA gel electrophoresis results for Cal04 #1 hairpi ns 7-1 and 7-3 69 3.5 DNA gel electrophoresis results for Cal04 #1 hairpi ns 5 and 7-2 70 and Cal04 #2 hairpin 7-3
3.6 DNA gel electrophoresis results for Cal04 #2 hairpi ns 5, 7-1, 71 and 7-2 3.7 BLAST result for the forward primer for hairpin 7-2 72 3.8 BLAST result for the reverse primer for hairpin 7-2 73 3.9 Locations of primers complementary to regions of ha irpin 7-1 75 3.10 Sequencing results for cDNA amplified by primer 7-1 3-1 76 3.11 Cal04 #1 DNA gel electrophoresis results for cDNAs amplified 77 by primer 7-1 3-1 3.12 Cal04 #2 DNA gel electrophoresis results for cDNAs amplified 77 by primer 7-1 3-1 3.13 BLAST result for primer 7-1 3-1 78 3.14 BLAST result for cDNA amplified by qPCR with primer 7-1 3-1 79 3.15 MiRNA-mRNA base pairing 80 3.16 QPCR results for Cal04 #1 primer 7-1 3-1 82 3.17 QPCR results for Cal04 #2 primer 7-1 3-1 83 3.18 Locations of primers complementary to regions of ha irpin 5 84 3.19 Sequencing results for cDNAs amplified by primers 5 3-1 85 and 5 3-2 3.20 Cal04 #1 DNA gel electrophoresis results for cDNAs amplified 85 by primers 5 3-1 and 5 3-2 3.21 Cal04 #2 DNA gel electrophoresis results for cDNAs amplified 86 by primers 5 3-1 and 5 3-2 3.22 QPCR results for Cal04 #1 primers 5 3-1 and 5 3-2 86 3.23 QPCR results for Cal04 #2 primers 5 3-1 and 5 3-2 87 3.24 Sequencing results for cDNA amplified by primer 7-1 3-2 89
3.25 QPCR results for Cal04 #1 primer 7-2 3-2 90 3.26 QPCR results for Cal04 #2 primer 7-2 3-2 90 3.27 BLAST result for primer 7-1 3-2 91 3.28 Cal04 #2 DNA gel electrophoresis results for cDNAs amplified 92 by primer 7-1 3-2 3.29 Locations of primers complementary to regions of ha irpin 7-2 93 3.30 Sequencing results for cDNAs amplified by primers 7 -2 3-1, 94 7-2 3-4, and 7-2 3-5 3.31 Cal04 #1 DNA gel electrophoresis results for cDNAs amplified 95 by primer 7-2 3-1 3.32 Cal04 #1 DNA gel electrophoresis results for cDNAs amplified 95 by primer 7-2 3-4 3.33 Cal04 #1 DNA gel electrophoresis results for cDNAs amplified 96 by primer 7-2 3-5 3.34 Cal04 #2 DNA gel electrophoresis results for cDNAs amplified 96 by primers 7-2 3-1, 7-2 3-4, and 7-2 3-5 3.35 BLAST result for the sequence of the cDNA amplified by 97 primer 7-2 3-1 3.36 BLAST result for the sequence of the cDNA amplified by 97 primer 7-2 3-4 3.37 BLAST result for the sequence of the cDNA amplified by 98 primer 7-2 3-5 3.38 QPCR Results for Cal04 #1 Primers 7-2 3-1, 7-2 34, and 7-2 3-5 99 3.39 QPCR results for Cal04 #2 primers 7-2 3-1, 7-2 34, and 7-2 3-5 100 3.40 QPCR results for Cal04 #1 primer 7-1 5-1 102 3.41 QPCR results for Cal04 #2 primer 7-1 5-1 102
3.42 Cal04 #1 DNA gel electrophoresis results for cDNAs amplified 103 by primers 7-3 5-1 and 7-3 5-2 3.43 Fluorescent microscopy images of 293T cells after t ransfection 104 with GIPz 3.44 Fluorescent microscopy images of A549 cells after t ransfection 104 with GIPz 3.45 QPCR results of the expression levels of Dicer 105
THE SEARCH FOR MICRORNAS ENCODED BY THE INFLUENZA A VIRUS Kathleen Maxwell New College of Florida, 2011 ABSTRACT MicroRNAs are small non-coding regulatory RNA mole cules. They regulate protein translation by binding to mRNAs preventing the mRNAs from being translated into proteins. MicroRNAs have been found in most eu karyotic cells and in DNA viruses such as the herpesviruses; they have not yet been f ound in RNA viruses. Influenza A is a RNA virus that causes a huge economic and health im pact worldwide. This study identified one possible mature microRNA encoded by the 2009 Swine Flu strain of the Influenza A virus. Several pre-miRNA hairpins were predicted using a computational program designed to predict pre-miRNAs within viral genomes. Possible microRNAs within these hairpins were identified and examined using qPCR, DNA gel electrophoresis, and sequencing. One possible matur e microRNA was identified that was in the proper location on a predicted pre-miRNA hai rpin and was the proper length to be a mature miRNA. Dr. Katherine Walstrom Division of Natural Sciences
Chapter 1: Introduction The Influenza A virus is found worldwide, and each year there are over 35,000 flu-related deaths (Thompson et al. 2003). In addit ion to being implicated in a large number of deaths, Influenza A has a huge economic i mpact. Every year millions of dollars are spent on surveillance of wild and farme d poultry and swine because these animals carry strains of influenza that may spread to humans. Each time an influenza outbreak is detected in farmed poultry, the poultry industry loses from $25 million to over $100 million to euthanize infected animals and test and quarantine nearby animals (Swayne and Akey 2005). However, these costs are sm all when compared to the potential cost of a pandemic. Metzer et al. (1999) estimated that a pandemic would cost the United States $71.3 billion to $166.5 billion. Another stu dy by Smith et al. (2009) estimated that a pandemic would cost the United Kingdom $10.9 bill ion to $21.7 billion if mortality rates were low and $71.6 billion to $95.5 billion i f mortality rates were high. Pandemics occur when a particularly virulent strain of the fl u emerges and spreads across the world. As travel among countries and continents has increa sed, the spread of diseases, such as Influenza A, has increased. In addition to devastat ing the economy of a country, these pandemics can affect the composition of the populat ion; the 1918 Spanish Flu pandemic, the spread of which was facilitated by World War I, lowered the life expectancy in the U.S. by 12 years (Palese 2004). The Influenza A virus has been and continues to be devastating because it mutates rapidly and is therefore able to spread through the same populations year after year. Some viruses mutate very slowly and so the host is able to develop antibodies against the virus
that decrease the risk of future infections by that virus. However, the Influenza A virus mutates quickly and generates a huge number of stra ins of the virus. Most strains are different enough that antibodies against one strain do not necessarily confer protection against another strain (Boni et al. 2008). The influenza virus is a ribonucleic acid (RNA) vir us, which means that it uses RNA, rather than deoxyribonucleic acid (DNA), as it s genetic material. RNA-dependent RNA polymerases that replicate RNA do not, unlike D NA polymerases, have a mechanism for proofreading. This means that errors are introduced into RNA genomes more frequently than DNA genomes during replication leading to a higher rate of mutation and the development of a large number of s trains of the virus. The variations among strains make it difficult to develop a vaccin e or antiviral medication that is universally effective (Wang et al. 1997). Like many other viruses, influenza subverts the hos t cell machinery to complete its life cycle. The exact methods that it uses to r egulate the host cell are not known. Additionally, how the virus regulates the expressio n of its own genes is not completely understood. It is possible that small non-coding re gulatory RNA molecules are involved. Small non-coding regulatory RNA molecules include m icroRNAs (miRNAs), small interfering RNAs (siRNAs) and piwi-interacting RNAs (piRNAs) (Carthew et al. 2009). MiRNAs were originally discovered in C. elegans in 1998 (Fire et al. 1998) and have now been found in the majority of eukaryotes and a number of DNA viruses (Cullen et al. 2009). The small non-coding regulatory RNA molecules encod ed by eukaryotes and DNA viruses regulate gene expression through RNA si lencing. MiRNAs regulate genes
post transcriptionally by directing the silencing c omplex that they are part of to bind to a messenger RNA (mRNA) and prevent the translation of a protein from the mRNA (Carthew et al. 2009). This allows the cell or viru s to selectively express proteins at certain times during its life cycle. It is possible that the small RNA molecule may, after receiving a later signal, release the mRNA and allo w its translation, a mechanism that would prevent waste by allowing the cell to repress and express mRNAs as needed without having to destroy and then synthesize new m RNAs each time (Kedde et al. 2007). One characteristic of the Influenza A virus that ma kes it a likely candidate for encoding miRNAs is the location where replication o f the influenza genome occurs. Unlike the majority of other RNA viruses which repl icate in the cytoplasm of the infected cell, Influenza A replicates in the nucleus of the infected cell. This is also where the initial processing of host cell miRNAs takes place. If the virus is able to use the proteins of the infected cell to generate its own miRNAs, as DNA viruses do, this process would have to take place in the nucleus (Wang et al. 1997 ). Additionally, the Influenza A virus, like other viruses, has a small genome that is used very efficiently. MiRNAs are a way for one section of the genome to serve multiple pur poses; one region can code for a protein and also for a miRNA. Recent findings from a deep sequencing study by Perez et al. (2010) of RNAs from cells infected with influen za A showed small RNAs from the ends of each viral segment at a higher concentratio n than other RNA present. The small RNAs indicate that the Influenza A virus is capable of generating small non-coding RNA molecules, although the method by which it generate s these is currently unknown. These
factors led to the question of whether or not Influ enza A encodes miRNAs. The research of this thesis is focused on this question. 1.1 Gene Expression Overview Gene expression is the term for creating a protein from a gene. To understand how microRNAs (miRNAs) and small interfering RNAs ( siRNAs) act to regulate gene expression, it is necessary to understand the basic s of gene expression. The pathway from gene to protein begins with transcription of a gene in the nucleus of the cell. Transcription creates a messenger RNA (mRNA) that i s then exported from the nucleus and translated into a protein (Cramer et al. 2001). 1.1.1 Transcription Transcription generates a pre-mRNA that is compleme ntary to the template DNA strand and, with the exception of the nucleotide su bstitution of uracil for thymine, has the same sequence as the coding DNA strand. DNA is doub le-stranded, and each strand has a 3 and a 5 end. The 5 end of DNA has a terminal p hosphate group, while the 3 end of DNA has a terminal hydroxyl group. The two strands of DNA run anti-parallel to each other (fig. 1.1). DNA template strands are read fro m the 3 to 5 end which means that mRNAs are synthesized from the 5 to 3 end (Hausne r and Thomm 2001).
Figure 1.1. DNA showing the antiparallel strands, the left str and runs from 5 to 3 while the right strand runs from 3 to 5. The nitrogeno us bases are also indicated. Guanine and Cytosine form three hydrogen bonds while Adenine an d Thymine form two. The alternating phosphate, five carbon sugar backbone i s clearly drawn as well and labeled as the phosphate-deoxyribose backbone. From Wikimedia commons. The process of transcription initiation is very com plicated, so the following description presents a simplified version of the pr ocess. In eukaryotes, transcription initiation of protein coding genes begins when prom oter sequences are recognized in the DNA (fig. 1.2). These promoter sequences occur 30, 75, and 90 nucleotides upstream from the site of initiation (Hausner and Thomm 2001 ). The most well known promoter sequence is the TATA box, 30 nucleotides upstream f rom the initiation site. The TATA box binds TATA binding protein (TBP), which then bi nds transcription factor (TF) IID. In addition to TFIID, TFIIA, TFIIB, TFIIE, TFIIF, a nd TFIIH are, in most cases,
necessary for initiation (Dynlacht et al. 1991). Th ese factors bind, along with RNA Polymerase II (RNAP II), and the pre Figure 1.2. The initiation of transcription begins with TFIID b inding to DNA when a promoter sequence is recognized. After TFIID binds, other transcription factors bind. Finally, RNAP II binds and transcription begins. necessary for initiation (Dynlacht et al. 1991). Th ese factors bind, along with RNA II), and the pre initiation complex (Hirose and Manly 2000). The initiation of transcription begins with TFIID b inding to DNA when a promoter sequence is recognized. After TFIID binds, other transcription factors bind. Finally, RNAP II binds and transcription begins. From Gilbert 2006. n necessary for initiation (Dynlacht et al. 1991). Th ese factors bind, along with RNA initiation complex (Hirose and Manly 2000). The initiation of transcription begins with TFIID b inding to DNA when a promoter sequence is recognized. After TFIID binds, other transcription factors bind.
While there are several RNA polymerases, protein co ding genes are transcribed by RNAP II (Cramer et al. 2001). In eukaryotes, RNA P II has 12 subunits (fig. 1.3). These subunits form a clamp that can swing over to hold the DNA in place, an active site containing two metal ions, and a carboxyl-terminal repeat domain where mRNA may exit the complex (Cramer et al. 2001). When RNAP II bind s the DNA, elongation of the transcript begins. There are several key elongation factors (EFs) that assist with the elongation of the mRNA chain (Conaway and Conaway 1 999). As the mRNAs are transcribed, a 5 cap is added. Following transcrip tion, a 3 polyA tail is added. If these modifications are not completed correctly, the mRNA will not be transported out of the nucleus. A number of RNA binding proteins (RBPs) bi nd to the mRNA forming a ribonucleoparticle (RNP) (Lei and Silver 2002). Whi le the termination of transcription is well understood in prokaryotes, termination in euka ryotes is not and there are several factors involved (Lykke-Andersen and Jensen 2007). Figure 1.3. RNAP II, showing the DNA-RNA hybrid with template D NA in blue and RNA in red. RNAP II surrounds the DNA-RNA hybrid. F rom Brueckner et al. 2009.
To form a functional mRNA, the transcribed pre-mRNA must be spliced. This can occur during or after transcription. The splicing i s performed by spliceosomes that are composed of small nuclear ribonucleoparticles (snRN Ps) and proteins from the serine/arginine-rich protein family. The spliceosom e removes the introns, or noncoding regions, of the pre-mRNA, leaving only exons, or co ding regions, that form the mature mRNA (Hirose and Manly 2000). After the mRNA is spl iced it still has regions on both the 5 and 3 ends that are not translated into a p rotein. These regions are known as untranslated regions (UTRs) (Lei and Silver 2002). Once transcription and processing have finished, t he mRNA must be exported from the nucleus. The mRNA is bound to a number of different proteins that form the RNP, but it is believed that the TAP/NXF family of proteins is the class of transport receptors responsible for exporting mRNA. The mRNA is exported from the nucleus 5 end first, indicating that the 5 cap may have some role in nuclear export (Lei and Silver 2002). If the mRNA is not transcribed or processed correctly, it will not be exported from the nucleus. It will instead remain near the site o f transcription until it is degraded (Brodsky and Silver 2000). 1.1.2 Translation The process of synthesizing a protein from an mRNA is called translation. The mRNA provides the code for amino acids that are lin ked together to form a protein. Three nucleotides form a codon which codes for one amino acid. While each prokaryotic mRNA may code for multiple proteins, each eukaryoti c mRNA typically encodes only one protein. This means that in eukaryotes, transla tion initiation usually begins when the
r 5 cap of the mRNA is recognized by eukaryotic init iation factor 4F (eIF4F). EIF4F is made up of three subunits: eIF4E, eIF4A, and eIF4G. EIF4E binds the cap while eIF4A and eIF4G help by, respectively, unwinding the seco ndary structure in the region where translation will take place and acting as scaffoldi ng for initiation. These proteins bind the 40S, or small ribosomal subunit, in place. The 40S subunit and the associated proteins move down the mRNA from the cap until they reach a start codon. Once at the start codon, a transfer RNA (tRNA) with the amino acid me thionine, which corresponds to the start codon, is brought to the translation site. Th en the 60S, or large ribosomal subunit, interacts with the 40S subunit to form the 80S, com plete, ribosome. Once the 60S subunit is in place, a guanosine-5-triphosphate (GTP) is h ydrolyzed, and the 60S subunit binds to the 40S subunit to form the complete ribosome (L pez-Lastra et al. 2005). Elongation of the protein then begins with tRNAs b ringing in the proper amino acids as specified by the codons of the mRNA. This process continues, with polypeptide bonds forming between subsequent amino acids, until a stop codon is reached. At the stop codon, release factors mediate the termination of t ranslation (Zhouravleva 1995). During translation, the amino acid chain is folded to form the secondary and tertiary structures of the protein. Following the termination of translati on, the quaternary structure of the protein is formed through interactions with other a mino acid chains. Chaperone proteins assist with the steps that lead to a mature protein (Mitchell and Tijan 1989). 1.1.3 Regulation of Gene Expression There are many points along the pathway from DNA t o protein where the process may be regulated. The first point of regulation inv olves access to the DNA. Usually,
DNA is tightly wound around proteins called histone s. Acetylation of histones loosens the wound DNA so that transcription can take place. Opposing acetylation is methylation at specific sites that bind to transcription repres sors that reduce transcription (Allfrey et al. 1964). Transcription factors, coactivators, and corepressors also regulate transcription. The proteins that regulate transcription generally contain a DNA binding domain and a domain that serves to activate or inhibit the trans criptional process (Mitchell and Tijan 1989). The mRNA may be modified to further or halt the pr ocess of gene expression. If the mRNA is not polyadenylated correctly, it will n ot be exported from the nucleus. Similarly, splicing can generate a functional mRNA or destroy a pre-mRNA. Once exported to the cytoplasm, the mRNA can be acted up on by miRNAs or siRNAs. MiRNAs bind to the 3 UTR of a target mRNA and prev ent its translation; siRNAs also prevent translation by binding in the 3 UTR of the target mRNA, but they act to induce cleavage of the mRNA while miRNAs typically do not induce cleavage (Yekta et al. 2004). Neely (2009) estimated that 30% of human gen es are regulated by miRNAs. MiRNAs may be involved in regulatory pathways with transcription factors. If miRNAs and transcription factors share targets, wit h transcription factors increasing the transcription of the gene and miRNAs preventing the translation of the protein, they may fine-tune the expression of that gene. Potential benefits of an miRNA and a TF that target the same gene include temporal regulation of gene expression and the prevention of wasted transcripts (Shalgi et al. 2007). Although this description of gene expression was ve ry brief, the process is complex. The multitude of ways that genes can be re gulated and the ways in which the
regulatory processes affect each other are still be ing investigated. It is clear that small non-coding regulatory RNA molecules such as miRNAs and siRNAs play a large role in the regulation of gene expression. 1.2 Influenza A virus 1.2.1 Basic Information According to the CDC, 36,171 influenza-related deat hs occur each year (Thompson et al. 2003). There are numerous complica tions that can occur in conjunction with or following influenza infections. These compl ications can include acute bronchitis and/or pneumonia. Influenza infections can also exa cerbate pre-existing conditions such as asthma (Kaiser et al. 2003). It is because of th e complications that mortality figures include all influenza-related deaths (Monto et al 2000). Occasionally, a particularly virulent strain of Inf luenza A emerges and causes a pandemic. Throughout history, influenza pandemics h ave occurred every 10-40 years (Table 1.1). Although pandemics earlier than the 19 18 Spanish Influenza are not well documented, it is believed that the 1889 Asiatic Fl u was caused by a strain of Influenza A (Nicholls 2006). Name Date # of deaths Swine Flu 2009 present 14,700 Hong Kong Flu 1968 1969 1 million Asian Flu 1957 1958 1.5 2 million Spanish Flu 1918 1920 50 million Asiatic Flu 1889 1890 1 million Table 1.1. Documented Influenza Pandemics Adapted from Hillem an 2002, Swine Flu information from World Health Organization 2010. The Influenza A virus is part of the Orthomyxoviridae family of viruses that includes Influenzas A, B, and C, as well as Thogotovirus and Isavirus The viruses in this
family are characterized as negative-sense, singlestranded, segmented RNA viruses. Negative-sense viruses have genetic material that i s complementary to mRNAs; mRNAs are defined as positive-sense. Segmented viruses ha ve a number of separate RNA or DNA segments that encode their genetic information (Krossy et al. 1998). Influenza A has eight segments that encode ten diff erent proteins. The seventh and eighth segments code for two proteins each; the mRNAs from these segments are alternatively spliced to form two proteins. In some strains, an eleventh protein, PB1-F2, is spliced from the second segment of the virus (Ch en et al. 2001). At the end of each segment are noncoding regions (NCRs). These regions form secondary structures that are important for the transcription of mRNA and for vir al replication (Perez et al. 2010). The Influenza A virus is pleomorphic which means th at it can exist in two different shapes: a spherical shape with a diameter of 100 nm or a filamentous shape with a length greater than 300 nm. The viral membrane is lipid-based and derived from the membrane of the host cell (Bourmakina and Garcia-Sa stre 2003). The viral particle has distinctive spikes that are 10-14 nm in length and are made up of the hemagglutinin (HA) and neuraminidase (NA) proteins. The variations in the HA and NA give Influenza A viruses their H and N numbers (Laver and Valentine 1969). 1.2.2 The Infection Process The HA protein is essential for binding the virus t o the outside of the host cell. The HA protein binds to acetylneuraminic acids (sia lic acids) on the outer surface of the cell. The HA proteins of viruses are specific to th eir host. Influenza viruses that primarily infect humans have HA proteins specific for acetyln euraminic acid attached to the
penultimate galactose sugar by an 2,6 glycosidic bond, the predominant type of linkag e on the human lung epithelial cells. The HA proteins of the avian flu viruses are specific for acetylneuraminic acid attached to the penultima te galactose sugar by 2,3 glycosidic bonds because in the avian gut, the linkages are pr imarily 2,3. However, both species possess a few cells with the different type glycosi dic bond; this is what allows the virus to spread among species (Glaser et al. 2005). Once the virus is bound to the host cell, the HA pr otein initiates pH-triggered fusion, and the virus enters the cell via clathrinmediated endocytosis (Sieczkarski and Whittaker 2002). The viral particle consists of the membrane coated with HA, NA, and matrix protein 2 (M2) and the core, or ribonucleopr otein (RNP) complex. The RNP is made up of all eight viral RNA segments, polymerase basic protein 1 (PB1), polymerase basic protein 2 (PB2), polymerase acid (PA), nucleo protein (NP) and nuclear export protein (NEP) (Martin and Heleniust 1991). Once inside the cell, viral protein M2 acts as an i on channel for protons that acidify the interior of the virus. This allows the RNP complex to dissociate from the rest of the viral components (Kochendoerfer et al. 1999) Influenza A is unique among RNA viruses because its replication is highly dependent on the nucleus of the host cell. The majority of other RNA viruses replicate in the cyto plasm, but the influenza virus must enter the nucleus to transcribe proteins and replic ate its genome. To enter the nucleus, NP interacts with the cellular protein karyopherinwhich binds to a nuclear localization signal (NLS) and the virus enters the nucleus via a n active transport mechanism (Wang et al. 1997).
Once in the nucleus, the RNA-dependent RNA polymera se (RdRp), composed of PA, PB1, and PB2, initiates transcription of viral mRNA (vRNA) and replication of the viral genome. To replicate the viral genome, copy R NA (cRNA) is synthesized and then transcribed back into vRNA (fig. 1.4) (Poon et al. 1998). Figure 1.4. A schematic showing the replication and transcript ion of negative sense vRNA. To synthesize proteins, the vRNA is transcrib ed into mRNA which is then translated into a protein. To replicate the viral g enome, cRNA is created from the vRNA. The cRNA is positive sense and is used to synthesiz e more vRNA so the virus may spread to other cells. From Racaniello 2009. The transcription of mRNA is dependent on PB1, PB2, PA, and the host cell RNA polymerase II (RNAP II). PB2 and PA steal the 5 ca p and 10-13 nt downstream from the cap from host pre-mRNAs for the viral mRNAs. The nu cleotides attached to the cap are then used as a primer for transcription (Perez et a l. 2010). PB1 catalyzes the addition of nucleotides during RNA chain elongation. The segmen t that is being transcribed forms a loop through the association of the 3 and 5 ends of the segment. The NCR of the 3 end associates with the RdRp and makes the RdRp stutte r as a uracil-rich region is transcribed. This creates the polyA tail of the mRN A (fig. 1.5). Once synthesized, viral
mRNAs from segments that form more than one protein have 5 and 3 splice sites that have the same consensus sequence as cellular exons and introns so the host cell splicing machinery can be used to splice the viral mRNAs (Po on et al. 1998). Figure 1.5. The transcription of viral mRNAs. Transcription is initiated by a 5 bound polymerase (P) that cannot transcribe the site to w hich it is attached so copies the uracil sequence multiple times. The transcribed mRNA is sh own as a dashed line. From Poon et al. 1998. To create more vRNAs, the virus must first synthesi ze cRNAs which can then be used to synthesize vRNAs. The synthesis of cRNAs fr om vRNAs and vRNAs from cRNAs are primer-independent processes. During cRNA synthesis, the secondary structure of the 3 NCR must not associate with the RdRp or a polyA tail will be transcribed. This lack of association and subsequen t failure to transcribe the polyA tail are what distinguish viral replication from viral t ranscription (Bui et al. 1999). The switch from viral transcription to viral replic ation occurs later in the infection process when the virus needs to synthesize sufficie nt genetic material to spread to other cells. The signal that causes the switch may be cau sed by changes in cRNA stability, changes to intracellular levels of nucleotides, cha nges in the soluble fractions of NP and
n the RdRp, small viral RNAs (svRNAs), or a combinati on of these factors (Perez et al. 2010). 1.2.3 Regulation of Gene Expression Although the virus has multiple ways of regulating its gene expression as evidenced by the preferential expression of certain proteins at different times after infection, the methods of control are not clearly u nderstood. Early in an infection, NP and non-structural protein 1 (NS1) are the most abundan tly synthesized proteins, while later HA, NA, and matrix protein 1 (M1) are more abundant (Shapiro et al. 1987). NS1 binds to the polyA tail of cellular and viral mRNAs and p revents the movement of mRNAs from the nucleus to the cytoplasm. This is one way of regulating gene expression by preventing the translation of proteins at certain p oints during infection (Qiu and Krug 1994). Additionally, M1 has been found to inhibit v iral transcription and play a role in slowing down transcription late in infection (Shapi ro et al. 1987). Another method of control may stem from the promote r at the 3 end of the viral RNA segments. This promoter is conserved in each se gment. However, there is a variation in the position 4 nt from the 3 end. PB1 PB2, and PA have a cytosine in that position while all other viral RNA segments have a uracil in that position. This difference may affect the transcription of the viral RNA molec ules. The virus also subverts the hostcell translation machinery so that viral mRNAs are preferentially translated before cellular mRNAs (Parvin et al. 1989). In addition to subverting the host cell machinery t o synthesize proteins and replicate the viral genome, the virus takes control of many other cellular processes. The
NS1 protein is responsible for inhibiting the inter feron immune response to the viral infection. M1 and M2 are involved in inducing apopt osis of the host cell. PB1-F2, the 11th protein that is encoded by some strains of the vir us, induces leukocyte apoptosis (Chen et al. 2001). Recent work by Perez et al. (2010) showed that Infl uenza A viruses generate small viral RNAs (svRNAs) of approximately 22-27 nt in length; these svRNAs are a key component of the switch from transcription to repli cation. The svRNAs were found via high-throughput sequencing of RNAs isolated from in fected cells. The RNA fraction that was sequenced contained RNAs 10-40 nt in length. Th e majority of the small RNAs that were sequenced were cellular miRNAs, but 0.12% were derived from the virus. Most of that 0.12% was composed of viral breakdown products However, 0.04% of the total number of reads was enriched for the 5 end of the viral segments. These were named svRNAs. The svRNAs associate with the RdRp and are hypothesized to block the secondary structure of the NCR of the 3 end during replication (fig. 1.5), thereby allowing replication, rather than transcription, to occur (Perez et al. 2010). These svRNAs are not miRNAs and do not have a simil ar function. Their existence does indicate that the Influenza A virus is capable of generating small RNAs. Additionally, these svRNAs are conserved in H1N1, H 3N2, and H5N1 viruses and are generated during infection in human, canine, and mu rine cells. This conservation across both species and viral strains indicates that certa in regions of the Influenza A virus cannot mutate rapidly without negatively affecting the vir al life cycle (Perez et al. 2010).
1.2.4 Spread of Viral Particles Once sufficient replication has occurred, the virus exits the infected cell and spreads to other cells. First, the virus must assem ble RNPs that can be released from the cell. Viral protein M1 is essential for the movemen t of RNPs from the nucleus, where they are assembled, to the cytoplasm (Bui et al. 19 99). To exit the cell, the virus forms budding particles from the cellular membrane. These particles are pinched off and infect new cells (Gmez-Puertas et al. 2000). 1.2.5 Mutations of the Influenza A Virus The influenza virus, like other RNA viruses, mutate s rapidly. One form of mutation is known as antigenic drift. Antigenic dri ft is caused by point mutations that lead to gradual changes of the virus. In contrast t o antigenic drift is antigenic shift which involves mutations major enough to change the HA or NA subtype. Antigenic shift is the reason that the 1957 Asian Flu, an H2N2 strain, cau sed so many deaths. The H2N2 strain had not circulated through the population in the ye ars preceding the pandemic, so people had no adaptive immunity to the H2N2 virus (Treanor 2004). The rapid mutation of the virus makes it difficult to develop effective vaccines. However, in July 2010, the NIH released information regarding the development of a universal influenza vaccine that is effective again st all strains of the Influenza A virus. The vaccine, which has now been tested in rats and non-human primates, stimulates the creation of antibodies against the stem of the HA p rotein. Previous vaccines have focused on the head of the protein which mutates at a much higher rate than the stem (Wei et al. 2010).
r The influenza infection process in human cells is s till being studied and more is being learned about it each year. The possibility t hat the influenza virus encodes miRNAs has not been ruled out and is possible because the virus replicates in the nucleus and is capable of generating small RNAs, as discussed earl ier (Perez et al. 2010). Further insights into the virus will improve the treatment and prevention of the disease and help distinguish features that could aid in the predicti on of pandemics. 1.3 Small Regulatory RNA Molecules Small non-coding regulatory RNA molecules are impor tant in many cellular pathways (table 1.2). There are signature expressio n profiles of miRNAs in certain cancers, and they are implicated in a number of dis eases (Calin and Croce 2006). Class Processed by Precursor Complex Targets Size (nt) Protein Association miRNA Drosha and Dicer 60-80 nt hairpins RISC or RITS mRNA or chromatin 21-24 Ago subfamily siRNA Dicer ds RNA RISC or RITS mRNA, chromatin, transposable elements, or viruses 21-24 Ago subfamily piRNA Piwi transcripts of retrotransposons piRNA complex transposable elements 24-31 Piwi subfamily Table 1.2. The main classes of small regulatory RNA molecules including their targets and proteins that they associate with. MiRNAs and s iRNAs have more in common with each other than with piRNAs, including how they are generated, their targets, and the proteins they associate with to form an active comp lex. SiRNA and miRNA information from Lau et al. 2009, piRNA information from Vagin et al. 2006. 1.3.1 MicroRNAs MicroRNAs (miRNAs) are transcribed in the nucleus b y RNA polymerase II (RNAP II) or RNA polymerase III (RNAP III) (Winter et al. 2009). The DNA regions
that encode miRNAs are usually located between clus ters of genes, or intergenic regions (Lee et al. 2004). About 25% are located in introns of mRNAs. These miRNAs have the same orientation as pre-mRNAs, indicating that they are spliced directly from the introns (Gregory et al. 2004). The transcription of the reg ion encoding an miRNA generates a primary miRNA (pri-miRNA). Sometimes multiple miRNA s are located in the same intergenic region. The transcription of these regio ns generates a pri-miRNA that includes several mature miRNAs (fig. 1.6) (Fazi and Nervi 20 08). It was originally believed that RNAP III transcribed pri-miRNAs because it is respo nsible for transcribing other small RNA molecules such as tRNAs and U6 snRNAs (Lee et a l. 2004). Figure 1.6. The transcription of miRNAs from various regions. M iRNAs may be polycistronic, which means that one transcribed pri -miRNA yields two or more mature miRNAs as shown on the left side of the figure, int ergenic as shown in the middle, or intronic/exonic as shown on the right side with exo ns as thick black bars. A transcription factor (TF) is shown regulating the transcription o f the polycistronic miRNAs and the miRNA from the intergenic region. TFs are not neces sary for transcription of miRNAs, but they can regulate the transcription process. Fr om Fazi and Nervi 2008.
Work by Lee et al. (2004) showed that the majority of pri-miRNAs have a 5 7methyl guanosine cap and a 3 poly(A) tail. These s tructures indicate processing by RNAP II. The work also showed a physical associatio n of RNAP II with an miRNA promoter. However, this work did not rule out trans cription of miRNAs by other RNA polymerases. Borchert et al. (2006) showed by chromatin immunopr ecipitation and cell-free transcription assays that RNAP III is sufficient fo r transcription of miRNAs from the human chromosome 19 miRNA cluster (C19MC). C19MC ha s a number of Alu repeats that signal transcription by RNAP III. Alu repeats are short interspersed elements within the human genome that do not code for proteins. The y are approximately 300 nt long and were originally characterized as Alu regions becaus e they are cut by the restriction endonuclease Alu I. Alu repeats are transcribed onl y by RNAP III. This work indicates that a significant percentage of miRNAs may be tran scribed by RNAP III rather than RNAP II (Borcert et al. 2006). The transcription process generates a transcript of approximately 1,000 nt (Murphy et al. 2008) that is spliced (Lee et al. 20 04) and folds to form a pri-miRNA with a characteristic hairpin structure (fig. 1.6). The hairpin structure contains a terminal loop and a region of intrastrand base pairs that form th e helical A-form of double-stranded RNA. Hairpins with loops larger than 10 nt are pref erentially recognized by the RNase III Drosha (RNASEN) (Zeng et al. 2005). Drosha forms tw o complexes in the nucleus. One complex (600 kDa) associates with 20 different prot eins with domains such as the DEAH-box family of RNA helicases, the double-strand ed RNA binding domain, the RNA recognition motif, and the zinc-finger domain. The other complex, the
microprocessor complex, is smaller (400 kDa) and is made up of Drosha and Pasha (DGCR8). While both complexes associate with pri-mi RNAs, the smaller complex is the main processor of pri-miRNAs, responsible for proce ssing eight times more pri-miRNAs than the larger complex (Gregory et al. 2004). Droshas cofactor, Pasha (DGCR8) contains two doubl e-stranded RNA-binding domains and a WW domain. The WW domain interacts wi th proline-rich proteins like those present in the N-terminal domain of Drosha. P asha increases the specificity of the complex and ensures that the complex only associate s with pri-miRNAs, not with other forms of double-stranded RNAs (Gregory et al. 2004) Drosha and Pasha cleave the primiRNA to form a 60-80 nt hairpin, the pre-miRNA. Th e pri-miRNA is cleaved at a site two turns of the RNA helix away from the terminal l oop (Zeng et al. 2005). The premiRNA has a 2 nt overhang at the 3 end (Zhang et a l. 2002) If the pre-miRNA is correctly formed, Exportin-5 (E xp5) will move the premiRNA from the nucleus to the cytoplasm. The moveme nt of the pre-miRNA is greatly decreased when RanGTP is inhibited, indicating a de pendence on RanGTP (Lund et al. 2004). Once in the cytoplasm, the pre-miRNA hairpin is recognized by the 220 kDa RNase III Dicer (Lima et al. 2009). Dicer recognize s the 2 nt overhangs on the 3 end of the pre-miRNA and cleaves the pre-miRNA approximate ly 22 nt away from the end (fig. 1.8). This generates a double-stranded duplex. The duplex has the same characteristic 2 nt overhang at both 3 ends (Zhang et al. 2004). Dicer has a double-stranded RNA binding domain that allows it to bind to the premiRNA and to assist in the incorporation of the miR NA or siRNA into the RNA induced silencing complex (RISC) (fig. 1.7). Dicer also pos sesses a PAZ domain similar to the
one present in Argonaute 2, which allows the protei ns to bind single-stranded nucleic acids. There are also two RNase III domains and a d omain homologous to ATPdependent RNA helicase. The separation between the PAZ and RNase III domains is approximately 22 nt long and accounts for the abili ty of Dicer to cleave pre-miRNA hairpins to form approximately 22 nt long miRNA dup lexes (Sashital and Doudna 2010). Figure 1.7. Processing pathways of a miRNA and a siRNA followin g transcription. As shown, pri-miRNAs are cleaved by the microprocessor complex and then cleaved again by Dicer before associating with the Argonaute prot ein. By contrast, siRNAs are not processed by the microprocessor complex. Similarly to miRNAs, they are cleaved by Dicer and then associate with the Argonaute protein The blue strand represents the guide strand while the green strand is the strand that is degraded in the cytoplasm. The characteristic 3 overhangs can be seen during proc essing by Dicer. Once incorporated into the RNA-induced silencing complex (RISC), miRN As and siRNAs silence genes by binding in the 3 UTR of the mRNA and repressing tr anslation or cleaving mRNAs. Adapted from Sashital et al. 2009.
Dicer assists the RNA-induced silencing complex (RI SC) in preferentially incorporating the strand that has weaker hydrogen b onding at its 5 end. The other strand, miRNA*, is rapidly degraded in the cytoplasm (Khvor ova et al. 2003). Transactivating response binding protein (TRBP) is also important i n the incorporation of the miRNA into the RISC. Together, Dicer and TRBP make up the RISC loading complex (RLC) (Murphy et al. 2008). The main protein of the RISC, regardless of whether the complex involves an miRNA or siRNA, is from the argonaute family of pro teins. Members of the argonaute protein family are characterized by having PAZ and PIWI domains (figs. 1.8 and 1.9) (Miyoshi et al. 2005). The PAZ domain is also found in proteins of the Dicer family (Fig. 1.8). The PAZ domain is composed of two subdomains one with an open -barrel with two helices at one end of the barrel, the other wit h a -hairpin and an -helix. The PAZ domain forms a fold similar to the OB-Fold that is known to bind single-stranded nucleic acids. This fold in the PAZ domain likely helps pro teins in the Dicer and Argonaute family bind single-stranded nucleic acids. The doma in interacts with the 3 end of the nucleic acids. This makes it ideal for interacting with the 3 overhangs of siRNAs and miRNAs (Song et al. 2003). Figure 1.8. A ribbon representation of the PAZ domain of Ago2 f rom a crystal structure of the Ago2 protein of Drosophila melanogaster -strands are purple, -helices are blue. From Song et al. 2003.
Figure 1.9. A ribbon representation of the Ago2 protein of Pyrococcus furiosus showing how the lower PIWI domain helps hold the mRNA in pl ace as the siRNA interacts with it. From Hutvagner and Simard 2008. In humans, argonaute protein 2 (Ago2) is responsibl e for the cleavage of mRNAs when there is perfect complementarity between the m iRNA and the mRNA (Liu et al. 2004 and Meister et al. 2004). In Drosophila both Ago2 and argonaute protein 1 (Ago1) possess slicer activity that allows them to cleave target mRNAs (Miyoshi et al. 2005). MiRNAs may induce the cleavage of the target mRNAs to decrease mRNA levels or may simply act to decrease translational efficiency. Gu o et al. (2010) showed that the main method of action (84%) to decrease the level of tar get mRNAs is to decrease the stability of the mRNAs, leading to their degradation. In anim als, cleavage of the target mRNA is not as common as in plants because the level of com plementarity required for cleavage is rarely present in animal mRNA targets. The RISC ins tead acts to decrease mRNA levels by destabilizing the target mRNAs by causing deaden ylation. The mRNA deadenylation leads to the de-capping and more rapid degradation of the mRNA through the normal processes (Guo et al. 2010).
n 1.3.2 Small Interfering RNAs While miRNAs come primarily from the genome, siRNAs have many sources. They can be processed from double-stranded RNA that is introduced from some external source or picked up from the environment. They can also be transcribed from centromeres, transposons (Lippman and Martienssen 2 004), or genomic transcripts (Vasquez et al. 2004). The distinction between endo genous and exogenous siRNAs can be made only by looking at the precursors. Endogeno us siRNAs will have a nuclear phase that exogenous siRNAs lack because they are p rocessed only in the cytoplasm (Lippman and Martienssen 2004). Once pre-siRNAs are generated, the processing is ve ry similar to that of premiRNAs. If they originate in the nucleus, they are exported to the cytoplasm where they are processed by Dicer. In the cytoplasm, ATP is us ed as an energy source to first unwind the siRNA duplex and then incorporate one strand in to the RISC (Miyoshi et al. 2005). Like miRNAs, the strand with the weaker hydrogen bo nding at the 5 end is preferentially incorporated into the RISC while the other strand is rapidly degraded in the cytoplasm (Khvorova et al. 2003). Unlike the majori ty of miRNAs, most siRNAs are perfectly complementary to the 3 UTR of their targ et mRNAs. This perfect complementarity induces the argonaute protein to cl eave the target mRNA (Steiner et al. 2009). Although differences in complementarity may exist b etween miRNAs and siRNAs and their targets, the distinguishing factor between the two is the precursor. MiRNAs come from stem-loop structures; siRNAs come from double-stranded sections of RNA that lack a loop structure at the end (Morri s et al. 2004).
Along with post-transcriptional gene regulation, si RNAs can regulate transcription of certain genes by acting as part of an RNA-induced initiation of transcriptional gene silencing (RITS) complex. The RITS complex regulates transcription by methylating the DNA and by inducing histone meth ylation. The complex contains the siRNA, an argonaute protein, and a heterochromatinassociated chromodomain protein (Verdel et al. 2004). MiRNAs, siRNAs, and piRNAs act in a number of diffe rent ways to regulate gene expression. The entire field of small non-coding re gulatory RNAs is very new and discoveries of their effects in the cell are even m ore recent. There may be methods of action that have still not been discovered, and the molecules may be more abundant than originally believed. These molecules have been foun d encoded by DNA viruses in the past ten years, and it is possible that they could be encoded by RNA viruses as well. This thesis will explore that possibility. 1.4 MicroRNAs and Viruses 1.4.1 MicroRNAs Encoded by DNA Viruses Virally encoded miRNAs were first found in the geno mes of viruses from the herpesvirus family. These viruses are double-strand ed DNA viruses with a life cycle that includes a latent stage during which the virus is n ot actively replicating and is not detectable by the immune system. This family includ es the Herpes simplex virus 1 (HSV1), Herpes simplex virus 2 (HSV-2), Epstein-Barr vi rus (EBV), Varicella zoster virus (VZV), Cytomegalovirus (CMV), Roseolovirus, and Kap osis sarcoma-associated herpesvirus (KSHV) (Bai et al. 2008), all of which encode miRNAs (Cullen 2009).
In addition to miRNAs encoded by viruses from the h erpesvirus family, miRNAs encoded by other DNA viruses have been identified. Primate polyoma viruses and human adenoviruses encode miRNAs, although far fewer miRN As encoded by these viruses have been identified (Seo et al. 2008). It is possi ble that viruses in the herpesvirus family encode more miRNAs in order to regulate their compl ex life cycle which includes a latent phase during which no active replication is occurri ng, and a lytic phase during which the virus is actively replicating (Cullen et al. 2009). In fact, some miRNAs encoded by the herpesviruses are involved in regulating the latent and lytic phases of the viral life cycle (Bai et al. 2008). However, it is also possible th at the large number of miRNAs encoded by the herpesviruses have been found because a disp roportionate amount of research has been focused on the herpesvirus family of viruses b ecause they were the first viruses found to encode miRNAs (Cullen et al. 2009). 1.4.2 Conservation of microRNAs in Viruses While conservation of miRNAs among several differen t viruses is not a typical technique for predicting viral miRNAs, it was succe ssful in predicting the existence of HSV-1 miR-3. This viral miRNA was predicted as a ho molog of HSV-2 miR-1. Both miRNAs lie within an exon of the latency-associated transcript (Bai et al. 2008). Conservation criteria were also used to identify a number of miRNAs encoded by murine cytomegalovirus (MCMV) after finding miRNAs encoded by human cytomegalovirus (HCMV) (Buck et al. 2007). These results show that, while not applicable to all viruses, it is possible to use conservation criteria to iden tify miRNAs encoded by viruses. It
r should be emphasized that HSV and CMV are DNA virus es, which, as discussed earlier, means that their mutation rate is much lower than t hat of RNA viruses. This may have implications for using conservation techniques with RNA viruses. 1.4.3 Roles of Virally Encoded microRNAs Virally encoded miRNAs serve a number of purposes. MiRNAs encoded by herpesviruses target cellular genes involved in the immune response (Bai et al. 2008). KSHV encodes a miRNA that is a homolog of cellular miR-155. This viral miRNA regulates the same genes that miR-155 does, genes i nvolved in cell growth and genes involved in the B cell life cycle. This homology in dicates that the viral miRNA evolved to exploit the already existing cellular pathways ( Gottwein et al. 2007). EBV encodes a miRNA that targets an apoptosis mediator in order t o prevent apoptosis of infected cells. Apoptosis of infected cells is a method frequently used by the immune system to stop the virus from using the cell for replication. The EBV miRNA prevents apoptosis in the cell until the virus has finished replicating and is rea dy to release from the cell and infect other cells (Choy et al. 2008). Other virally encoded miRNAs target viral mRNAs. Mi RNAs encoded by HSV-1 during the latent, inactive, phase of the viral lif e cycle target mRNAs of genes necessary for the virus to begin actively replicating. These miRNAs are partially responsible for initiating and maintaining the latent phase (Umbach et al. 2008).
1.4.4 Influenza A and MicroRNAs While it has not been shown that Influenza A encode s functional miRNAs, the infected host cell generates miRNAs that act to rep ress the expression of viral proteins. Song et al. (2010) showed that cellular miRNAs act on the H1N1 Influenza A virus to degrade the mRNAs transcribed from the PB1 gene. Th e three miRNAs that act on the PB1 mRNAs also have roles in regulating the express ion of cellular genes (Song et al. 2010). This regulation of viral gene expression by cellular miRNAs also occurs during vesicular stomatitis virus (VSV) infection (Otsuka et al. 2007) and primate foamy virus type 1 (PFV-1) infection (Lecellier et al. 2005). The discussion above is not inclusive of all virall y encoded miRNAs and their roles in cells. There are multiple roles for virall y encoded miRNAs and there are cellular miRNAs that target viral mRNAs as noted above. It c annot be conclusively decided that Influenza A does not encode miRNAs; the research of this thesis is focused on the possibility that Influenza A does in fact encode mi RNAs that may serve regulatory roles similar to those discussed above. 1.5 Methods for Identifying MicroRNAs 1.5.1 Computational Methods There are a number of programs that are used to pre dict pre-miRNA hairpins and mature miRNAs. These programs usually predict pre-m iRNAs and mature miRNAs based on several criteria. Research has demonstrate d that pre-miRNAs have a folding free energy that is lower than that of tRNAs or oth er RNAs. In addition, pre-miRNAs
always fold to form a hairpin structure. The third commonly used criterion is the conservation of the predicted miRNA across several species. Homology based searches use conservation criteria and the location and sequ ence of a known miRNA to predict the location and sequence of a miRNA in a different spe cies (Zhang et al. 2006). While there are a number of programs designed to pr edict novel miRNAs, most are at least partially based on the conservation of miRNAs among species. This approach will not work in most viruses because most viruses are only very distantly related and have evolved to be highly specific to their host an d a certain cell type within their host. Also, because viruses use their genetic material as efficiently as possible, viral miRNAs may overlap with open reading frames (ORFs) and the refore be overlooked by conservation criteria. With these ideas in mind, VM ir was developed (Grundhoff et al. 2006). VMir predicts pre-miRNA hairpins within the viral genome. VMir has low stringency criteria and therefore generates a numbe r of false-positives. Because the viral genomes are so small in comparison to the human gen ome, the false-positives can be eliminated by testing for the presence of the predi cted hairpins using high-throughput selection methods like polymerase chain reaction (P CR) (Grundhoff et al. 2006). The VMir program was used in this project to search for possible pre-miRNA hairpins in the Influenza A genome. VMir predicts pre-miRNA hairpins by sliding a wind ow of 500 nt across the input sequence and calculating the minimum free energy of folding for possible hairpins within that window. The predicted hairpins are compared to a reference set of known premiRNA hairpins and scored accordingly. In a diagnos tic trial, a score cutoff of 115 included 95% of a set of known pre-miRNAs. A window score is also generated based on
the number of windows that the hairpin folds in. Th e window is slid 10 nt at a time across the genome, each window is 500 nt, and pre-miRNA ha irpins are 60-80 nt. Therefore a stable hairpin would be expected to fold in more th an one window. Depending on the size of the viral genome, Sullivan and Grundhoff (2007) suggest a window count of 10 (least stringent) to 35 (most stringent). 1.5.2 Experimental Verification of Computationally Predicted MicroRNAs 22.214.171.124 Northern Blotting Northern blotting is a commonly used method to veri fy the existence of computationally predicted miRNAs. First, purified RNAs must be separated by size; thi s is commonly done using a gel. The RNAs are then tra nsferred to a membrane. The membrane is typically nylon with a positive charge that causes the negatively charged RNAs to stick to the membrane. Then, a radioactivel y labeled probe that can anneal to the RNA of interest is applied to the membrane. The membrane is washed to ensure that there is no nonspecific binding of the probe and RN A and then examined. Any RNA that is complementary to the probe will be detected (Gra d et al. 2003). This technique has been used to verify the existence of many miRNAs an d it allows for the quantitation of expression levels of the miRNA and size determinati on of the miRNA, but it has a low sensitivity (Vloczi et al. 2004). 126.96.36.199 Locked Nucleic Acids to Increase the Sensit ivity of Northern Blotting Locked nucleic acids (LNAs) are nucleic acids that have been modified to include a methylene bridge between the 2 oxygen and 4 car bon (fig. 1.10). This bond increases
the specificity of the nucleic acid, meaning that t he nucleic acid will only hydrogen bond to one specific nucleotide (adenine with uracil, gu anine with cytosine) rather than engaging in the nonspecific hydrogen bonding that c an occur between guanine and uracil (Vloczi et al. 2004). Figure 1.10. A locked nucleic acid, note the bond between the 2 oxygen and the 4 carbon. This bond locks the nucleic acid in the 3 endo conformation and increases the stability and specificity of the nucleic acid. From exiqon.com. Northern Blotting is used as discussed above to se arch for predicted miRNAs. However, every third nucleotide in the probe is a L NA. This technique has the advantage of higher specificity and sensitivity in addition t o the ability to quantitate the expression levels of the miRNAs (Vloczi et al. 2004). 188.8.131.52 Microarrays Microarrays provide a method that has a sensitivity comparable to Northern Blotting, but is much more time efficient when exam ining the expression of a large number of miRNAs. Microarrays are specific enough t o distinguish between primary and pre-miRNAs and mature miRNAs. A total RNA sample is covalently labeled with fluorophores and then hybridized to a membrane of a chemical matrix attached to a glass or silicon chip containing probes for the miRNAs of interest. Mature miRNAs hybridize strongly to the probes on the membrane, but precurs or miRNAs do not because they are
much longer. While the region that hybridizes to th e probe is the same length, the precursor miRNAs have extra nucleotides that are no t hybridized to the probe and so the precursor miRNAs are removed by a washing step afte r the hybridization step. When this process is repeated with RNA from multiple time poi nts following infection or other stimulus, a temporal expression pattern can be obta ined for the miRNAs of interest (Babak et al. 2004). 184.108.40.206 High-throughput Sequencing High-throughput sequencing is a technique that can be applied to quantify and determine the locations of any small RNAs present i n the given RNA sample. To find miRNAs, the small RNA fraction is isolated. Then, a library of possible RNAs is created. The RNAs are reverse transcribed and compared to th e library. The relative quantities of the RNAs, the lengths, and the locations of the RNA s on the genome can then be determined (Hall 2007). Although high-throughput sequencing can be very sen sitive and may eliminate the problem of missing miRNAs present at very low conce ntrations, it is quite expensive. The technique, with few exceptions (Ozsolak et al. 2010), involves the reverse transcription of RNA to cDNA and then the amplifica tion of the cDNA. Each reverse transcription and amplification technique has the p ropensity to favor a certain secondary structure, or lack of secondary structure, introduc ing bias into the sequencing results (Dohm et al. 2008).
220.127.116.11 Polymerase Chain Reaction PCR is a way of amplifying and determining the pres ence of specific DNA molecules (fig. 1.11). If it is the RNA that is of interest, it must first be converted to cDNA by reverse transcription. The sequence of the oligonucleotide of interest must be known. Then, a forward primer complementary to the beginning of the sequence of interest and a reverse primer complementary to the end of the sequence of interest are designed. The DNA is mixed with the forward and re verse primer, deoxynucleotide triphosphates (dNTPs), Taq Polymerase (or another thermostable polymerase with a temperature optimum of approximately 70 oC), and a buffer solution. This solution is heated which causes the double-stranded DNA to dena ture. As the reaction is cooled, the primers anneal to the DNA; because the primer conce ntration is much higher than the DNA concentration the primers anneal rather than th e two strands of DNA reannealing. The polymerase and the dNTPS polymerize the section of DNA of interest from the start of the primer. The cycle of heating and cooling is then repeated and the amount of DNA increases exponentially (Pavlov et al. 2004). With the exception of a platform developed by Helico, all high-throughput sequencing technique s use a PCR amplification step that can introduce a bias into the sequencing results (K apranov et al. 2010).
n Figure 1.11. The PCR process, including the dissociation, anneal ing, and extension steps. The sequence of interest is amplified exponentially From scienceblogs.com 18.104.22.168 Quantitative Polymerase Chain Reaction Quantitative PCR (qPCR) is a way of quantitating th e amount of starting DNA. There are two methods for this. One method uses pri mers labeled with a probe that fluoresces when it binds to the target DNA. The flu orescence is measured after each cycle of the reaction and compared to a standard cu rve. The other method uses SYBR Green which fluoresces in the presence of any doubl e-stranded DNA. This method is prone to giving false-positives because if the prim ers anneal to each other and a primerdimer is formed, SYBR Green will fluoresce. However SYBR Green is much less
expensive than fluorescently labeled probes and pri mer-dimers can be detected by examining a dissociation curve. The dissociation cu rve measures the temperature at which the double-stranded DNA dissociates. This tem perature is lower for primer-dimers than it is for longer segments of double-stranded D NA (Heid et al. 1996). 22.214.171.124 Stem-Loop Reverse Transcriptase quantitativ e PCR Stem-loop reverse transcriptase qPCR has the advan tages of sensitivity and low cost, but has not been used frequently because it i s a relatively new technique. The sequence of the miRNA or predicted miRNA of interes t must be known. A stem-loop primer with a sequence complementary to the end of the miRNA is designed. The stemloop hybridizes onto the end of the miRNA and the m iRNA is reverse transcribed by a reverse transcriptase (fig. 1.12). Then qPCR is per formed in the usual manner. Figure 1.12. A schematic illustrating the hybridization of the stem-loop primer to the miRNA, the extension of the strand complementary to the miRNA to create cDNA, and the location of the PCR forward and reverse primers that would function to quantitate the miRNA. From Chen et al. (2005).
The stem-loop reverse transcriptase primers provide greater sensitivity than conventional primers because base stacking of the s tem-loop structure improves the thermal stability. Also, the shape of the stem-loop primer prevents it from binding to double-stranded genomic DNA. The stem-loop primer h ybridizes with primary or premiRNA at a much lower percentage than it does with the mature miRNA because of the structure of the primer (Chen et al. 2005). 126.96.36.199 Primer Extension QPCR The primer extension method of reverse transcribin g miRNAs into cDNA that can be used for qPCR is similar to the stem-loop method However, in this method, the reverse transcriptase primer is a linear primer and therefore may hybridize to primary miRNAs or pre-miRNAs (fig. 1.13). This would lead t o the levels of miRNAs being recorded as higher than they actually are (Raymond et al. 2005). This method was used for the reverse transcription step of this thesis. It did not provide adequate sensitivity to quantitate levels of miRNAs. As will be discussed l ater, future research might utilize the stem-loop real time primer technique discussed in s ection 188.8.131.52.
r Figure 1.13. The Primer Extension method for reverse transcribin g miRNAs to create cDNA. The forward primer for qPCR is complementary to the miRNA while the reverse primer is complementary to the tail that was added to the miRNA. Adapted from systembio.com.
Chapter 2: Materials and Methods 2.1 Overview This section will provide a brief overview of the e xperimental procedures (fig. 2.1). The entire procedure is described in more det ail below. The goal of this project was to find small RNAs, possible miRNAs, encoded by the Influenza A virus. First, premiRNA hairpins were predicted computationally and m ature miRNAs were predicted from those hairpins. Then human lung epithelial cel ls were infected with the Influenza A virus and RNA was harvested. This RNA was reverse t ranscribed and used to run quantitative PCR (qPCR) to look for and quantitate any pre-miRNA hairpins or small RNA molecules encoded by the Influenza A virus that could be possible miRNAs. Then the size of the small RNA molecules was determined using DNA agarose gel electrophoresis, and the qPCR products were sequenc ed to determine whether they were qPCR products of possible virally encoded miRNAs or of cellular miRNAs.
Figure 2.1. A flowchart describing the process for finding smal l, virally encoded RNAs !"# # $% !" $ !"% !" !" () *%)( !"# !"%# !"#'+) %,!"% !"() *% )(( !" -,!".#% ) (/ )() (0 !"1
2.1.1 Selection of Influenza Strain The research was conducted using one strain of Infl uenza: A/California/04/09(H1N1). This pandemic influenza s train was isolated form an initial index case during the pandemic outbreak of the swin e-origin influenza strain in 2009. This virus strain continues to circulate today, and is representative of the predominant H1N1 virus that circulates globarlly (Nelson et al. 2008). Because it is an H1N1 strain, rather than a highly pathogenic H5N1 strain, it can be worked with in biological safety level (BSL) 2 labs, rather than requiring that the work be conducted in the BSL3 lab. 2.1.2 Use of VMir for Pre-miRNA Hairpin Prediction VMir, a program that has successfully predicted the hairpins formed by premiRNAs in DNA viruses, was used to predict pre-miRN A hairpins in the Influenza A virus. The program uses thermodynamic calculations to predict stable hairpins that could be processed by Dicer into functional miRNAs (Grund hoff et al. 2006). The program gives a graphic output of the sequence of the predicted pre-miRNA hairpin, any subhairpins (predicted pre-miRNA hairp ins that are shorter than the main hairpin), and the regions of the base pairing withi n the hairpin (fig. 2.2).
Figure 2.2. A view of the output of VMir. The location of the h airpin within the input segment is indicated by the red line and the blue b ar on the right hand side. The sequence of the hairpin and the exact location of the hairpi n are given. The subhairpins are also shown, and labeled by the name of the hairpin_S1 et c. The parentheses underneath of the sequence indicate regions of base pairing between t he two arms of the predicted hairpin. The open parentheses indicate the 5 arm of the hai rpin and the close parentheses indicate the 3 arm of the hairpin. A larger version of the output is shown below (fig. 2.3). 2.1.3 Selection of Hairpins The sequences of each segment of California04 were loaded into VMir v1.5. VMir was used to predict hairpin loop structures, p ossible pre-miRNAs, in each segment.
As discussed earlier, VMir is designed for viral ge nomes which are much smaller than typical eukaryotic genomes (Grundhoff et al. 2006). A minimum hairpin score of 115 and a minimum window score of 15 were required for a ha irpin to be considered. California04 only had four predicted hairpins (see A.1 for sequences ) so it was feasible to generate primers to amplify all of the potential ha irpins with qPCR (fig. 2.3). Figure 2.3. A hairpin predicted by Vmir is shown. This particul ar hairpin has three possible sub hairpins. The blue bars above the nucl eotide sequence show the primer design. Each blue bar indicates a forward primer th at would amplify that specific sequence if it is present in the small RNA fraction The red boxes indicate the typical location of mature miRNAs within the pre-miRNA hair pin. Special attention was paid to design a primer that covered those locations in eac h hairpin.
2.2 Preparation of Viral RNA 2.2.1 Cell Growth and Infection Wells of immortalized adenocarcinomic human alveola r basal epithelial (A549) cells were plated from cells that had been growing in Opti-mem1 Reduced Serum Medium 1X with HEPES, 2.4 g/L Sodium Bicarbonate, a nd L-Glutamine (Invitrogen Carlsbad, CA), 1X Antibiotic-Antimycotic (Invitroge n Carlsbad, CA), and 10% Fetal Bovine Serum (FBS) (Invitrogen Carlsbad, CA) (th e combination of these components is called growth media) at 37 oC under 5% CO2 in 225 cm2 vented cell culture flasks (Corning Union City, CA). To plat e the cells, the media was aspirated and the cells were rinsed with approximately 10 ml 1X Phosphate Buffered Saline (PBS, Invitrogen Carlsbad, CA). The PBS was aspirated and 4 ml of Trypsin 0.25% (1X) with Ethylenediaminetetraacetic acid (EDTA) (Invitr ogen Carlsbad, CA) was added to release the cells from the surface of the flask. Th e cells and trypsin were incubated at 37 oC under 5% CO2 for approximately 2 minutes. Following the incubat ion period, 12 ml of growth media was added to the cells and mixed by pi petting. The cells were then counted. The number of cells/we ll varied from one infection study to another (table 2.1). However, prior to inf ection, the cells were allowed to grow until they were confluent in each well. For the 6 w ell plates that the A549 cells were grown in, 3 ml of media including cells was used fo r each well. After plating, the cells were grown in an incubator at 37 oC under 5% CO2 until confluent. Once confluent, the cells were washed once with approximately 3 ml PBS/ well and then infected.
n Infection MOI Viral Titer (PFU/ml) # of cells/ well # of virions/ well ml virions ml infection media Passage number Cal 04 #1 1 2.0 x 10 8 1.8 x 106 1.8 x 10 6 0.0090 3.384 105 Cal 04 #2 1 2.0 x 10 8 1.3 x 106 1.3 x10 6 0.0065 1.961 99 Table 2.1. Data for each infection study including number of c ells, multiplicity of infection (MOI), the number of virions/well and the passage number of the cells. Although the number of cells/well varied between #1 and #2, the MOI was the same. The viral infection media was 1X Antibiotic-Antimyc otic and Opti-mem 1 Reduced-Serum Medium (1X) containing HEPES, 2.4 g/L Sodium biocarbonate and LGlutamine (infection media). The growth media was s uctioned off of the cells and 300 l/well of infection media containing virus was adde d to each well. The 0 h uninfected cells received an equal amount of media that did no t contain virus. The plated cells were placed in a 5% CO2 chamber on a rocker at 37 oC for 1 h. 2.2.2 RNA Isolation After 1 h, the media was aspirated off of all wells Three ml of media without virus was added to each later time point well. To i solate RNA, 500 l TRIzol Reagent (Invitrogen Carlsbad, CA) was added to each well pipetted up and down, and then collected. This was repeated with a second aliquot of TRIzol Reagent that was collected in the same vial. After the TRIzol Reagent was collected, 1/10 volume of BCP Phase Separation Reagent (Molecular Research Center Cincinnati, O hio) was added and the samples were vortexed for 30 s and then spun in a microcent rifuge at 13,200 rpm at 4 oC for 15 min. Following centrifugation, the aqueous phase wa s transferred to a new vial and an equal volume of isopropanol was added. This was mix ed by inverting the vial 30 times
and then spinning at 13,200 rpm at 4 oC for 15 min. The supernatant was decanted off and 1 ml of 75% ethanol was added to wash the sampl e. This was mixed by inverting 30 times and then centrifuged at 13,200 rpm at 4 oC for 15 min. The supernatant was pipetted off and the samples were allowed to dry un til the RNA pellet was clear. Once the pellets were dry, 50 l H2O was added to each sample and mixed by pipetting a nd the RNA concentration was determined using a NanoDrop 8 000 Spectrophotometer (Thermo Scientific Pittsburgh, PA). 2.2.3 Small RNA Isolation Small RNA was isolated from total RNA for each time point. The mirVanaTM miRNA Isolation Kit (Ambion Foster City, CA) was used and the procedure in the manual was followed. All centrifugation steps were done in a microcentrifuge. Each total RNA sample was mixed with 5 volumes of Lysis/Bindin g Buffer by inverting the tube several times. Following mixing, 1/10 volume of miRNA Homogenate Additive was added, the samples were vortexed, and then put on i ce for 10 min. Then, 1/3 volume of 100% ethanol was added to bring the ethanol concent ration to 25%, and the samples were thoroughly mixed by inverting the tubes several tim es. The samples were then transferred to filter cartridges and spun for 1 min at 5,000 rp m. The filter cartridges contained a glass-fiber filter that immobilized the RNAs. At a 25% ethanol concentration, the RNAs >200 bp were immobilized in the filter and the RNAs <200 bp passed through the column to the filtrate. The filtrates, containing the small RNA fraction, w ere mixed with 2/3 volume 100% ethanol at room temperature to bring the ethan ol concentration to 55% so that the
small RNA fraction was immobilized in the filter. T he mixtures were put onto filter cartridges and spun for 1 min at 5,000 rpm. The flo w-through was discarded and 700 l miRNA Wash Solution 1 was added to each filter cart ridge. The samples were again spun for 1 min at 5,000 rpm and the flow-through from ea ch sample was discarded. To further purify the RNA, 500 l Wash Solution 2/3 was added to each of the filter cartridges which were spun for 1 min at 5,000 rpm. This was re peated with a second aliquot of Wash Solution 2/3. The flow-through from each sampl e was discarded and the filter cartridges were spun for 1 min at 10,000 rpm to pul l any additional wash solution out of the filter. Following the centrifugation, the filte r cartridges were transferred into new collection tubes. Fifty l of 95 oC Elution Solution was added to the center of each filter cartridge, and the samples were incubated at room t emperature for 2 min. Following this incubation period, the samples were centrifuged for 1 min at 10,000 rpm. The process of adding elution solution, centrifuging, and saving t he flow-through was repeated using the same collection tube and the eluted RNA samples wer e stored in a -20 oC freezer. 2.3 Preparation of QPCR Components 2.3.1 Primer Design QPCR was the initial method for finding and quantif ying small RNAs that were present during infection. A kit specially designed for finding small RNAs, QuantiMir RT Kit (System Biosciences Mountain View, CA), was used for the PCR preparation and reactions. The kit provides reagents to polymerase a polyA tail onto all RNA molecules. Then, an oligo dT adaptor with a specific sequence of nucleotides (see A.2 for sequence) following the segment of thymines was hybridized to the polyA tail and reverse
r transcription was used to create the first strand o f cDNAs. This process generated a pool of cDNAs, with the sequence of the anchor-tailed RN A molecules, that could be used for qPCR with a user designed forward primer specific t o the RNA of interest and a universal reverse primer (see A.3 for sequence) (fig. 1.13). Forward primers were designed for each hairpin in C alifornia 04 (fig. 2.3) (see A.4 for sequences of forward primers). The usual lo cation of mature miRNAs within the pre-miRNA hairpin varies from the 5 side to the 3 side. On the 5 side, mature miRNAs usually end at the start of the loop. On the 3 sid e, mature miRNAs usually begin 2 nt after the loop (Grundhoff et al. 2006). However, th ese locations were determined by examining animal miRNAs; similar data is not availa ble for viral miRNAs. Additionally, these are the typical locations of the mature miRNA s, but mature miRNAs may be cleaved from other locations within the hairpin (Gr undhoff et al. 2006). Because mature miRNAs are not always cleaved from those locations and because there is no data on the locations of mature miRNAs within viral pre-miRNA h airpins, forward primers were designed to overlap each other and to cover the ent ire hairpin. The forward primers (Integrated DNA Technologies San Diego, CA) were designed to overlap by at least 10 nt, and to have a similar melting point so that they could be run on the same qPCR plate. The melting point was determined using idtDN A SciTools OligoAnalyzer 3.1 with the default settings (idtdna.com/analyzer/Applicati ons/OligoAnalyzer/). 2.3.2 Reverse Transcription of RNA To look for viral pre-miRNA hairpins, cDNAs were sy nthesized from the isolated small RNAs using the procedure from the Quantitect Reverse Transcription Kit (Qiagen
, Valencia, CA). The template RNAs were thawed on ice. Quantiscript RT Buffer, RT Primer Mix, and gDNA Wipeout Buffer were thawed at room temperature, gently mixed, centrifuged down, and placed on ice. The Quantiscri pt RT Buffer contained dNTPs and served as a buffer during the reverse transcription reaction. The RT Primer mix contained a blend of oligo-dTs and random primers dissolved i n water. The Quantiscript Reverse Transcriptase was a reverse transcriptase isolated from E. Coli The gDNA Wipeout Buffer was a proprietary blend from Qiagen that rem oved any contaminating genomic DNA from the RNA sample; the buffer likely containe d DNases. Two l gDNA Wipeout Buffer, 7X, 2 l template RNA, and 10 l H2O were mixed together for each reaction. The samples were then incubated at 42 oC for 2 min and immediately placed on ice. A reverse transcription master mix was prepared on ice. The master mix contained 1 l Reverse Transcriptase, 4 l RT Buffer (5X), and 1 l RT primer mix for each reaction. Six l of the master mix was then added to each RNA sample, and the solutions were incubated at 42 oC for 15 min and then at 95 oC for 3 min. The products were diluted 1:10 in H2O and the concentration and purity of the cDNA samples were determined using a NanoDrop 8000 Spect rophotometer (table 2.2). ID of infection experiment Concentration (ng/ l) A260/A280 A260/A230 Cal04 #1 0h 197.1 1.76 2.28 Cal04 #1 24h 192.4 1.77 2.27 Cal04 #1 48h 181.2 1.77 2.25 Cal04 #2 0h 194.7 1.78 2.26 Cal04 #2 24h 190.1 1.77 2.27 Cal04 #2 48h 193.2 1.78 2.25 Table 2.2. The concentrations and purity of the cDNA used to l ook for pre-miRNA hairpins using the qPCR method, an A260/A280 value of 1.7-2.0 indicates pure DNA, an A260/A230 value of approximately 1.8 indicates pure DNA. The A260/A280 value is used to determine whether there are protein contami nants. The A260/A230 value is used to determine if there are organic solvent contamina nts.
2.3.3 Creation of a pool of anchor-tailed cDNAs To look for small, virally encoded RNAs, cDNAs were synthesized from the isolated RNAs using the procedure from the QuantiMi r RT Kit Small RNA Quantitation System (System Biosciences Mountain View, CA), w hich adds a polyA tail to all RNAs, anneals an anchor dT adaptor and then synthes izes cDNAs (Fig. 1.13). For each reaction, 5 l RNA, 2 l 5X PolyA Buffer, 1 l 25 m M MnCl2, 1.5 l 5 mM ATP, and 0.5 l PolyA Polymerase were mixed by gentle pipett ing in a PCR tube. Each tube was incubated for 30 min at 37 2C. Following the incubation, 0.5 l Oligo dT Adapto r was added to each sample and the samples were heated fo r 5 min at 60 2C and then allowed to cool to room temperature for 2 min. To each sample, 4 L 5X RT Buffer, 2 l dNTP mix, 1.5 l 0.1 M DTT, 1 l reverse transcriptase, and 1 .5 l RNAse free H2O were added. The samples were incubated for 60 min at 42 2C and then for 10 min at 95 2C. The cDNA was then diluted 1:100 in H2O, and the concentration and purity of the cDNA sam ples were determined using a NanoDrop 8000 Spectrophotom eter (table 2.3). ID of infection experiment Concentration (ng/ l) A260/A280 A260/A230 Cal04 #1 0h 34.61 1.61 2.05 Cal04 #1 24h 28.4 1.71 2.11 Cal04 #1 48h 25.97 1.68 1.96 Cal04 #2 0h 27.95 1.66 2.23 Cal04 #2 24h 27.85 1.70 2.03 Cal04 #2 48h 26.40 1.64 1.94 Table 2.3: The concentrations and purity of the cDNA used to l ook for small RNAs using the qPCR method, an A260/A280 value of 1.7-2. 0 indicates pure DNA, an A260/A230 value of approximately 1.8 indicates pure DNA. The A260/A280 value is used to determine whether there are protein contami nants. The A260/A230 value is used to determine if there are organic solvent contamina nts.
2.3.4 PCR Procedure The procedure for real-time qPCR Reactions using Qu antiMir cDNAs was slightly modified. Each qPCR reaction to look for s mall RNAs, possible miRNAs, (total volume of 15 l) contained 7.5 l 2x QuantiTect SYB R Green Rt-PCR Master Mix (Qiagen Valencia, CA), 0.25 l Universal Reverse Primer (10 M) (System Biosciences Mountain View, CA), 1 l cDNA from t he appropriate time point, and 2 l forward primer (5 M). Each qPCR reaction to loo k for pre-miRNA hairpins (total volume of 15 l), contained 7.5 l 2x QuantiTect SY BR Green RT-PCR Master Mix, 3 l cDNA, 2 l forward primer (2.5 M), 2 l reverse primer (2.5 M) (see A.5 for sequence of forward and reverse primers). QPCR reac tion conditions were as follows: 1 cycle at 95 2C for 3 min, 40 cycles at 95 2C for 15 s, 51 2C for 30 s, 60 2C for 45 s, and 1 cycle at 95 2C for 15 s, 60 2C for 30 s, 95 2C for 15 s. The qPCR was run on an Applied Biosystems 7300 Real-Time PCR System (Applied Biosy stems Carslbad, CA). QPCR reactions were done in triplicate. For the small R NAs a positive control was provided by the QuantiMir Kit: a forward primer for miR-16, an endogenous miRNA present in A549 cells (see A.6 for sequence of miR-16 forward prime r). Non-template control (NTC) qPCR reactions, composed of 7.5 l 2x QuantiTect SYBR Green Rt-PCR Master Mix, 0.25 l Universal Rev erse Primer (10 M) and 1 l cDNA (prepared according to the above procedurepr eparation of a pool of anchor-tailed cDNAs, section 2.3.3) (total volume 15 l), were ru n with anchor-tailed cDNAs from each time point to ensure that the reverse primer w as not hybridizing to itself. Because there were so many user-designed forward primers, i t was not feasible to run a NTC
reaction for each forward primer to check for prime r-dimers (when primers hybridize to themselves), as is typically done for qPCR NTCs. In stead, the dissociation curve, which tracks the temperature at which the product dissoci ates, was examined carefully for each primer. Primer-dimers dissociate at a lower tempera ture than actual primer-cDNA hybridized pairs do. The primers that formed primer -dimers are noted in A.11. Additionally, the DNA gel electrophoresis of the qP CR products showed any primerdimers that had formed during the qPCR. NTC qPCR reactions composed of 7.5 l 2x QuantiTect SYBR Green RT-PCR Master Mix, 3 l cDNA (prepared according to the ab ove procedurereverse transcription of RNA, section 2.3.4), 2 l forward primer (2.5 M), 2 l reverse primer (2.5 M), were run with the forward and reverse pri mers for the hairpins to ensure that those primers were not forming primer-dimers. 2.4 Further Analysis of QPCR Products 2.4.1 DNA Gel Electrophoresis Primers that showed amplification of the cDNA from 24 and 48 hpi, but no amplification of cDNA from 0 hpi, when the cells ha d not yet been exposed to virus, were run on a 3.5% agarose gel at 80 V next to a 20 bp ladder and the qPCR product from a reaction containing the forward primer for m iR-16 and the universal reverse primer as a positive control. Five l of cDNA was m ixed with 2 l of DNA Gel Loading Buffer (Fisher Scientific Gaithersburg, MD) and loaded into the agarose gel. If only two of the three replicates showed amplification du ring qPCR, the cDNA from one of those wells was run on the gel. If the cDNA amplifi ed by that primer was chosen for
sequencing, the cDNA from the well that did not sho w amplification was not combined, purified, and sequenced. CDNA that was similar in l ength to miR-16 at 24 and 48 hpi, but not at 0 hpi was ligated into a plasmid and seq uenced. 2.4.2 PCR Product Purification The cDNAs from the qPCR reactions were purified usi ng the procedure from MinElute PCR Purification kit (Qiagen Valencia, CA). MinElute filter columns contained silica filters. DNA binds to silica when the concentration of chaotropic salts is high and the pH is approximately 7.5. In each MinEl ute filter column, 20 l qPCR Product (replicate wells were combined to give a to tal volume of 20 l) and 100 l Buffer PB were mixed and then centrifuged for 1 min at 13,000 rpm in a microcentrifuge. Buffer PB provided the necessary concentration of s alts and the correct pH to maximize the adsorption of DNA to the silica filter. After c entrifugation, the flow through from each tube was discarded and 0.75 ml Buffer PE was a dded to wash the salts from the filter. The tubes were centrifuged for 1 min at 13, 000 rpm and the flow through from each tube was discarded. The tubes were centrifuged for an additional 1 min at 13,000 rpm and the filters were placed into clean 1.5 ml m icrocentrifuge tubes and 10 l Buffer EB was added. Buffer EB provided a low salt concent ration and basic pH that efficiently eluted the DNA from the filter. The tubes were allo wed to stand for 1 min and then centrifuged for 1 min at 13,000 rpm, after which th e filters were discarded and the purified PCR product was obtained and frozen at -20 oC.
2.4.3 Plasmid Creation A plasmid was created from each purified product us ing a modified procedure from the PCR cloning kit (Qiagen Valencia, CA). The PCR cloning kit used a UAbased ligation. The UA-based ligation used the sing le A overhang present in PCR products to hybridize to the pDrive Cloning Vector that had a U overhang at each end. The vector was prepared for ampicillin selection an d contained a T7 promoter on one side of the cloning site which is used later for sequenc ing (see A.7 for vector information). In a PCR tube, 0.4 l pDrive Cloning Vector, 1.6 l PC R Product, and 2 l Ligation Master Mix, 2X were added and then incubated for 2 hr at 1 6 oC. Following the incubation, the samples in the tubes were mixed gently. 2.4.4 Transfection of E. coli Each plasmid was transfected into E. coli using a modified procedure from the High Efficiency transformation protocol from NEB 10 -beta Competent E. coli (New England BioLabs Ipswich, MA). The E. coli contained the lac operon that allowed for blue/white screening using the X-gal and IPTG metho d described below. A transformation tube was pre-chilled on ice and 25 l E. coli and 1 l plasmid DNA were added. The tube was carefully flicked 4-5 times and placed on ice for 30 minutes. The tube was then heat shocked for 30 s at 42 oC and placed back on ice. After 30 min, 960 l SOC (New England BioLabs Ipswich, MA) was added and the tube was incubated and shaken for 60 min at 37 oC. To allow colonies that had received the plasmid to be distinguished from those that had not 20 l X-gal (Promega Madison,
n WI) and 10 l IPTG (Promega Madison, WI) were ad ded to each mixture after incubation. The IPTG and X-gal allowed for the blue /white screening of the colonies. Colonies that had not received the vector insert ha d an intact lac operon that was induced to produce functional -galactosidase enzyme by IPTG. -galactosidase metabolizes Xgal to form a bright blue, insoluble prodcut. There fore, the white colonies contained the vector insert, while the blue colonies did not. Fif ty l of each mixture was then spread onto a Lysogeny Broth (LB) agar plate containing 10 0 g/ml ampicillin. The plates were incubated overnight (approx. 16 hrs) at 37 oC. After the overnight incubation, 1 white colony was selected from each plate and placed into 2 ml of liquid LB media. The tubes were incubated in a shaking incubator for 12-16 hrs at 37 oC. 2.4.5 Purification of DNA for Sequencing After incubation, the procedure from the QIAprep mi niprep kit (Qiagen Valencia, CA) was followed to isolate the plasmid D NA. Each tube was centrifuged at 4,000 g for 15 min, the supernatant was discarded, and the pelleted bacterial cells were resuspended in 250 l Buffer P1 and transferred to a microcentrifuge tube. Two hundred and fifty l of Buffer P2 was added to each sample and mixed thoroughly by inverting the tubes 4-6 times. Buffer P2 served to lyse the b acterial cells by creating alkaline conditions. Three hundred and fifty l of Buffer N3 was added to each sample and mixed immediately and thoroughly by inverting the tubes 4 -6 times. Buffer N3 neutralized Buffer P2 and increased the salt conditions so the DNA would bind to the silica filter. The tubes were then centrifuged for 10 min at 13,00 0 rpm in a microcentrifuge. The supernatant from each tube was applied to a QIA prep spin column that was
centrifuged for 1 min at 13,000 rpm. The flow throu gh from each tube was discarded and 0.5 ml of Buffer PB was added to each sample. The t ubes were centrifuged for 60 s at 13,000 rpm and the flow through from each tube was discarded. The QIAprep spin columns were washed by adding 0.75 ml Buffer PE and centrifuging for 1 min. The flow through was again discarded and the tubes were cent rifuged for 1 min. The QIAPrep columns were placed in 1.5 ml microcentrifuge tubes and 50 l Buffer EB was added to the center of each spin column. Buffer EB was a low -salt buffer with a pH of approximately 7.5 that released the DNA from the si lica membrane. The tubes were allowed to stand for 1 min, and then centrifuged fo r 1 min at 13,000 rpm. The columns were discarded and the plasmid DNA concentration wa s determined using a NanoDrop 8000 Spectrophotometer. The samples were diluted or concentrated, as needed, to a concentration of 150-300 ng/l and sent to Operon ( Eurofins MWG Operon Huntsville, AL) for sequencing. 2.5 Procedure to Decrease the Expression Levels of Dicer 2.5.1 Procedure for Dicer shRNA To examine the processing of the small, virally en coded RNAs, an attempt was made to decrease the expression level of the protei n Dicer; Dicer processes pre-miRNA hairpins into mature miRNAs. The goal was to introd uce a short-hairpin RNA (shRNA) into the cells that would be processed by Dicer to form a siRNA complementary to the 3 UTR of the mRNA of Dicer. The siRNA would reduce th e expression levels of Dicer. The knockout of Dicer was incomplete because Dicer is required to process shRNAs
(Sashital et al. 2009) and the A549 cells that were used for the infection studies were not easily transfected (see results section 3.3). The Dicer shRNA was received in a plasmid in ampici llin resistant E. coli in a stab culture. A sterile pipette tip was touched to the culture and streaked on a LB Agar plate containing ampicillin (100 ng/l). The plate was incubated overnight at 37 oC. A single colony was selected from the plate and grown in 50 ml liquid LB for 16 hours in a 37 oC shaking incubator. The PureYield Plasmid Maxiprep System (Promega M adison, WI) was used to isolate and purify the DNA from the E. coli cells. The cells were pelleted by centrifuging at 5000 g for 10 minutes. The supernatant was disca rded and the cell pellet was resuspended in 12 ml of Cell Resuspension Solution. Following resuspension, 12 ml of Cell Lysis Solution was added and the solution was inverted gently 5 times to mix and incubated at room temperature for 3 min. Following the incubation, 12 ml of Neutralization Solution was added and the mixture w as inverted gently 10 times to mix. Neutralization Solution acted to increase the salt concentration and neutralize the cell lysis solution. The lysate was centrifuged at 7,000 g for 30 min at room temperature. To purify the plasmid, a blue PureYieldTM Clearing Column and white PureYieldTM Maxi Binding Column were assembled in a stack with the clearing column on top. The column stack was placed on the vacuum m anifold. The filter was a silica membrane that bound the plasmid DNA. One half the l ysate was poured into the clearing column and the maximum vacuum was applied until the lysate passed through both columns. The remaining lysate was added and the pro cess was repeated. The vacuum was then released and the clearing column was discarded The binding column was left on the
r vacuum manifold and 5 ml of Endotoxin Removal Wash was applied to the column. The vacuum was applied until the solution was pulled th rough the column. To wash the column, 20 ml of Column Wash was applied and the va cuum was used to pull the solution through the column. The membrane was then dried by applying the vacuum for 5 min. The binding column was removed from the vacuu m manifold and the tip was tapped dry on a paper towel. The column was placed in a new 50 ml centrifuge tube and 1.5 ml of nuclease-free water were applied. The sam ple was then centrifuged at 2,000 g in a swinging bucket rotor at room temperature for 5 min. The filtrate containing the DNA was collected, transferred to a 1.5 ml tube, an d stored at -20 oC. 2.5.2 Transfection of cells To determine whether the transfection of the Dicer shRNA would work to effectively reduce the levels of Dicer, the transfe ction was first performed on Human Embryonic Kidney 293 (293T) cells, which are easier to transfect than A549 cells. The cells were transfected with the Dicer shRNA and wit h a GIPZ lentiviral shRNAmir (Thermo Scientific Pittsburgh, PA). The GIPZ len tiviral shRNAmir caused the cells to express Green Fluorescent Protein (GFP) when the in serted shRNA was processed. GIPZ acted as a positive control for to show that the tr ansfection was effective and that the shRNAs were processed by the cell and as a negative control because it should not have affected the levels of Dicer present in the cells. Following the transfection of 293T cells, the same procedure was applied to transfect A549 ce lls that were later infected with Influenza A. The transfection reagent used was FuGE NE HD Transfection Reagent (Roche Indianapolis, IN). The transfection reage nt was a proprietary blend of lipids in
n 80% ethanol that acted by binding to the DNA and th en entering the cell through the plasmid membrane. The plasmids and FuGENE were mixed in Opti-MEM 1 wi thout serum and then mixed together to give concentrations that depended on the number of cells that were going to be transfected (table 2.4). Two hundred l was then applied to each well of a plate of cells prepared as described above, and the plate was gently shaken for a few seconds then put back into the incubator at 37 oC under 5% CO2. In 24 h, cells transfected with GIPZ shRNAs showed GFP under a fluorescent mic roscope. Cell Type Plate Type # of cells/well Amount of Transfection reagent/well (l) Amount of plasmid/well (ng) Amount of media added /well (l) 293T 12 well 2.5 x 10 5 1.5 500 200 A549 6 well 6 x 10 5 3 1000 200 Table 2.4. Amounts of reagents used for transfection of Dicer shRNA and GIPz shRNA. RNA was harvested from the cells at 24h and 48h aft er transfection following the trizol procedure given above (section 2.2.2). The Q uantitect Reverse Transcription Kit was used to synthesize cDNA (section 2.3.2). QPCR w as run using the cDNA from 24h and 48h. Each PCR reaction (total volume 15 l) co ntained 7.5 l 2x QuantiTect SYBR Green RT-PCR Master Mix, 3 l cDNA, 2 l forward pr imer (2.5 M), and 2 l reverse primer (2.5 M) (see A.8 for sequence of forward an d reverse primers for Dicer). In A549 cells, a transfection with the Dicer shRNA and pGIPz was performed prior to infecting the cells with California 04. Th e transfection was performed as above and 48 h later the infection was performed as descr ibed previously. The procedures
n described above were followed to prepare and run qP CR using the RNA from the transfected cells. 2.6 QPCR Data Analysis Because the goal of this research was to determine whether the Influenza A virus encodes small RNA molecules, quantification of the levels of the small RNA molecules was not necessary. The presence of a qPCR product o f the appropriate size at 24 hpi and 48 hpi and an absence at 0 hpi along with sequencin g results was considered sufficient to confirm the presence of small, virally encoded RNA molecules. To analyze the qPCR results, the results from each primer were normaliz ed only to themselves. They could not be normalized to other primers because the efficien cy of the qPCR varied from one primer to the next (data not shown). First, the Ct values for replicates were averaged. If two of the three replicates showed amplification, while one did not, the two Ct values were averaged. If only one of the three replicates showed amplification, the aver age Ct value was recorded as undetermined (see A.12 for primers for which this o ccurred). These samples are marked by an asterisk to the right of the time point on th e graph. The expression levels of each primer were normalized to the Ct value of the 24 hp i cDNA: Relative Expression Level = 2^(Ct value of 24 hpi cDNA Ct value of 0, 24, or 48 hpi cDNA) This automatically set the expression level at 24 hpi equal to 1. If there was no amplification of 24 hpi cDNA for a certain primer, then the data for that primer was
n normalized to the Ct value of the 48 hpi cDNA. This occurred for Cal04 #1 primer 7-2 3-4 (fig. 3.38) and Cal04 #2 primers 5 5-1, 5 52, 7-1 5-2, 7-2 3-1 (fig. 3.39), 7-2 3-2, 7-2 3-4 (fig. 3.39), 7-3 5-1, and 7-3 5 -2. This method of data analysis was used because the development of a standard curve to determine the amplification efficiency for each primer set would have been time consuming and would not have significantly improved the results. To create a standard curve, an oligonucleotide with the sequence of one hairpin would have to be synthesized. Then, a known copy number would have to be serially diluted to create known concentrations of the hairpin. The Ct values of ea ch dilution would be used to determine the amplification efficiency.
n Chapter 3: Results and Discussion While the general structure of scientific papers an d, by extension, theses, includes a separate results and discussion section, for the purposes of clarity the two sections are combined in this thesis; this will allow further el aboration on the results when they are presented. Elaboration is necessary for several rea sons. First, the methods of finding mature miRNAs differ from those typically used in p apers presenting novel miRNAs. Also, there was a lack of correlation between the q PCR results and the presence of a mature miRNA. Therefore, a discussion of the DNA ge l electrophoresis, sequencing, and qPCR results at the time they are presented will se rve to clarify the significance of each of the results. The results from the possible pre-miRNA hairpins wi ll be presented first because they provided the impetus to continue the search fo r mature miRNAs cleaved from the hairpins. Then the results from the possible mature miRNA will be presented followed by the results that did not yield mature miRNAs. Two i dentical experiments were conducted, and the results from both are presented in conjunct ion below. The entire process, beginning with plating and infection of cells, was repeated meaning that these biological replicate results were not generated from the same RNA samples. The two experiments, performed using A/California/04/2009, are referred to as Cal04 #1 and Cal04 #2. Additionally, when presenting the sequencing result s, it is important to note that the cDNAs generated by qPCR were sequenced, but that th e sequence of the cDNA was either identical to, or complementary to, the origi nal RNA sequence. The primers for the small RNAs were labeled according to the segment of the viral genome where the
n predicted hairpin was located (5 or 7), the number of the hairpin within that segment (7-1, 7-2, or 7-3), and the location on that hairpin (51, 5-2, 3-1, 3-2, etc.) with 3 or 5 indicating the arm of the hairpin where the primer was located. Following a presentation and discussion of the results, possible improvement s to the experiment and future possible directions for this project will be discus sed. 3.1 Pre-miRNA Hairpins 3.1.1 Predicted Hairpins Pre-miRNA hairpins were predicted by VMir, a progra m designed to predict hairpins within viral genomes. VMir has been used s uccessfully to predict pre-miRNA hairpins in several viruses from the herpesvirus fa mily (Grundhoff et al. 2006). Figure 3.1. A diagram indicating the location of the predicted pre-miRNA hairpins in the California04 influenza virus genome. There are thre e predicted hairpins in the M1/M2 segment: hairpins 7-1, 7-2, and 7-3. There is one p redicted hairpin in the NP segment, hairpin 5.
n There were four hairpins predicted by the program V Mir in the Cal04 genome. There was 1 predicted hairpin in segment 5 (NP) and 3 predicted hairpins in segment 7 (M1/M2) (fig. 3.1) (see A.1 for sequences of the pr edicted hairpins and A.5 for sequences of primers used to amplify the hairpins). 3.1.2 QPCR Results QPCR was a preliminary test to determine whether th e predicted pre-miRNA hairpins were expressed by the virus. Figure 3.2. QPCR of pre-miRNA hairpins in Cal04 #1. The first number (5 or 7) indicates which segment of the viral genome encodes the predicted hairpin. The number following the dash (1, 2, or 3) indicates the numbe r of that hairpin in that segment. MiR16 is a cellular miRNA that was used as a positive control. Samples were collected before infection (0h) and then 24h and 48 hr post i nfection (hpi). n ###############nrr r # # # n333 3
nn Figure 3.3. QPCR of pre-miRNA hairpins in Cal04 #2. The first number (5 or 7) indicates which segment of the viral genome encodes the predicted hairpin. The number following the dash (1, 2, or 3) indicates the numbe r of that hairpin in that segment. MiR16 is a cellular miRNA that was used as a positive control. The results of Cal04 #1 (fig. 3.2) and Cal04 #2 should be very similar because th e experiments are biological replicates. The qPCR results from Cal04 #1 and #2 showed that a ll of the predicted premiRNA hairpin sequences were expressed 24 and 48 ho urs post infection (hpi), but not 0 hpi (figs. 3.2 and 3.3). In the second replicate of the experiment, Cal04 #2, the expression levels for each predicted pre-miRNA hairpin were gr eater at 48 hpi than at 24 hpi. By 24 hpi with a MOI of 1, the virus had gone through sev eral cycles of infecting cells, replicating, and infecting new cells. Although an M OI of one indicates that there was one virion per cell, unpublished studies by the Harrod lab indicate that not all the cells are infected initially at this MOI (Dr. Jennifer Tipper personal communication). The uninfected cells are infected at a later time point as the virus replicates in and then spreads from infected cells. If the pre-miRNA hairp ins are not cleaved to form miRNAs, n ###############nrr r # # # n333 3
n the increased expression levels at 48 hpi could be explained by the accumulation of the pre-miRNA hairpins over the course of the infection process. In the first replicate of the experiment, Cal04 #1, the expression levels for the predicted pre-miRNA hairpins were less consistent. Hairpin 7-2 was the only predicted hairpin with a higher concentration at 48 hpi as co mpared to 24 hpi. The lower expression of the other small RNAs in the Cal04 #1 experiment can be partially explained by looking at the lower concentration of cDNA that was added to the Cal04 #1 48h qPCR reaction (table 3.1). However, there were likely ot her factors affecting the expression levels of the hairpins at 48 hpi because the cDNA c oncentration differences alone are not enough to account for the decreased expression leve ls at 48 hpi in the Cal 04 #1 experiment compared to the Cal 04 #2 experiment. ID of infection experiment Concentration (ng/ l) A260/A280 A260/A230 Cal04 #1 0h 197.1 1.76 2.28 Cal04 #1 24h 192.4 1.77 2.27 Cal04 #1 48h 181.2 1.77 2.25 Cal04 #2 0h 194.7 1.78 2.26 Cal04 #2 24h 190.1 1.77 2.27 Cal04 #2 48h 193.2 1.78 2.25 Average 191.5 1.77 2.26 Standard Deviation 5.5 0.008 0.02 Standard Error 2.3 0.003 0.005 Table 3.1. The concentration of cDNA used in each qPCR experi ment is indicated. An A260/A280 value of approximately 1.8 indicates pure DNA, while deviations from this value indicate protein contamination. An A260/A230 value of 1.8-2.2 indicates pure DNA, and deviations from this indicate organic solv ent contamination (Thermo Fisher Scientific 2008). For Cal04 #1 the concentration of cDNA 48 hpi was l ower than the concentration of cDNA 0 or 24 hpi. The number of ng of cDNA added to each qPCR reaction was not normalized. However, the number of cells plated for each time point in the Cal04 #1 and
n Cal04 #2 infections was consistent. This should yie ld comparable levels of cDNA for 0, 24, and 48 hpi. The number of cells plated was not consistent from Cal04 #1 to Cal04 #2, although in each study the cells were allowed to gr ow until they were confluent prior to infection. A further discussion of this issue follo ws in section 3.5.1. Although the difference in concentration between th e mean and Cal04 #1 48 hpi was 10.3 ng/l, a 5% difference and nearly two stan dard deviations, the difference is still not large enough to account for the differences see n in expression levels from 48 hpi Cal04 #1 as compared to Cal04 #2. This indicates th at there were other factors affecting the Cal04 #1 48 hpi qPCR results. One possible fact or is the degradation of the Cal04 #1 48 hpi RNA or cDNA; degraded RNA or cDNA would stil l give a similar concentration reading from a spectrometer, but the qPCR results w ould show less intact cDNA present. The RNA samples were not run on a gel prior to qPCR so the integrity of each sample is not known. The difference in cDNA concentration and other fact ors contributing to the qPCR results for Cal04 #1 48 hpi were reflected in the e xpression levels of miR-16 as well. Cal04 #2 showed a significant increase in the expre ssion level of miR-16 from 24 hpi to 48 hpi while Cal04 #1 showed a slight decrease betw een 24 and 48 hpi. Given that the concentrations of cDNA were more consistent for Cal 04 #2, it appears that the second measurement was more reliable and that the level of miR-16 did increase slightly during the infection process. MiR-16 regulates the expression of BCL2, a gene dir ectly involved in inhibiting cellular apoptosis. During several types of cancers miR-16 expression is decreased leading to an increase in BCL2 expression (Bonci et al. 2008). Wang et al. (2009) showed
nr that during infection of chicken lung cells with a strain of avian influenza, the levels of miR-16 decreased by 80%. If the same is true for hu man lung cells, though this is contrary to what the qPCR data indicates, then the expression levels of miR-16 cannot be used as a consistently expressed endogenous control to accurately examine the amount of cDNA added to each reaction (which was the original goal). Despite the differences in expression levels, miR-16 still serves as a positiv e control for the size and amplification of a small RNA. 3.1.3 DNA Gel Electrophoresis Results Figure 3.4. DNA Gel Electrophoresis Results for Cal04 #1 hairpi ns 7-1 and 7-3. Because of the poor quality of the gel image, the gel image was inverted and there are arrows indicating the bands for hairpin 7-3. The gel was i maged on the same equipment using the same procedure as all the other gels. There are bands present at 24 and 48 hpi but not at 0 hpi for both hairpins, indicating that the qPC R amplified RNA of viral origin.
Figure 3.5. DNA Gel Electrophoresis Results for Cal04 #1 hairpi ns 5 and 7-2 and Cal04 #2 hairpin 7-3. The band at 0 hpi for Cal04 #1 hair pin 7-2 indicates that the qPCR amplified something derived from the cell rather th an the virus because the cell had not been exposed to virus at 0 hpi.Cal04 #2 hairpin 7-3 and Cal04 #1 hairpin 5 both showed bands 24 and 48 hpi, but not 0 hpi.
Figure 3.6. DNA Gel Electrophoresis Results for Cal04 #2 hairpi ns 5, 7-1, and 7-2. There are bands at 24 and 48 hpi for hairpins 5, 71, and 7-2 and not at 0 hpi. This is different than the results seen in fig. 3.5 for Cal 04 #1 hairpin 7-2, which had a band present at 0 hpi. The qPCR products from the pre-miRNA hairpins were run on agarose electrophoresis gels. They were run next to a 20 bp ladder and the cDNA amplified by a primer for miR-16 (figs. 3.4-3.6). In order to be c onsidered a virally derived hairpin, there had to be a band of approximately the same si ze as the predicted qPCR product from the 24 and 48 hpi samples and no band from the 0 hpi sample. For Cal04 #1 and #2, hairpins 5, 7-1, and 7-3 showe d no band at 0 hpi and bands at 24 and 48 hpi. The approximate lengths of hairpi ns 5, 7-1, and 7-3 were consistent with the predicted lengths of 61, 69, and 40 bp res pectively, based on the location of the primers on the hairpin (see A.5 for size of PCR pro ducts based on primer location).
Based on the qPCR and DNA gel electrophoresis data, it appears that hairpins 5, 7-1, and 7-3 are virally encoded potential pre-miRNA hairpin s. For hairpin 7-2, the results from Cal04 #1 conflict slightly with the results from Cal04 #2. Cal04 #1 showed a band of approximately 8 0 bp at 0, 24, and 48 hpi while Cal04 #2 showed a band of the same approximate leng th only at 24 and 48 hpi. The band shown at 0 hpi from Cal04 #1 indicated that the pri mers for hairpin 7-2 may have amplified cellular RNA. It appears that this did no t happen in the Cal04 #2 0h sample. However, because the cDNA amplified by primers for hairpin 7-2 appeared to be the same length for Cal04 #1 (0, 24, and 48 hpi) and #2 (24 and 48 hpi), the primers may have amplified the same cellular RNA in each case. Figure 3.7. The top BLAST result for the forward primer for th e predicted viral hairpin 7-2. The top matches (100% complementarity) were al l to segment 7 in Influenza A H1N1 viruses, including multiple matches to the 200 9 Swine virus, of which Cal04 is one strain. Influenza A virus (A/Russia/01-MA/2009(H1N1)) segme nt 7 matrix protein 2 (M2) and matrix protein 1 (M1) genes, com plete cds Length=1060 Score = 36.2 bits (18), Expect = 0.88 Identities = 18/18 (100%), Gaps = 0/18 (0%) Strand=Plus/Plus Query 1 TAGGCAGATGGTACATGC 18 |||||||||||||||||| Sbjct 671 TAGGCAGATGGTACATGC 688
Figure 3.8. The top BLAST result for the reverse primer for th e predicted viral hairpin 72. The top matches (100% complementarity) were all to segment 7 in Influenza A H1N1 viruses, including multiple matches to the 2009 Swi ne virus, of which Cal04 is one strain. The primer sequences were tested to determine if th ey could amplify a ~80 bp PCR product from human cDNA. First, the NCBI rever se electronic PCR (e-PCR) tool was used to perform a reverse e-PCR search against the human transcriptome (Dataset  Homo sapiens transcriptome snapshot 2010/10/ 29). In the reverse e-PCR test, primer sequences are entered with an expected PCR p roduct size, and the entire genome is tested to see if the primers could make a PCR pr oduct near the appropriate size. For this test, the criteria were set to the least strin gent, allowing 2 mismatches, 2 gaps in the primer matches, and an allowed PCR product size dev iation of 50 bp. The search did not match the forward and reverse primers for hairpin 7 -2 to any potential template sequence from the human genome.Also, the sequences of the fo rward and reverse primers for hairpin 7-2 were compared to the human genome to de termine if they annealed to any cellular sequences (figs. 3.7 and 3.8). This search did not return any complementarity between regions in the human genome and the primers Since the hairpin 7-2 PCR Influenza A virus (A/Russia/01-MA/2009(H1N1)) segme nt 7 matrix protein 2 (M2) and matrix protein 1 (M1) genes, com plete cds Length=1060 Score = 40.1 bits (20), Expect = 0.056 Identities = 20/20 (100%), Gaps = 0/20 (0%) Strand=Plus/Minus Query 1 GCTAGGATGAGTCCCAATAG 20 |||||||||||||||||||| Sbjct 716 GCTAGGATGAGTCCCAATAG 697
product was not sequenced, it is impossible to say if the qPCR product resulted from cellular genes. Since the 7-2 primers dont anneal to the host cell genome, its possible that the PCR product in Cal #1 0h was from nonspeci fic amplification or a result of contamination of the sample by viral cDNA. There was also a discrepancy between the predicted size of the amplicon of hairpin 7-2 and the size estimated from the gel. Th e gel showed an approximate size of 80 bp while the predicted size of the amplicon was 45 bp. A BLAST search showed that the primers were complementary to regions within th e hairpin that would generate an amplicon of 45 bp; the primers did not have a high percentage of complementarity to other regions in the genome. Based on the qPCR and DNA gel electrophoresis resul ts, hairpins 5, 7-1, and 7-3 are expressed. The qPCR showed amplification 24 and 48 hpi and no amplification 0 hpi. The DNA gel electrophoresis demonstrated that the q PCR products were of the approximately correct size based on the location of the primers on the predicted hairpins. Because the primers for hairpin 7-2 amplified somet hing 0 hpi for Cal04 #1, but not for Cal04 #2, and the size of the amplicon varied from the predicted size, it is difficult to conclude whether the predicted pre-miRNA hairpin 72 was expressed. However, work on possible miRNAs cleaved from hairpin 7-2 was con tinued. 3.2 The Search for miRNAs Once the information regarding the predicted pre-mi RNAs had been gathered and analyzed, the project moved on to examine the possi ble mature miRNAs that could be cleaved from the pre-miRNA hairpins. For the small RNA that appears to be a mature
miRNA, the DNA gel electrophoresis data will be pre sented first, followed by the sequencing data, and finally the qPCR. Then the dat a for the cDNAs that were sequenced but did not appear to be mature miRNAs will be disc ussed in the same order. A brief discussion of the DNA gel electrophoresis and qPCR results that did not yield possible miRNAs will follow. 3.2.1 Mature MicroRNAs In Hairpin 5, 7-2, and 7-3, no possible mature miRN As were found. The hairpins were predicted using VMir and then various possible mature miRNAs from these hairpins were tested for using the qPCR, DNA gel electrophor esis, and sequencing approach that covered the majority of possible locations for miRN As within the predicted hairpin. After reviewing the qPCR and DNA gel electrophoresis resu lts, several qPCR products from hairpins 5 and 7-2 were sequenced, but the sequenci ng results were not consistent with the size or location of mature miRNAs. The possibl e miRNA from hairpin 7-1 will be presented first. Figure 3.9. Locations of primers complementary to regions of H airpin 7-1. The typical locations of mature miRNAs are boxed in red while t he locations of the forward primers are indicated by blue bars with the name of the pri mer above the bar. All the primers shown here were designed using the longest predicte d pre-miRNA hairpin, hairpin 7-1.
n Figure 3.10. Sequencing results for cDNA amplified by primer 7-1 3-1. Hairpin 7-1 is displayed, and the sequence from the sequencing res ults is indicated by a green bar. The forward primer location is indicated by a blue bar, and the location where miRNAs are usually cleaved from is indicated by a red box. In Hairpin 7-1, one possible mature miRNA was ident ified. However, the cDNAs generated from primers 7-1 3-1 and 7-1 3-2 (fig. 3.9) were sequenced. Primer 7-1 3-2 will be discussed in the next section because it do es not appear to be a mature miRNA. The sequence of the cDNA generated in the qPCR reac tion with primer 7-1 3-1 included the entire sequence complementary to primer 3-1 an d 3 additional nucleotides from the hairpin (fig. 3.10). This gave the sequenced cDNA a total length of 20 bp, putting it within the range expected for a mature miRNA. The c DNA sequence from 24 and 48 hpi and from Cal04 #1 and #2 was identical. Primer 7-1 3-1 covers the area of the hairpin
where it is likely a mature miRNA would be located (shown by the red box in fig. 3.10) although the sequencing data indicated that the miR NA started 4 nt after the end of the loop part of the hairpin, instead of the 2 nt spaci ng that is most commonly found (Grundhoff et al. 2006). Figure 3.11. Cal04 #1 DNA Gel electrophoresis results for cDNAs amplified by primer 7-1 3-1. There is amplification 24 and 48 hpi, but no amplification 0 hpi. The band at 24 hpi is blurred and could indicate multiple qPCR pro ducts or simply poor gel preparation. The primers are visible at the bottom of the gel in the 0, 24, and 48 hpi lanes. Figure 3.12. Cal04 #2 DNA Gel electrophoresis results forcDNAs amplified by primer 71 3-1. Multiple bands are visible 24 and 48 hpi. O nly the shorter fragment of cDNA was obtained from the sequencing reactions; those resul ts are discussed below. The DNA gel electrophoresis of the cDNA amplified b y primer 7-1 3-1 showed bands of the same length as the positive control, m iR-16 (figs. 3.11 and 3.12). For Cal04
#1, the band at 24 hpi was not strong and there may be cDNAs of more than one length. Rather than cutting the DNA bands from the gel and purifying for sequencing, the qPCR products were purified on a column and cloned. Ther efore it was random which size PCR product was inserted into a vector and cloned. Only one clone was sequenced. The sequencing results for the clone that was sequenced indicate that it is a possible miRNA. However, this sequencing method was not ideal becau se it did not determine the origin of the other bands present on the gel and is discussed further in section 3.5.2. Figure 3.13. BLAST results for primer 7-1 3-1. The top matches were to this location in segment 7 in a number of different Influenza A viru ses, including H3 and H9 viruses. However, the primer also matches to Cal04 Cal04 #1 showed a single, strong band at 48 hpi. Ca l04 #2 showed two distinct bands at 24 and 48 hpi. Again, non-specific amplifi cation or amplification of the entire hairpin are possible explanations for these bands. Also, degradation of the viral genome is not random. Areas of the genome that form long s tems, such as hairpin pre-miRNA structures, degrade more rapidly than other areas o f the genome, generating fragments of the viral genome that are at a higher concentration than other fragments. Therefore, the primer may have amplified a degradation product (Br ower-Sinning 2009). Because there was no band present 0 hpi and there was no amplific ation indicated by the qPCR results 0 hpi, it is likely that the second band was amplifie d from virally derived cDNA. If the two cDNA products shown by the gels were present at equ al concentrations, they had an
r equal chance of being cloned and sequenced because the method for cloning and sequencing was simply purifying all the qPCR produc ts and then cloning and sequencing rather than cutting the band from the DNA gel and t hen cloning and sequencing. However, the sequencing results for 24 and 48 hpi f or Cal04 #1 and #2 were identical and complementary to primer 7-1 3-1. To determine the origin of the second band, it would be necessary to repeat the cloning and sequencing u ntil the other cDNA that was present was inserted into the vector and sequenced. Alterna tively, the larger band could be cut from the agarose gel and cloned and sequenced. Figure 3.14. A BLAST of the sequencing results from the cDNA amp lified by primer 7-1 3-1. The results show that the sequence is derived from the viral genome. A BLAST (fig. 3.14) of the sequencing results of th e PCR product amplified by primer 7-1 3-1 confirmed that the sequence was der ived from the H1N1 California 04 genome. To determine human mRNA targets of the poss ible miRNA, two methods were used: a Blast Local Alignment Tool (BLAT) (genome.u csc.edu/cgi-bin/hgBlat) search and TargetScanHuman Custom. BLAT aligns the input m iRNA sequence to any human genomic sequence. A BLAT search of the entire seque nce of the possible miRNA amplified by primer 7-1 3-1 did not show any match es to the Feb. 2009 (GRCh37/hg19) assembly of the human genome.
TargetScanHuman Custom (v5.1 released April 2009) a llows the user to input the seed sequence (nt 2-8) of a novel miRNA that is the n matched to 3' UTRs of mRNAs of the selected species, in this case human. Base-pai ring of the seed sequence is important for miRNA function; the majority of the seed sequen ce must be complementary to the mRNA for the miRNA to effectively bind. It has not been determined what pairing outside of the seed sequence region is necessary fo r miRNAs to function; therefore, seed sequence base pairing is necessary, but may not be sufficient for an effective repression of translation (Chorn et al. 2010). The TargetScan prediction method returned 148 targets that were conserved across several species: human, mouse, rat and dog. The conservation of a miRNA target sequence across several species i s one criterium that is commonly used by computational prediction programs to determ ine whether a certain miRNA acts on a specific mRNA (Friedman et al. 2009). Gene ontology, a method for categorizing genes acco rding to their functions as a protein, was performed on the 148 genes using the D AVID Functional Annotation Tool (v6.7) (david.abcc.ncifcrf.gov/summary.jsp). Of th e 148 genes, 19.6% (p-value 9.8 x 102) of the genes were categorized as involved in dise ases. Specific diseases were not noted in the gene ontology output. Other categories were returned as well, but the general heading of disease was most relevant to this thesis Figure 3.15. MRNA-miRNA base pairing between the possible miRNA amplified by primer 7-1 3-1 and the DCUN1D4 mRNA. The base pair ing between nt 2-8 of the miRNA and the mRNA is the seed sequence.
The sequences of each of the 148 mRNA targets with a matching seed sequence were examined for possible base pairing to the rest of the possible miRNA amplified by primer 7-1 3-1. From the 148 mRNAs, 11 of them had greater than 50% complementarity to the miRNA (fig. 3.15). Gene onto logy on these 11 showed that 36.4% (p-value 7.8 x 10-3) of the mRNAs created proteins that were involved in disease (see A.10 for list of mRNAs). Current prediction programs that match miRNAs with possible target mRNAs use several criteria: the conservation across several s pecies of the region in the 3' UTR that pairs with the seed sequence of the miRNA, the degr ee to which the seed sequence of the miRNA is complementary to the mRNA, the composition of the nearby bases of the mRNA, the secondary structure of the mRNA, and the location in the 3' UTR of the region where the miRNA could bind (Wang and El Naqa 2007). The more bases in the seed sequence that pair with the 3' UTR of the mRNA the more likely it is that mRNA is regulated by that miRNA. The composition of the nea rby nucleotides of the mRNA to the binding site of the miRNA also indicates the likeli hood that it is regulated by that miRNA; AU-rich regions are an indication that the m iRNA regulates that mRNA (Grimson et al. 2007). Also, the secondary structur e of the mRNA affects its regulation by miRNAs; if the binding site for the miRNA is not accessible, that miRNA cannot bind to the mRNA. An additional distinguishing feature o f mRNA-miRNA duplexes is a bulge directly after the seed region (Kertesz et al. 2007 ). The complexity of these criteria make it difficult to accurately predict targets of novel miRNAs, but it is possible that the mRNAs ide ntified by the searches discussed above are targets of the viral miRNA. More research is necessary to validate these results.
The levels of the possible target mRNAs could be mo nitored during infection and during infection when a locking nucleic acid complementary to the possible miRNA is used to prevent the miRNA from acting to silence the mRNAs. Figure 3.16. QPCR results for experiment #1 with Cal04 (Cal04 # 1) primer 7-1 3-1. Primer 7-1 3-1 showed amplification 24 and 48 hpi, but not 0 hpi, indicating that a virally derived RNA was amplified. The relative exp ression levels of miR-16 as compared to primer 7-1 3-1 do not yield informatio n regarding the actual levels of the RNAs that were amplified because of possible nonspe cific amplification and a lack of information about the efficiency of the primers. Ho wever, the relative expression level of 7-1 3-1 shows that there was more RNA that was amp lified by the primer 7-1 3-1 at 24 hpi as compared to 48 hpi. n ######nrr # # # n 4
Figure 3.17. QPCR results for Cal04 #2 primer 7-1 3-1.Primer 7 -1 3-1 showed amplification 24 and 48 hpi, but not 0 hpi, indicat ing that a virally derived RNA was amplified. The relative expression levels of miR-16 as compared to primer 7-1 3-1 do not yield information regarding the actual levels o f the RNAs that were amplified because of possible nonspecific amplification and a lack of information about the efficiency of the primers. However, the relative ex pression level of 7-1 3-1 shows that there was more RNA that was amplified by the primer 7-1 3-1 at 48 hpi as compared to 24 hpi. The qPCR results for Cal04 #1 and #2 primer 7-1 3 -1 (figs. 3.16 and 3.17) did not yield any quantitative information about the po ssible miRNA. This lack of information was because of the lack of standards, l ack of information regarding the efficiency of the primer and amplification, and the presence of multiple PCR products as indicated by DNA gel electrophoresis. The relative quantitation from 24 to 48 hpi mirrored what was shown by miR-16 and by hairpin 71: lower concentrations at 48 hpi as compared to 24 hpi for Cal04 #1 and the reverse for Cal04 #2. Because there was no amplification 0 hpi and amplification 24 and 48 hpi the qPCR results did serve to indicate that this amplicon should be examined clos ely as a real virally-encoded miRNA. n ######nrr # # # n 4
3.2.2 Other Sequencing Results Other qPCR products were also sequenced although t hey did not return any possible miRNAs. These results are discussed below (see A.9 for complete results of all qPCR products that were sequenced). 184.108.40.206 Hairpin 5 Primers 3-1 and 3-2 Hairpin 5 showed consistent data from Cal04 #1 to C al04 #2 and was considered a good candidate as a pre-miRNA hairpin from which viral miRNAs might be cleaved. Figure 3.18. Locations of primers complementary to regions of H airpin 5. The typical locations of mature miRNAs are boxed in red while t he locations of the forward primers are indicated by blue bars with the name of the pri mer above the bar.
Figure 3.19. Sequencing results for cDNAs amplified by primers 5 3-1 and 5 3-2. Hairpin 5 is displayed, and the sequences from the sequencing results are indicated by green bars. The dark green bar indicates the sequen ce amplified by primer 5 3-1 and the dark blue bar indicates the location of primer 5 3 -1. The light green bar indicates the sequence amplified by primer 5 3-2 and the light b lue bar indicates the location of primer 5 3-2. The location where miRNAs are usuall y cleaved from is indicated by a red box. Both sequences continued 9 nt past the end of the hairpin to the end of segment 5. This is not shown in this figure. Figure 3.20. Cal04 #1 DNA gel electrophoresis results for cDNAs amplified by primers 5 3-1 and 5 3-2. There are clear bands 24 and 48 hpi for primers 5 3-1 and 5 3-2 and no band 0 hpi. This indicates that the qPCR amplifi ed virally derived RNA.
n Figure 3.21. Cal04 #2 DNA Gel electrophoresis results for cDNAs amplified by primers 5 3-1 and 5 3-2. The results for Cal04 #2 are ver y similar to those of Cal04 #1 (fig. 3.18) with bands at 24 and 48 hpi and no band at 0 hpi for both primers. Figure 3.22. QPCR results for Cal04 #1 primers 5 3-1 and 5 32. Primers 5 3-1 and 32 showed amplification 24 and 48 hpi, but not 0 hpi indicating that a virally derived RNA was amplified. The relative expression levels a mong different primers do not yield information regarding the actual levels of the RNA that was amplified because of possible nonspecific amplification seen with DNA ge l electrophoresis (fig. 3.20). n # # # # # # # # # nrr # # # n4 4
Figure 3.23. QPCR results for Cal04 #2 primers 5 3-1 and 5 32. Primers 5 3-1 and 32 showed amplification 24 and 48 hpi, but not 0 hpi indicating that a virally derived RNA was amplified. The relative expression level do es not yield information regarding the actual levels of the RNA that was amplified bec ause of possible nonspecific amplification seen with DNA gel electrophoresis (fi g. 3.21). CDNAs generated from reactions with primers 5 3-1 and 5 3-2 (fig. 3.18 and 3.19) appeared to be the same size as the positive control, miR-16, when run on a DNA electrophoresis gel (figs. 3.20 and 3.21) and durin g qPCR showed amplification at 24 and 48 hpi and no amplification at 0 hpi (figs. 3.22 an d 3.23). Despite the promising results based on location wit hin the hairpin, clear, single bands of the proper size on DNA gels, and qPCR resu lts indicating viral origin, the sequences of the cDNAs generated by reactions with primers 5 31 and 5 3-2 do not appear to be mature miRNAs. The cDNAs generated from the reactions with primers 5 3-1 and 5 3-2 continued to the end of the hairpin (fig. 3.19) and a few nucleotides into the genome past the end of the predicted hairpin to the end of segm ent 5. These results indicated that the RNA amplified by the qPCR was merely the end of the hairpin or the end of segment 5 of n # # # # # # # # # nrr # # # 4#(5n6r 4#(56 n4 4
the virus. While the viral genomic segments should have been too large to be purified in the small RNA fraction, degradation of the segments does occur and there may have been a fragment that included the end of the genome that was amplified by the qPCR primer set near the end of the genome. Autopriming could c ause this amplification; the ends of genome segments can loop back on themselves forming a double-stranded region that could have acted as a reverse primer and amplified that area along with the user-designed forward primer. It is also possible that hairpin 5 exists and serves some purpose in the viral life cycle, but it doesnt appear that it gen erates mature miRNAs. The RNA amplified by primer 5 3-1 was 36 bp long, which is beyond the length of mature miRNAs. The RNA amplified by primer 5 3-2 was 29 b p long, also much longer than a mature miRNA. Previously completed, unpublished data from the Har rod Lab from an identical experiment with A/NewCaledonia/20/99 (H1N1) showed that NewCaledonia20 encodes a predicted hairpin in the same location. Primers w ere designed for the hairpin, but sequencing results returned results identical to th e ones for California04: amplification of RNA too long to be a miRNA. The sequencing results included the entire end of the predicted pre-miRNA hairpin and the end of the geno me (Jesse VanWestrienen, personal communication). It appears there is a predicted pre -miRNA-like structure in both strains, but its unknown if the potential pre-miRNA hairpin is cleaved from the viral RNA since both sequencing results showed that the hairpin was intact with the end of the genomic sequence.
r 220.127.116.11 Hairpin 7-1 Primer 7-1 3-2 In addition to the previously discussed cDNA genera ted from the qPCR with primer 7-1 3-1, qPCR with primer 7-1 3-2 showed p romising DNA gel electrophoresis and qPCR results, but it was determined that it was not a viral miRNA based on the sequencing results. Figure 3.24. Sequencing results for cDNA amplified by primer 71 3-2. Hairpin 7-1 is displayed, and the sequence from the sequencing res ults is indicated by a green bar. The primer location is indicated by a blue bar, and the location where miRNAs are usually cleaved from is indicated by a red box.
r Figure 3.25. QPCR results for Cal04 #1 Primer 7-1 3-2. There i s a small bar present at 0 hpi. Because of the scale of the graph it is not vi sible, but the small amount of amplification present indicates that the reaction w ith primer 7-1 3-2 amplified cellular or nonspecific RNA. However, the much greater amplific ation at 24 hpi indicates that primer 7-1 3-2 may also have amplified viral RNA. Amplification at 48 hpi was low. These results differ from those seen for Cal04 #2 p rimer 7-1 3-2 (fig. 3.26). Figure 3.26. QPCR results for Cal04 #2 primer 7-1 3-2. Unlike Cal04 #1 primer 7-1 32 (fig. 3.25) there is no amplification 0 hpi for p rimer 7-1 3-2. There is amplification 24 and 48 hpi, indicating the amplification of viral R NA. n ######nrr # # # n 4 n # # # # # # nrr # # # n 4
r The qPCR reaction with primer 7-1 3-2 generated cD NAs from 24 and 48 hpi RNA but not 0 hpi from Cal04 #2 (figs. 3.9, 3.24-3. 26). The cDNA was sequenced and the results returned sequence of the proper size to be a mature miRNA. The sequence was complementary to primer 7-1 3-2, giving a total le ngth of 19 nt. This is slightly shorter than the 20-22 nt length of mature miRNAs. Also, th is miRNA would overlap with the possible miRNA amplified by primer 7-1 3-1 (discus sed in section 3.2.1). If miRNAs were cut from overlapping locations within the same hairpin it would not be an efficient use of the hairpin because the same hairpin could n ot be used to generate the two miRNAs. Primer 7-1 3-2 covered the end of the pred icted hairpin and therefore could have been binding to and amplifying the end of the hairpin, not a cleaved miRNA. Figure 3.27. BLAST results for primer 7-1 3-2. The results wer e all 100% matches to segment 7 of various H1N1 strains. This, along with ePCR results, means that it is likely that nonspecific amplification of cellular RNA is r esponsible for the weak signal for Cal04 #1 0 hpi. The qPCR results for Cal04 #2 primer 7-1 3-2 were not replicated by Cal04 #1. Cal04 #1 primer 7-1 3-2 showed low levels of ampli fication at 0 hpi during qPCR (fig. 3.25). For this reason, the cDNA from this first ex periment was not run on a gel and was not sequenced. It is unclear why this discrepancy e xists between Cal04 #1 and #2. The amplification at 0 hpi in Cal04 #1 was much less th an the amplification 24 and 48 hpi. It is possible that nonspecific amplification of cellu lar RNA occurred at 0 hpi while specific amplification of viral RNA occurred at 24 and 48 hp i, accounting for the large difference
r in relative expression levels between 0 and 24 and 48 hpi. Nonspecific amplification of cellular RNA is the most likely explanation for the amplification at 0 hpi because the BLAST results (fig. 3.27) do not indicate that the primers were complementary to parts of the human genome. A reverse ePCR search of the huma n transcriptome (Dataset  Homo sapiens transcriptome snapshot 2010/10/29) wit h the settings on the least stringent, 2 gaps and 2 mismatches allowed, and an allowed PCR product size deviation of 50 bp, did not indicate complementarity to the human trans criptome. Figure 3.28. Cal04 #2 DNA Gel electrophoresis results for cDNAs amplified by primer 7-1 3-2. The primers are visible at the bottom of the gel. The bands of varying lengths from primers 7-1 3-1 and 7-1 3-2 24 and 48 hpi in dicate amplification of multiple cDNAs during qPCR. Both Cal #2 7-1 3-1 and 3-2 sh ow amplification at 24 and 48 hpi and no amplification 0 hpi. The DNA gel electrophoresis results for Cal04 #2 pr imer 7-1 3-2 (fig. 3.28) showed multiple bands at 24 and 48 hpi and no band 0 hpi. At 24 and 48 hpi, there was a band showing cDNA similar in length to miR-16, the positive control, but there were other bands indicating the presence of longer cDNA as well. At 24 hpi, the longer band was approximately 140 bp. This is too long to be th e primer binding to and amplifying the hairpin. It is possible that the primer bound t o a fragment of the genome and amplified that.At 48 hpi the longer band is approxi mately 100 bp. Again, because primer
r 7-1 3-2 sits at the end of the hairpin, this is al so too long to be caused by amplification of the hairpin. The degradation of the viral genome th at occurs naturally is non-random; the areas that contain hairpin structures are more susc eptible to degradation and may therefore be present at higher concentrations than other degradation products. The nonrandom degradation could create fragments at a high enough concentration to be amplified by qPCR (Brower-Sinning 2009). The seque ncing results (Fig. 3.24) indicated that the virally derived cDNA similar in length to miR-16 was inserted into the vector and cloned from the small band. In order to determi ne what the longer bands are, the cloning and sequencing would have to be repeated us ing the longer band cut from the DNA gel and cloned and sequenced. As with primer 71 3-1, the qPCR results provided little quantitative information (figs. 3.25 and 3.2 6), but did give confirmation that research on this specific primer amplicon should co ntinue for Cal04 #2. 18.104.22.168 Hairpin 7-2 Although multiple cDNAs from hairpin 7-2 were ident ified as being generated from possible miRNAs through qPCR and DNA gel elect rophoresis, the sequencing results determined that these cDNAs were not from m ature miRNAs. Figure 3.29. Locations of primers complementary to regions of H airpin 7-2. The typical locations of mature miRNAs are boxed in red while t he locations of the forward primers are indicated by blue bars with the name of the pri mer above the bar. Primers 5-1 and 35 continue past the ends of the predicted hairpin.
r Figure 3.30. Sequencing results for cDNAs amplified by primers 7-2 3-1, 7-2 3-4, and 7-2 3-5. Hairpin 7-2 is displayed, and the sequenc ing results are indicated by green bars. The dark green bars indicate the sequences amplifie d by primers 7-2 3-1 and 7-2 3-5 and the dark blue bars indicate the locations of pr imers 7-2 3-1 and 7-2 3-5. The light green bar indicates the sequence amplified by prime r 7-2 3-4 and the light blue bar indicates the location of primer 7-2 3-4. The loca tion where miRNAs are usually cleaved from is indicated by a red box. The sequences of 72 3-4 and 7-2 3-5 continued past the end of the hairpin to the end of segment 7-2. The c DNAs generated from three separate qPCR reactions were cloned and sequenced. These wer e from primers 7-2 3-1, 7-2 3-4, and 7-2 3-5 (figs. 3.29 and 3.30).
r Figure 3.31. Cal04 #1 DNA gel electrophoresis results for cDNAs amplified by primer 7-2 3-1. Because of the poor quality of the gel im age, the gel image was inverted and the contrast was increased. The gel was imaged on the s ame equipment using the same procedure as all the other gels. There are bands pr esent at 24 and 48 hpi but not at 0 hpi for Cal #1 7-2 3-1, indicating that the qPCR with primer 7-2 3-1 amplified RNA of viral origin. Figure 3.32. Cal04 #1 DNA gel electrophoresis results for cDNAs amplified by primer 7-2 3-4. Because of the poor quality of the gel im age, the gel image was inverted and the contrast was increased. The gel was imaged on the s ame equipment using the same procedure as all the other gels. There are bands pr esent at 24 and 48 hpi but not at 0 hpi for Cal #1 7-2 3-4, indicating that the qPCR with primer 7-2 3-4 amplified RNA of viral origin.
rn Figure 3.33. Cal04 #1 DNA gel electrophoresis results for cDNAs amplified by primer 7-2 3-5. There are bands present at 24 and 48 hpi but not at 0 hpi for Cal #1 7-2 3-5, indicating that the qPCR with primer 7-2 3-5 ampli fied RNA of viral origin. Figure 3.34. Cal04 #2 DNA gel electrophoresis results for cDNAs amplified by primers 7-2 3-1, 3-4, and 3-5. There are bands present a t 24 and 48 hpi and no bands present at 0 hpi for primers 7-2 3-1 and 7-2 3-5, indicating that the qPCR amplified RNA of viral origin. For primer 7-2 3-4, there are no bands at 0 and 24 hpi, but there is a band at 48 hpi. This may indicate viral RNA that is at such a low concentration that it is not present at a concentration high enough to be efficiently re verse transcribed to form cDNA that is amplified by the qPCR.
r Figure 3.35. The top BLAST result for the sequencing results of the cDNA amplified by primer 7-2 3-1. Although the top BLAST result indi cates 100% identity to segment 7 of a different strain of the virus, the sequence is al so 100% identical to the 2009 Cal04 strain used in this research. The DNA gel electrophoresis results (figs. 3.31-3.3 4) were fairly consistent between Cal04 #1 and #2. Primer 7-2 3-1 showed ban ds of the appropriate length (figs. 3.31 and 3.34) at 24 and 48 hpi and no band at 0 hp i. The sequencing results revealed that the primer was likely amplifying a fragment of the viral genome (fig. 3.35). The length of the RNA was between 26 and 28 nt (results varied sl ightly between 24 and 48 hpi), too long to be a mature miRNA. Also, only one of the th ree qPCR replicates in the Cal04 #1 qPCR showed amplification (fig. 3.38). This indicat es that a very low concentration of the target RNA was present, a concentration so low that the reverse transcription reaction performed prior to qPCR did not work at a high enou gh efficiency to generate cDNA that could be amplified by qPCR. Figure 3.36. The top BLAST result for the sequencing results of the cDNA amplified by primer 7-2 3-4. Although the top BLAST result indi cates 100% identity to segment 7 of a different strain of the virus, the sequence is al so 100% identical to the 2009 Cal04 strain used in this research.
r Primer 7-2 3-4 amplified a cDNA of the appropriate length (Figs. 3.31 and 3.33) although this band was only present at 48 hpi for C al04 #2. The cloning reaction and subsequent sequencing did not work for Cal04 #1 24 hpi, which may have been caused by a low concentration of the qPCR product. This ma y also be the reason that no band was visible for Cal04 #2 at 24 hpi. The sequencing results from Cal04 #1 and #2 48 hpi indicated an RNA of 31 nt present at 48 hpi. This i s too long to be a mature miRNA. As further evidence that this is not a mature miRNA, t he entire end of the hairpin was included in the sequence (figs. 3.30 and 3.36). Thi s suggests that the primer annealed to the entire hairpin or a fragment of the genome, rat her than a specific miRNA. Figure 3.37. The top human BLAST result for the cDNA amplified by primer 7-2 3-5. Although the results indicate 100% identity with a Mexico City strain of the 2009 H1N1 influenza, the results are also 100% identical to t he Cal04 strain used in this research. Hairpin 7-2 3-5 also does not appear to be a matur e miRNA. This was predicted by the DNA gel electrophoresis results (figs. 3.33 and 3.34) as they differed slightly from Cal04 #1 to Cal04 #2. Cal04 #1 showed bands at 24 a nd 48 hpi slightly longer than miR16 while Cal04 #2 showed bands of the same approxim ate length as miR-16. The sequencing results confirmed that while both were v irally derived, the length was different. The qPCR product from Cal04 #1 24 and 48 hpi was 33 bp long (fig. 3.37) and included part of the viral genome past the end of t he predicted hairpin, while the qPCR product from Cal04 #2 24 and 48 hpi was 21 bp long and was cut off before the end of
rr the hairpin. Had the results been consistent betwee n Cal04 #1 and #2, 7-2 3-5 could be considered a possible miRNA. However, the variation between the two replicates with one yielding a product far too long to be a mature miRNA, indicates that this is not a mature miRNA. Additionally, the RNA amplified by pr imer 7-2 3-5 was not in the typical location on the pre-miRNA hairpin for a mat ure miRNA. Figure 3.38. QPCR results for Cal04 #1 primers 7-2 3-1, 7-2 3 -4, and 7-2 3-5. The results show amplification from primer 7-2 3-4 at 48 hpi and from primer 7-2 3-5 at 24 and 48 hpi with no amplification at 0 hpi. Unlike C al04 #2 (fig. 3.39) primer 7-2 3-1 did not amplify anything. The asterisk by 24 and 48 hpi for primer 7-2 3-1 indicate that one of the three qPCR replicates did show amplification but two did not. n # # # # # # # # # # # # nrr # # # n4 44 77
Figure 3.39. QPCR results for Cal04 #2 for primers 7-2 3-1, 72 3-4, and 7-2 3-5. The results show amplification 48 hpi for primers 7-2 3 -1 and 7-2 3-4 and amplification 24 and 48 hpi for primer 7-2 3-5 with no amplificatio n 0 hpi. These results indicate that the amplified cDNA was from virally derived RNA. The as terisk by 24 hpi for primer 7-2 31 indicates that one of the three qPCR replicates d id show amplification while the other two did not. The qPCR results do not reveal any additional info rmation regarding hairpin 7-2 3-1, 3-4, or 3-5 (figs. 3.38 and 3.39). As noted above, only one of the three replicates within the Cal04 #1 qPCR reaction showed amplificat ion for 24 and 48 hpi for primer 31. Because the Ct values were averaged to give the values on the graph below, 24 and 48 hpi show 0 as the relative expression level. A simi lar situation occurred for Cal04 #2 3-1 24 hpi. For 48 hpi, two of the three replicates sho wed amplification so the average Ct value was obtained. The qPCR results were consisten t with the DNA gel electrophoresis results. If one replicate showed amplification, a b and was visible on the DNA gel. If another replicate from the same primer that did not show amplification was run on a DNA gel, no band was visible. The work discussed above identified one possible m ature miRNA, the RNA amplified by primer 7-1 3-1. Multiple other possib le miRNAs from predicted hairpins 5, n # # # # # # # # # # # # nrr n4 44 7
7-1, and 7-2 were cloned and sequenced following qP CR and DNA gel electrophoresis. After the sequencing results were examined, most po ssible miRNAs were eliminated for a variety of reasons. If the sequencing results ret urned a sequence longer or shorter than the typical length of miRNAs, it was no longer cons idered a possible mature miRNA. Also, if the sequencing result included nucleotides past the end of the predicted premiRNA hairpin, that RNA fragment was eliminated. 3.2.3 Elimination of Possible miRNAs The previous sections described the results for the possible mature miRNA that was identified (section 3.2.1) and of possible miRN As that were determined not to be miRNAs after sequencing (section 3.2.2); this secti on will discuss possible miRNAs that were eliminated from further consideration early in the study. QPCR served as the first step to eliminating possible miRNAs. If a certain p rimer amplified something at 24 and 48 hpi, but not 0 hpi, the qPCR product was run on a DNA gel. If a primer amplified something at 0 hpi or did not amplify anything 0, 2 4, or 48 hpi, it was not run on a gel. If DNA gel electrophoresis of a qPCR product showed pr oducts of a drastically different size than miR-16 then it was eliminated from furthe r study.
Figure 3.40. QPCR results for Cal04 #1 primer 7-1 5-1. The res ults show a relatively high experession level at 0 hpi when compared to 24 and 48 hpi for primer 7-1 5-1. Any amplification at 0 hpi indicates that cellular RNA was amplified. This meant that the cDNA amplified by this primer was considered cellul ar and so no further work was done with the qPCR product. Figure 3.41. QPCR results for Cal04 #2 primer 7-1 5-1. The res ults show amplification of something at 0 hpi as well as 24 and 48 hpi by p rimer 7-1 5-1. While the relative expression level 0 hpi is not as large compared to 24 and 48 hpi as it was for Cal04 #1 (fig. 3.40) any expression at 0 hpi, especially wit h consistent results between Cal04 #1 and #2 indicated amplification of cellular RNA, rat her than viral and eliminated that qPCR product from further work. n # # # # # # nrr # # # n 4 6 6 6 6 # # # # # # nrr # # # n 4
If a certain primer amplified something at 0 hpi, i t was eliminated from the process (figs. 3.40 and 3.41). Amplification of 0 h pi RNA indicated that a cellular RNA had been amplified because there was no viral RNA p resent 0 hpi because the cells had not yet been exposed to the virus. Figure 3.42. Cal04 #1 DNA gel electrophoresis results for cDNAs amplified by primers 7-3 5-1 and 7-3 5-2. The multiple bands of cDNA c onsiderably longer than the desired size (demonstrated by miR-16) eliminated the cDNA a mplified by primers 7-3 5-1 and 7-3 5-2. The primers are visible at the bottom of the gel under Cal #1 7-3 5-2. If DNA gel electrophoresis revealed multiple cDNAs indicating nonspecific amplification, or bands that are too long to indica te a cDNA copy of a miRNA, the reaction was eliminated from further studies (fig. 3.42). Also, if the 24 and 48 hpi showed markedly differently sized cDNAs, the primer was el iminated from further studies. 3.3 The Attempt to Decrease Expression Levels of Di cer Dicer is a protein essential to the processing of miRNAs. It completes the final step in miRNA processing by cleaving the miRNA from the pre-miRNA hairpin (Lima et al. 2009). In order to decrease the expression leve ls of Dicer, a shRNA was going to be transfected into the cells and then processed by Di cer to yield a siRNA that would bind in the 3 UTR of the Dicer mRNA and prevent its transl ation, decreasing the levels of
functional Dicer in the cell (Kumar et al. 2007). T his experiment was supposed to test whether the predicted miRNAs required Dicer activit y to be formed. Figure 3.43. Fluorescent microscopy images of 293T cells after transfection with GIPZ which causes the cells to express green fluorescent protein (GFP). GFP shows up green under a fluorescent microscope. The left image show s cells 24 hours after transfection while the right shows cells 48 hours after transfec tion. It appears that there is slightly less fluorescence 48 hours after transfection, likely ca used by the rapid growth and subsequent death of the 293T cells. Figure 3.44. Fluorescent microscopy images of adenocarcinomic h uman alveolar basal epithelial (A549) cells after transfection with GIP Z which causes the cells to express GFP. From left to right 12, 36, and 72 hours post t ransfection. When compared to fig. 3.40, the efficiency of transfection was very low. This is likely because the A549 cells are much more difficult to transfect. The greatest amou nt of fluorescence was seen 36 hours after transfection. The first step of this process was the transfection As discussed in the section 2.5, the efficiency of the transfection was monitored us ing GIPZ which causes the transfected cells to express Green Fluorescent Protein (GFP). T his transfection was tested on Human Embryonic Kidney (293T) cells. The 293T cell line i s known to be more easily
transfected than most other cell lines. The transfe ction was much more efficient in the 293T cells than the A549 cells that had been used f or the infection studies (figs. 3.43 and 3.44). Figure 3.45. QPCR results for the expression levels of Dicer fo llowing transfection of the cells with a shRNA that was designed to decreas e the expression levels of Dicer. Dicer indicates cells that were transfected with th e Dicer shRNA and GIPZ indicates cells that were transfected with GIPZ, which causes cells to express GFP. GIPZ was a negative control. There is no data for 72h GIPZ cells. At 48 h the relative expression level of Dicer was less in the cells transfected with GIPZ which s hould not affect the levels of Dicer than in the cells transfected with the shRNA that s hould reduce the expression levels of Dicer. The shRNA that targeted Dicer was tested in 293T ce lls using qPCR (fig. 3.45). At 48 hours post transfection (hpt), the expression levels of Dicer were not affected. In fact, it appears that the cells transfected with th e Dicer shRNA had higher levels of Dicer than the cells transfected with GIPZ. At 72 hours p ost transfection (hpt) it was impossible to determine if the expression levels of Dicer were affected by the shRNA. This is because the 72 hpt cells that had been transfected with GIPZ stopped adhering to the cell culture well and were removed when the media was re moved. It does appear, when compared to 48 hpt, that the expression level of Di cer was decreased 72 hpt. However, it 6 6 6n 6 6 ####7nrr"#r$ %nrrr&r ,8$9,8$9
n also appeared that the GIPZ transfected cells had a decreased expression level of Dicer 48 hpt as compared to the cells transfected with the D icer shRNA. This makes it difficult to determine whether the decrease seen 72 hpt represen ts an actual decrease in the expression levels of Dicer. Because the transfection efficiency appeared to be very low in the A549 cells, and because it was unclear whether the Dicer shRNA redu ced the expression levels of Dicer, the transfection was conducted in conjunction with the infection, but qPCR was not run. 3.4 Future Work This thesis research identified one possible matur e miRNA, a miRNA amplified by primer 7-1 3-1. In order to verify this as an m iRNA, some information regarding its processing or efficacy in the cell is necessary. Ad ditionally, few papers are published with qPCR and sequencing as the only proof of the e xistence of a miRNA. In section 1.5.2 some commonly published methods for proving t he existence of an miRNA are discussed. These methods include northern blotting (Vloczi et al. 2004), highthroughput sequencing (Dohm et al. 2008), and micro arrays (Babak et al. 2004). Additionally, papers generally include information about the role of the miRNA in the cell or demonstrate that the miRNA is processed by Dicer and Drosha to prove that it is an actual miRNA (Vloczi et al. 2004) To gain information regarding the processing of the miRNA, a transfection of a shRNA that would decrease the expression levels of Dicer was attempted. However, this transfection did not work for several reasons that will be discussed in more detail in section 3.5.2. A transfection of a siRNA that targe ted Dicer or the transfection of an
shRNA or siRNA that targeted a protein involved in a different stage of the processing of miRNAs would help determine how the possible viral miRNA is processed. An additional method to verify the existence of th e possible miRNA would be to show its efficacy. Methods for showing the efficacy of miRNAs include using locked nucleic acids (LNAs) that are complementary to the miRNA of interest. The LNA anneals to the miRNA, blocking its function. Then, changes that appear in the cells are monitored with and without the locked nucleic acid in order to determine if the possible miRNA has an effect on the cell and/or virus. For t he virally derived miRNA it would be possible to monitor basic changes in replication ca used by the presence or lack of the possible miRNA by using a locked nucleic acid and t hen collecting RNA samples at various time points after the delivery of the locke d nucleic acid and quantitating the amount of virus present at each time point (Vloczi et al. 2004). 3.5 Improvements Possible improvements to this project fall into tw o basic categories: improvements to techniques and improvements in the overall procedure and planning of the experiment. In terms of overall technique, many of the errors were due to inexperience in the laboratory. As the thesis resea rch progressed, techniques improved. 3.5.1 Basic Technique The concentration of cDNA was not normalized among 0, 24, and 48 hpi because it is possible that the infection affected the numb er of cells or amount of RNA that was present. In other words, it is possible that later in infection (at 48 hpi for example) there
was less cellular RNA than at 0 hpi because of cell death caused by the infection. This meant that normalizing the concentration of cDNA us ed in qPCR, which should be directly proportional to the concentration of RNA, could have introduced some bias. Instead the number of cells/well and the number of virions added to each well were normalized within Cal04 #1 and Cal04 #2. When the cells for Cal04 #1 and #2 were plated, dif ferent numbers of cells were plated each time. The cells were then allowed to gr ow until confluent, but it would have been more consistent to plate the same number of ce lls and then allow the cells to grow for the same time period. The number of virions use d in the infection was calculated based on a multiplicity of infection of one and dep ended on the original number of cells plated. This was also inconsistent because the cell s were allowed to grow until confluent. This meant that by the time they were infected, the re was approximately the same number of cells for each infection study, but the n umber of virions used was different. For other reasons, which will be discussed later, t he quantitation of the small RNAs did not work. This means that the varied level s of cells likely had little affect on the results and data. Other errors only affected the presentation of the data. Many of the agarose electrophoresis gels were too dark and lacked adequ ate separation of the ladder. The appearance of these gels could have been improved b y allowing the gels to run for a longer period of time, thereby increasing the separ ation of the DNA. The background staining could have been reduced by placing the gel s in water after they were run and rocking gently for 10-15 minutes. This would have r emoved the loading dye, which would eliminate the dark bands that occasionally ob scured the DNA bands. In all cases,
r the DNA gels provided sufficient information for de termining whether or not to continue and clone and sequence a particular cDNA, but the q uality of the gels was not as high as is expected for scientific journal articles. 3.5.2 Procedure The qPCR results did not correlate with the existe nce of mature miRNAs and did not serve to quantitate the levels of the RNAs that were amplified. This aspect of the qPCR failed for several reasons. First, primer-dime rs formed. This was due to the user designed forward primers that had to fit in certain locations along the hairpin. Using probes on the primers rather than SYBR green detect ion would have eliminated any fluorescence due to primer-dimers although this app roach would have been more expensive. In addition to the primer-dimers, non-sp ecific amplification occurred. This can be seen on the DNA gels that contain multiple b ands, indicating multiple RNAs were amplified. Also, the results among qPCR replicates were sometimes inconsistent. This was because of the very low concentration of small RNAs that were subject to an inefficient tailing reaction to generate cDNAs. The last two problems, non-specific amplification and inconsistent results, could possibly be fixed by using a different method to ge nerate cDNA from the RNA. A number of techniques were discussed in section 1.5. 2. One that would be a simple modification to the procedure is the tailing of the RNAs with a hairpin rather than a linear tail (Chen et al. 2005). Using a hairpin has severa l advantages over a linear tail. These advantages are discussed in detail in section 1.5.2 .7. One particularly relevant advantage, given some of the sequencing results and the non-sp ecific amplification seen on the DNA
gels, is that the hairpins hybridize with prior p re-miRNAs at a much lower rate than the linear tails do. The hairpins also hybridize to mat ure miRNAs with a greater efficiency than the linear tails (Chen et al. 2005). This woul d improve the consistency among identical qPCR replicates. When non-specific amplification occurred, it would have been necessary to use a different approach for cloning and sequencing to de termine the origin of the qPCR products. When multiple bands were observed in the DNA gels, the different bands could have been cut from the DNA gel and then the DNA fro m each band could have been cloned and then sequenced. Alternatively, because t he sequencing was random, more clones could have been sequenced until all the qPCR products had been identified. After identifying a possible miRNA, an attempt was made to decrease the expression levels of Dicer, a protein essential for processing cellular miRNAs. This attempt failed for several reasons. First, the tran sfection efficiency of RNAs into adenocarcinomic human alveolar basal epithelial (A5 49) cells, as measured by the fluorescence induced by the successful transfection of GIPZ into the A549 cells, was very low. The transfection had a much higher efficiency in Human Embryonic Kidney (293T) cells. The 293T cells are known for being easily tr ansfected. In order to achieve a higher transfection efficiency in the A549 cells, the amou nt of the reagents used would need to be varied until the optimum amount of transfection reagent and GIPZ was obtained. Another problem arose from the attempt to decrease the expression levels of Dicer using a shRNA that must be processed by Dicer to be functional. This problem could be remedied by transfecting a mature siRNA into the ce ll. However, Kumar et al. (2007) showed that the shRNA that was used in this experim ent was effective at reducing the
levels of Dicer by 80% in mouse lung adenocarcinoma l cells. This result indicates that the Dicer shRNA can effectively reduce the expressi on levels of Dicer in certain cell types. While Dicer processes pre-miRNAs to form mature miR NAs, there are other proteins involved in miRNA processing. The other pr oteins involved, including Drosha, the protein responsible for cleaving pri-miRNA to f orm pre-miRNA hairpins (Zeng et al. 2005), are also possible targets for siRNAs that co uld generate information regarding the origin and processing of the possible mature miRNA identified by the work of this thesis. Once the expression levels of Dicer were decreased, it would be necessary to use a technique that allowed for accurate quantitation of the levels of the small RNAs. Otherwise, the decrease in expression levels of Dic er would not yield any information regarding the levels of the small RNAs unless they completely eliminated them. This is also true if another protein that processes miRNAs was targeted using an shRNA or siRNA.
Chapter 4: Conclusions Through methods of qPCR, DNA gel electrophoresis, and sequencing, a potential virally derived miRNA was identified. While miRNAs encoded by DNA viruses have been found, there is no published data indicating t hat RNA viruses encode miRNAs. The evidence presented in this thesis identified a vira lly derived RNA approximately the size of a mature miRNA that is potentially derived from a predicted pre-miRNA hairpin. To identify this as a miRNA it would be necessary to d emonstrate a role in the viral life cycle, either targeting cellular or viral genes. Ho wever, this identified small RNA likely serves some role for the virus even if it does not function as a miRNA. It is present at higher levels during the infection process than a p roduct of the degradation of the viral RNA would be, indicating that may function in some way during the viral life cycle. The implications of a mature miRNA encoded by Influ enza A, a RNA virus, are huge. The more that is understood about how the vir us replicates and infects cells, the greater our ability to predict pandemics and predic t which strains will be particularly virulent and damaging to the world health and econo my. Also, if other RNA viruses such as Human Immunodeficiency Virus (HIV) encode miRNAs the miRNAs may be used as potential targets of antiviral therapies. Although there is more work to conduct to confirm the existence of the potential miRNA identified by this thesis, now there are defi nitive targets to work towards. The sequence and location of a potential miRNA have bee n identified, making it much easier to determine the role of this miRNA in the viral li fe cycle.
Appendix A.1 Sequence and diagram of Cal04 hairpins and subh airpins Hairpin 5: TCGTGCCTTCCTTTGACATGAGTAATGAAGGGTCTTATTTCTTCGGAGACA AT GCAGAGGAGTATGA Hairpin 7-1: TTGGTCTAGTGTGTGCCACTTGTGAACAGATTGCTGATTCACAGCATCGGT CT CACAGACAGATGGCTACTACCACCAA Hairpin 7-2: AGGCAGCGGAGGCCATGGAGGTTGCTAATCAGACTAGGCAGATGGTACATG C AATGAGAACTATTGGGACTCATCCTAGCTCCAGTGCTGGTCTGAAAGATGA C CTTCTTGAAAATTTGCAGGCCT
Hairpin 7-3: GATCGTCTTTTTTTCAAATGTATTTATCGTCGCTTTAAATACGGTTTGAAA AG AGGGCCTTC A.2 Sequence of oligoDT adaptor TTTTTTTTTTTGGATATCACTCAGCATAATTAAGACACGAGCGT A.3 Sequence of Universal Reverse Primer 5ACGCTCGTGTCTTAATTATGCTGAGTGATATCC -3 Tm = 60.8 A.4 Forward primers for predicted hairpins in Californi a 04 Hairpin 5, located 1471-1537 nt from the start of s egment 5 Primer Name Sequence Tm (oC) Length (nt) Location within Segment 5 5-1 CTT CCT TTG ACA TGA GTA ATG AA 50.8 23 1476-1499 5-2 GAC ATG AGT AAT GAA GGG TCT T 52.1 22 1485-1507 3-1 CTT CGG AGA CAA TGC AGA 52.0 18 1511-1529 3-2 GAC AAT GCA GAG GAG TAT GA 51.7 20 1517-1537 Hairpin 7-1, located 478-556 nt from the start of s egment 7 Primer Name Sequence Tm (oC) Length (nt) Location within segment 7 5-1 GTC TAG TGT GTG CCA CTT 51.9 18 481-499 5-2 TGC CAC TTG TGA ACA GAT 51.6 18 491-509 5-3 GAA CAG ATT GCT GAT 50.6 20 500-520
TCA CA 5-4 GAT TGC TGA TTC ACA GCA T 50.8 19 505-524 3-1 TCG GTC TCA CAG ACA GAT 52.4 18 523-541 3-2 GAT GGC TAC TAC CAC CAA 51.2 18 538-556 Hairpin 7-2, located 648-765 nt from the start of s egment 7 Primer Name Sequence Tm (oC) Length (nt) Location within Segment 7 5-1 GAG GCC ATG GAG GTT 51.0 15 654-669 5-2 TTG CTA ATC AGA CTA GGC A 50.5 19 669-688 5-3 GCA GAT GGT ACA TGC AAT 50.2 18 684-702 5-4 TGG TAC ATG CAA TGA GAA CTA T 51.7 22 690-712 3-1 AAC TAT TGG GAC TCA TCC TAG 50.9 21 706-727 3-2 ATC CTA GCT CCA GTG CT 52.4 17 720-737 3-3 AGT GCT GGT CTG AAA GAT 50.7 18 730-748 3-4 CTG AAA GAT GAC CTT CTT GAA AAT T 52.0 25 739-764 3-5 GAC TTC TTG AAA ATT TGC AGG 50.2 21 744-765 Hairpin 7-3, located 865-926 nt from the start of s egment 7 Primer Name Sequence Tm (oC) Length (nt) Location within Segment 7 5-1 GAT CGT CTT TTT TTC AAA TGT ATT TAT CG 52.2 29 865-894 5-2 GTA TTT ATC GTC GCT TTA AAT ACG G 52.2 25 884-909 3-1 AAT ACG GTT TGA AAA GAG GG 50.4 20 900-920
n A.5 Primers for California 04 hairpins Primer Name Sequence Tm (oC) Length (nt) Location on hairpin Amplicon Size (nt) Forward 5 GTG CCT TCC TTT GAC ATG AG 53.8 20 2-22 61 Reverse 5 ACT CCT CTG CAT TGT CTC 51.4 18 45-63 61 Forward 7-1 GTC TAG TGT GTG CCA CTT 51.5 18 4-22 69 Reverse 7-1 GTA GTA GCC ATC TGT CTG TG 51.9 20 53-73 69 Forward 7-2 TAG GCA GAT GGT ACA TGC 51.2 18 35-53 45 Reverse 7-2 GCT AGG ATG AGT CCC AAT AG 51.5 20 60-80 45 Forward 7-3 TTT CAA ATG TAT TTA TCG CT 45.1 20 12-32 40 Reverse 7-3 TTT CAA ACC GTA TTT AAA GC 46.6 20 32-52 40 A.6 Sequence of miR-16 forward primer 5TAGCAGCACGTAAATATTGGCG -3 Tm = 58.9, amplicon size = 68 bp
A.7 Vector Information
Figures adapted from Qiagen PCR Cloning Handbook
r A.8 Primers for Dicer (product is 194 nt) Primer Name Sequence Tm ( o C) Length (nt) Dicer forward CAGTCCCTAGGATGATGTAG 50.9 20 Dicer reverse CTGTCCCTTACTATCAGGTG 51.4 20 A.9 Sequencing Results for all small, virally encod ed RNAs Primer Name and Infection Study # Sequencing Result Cal #1 5 3-1 24h TCTTCGGAGACAATGCAGAGGAGTATGACAGTT GAGG Cal #1 5 3-1 48h TCTTCGGAGACAATGCAGAGGAGTATGACAGTT GAGG Cal #2 5 3-1 24h TCTTCGGAGACAATGCAGAGGAGTATGACAGTT GAGG Cal #2 5 3-1 48h TCTTCGGAGACAATGCAGAGGAGTATGACAGTT GAGG Cal #1 5 3-2 24h GACAATGCAGAGGAGTATGACAGTTGAGG Cal #1 5 3-2 48h GACAATGCAGAGGAGTATGACAGTTGAGG Cal #2 5 3-2 24h GACAATGCAGAGGAGTATGACAGTTGAGG Cal #2 5 3-2 48h GACAATGCAGAGGAGTATGACAGTTGAGG Cal #1 7-1 3-1 24h CGGTCTCACAGACAGATGGC Cal #1 7-1 3-1 48h CGGTCTCACAGACAGATGGC Cal #2 7-1 3-1 24h CGGTCTCACAGACGATGGC Cal #2 7-1 3-1 84h CGGTCTCACAGACGATGGC Cal #2 7-1 3-2 24h GATGGCTACTACCACCAA Cal #2 7-1 3-2 48h GATGGCTACTACCACCAA Cal #1 7-2 3-1 24h AACTATTGGGACTCATCCTAGCTCC Cal #1 7-2 3-1 48h AACTATTGGGACTCATCCTAGCTCCAGT Cal #2 7-2 3-1 48h AACTATTGGGACTCATCCTAGCTCCAGT Cal #1 7-2 3-4 48h CTGAAAGATGACCTTCTTGAAAATTTGCAGG Cal #2 7-2 3-4 48h CTGAAAGATGACCTTCTTGAAAATTTGCAG G Cal #1 7-2 3-5 24h TGTCTTCTTGAAAATTTGCAGGCCTACCAGA AG Cal #1 7-2 3-5 48h GACTTCTTGAAAATTTGCAGGCCTACCAGAA GC Cal #2 7-2 3-5 24h GACTTCTTGAAAATTTGCAGGCCTACC Cal #2 7-2 3-5 48h GACTTCTTGAAAATTTGCAGG A.10 MRNAs with Seed Sequence Match and Greater tha n 50% Complementarity to the Possible Viral MiRNA Amplified by Primer 7-1 3-1 ATP7B, DCUN1D4, MRFAP1, SEPT8, SLC11A2, SLC2A1, SV2 A, C18orf32, WDR57, GRIK3, UBTD2
A.11 Primers that formed primer-dimers Cal04 #1 Cal04 #2 7-1 3-1 7-1 3-1 7-1 3-2 7-1 3-2 7-1 5-2 7-1 5-2 7-2 3-2 7-2 3-2 7-3 5-2 7-3 5-2 A.12 Primers with inconsistent amplification betwee n PCR replicates Cal04 #1 Primer Cal04 #2 Primer 5 5-1 48h 5 5'-1 48h 5 5-2 48h 5 5'-2 24h 7-1 5-2 48h 7-1 3'-2 24h 7-2 3-1 24h 7-1 5'-2 24h 7-2 3-1 48h 7-2 3'-1 24h 7-2 3-2 48h 7-2 3'-3 48h 7-2 3-3 24h 7-2 3'-4 24 7-2 3-3 48h 7-2 5'-1 48h 7-2 3-4 48h 7-2 5'-2 24h 7-2 5-2 48h 7-3 3'-1 24h 7-2 5-4 24h 7-3 3'-1 48h 7-3 3-1 24h 7-3 5'-1 24h 7-3 3-1 48h 7-3 5'-2 0h 7-3 5-2 48h 7-3 5'-2 48h
References Pandemic (H1N1) 2009 update 106 [Internet]: World Health Organization; c2010 [cited 2010 July 6, 2010]. Available from: http://www.who.int/csr/don/2010_06_25/en/index.html The Deadly Virus The Influenza Epidemic of 1918 [Internet]: National Archives and Records Administration [cited 2010 7/20]. Available from: http://www.archives.gov/exhibits/influenza-epidemic / Allfrey VG, Faulkner R, Mirsky AE. 1964. Acetylatio n and methylation of histones and their possible role in the regulation of RNA synthe sis. Proc Natl Acad Sci U S A 51(5):786-94. Ansari AZ. 2007. Chemical crosshairs on the central dogma. Nature Chemical Biology 3(1):2-7. Babak T, Zhang W, MORRIS Q, Blencowe BJ, Hughes TR. 2004. Probing microRNAs with microarrays: Tissue specificity and functional inference. RNA 10(11):1813. Bai Z, Lei X, Wang L, Gao S. 2008. Identification a nd function of MicroRNAs encoded by herpesviruses. Virol Sin 23(6):459-72. Bonci D, Coppola V, Musumeci M, Addario A, Giuffrid a R, Memeo L, D'Urso L, Pagliuca A, Biffoni M, Labbaye C. 2008. The miR-15a miR-16-1 cluster controls prostate cancer by targeting multiple oncogenic act ivities. Nat Med 14(11):1271-7. Boni MF, Zhou Y, Taubenberger JK, Holmes EC. 2008. Homologous recombination is very rare or absent in human influenza A virus. J V irol 82(10):4807. Borchert GM, Lanier W, Davidson BL. 2006. RNA polym erase III transcribes human microRNAs. Nature Structural & Molecular Biology 13 (12):1097-101. Bourmakina SV and Garcia-Sastre A. 2003. Reverse ge netics studies on the filamentous morphology of influenza A virus. J Gen Virol 84(3): 517. Brodsky AS and Silver PA. 2000. Pre-mRNA processing factors are required for nuclear export. RNA 6(12):1737. Brower-Sinning R, Carter DM, Crevar CJ, Ghedin E, R oss TM, Benos PV. 2009. The role of RNA folding free energy in the evolution of the polymerase genes of the influenza A virus. Genome Bio 10(2):18.
Brueckner F, Armache K, Cheung A, Damsma GE, Ketten berger H, Lehmann E, Sydow J, Cramer P. 2009. Structure-function studies of th e RNA polymerase II elongation complex. Biol Cryst D 65:112-20. Buck AH, Santoyo-Lopez J, Robertson KA, Kumar DS, R eczko M, Ghazal P. 2007. Discrete clusters of virus-encoded micrornas are as sociated with complementary strands of the genome and the 7.2-kilobase stable i ntron in murine cytomegalovirus. J Virol 81(24):13761. Bui M, Wills EG, Helenius A, Whittaker GR. 2000. Ro le of the influenza virus M1 protein in nuclear export of viral ribonucleoprotei ns. J Virol 74(4):1781. Calin GA and Croce CM. 2006. MicroRNA-cancer connec tion: The beginning of a new tale. Cancer Res 66(15):7390. Carthew RW and Sontheimer EJ. 2009. Origins and mec hanisms of miRNAs and siRNAs. Cell 136(4):642-55. How Well Does the Seasonal Flu Vaccine Work? [Inter net]: Centers for Disease Control and Prevention; c2009 [cited 2010 June 29, 2010]. A vailable from: http://www.cdc.gov/flu/about/qa/vaccineeffect.htm Chen C, Ridzon DA, Broomer AJ, Zhou Z, Lee DH, Nguy en JT, Barbisin M, Xu NL, Mahuvakar VR, Andersen MR. 2005. Real-time quantifi cation of microRNAs by stem-loop RT-PCR. Nucleic Acids Res 33(20):e179. Chen W, Calvo PA, Malide D, Gibbs J, Schubert U, Ba cik I, Basta S, O'Neill R, Schickli J, Palese P. 2001. A novel influenza A virus mitoch ondrial protein that induces cell death. Nat Med 7(12):1306-12. Chorn G, Zhao L, Sachs AB, Flanagan M, Lim LP. 2010 Persistence of a seed-based activity following segmentation of a microRNA guide strand. RNA 16(12):2336. Choy EYW, Siu KL, Kok KH, Lung RWM, Tsang CM, To KF Kwong DLW, Tsao SW, Jin DY. 2008. An epstein-barr virusencoded microRN A targets PUMA to promote host cell survival. J Exp Med 205(11):2551. Conaway JW and Conaway RC. 1999. Transcription elon gation and human disease. Annu Rev Biochem 68(1):301-19. Cramer P, Bushnell DA, Kornberg RD. 2001. Structura l basis of transcription: RNA polymerase II at 2.8 angstrom resolution. Science 2 92(5523):1863. Cullen BR. 2009. Viral and cellular messenger RNA t argets of viral microRNAs. Nature 457(7228):421-5.
Dohm JC, Lottaz C, Borodina T, Himmelbauer H. 2008. Substantial biases in ultra-short read data sets from high-throughput DNA sequencing. Nucleic Acids Res Dynlacht BD, Hoey T, Tjian R. 1991. Isolation of co activators associated with the TATA-binding protein that mediate transcriptional a ctivation. Cell 66(3):563-76. Fazi F and Nervi C. 2008. MicroRNA: Basic mechanism s and transcriptional regulatory networks for cell fate determination. Cardiovasc Re s 79(4):553. Fire A, Xu SQ, Montgomery MK, Kostas SA, Driver SE, Mello CC. 1998. Potent and specific genetic interference by double-stranded RN A in caenorhabditis elegans. Nature 391(6669):806-11. Friedman RC, Farh KK, Burge CB, Bartel DP. 2009. Mo st mammalian mRNAs are conserved targets of microRNAs. Gen Res 19:92. Gallie DR. 1991. The cap and poly (A) tail function synergistically to regulate mRNA translational efficiency. Genes Dev 5(11):2108. Gilbert SF. 2006. Promoter structure and the method s of transcription complex assembly. In: Developmental biology. 8th ed. Sunderland, MA: Sinauer and Associates. Glaser L, Stevens J, Zamarin D, Wilson IA, Garcia-S astre A, Tumpey TM, Basler CF, Taubenberger JK, Palese P. 2005. A single amino aci d substitution in 1918 influenza virus hemagglutinin changes receptor binding specif icity. J Virol 79(17):11533. Gmez-Puertas P, Albo C, Perez-Pastrana E, Vivo A, Portela A. 2000. Influenza virus matrix protein is the major driving force in virus budding. J Virol 74(24):11538. Gottwein E, Mukherjee N, Sachse C, Frenzel C, Major os WH, Chi JTA, Braich R, Manoharan M, Soutschek J, Ohler U. 2007. A viral mi croRNA functions as an orthologue of cellular miR-155. Nature 450(7172):10 96-9. Grad Y, Aach J, Hayes GD, Reinhart BJ, Church GM, R uvkun G, Kim J. 2003. Computational and experimental identification of C. elegans microRNAs. Mol Cell 11(5):1253-63. Gregory RI, Yan K, Amuthan G, Chendrimada T, Dorato taj B, Cooch N, Shiekhattar R. 2004. The microprocessor complex mediates the genes is of microRNAs. Nature 432(7014):235-40. Grimson A, Farh KK, Johnston WK, Garrett-Engele P, Lim LP, Bartel DP. 2007. MicroRNA Targeting specificity in mammals: determin ants beyond seed pairing. Mol Cell 27(1):91-105.
Grundhoff A, Sullivan CS, Ganem D. 2006. A combined computational and microarraybased approach identifies novel microRNAs encoded b y human gammaherpesviruses. RNA 12(5):733. Guo H, Ingolia NT, Weissman JS, Bartel DP. Mammalia n microRNAs predominantly act to decrease target mRNA levels. Nature 466(7308):83 5-40. Hall N. 2007. Advanced sequencing technologies and their wider impact in microbiology. J Exp Bio 210:1518. Hausner W and Thomm M. 2001. Events during initiati on of archaeal transcription: Open complex formation and DNA-protein interactions. J B acteriol 183(10):3025. Heid CA, Stevens J, Livak KJ, Williams PM. 1996. Re al time quantitative PCR. Genome Res 6(10):986. Hilleman MR. 2002. Realities and enigmas of human v iral influenza: Pathogenesis, epidemiology and control. Vaccine 20(25-26):3068-87 Hirose Y and Manley JL. 2000. RNA polymerase II and the integration of nuclear events. Genes Dev 14(12):1415. Hutvagner G and Simard MJ. 2008. Argonaute proteins : Key players in RNA silencing. Nature Reviews Molecular Cell Biology 9(1):22-32. OligoAnalyzer [Internet]: Integrated DNA Technologi es Inc; c2010 [cited 2010 6/4/10]. Available from: http://www.idtdna.com/analyzer/Applications/OligoAn alyzer/ SOLiDTM Total RNA-Seq Kit [Internet]; c2011 [cited 2011 220]. Available from: http://www.invitrogen.com/site/us/en/home/ProductsandServices/Applications/Nucleic-Acid-Amplification-an d-Expression-Profiling/DNASequencing/next-generation-sequencing/solid-total-r na-seq-kit.html Kaiser L, Wat C, Mills T, Mahoney P, Ward P, Hayden F. 2003. Impact of oseltamivir treatment on influenza-related lower respiratory tr act complications and hospitalizations. Arch Intern Med 163(14):1667. Kapranov P, Ozsolak F, Kim SW, Foissac S, Lipson D, Hart C, Roels S, Borel C, Antonarakis SE, Monaghan AP, John B, Milos PM. 2010 New class of genetermini-associated human RNAs suggests a novel RNA copying mechanism. Nat Let 466: 642. Kedde M, Strasser MJ, Boldajipour B, Vrielink JAF, Slanchev K, le Sage C, Nagel R, Voorhoeve PM, van Duijse J, rom UA. 2007. RNA-bind ing protein Dnd1 inhibits microRNA access to target mRNA. Cell 131(7):1273-86
Kertesz M, Iovino N, Unnerstall U, Gaul U, Segal E. 2007. The role of site accessibility in microRNA target recognition. Nat Gen 39: 1278-12 84. Khvorova A, Reynolds A, Jayasena SD. 2003. Function al siRNAs and miRNAs exhibit strand bias. Cell 115(2):209-16. Kochendoerfer GG, Salom D, Lear JD, Wilk-Orescan R, Kent SBH, DeGrado WF. 1999. Total chemical synthesis of the integral membrane p rotein influenza A virus M2: Role of its C-terminal domain in tetramer assembly. Biochemistry (N Y ) 38(37):11905-13. Krossoy B, Hordvik I, Nilsen F, Nylund A, Endresen C. 1999. The putative polymerase sequence of infectious salmon anemia virus suggests a new genus within the orthomyxoviridae. J Virol 73(3):2136. Kumar MS, Lu J, Mercer KL, Golub TR, Jacks T. 2007. Impaired microRNA processing enhances cellular transformation and tumorigenesis. Nat Genet 39(5):673-7. Lau N, Robine N, Martin R. 2009. Abundant primary p iRNAs, endo-siRNAs and microRNAs in. Laver W and Valentine R. 1969. Morphology of the is olated hemagglutinin and neuraminidase subunits of influenza virus. Virology 38(1):105-19. Lecellier CH, Dunoyer P, Arar K, Lehmann-Che J, Eyq uem S, Himber C, Saib A, Voinnet O. 2005. A cellular microRNA mediates antiv iral defense in human cells. Science 308(5721):557. Lee Y, Kim M, Han J, Yeom KH, Lee S, Baek SH, Kim V N. 2004. MicroRNA genes are transcribed by RNA polymerase II. EMBO J 23(20):405 1-60. Lei EP and Silver PA. 2002. Protein and RNA export from the nucleus. Developmental Cell 2(3):261-72. Lima WF, Murray H, Nichols JG, Wu H, Sun H, Prakash TP, Berdeja AR, Gaus HJ, Crooke ST. 2009. Human dicer binds short single-str and and double-strand RNA with high affinity and interacts with different reg ions of the nucleic acids. J Biol Chem 284(4):2535. Lippman Z and Martienssen R. 2004. The role of RNA interference in heterochromatic silencing. Nature 431(7006):364-70. Liu J, Carmell MA, Rivas FV, Marsden CG, Thomson JM Song JJ, Hammond SM, Joshua-Tor L, Hannon GJ. 2004. Argonaute2 is the ca talytic engine of mammalian RNAi. Science's STKE 305(5689):1437.
n Lpez-Lastra M, Rivas A, Barra MI. 2005. Protein s ynthesis in eukaryotes: The growing biological relevance of cap-independent translation initiation. Biol Res 38:121-46. Lund E, Guttinger S, Calado A, Dahlberg JE, Kutay U 2004. Nuclear export of microRNA precursors. Science 303(5654):95. Lykke-Andersen S, Jensen TH. 2007. Overlapping path ways dictate termination of RNA polymerase II transcription. Biochimie 89(10):1177. Martin K and Heleniust A. 1991. Nuclear transport o f influenza virus ribonucleoproteins: The viral matrix protein (M1) promotes export and i nhibits import. Cell 67(1):11730. Meister G, Landthaler M, Patkaniowska A, Dorsett Y, Teng G, Tuschl T. 2004. Human Argonaute2 Mediates RNA Cleavage Targeted by miRNAs and siRNAs. Mol Cell 15(2):185-197. Metzer MI, Cox NJ, Fukuda K. 1999. The economic imp act of pandemic influenza in the united states: Priorities for intervention. Emergin g Infectious Diseases 5(5):659. Mitchell PJ and Tjian R. 1989. Transcriptional regu lation in mammalian cells by sequence-specific DNA binding proteins. Science 245 (4916):371. Miyoshi K, Tsukumo H, Nagami T, Siomi H, Siomi MC. 2005. Slicer function of drosophila argonautes and its involvement in RISC f ormation. Genes Dev 19(23):2837. Monto AS, Gravenstein S, Elliott M, Colopy M, Schwe inle J. 2000. Clinical signs and symptoms predicting influenza infection. Arch Inter n Med 160(21):3243. Morris KV, Chan SWL, Jacobsen SE, Looney DJ. 2004. Small interfering RNA-induced transcriptional gene silencing in human cells. Scie nce 305(5688):1289. Murphy D, Dancis B, Brown JR. 2008. The evolution o f core proteins involved in microRNA biogenesis. BMC Evol Biol 8:92. Neely LA. 2009. 13 methods to quantify microRNA gen e expression. Regulation of Gene Expression by Small RNAs :241. Nelson MI, Viboud C, Simonsen L, Bennett RT, Griese mer SB, George KS, Taylor J, Spiro DJ, Sengamalay NA, Ghedin E. 2008. Multiple r eassortment events in the evolutionary history of H1N1 influenza A virus sinc e 1918. PLoS Pathogens 4(2). Nicholls H. 2006. Pandemic influenza: The inside st ory. PLoS Biology 4(2).
Otsuka M, Jing Q, Georgel P, New L, Chen J, Mols J, Kang YJ, Jiang Z, Du X, Cook R. 2007. Hypersusceptibility to vesicular stomatitis v irus infection in Dicer1-deficient mice is due to impaired miR24 and miR93 expression. Immunity 27(1):123-34. Ozsolak F, Ting DT, Wittner BS, Brannigan BW, Paul S, Bardeesy N, Ramaswamy S, Milos PM, Haber DA. 2010. Amplification-free digita l gene expression profiling from minute cell quantities. Nature Methods 7(8):61 9-21. Palese P. 2004. Influenza: Old and new threats. Nat Med 10(12):82. Parvin JD, Palese P, Honda A, Ishihama A, Krystal M 1989. Promoter analysis of influenza virus RNA polymerase. J Virol 63(12):5142 Pavlov AR, Pavlova NV, Kozyavkin SA, Slesarev AI. 2 004. Recent developments in the optimization of thermostable DNA polymerases for ef ficient applications. Trends Biotechnol 22(5):253-60. Perez JT, Varble A, Sachidanandam R, Zlatev I, Mano haran M, Garca-Sastre A, tenOever BR. 2010. Influenza A virus-generated smal l RNAs regulate the switch from transcription to replication. Proceedings of t he National Academy of Sciences 107(25):11525. Poon LLM, Pritlove DC, Fodor E, Brownlee GG. 1998. Direct evidence that the poly (A) tail of influenza A virus mRNA is synthesized by re iterative copying of a U track in the virion RNA template. J Virol 73(4):3473. Qiu Y and Krug RM. 1994. The influenza virus NS1 pr otein is a poly (A)-binding protein that inhibits nuclear export of mRNAs containing po ly (A). J Virol 68(4):2425. Racaniello V. Influenza Viral RNA Synthesis [Intern et]: Virology Blog; c2009 [cited 2010 7/12]. Available from: http://www.virology.ws/ 2009/05/08/influenza-viral-rnasynthesis/ Raymond CK, Roberts BS, Garrett-Engele P, Lim LP, J ohnson JM. 2005. Simple, quantitative primer-extension PCR assay for direct monitoring of microRNAs and short-interfering RNAs. RNA 11(11):1737. Sashital DG and Doudna JA. 2010. Structural insight s into RNA interference. Curr Opin Struct Biol 20(1):90-97. Seo GJ, Fink LHL, OHara B, Atwood WJ, Sullivan CS. 2008. Evolutionarily conserved function of a viral microRNA. J Virol 82(20):9823-2 8. Shalgi R, Lieber D, Oren M, Pilpel Y. 2007. Global and local architecture of the mammalian microRNA-transcription factor regulatory network. PLoS Comput Biol 3(7):e131.
Shapiro G, Gurney Jr T, Krug R. 1987. Influenza vir us gene expression: Control mechanisms at early and late times of infection and nuclear-cytoplasmic transport of virus-specific RNAs. J Virol 61(3):764. Sieczkarski SB and Whittaker GR. 2002. Influenza vi rus can enter and infect cells in the absence of clathrin-mediated endocytosis. J Virol 7 6(20):10455. Smith RD, Keogh-Brown MR, Barnett T, Tait J. 2009. The economy-wide impact of pandemic influenza on the UK: A computable general equilibrium modelling experiment. Br Med J 339(nov19 1):b4571. Song JJ, Liu J, Tolia NH, Schneiderman J, Smith SK, Martienssen RA, Hannon GJ, Joshua-Tor L. 2003. The crystal structure of the Ar gonaute2 PAZ domain reveals an RNA binding motif in RNAi effector complexes. Natur e Structural & Molecular Biology 10(12):1026-32. Song L, Liu H, Gao S, Jiang W, Huang W. 2010. Cellu lar MicroRNAs inhibit replication of the H1N1 influenza A virus in infected cells. J Virol 84(17):8849. Steiner FA, Okihara KL, Hoogstrate SW, Sijen T, Ket ting RF. 2009. RDE-1 slicer activity is required only for passenger-strand clea vage during RNAi in caenorhabditis elegans. Nature Structural & Molecul ar Biology 16(2):207-11. Sullivan CS and Grundhoff A. 2007. Identification o f viral microRNAs. Meth Enzymol 427:1-23. Swayne D and Akey B. 2005. Avian influenza control strategies in the united states of america. Avian Influenza: Prevention and Control :1 13. Thermo Fisher Scientific. 2009. NanoDrop 8000 spect rophotometer user manual. Washington, DE: Thermo Fisher Scientific. Thompson WW, Shay DK, Weintraub E, Brammer L, Cox N Anderson LJ, Fukuda K. 2003. Mortality associated with influenza and respi ratory syncytial virus in the united states. JAMA 289(2):179. Tipper J. 2008. Study of cell cultures and viral ti ters. Personal Communication. Treanor J. 2004. Influenza vaccine--outmaneuvering antigenic shift and drift. N Engl J Med 350(3):218. Umbach JL, Kramer MF, Jurak I, Karnowski HW, Coen D M, Cullen BR. 2008. MicroRNAs expressed by herpes simplex virus 1 durin g latent infection regulate viral mRNAs. Nature 454(7205):780-3.
r Vagin VV, Sigova A, Li C, Seitz H, Gvozdev V, Zamor e PD. 2006. A distinct small RNA pathway silences selfish genetic elements in th e germline. Science 313(5785):320. Valoczi A, Hornyik C, Varga N, Burgyan J, Kauppinen S, Havelda Z. 2004. Sensitive and specific detection of microRNAs by northern blo t analysis using LNA-modified oligonucleotide probes. Nucleic Acids Res 32(22):e1 75. VanWestrienen J. 2008. Study of Possible miRNAs enc oded by Influenza A/NewCaledonia/20/1999. Personal Communication. Vazquez F, Vaucheret H, Rajagopalan R, Lepers C, Ga sciolli V, Mallory AC, Hilbert JL, Bartel DP, Crt P. 2004. Endogenous trans-acting s iRNAs regulate the accumulation of arabidopsis mRNAs. Mol Cell 16(1):6 9-79. Verdel A, Jia S, Gerber S, Sugiyama T, Gygi S, Grew al SIS, Moazed D. 2004. RNAimediated targeting of heterochromatin by the RITS c omplex. Science 303(5658):672. Wang P, Palese P, O'Neill RE. 1997. The NPI-1/NPI-3 (karyopherin alpha) binding site on the influenza a virus nucleoprotein NP is a nonc onventional nuclear localization signal. J Virol 71(3):1850. Wang X and El Naqa IM. 2007. Prediction of both con served and nonconserved microRNA targets in animals. Bioinformatics 24(3):3 25. Wang Y, Brahmakshatriya V, Zhu H, Lupiani B, Reddy SM, Yoon BJ, Gunaratne PH, Kim JH, Chen R, Wang J, et al. 2009. Identification of differentially expressed miRNAs in chicken lung and trachea with avian influ enza virus infection by a deep sequencing approach. BMC Genomics 10:512. Wei CJ, Boyington JC, McTamney PM, Kong WP, Pearce MB, Xu L, Andersen H, Rao S, Tumpey TM, Yang ZY. 2010. Induction of broadly n eutralizing H1N1 influenza antibodies by vaccination. Science Winter J, Jung S, Keller S, Gregory RI, Diederichs S. 2009. Many roads to maturity: MicroRNA biogenesis pathways and their regulation. Nat Cell Biol 11(3):228-34. Yao Y, Zhao Y, Smith LP, Watson M, Nair V. 2009. No vel microRNAs (miRNAs) encoded by herpesvirus of turkeys: Evidence of miRN A evolution by duplication. J Virol 83(13):6969. Yekta S, Shih I, Bartel BP. 2004. MicroRNA-Directed cleavage of HOXB8 mRNA. Science 304:594.
Zeng Y, Yi R, Cullen BR. 2005. Recognition and clea vage of primary microRNA precursors by the nuclear processing enzyme drosha. EMBO J 24(1):138. Zhang B, Pan X, Wang Q, Cobb GP, Anderson TA. 2006. Computational identification of microRNAs and their targets. Computational Biolo gy and Chemistry 30(6):395407. Zhang H, Kolb FA, Jaskiewicz L, Westhof E, Filipowi cz W. 2004. Single processing center models for human dicer and bacterial RNase I II. Cell 118(1):57-68. Zhang H, Kolb FA, Brondani V, Billy E, Filipowicz W 2002. Human dicer preferentially cleaves dsRNAs at their termini without a requireme nt for ATP. EMBO J 21(21):5875-85. Zhouravleva G, Frolova L, Le Goff X, Le Guellec R, Inge-Vechtomov S, Kisselev L, Philippe M. 1995. Termination of translation in euk aryotes is governed by two interacting polypeptide chain release factors, eRF1 and eRF3. EMBO J 14(16):4065