Discovery and Biochemical Characterization of Casl3b, a Type VI-B CRISPR-Associated
RNA-Guided RNase
by
Neena Kenton Pyzocha
B.S. Biology, Rensselaer Polytechnic Institute (2011)
Submitted to the Department of Biology
in Partial Fulfillment of the Requirements for the Degree of
Doctor of Philosophy
at the
MASSACHUSETTS INSTITUTE OF TECHNOLOGY
June 2017
@2017 Massachusetts Institute of Technology. All rights reserved.
A'!11--111
Signature of Author: __
Certified by:
Signature redacted
fl/ 9 /1 t
ignature redacte
U
Associate Professor of Biological
Department of Biology
May 18, 2017
Feng Zhang
Engineering and Brain and Cognitive Sciences
Thesis Supervisor
Signature redacted
Amy Keating
Professor of Biology
Co-Chair, Biology Graduate CommitteeI
1
Accepted by:
MASSACHUSETTS INSTITUTEOF TECHNOLOGY
MAY 2 3 2017
LIBRARIES

Discovery and Biochemical Characterization of Casl3b, a Type VI-B CRISPR-Associated RNA-
Guided RNase
by
Neena Pyzocha
Submitted to the Department of Biology, on May 18, 2017 in partial fulfillment of the requirements
for the degree of Doctor of Philosophy
ABSTRACT
Targeted methods to perturb nucleic acid sequences or epigenetic modifications enable better
understanding of nucleic acid function or dysfunction in disease. Particular CRISPR-associated
(Cas) proteins are used for targeted DNA editing because they are easy to use, precise, and specific.
CRISPR-Cas systems function in microbes as adaptive immune systems and are composed of
endonuclease protein(s) with targeting guide RNAs that together provide interference against
foreign nucleic acids. CRISPR-Cas systems exhibit wide diversity, with the ability to target DNA or
RNA, and some large single effector Class 2 Cas proteins have been harnessed for genome
engineering. There is an interest to find and characterize additional Class 2 Cas proteins, especially
ones that target RNA, to enable further targeting capabilities. A large computational database
search was conducted to find a large uncharacterized protein within 10 kilobases of a CRISPR array
and from this search casl3b was identified. The casl3b loci may contain the genes csx27 or csx28
and one or two CRISPR arrays with canonical or long direct repeats. Experimentally, purified
Casl3b processes its own CRISPR array(s) into crRNA(s) and Casl3b with either crRNA
architecture targets single stranded RNA cleavage using its HEPN domains. SsRNA cleavage occurs
at pyrimidine residues and is constrained by 5' and 3' protospacer-flanking sequences on the target.
Casl3b also displays the collateral effect, a non-specific cleavage of ssRNAs after targeted cleavage.
Casl3b with mutated HEPN domains lose ssRNA cleavage activity but maintain strong, targeted
binding capacity. Genetically, Casl3b-mediated RNA interference occurs with Casl3b alone and is
repressed by Csx27 or enhanced by Csx28. Casl3b is characterized as a Type VI-B CRISPR-Cas
system and represents the second Class 2 Cas protein to target RNA, the other being Casl3a (C2c2).
Future studies of Casl3b are warranted to better understand its functional mechanisms, specificity,
role of small proteins, and acquisition. Casl3b could be developed into a suite of tools for
transcriptome engineering to mediate RNA translation, splicing, or deposition of epitranscriptomic
marks. Casl3b could also be utilized for RNA diagnostic or RNA imaging assays. Identifying and
characterizing novel CRISPR-Cas systems opens new opportunities for utilizing Cas enzymes for
biomedical advances.
Thesis Supervisor: Feng Zhang
Title: Associate Professor of Biological Engineering and Brain and Cognitive Sciences
3
ACKNOWLEDGEMENTS
There are many people I would like to acknowledge for their scientific or personal support who
have helped me throughout graduate school.
Scientifically, I would like to thank my PI Feng Zhang for everything he has helped me with
over the past five years as a member of his lab. It has truly been an honor to work with such a
forward-thinking and energetic young principal investigator who has assembled a truly amazing
group of laboratory personnel. I'd next like to thank my committee members: Steve Bell for being
my committee chair, providing thoughtful suggestions for my work, and early support during
graduate school; Mary Gehring for being on my committee my whole graduate career and helpful
scientific and personal advice, and lastly Kevin Eggan for serving as my external committee
member. I'd next like to thank those students and post-docs who co-authored papers with me.
First, I'll thank David Cox for his hard work, critical evaluation of science, and great friendship. I'd
also like to thank Aaron Smargon for his enthusiasm and interest for embarking on the discovery of
Casl3b with me. Lastly, I will thank Kaijie Zheng, Ian M. Slaymaker, Jonathan S. Gootenberg, Omar
A. Abudayyeh, Patrick Essletzbichler, Fei Ann Ran, and Patrick D. Hsu for their help with publishing
my first author papers. I also want to thank Sergey Shmakov, Kira S. Makarova, and Eugene V.
Koonin, our helpful collaborators. Also to thank is the whole Zhang lab. The members of the Z-lab
are the most persistent, dedicated, hard working, intelligent, and talented group of people I've ever
met. It has been an honor to work amongst them, albeit the most humbling 5 years of my life. I'd
also like to thank Michael Yim, Rich Belliveau, Erin Blackwell, and Rhiannon Macrae, the best
support staff a lab could have. In addition to those who directly contributed to my published work,
I'd like to thank several others who helped me academically. First, thank you to all my former
teachers, especially Mr. Richard Parent, whose class in high school was one big contributor in me
pursuing a Ph.D. and doing research. I will also thank Eric Rutledge, George Makhatadze, George
Edick, Erin Piazza, Mandana Sassanfar, Krishanu Saha, and Rudolf Jaenisch for support during
undergraduate or summer research. I'd also like to thank the NSF GFRP for selecting me as a fellow
and for financial support. I'd also like to thank scholarship funds that allowed me to go to RPI.
There are many other people who provided personal support during my graduate tenure at
MIT. First, I would like to thank my family: Mom, Dad, Natasha, and Nicholas for all the love and
support from the day I was born and throughout graduate school. Also the Pyzocha family dogs
Maddie, Meeka, Milo, Mixie, and Max (and Ray the cat) just always make me smile. I would next like
to thank the person who made the greatest positive difference in my life during graduate school,
and that is my fianc6 Colin M. Smith who provided daily love, support, and encouragement. I'd also
like to thank my extended family: grandparents Nana, Papa, RIP Babci, as well as uncles, aunts,
cousins, and soon to be family-in-law including Bruce and Diane Smith. I'd also like to thank my
closest friends Lindsay, Jen, Laura K., Brian, Andy, Chris M., Krissy, Chris W., Mike, Cait, and
roommates Steph and Chris B. I'd also like to thank supportive classmates including Laura F., Shuyu,
and Xinchen. I'd also like to thank particular lab members including Naomi Habib for her support,
care, and willingness to discuss life and family as a female scientist; Sara Jones for her support,
humor, and an ear to always listen; David Cox for his friendship, humor, and support; Ann Ran for
her help and care in the early stages of my graduate career; Bernd Zetsche for his humor and care;
Michael Yim for his friendship and willingness to listen and Sourav Choudry for career and personal
support. I'd also like to thank Frank Solomon and Anne Welsh.
4
BIOGRAPHICAL NOTE
Undergraduate Education:
2007-2011: B.S. (Biology), Rensselaer Polytechnic Institute, Troy, NY.
Cumulative GPA: 3.93/4.0. Graduated Summa cum laude
Awards:
2011: NSF Graduate Research Fellow
2013: Teresa Keng Graduate Teaching Prize (Biology Department, MIT)
Publications:
Smargon, A*; Cox, D*; Pyzocha, N*; Zheng, K; Slaymaker, I; Gootenberg, J; Abudayyeh, 0;
Essletzbichler, P; Shmakov, S; Makarova, K; Koonin, E; Zhang, F. Casl3b is a Type VI-B CRISPR-
associated RNA-Guided RNAse differentially regulated by accessory proteins Csx27 and Csx28.
Molecular Cell 2017, 65, 1-13. (* Co-first authors)
Shmakov, S; Smargon, A; Scott, D; Cox, D; Pyzocha, N; Yan,W; Abudayyeh, 0; Gootenberg, J;
Makarova, K; Wolf, Y; Severinov, K; Zhang, F; Koonin, E. Diversity and evolution of Class 2 CRISPR-
Cas systems Nature Reviews Microbiology 2017, 15, 169-182.
Pyzocha, N; Ran, F; Hsu, P; Zhang, F. RNA-guided Genome Editing of Mammalian Cells. Methods in
molecular biology 2013, 1114, 269-277.
Saha, K; Mei, Y; Reisterer, C; Pyzocha, N; Yang, J; Muffat, J; Davies, M; Alexander, M; Langer, R;
Anderson, D; Jaenisch, R. Surface-engineered substrates for improved human pluripotent stem cell
culture under fully defined conditions. Proceedings of the National Academy of Sciences of the United
States ofAmerica 2011, 108, 18714-9.
Mei, Y; Saha, K; Bogatyrev, S; Yang, J; Hook, A; Kalcioglu, Z; Cho, S; Mitalipova, M; Pyzocha, N; Rojas,
F; Van Vliet, K; Davies, M; Alexander, M; Langer, R; Jaenisch, R; Anderson, D. Combinatorial
development of biomaterials for clonal growth of human pluripotent stem cells. Nature
Materials 2010, 9, 768-78.
Tsamaloukas, A; Pyzocha, N; Makhatadze, G. "Pressure Perturbation Calorimetry of Unfolded
Proteins." The Journal of Physical Chemistry B 2010, 114.49, 16166-16170.
Teaching Experience and Certifications:
2016: Contributing author for Introductory Biology online textbook secretoflife.org. Wrote, edited,
and composed figures for several subchapters.
Spring 2015: Teaching Assistant, Introduction to Experimental Biology and
Communication (7.02), MIT
Spring 2015: Volunteer Science Fair Judge at the Massachusetts State Science & Engineering
Fair, Pioneer Charter School of Science in Everett, MA, and Boston Latin Academy in
Boston, MA.
Spring 2014: Kaufman Teaching Certificate Program, MIT
Spring 2013: Teaching Assistant, Introductory Biology (7.013), MIT
2011-2013: Volunteer tutoring, Tutoring Plus, Cambridge, MA
5
TABLE OF CONTENTS
Chapter 1 ------------------------------------------------------------------------------ Page 7
Introduction: CRISPR-Cas Systems And Their Use In Genome And
Transcriptome Engineering
Chapter 2 ---------------------------------------------------------------------------- Page 33
Thesis Work: Casl3b Is A Type VI-B CRISPR-Associated RNA-Guided Rnase
Differentially Regulated By Accessory Proteins Csx27 And Csx28
Chapter 3 ---------------------------------------------------------------------------- Page 71
Future Directions: Basic Biology And Engineering Applications Of Casl3b
6
CHAPTER 1
INTRODUCTION
CRISPR-CAS SYSTEMS AND THEIR USE IN GENOME AND TRANSCRIPTOME
ENGINEERING
7
CRISPR-Cas editing technologies have revolutionized modern molecular biology by
making targeted DNA edits simple and scalable for use in mammalian genome engineering
and other biological assays. CRISPR-Cas technologies are harnessed from naturally
occurring microbial systems that display wide diversity for targeted nucleic acid cleavage.
Several CRISPR-Cas single effector enzymes have been characterized and engineered for use
in mammalian cells, and their unique properties can make a critical difference in
experimental use or targeting specificity. Currently, there exist tools to efficiently target
DNA, and tools to easily modify RNA illustrate an unmet need. Increasing our understanding
of novel CRISPR-Cas systems will expand the opportunities for utilizing them for biomedical
advances. Therefore there exists an interest to continue to identify, characterize, and
engineer CRISPR-Cas enzymes and further develop them for genome or transcriptome
editing.
Part I: CRISPR Origins, Function, and Categorization
The evolutionary web of life represents a complex and competitive environment in which
organisms must continually evolve for better adaptation and survival. The microbial world displays
this concept notably: phage (viruses that infect bacteria) can outnumber bacteria 10:1 in some
environments (Suttle, 2005) and parasitic mobile genetic elements such as plasmids cause
unnecessary energetic or genomic burdens (Doolittle and Sapienza, 1980). To combat these forces,
bacteria and archaea have evolved several different forms of resistance to foreign genetic material.
A collective list of immune systems used by bacteria and archaea include physical blockage,
restriction enzymes, abortive infection, bacteriophage exclusion, and CRISPR-Cas systems, but vary
in their representation in specific bacterial or archaeal species. Physical blockage mechanisms
result in physically altered cell surfaces to reduce phage entry. Two specific physical blockage
mechanisms include adsorption resistance, where surface receptors mutate to inhibit
bacteriophage interaction, and penetration blocks where surface receptors are physically
obstructed from bacteriophages, for example using extracellular polymeric substances (Labrie et
al., 2010). Alternatively, numerous restriction modification system enzymes have sequence specific
endonuclease activity to cleave unmethylated phage DNA leaving methylated host DNA protected
(Meselson and Yuan, 1968). Bacteriophage exclusion immunity mechanisms specifically inhibit
phage DNA replication through unknown mechanisms and protect host cell DNA specifically based
on host cell DNA methylation (Goldfarb et al., 2015). Abortive infection systems are a suite of
multiple different enzymatic approaches to eliminate phage infection by targeting phage
8
replication, transcription, or translation machinery but results in phage and the organisms' own
death to prevent bacteriophage infection of the greater bacterial community (Snyder, 1995). Lastly,
CRISPR-Cas adaptive immune systems protect bacteria and archaea against foreign invading
genetic elements such as viruses or plasmids (Marraffini and Sontheimer, 2008) by encoding those
foreign DNA fragments into the host genome and later target the foreign nucleic acid for cleavage
based on complementary association with the encoded sequence (Barrangou et al., 2007).
Basic Components of CRISPR-Cas Systems
The acronym CRISPR-Cas describes the basic components in the microorganisms' genomic
locus necessary for the adaptive immunity process. CRISPR stands for Clustered Regularly
Interspaced Short Palindromic Repeats (Jansen et al., 2002), which describes the repetitive CRISPR
array targeting moiety, and Cas is short for CRISPR-associated, which describes the diverse cas
genes whose proteins perform various functions required for CRISPR-Cas based immunity (Jansen
et al., 2002). The composition of the CRISPR array and function of the Cas proteins are described in
further detail below and illustrated in figure 1.
The CRISPR array has a regulatory leader sequence upstream of the alternating short
variable protospacer (also known as spacer) sequences and short direct repeat (DR) sequences.
The A-T rich regulatory leader sequence directly upstream of the CRISPR array contains the
promoters for CRISPR array transcription (Pougach et al., 2010). Unique spacers are sequences
identical to sequences from invading foreign genetic elements, which can originate from multiple
different phages or mobile genetic elements such as plasmids(Bolotin et al., 2005; Mojica et al.,
2005; Pourcel et al., 2005). The partially palindromic direct repeats within a single CRISPR-Cas
system usually have the same sequence and when transcribed, this direct repeat individually, or in
some cases, complexed to another unique trans-activating RNA (tracrRNA) sequence (Deltcheva et
al., 2011), forms a unique hairpin-like structure that is recognized by the Cas protein(s) from its
locus (Brouns et al., 2008). The CRISPR array is not translated and its final functional form is
processed individual CRISPR RNA (crRNA) units consisting of a spacer and direct repeat.
9
tracrRNA
SpCas9 locus . cas9
Cas gene operon
casi cas2
CRISPR array
csn2
Spacers
Leader I I I |
Direct repeats
Figure 1: Schematic of a CRISPR-Cas locus. Although the loci will vary depending on the CRISPR-
Cas type, the Type II CRISPR-Cas9 locus from the bacteria Streptococcus pyogenes is shown here as a
representative. Figure adapted from Cong et al., 2013.
The Cas genes are organized in an operon expression system and have diverse roles to result in
adaptive immunity. There are four distinct functional modules that classify Cas protein function
including spacer acquisition, crRNA processing, target cleavage, and ancillary roles. These proteins
play critical roles in the three phases of CRISPR adaptive immunity: spacer acquisition proteins
mediate adaptation, crRNA processing proteins are involved in creating the CRISPR-Cas functional
complex, and target cleavage proteins lead to interference of foreign nucleic acid propagation and
immunity. The three phases of CRISPR adaptive immunity are described in further detail in the
next section and are summarized in Figure 2.
Proka
Stage 1:Foreign DNA acquisition
ryotic cell
cas genes
host genome
CRISPR locus
CRISPR locus transcription
Ds/
pre-crRNA
Stage 2: CRISPR RNA processing
Figure 2: Schematic summarizing the three stages of CRISPR-Cas adaptive immunity. Image
from http://rna.berkeley.edu/crispr.html.
10
Stage 3:
RNA-guided
targeting of
viralelement
The three phases of CRISPR adaptive immunity
Adaptation is first phase of CRISPR adaptive immunity and is the process by which foreign
nucleic acids are encoded into the CRISPR array (Barrangou et al., 2007). Acquisition of DNA
protospacers requires the Cast and Cas2 proteins (Datsenko et al., 2012; Nunez et al., 2014; Yosef
et al., 2012), which are conserved across different CRISPR-Cas systems (Makarova et al., 2011).
Cast functions as an integrase, contains the catalytic residues responsible for inserting new spacers
into the CRISPR array (Nunez et al., 2014), and is the most highly conserved Cas protein in all
CRISPR systems (Takeuchi et al., 2012). The Cas2 protein forms a complex with the Cast protein
(Nunez et al., 2014), but its specific function is less well understood. Other proteins may also form
a complex with Cast and Cas2 to assist in acquisition in certain CRISPR-Cas types, for example Cas9
in Type II systems, which is the key protein involved in target cleavage (Heler et al., 2015). In
addition to protein requirements, many systems have known target sequence requirements that
must be present to insert a DNA fragment into the CRISPR array. This sequence, known as the
protospacer adjacent motif (PAM), is adjacent to the spacer sequence originating from the phage
and not inserted into the CRISPR array (Deveau et al., 2008; Mojica et al., 2009). It is the preferred
recognition site for Cas proteins involved in acquisition (Heler et al., 2015; Yosef et al., 2012).
Spacers are also added in a directional fashion, with the newest spacers inserted in between the
leader sequence and the first repeat concurrently with direct repeat duplication (Yosef et al., 2012).
Acquisition is critical because it allows bacteria to adapt to new foreign genetic insults.
Expression is the second phase of CRISPR adaptive immunity, during which the components
for targeted nucleic acid cleavage are synthesized, processed, and complexed to form the RNA-
guided endonuclease complex. In the expression phase, cas genes in the operon are transcribed and
translated into numerous Cas proteins. The CRISPR array is transcribed as a single, long, immature
pre-CRISPR RNA (pre-crRNA) transcript containing multiple spacers and direct repeats. It is
processed into its final form, the crRNA, which consists of one spacer and one direct repeat, by
either an endonuclease subunit of a Cas protein (Carte et al., 2008) or other cellular RNases, such as
bacterial RNase III (Deltcheva et al., 2011). The individual crRNAs may be further processed to
remove some RNA bases into a shortened mature form by nuclease activity (Deltcheva et al., 2011).
The targeting RNAs (crRNA only or crRNA+tracrRNA) interact with Cas protein(s) to form a
targeting nuclease complex (Brouns et al., 2008).
Interference is the final phase of CRISPR adaptive immunity and results in cleavage of the
targeted foreign nucleic acid. Interference mechanisms vary between different CRISPR types but in
all systems, targeted nucleic acid cleavage depends on the proper base pairing between the crRNA
11
and the target sequence in the protospacer and the protospacer flanking region(s). In systems that
utilize a PAM for adaptation, the presence of the correct PAM on the target DNA is a requirement for
interference (Mojica et al., 2009). The CRISPR-Cas RNA-guided endonuclease complex first
recognizes and binds the PAM sequence on the target DNA (Sashital et al., 2012; Sternberg et al.,
2014). Subsequently, the target DNA and the crRNA protospacer sequence form Watson-Crick base
pairs at the PAM proximal or "seed region" and continue outward to the PAM distal portion of the
targeted DNA (Sternberg et al., 2014; Wiedenheft et al., 2011). With sufficient pairing, the cognate
nucleic acid is cleaved (Gasiunas et al., 2012; Semenova et al., 2011), resulting in interference. The
PAM sequence serves an important role during interference and because the sequence requirement
is outside the protospacer sequence present in the CRISPR array, it ensures that the CRISPR-Cas
system will not cleave its own genomic DNA. In systems that do not utilize a PAM for adaptation,
target cleavage is only achieved in the absence of base pairing proximal to the protospacer
(Marraffini and Sontheimer, 2010). Perfect base pairing proximal to the protospacer would only
occur if the RNA-targeting nuclease complex were targeting its own DNA (Marraffini and
Sontheimer, 2010; Samai et al., 2015). The sequence binding requirements for interference are
summarized in Figure 3.
a Unbound state b Pm recognmon
cgRNA Targeting of non-self DNA
DgNA 'A"'ngn
Guide 3'wW~~
Seed
DNA 5''''~'" 3' Tre
PAM
DNA melting and
Seed pairig
" uProtection of self DNAd e
Extensive pairig ard cleavage Partial pairilg ard no cleavage
=__ _= repeat ISpaceri repeat]
Figure 3: Base pairing between protospacer adjacent sequences during interference. Left: For
CRISPR-Cas systems that require a PAM (such as Cas9 shown here), the PAM serves as an initial
binding site, which leads to DNA melting and cleavage with extensive cleavage. Image from Wu et
al., 2014. Right: For CRISPR-Cas systems that do not utilize a PAM, this schematic illustrates how
crRNA binding helps the enzyme discriminate between self and non-self. Image from Marraffini and
Sontheimer 2010.
A majority of the known Cas proteins function in one of the three phases of CRISPR adaptive
immunity, but some CRISPR systems have ancillary proteins, that conduct various but generally less
12
well characterized roles. These rare ancillary proteins are believed to provide regulatory,
modulating, or additional functions. For example, Cas4 has known functions including 5' to 3'
ssDNA exonuclease activity and involvement with acquisition processes (Zhang et al., 2012), but it
is also hypothesized to have roles in CRISPR-Cas-coupled programmed cell death.(Makarova et al.,
2014) Also, the s117009 gene has been implicated as a negative regulator in a specific CRISPR-Cas
system in Synechocystis sp (Hein et al., 2013). Some ancillary Cas proteins have a CRISPR-associated
Rossman fold (CARF) domain, which is predicted to bind nucleotides. Some Cas proteins with CARF
domains also exhibit nucleic acid cleavage activity (Makarova et al., 2014). Two Cas proteins with
CARF domains, Csx1 and Csm6, play key roles in certain RNA-targeting CRISPR-Cas systems.
Further analysis of ancillary proteins could help elucidate specific regulation or function of certain
CRISPR-Cas systems.
CRISPR Nomenclature
CRISPR-Cas systems in bacteria or archaea exhibit notable diversity and are organized
according to a specific classification scheme. Approximately 47% of analyzed bacterial genomes
and 87% of analyzed archaea genomes have at least one CRISPR system and many contain multiple
CRISPR-Cas systems (Makarova et al., 2015). CRISPR-Cas protein classification is based on
phylogenetic, comparative genomic, and protein structural analyses (Makarova et al., 2015;
Shmakov et al., 2017). Once cas genes are examined to determine the homology to other known cas
genes, categorization is determined by the presence of a specific signature protein directly involved
in target nucleic acid cleavage and further classified into subgroups based on the presence of
additional signature genes or specific gene arrangements. Currently, there are two classes of
CRISPR-Cas systems composed of six major types, which are further divided into nineteen subtypes
(Makarova et al., 2015; Shmakov et al., 2017). It is important to note however, that the naming of
signature proteins and CRISPR-Cas types is based on the timeline of characterization or
experimental validation and therefore does not have a sequential naming scheme based on
characterization alone (Makarova et al., 2011; Makarova et al., 2015; Shmakov et al., 2015;
Shmakov et al., 2017).
CRISPR-Cas systems are most broadly characterized as either Class 1 or Class 2 (Makarova
et al., 2015). Class 1 systems require multiple Cas proteins to come together in a complex to
mediate interference against foreign genetic elements. Class 1 systems occur in both bacteria and
archaea and are the most common CRISPR-Cas systems utilized in microbial adaptive immunity
(Makarova et al., 2015). Class 1 systems are further divided into three CRISPR-Cas types based on
13
the presence of a specific signature protein: Type I contains Cas3, Type III contains CasiO, and the
putative Type IV contains Csfl. Type I mediates interference through the multisubunit Cascade
complex containing Cas3. Type III systems are characterized based on the presence of CasiO but
notably form complexes with additional Cmr or Csm proteins. Type IV is putative due to its
computational characterization, but lacks experimental evidence confirming targeted nucleic acid
cleavage. In contrast to Class 1 systems, Class 2 systems utilize a large single Cas enzyme to
mediate interference. Class 2 systems are less common than Class 1 systems and occur almost
exclusively in the bacterial domain of life. Sequencing studies have found only four species of
archaea with Class 2 systems (Burstein et al., 2017; Shmakov et al., 2017). Class 2 systems are
further divided into three CRISPR-Cas types based on the presence of other specific signature
proteins: Type I contains Cas9, Type V contains Cas12 (also named Cpfl, C2cl, and C2c3), and Type
VI contains Cas13 (also named C2c2). Figure 4 shows a schematic of the two classes and six major
subtypes of CRISPR-Cas systems with their associated Cas proteins involved with expression,
interference, adaptation, and ancillary functions.
Expression interference Adaptation Ancillary
Pre-crRNA Effector module (crRNA and target binding) Target cleavage Spacer Insertion Regulation, Helper, role
procesirydormancy unknown
Type I 1C1S6 Casl Ca&5 Cas8BLS HODCas3" /ICasT.. Casi
M Type Illinm m casioLS -
Type IV? Muin
Typell .1 13
Type V
Type VI
Figure 4: Schematic of the six CRISPR-Cas types and the proteins involved in CRISPR-Cas
mediated adaptive immunity. Image from Mohanraju et al., 2016 and slightly modified to reflect
recent findings. Cas3 may be expressed as a single protein or two separate proteins (Cas3' and
Cas3"). The acronym SS means small subunit and refers to other small proteins required for
complex formation in Class 1 systems whereas LS means large subunit.
Each CRISPR system cleaves specific nucleic acids
CRISPR-Cas systems display a range of targeted nuclease activity; they either cleave DNA,
RNA, or both RNA and DNA based on the presence of specific nucleic acid cleavage domains
observed on the signature Cas proteins. Although there are several different signature Cas proteins,
14
there are a few recurrent functional catalytic nuclease domains. In this section, the nucleic acid
target and its associated active nuclease domain(s) will be discussed for each type of CRISPR-Cas
system. Table 1 summarizes these conclusions.
Substrate CRISPR type Protein(s) with nuclease activity Nuclease Domain(s)
dsDNA Type I Cas3 HD
Type II Cas9 RuvC and HNH
Type V Cas12 (Cpfl, C2cl, C2c3) RuvC and Nuc (analogous to HNH)
ssRNA Type VI Cas13a (C2c2) HEPN
ssDNA & Type III Cas10 HD
ssRNA Csm3/Cmr4 HEPN
Csm6/Csxl HEPN
Table 1: A summary of CRISPR-Cas targeted substrates and associated CRISPR types,
nuclease proteins, and associated nuclease domains. With multiple different CRISPR types and
unique proteins with nuclease activity, certain nuclease domains are common and dictate substrate
specificity.
The CRISPR-Cas types that exclusively cleave dsDNA are Type I, II, and V, and the DNase
domains utilized vary with respect to whether the system is a Class 1 or Class 2 CRISPR system. The
Type I signature target cleaving Cas protein, Cas3, cleaves DNA residues using the Histidine-
aspartate (HD) nuclease (Huo et al., 2014; Sinkunas et al., 2011). Mutagenesis of the Cas3 HD
nuclease catalytic domains eliminates DNase activity (Sinkunas et al., 2011). On the other hand,
Type II and V CRISPR-systems are Class 2 large single effector enzymes and both of these DNA
nucleases utilize two separate nuclease domains. The two DNA nuclease domains include RuvC-like
and either HNH-like, present in Type II systems, or Nuc, present in Type V systems and functionally
analogous to HNH. Crystal structure evidence illustrates the presence of these nuclease domains in
Cas9 (Jinek et al., 2014; Nishimasu et al., 2015; Nishimasu et al., 2014) and Cas12 (specifically Cpfl
and C2c1) (Dong et al., 2016; Liu et al., 2017a; Yamano et al., 2016; Yang et al., 2016). The RuvC and
HNH nuclease domains are also catalytic domains present in other DNAses: the RuvC domain has a
similar structure and function to the RuvC holiday junction resolvase (Gorecka et al., 2013;
Nishimasu et al., 2014) and the HNH nuclease has a 1pa-metal fold tertiary structure (Nishimasu et
al., 2014) that is present in phage T4 endonuclease VII (Biertumpfel et al., 2007). Genetic evidence
has also identified specific residues involved in DNA cleavage activity. For example in Cas9, both
nuclease domains must be mutated to create a catalytically inactive protein because each nuclease
15
domain cleaves one of the two strands of DNA to result in a DSB (Gasiunas et al., 2012). Mutating
either HNH or RuvC alone creates a nickase enzyme that cleaves just one of the two strands of DNA
(Gasiunas et al., 2012; Jinek et al., 2012).
The C2c2 protein of the Type VI CRISPR-Cas system exclusively cleaves ssRNA using HEPN
domains (Abudayyeh et al., 2016). HEPN (Higher Eukaryotes and Prokaryotes Nucleotide-binding)
(Grynberg et al., 2003) domains exhibit wide diversity of amino acid sequences and occur in
proteins involved in several forms of bacterial immunity including abortive infection, restriction
modification, and CRISPR-Cas systems (Anantharaman et al., 2013). HEPN domains are the
catalytic domains of many different RNAses including RNase LS and Irel (Anantharaman et al.,
2013; Lee et al., 2008; Otsuka and Yonesaki, 2005). HEPN domains generally consist of a-helical
structures and have four commonly observed amino acids: E upstream of an R(X 4 .6 )H motif where
the amino acid after R is typically polar (often N, D, or H) (Anantharaman et al., 2013). Mutagenesis
studies in the HEPN domain of Irel demonstrated that arginine and histidine are essential for
nuclease activity (Lee et al., 2008) and mutagenesis of arginine in either HEPN domain of C2c2
resulted in a catalytically dead enzyme (Abudayyeh et al., 2016). To function as an active RNase,
two HEPN domains need to associate together. Two HEPN domains from the same protein can
come together, as is the case for C2c2 (Liu et al., 2017b), or a single HEPN domain can be present
but require two individual proteins to dimerize for activity (Lee et al., 2008).
The Type III CRISPR-Cas systems cleave both RNA and DNA co-transcriptionally (Peng et al.,
2015; Samai et al., 2015) using Cas proteins with RNase and DNase domains previously described.
The CasiO family of proteins (also known as Csml or Cmr2) are characteristic for Type III CRISPR
systems and usually have a HD nuclease domain, similar to the Type I systems, which are
responsible for cleaving DNA (Elmore et al., 2016). As a Class 1 CRISPR-system, Type III systems
consist of many proteins, and other Cas proteins within the locus have a HEPN domain that cleaves
RNA. Notably, there are two different mechanisms by which RNA is cleaved in Type III systems:
crRNA-guided RNAse activity or crRNA-independent RNAse activity. In guided RNase activity,
Csm3 in 1Il-A systems (Samai et al., 2015; Staals et al., 2014; Tamulaitis et al., 2014) or Cmr4 in III-B
systems (Benda et al., 2014; Ramia et al., 2014) associates with the target crRNA and cleaves the
RNA transcript. RNase activity independent of the crRNA is mediated through Csm6 in Type Il1-A
(Jiang et al., 2016; Niewoehner and Jinek, 2016) and Csxl, which is a Csm6 ortholog in Type III-B
(Kim et al., 2013; Lintner et al., 2011; Sheppard et al., 2016), and no stable complex formation with
the crRNA or crRNA binding proteins is observed. Both targeted and non-targeted RNAses help
mediate interference against foreign nucleic acids although the mechanism by which trans acting
16
systems help mediate interference is still under investigation. There are a few other notable
features of Type III systems relating to its activity as an adaptive immune system. Many Type III
systems do not have a casi gene within its locus (Makarova et al., 2015) and targeting of RNA is
often PAM-independent (Hale et al., 2009). As mentioned before, target cleavage in PAM-
independent systems is achieved due to the absence of base pairing proximal to the protospacer.
Lastly, since Cast is the only known protein required for adaptation, it is not known whether some
Type III systems lack adaptation capabilities, utilize Cast in trans, or utilize other unknown
mechanisms.
Part II: Genome Engineering using Class 2 CRISPR-Cas Systems
In the previous section, the basic biology of several different CRISPR-Cas systems was
introduced to provide a background understanding of CRISPR-Cas origins, function, and
categorization. In this part, CRISPR-Cas systems will be discussed with respect to their utility as
genome or transcriptome editing tools. The ability of CRISPR-Cas systems to undergo targeted
nucleic acid cleavage within bacteria or archaea represents a system that can be repurposed for
targeted editing of nucleic acids in other organisms. Harnessed CRISPR-Cas systems can provide a
simple system for genome or transcriptome modulation and can also be applied for multiple uses in
other biological assays. Class 2 CRISPR-Cas systems are employed for genome engineering or assay
development simply because there are fewer components to engineer compared to Class 1 CRISPR-
Cas systems. The targeting realm of useful genome or transcriptome editing CRSIPR-Cas
technologies and related assays are dependent on the diversity of characterized Class 2 CRISPR-Cas
systems.
Genome editing and predecessors to CRISPR-Cas systems
Traditional genome editing occurs by targeting DNA double stranded breaks (DSBs) and
utilizing natural repair processes within cells to insert, delete, or alter the genetic code. Although
genome editing can happen spontaneously by simply providing an edited template (Smithies et al.,
1985; Thomas et al., 1986), this process is exceedingly rare and can even result in higher
incorporation of the fragment in other parts of the genome (Lin et al., 1985). Introduction of a DSB
near the site of interest increases targeted genome editing efficiency (Bibikova et al., 2001;
Bibikova et al., 2002) because it stimulates efficient cellular DSB repair mechanisms known as non-
homologous end joining (NHEJ) and homology-directed repair (HDR). NHEJ is the predominant
17
repair pathway often guided by microhomology of the DNA (Roth and Wilson, 1986), but is an
error-prone process and frequently results in small insertions or deletions (indels) (Roth et al.,
1985). Indels occurring within a protein-coding gene have a 2/3 chance of resulting in a frameshift
mutation leading to a premature stop codon and gene knockout. Alternatively, HDR leads to precise
gene editing. However, HDR requires a homology template and does not function in non-dividing
cells (Symington and Gautier, 2011).
Genome engineering has origins in a three other systems: homing endonucleases
(H Es)(Smith et al., 2006; Stoddard, 2005), Zinc Finger Nucleases (ZFs) (Kim et al., 1996; Miller et al.,
2007; Porteus and Baltimore, 2003), and transcription activator-like effector nucleases
(TALENs)(Christian et al., 2010; Miller et al., 2011; Zhang et al., 2011). HEs are sequence specific
endonucleases that originate from microbial mobile genetic elements (Smith et al., 2006). ZFs are
derived from modular eukaryotic transcription factor DNA binding domains and fused to the non-
specific DNA cleavage domain of the FokI endonuclease (Kim et al., 1996). TALENs are modular
DNA binding domains derived from a family of proteins called the transcription activator-like
effectors (TALE) from the plant pathogen Xanthomonas genus, and, like ZFs, the TALE binding
module is fused to the FokI endonuclease domain to create the TALEN (Christian et al., 2010).
Although these systems have some limitations with respect to either specificity for HEs (Smith et
al., 2006) or module interference for ZFs (Maeder et al., 2008) and TALENs (Juillerat et al., 2014), a
key distinction is that these three methods employ protein-based DNA targeting modalities in
which sequence specificity is conferred through direct protein-DNA interactions. Therefore, a
unique protein must be cloned, expressed, and tested for each specific targeted DNA sequence,
which requires a lengthy cloning process that is not scalable. CRISPR-Cas systems offer a notable
advantage over protein based targeting modalities because sequence specificity is determined by
RNA-DNA (or in the case of RNA-targeting CRISPR-Cas systems RNA-RNA) interactions and guide
RNA synthesis is simple and easy to complete compared to engineering a whole protein. CRISPR-
Cas systems also offer a promising potential to develop RNA based transcriptome engineering tools
that can mediate RNA modification other than knockdown which can be accomplished using
existing siRNA or miRNA technologies.
Currently developed and promising CRISPR-Cas systems in genome and transcriptome engineering
To use CRISPR-Cas systems in heterologous genome or transcriptome editing contexts, the
system must be identified, characterized, harnessed to function in the organism of choice, and
analyzed for genome-wide targeted efficiency or developed into molecular tools. The Class 2
18
enzymes described earlier fall along this spectrum: some have limited characterization whereas
others have undergone several targeted cleavage analyses. For an illustration of Class 2 CRISPR-Cas
systems and their development for molecular tools, see figure 5. Currently, there are only a few
CRISPR-Cas systems harnessed for mammalian genome editing and they exclusively target DNA. In
this section, the Class 2 CRISPR-Cas enzymes with respect to their functional understanding and
repurposing as a genome-editing tool will be described.
Functional Understanding and Repurposing as a
Genomne or Transcriptomne Engineering Tool
Identified Characterized ~Harnessed
Figure 5: Functional characterization of Class 2 CRISPR-Cas systems for use as genome
editing or transcriptome editing tools.
The first known CRISPR-Cas systems were discovered empirically due to their high
abundance in nature whereas other more rare CRISPR-Cas systems were discovered through large
computational searches. The first empirically discovered CRISPR-Cas systems were Types I, II, and
III which are relatively common (Makarova et al., 2011). Next, while analyzing the genomic DNA of
the intracellular pathogen Franciscella tularensis, the Type V cpfl gene was first discovered as an
uncharacterized gene next to casi, cas2, cas4, and a CRISPR array (Schunder et al., 2013) and later
found in other bacterial species (Makarova et al., 2015). With increasing diversity of CRISPR-Cas
systems and successful repurposing of Cas9 as a genome-engineering tool, an initial comprehensive
search for uncharacterized Class 2 CRISPR proteins was seeded on the presence of casl(Shmakov et
al., 2015). It was seeded on casi because it encodes the most highly conserved Cas protein
(Takeuchi et al., 2012) and is present in most CRISPR-Cas loci (Makarova et al., 2011; Makarova et
al., 2015). From this search, the three large single effector Cas enzymes C2cl, C2c2, and C2c3 were
found (Shmakov et al., 2015) and later categorized as Types V or VI. After computational discovery,
the system can be further characterized to ascertain predicted functional nuclease domains, tested
for expression as well as processing of the CRISPR array, screened for protospacer adjacent
19
sequence requirements (such as a PAM) for targeting, and illustrate targeted nucleic acid cleavage
biochemically or genetically (Abudayyeh et al., 2016; Shmakov et al., 2015; Zetsche et al., 2015).
To utilize CRISPR-Cas systems for targeted genome engineering in heterologous organisms
such as mammals, both the Cas proteins as well as the guide RNAs must undergo specific
modifications for proper expression and localization in the cell type of interest. Guide RNAs need to
be expressed by a promoter recognized by endogenous transcription machinery. In mammalian
cells, the RNA polymerase III U6 promoter is used because it is short and able to transcribe short
RNAs (Miyagishi and Taira, 2002). Second, if a tracrRNA is required, such as for Cas9, a chimeric
crRNA-tracrRNA hybrid RNA can be used to reduce the number of expressed RNAs required for
targeted genome editing (Jinek et al., 2012). Lastly, the spacer sequence of the guide RNA must be
chosen in the genomic area of interest with the appropriate flanking PAM sequence.
Several modifications are also required for Cas protein function in a mammalian cell.
Different orthologs of CRISPR-Cas nucleases display a range of catalytic activity as assayed by indel
frequency and PAM sequence requirements (Ran et al., 2015; Zetsche et al., 2015). Orthologs in
mammalian cells that express well and display robust as well as consistent levels of targeted DNA
cleavage have traditionally been selected for genome editing purposes. PAM sequence
requirements can also vary across multiple orthologs (Cong et al., 2013; Ran et al., 2015) and may
play a role in the selection of a particular Cas protein for genome editing. After an active and
specific Cas ortholog is identified, the cas gene is codon optimized for efficient translation in the
specific organism and nuclear localization sequence tags are added so that the Cas9 protein
localizes to the nucleus (Cong et al., 2013). Varying promoters can be used to allow for tissue
specific or ubiquitous constitutive expression.
The first CRISPR associated single effector enzyme to be repurposed as a mammalian
genome-editing tool was Cas9 (Cong et al., 2013; Mali et al., 2013b). The first and most well
characterized Cas9 ortholog comes from the species Streptococcus pyogenes SF370, termed SpCas9,
which is a large 1368aa protein with a short, permissive NGG PAM (Figure 1). Several other
orthologs of Cas9 have been harnessed for mammalian genome editing including Staphylococcus
aureus, a notable smaller ortholog of 1053aa but a longer PAM of NNGRR(T) (Ran et al., 2015).
Orthologs of Cas9 are widely used to mediate genome editing within in vitro or in vivo
environments and have undergone several genome-wide off-target cleavage analyses in
mammalian cells to better understand their on-target and off-target cleavage patterns (Kim et al.,
2015; Tsai et al., 2015).
20
In addition to its use as a way to edit the genome, its specific activity, ease of targeting, and
modularity have enabled Cas9 to be repurposed for many other assays, summarized in figure 6.
Cas9 can be used in in genome-wide knockout screens in which a pooled library of guide RNAs can
be delivered to a cell population through lentiviral vectors and then cells can undergo positive or
negative selection (Shalem et al., 2014). Cas9 based screening functions have similar applications
to other screening methodologies such as RNA interference (RNAi), but results in protein knockout
instead of knockdown and can have improved specificity. Alternatively, catalytically inactive Cas9,
or dCas9, can be used as a programmable DNA binding protein. For targeted transcription
initiation, dCas9 is either directly fused to a transcriptional activator such as VP64 (Maeder et al.,
2013; Mali et al., 2013a; Perez-Pinera et al., 2013) and/or the guide RNA can undergo structure-
guided alterations to direct localization of VP64. A MS2 RNA stem loop, from the MS2
bacteriophage, can be added to a permissive area of the guide RNA and co-transfected with an
engineered MS2 coat protein fused to the VP64 protein, in which the MS2 coat protein strongly and
specifically binds to the MS2 RNA stem loop (Konermann et al., 2015; Zalatan et al., 2015). Also,
dCas9 has also been used for targeted DNA imaging (Tanenbaum et al., 2014) or to alter epigenetic
genetic markers such as targeted histone demethylase LSD1 (Kearns et al., 2015).
21
G H
TunsdptIonaW net AIC
ControW "W"On DNA labeling Inducible regulation
Figure 6: Applications of CRISPR-Cas9. Figure from (Hsu et al., 2014). A. Cas9 can be used for
targeted genomic edits as a nuclease, nickase or B. double nickase. C. All components for Cas9 can
be delivered on a plasmid or D. direct injection of purified Cas9 protein and guide RNA. E. Viral
vectors can be used to deliver CRISPR reagents for in vivo targeting. F. Cas9 is useful in genome-
wide screening assays, G. activating endogenous transcription, mediating epigenetic modulation,
and H. live-imaging labeling of DNA loci. I. Cas9 proteins have been engineered in various ways,
including light-inducible regulation.
Cpfl has also been studied and harnessed for mammalian genome editing (Zetsche et al.,
2015). Cpfl has unique differences compared to Cas9, which can be exploited and used to
complement Cas9. Unlike Cas9 enzymes, which have G-rich PAM sequences, Cpfl enzymes have a T-
rich PAM as illustrated by the ortholog Francisella novicida U112 with a 5'-TTN PAM. Cpfl is also
naturally targeted to the cognate DNA using only the single crRNA which is simpler than the crRNA-
tracrRNA duplex used by Cas9 and shorter than the chimeric Cas9 guide RNA. Cpfl is also unique in
that the enzyme itself cleaves its own CRISPR array into individual crRNA units which is especially
promising for multiplexed genome editing because only the CRISPR array containing the targeting
22
A BJI
Nuolease Nickase HNH+ Nickase RuvC+
Double nickng
C D E
PCC
Guide Cloning Mutipe Enrichment
li brary guide RNA Lentips lba n s of desired
synthesis plainld lbrary &nO* ;M phenotype
F
spacers is required for Cpf1. Lastly, Cpfl cleavage results in a staggered cut unlike the blunt cut of
Cas9, which may aid in directing homologous recombination. Although Cpfl exhibits notable
differences compared to Cas9, it could be just as useful in many of the same applications as Cas9
such as screening or imaging.
There exist more opportunities to develop more DNA editing tools beyond the
characterized orthologs of Cas9 and Cpfl. First, uncharacterized orthologs or engineered versions
of characterized enzymes could be explored. Characterizing other orthologs has produced genome
engineering tools with reduced DNA sizes such as SaCas9 (Ran et al., 2015) that are ideal for viral
delivery methods with packaging limits. Structure-guided engineering has produced genome
engineering tools with altered PAM specificity (Kleinstiver et al., 2015) and engineered Cas9
enzymes with enhanced cutting specificity (Kleinstiver et al., 2016; Slaymaker et al., 2016).
Alternatively, there are other Type V and Type V-like Class 2 enzymes that could be adapted for
genome or transcriptome editing including C2c1, C2c3, CasX, and CasY. C2c1 and C2c3 are formally
categorized as Type V whereas CasX and CaxY have similar predicted domain organization.
Available data suggests that these four unique enzymes may behave similarly to Cpfl. For example,
solved crystal structures show that C2c1 is structurally similar to Cpfl and biochemically known to
generate staggered DNA cuts distal to its T-rich PAM (Liu et al., 2017a; Shmakov et al., 2015; Yang
et al., 2016). CasY has local amino acid alignment similarities to C2c3, and cuts DNA with a TA PAM.
CasX also cuts DNA with a T-rich TTC PAM (Burstein et al., 2017). Although known differences
include that C2c1 uses a tracrRNA and CasX is a small 980 amino acid protein, identifying additional
unique properties about each protein could lead to different applications.
Repurposing the first characterized single effector CRISPR-associated RNase C2c2 would
enable an entirely new area of RNA or transcriptome engineering. No existing CRISPR-Cas system
can simply target endogenous mammalian RNA; for example, Cas9 is capable of targeting RNA but
requires an unnatural dsDNA at the PAM (O'Connell et al., 2014) RNA-targeting CRISPR-systems
could be used for endogenous RNA knockdown, translation control, splicing control, localization
studies or in other applications for targeted RNA-binding assays. C2c2 has two individual HEPN
domains that come together to form the RNase active site (Liu et al., 2017b) and validating in vitro
cleavage data with purified protein shows cleavage of ssRNA at the RNA base Uracil (Abudayyeh et
al., 2016). C2c2 also cleaves its own CRISPR array (East-Seletsky et al., 2016) similar to Cpfl, which
could aid in multiplexed targeting of multiple RNA sequences. C2c2 also exhibits a cleavage
phenomenon called the collateral effect, which is promiscuous RNase cleavage activity after initial
targeted RNA cleavage (Abudayyeh et al., 2016). Although Type III systems also have non-targeted
23
RNA cleavage activity, C2c2 is unique in that the protein itself first undergoes targeted and then
nonspecific cleavage. This unique property has been proposed to be useful in development for RNA
detection assays (East-Seletsky et al., 2016) where the RNA for detection is mixed with the C2c2
protein, targeting crRNA, and non-targeted fluorescent-quenched RNAs such that when C2c2
cleaves its target, it also cleaves the fluorescent-quenched RNAs to enable detection. Although C2c2
is a promising prospect for targeted RNA modulation or detection, there are only 30 known
orthologs, a rarity compared to the 3,822 Cas9 orthologs (Shmakov et al., 2017), and fewer
orthologs reduces the exploitable diversity offered by the system. Identifying additional C2c2
orthologs or other unique RNA-targeting single effector Cas enzymes would increase the diversity
of the RNA targeting CRISPR-Cas toolbox.
Based on the utility of characterizing and harnessing unique Cas proteins and orthologs,
there exists a desire to conduct an exhaustive search to find additional Class 2 CRISPR associated
enzymes for genome or transcriptome engineering purposes. We hypothesized that there may be
additional uncharacterized CRISPR-Cas systems and that additional Cas single effector enzymes
could be found by simply seeding a search on the presence of a large protein near a CRISPR array.
As seen in some Type III systems, casi is sometimes not present in active CRISPR-Cas loci and
therefore are not required for CRISPR-Cas interference activity. Any uncharacterized protein
identified through a search can be computationally analyzed to identify any conserved catalytic
domains and then biochemically or genetically characterized in the lab.
24
References:
Abudayyeh, 0.0., Gootenberg, J.S., Konermann, S., Joung, J., Slaymaker, I.M., Cox, D.B.,
Shmakov, S., Makarova, K.S., Semenova, E., Minakhin, L., et al. (2016). C2c2 is a single-
component programmable RNA-guided RNA-targeting CRISPR effector. Science 353,
aaf5573.
Anantharaman, V., Makarova, K.S., Burroughs, A.M., Koonin, E.V., and Aravind, L. (2013).
Comprehensive analysis of the HEPN superfamily: identification of novel roles in intra-
genomic conflicts, defense, pathogenesis and RNA processing. Biol Direct 8, 15.
Barrangou, R., Fremaux, C., Deveau, H., Richards, M., Boyaval, P., Moineau, S., Romero, D.A.,
and Horvath, P. (2007). CRISPR provides acquired resistance against viruses in
prokaryotes. Science 315, 1709-1712.
Benda, C., Ebert, J., Scheltema, R.A., Schiller, H.B., Baumgartner, M., Bonneau, F., Mann, M.,
and Conti, E. (2014). Structural model of a CRISPR RNA-silencing complex reveals the RNA-
target cleavage activity in Cmr4. Mol Cell 56, 43-54.
Bibikova, M., Carroll, D., Segal, D.J., Trautman, J.K., Smith, J., Kim, Y.G., and Chandrasegaran,
S. (2001). Stimulation of homologous recombination through targeted cleavage by chimeric
nucleases. Mol Cell Biol 21, 289-297.
Bibikova, M., Golic, M., Golic, K.G., and Carroll, D. (2002). Targeted chromosomal cleavage
and mutagenesis in Drosophila using zinc-finger nucleases. Genetics 161, 1169-1175.
Biertumpfel, C., Yang, W., and Suck, D. (2007). Crystal structure of T4 endonuclease VII
resolving a Holliday junction. Nature 449, 616-620.
Bolotin, A., Quinquis, B., Sorokin, A., and Ehrlich, S.D. (2005). Clustered regularly
interspaced short palindrome repeats (CRISPRs) have spacers of extrachromosomal origin.
Microbiology 151, 2551-2561.
Brouns, S.J., Jore, M.M., Lundgren, M., Westra, E.R., Slijkhuis, R.J., Snijders, A.P., Dickman,
M.J., Makarova, K.S., Koonin, E.V., and van der Oost, J. (2008). Small CRISPR RNAs guide
antiviral defense in prokaryotes. Science 321, 960-964.
Burstein, D., Harrington, L.B., Strutt, S.C., Probst, A.J., Anantharaman, K., Thomas, B.C.,
Doudna, J.A., and Banfield, J.F. (2017). New CRISPR-Cas systems from uncultivated
microbes. Nature 542, 237-241.
Carte, J., Wang, R., Li, H., Terns, R.M., and Terns, M.P. (2008). Cas6 is an endoribonuclease
that generates guide RNAs for invader defense in prokaryotes. Genes Dev 22, 3489-3496.
Christian, M., Cermak, T., Doyle, E.L., Schmidt, C., Zhang, F., Hummel, A., Bogdanove, A.J., and
Voytas, D.F. (2010). Targeting DNA double-strand breaks with TAL effector nucleases.
Genetics 186, 757-761.
Cong, L., Ran, F.A., Cox, D., Lin, S., Barretto, R., Habib, N., Hsu, P.D., Wu, X., Jiang, W.,
Marraffini, L.A., et aL. (2013). Multiplex genome engineering using CRISPR/Cas systems.
Science 339, 819-823.
25
Datsenko, K.A., Pougach, K., Tikhonov, A., Wanner, B.L., Severinov, K., and Semenova, E.
(2012). Molecular memory of prior infections activates the CRISPR/Cas adaptive bacterial
immunity system. Nat Commun 3, 945.
Deltcheva, E., Chylinski, K., Sharma, C.M., Gonzales, K., Chao, Y., Pirzada, Z.A., Eckert, M.R.,
Vogel, J., and Charpentier, E. (2011). CRISPR RNA maturation by trans-encoded small RNA
and host factor RNase III. Nature 471, 602-607.
Deveau, H., Barrangou, R., Garneau, J.E., Labonte, J., Fremaux, C., Boyaval, P., Romero, D.A.,
Horvath, P., and Moineau, S. (2008). Phage response to CRISPR-encoded resistance in
Streptococcus thermophilus. J Bacteriol 190, 1390-1400.
Dong, D., Ren, K., Qiu, X., Zheng, J., Guo, M., Guan, X., Liu, H., Li, N., Zhang, B., Yang, D., et al.
(2016). The crystal structure of Cpf1 in complex with CRISPR RNA. Nature 532, 522-526.
Doolittle, W.F., and Sapienza, C. (1980). Selfish genes, the phenotype paradigm and genome
evolution. Nature 284, 601-603.
East-Seletsky, A., O'Connell, M.R., Knight, S.C., Burstein, D., Cate, J.H., Tjian, R., and Doudna,
J.A. (2016). Two distinct RNase activities of CRISPR-C2c2 enable guide-RNA processing and
RNA detection. Nature 538, 270-273.
Elmore, J.R., Sheppard, N.F., Ramia, N., Deighan, T., Li, H., Terns, R.M., and Terns, M.P.
(2016). Bipartite recognition of target RNAs activates DNA cleavage by the Type Ill-B
CRISPR-Cas system. Genes Dev 30, 447-459.
Gasiunas, G., Barrangou, R., Horvath, P., and Siksnys, V. (2012). Cas9-crRNA
ribonucleoprotein complex mediates specific DNA cleavage for adaptive immunity in
bacteria. Proc Natl Acad Sci U S A 109, E2579-2586.
Goldfarb, T., Sberro, H., Weinstock, E., Cohen, 0., Doron, S., Charpak-Amikam, Y., Afik, S.,
Ofir, G., and Sorek, R. (2015). BREX is a novel phage resistance system widespread in
microbial genomes. EMBO J 34, 169-183.
Gorecka, K.M., Komorowska, W., and Nowotny, M. (2013). Crystal structure of RuvC
resolvase in complex with Holliday junction substrate. Nucleic Acids Res 41, 9945-9955.
Grynberg, M., Erlandsen, H., and Godzik, A. (2003). HEPN: a common domain in bacterial
drug resistance and human neurodegenerative proteins. Trends Biochem Sci 28, 224-226.
Hale, C.R., Zhao, P., Olson, S., Duff, M.O., Graveley, B.R., Wells, L., Terns, R.M., and Terns, M.P.
(2009). RNA-guided RNA cleavage by a CRISPR RNA-Cas protein complex. Cell 139, 945-
956.
Hein, S., Scholz, I., Voss, B., and Hess, W.R. (2013). Adaptation and modification of three
CRISPR loci in two closely related cyanobacteria. RNA Biol 10, 852-864.
Heler, R., Samai, P., Modell, J.W., Weiner, C., Goldberg, G.W., Bikard, D., and Marraffini, L.A.
(2015). Cas9 specifies functional viral targets during CRISPR-Cas adaptation. Nature 519,
199-202.
Hsu, P.D., Lander, E.S., and Zhang, F. (2014). Development and applications of CRISPR-Cas9
for genome engineering. Cell 157, 1262-1278.
26
Huo, Y., Nam, K.H., Ding, F., Lee, H., Wu, L., Xiao, Y., Farchione, M.D., Jr., Zhou, S., Rajashankar,
K., Kurinov, I., et aL. (2014). Structures of CRISPR Cas3 offer mechanistic insights into
Cascade-activated DNA unwinding and degradation. Nat Struct Mol Biol 21, 771-777.
Jansen, R., Embden, J.D., Gaastra, W., and Schouls, L.M. (2002). Identification of genes that
are associated with DNA repeats in prokaryotes. Mol Microbiol 43, 1565-1575.
Jiang, W., Samai, P., and Marraffini, L.A. (2016). Degradation of Phage Transcripts by
CRISPR-Associated RNases Enables Type III CRISPR-Cas Immunity. Cell 164, 710-721.
Jinek, M., Chylinski, K., Fonfara, I., Hauer, M., Doudna, J.A., and Charpentier, E. (2012). A
programmable dual-RNA-guided DNA endonuclease in adaptive bacterial immunity.
Science 337, 816-821.
Jinek, M., Jiang, F., Taylor, D.W., Sternberg, S.H., Kaya, E., Ma, E., Anders, C., Hauer, M., Zhou,
K., Lin, S., et aL. (2014). Structures of Cas9 endonucleases reveal RNA-mediated
conformational activation. Science 343, 1247997.
Juillerat, A., Dubois, G., Valton, J., Thomas, S., Stella, S., Marechal, A., Langevin, S., Benomari,
N., Bertonati, C., Silva, G.H., et aL. (2014). Comprehensive analysis of the specificity of
transcription activator-like effector nucleases. Nucleic Acids Res 42, 5390-5402.
Kearns, N.A., Pham, H., Tabak, B., Genga, R.M., Silverstein, N.J., Garber, M., and Maehr, R.
(2015). Functional annotation of native enhancers with a Cas9-histone demethylase fusion.
Nat Methods 12, 401-403.
Kim, D., Bae, S., Park, J., Kim, E., Kim, S., Yu, H.R., Hwang, J., Kim, J.I., and Kim, J.S. (2015).
Digenome-seq: genome-wide profiling of CRISPR-Cas9 off-target effects in human cells. Nat
Methods 12, 237-243, 231 p following 243.
Kim, Y.G., Cha, J., and Chandrasegaran, S. (1996). Hybrid restriction enzymes: zinc finger
fusions to Fok I cleavage domain. Proc Natl Acad Sci U S A 93, 1156-1160.
Kim, Y.K., Kim, Y.G., and Oh, B.H. (2013). Crystal structure and nucleic acid-binding activity
of the CRISPR-associated protein Csx1 of Pyrococcus furiosus. Proteins 81, 261-270.
Kleinstiver, B.P., Pattanayak, V., Prew, M.S., Tsai, S.Q., Nguyen, N.T., Zheng, Z., and Joung, J.K.
(2016). High-fidelity CRISPR-Cas9 nucleases with no detectable genome-wide off-target
effects. Nature 529, 490-495.
Kleinstiver, B.P., Prew, M.S., Tsai, S.Q., Topkar, V.V., Nguyen, N.T., Zheng, Z., Gonzales, A.P.,
Li, Z., Peterson, R.T., Yeh, J.R., et aL. (2015). Engineered CRISPR-Cas9 nucleases with altered
PAM specificities. Nature 523, 481-485.
Konermann, S., Brigham, M.D., Trevino, A.E., Joung, J., Abudayyeh, 0.0., Barcena, C., Hsu,
P.D., Habib, N., Gootenberg, J.S., Nishimasu, H., et aL. (2015). Genome-scale transcriptional
activation by an engineered CRISPR-Cas9 complex. Nature 517, 583-588.
Labrie, S.J., Samson, J.E., and Moineau, S. (2010). Bacteriophage resistance mechanisms. Nat
Rev Microbiol 8, 317-327.
Lee, K.P., Dey, M., Neculai, D., Cao, C., Dever, T.E., and Sicheri, F. (2008). Structure of the dual
enzyme Irel reveals the basis for catalysis and regulation in nonconventional RNA splicing.
Cell 132, 89-100.
27
Lin, F.L., Sperle, K., and Sternberg, N. (1985). Recombination in mouse L cells between DNA
introduced into cells and homologous chromosomal sequences. Proc Natl Acad Sci U S A 82,
1391-1395.
Lintner, N.G., Frankel, K.A., Tsutakawa, S.E., Alsbury, D.L., Copie, V., Young, M.J., Tainer, J.A.,
and Lawrence, C.M. (2011). The structure of the CRISPR-associated protein Csa3 provides
insight into the regulation of the CRISPR/Cas system. J Mol Biol 405, 939-955.
Liu, L., Chen, P., Wang, M., Li, X., Wang, J., Yin, M., and Wang, Y. (2017a). C2c1-sgRNA
Complex Structure Reveals RNA-Guided DNA Cleavage Mechanism. Mol Cell 65, 310-322.
Liu, L., Li, X., Wang, J., Wang, M., Chen, P., Yin, M., Li, J., Sheng, G., and Wang, Y. (2017b). Two
Distant Catalytic Sites Are Responsible for C2c2 RNase Activities. Cell 168, 121-134 e112.
Maeder, M.L., Linder, S.J., Cascio, V.M., Fu, Y., Ho, Q.H., and Joung, J.K. (2013). CRISPR RNA-
guided activation of endogenous human genes. Nat Methods 10, 977-979.
Maeder, M.L., Thibodeau-Beganny, S., Osiak, A., Wright, D.A., Anthony, R.M., Eichtinger, M.,
Jiang, T., Foley, J.E., Winfrey, R.J., Townsend, J.A., et aL. (2008). Rapid "open-source"
engineering of customized zinc-finger nucleases for highly efficient gene modification. Mol
Cell 31, 294-301.
Makarova, K.S., Anantharaman, V., Grishin, N.V., Koonin, E.V., and Aravind, L. (2014). CARF
and WYL domains: ligand-binding regulators of prokaryotic defense systems. Front Genet
5, 102.
Makarova, K.S., Haft, D.H., Barrangou, R., Brouns, S.J., Charpentier, E., Horvath, P., Moineau,
S., Mojica, F.J., Wolf, Y.I., Yakunin, A.F., et aL. (2011). Evolution and classification of the
CRISPR-Cas systems. Nat Rev Microbiol 9, 467-477.
Makarova, K.S., Wolf, Y.I., Alkhnbashi, O.S., Costa, F., Shah, S.A., Saunders, S.J., Barrangou, R.,
Brouns, S.J., Charpentier, E., Haft, D.H., et aL. (2015). An updated evolutionary classification
of CRISPR-Cas systems. Nat Rev Microbiol 13, 722-736.
Mali, P., Aach, J., Stranges, P.B., Esvelt, K.M., Moosburner, M., Kosuri, S., Yang, L., and Church,
G.M. (2013a). CAS9 transcriptional activators for target specificity screening and paired
nickases for cooperative genome engineering. Nat Biotechnol 31, 833-838.
Mali, P., Yang, L., Esvelt, K.M., Aach, J., Guell, M., DiCarlo, J.E., Norville, J.E., and Church, G.M.
(2013b). RNA-guided human genome engineering via Cas9. Science 339, 823-826.
Marraffini, L.A., and Sontheimer, E.J. (2008). CRISPR interference limits horizontal gene
transfer in staphylococci by targeting DNA. Science 322, 1843-1845.
Marraffini, L.A., and Sontheimer, E.J. (2010). Self versus non-self discrimination during
CRISPR RNA-directed immunity. Nature 463, 568-571.
Meselson, M., and Yuan, R. (1968). DNA restriction enzyme from E. coli. Nature 217, 1110-
1114.
Miller, J.C., Holmes, M.C., Wang, J., Guschin, D.Y., Lee, Y.L., Rupniewski, I., Beausejour, C.M.,
Waite, A.J., Wang, N.S., Kim, K.A., et aL. (2007). An improved zinc-finger nuclease
architecture for highly specific genome editing. Nat Biotechnol 25, 778-785.
28
Miller, J.C., Tan, S., Qiao, G., Barlow, K.A., Wang, J., Xia, D.F., Meng, X., Paschon, D.E., Leung, E.,
Hinkley, S.J., et al. (2011). A TALE nuclease architecture for efficient genome editing. Nat
Biotechnol 29, 143-148.
Miyagishi, M., and Taira, K. (2002). U6 promoter-driven siRNAs with four uridine 3'
overhangs efficiently suppress targeted gene expression in mammalian cells. Nat
Biotechnol 20, 497-500.
Mojica, F.J., Diez-Villasenor, C., Garcia-Martinez, J., and Almendros, C. (2009). Short motif
sequences determine the targets of the prokaryotic CRISPR defence system. Microbiology
155, 733-740.
Mojica, F.J., Diez-Villasenor, C., Garcia-Martinez, J., and Soria, E. (2005). Intervening
sequences of regularly spaced prokaryotic repeats derive from foreign genetic elements. J
Mol Evol 60, 174-182.
Niewoehner, 0., and Jinek, M. (2016). Structural basis for the endoribonuclease activity of
the type 11-A CRISPR-associated protein Csm6. RNA 22, 318-329.
Nishimasu, H., Cong, L., Yan, W.X., Ran, F.A., Zetsche, B., Li, Y., Kurabayashi, A., Ishitani, R.,
Zhang, F., and Nureki, 0. (2015). Crystal Structure of Staphylococcus aureus Cas9. Cell 162,
1113-1126.
Nishimasu, H., Ran, F.A., Hsu, P.D., Konermann, S., Shehata, S.I., Dohmae, N., Ishitani, R.,
Zhang, F., and Nureki, 0. (2014). Crystal structure of Cas9 in complex with guide RNA and
target DNA. Cell 156, 935-949.
Nunez, J.K., Kranzusch, P.J., Noeske, J., Wright, A.V., Davies, C.W., and Doudna, J.A. (2014).
Cas1-Cas2 complex formation mediates spacer acquisition during CRISPR-Cas adaptive
immunity. Nat Struct Mol Biol 21, 528-534.
O'Connell, M.R., Oakes, B.L., Sternberg, S.H., East-Seletsky, A., Kaplan, M., and Doudna, J.A.
(2014). Programmable RNA recognition and cleavage by CRISPR/Cas9. Nature 516, 263-
266.
Otsuka, Y., and Yonesaki, T. (2005). A novel endoribonuclease, RNase LS, in Escherichia coli.
Genetics 169, 13-20.
Peng, W., Feng, M., Feng, X., Liang, Y.X., and She, Q. (2015). An archaeal CRISPR type Ill-B
system exhibiting distinctive RNA targeting features and mediating dual RNA and DNA
interference. Nucleic Acids Res 43, 406-417.
Perez-Pinera, P., Kocak, D.D., Vockley, C.M., Adler, A.F., Kabadi, A.M., Polstein, L.R., Thakore,
P.I., Glass, K.A., Ousterout, D.G., Leong, K.W., et al. (2013). RNA-guided gene activation by
CRISPR-Cas9-based transcription factors. Nat Methods 10, 973-976.
Porteus, M.H., and Baltimore, D. (2003). Chimeric nucleases stimulate gene targeting in
human cells. Science 300, 763.
Pougach, K., Semenova, E., Bogdanova, E., Datsenko, K.A., Djordjevic, M., Wanner, B.L., and
Severinov, K. (2010). Transcription, processing and function of CRISPR cassettes in
Escherichia coli. Mol Microbiol 77, 1367-1379.
29
Pourcel, C., Salvignol, G., and Vergnaud, G. (2005). CRISPR elements in Yersinia pestis
acquire new repeats by preferential uptake of bacteriophage DNA, and provide additional
tools for evolutionary studies. Microbiology 151, 653-663.
Ramia, N.F., Spilman, M., Tang, L., Shao, Y., Elmore, J., Hale, C., Cocozaki, A., Bhattacharya, N.,
Terns, R.M., Terns, M.P., et aL. (2014). Essential structural and functional roles of the Cmr4
subunit in RNA cleavage by the Cmr CRISPR-Cas complex. Cell Rep 9, 1610-1617.
Ran, F.A., Cong, L., Yan, W.X., Scott, D.A., Gootenberg, J.S., Kriz, A.J., Zetsche, B., Shalem, 0.,
Wu, X., Makarova, K.S., et aL. (2015). In vivo genome editing using Staphylococcus aureus
Cas9. Nature 520, 186-191.
Roth, D.B., Porter, T.N., and Wilson, J.H. (1985). Mechanisms of nonhomologous
recombination in mammalian cells. Mol Cell Biol 5, 2599-2607.
Roth, D.B., and Wilson, J.H. (1986). Nonhomologous recombination in mammalian cells:
role for short sequence homologies in the joining reaction. Mol Cell Biol 6, 4295-4304.
Samai, P., Pyenson, N., Jiang, W., Goldberg, G.W., Hatoum-Aslan, A., and Marraffini, L.A.
(2015). Co-transcriptional DNA and RNA Cleavage during Type III CRISPR-Cas Immunity.
Cell 161, 1164-1174.
Sashital, D.G., Wiedenheft, B., and Doudna, J.A. (2012). Mechanism of foreign DNA selection
in a bacterial adaptive immune system. Mol Cell 46, 606-615.
Schunder, E., Rydzewski, K., Grunow, R., and Heuner, K. (2013). First indication for a
functional CRISPR/Cas system in Francisella tularensis. Int J Med Microbiol 303, 51-60.
Semenova, E., Jore, M.M., Datsenko, K.A., Semenova, A., Westra, E.R., Wanner, B., van der
Oost, J., Brouns, S.J., and Severinov, K. (2011). Interference by clustered regularly
interspaced short palindromic repeat (CRISPR) RNA is governed by a seed sequence. Proc
Natl Acad Sci U S A 108, 10098-10103.
Shalem, 0., Sanjana, N.E., Hartenian, E., Shi, X., Scott, D.A., Mikkelsen, T.S., Heckl, D., Ebert,
B.L., Root, D.E., Doench, J.G., et aL. (2014). Genome-scale CRISPR-Cas9 knockout screening in
human cells. Science 343, 84-87.
Sheppard, N.F., Glover, C.V., 3rd, Terns, R.M., and Terns, M.P. (2016). The CRISPR-associated
Csx1 protein of Pyrococcus furiosus is an adenosine-specific endoribonuclease. RNA 22,
216-224.
Shmakov, S., Abudayyeh, 0.0., Makarova, K.S., Wolf, Y.I., Gootenberg, J.S., Semenova, E.,
Minakhin, L., Joung, J., Konermann, S., Severinov, K., et aL. (2015). Discovery and Functional
Characterization of Diverse Class 2 CRISPR-Cas Systems. Mol Cell 60, 385-397.
Shmakov, S., Smargon, A., Scott, D., Cox, D., Pyzocha, N., Yan, W., Abudayyeh, 0.0.,
Gootenberg, J.S., Makarova, K.S., Wolf, Y.I., et aL. (2017). Diversity and evolution of class 2
CRISPR-Cas systems. Nat Rev Microbiol.
Sinkunas, T., Gasiunas, G., Fremaux, C., Barrangou, R., Horvath, P., and Siksnys, V. (2011).
Cas3 is a single-stranded DNA nuclease and ATP-dependent helicase in the CRISPR/Cas
immune system. EMBO J 30, 1335-1342.
30
Slaymaker, I.M., Gao, L., Zetsche, B., Scott, D.A., Yan, W.X., and Zhang, F. (2016). Rationally
engineered Cas9 nucleases with improved specificity. Science 351, 84-88.
Smith, J., Grizot, S., Arnould, S., Duclert, A., Epinat, J.C., Chames, P., Prieto, J., Redondo, P.,
Blanco, F.J., Bravo, J., et aL. (2006). A combinatorial approach to create artificial homing
endonucleases cleaving chosen sequences. Nucleic Acids Res 34, e149.
Smithies, 0., Gregg, R.G., Boggs, S.S., Koralewski, M.A., and Kucherlapati, R.S. (1985).
Insertion of DNA sequences into the human chromosomal beta-globin locus by homologous
recombination. Nature 317, 230-234.
Snyder, L. (1995). Phage-exclusion enzymes: a bonanza of biochemical and cell biology
reagents? Mol Microbiol 15, 415-420.
Staals, R.H., Zhu, Y., Taylor, D.W., Kornfeld, J.E., Sharma, K., Barendregt, A., Koehorst, J.J.,
Vlot, M., Neupane, N., Varossieau, K., et aL. (2014). RNA targeting by the type I1-A CRISPR-
Cas Csm complex of Thermus thermophilus. Mol Cell 56, 518-530.
Sternberg, S.H., Redding, S., Jinek, M., Greene, E.C., and Doudna, J.A. (2014). DNA
interrogation by the CRISPR RNA-guided endonuclease Cas9. Nature 507, 62-67.
Stoddard, B.L. (2005). Homing endonuclease structure and function. Q Rev Biophys 38, 49-
95.
Suttle, C.A. (2005). Viruses in the sea. Nature 437, 356-361.
Symington, L.S., and Gautier, J. (2011). Double-strand break end resection and repair
pathway choice. Annu Rev Genet 45, 247-2 71.
Takeuchi, N., Wolf, Y.I., Makarova, K.S., and Koonin, E.V. (2012). Nature and intensity of
selection pressure on CRISPR-associated genes. J Bacteriol 194, 1216-1225.
Tamulaitis, G., Kazlauskiene, M., Manakova, E., Venclovas, C., Nwokeoji, A.O., Dickman, M.J.,
Horvath, P., and Siksnys, V. (2014). Programmable RNA shredding by the type I1-A CRISPR-
Cas system of Streptococcus thermophilus. Mol Cell 56, 506-517.
Tanenbaum, M.E., Gilbert, L.A., Qi, L.S., Weissman, J.S., and Vale, R.D. (2014). A protein-
tagging system for signal amplification in gene expression and fluorescence imaging. Cell
159, 635-646.
Thomas, K.R., Folger, K.R., and Capecchi, M.R. (1986). High frequency targeting of genes to
specific sites in the mammalian genome. Cell 44, 419-428.
Tsai, S.Q., Zheng, Z., Nguyen, N.T., Liebers, M., Topkar, V.V., Thapar, V., Wyvekens, N.,
Khayter, C., lafrate, A.J., Le, L.P., et aL. (2015). GUIDE-seq enables genome-wide profiling of
off-target cleavage by CRISPR-Cas nucleases. Nat Biotechnol 33, 187-197.
Wiedenheft, B., van Duijn, E., Bultema, J.B., Waghmare, S.P., Zhou, K., Barendregt, A.,
Westphal, W., Heck, A.J., Boekema, E.J., Dickman, M.J., et aL. (2011). RNA-guided complex
from a bacterial immune system enhances target recognition through seed sequence
interactions. Proc Natl Acad Sci U S A 108, 10092-10097.
31
Yamano, T., Nishimasu, H., Zetsche, B., Hirano, H., Slaymaker, I.M., Li, Y., Fedorova, I.,
Nakane, T., Makarova, K.S., Koonin, E.V., et al. (2016). Crystal Structure of Cpf1 in Complex
with Guide RNA and Target DNA. Cell 165, 949-962.
Yang, H., Gao, P., Rajashankar, K.R., and Patel, D.J. (2016). PAM-Dependent Target DNA
Recognition and Cleavage by C2c1 CRISPR-Cas Endonuclease. Cell 167, 1814-1828 e1812.
Yosef, I., Goren, M.G., and Qimron, U. (2012). Proteins and DNA elements essential for the
CRISPR adaptation process in Escherichia coli. Nucleic Acids Res 40, 5569-5576.
Zalatan, J.G., Lee, M.E., Almeida, R., Gilbert, L.A., Whitehead, E.H., La Russa, M., Tsai, J.C.,
Weissman, J.S., Dueber, J.E., Qi, L.S., et al. (2015). Engineering complex synthetic
transcriptional programs with CRISPR RNA scaffolds. Cell 160, 339-350.
Zetsche, B., Gootenberg, J.S., Abudayyeh, 0.0., Slaymaker, I.M., Makarova, K.S.,
Essletzbichler, P., Volz, S.E., Joung, J., van der Oost, J., Regev, A., et aL. (2015). Cpfl is a single
RNA-guided endonuclease of a class 2 CRISPR-Cas system. Cell 163, 759-771.
Zhang, F., Cong, L., Lodato, S., Kosuri, S., Church, G.M., and Arlotta, P. (2011). Efficient
construction of sequence-specific TAL effectors for modulating mammalian transcription.
Nat Biotechnol 29, 149-15 3.
Zhang, J., Kasciukovic, T., and White, M.F. (2012). The CRISPR associated protein Cas4 Is a
5' to 3' DNA exonuclease with an iron-sulfur cluster. PLoS One 7, e47232.
32
CHAPTER 2
CAS13B IS A TYPE VI-B CRISPR-ASSOCIATED RNA-GUIDED RNASE
DIFFERENTIALLY REGULATED BY ACCESSORY PROTEINS CSX27 AND CSX28
This chapter is adapted from the following article:
Smargon, A.A., Cox, D.B., Pyzocha, N.K., Zheng, K., Slaymaker, I.M., Gootenberg, J.S., Abudayyeh, O.A.,
Essletzbichler, P., Shmakov, S., Makarova, K.S. and Koonin, E.V., 2017. Casl3b is a Type VI-B CRISPR-
associated RNA-Guided RNase differentially regulated by accessory proteins Csx27 and
Csx28. Molecular Cell. 16 Feb 2017
Contributions: A.A.S., D.B.C, and N.K.P. are co-first authors; A.A.S., D.B.C., N.K.P. designed and
performed all experiments and analyzed data under the guidance of F.Z; A.A.S. led the design and
implementation of the CRISPR discovery pipeline with help from K.Z and N.K.P; A.A.S. and D.B.T.C.
performed screening experiments and computational analysis. D.B.C. led the microbiological assays
for characterizing in vivo function of Casl3b, Csx27, and Csx28 with help from J.S.G., O.A.A., and P.E;
N.K.P. purified proteins with help from I.M.S; N.K.P. completed the biochemical characterizations of
Casl3b activity. S.S., K.S.M., and E.V.K provided input on annotation, classification, and naming of
Casl3b, Csx27, and Csx28. A.A.S., D.B.C., N.K.P., and F.Z. wrote the manuscript with input from all
authors.
33
CRISPR-Cas adaptive immune systems defend microbes against foreign nucleic acids via
RNA-guided endonucleases. Using a computational sequence database mining approach, we
identify two Class 2 CRISPR-Cas systems (subtype VI-B) that lack Cast and Cas2 and
encompass a single large effector protein, Casl3b, along with one of two previously
uncharacterized associated proteins, Csx27 or Csx28. We establish that these CRISPR-Cas
systems can achieve RNA interference when heterologously expressed. Through a
combination of biochemical and genetic experiments, we show that Casl3b processes its
own CRISPR array with short and long direct repeats, cleaves target RNA, and exhibits
collateral RNase activity. Using an E. coli essential gene screen, we demonstrate that Casl3b
has a double-sided protospacer-flanking sequence. We also find that Csx27 represses,
whereas Csx28 enhances, Casl3b-mediated RNA interference. Characterization of these
CRISPR systems creates opportunities to develop tools to manipulate and monitor cellular
transcripts.
Introduction
CRISPR-Cas (Clustered Regularly Interspaced Short Palindromic Repeats and CRISPR-
associated proteins) systems are divided into two classes, Class 1 systems, which utilize multiple
Cas proteins and CRISPR RNA (crRNA) to form an effector complex, and the more compact Class 2
systems, which employ a large, single effector with crRNA to mediate interference (Makarova et al.,
2015). CRISPR-Cas systems display a wide evolutionary diversity, involving distinct protein
complexes and different modes of operation, including the ability to target RNA (Abudayyeh et al.,
2016; East-Seletsky et al., 2016; Hale et al., 2009; Jiang et al., 2016; Staals et al., 2013; Staals et al.,
2014; Tamulaitis et al., 2014).
Computational sequence database mining for diverse CRISPR-Cas systems has been carried
out by searching microbial genomic sequences for loci harboring the casi gene, the most highly
conserved cas gene involved in the adaptation phase of CRISPR immunity (Marraffini, 2015).
Among other findings, this approach led to the discovery of the Class 2 subtype VI-A system with its
signature effector Casl3a (previously known as C2c2), which targets RNA (Abudayyeh et al., 2016;
East-Seletsky et al., 2016; Shmakov et al., 2015). Since distinct variants of functional Class 1 CRISPR
systems have been characterized that lack casi (Makarova et al., 2015), we sought to identify Class
2 CRISPR-Cas systems lacking casi by modifying the computational discovery pipeline so that it is
not seeded on Casi. Here we report the characterization of a Class 2 subtype, VI-B, which was
34
discovered through this computational approach, and demonstrate that the VI-B effector, Casl3b, is
an RNA-guided RNase.
Results:
Computational discovery of Class 2 subtype VI-B CRISPR systems
We designed a computational pipeline to search specifically for putative CRISPR-Cas loci
lacking Cast and Cas2 (Figure 1A). Fully assembled microbial genomes were searched for all
proteins within 10kb of CRISPR arrays (Edgar, 2007; Yates et al., 2016). The list of identified loci
was further narrowed down using the following criteria: no more than one neighboring protein
larger than 700aa (to eliminate Class 1 system false positives), presence of a putative single effector
of size 900aa to 1800aa (informed by the size distribution of previously classified Class 2 effectors),
and absence of casi and cas2 genes within 10kb of the CRISPR array (Method Details). Candidate
effectors were grouped into families according to homology (Camacho et al., 2009; Hildebrand et al.,
2009; Remmert et al., 2012), and discarded if they matched previously identified CRISPR-Cas
systems (Makarova et al., 2015). To focus on likely functional CRISPR loci, we limited the candidate
list to families of at least 10 non-redundant effectors in which the putative effector was near a
CRISPR array for at least 50% of the members.
Among the candidates, we identified two genetically diverse putative Class 2 CRISPR-Cas
systems (105 genomic loci, 81 containing a unique entry Casl3b in the non-redundant NCBI protein
database, and 71 of these 81 containing an annotated CRISPR array) represented in Gram-negative
bacteria (Figure 2A). For some genera, in particular Porphyromonas and Prevotella, Cas13b proteins
are encoded in several unique sequenced loci, and, occasionally, in the same sequenced genome.
These systems often co-occur with other CRISPR-Cas systems. Of the 81 type VI-B loci found across
complete and incomplete bacterial genomes, 62 also possess at least one other CRISPR-Cas locus
that includes the key adaptation endonuclease, Cast. However, three complete genomes carrying
the type VI-B locus (Flavobacterium branchiophilum_FL_15_GCA_000253275.1,
Paludibacter-propionicigenesWB4_GCA_000183135.1, and
Porphyromonas-gingivalisAJW4_GCA_001274615.1) may lack Cast altogether.
35
B
1. Compile assembled Ensembi genomes
2. PILER-CR CRISPR locus discovery
3. Nearby protein discovery (within 10kb)
4. Filter CRISPR arrays and proteina
-Arrays with only one neighboring protein >70o a
-Neigboring proteins between O0aa and 1800aa
-No Cast or CaS2 within 10kb of array
5. Classify and select Class 2 candidates
-Pertains BLAST on fitered proteins
-Group into famiies according to homoogy
-Ensure >50% of tamily membe rs ane nearby array
-HHPred and locus analysis tore cassificabion
(19) DR 
Spacer
-1.11.110- Prevotella sp. P5-125
...........- Capnocytophaga canimorsus Cc5
. ------ Phaeodactylibacter xiamenensis
(12)
wmNsu . Porphyromonas gingivalis ATCC 33277
(18)
-0- Bacteroides pyogenes JCM 10003
Alistipes sp. ZOR0009
ft Flavobacterium branchiophilum FL-15
(7)
-inuouauS.- Prevotella sp. MA2016
(9)
Bergeyelfa zoohelcum ATCC 43767
-USUU W Psychroflexus torquis ATCC 700755
(7)
N W Flavobacterium columnare ATCC 49512
(12)
-UUS.W- Riemerella anatipestiferATCC 11845
(5)
-USUSEW Prevotella saccharolytica JCM 17484
(t6) Prevotella buccae ATCC 33574
_)__ Porphyromrtonas gulae
Figure 1: Discovery of two Class 2 CRISPR-Cas systems, subtype VI-B1 and VI-B2, containing
Casl3b. (A) Bioinformatic pipeline to discover putative Class 2 CRISPR loci lacking Casi and Cas2.
(B) A schematic phylogenetic tree of the subtype VI-B loci. Loci with Csx27 (brown) comprise
variant VI-Bi; loci with Csx28 (gold) comprise variant VI-B2. Loci may contain either Csx27, Csx28,
or neither protein in its locus.
All these loci encode a large (-1100aa) candidate effector protein and, in about 80% of the
cases, one different additional small (-200aa) protein (Figures 1B and Figure 2A). The putative
effector proteins contain two predicted HEPN domains (Anantharaman et al., 2013) at their N- and
C-termini (Figure 2B), similar to the domain architecture of the large effector of subtype VI-A
(Casl3a) (Shmakov et al., 2015). Beyond the occurrence of two HEPN domains, however, there is no
significant sequence similarity between the predicted effector and Casl3a. These systems were also
identified by a generalized version of the pipeline described above as part of a comprehensive
analysis of Class 2 CRISPR-Cas systems, and were classified into subtype VI-B, with predicted
effector protein Casl3b (Shmakov et al., 2017).
36
A
Phylogenetic tree of 81 non-redundant
2M!"! "O"'" Ammi62) I CO.slO6Ano.10116 C022A6MeO.62 e.NGI - C6nl1ng2.6 C-2s13im>()
C
L
Casl3b Csx28
HEPN HEPN HEPN
Cono x69 1 x x 6 - x r Cansensus x x x x x J m x
sequence2 : g Seqen A
Identity o -...-.-- o - Identity t -
Figure 2: Phylogenetic tree of Casl3b bifurcates into two variants of subtype VI-B CRISPR
loci. (A) A phylogenetic tree (alignment generated by BLOSUM62) of non-redundant Casl3b
effectors, with the full type VI-B locus depicted in every instance. Accession numbers for genome,
Cas13b (blue), and Csx27 (brown)/Csx28 (gold) are included, as well as number of nearby spacers
detected by PILER-CR, the presence of Casi in the sequenced genome, and the size of Casl3b. (B)
Two HEPN sequences identified via multiple sequence alignment (BLOSUM62) of putative non-
redundant Casl3b proteins. (C) Divergent HEPN sequence identified via multiple sequence
alignment (BLOSUM62) of putative non-redundant Csx28 proteins.
37
A
Cas13b effectors with full gene
22AM SP WO NZJXOIM02121)6
22626462 W. P5119 NZJ222.22A2
IO 266296028272NCL22680 2)
-22-42 3 042I22260226NZ.J2026022040222
22622222222 2.221226ATC 2222NCQI072226 )
P2626228222663222226262226142V22222622 2)
2.26 40022.622QN262W7NZ_.262621)
26622,2626CM 10200 .80)20226201)
AR6226op222222602222...LD1~ 22M91
it 22266266626826222 bfM22..2L65)NC-MM" I6
MyiK 2220266226222 22210202222)OODOT
UG 3822626220226 2222202000620 )
2682626260226222662
h3282M IN2Z-C03W2
06628226222 Y02222Nt2N63217 6W26229&1
R628"6 80222622 264C2222.26 10521
N8262226262""2'-..L2206222600221)
N W= wk2 N2262822262)
206822682 282F8202A262411010 I)RWMM *WM6622ti222 26C 11M 622IS6208222 91
Film262 tii262232 U02022222
U" anate A26P20POO50 1
PA2268 WOUPO622 26 2 222 O .26=222 1)662
R-Ma 232262202.JR4T220 2222 1
P2222e 2666 2460 262.464(NZJ22622 11
2-M 2b-- 2 20283226222OOODDI I
22622226682262267 NZ-4222631121
P-h6220 1-26DI67N246A6230201)20222628 1 426 M NZX73 W-ALJOGOW022 2)
220226206 2226 2TO NZ.M 2242822I 022
Ptsw22ltaO-W266O2W11 - OM2D78 N.AE622 M71
P2626 ". " 6222262T OOI 02222)027
72222aWdrm"17 0222222
2262228226620.Z.022622822
226226622firn6222014162262226
P2~y22222220NZ.L0WL2I
B
_V'- ,
WP04407478D I
WP044=147.1
WP_044005M 1
WP.04018160 I
WP-0130WWI I
WPjNQI8M I
WP_006"11.1
W-WISM75 I
VPM24WI51A
EFUI"W71
M-WM6571
W 023540M 1
E .1
EPj8rM
VVPWIM2,1
YWAM40310.1
W 01111591W. 
IWIP0304450551
EFIIB17001
M. 0345422911
W'P-_."'4064@W' "
WP-039RMMI
EHOONal
EKSM4.1
VW 0005"W 
IM-.414,
EHOOWMII
W-08WOMIN, I
EKOWS&I
W1041911MI 1
WIP-00200"M 1
WP_045960ml
WIP-0111OWNS I
VW-C"410541. I
W-OMMMON, 1
W 003744070 I
W I
Z 074,17981
WP_0040wm 1
W0115345111M 1
W-OM54=1
W081711213111
WP004970W 1
EKYODOIS I
WP.jMISM4841
EFU31M.1
WP_90 I
WP004343M I
YWW412103.1
EGOIN" I
WP000044633.1
WP030800ml
W 08188116M I
AFJDrM I
WF0-050VAM I
SAUISM.1
Kale" I
WP 02500000A
WPCel I
WPGMM4K I
W038884M.'
WPIM94INIZ,
WP039420" 
1WiMP"217i I
WP-03mir/al
WVP O.IW6 I
WP:W0Q48W I
WP 036410MZI
VP_0 76.1
VP _O Iw I
.0131"61. 1
*P.0124594141
WP .05MIg2501
FJOAMM1
wp-OM74,95.1
VW-05MV112.1
WP_053 17' 1
VPO Irm.l
WP.OOIINM7a1
NA
NA
NA
NA
WP OIMM74A
WP-O"2iml I
W-MMMIS'l
WP021804761
WP-012MIS21
ERMISM I
WIP-021arAM 1
W005117WIS 1
EPLJOSM I
EFLA73X,
APOOW-1-
WP0210=9&1
M00597351&1
WP0394451M I
ERIMISM.11
WP0345422791
MA
V& 014084M 1
NA
aKXWWI
EKBOBMSI
OP006205 01
WP-OOBM4151
EHOO47621
WPOWM4 5.1
EICE164194A
WPW1905781
WR W49OW0.1
NA
NA
NA
NA
NA
NA
NA
VPOD4OIVM I
IW-OD4919M.1
W01149197581
W081710139 1
WP-MOGMM I
BCVOOM I
KA
EFLMMIZI
W0043439741
WP-004343WZi
WP-0311,02rMl
830184431
WPOWMEQ()O I
W0507952DO
MA
AFJO73WI
%W'M47 1
BAUIM124
Milem 1
W036811"O
WPWI504705
M10241MIT121
WP0132734&1
WP-052YrJ447
WP-MW357&;
VMOMMM I
W0401110411
VR04=1041 1
W 030434OW I
WP-OM78D41 I
WFOOMAN1721
WF _ WMI
wp-oQNma5 I
W-CM20MI31
WP-043"84MI
EQAIOM&I
Vot-010056MI I
W01005MI 1
WP-04380"M I
WF _021805110& I
WP-OMSBWO I
C
-1
CRISPR-Casl3b loci contain small accessory proteins
The identity of the putative accessory protein correlates with the two distinct branches in
the phylogenetic tree of Casl3b (Figures 1B and 2A) (Henikoff and Henikoff, 1992), indicative of the
existence of two variant systems, which we denote VI-B1 (accessory protein referred to as Csx27)
and VI-B2 (accessory protein referred to as Csx28). While subtype VI-B2 systems almost invariably
contain csx28, csx27 is less consistently represented in VI-B1 loci. The protein sequences of Csx27
and Csx28 show no significant similarity to any previously identified Cas proteins. Both putative
accessory proteins were predicted to contain one or more transmembrane segments (Figure 3A)
(Moller et al., 2001). However, Csx27 of Bergeyella zoohelcum and Csx28 of Prevotella buccae tagged
with RFP at either the N- or C-terminus did not show membrane localization when expressed in E.
coli (Figure 3B). In addition to the predicted hydrophobic domains, analysis of the multiple
sequence alignment of Csx28 proteins indicated the presence of a divergent HEPN domain (Figure
2C).
A
TMHMM posterior probabilities of B. zoohelcum Csx27
1.2
1 .
2 BC024 B. zoohelcum Csx27 P buccae Csx28
0 CLa.
50 100 10 200 Li.
transmembrane - inside outside R FP
TMHMM posterior probabilities of P buccae Csx78
12 Z
0.6 L U
0.2
2 40 0 so 100 120 140 160
transmnmbrane -- inside - outside - -
Figure 3 I Predicted transmembrane domains of Csx2 7 and Csx2 8 not validated
experimentally. (A) Transmembrane domain prediction in Csx27 of B. zoohelcum and Csx28 of P.
buccae using TMHMM v2. (B) N- and C-terminally fused RFP imaging of Csx27 of B. zoohelcum and
Csx28 of P. buccae.
Casl3b-associated CRISPR arrays display unique features
In contrast to their differing putative accessory proteins, both variants of subtype VI-B
systems show distinct, conserved features in the CRISPR arrays. The direct repeats in the CRISPR
arrays are conserved in size, sequence, and structure, with a length of 36 nt, a poly-U stretch in the
38
A CRISPR Class 2 subtype VI-B1 direct repeat RNA folds - 36 nt*
Pre steft p5-l1 Capnocytophaga anummors PoAphymnaz ging al'sPmvote4sp P5-1251 Capnocyophaga cyodegmi 13 Porphyronionas g1ulae Akshpes sp. ZOR0009
F Psy Flavobactenum columnare Paledkbacter propionieglenes Phaeodatylbacter xamenensis
*37nt direct repeat
B Ctorofdes pyogensC, Prc tefla sp MA2016 Prmoill sp P5-60 4 Myroies doratimmit sd
B CRISPR Class 2 subtype VI-B2 direct repeat RNA folds - 36 nt
Pophyromonas gLae Prevotella auraniao Prevotella busae Prow te.o e
Pofphyrmms sp. COT-052 RPevoeeila saccharlyrisa 13 Prevatella MSX73 1 Prevstela intennedsPS1
Porphyramonas gingivah Prevotela falsemi Prevotelle pl,0 r td5s Remeella anapeshtfer
C Conservation of 36 nt direct repeats (24 unique)
GQAAciqU 1 -s- G
Direct repeat
D
Flanking regions of protospacers mapped to phage genomes (36 unique)
0.5-
0.4 -
0.3 - T
M 0.2-T
0 U AU A A
0.0 - g
5' Protospacer
Figure 4: Predicted sequence and secondary structure of type VI-B direct repeats; predicted
protospacer flanking sequences. (A) Predicted secondary structure folds of structurally unique
CRISPR class 2 type VI-B1 direct repeats (Vienna RNAfold). (B) Predicted secondary structure folds
of structurally unique CRISPR Class 2 type VI-B2 direct repeats. (C) Weblogo of all unique VI-B
direct repeat sequences of length 36 nt, taken as the same transcriptional orientation as Casl3b.
(D) Weblogo of all unique VI-B protospacer flanking sequences from CRISPRTarget mapping of
protospacers to phage databases.
39
open loop region, and complementary sequences 5'-GUUG and CAAC-3' at the ends of the repeat
predicted to yield a defined secondary structure mediated by intramolecular base-pairing (Figure
4) (Lorenz et al., 2011). Our analysis revealed 36 Casl3b spacers mapped with greater than 80%
homology to unique protospacers in phage genomes. Twenty-seven of the identified Casl3b
spacers targeted the coding strand of phage mRNA, while seven spacers targeted the noncoding
strand and two spacers targeted regions of the phage genome without predicted transcripts.
Although the composite of these imperfect mappings revealed no consensus flanking region
sequence (Figure 4D) (Biswas et al., 2013), the well-conserved protospacer length of 30 nt,
combined with the conserved direct repeat sequence and length, suggests that the nucleic acid
targeting rules may be similar among different VI-B loci.
RNA sequencing of the total RNA from B. zoohelcum (subtype VI-Bi) showed processing of
the pre-crRNA into a 66-nt mature crRNA, with the full 30-nt 5' spacer followed by the full 36-nt 3'
direct repeat (Figure 5A) (Heidrich et al., 2015; Li and Durbin, 2009; Shmakov et al., 2015). A longer
118-nt crRNA, distal to the 36-nt crRNAs in the CRISPR array and with a direct repeat consisting of
5' and 3' fragments of the 36-nt direct repeat sequence interrupted by an intervening repeat
sequence, was also processed. This phenomenon was computationally predicted to occur in
additional VI-B loci, such as those from Capnocytophaga canimorsus, Myroides odoratimimus, and
Riemerella anatipestifer. Other CRISPR Class 2 effectors are known to process their arrays without
involvement of additional RNases (East-Seletsky et al., 2016; Zetsche et al., 2015). Similarly, we find
that purified BzCas13b is capable of cleaving its CRISPR array, generating mature crRNAs with
short or long direct repeats, an activity which is not affected by mutation of the predicted catalytic
residues of the HEPN domain (Figures 5B and 6). This suggests that an independent RNase domain
cleaves the CRISPR array into individual crRNA units and does not involve the HEPN domains.
Additionally, spacers in the crRNA are not further processed beyond 30 nt, which is the spacer
length in the full CRISPR array.
40
B. zoohelcum ATCC 43767 1201
Read count I&i I ILI
66nt crRNA
-Ont spacer 36nt DR
AC A U
5 -NNNNNNNNNNNNNNNNNNNNNNNNNNNNNNGUUG-GA UGCUCUC U
liii 11 1.1.111 U
3 '-CAAC ACU A AUGGGAG G U
3Ont spacer
5' - NNNNNNNNNNNNNNNNNNNNNNNNNNNNNNGUUG-GA UGCUCUC
11' 1 -C 1A1 111
3' -CAAC A CU A AUGGGAG
S 322nt
160nt
26 13n
U 88nt DR
A U UU C A UU
AUUCGUGAGUUUUU UUG u
G-C GUAAGUACUCAAGGA-AAC U U
U-A
U-A
U-A
A G
C
L =763nt
1 18nt
236ni
HEPN mutants
WT D1 D2 Q
Ladder S S S S S
HEPN mutants
WT D1 D2 Q
L L L L L
* = 36nt DR
= 88nt DR
* = 30nt spacer
4 118nt
4 66nt
Figure 5: Casl3b from the VI-B1 locus processes a CRISPR array with two direct repeat
variants. (A) RNA-Sequencing of the native VI-B1 locus from Bergeyella zoohelcum ATCC 43767.
(B) Denaturing gel showing cleavage products of in vitro synthesized short-DR containing or long-
DR containing CRISPR arrays from the B. zoohelcum genome by either wildtype or HEPN mutant
BzCas13b (D1, R116A/H121A; D2, R1177A/H1182A; Q, R116A/H121A/R1177A/H1182A). The
schematic shows fragment lengths of a cleaved CRISPR array.
41
A
11 8nt crRNA
B
BzCasl3b
CRISPR array
1000 nt
500
300
150
80
50
**b
h7t-_cY97 hyna-Q 1Ah
& 0
Purified Casl3b HEPN mutants
L WT L D1 L D2 L Q
160kDa 0owm Opeo 0vo op 4 BzCasl3b
* *
* 040
BzCasl3b
HEPN Mixed a/P HEPN
R; 6  H;21 R11 77 H1 82
I I I J
Predicted catalytic residues
D1 (R116A/H121A) D2 (R1177A/H1182A) 0 (D1 & D2)
Figure 6: Protein gels of purified WT BzCas13b and three mutant BzCas13b proteins.
Denaturing protein gels of B. zoohelcum wildtype, D1 (R116A/H121A mutant), D2
(R1177A/H1182A) mutant, and Q (R116A/H121A/R1177A/H1182A) mutant Casl3b.
An E. coli essential gene screen reveals targeting rules for BzCasl3b
To validate the expected interference activity of the VI-B system and to determine the
targeting rules for the VI-B1 locus from B. zoohelcum, we developed an E. coli essential gene screen
(Figure 7A). For this negative selection screen, we generated a library of 54,600 unique spacers
tiled with single-nucleotide resolution over the coding region of 45 monocistronic essential genes
(Baba et al., 2006; Gerdes et al., 2003), plus 60 nt into the 5' and 3' UTRs. We also included 1100
randomly generated non-targeting spacers to establish baseline activity. We then transformed this
library with plasmids carrying bzcasl3b (casl3b gene from B. zoohelcum) and bzcsx27, just
bzcasl3b, or a control empty vector. After quality-control-filtering of all screened spacers, we found
a statistically significant depletion of targeting spacers over non-targeting spacers, indicating that
Casl3b, alone or with Csx27, can achieve nucleic acid interference (Figure 7B).
42
(i) Design spacers targeting (ii) Create library of (iii) Transform E. coli with (iv) Select for dual (v) Isolate and sequence
essential E. co/i genes spacer plasmids bzcas13b & bzcsx27 antibiotic resistance spacers from
or bzcasl3b surviving colonies
or
-~~ Fi..
____7__ ....-
- -0 C)l00
pLac pj23150 pLac
mapped over genes Spacer depletions aggregated Spacer depletions mapped over genes Spacer depletions aggregated
across gene distance across gene distance
10'
3 4 5 0 0.2 0.4 0.6 0.8 1 0 1 2 3 4 5 0 02 04 06 08 1
in E coliscreen (x10") Normalized distance along gene Index of spacer in in E. coli screen (x10 ) Normalized distance along gene
Weblogo of strongly depleted sacers (bzcasl3b & bzcsx27) Weblogo of strongly depleted spacers (bzcasl3b)0.5- 0.5-
0.4- 0.4-
e 0.3- 0.3-
0.2- A l 0.2-
0.1- 'A A G o.1- GC
.- - . -
5 PFS Pmrospaier 3: PFS 5' PFS Proospar 3 PFS
Normalized PFS score from safely depleted spacers (bzcasl3b & bzcsx27) Normalized PFS score from safely depleted spacers (bzcasl3b)
A
rnuG
C
AA AG AU AC GA GG GU GC UA UG UU UC CA CG CU CC AA AG AU AC GA GG GU GC A UG UU UC CA CG CU CC
3' PFS 3' PFS
0 0.2 0.4 0.6 0.8 1
5'PFS Protospacer 3' PFS
~NN. -31
Rulel1 Rule 2
D (A, U, or G) NAN or NNA
Composition of spacers in Kan screening library
rules (13 0%)
Non-targeting One rule (14.6%)
(50.3%)
No rules (22.0%)
F
pLac pJ23119 pJ23119
Plate on Cm + Kan
G
10-2
0
is ~ 00o *4,2 00
16 o-4 -1
Ct s-6 abundancer314Control spacer abundanc
Figure 7: Heterologous expression of Casl3b mediates knockdown of E. coli essential genes
by a double-sided PFS. (A) Design of E. coli essential gene screen to determine targeting rules of
nucleic acid interference. (B) Manhattan plots of mean spacer depletions mapped over 45 genes
and aggregated across normalized gene distance for either the full B. zoohelcum VI-B1 locus (left) or
casl3b alone (right), with non-targeting spacers in gray, safely depleted spacers (>So above mean
depletion of non-targeting spacers) above blue line, and strongly depleted spacers (top 1%
depleted) above red line. For the full locus, 36,142 targeting spacers and 630 non-targeting spacers
passed QC filter. Of the targeting, 367 are strongly depleted and 1672 are safely depleted. For
casl3b alone, 35,272 targeting spacers and 633 non-targeting spacers passed QC filter. Of the
targeting, 359 are strongly depleted and 6374 are safely depleted. (C) Weblogo of sequence motifs
of strongly depleted B. zoohelcum spacers. (D) Normalized PFS score matrix, where each score is
the ratio of number of safely depleted B. zoohelcum spacers to total number of spacers for a given
43
A
B
10-
1C
C
Spacer depletions
0 1 2
Index of spacer in
D
E
PFS, scaled so that maximum PFS score is 1. (E) Spacers targeting kanamycin to validate PFS
targeting rules of 5' PFS (D) and 3' PFS (NAN or NNA). (F) Schematic of kanamycin validation
screen for B. zoohelcum casl3b in E. coli. (G) Results from kanamycin validation screen; spacer
abundances versus control for individual B. zoohelcum spacers, with abundances colored by type of
spacer.
To assess the targeting rules for Casl3b, we established two spacer depletion levels:
strongly depleted (top 1% of depleted spacers) and safely depleted (spacers depleted So above the
mean depletion of the filtered non-targeting spacers). From spacers passing the strongly depleted
cutoff we derived sequence motifs qualitatively identifying a double-sided protospacer flanking
sequence (PFS) (Figure 7C) (Crooks et al., 2004). Because each position in a sequence motif is
assumed to be independent, we developed a more quantitative, base-dependent PFS score defined
as the ratio of the number of safely depleted spacers to the number of all spacers with a given PFS,
normalized across all PFS scores (Figure 7D).
The normalized PFS scores revealed a 5' PFS of D (A, U, or G) and 3' PFS of NAN or NNA,
consistent for Casl3b with Csx27, as well as for Casl3b alone. To validate these sequence-targeting
rules, we performed an orthogonal depletion screen with Casl3b alone, targeting the Kanamycin
resistance gene (Figures 7E and 7F). Four classes of spacers were created: non-targeting, targeting
with both 5' and 3' PFS rules, targeting with only the 5' or 3' PFS rule, and targeting with neither
rule. Consistent with our findings from the E. coli essential gene screen, the combined 5' and 3' PFS
spacers resulted in the highest Kanamycin sensitivity (Figure 7G). The kanamycin screen also
provides an additional control compared to the E. coli essential gene screen because it targets the
same transcript with consistent expression levels and consistent local RNA secondary structure.
BzCasl3b cleaves single-stranded RNA and exhibits collateral activity in vitro
Based on the presence of the computationally predicted HEPN domains that function as
RNases in other CRISPR-Cas systems, including VI-A and some Class 1 systems (Abudayyeh et al.,
2016; Kim et al., 2013; Sheppard et al., 2016; Staals et al., 2014), we anticipated that Cas13b
interferes with RNA. We confirmed this by demonstrating that purified Cas13b exclusively cleaves
single-stranded RNA with both direct repeat architectures (Figures 8A and 9A). We then validated
the PFS targeting rules biochemically, showing that a 5' PFS of C greatly inhibits single-stranded
RNA cleavage (Figure 8B), whereas a 3' PFS of NAN or NNA enhances this activity (Figure 8C).
44
A B
Body-labeled ssRNA 2 10
5 11
3 6 12
2 7
14 8
. 99 9 99 9 99 9 #9 9 51 PFS
AGU C
5' labeled Lys ssRNA target
3'-UCCUUCYAGUAUG "A U
5' - CAAMI GAAG UCA- -ACl U
u UACUCUCGUCAAG-GUUGUC tAC LCC UAAUCGCUACACCdLLi:GI -5-
UGGAGGGUA-JCACAAC-3' 30nt spacer
Short DR Long DR
BzCasl3b + + - - + + - -
crRNA Ladder + - + - + - + -
Uncleaved C 75nt - s 0 me so
50nt.
Cleaved
Body-labeled ssRNA 3
5' PFS 3'PFS
5' - AAGGGA MAAGAACAA ..- 3'
9UACUCUCGUCAG-GUUGU-ACUCCUAAUCGCUACA 
CUCUGUUGi -5'
5
UGGAGGGUAA-UCACAAC-3' 30nt spacer
AA AG AU AC GA GG GU GC UA UG UU UC CA CG CU CC
D
A G U C
12 10 1 2 6 8 4 7 11 5 9 3 -
Body-labeled ssRNA 4 3'- UUCUUCUAGUUAUGN N
Nl
5'- C ACN
UACUCUCGUCAAG-GJUGUGLItAC-UAA'VGG(UACAGCUUCUGUUGIIJ 5
UGGAGGGUAA-CACC- 3 30nt spacer
Variable loop base G C A U G C A U
crRNA + + + + - - - -
Uncleaved C 0 4a n
CeI
Cleaved[
I
Figure 8: Casl3b is a programmable single-stranded RNase. (A) Schematic showing the RNA
secondary structure of the cleavage target in complex with a targeting 30-nt spacer connected to
short direct repeat (top). Denaturing gel demonstrating short direct repeat and long direct repeat
crRNA-mediated ssRNA cleavage (bottom). Reactions were incubated for 10 minutes. The ssRNA
target is 5'labeled with IRDye 800. Three cleavage sites are observed. (B) Schematic showing three
numbered protospacers for each colored 5' PFS on a body-labeled ssRNA target (top). Denaturing
gel showing crRNA-guided ssRNA cleavage activity demonstrating the requirement for a D 5' PFS
(not C) (bottom). Reactions were incubated for 60 minutes. crRNAs correspond to protospacer
numbered from the 5' to the 3' end of the target. Gel lane containing RNA ladder not shown. (C)
Schematic of a body-labeled ssRNA substrate being targeted by a crRNA (top). The protospacer
region is highlighted in blue, and the orange bars indicate the 5' PFS and 3' PFS sequences. The
orange letters represent the altered sequences in the experiment. Denaturing gel showing crRNA-
guided ssRNA cleavage activity after 60 minutes of incubation, with the 5' PFS tested as A, and the
3' PFS tested as ANN (bottom). The orange 3' PFS letters represent the RNA bases at the second and
third 3' PFS position within each target ssRNA. Gel lane containing RNA ladder not shown. Dashed
line indicates two separate gels shown side by side. (D) Schematic showing the secondary structure
of the body labeled ssRNA targets used in the denaturing gel. The variable loop of the schematic
(represented as N 5) is substituted with five monomers of the variable loop base in the gel (top).
Denaturing gel showing cleavage bands of the homopolymer variable loop base (bottom). The
targets were incubated for 30 minutes. Dashed line indicates where the image was stitched
together to remove U/A heteropolymer RNA lanes (shown in Figure 9B). Gel lane containing RNA
ladder not shown.
45
5' PFS
crRNA
Uncleaved i
Cleaved
C
3' PFS
Uncleaved
Cleaved
~;-
a~U U
1 m.
.*. , 0w
S '
L
owe I* a as a 6d I,4 aa
4064 iotsa& "laab am a* -
A
dsRNA
Short DR Long DR
BzCasl3b + + - - + + - -
crRNA + - + - + - + -
Uncleaved 1 U U U U U U
Cleaved
B
Variable loop base G C A U U/A G C A U U/A
crRNA + + + + + - - - -
Uncleaved U No 10 " U U 0 0 0 1
Cleaved[
C
ssDNA
Short DR Long DR
+ + - - + + - -
+ - + - + - + -
BzCasl3b + + + - -
crRNA + + - + -
EDTA - + + - -
Uncleaved 1' 4
Cleaved
dsDNA
Short DR Long DR
+ + - - + + --
+ - + - + - + -
D
PbCasl3b + + - -
crRNA + - + -
Uncleaved t I
Cleaved
WT On target
nM
0.594 594
WT Off target
nM
0.594 594
fttWW u
1 10 100
Casl3b+crRNA complex (nM)
Figure 9: Casl3b cleaves and binds to single-stranded RNA. (A) Denaturing gels demonstrating
no cleavage of dsRNA, ssDNA, or dsDNA by BzCasl3b with either the short DR or long DR. Reactions
were incubated for 10 minutes, the same amount of time which results in robust ssRNA cleavage for
this target and crRNA pair. The ssDNA and top strand of the dsDNA target is 5' labeled with IRDye
800. The dsRNA target is body labeled. Gel lane containing RNA ladder not shown. (B) Denaturing
gel showing cleavage bands from the variable loop target as shown in Figure 8D. The U/A
46
D1 On targetE
F
D2 On target 0 On target
W
~wwww
* WT Off Target
* WT On Target
M a I1.0
0.8
0.6
0.4-
0.2
0.0-
0.1
LL
M
UU U 0 0 0
1000
heteropolymer consists of the Ns variable loop of alternating U and A residues (5' AUAUA 3'). (C)
ssRNA cleavage requires BzCas13b and a targeting crRNA, and this cleavage activity is abolished by
addition of EDTA. Gel lane containing RNA ladder not shown. (D) Denaturing gel showing PbCas13b
cleavage activity of an ssRNA targeted substrate. The ssRNA is 5' labeled with IRDye 800 and
incubated for 30 minutes. Gel lane containing RNA ladder not shown. (E) EMSA gels that were used
to quantify the KD of the WT and mutant BzCas13b proteins, using an on-target crRNA
complementary to the targeted ssRNA. (F) EMSA gel of WT BzCas13b with an off-target crRNA. The
off-target crRNA is non-complementary to the targeted ssRNA.
Other HEPN domain-containing CRISPR-Cas RNA-targeting systems, such as Csx1 from the
Type Ill-B CRISPR-Cas systems, preferentially cleave targets containing specific single-stranded
nucleotides (Sheppard et al., 2016). To determine if Casl3b exhibits such a preference, we tested an
RNA substrate with a variable homopolymer loop outside of the spacer:protospacer duplex region
(Figure 8D). A heteropolymer loop consisting of alternating A then U was also tested (Figure 9B).
We observed cleavage at pyrimidine residues, with a strong preference for uracil. This activity is
abolished in the presence of EDTA (Figure 9C), suggesting a divalent metal ion-dependent
mechanism for RNA cleavage akin to that of a similar HEPN-containing, Class 2 effector protein,
Casl3a (Abudayyeh et al., 2016; East-Seletsky et al., 2016).
Given that Cas13a has also been reported to cleave RNA non-specifically once activated by
interaction with the target ("collateral effect") (Abudayyeh et al., 2016; East-Seletsky et al., 2016),
we sought to test the ability of Cas13b to cleave a second, non-specific substrate following target
cleavage. Using an in vitro assay similar to the one we previously used with Casl3a (Abudayyeh et
al., 2016), we incubated Cas13b-crRNA complexes with both a target and non-target RNA substrate.
We observed collateral cleavage of the non-targeted RNA, but only in the presence of the target
RNA (Figure 10).
47
Labeled 1 1 1 1 1 1 1 1 2 2 2 2 2 2 2 2
Unlabeled 2 2 2 2 - - 2 2 1 1 1 1 - - 1 1
crRNA 1A 1B 2A 2B 2A 2B - - 2A 2B 1A 1B 1A 1B - -
Unlabeled ssRNA 1 BzCasl3b + + + + + + + - + + + + + + + -
Spacer 1A Spacer 1B
Body-labeled ssRNA 2
5'- 99 999 9 99 9 99999 -IP ,
Cleaved I
Spacer 2A Spacer 2B LI
Figure 10: Casl3b exhibits collateral activity. Denaturing gel showing BzCasl3b collateral
cleavage activity after 30 minutes of incubation, with schematic of cleavage experiment to the right.
Two crRNAs (A and B) target substrate 1 (1A and 1B) or substrate 2 (2A and 2B). Gel lane
containing RNA ladder not shown. Dashed line indicates two separate gels shown side by side.
Casl3b shows robust HEPN-dependent interference and is repressed by Csx27 activity
To validate RNA interference in vivo, we assayed interference against the lytic, single-
stranded RNA bacteriophage MS2, whose life cycle contains no DNA intermediates. We performed
an MS2 drop plaque assay at serial dilutions of phage for both bzcasl3b with bzcsx27 and bzcasl3b
alone with three spacers targeting the MS2 genome, two at the lys-rep interface and one in rep, as
well as one non-targeting spacer (Figure 11A). We observed substantial reduction in plaque
formation for all targeting spacers compared to the non-targeting spacer, confirming sequence-
specific RNA targeting by VI-B1 systems. (Figures 11A and 12B). Notably, the presence of bzcsx27
weakened RNA interference by bzcasl3b for all three targeting spacers.
To confirm the lack of DNA interference in vivo, we modified an existing plasmid
interference assay with a protospacer placed either in-frame at the 5' end of the bla ampicillin-
resistance gene (transcribed target) or upstream of the bla gene promoter on the opposite strand
(non-transcribed target). Bacteria co-transformed with bzcasl3b and spacer as well as the non-
transcribed target plasmid survived at a comparable rate to co-transformation of the same target
with the empty vector on dual antibiotic selection. For bacteria co-transformed with the
transcribed target, the colony forming unit rate under dual antibiotic selection was reduced by -2
orders of magnitude in the presence of bzcasl3b, corroborating that Casl3b exclusively targets
RNA in vivo (Figure 11B).
48
M O -go
I'
Unnevedi I-
A
MS2 ssRNA genome j
5'-PFS protoapacer 1 3'-PFS
rep/ys S Cuo U 0 AW.W~iCU CMUV..~tW
rep/dys s - C GuAusC5  A e ec a 2 
Aacelcu A A AU CU. S55-PFS protospacer23 3 -PS
rep S' GUAGGGMAO V#XGAUICA.WVCOASAIJAdIAAAUA
pLac PJ23119
pLac+
Spacer 1
Spacer 2
Spacer 3
Non-
targeting
pBzCasl3b pLA
p~ac p23115Transcribed target
or
+ pL
Non-transcribed target
Plate on Cm + Amp
102-
10 - Transcribed target
10
0 
- Non-transcribed target
Empty vector pBzCasl3b
pLac PJ23119
rE-Emr W
PFU PFU
5.6 x 102  5.6 x 09 5.6 x 102 5.6 x 10 9
C
BzCasl3b
HEPN Mixed Ji HEPN
R'6 H 121 R177 H12
Predicted catalytic residues
D1 (R116A/H121A) D2 (R1177A/H1182A) 0 (D1 & D2)
BzCasl3b WT D1 D2 Q - WT D1 D2 Q
crRNA + + + + +
Uncleaved E
Cleaved
E
Kovalues:
+ WT: 33nM
4 D1: 38nM
-+ D2: 42nM
-- Q: 27nM
W 0
1 10
Casl3b+crRNA complex (nM)
100 1000
1 Spacer I
E Spacer 3
* Non-targeting
io4
10 2
a WT Q WT a WT OWT
Figure 11: HEPN domains mediate RNA cleavage by Casl3b, whose activity is repressed by
Csx27. (A) Protospacer design for MS2 phage plaque drop assay to test RNA interference (left).
Plaque drop assay for full B. zoohelcum VI-B1 locus (center) and bzcasl3b (right). (B) DNA
interference assay schematic (top) and results (bottom). A target sequence is placed in frame at the
start of the transcribed bla gene that confers ampicillin resistance or in a non-transcribed region on
the opposite strand of the same target plasmid. Target plasmids were co-transformed with
bzcasl3b plasmid or empty vectors conferring chloramphenicol resistance and plated on double
selection antibiotic plates. (C) Schematic (top) and denaturing gel (bottom) showing ssRNA
cleavage activity of WT and HEPN mutant BzCasl3b. The protein and targeting crRNA complexes
were incubated for 10 minutes. Gel lane containing RNA ladder not shown. (D) Electrophoretic
Mobility Shift Assay (EMSA) graph showing the affinity of BzCasl3b proteins and targeting crRNA
complex to a 5' end labeled ssRNA. The EMSA assay was performed with supplemental EDTA to
reduce any cleavage activity. (E) Quantification of MS2 phage plaque drop assay with B. zoohelcum
wildtype and Q (R116A/H121A/R1177A/H1182A) mutant Casl3b.
49
B
4.
D
i
I
1.0-
0.8-
0.6-
0.4-
0.2
0.01
0.1
AD
LL
5' PFS Protospacer 3' PFS
5 '-. NNINNNNNNNN NNNNNNNNNNNNNNNNNNNNN. .-3,
Rule 2
NAN or NNA
Composition of spacers in Kan screening library
_ Both rules (13.0%)
a
4)
aQ.
x
(D
72,
I One rule (14.6%)
No rules (22.0%)
10-2
10.-
10-6
B. zoohelcum Cas1 3b
Kanamycin screen biorep 2
00
0 C> 0
0
-6
-4 10-2
Control spacer abundance
MS2 interference assay
Protospacer 1 Protospacer 2 Protospacer 3
rep" MP
pLac J-CM-
pLac pJ23119
pLac pJ23119
+
s5,6 10' s,6 x JW
BR 1
BR2
BR3
PFU
10' 5 6 x 10
BR1
BR3
PFU
56x10 5.6x 10'
-I-IBR1
BR2
BR3
BR
BR
BR
p l1
PR)
1.2 x 10 1 x 10'
BR1
Spacer 3 BR2
BR3
BR1
Non-targeting BR2
BR3
Figure 12: Targeting rule validation of BzCasl3b and MS2 interference assay of BzCasl3b
and PbCasl3b. (A) Spacers targeting kanamycin to validate PFS targeting rules of 5' PFS (D) and 3'
PFS (NAN or NNA) (left). Second kanamycin validation screen bioreplicate of spacer abundances
versus control for individual B. zoohelcum spacers, with abundances colored by type of spacer
(right). (B) Plaque drop assay with bioreplicates for B. zoohelcum VI-B1 locus and casl3b, for P.
buccae VI-B2 locus and casl3b, and for P. buccae casl3b with pUC19, B. zoohelcum csx27, and P.
buccae csx28.
50
A
Rule 1
D (A, U, or G)
Non-targeting
(50.3%)
B
Non-targeting spacer
ePac pLac +
PEU
L6x 1w
,T-MIAX-
IMM-K IT 11
We next tested if predicted catalytic residues in the HEPN domains were responsible for
RNA cleavage by Casl3b. Three HEPN mutants were obtained by replacing the conserved catalytic
arginines and histidines in the two HEPN domains with alanines (R116A/H121A, termed domain 1
(Dl); R1177A/H1182A, termed domain 2 (D2); and R116A/H121A/R1177A/H1182A, termed
quadruple (Q)) (Figure 6). All mutants lacked observable cleavage activity (Figure 11C), yet
retained RNA binding capacity in vitro (Figures 11D and 9E). The wildtype and all three HEPN
mutant Casl3b proteins showed comparable binding affinities for a single-stranded target RNA
substrate, with KD values ranging from 27nM to 42nM (Figures 11D and 9E). The KD for off-target
binding was found to be greater than 188nM (Figure 9F).
We confirmed the involvement of the HEPN domains in RNA interference in vivo, finding
-5.5 orders of magnitude decrease in resistance to MS2 phage in the quadruple HEPN mutants
versus wildtype Casl3b (Figures 11E and 12B). Interestingly, quadruple mutant Casl3b with
spacers 2 and 3 still showed weak phage resistance, potentially due to catalytically inactive Casi3b
binding to phage genomic RNA, leading to reduced phage replication.
CRISPR-Casl3b effectors are differentially regulated by Csx27 and Csx28
To determine if the established RNA targeting rules generalize across the subtype VI-B
systems from diverse bacteria, we characterized the subtype VI-B2 locus from P. buccae. RNA
sequencing of the CRISPR array revealed processing effectively identical to that of B. zoohelcum,
excluding the long crRNA (Figure 13A). The E. coli essential gene screen with pbcasl3b and pbcsx28
or pbcas13 alone led to the identification of a PFS matrix similar to that of B. zoohelcum, with
certain PFS's disfavored (Figures 14A, 13B, and 13C). Similar to BzCas13b, PbCasl3b was found to
cleave targeted single-stranded RNA in vitro (Figure 9D). As with bzcsx27, the presence of pbcsx28
did not appreciably alter the PFS. We further explored the apparent reduced activity of pbcasl3b
alone relative to the respective full CRISPR-Cas locus using the MS2 phage plaque drop assay and
found that pbcsx28 enhances MS2 phage interference by up to four orders of magnitude (Figures
14B and 12B). The differential ability of csx27 to repress and csx28 to enhance casl3b activity
generalizes across thousands of spacers in the E. coli essential gene screen (Figure 14C),
highlighting the distinctive regulatory modes of the two variants of subtype VI-B CRISPR-Cas
systems.
51
A R buccae ATCC 33574 30nt spacer 36nt DR
2988 A____________ _ cUc
Read count 5 ' -NNNNNNNNNNNNNUNNNNNN NNNNN UC UU
3 -CAAC A AUGGAAAGU Uin____AA---- ---
66nt crRNA - -
P buccae cas13b & 2 P buccae casl3b
Spacer depletions mapped over genes Spacer depletions aggregated Spacer depletions mapped over genes Spacer depletions aggregated
across gene distance ars eedsac
10, 10 -
0 __ _1a n e jn-
0 1 2 3 4 0 0 020. 4o 6e08 1 0 1 2 3 4 5 0 020.40.6 0.8
Index ot spacer In E. coliacreen (x104) Normalized distance along gene lndeo ot spacer In E. coli screen (x104) Normalized distance along gene
C
0.s Weblogo of strongly depleted spacers (pbcasl3b & 0.5- Weblogo of strongly depleted spacers (pbcasl3b)
0.4 0.4-0.3  0.3-
E 0.2 0 .2 -
0.0 -I CfA 02-;R;
5 PFS P3rPso: 5FS Pro3PFOsaPFS
Figure 13: RNA-targeting of P. buccae VI-B2 CRISPR locus. (A) RNA-Sequencing of
heterologously expressed VI-B2 locus from P. buccae ATCC 33574 in E. coli. (B) Manhattan plots of
spacer depletions mapped over 45 genes and aggregated across normalized gene distance for full P.
buccae VI-B2 locus (left) and casl3b (right), with non-targeting spacers in gray, safely depleted
(>5Q above mean depletion of non-targeting spacers) spacers above blue line, and strongly
depleted (top 1% depleted) spacers above red line. For the full locus, 36,141 targeting spacers and
859 non-targeting spacers passed QC filter. Of the targeting, 370 are strongly depleted and 8065 are
safely depleted. For casl3b alone, 41,126 targeting spacers and 824 non-targeting spacers passed
QC filter. Of the targeting, 419 are strongly depleted and 3295 are safely depleted. (C) Sequence
weblogos of strongly depleted P. buccae spacers, revealing double-sided PFS (protospacer flanking
sequence).
To further explore the ability of the small accessory proteins to modulate Casl3b activity,
we tested if Csx27 can also repress PbCasl3b using the MS2 drop plaque assay. Cells co-
transformed with pbcasl3b and bzcsx27 expression plasmids exhibited a 105 fold reduction in
interference activity relative to pbcasl3b expression plasmid and pUC19 empty vector, indicating
that Csx27 exerts an inhibitory effect on PbCasl3b (Figures 14D and 12B). The ability of Csx27 to
modulate the interference activity of BzCas13b and PbCasl3b suggests that it is a modular protein
that can function across multiple VI-B loci.
52
ALO, U
AA AG AU AG GA GG GU GC UA UG UU UC CA CG CU CC
3' PFS
Normalized PFS score from safely depleted spacers (pbcast3b)
A
LG
3 PFS
0 0.2 0.4 0.6 0.8 1
B
pLc PJ23119
pac +
pLac pj23119
PFU
5.6 x101 5.6 x101
Spacer 1
Spacer 2
Spacer 3
Non-_________
targeting
PFU
5.6 x 102 5.6 x 01
Spacer 1
Spacer 2
Spacer 3
Non-
targeting
D
Depletion modulations of csx27 vs
coo 0
100 10'
cas13b
pLac PJ23119
pLac pLac pLa,
ror r_" or M 3
0L
106
104
102
102
4
Spacer 3
I pUC19
PbCsx28
BZCsx27
Non-targeting
Figure 14: Class 2 type VI-B systems are differentially regulated across two loci by Csx27 and
Csx28. (A) Normalized PFS matrix, for P. buccae VI-B2 locus (left) and pbcasl3b (right). (B) MS2
Plaque drop assay for full P. buccae VI-B2 locus (left) and pbcasl3b (right). (C) Spacer depletions of
bzcasl3b with and without bzcsx27 (brown), as compared to pbcasl3b with and without pbcsx28
(gold). (D) Fold resistance to MS2 infection for cells co-transformed with pbcasl3b and the
indicated csx expression plasmid.
Discussion
Here we describe two RNA-targeting CRISPR Class 2 systems of subtype VI-B (VI-B1 and VI-
B2), containing the computationally discovered RNA-guided RNase Casl3b. Type VI-B systems
show several notable similarities to the recently characterized VI-A system. The single protein
effectors of both systems cleave single-stranded RNA via HEPN domains, process their CRISPR
arrays independent of the HEPN domains, and exhibit collateral RNase activity. Casl3b proteins,
however, show only limited sequence similarity to Casl3a, and the common ancestry of the two
type VI subtypes remains uncertain. Furthermore, the type VI-B systems differ from VI-A in several
other ways, including the absence of both casi and cas2, which are involved in spacer acquisition in
other CRISPR-Cas systems (Mohanraju et al., 2016). The VI-B CRISPR arrays contain multiple
spacers that differ among closely related bacterial strains, suggesting that acquisition does occur,
either autonomously or possibly in trans, by recruiting Cast and Cas2 encoded in other CRISPR-Cas
53
C
102
S10'
to,
C100
-1
loci from the same genome. In trans utilization of adaptation modules of other CRISPR-Cas systems
is compatible with the finding that the great majority of type VI-B systems co-occur in the same
bacterial genome as other CRISPR-Cas loci that include casi and cas2 genes; conceivably, the three
VI-B-carrying genomes that lack adaptation modules have lost them recently. Additionally, VI-B
systems differ from VI-A systems by the presence of the small accessory proteins Csx27 (VI-B1
systems) and Csx28 (VI-B2 systems), which exhibit opposing regulatory effects on Cas13b activity.
Repression of Casl3b by Csx27 in VI-B1 systems could be part of an important regulatory
mechanism of phage interference. The ability of Csx27 to repress Casl3b activity may be a general
property, as we found that it can also repress PbCasl3b (subtype VI-B2). In the case of type VI-B2
systems, Csx28 might enhance the collateral activity of Cas13b to inactivate numerous transcripts
of invading bacteriophages or to promote programmed cell death. Both Csx27 and Csx28 contain
predicted long, hydrophobic a-helices that might enable them to interact physically with Cas13b,
but this remains to be determined. We did not find homologs of Csx27 or Csx28 encoded in any
CRISPR-Cas loci other than type VI-B loci, suggesting that, at least in the CRISPR-Cas context, these
proteins might function in tight association with Casl3b.
As with previously characterized Class 2 CRISPR-Cas effectors, such as Cas9 and Cpfl, there
is enormous potential to harness Casl3b for use as a molecular tool (Cong et al., 2013; Mali et al.,
2013; Wright et al., 2016). A holistic understanding of the factors that affect target selection is
essential to the success of any such tools, particularly those that target RNA, where secondary
structure will likely impact activity. We therefore developed an E. coli essential gene screen to
explore the targeting rules of Casl3b more fully. This E. coli screen offers several advantages by
increasing the number of guides testable in a single experiment to explore how diverse spacer and
flanking sequences may affect Cas13b activity. This screen revealed a double-sided PFS in VI-B
systems, which may give insight into Cas13b protein-RNA interactions, and could help improve
specificity by expanding sequence targeting constraints (Ran et al., 2015).
The characterization of Casl3b and other RNA-targeting CRISPR systems raises the
prospect of a suite of precise and robust in vivo RNA manipulation tools for studying a wide range of
biological processes (Abil and Zhao, 2015; Filipovska and Rackham, 2011; Mackay et al., 2011). The
ability of Casl3b to process its own CRISPR array could be extended to multiplex transcriptome
engineering. In addition, the VI-B functional long direct repeats could be altered to incorporate
stem loops akin to the Cas9-SAM system (Konermann et al., 2015). Like Cas9 and Cpf1, Casl3a and
Cas13b may be utilized for complementary applications in science and technology.
54
Acknowledgements
We would like to thank R. Belliveau for overall research support; R. Macrae for critical reading of
the manuscript; and the entire Zhang laboratory for support and advice. D.B.T.C. is supported by
award number T32GM007753 from the National Institute of General Medical Sciences. The content
is solely the responsibility of the authors and does not necessarily represent the official views of the
National Institute of General Medical Sciences or the NIH. I.M.S. is supported by the Simons Center
for the Social Brain. J.S.G. is supported by a D.O.E. Computational Science Graduate Fellowship.
O.A.A. is supported by a Paul and Daisy Soros Fellowship, a Friends of the McGovern Institute
Fellowship, and the Poitras Center for Affective Disorders. K.S.M. and E.V.K. and S.S. are supported
by the intramural program of the US Department of Health and Human services (to the National
Library of Medicine). S.S. is also supported by a Skoltech-MIT Next Generation Program grant to K.
Severinov and F.Z. F.Z. is supported by the NIH through NIMH (5DP1-MH100706 and 1RO1-
MH110049); NSF; the New York Stem Cell Foundation; the Allen Distinguished Investigator
Program, through The Paul G. Allen Frontiers Group; the Simons and Vallee Foundations; the
Howard Hughes Medical Institute; the Skoltech-MIT Next Generation Program; James and Patricia
Poitras and the Poitras Center for Affective Disorders; Robert Metcalfe; and David Cheng. F.Z. is a
New York Stem Cell Foundation-Robertson Investigator. A patent application has been filed related
to this work and the authors plan to make the reagents widely available to the academic community
through Addgene and to provide protocols and software tools via the Zhang lab website
(www.genome-egineering.org) and GitHub (www.github.com/fengzhanglab).
55
References:
Abil, Z., and Zhao, H. (2015). Engineering reprogrammable RNA-binding proteins for study and
manipulation of the transcriptome. Mol Biosyst 11, 2658-2665.
Abudayyeh, 0.0., Gootenberg, J.S., Konermann, S., Joung, J., Slaymaker, I.M., Cox, D.B., Shmakov, S.,
Makarova, K.S., Semenova, E., Minakhin, L., et al. (2016). C2c2 is a single-component programmable
RNA-guided RNA-targeting CRISPR effector. Science 353, aaf5573.
Anantharaman, V., Makarova, K.S., Burroughs, A.M., Koonin, E.V., and Aravind, L. (2013).
Comprehensive analysis of the HEPN superfamily: identification of novel roles in intra-genomic
conflicts, defense, pathogenesis and RNA processing. Biol Direct 8, 15.
Baba, T., Ara, T., Hasegawa, M., Takai, Y., Okumura, Y., Baba, M., Datsenko, K.A., Tomita, M., Wanner,
B.L., and Mori, H. (2006). Construction of Escherichia coli K-12 in-frame, single-gene knockout
mutants: the Keio collection. Mol Syst Biol 2, 2006 0008.
Bernhart, S.H., Hofacker, I.L., and Stadler, P.F. (2006). Local RNA base pairing probabilities in large
sequences. Bioinformatics 22, 614-615.
Biswas, A., Gagnon, J.N., Brouns, S.J., Fineran, P.C., and Brown, C.M. (2013). CRISPRTarget:
bioinformatic prediction and analysis of crRNA targets. RNA Biol 10, 817-827.
Camacho, C., Coulouris, G., Avagyan, V., Ma, N., Papadopoulos, J., Bealer, K., and Madden, T.L. (2009).
BLAST+: architecture and applications. BMC Bioinformatics 10, 421.
Cong, L., Ran, F.A., Cox, D., Lin, S., Barretto, R., Habib, N., Hsu, P.D., Wu, X., Jiang, W., Marraffini, L.A., et
al. (2013). Multiplex genome engineering using CRISPR/Cas systems. Science 339, 819-823.
Crooks, G.E., Hon, G., Chandonia, J.M., and Brenner, S.E. (2004). WebLogo: a sequence logo
generator. Genome Res 14, 1188-1190.
East-Seletsky, A., O'Connell, M.R., Knight, S.C., Burstein, D., Cate, J.H., Tjian, R., and Doudna, J.A.
(2016). Two distinct RNase activities of CRISPR-C2c2 enable guide-RNA processing and RNA
detection. Nature.
Edgar, R.C. (2007). PILER-CR: fast and accurate identification of CRISPR repeats. BMC
Bioinformatics 8, 18.
Filipovska, A., and Rackham, 0. (2011). Designer RNA-binding proteins: New tools for manipulating
the transcriptome. RNA Biol 8, 978-983.
Gerdes, S.Y., Scholle, M.D., Campbell, J.W., Balazsi, G., Ravasz, E., Daugherty, M.D., Somera, A.L.,
Kyrpides, N.C., Anderson, I., Gelfand, M.S., et al. (2003). Experimental determination and system
level analysis of essential genes in Escherichia coli MG1655. J Bacteriol 185, 5673-5684.
Hale, C.R., Zhao, P., Olson, S., Duff, M.O., Graveley, B.R., Wells, L., Terns, R.M., and Terns, M.P. (2009).
RNA-guided RNA cleavage by a CRISPR RNA-Cas protein complex. Cell 139, 945-956.
Heidrich, N., Dugar, G., Vogel, J., and Sharma, C.M. (2015). Investigating CRISPR RNA Biogenesis and
Function Using RNA-seq. Methods Mol Biol 1311, 1-2 1.
56
Henikoff, S., and Henikoff, J.G. (1992). Amino acid substitution matrices from protein blocks. Proc
Natl Acad Sci U S A 89, 10915-10919.
Hildebrand, A., Remmert, M., Biegert, A., and Soding, J. (2009). Fast and accurate automatic
structure prediction with HHpred. Proteins 77 Suppl 9, 128-132.
Jiang, W., Samai, P., and Marraffini, L.A. (2016). Degradation of Phage Transcripts by CRISPR-
Associated RNases Enables Type III CRISPR-Cas Immunity. Cell 164, 710-721.
Kim, Y.K., Kim, Y.G., and Oh, B.H. (2013). Crystal structure and nucleic acid-binding activity of the
CRISPR-associated protein Csx1 of Pyrococcus furiosus. Proteins 81, 261-270.
Konermann, S., Brigham, M.D., Trevino, A.E., Joung, J., Abudayyeh, 0.0., Barcena, C., Hsu, P.D., Habib,
N., Gootenberg, J.S., Nishimasu, H., et al. (2015). Genome-scale transcriptional activation by an
engineered CRISPR-Cas9 complex. Nature 517, 583-588.
Li, H., and Durbin, R. (2009). Fast and accurate short read alignment with Burrows-Wheeler
transform. Bioinformatics 25, 1754-1760.
Lorenz, R., Bernhart, S.H., Honer Zu Siederdissen, C., Tafer, H., Flamm, C., Stadler, P.F., and Hofacker,
I.L. (2011). ViennaRNA Package 2.0. Algorithms Mol Biol 6, 26.
Mackay, J.P., Font, J., and Segal, D.J. (2011). The prospects for designer single-stranded RNA-binding
proteins. Nat Struct Mol Biol 18, 256-261.
Makarova, K.S., Wolf, Y.I., Alkhnbashi, O.S., Costa, F., Shah, S.A., Saunders, S.J., Barrangou, R., Brouns,
S.J., Charpentier, E., Haft, D.H., et al. (2015). An updated evolutionary classification of CRISPR-Cas
systems. Nat Rev Microbiol 13, 722-736.
Mali, P., Yang, L., Esvelt, K.M., Aach, J., Guell, M., DiCarlo, J.E., Norville, J.E., and Church, G.M. (2013).
RNA-guided human genome engineering via Cas9. Science 339, 823-826.
Marraffini, L.A. (2015). CRISPR-Cas immunity in prokaryotes. Nature 526, 55-61.
Mohanraju, P., Makarova, K.S., Zetsche, B., Zhang, F., Koonin, E.V., and van der Oost, J. (2016).
Diverse evolutionary roots and mechanistic variations of the CRISPR-Cas systems. Science 353,
aad5147.
Moller, S., Croning, M.D., and Apweiler, R. (2001). Evaluation of methods for the prediction of
membrane spanning regions. Bioinformatics 17, 646-653.
Ran, F.A., Cong, L., Yan, W.X., Scott, D.A., Gootenberg, J.S., Kriz, A.J., Zetsche, B., Shalem, 0., Wu, X.,
Makarova, K.S., et al. (2015). In vivo genome editing using Staphylococcus aureus Cas9. Nature 520,
186-191.
Remmert, M., Biegert, A., Hauser, A., and Soding, J. (2012). HHblits: lightning-fast iterative protein
sequence searching by HMM-HMM alignment. Nat Methods 9, 173-175.
Sheppard, N.F., Glover, C.V., 3rd, Terns, R.M., and Terns, M.P. (2016). The CRISPR-associated Csx1
protein of Pyrococcus furiosus is an adenosine-specific endoribonuclease. RNA 22, 216-224.
57
Shmakov, S., Abudayyeh, 0.0., Makarova, K.S., Wolf, Y.I., Gootenberg, J.S., Semenova, E., Minakhin, L.,
Joung, J., Konermann, S., Severinov, K., et al. (2015). Discovery and Functional Characterization of
Diverse Class 2 CRISPR-Cas Systems. Mol Cell 60, 385-397.
Shmakov, S., Smargon, A., Scott, D., Cox, D., Pyzocha, N., Yan, W., Abudayyeh, 0.0.,
Gootenberg, J.S., Makarova, K.S., Wolf, Y.I., et aL. (2017). Diversity and evolution of class 2
CRISPR-Cas systems. Nat Rev Microbiol.
Staals, R.H., Agari, Y., Maki-Yonekura, S., Zhu, Y., Taylor, D.W., van Duijn, E., Barendregt, A., Vlot, M.,
Koehorst, J.J., Sakamoto, K., et al. (2013). Structure and activity of the RNA-targeting Type Ill-B
CRISPR-Cas complex of Thermus thermophilus. Mol Cell 52, 135-145.
Staals, R.H., Zhu, Y., Taylor, D.W., Kornfeld, J.E., Sharma, K., Barendregt, A., Koehorst, J.J., Vlot, M.,
Neupane, N., Varossieau, K., et al. (2014). RNA targeting by the type IIl-A CRISPR-Cas Csm complex
of Thermus thermophilus. Mol Cell 56, 518-530.
Tamulaitis, G., Kazlauskiene, M., Manakova, E., Venclovas, C., Nwokeoji, A.O., Dickman, M.J., Horvath,
P., and Siksnys, V. (2014). Programmable RNA shredding by the type I1-A CRISPR-Cas system of
Streptococcus thermophilus. Mol Cell 56, 506-517.
Wright, A.V., Nunez, J.K., and Doudna, J.A. (2016). Biology and Applications of CRISPR Systems:
Harnessing Nature's Toolbox for Genome Engineering. Cell 164, 29-44.
Yates, A., Akanni, W., Amode, M.R., Barrell, D., Billis, K., Carvalho-Silva, D., Cummins, C., Clapham, P.,
Fitzgerald, S., Gil, L., et al. (2016). Ensembl 2016. Nucleic Acids Res 44, D710-716.
Zetsche, B., Gootenberg, J.S., Abudayyeh, 0.0., Slaymaker, I.M., Makarova, K.S., Essletzbichler, P.,
Volz, S.E., Joung, J., van der Oost, J., Regev, A., et al. (2015). Cpfl is a single RNA-guided endonuclease
of a class 2 CRISPR-Cas system. Cell 163, 759-771.
58
Exuerimental Model And Subiect Details
P. buccae (Holdeman et al.) Shah and Collins: P. buccae was not grown in this study.
B. zoohelcum (Holmes et al.) Vandamme et al.: B. zoohelcum ATCC 43767 was grown in ATCC
medium 44 (Brain Heart Infusion broth) at 37'C at 250 rpm overnight.
E. coli (Migula) Castellani and Chalmers (C3000): E. coli was grown in LB at 37'C at 250 rpm
overnight.
One Shot Stbl3'" E. coli: E. coli was grown in LB at 37'C at 250 rpm overnight.
NEB@ 10-beta Competent E. colf (High Efficiency): NEB@ 10-beta Competent E. coli was
transformed on LB agar at 37'C overnight.
MegaX DH10B'" T1R Electrocomp" Cells: NEB@ 10-beta Competent E. coli was transformed on
LB agar at 37*C overnight.
One Shot@ BL21(DE3)pLysE Chemically Competent E. coil: The BzCas13b expression construct
was transformed into One Shot@ BL21(DE3)pLysE (Invitrogen) cells. 25 mL of 6hr growing culture
were inoculated into 2 liters of Terrific Broth 4 growth media (12 g/L tryptone, 24 g/L yeast
extract, 9.4 g/L K2HPO, 2.2 g/L KH2PO4, Sigma). Cells were then grown at 37*C to a cell density of
0.6 OD600, and then SUMO-BzCas13b expression was induced by supplementing with IPTG to a
final concentration of 500 pM. Induced culture was grown for 16-18 hours before harvesting cell
paste, which was stored at -80'C until subsequent purification. For each BzCas13b mutant, 1 L of
Terrific Broth was used to generate cell paste and all other reagents were scaled down accordingly.
Protein purification was performed using the same protocol as wild-type Casl3b. PbCasl3b was
cloned into the same pET based vector and purified using a similar protocol as BzCas13b with the
following differences: cells were grown at 21*C for 18 hours.
59
Method Details
Computational Sequence Analysis
From complete compiled Ensembl Release 27 genomes (Yates et al., 2016), CRISPR repeats were
identified using PILER-CR (Edgar, 2007). Proteins within 10kb of identified CRISPR arrays were
clustered into loci, with loci rejected if more than one protein of size 700 amino acids or larger or if
either Cast or Cas2 were present. For candidate Class 2 effectors, only proteins in these remaining
loci of size 900aa to 1800aa were selected. These candidate effectors were subjected to the BLASTP
(Camacho et al., 2009) search against the NCBI non-redundant (NR) protein sequence database
with an E-value cutoff of le-7. All discovered proteins were then grouped into putative families via
a nearest-neighbor grouping with the same E-value cutoff. Only putative families with at least ten
candidate effectors and more than 50% of candidate effectors within 10kb of CRISPR arrays were
considered. HHpred (Remmert et al., 2012) and existing CRISPR locus classification rules
(Makarova et al., 2015) were used to classify each family, leaving Casl3b as the only unclassified
family. Additional Casl3b proteins in the family were found through a nearest-neighbor search of
previously discovered Csx27/Csx28 against the NCBI non-redundant (NR) protein sequence
database with an E-value cutoff of 1e-7, and then by searching in genomes within 1kb of any newly
discovered Csx27/Csx28. Within this Casl3b family, truncated or suspected partially sequenced
effectors were discarded, leaving 105 loci, and 81 with a unique protein accession number in the
NCBI non-redundant (NR) protein sequence database. Multiple sequence alignments on these 81
proteins (as well as the accessory Csx27 and Csx28 proteins) were performed using BLOSUM62
(Henikoff and Henikoff, 1992) to identify the HEPN domains and to sort the loci into phylogenetic
trees. Loci represented in the tree of 81 non-redundant proteins were selected first for annotated
Csx27/Csx28 within 1kb of Casl3b, and next for annotated CRISPR array within 10kb of Casl3b.
Vienna RNAfold (Lorenz et al., 2011) was used to predict the secondary structure of each direct
repeat, whose transcriptional orientation was chosen as identical to that of Casl3b in its locus.
CRISPRTarget (Biswas et al., 2013) was used to search the spacers in each locus against NCBI phage
and plasmid genomes. Weblogos were generated for all unique direct repeats and protospacer
flanking sequences (Crooks et al., 2004). TMHMM Server v. 2.0 (Moller et al., 2001) was used to
predict the transmembrane helices in Csx27 and Csx28.
60
Nucleic Acid Preparation
For in vitro synthesis of RNA, a T7 DNA fragment must be generated. To create T7 DNA fragments
for crRNAs, top and bottom strand DNA oligos were synthesized by IDT. The top DNA oligo
consisted of the T7 promoter, followed by the bases GGG to promote transcription, the 30-nt target
and then direct repeat. Oligos were annealed together using annealing buffer (30 mM HEPES pH
7.4, 100 mM potassium acetate, and 2 mM magnesium acetate). Annealing was performed by
incubating the mixture for 1 minute at 95*C followed by a -1*C/minute ramp down to 23'C. To
create ssRNA targets, short targets (Trunc2, 3, 4) were synthesized as top and bottom strand oligos
containing the T7 promoter. For long ssRNA targets (El, E2, S and L CRISPR Arrays), DNA primers
with a T7 handle on the forward primer were ordered and the DNA fragment was amplified using
PCR. T7 DNA constructs for RNA generation without body labeling were incubated with T7
polymerase overnight (10-14 hours) at 30*C using the HiScribe T7 Quick High Yield RNA Synthesis
kit (New England Biolabs). Body-labeled constructs were incubated with Cyanine 5-UTP (Perkin
Elmer) and incubated with T7 polymerase overnight at 30'C using the HiScribe T7 High Yield RNA
Synthesis kit (New England Biolabs). For a complete list of crRNAs and target ssRNAs used in this
study, visit http://www.sciencedirect.com/science/article/pii/S1097276516308668. 5' end
labeling was accomplished using the 5' oligonucleotide kit (VectorLabs) and with a maleimide-
IR800 probe (LI-COR Biosciences). 3' end labeling was performed using a 3' oligonucleotide
labeling kit (Roche) and Cyanine 5-ddUTP (Perkin Elmer). RNAs were purified using RNA Clean and
Concentrator columnsTM-5 (Zymo Research). Body-labeled dsRNA substrates were prepared by T7
DNA fragments for the bottom and top RNA strand. After synthesis, 1.3-fold excess of non-labeled
bottom strand ssRNA was added and re-annealed to ensure the top strand would be annealed to a
bottom strand by incubating the mixture for 1 minute at 95*C followed by a -1*C/minute ramp
down to 23*C.
BzCasl3b Protein Purification
The mammalian codon-optimized gene for Casl3b (B. zoohelcum) was synthesized (GenScript) and
inserted into a bacterial expression vector (6x His/Twin Strep SUMO, a pET based vector received
as a gift from Ilya Finkelstein) after cleaving the plasmid with the BamHI and Notl restriction
enzymes and cloning in the gene using Gibson Assembly@ Master Mix (New England Biolabs). The
BzCas13b expression construct was transformed into One Shot@ BL21(DE3)pLysE (Invitrogen)
cells. 25 mL of 6hr growing culture were inoculated into 2 liters of Terrific Broth 4 growth media
(12 g/L tryptone, 24 g/L yeast extract, 9.4 g/L K2HPO, 2.2 g/L KH2PO4, Sigma). Cells were then
61
grown at 37*C to a cell density of 0.6 OD600, and then SUMO-BzCas13b expression was induced by
supplementing with IPTG to a final concentration of 500 ptM. Induced culture was grown for 16-18
hours before harvesting cell paste, which was stored at -80'C until subsequent purification. Frozen
cell paste was crushed and resuspended via stirring at 4*C in 500 mL of Lysis Buffer (50mM
NaH 2 PO4 pH 7.8, 400mM NaCl) supplemented with protease inhibitors (cOmplete, EDTA-free,
Roche Diagnostics Corporation) and 1250U of benzonase (Invitrogen). The resuspended cell paste
was lysed by a LM20 microfluidizer at 18,000 psi (Microfluidics). Lysate was cleared by
centrifugation at 10,000g for 1 hour. Filtered lysate was incubated with StrepTactin Sepharose
High Performance (GE Healthcare Life Sciences) at 4C for 1 hour with gentle agitation, and then
applied to an Econo-column chromatography column (Bio-Rad Laboratories). Resin was washed
with Lysis Buffer for 10 column volumes. One column volume of fresh Lysis Buffer was added to the
column and mixed with 10 units of SUMO protease (Invitrogen) and incubated overnight. The
eluate was removed from the column, SUMO cleavage was confirmed by SDS-PAGE and BlueFast
protein staining (Eton Bioscience), and the sample was concentrated via Centrifugal Filter Unit to 2
mL. Concentrated sample was loaded onto a HiTrap Heparin HP column (GE Healthcare Life
Sciences) via FPLC (AKTA Pure, GE Healthcare Life Sciences) and eluted over a gradient with an
elution buffer with salt concentration of 1.2 M. The resulting fractions were tested for presence of
BzCas13b protein by SDS-PAGE; fractions containing BzCas13b were pooled, and concentrated via
Centrifugal Filter Unit to 1 mL. Concentrated sample was loaded a gel filtration column (HiLoad
16/600 Superdex 200, GE Healthcare Life Sciences) via FPLC (AKTA Pure, GE Healthcare Life
Sciences) with buffer 500 mM NaCl, 50 mM Tris-HCl pH 7.5, 1 mM DTT.
BzCasl3b HEPN Mutant Protein Purification
Alanine mutants at each of the HEPN catalytic residues were generated using the Q5® site-directed
mutagenesis kit (New England Biolabs) and transformed into One Shot@ BL21(DE3)pLysE cells
(Invitrogen). For each mutant, 1 L of Terrific Broth was used to generate cell paste and all other
reagents were scaled down accordingly. Protein purification was performed using the same
protocol as wild-type Casl3b.
PbCasl3b Protein Purification
PbCasl3b (Prevotella buccae) was cloned into the same pET based vector and purified using a
similar protocol as BzCas13b. However, these bacterial cells were grown at 21'C for 18 hours. The
frozen cell paste was resuspended into 500 mM NaCl, 50 mM HEPES 7.5 and 2 mM DTT prior to
62
breaking cells in the microfluidizer. The Superdex 200 column was run in 500 mM NaCl, 10 mM
HEPES 7.0, and 2 mM DTT.
Nuclease Assay
Nuclease assays were performed with equimolar amounts of end-labeled or body-labeled ssRNA
target, purified protein, and crRNA, for targeted ssRNA cleavage. For CRISPR array cleavage,
protein was supplied in a four times molar excess of the CRISPR array. Reactions were incubated in
nuclease assay buffer (10 mM TrisHCl pH 7.5, 50 mM NaCl, 0.5 mM MgCI 2 , 20U SUPERase InTM
(ThermoFisher Scientific), 0.1% BSA). Reactions were allowed to proceed at 37*C for times
specified in the figure legends. After incubation, samples were then quenched with 0.8U of
Proteinase K (New England Biolabs) for 15 minutes at 25'C. The reactions were mixed with equal
parts of RNA loading dye (New England Biolabs) and denatured at 95*C for 5 minutes and then
cooled on ice for 2 minutes. Samples were analyzed by denaturing gel electrophoresis on 10%
PAGE TBE-Urea (Invitrogen) run at 45*C. Gels were imaged using an Odyssey scanner (LI-COR
Biosciences).
EMSA Assay
For the Electrophoretic Mobility Shift Assay (EMSA), binding experiments were performed with a
series of half-log complex dilutions (crRNA and BzCas13b) from .594 to 594 nM. Binding assays
were performed in nuclease assay buffer (without MgCl 2) supplemented with 10 mM EDTA to
prevent cutting, 5% glycerol, and 5ig/mL heparin in order to avoid non-specific interactions of the
complex with target RNA. Protein was supplied at two times the molar amount of crRNA. Protein
and crRNA were preincubated at 37*C for 15 minutes, after which the 5'-labeled target was added.
Reactions were then incubated at 37'C for 10 minutes and then resolved on 6% PAGE TBE gels
(Invitrogen) at 4C (using 0.5X TBE buffer). Gels were imaged using an Odyssey scanner (LI-COR
Biosciences). Gel shift of the RNA targets was quantified from an EMSA gel using Imagej (Wayne
Rasband, NIH) and plotted in GraphPad Prism version 7 (GraphPad Software, La Jolla California
USA). Line regression was performed in Prism 7 using nonlinear fit with one-site binding
hyperbola. KD values are calculated by GraphPad Prism based on regression analysis of data.
63
RFP-Tagged Protein Fluorescent Imaging
One Shot Stbl3 Chemically Competent E. coli were transformed with plasmids containing RFP
(negative control) or RFP fused to the N- or C-terminus of Csx27 of B. zoohelcum or Csx28 of P.
buccae. Clones were cultured up in 5mL of antibiotic LB overnight, then spun down at 5000g and
resuspended in PBS with 1% methanol-free formaldehyde. After 30 minutes fixation, cells were
washed once with PBS and then diluted 1:2 in PBS. 5uL of sample was pipetted onto a silane-coated
slide, which was covered with a coverslip. Fluorescent imaging was performed in a 63x objective
microscope with oil immersion.
Bacterial RNA-Sequencing
RNA was isolated and prepared for sequencing using a modification of a previously described
protocol (Heidrich et al., 2015; Shmakov et al., 2015). RNA was isolated from 5 mL of stationary
phase of bacterial cultures by resuspending pelleted cells in 1mL of TRIzol (ThermoFisher
Scientific) and then homogenizing with 300 uL zirconia/silica beads (BioSpec Products) in a
BeadBeater (BioSpec Products) for 7 1-minute cycles. 200 uL of chloroform was added to the
homogenized sample and then samples were centrifuged for 15 min. (12000xg, 4C). The aqueous
phase was then used for input into the Direct-Zol RNA miniprep kit (Zymo). Purified RNA was
DNase treated with TURBO DNase (Life Technologies) and 3' dephosphorylated/5' phosphorylated
with T4 Polynucleotide Kinase (New England Biolabs). rRNA was eliminated using the bacterial
Ribo-Zero rRNA removal kit (Illumina). Next, RNA was treated with RNA 5' polyphosphatase
(Epicentre Bio) to convert 5'-triphosphates to 5'-monophosphates for adapter ligation. Samples
were then polyA tailed with E. coli Poly(A) polymerase (New England Biolabs), and a 5' RNA
Illumina sequencing adapter ligated to cellular RNA using T4 RNA Ligase 1 (ssRNA ligase) (New
England Biolabs). RNA was reverse transcribed using AffinityScript cDNA synthesis kit (Agilent
Technologies) and an oligo-dT primer. cDNA was amplified with Herculase II polymerase (Agilent
Technologies) and barcoded primers. The prepared cDNA libraries were sequenced on a MiSeq
(Illumina)
For RNA sequencing of native B. zoohelcum ATCC 43767, we repeated the experiment with a
modified protocol, omitting RNA 5' polyphosphatase prior to 5' adapter ligation, to promote
enrichment of processed transcripts originating from the CRISPR array. For heterologous P. buccae
ATCC 33574 RNA sequencing in E. coli, we cloned the locus into pACYC184. Reads from each sample
were identified on the basis of their associated barcode and aligned to the appropriate RefSeq
reference genome using BWA (Li and Durbin, 2009). Paired-end alignments were used to extract
64
entire transcript sequences using Galaxy (https://usegalaxy.org), and these sequences were
analyzed using Geneious 8.1.8.
E. coli Essential Gene Screen Experiment
The intersection of two E. coli DH10B strain essential gene studies (Baba et al., 2006; Gerdes et al.,
2003) was taken, and further pared down to 45 genes by only selecting genes exclusive to their
respective operons. Over these 45 genes 54,600 spacers were designed to tile at single resolution
across the coding region, as well as to extend 60 nt into the 5' UTR and 3'UTR. In addition, 1100
non-targeting, pseudorandomly generated spacers with no precise match to the E. coli DH10B
strain genome were added to the library as a non-targeting negative control. The library of spacers
was cloned into a B. zoohelcum or P. buccae direct repeat-spacer-direct repeat backbone containing
a chloramphenicol resistance gene using Golden Gate Assembly (NEB) with 100 cycles, and then
transformed over five 22.7cm x 22.7cm chloramphenicol LB Agar plates. Libraries of transformants
were scraped from plates and DNA was extracted using the Macherey-Nagel Nucleobond Xtra
Midiprep Kit (Macherey-Nagel). 5Ong of library plasmid and equimolar gene plasmid containing an
ampicillin resistance gene (bzcasl3b, bzcasl3b & bzcsx27, pbcasl3b, pbcasl3b & pbcsx28, empty
vector pBR322) were transformed into MegaX DH10BTM TiR ElectrocompT M Cells (ThermoFisher)
according to manufacturer's protocol, with four separate 22.7cm x 22.7cm carbenicillin-
chloramphenicol LB Agar plates per bioreplicate, and three bioreplicates per condition (twelve
transformations total per condition). Eleven hours post-transformation, libraries of transformants
were scraped from plates and DNA extracted using the Macherey-Nagel Nucleobond Xtra Maxiprep
Kit (Macherey-Nagel).
E. coli Essential Gene Screen Analysis
Prepared DNA libraries were sequenced on a NextSeq (Illumina), with reads mapped to the input
library of spacers. Spacer depletions were calculated as the read abundance of a spacer in the
empty vector condition divided by read abundance in each gene plasmid condition. Mean
depletions over three bioreplicates were calculated. We imposed a two-step quality-control filter on
the data: a maximum coefficient of variation of 0.2 for depletion over three bioreplicates, and a
minimum spacer read abundance of 1/3N in each bioreplicate, where N = 55,700. Weblogos of the
strongly depleted (top 1% depleted) spacers were generated (Crooks et al., 2004), and from each
identified PFS, heatmaps of the ratio of moderately depleted (referred to as safely depleted) (>5
above mean depletion of non-targeting spacers) spacers to all spacers in the screen were generated.
65
For spatial analysis via empirical cumulative distribution functions, safely depleted spacers were
aggregated across the first or last 250 nt of genes.
For secondary structure analysis, we utilized the RNA accessibility model from Vienna RNAplfold
(Bernhart et al., 2006).
Kanamycin Validation Screen Experiment
A total of 160 kanamycin-targeting spacers was selected, 42 of which contain both PFS rules, 47 of
which contain one rule, and 71 of which contain no rules, to which 162 non-targeting control
spacers were added. The library of spacers was cloned into either a bzcasl3b and B. zoohelcum
direct repeat-spacer-direct repeat backbone or simply a B. zoohelcum direct repeat-spacer-direct
repeat backbone containing a chloramphenicol resistance gene using Golden Gate Assembly (NEB)
with 100 cycles, and then transformed over one 22.7cm x 22.7cm carbenicillin LB Agar plate. The
two cloned library plasmids were then re-transformed with over a 22.7cm x 22.7cm
chloramphenicol LB Agar plate or a 22.7cm x 22.7cm kanamycin-chloramphenicol LB Agar plate.
Libraries of transformants were scraped from plates and DNA extracted using the Qiagen Plasmid
Plus Maxi Kit (Qiagen). 100 ng of library DNA and 100 ng of pMAX-GFP (Lonza), containing a
kanamycin resistance gene were added to 50 uL of chemically competent 10-beta cells (NEB) and
transformed according to the manufacturer's protocol.
Kanamycin Validation Screen Analysis
Prepared DNA libraries were sequenced on a NextSeq (Illumina), with reads mapped to the input
library of spacers. For normalizing the abundance of spacers of two separate clonings, the corrected
experimental read abundance of a given spacer was calculated as the read abundance of that spacer
in the bzcasl3b plasmid (kanamycin-chloramphenicol transformation) multiplied by the ratio of the
read abundance ratio of that spacer in the non-bzcasl3b plasmid (chloramphenicol-only
transformation) to the read abundance ratio of that spacer in the bzcasl3b plasmid
(chloramphenicol-only transformation).
MS2 Phage Drop Plaque Assay
Individual spacers for bacteriophage MS2 interference were ordered as complementary
oligonucleotides containing overhangs allowing for directional cloning in between two direct
repeat sequences in vectors containing casl3b. 10 uM of each complementary oligo were annealed
66
in 1oX PNK Buffer (NEB), supplemented with 10mM ATP and 5 units of T4PNK (NEB). Oligos were
incubated at 37*C for 30 min., followed by heating to 95'C for 5 min. and then annealed by cooling
to 4C. Annealed oligos were then diluted 1:100 and incubated with 25 ng of Eco3lI digested
casl3b vector in the presence of Rapid Ligation Buffer and T7 DNA ligase (Enzymatics). Individual
plasmids were prepared using the QlAprep Spin Miniprep Kit (Qiagen), sequence confirmed and
then transformed into C3000 (ATCC 15597) cells made competent using the Mix & Go E. coli
Transformation Kit (Zymo). In the case of experiments using csx27 or csx28, C3000 cells harboring
csx plasmids were made competent and then transformed with casl3b direct repeat-spacer-direct
repeat plasmids. Following transformation, individual clones were picked and grown overnight at
37'C in LB containing the appropriate antibiotics. The following morning, cultures were diluted
1:100 and grown to an OD600 of 2.0 by shaking at 37'C with 5% C02 at 250 rpm, then mixed with
4mL of antibiotic containing Top Agar (10 g/L tryptone, 5 g/L yeast extract, 10 g/L sodium
chloride, 5 g/L agar) and poured on to LB-antibiotic base plates. 10 fold serial-dilutions of MS2
phage (ATCC 15597-B1) were made in LB and then spotted onto hardened top agar with a multi-
channel pipette. Plaque formation was assessed after overnight incubation of the spotted plates at
370 C. For assessing interference levels, samples were blinded using a key and the lowest dilution
of phage at which plaque formation occurred was compared to a pACYC condition by eye, where the
lowest dilution of MS2 that formed plaques on pACYC was set to 1. The lowest dilution of phage
used was 1.05*108 pfu.
DNA Interference Assay
A 34-nt target sequence consisting of a 30-nt protospacer and a permissive PFS (5'-G, 3'-AAA) was
cloned into pUC19 in two locations. For the transcribed target, the target sequence was cloned into
the coding strand of the bla gene, in frame immediately after the start codon, with the G of the start
codon serving as the 5' PFS. For the non-transcribed target the identical target sequence
(protospacer and PFS) were cloned into the AatII site of pUC19, so that the protospacer appears on
the non-transcribed strand with respect to the pBla and pLac promoters. To determine
interference, 25 ng of the ampicillin resistant target plasmid and 25 ng of the chloramphenicol
resistant bzcasl3b or empty vector (pACYC) were added to 5 uL of NovaBlue GigaSingle cells
(Novagen). The cells were incubated for 30 minutes on ice, heatshocked for 30 seconds at 42'C and
incubated on ice for 2 minutes. Then, 95 uL of SOC was added to cells and they were incubated with
shaking at 37'C for 90 minutes, before plating the entire outgrowth (100 uL) on plates containing
both chloramphenicol and ampicillin.
67
Quantification And Statistical Analysis
MS2 Interference Assay-HEPN Mutants
Three bioreplicates of the MS2 interference assay were performed for the fold resistance
quantification. For assessing interference levels, samples were blinded using a key and the highest
dilution of phage at which plaque formation occurred was compared to a vector only condition by
eye, where the highest dilution of MS2 that formed plaques on pACYC was set to 1. The error bars
are the standard deviation of the fold-resistance for each condition.
DNA Interference Assay
Three bioreplicates of the DNA interference assay were performed for the colony forming unit
quantification. The mean values were taken from the mean of number of colony forming units from
a standard colony forming unit count, and the standard deviation values accordingly from the same
standard count.
E. coli Essential Gene Screen
Spacer depletions from the screen were calculated as the read abundance of a spacer in the empty
vector condition divided by read abundance in each gene plasmid condition. Mean depletions over
three bioreplicates were calculated. We imposed a two-step quality-control filter on the data: a
maximum coefficient of variation of 0.2 for depletion over three bioreplicates, and a minimum
spacer read abundance of 1/3N in each bioreplicate, where N = 55,700. This reduced the number of
guides represented from N to approximately 30,000-40,000.
KD Calculations
Gel shift of the RNA targets was quantified from an EMSA gel using ImageJ (Wayne Rasband, NIH)
and plotted in GraphPad Prism version 7 (GraphPad Software, La Jolla California USA). Line
regression was performed in Prism 7 using nonlinear fit with one-site binding hyperbola. KD values
are calculated by GraphPad Prism based on analysis of regression data.
Data And Software Availability
Data Resources
Data have been deposited in the following resources:
68
Next-Generation Sequencing for bacterial RNA-sequencing, E. coli essential gene screen, kanamycin
validation screen: https://www.ncbi.nlm.nih.gov/bioproject/PRINA358111
Supplemental Tables are available online at
http://www.sciencedirect.com/science/article/pii/S1097276516308668
Table S1j All crRNAs, nucleic acid targets, and primers used in biochemical experiments.
Table S2 | All Casl3b plasmids used in this study.
Table S3
spacers.
I E. coli essential genes represented in E. colf essential gene screen library of
Table S4 j Spacers from E. colf essential gene screen.
Table S S Spacers from kanamycin validation screen.
Table S6 | Spacers targeting MS2 and pBLA plasmids.
Table S7 | EMSA raw data.
69
70
CHAPTER 3
FUTURE DIRECTIONS
BASIC BIOLOGY TO INVESTIGATE AND POTENTIAL ENGINEERING
APPLICATIONS OF CAS13B
71
The identification and characterization of Casl3b increases the known diversity of
CRISPR-Cas adaptive immune systems and provides an additional single effector Cas enzyme
that may be harnessed for tool development. Casl3b represents the second unique single
effector RNA-guided RNAse that has both similar and unique properties compared to the
other single effector Cas RNase C2c2. The discovery of Casl3b following characterization of
C2c2 is analogous to the discovery of Cpfl after Cas9 characterization. Both Cpfl and Cas9
have been harnessed as effective genome editing tools that can be used for similar or
complementary applications and Casl3b and C2c2 may later exhibit a similar relationship.
In order to harness Casl3b for transcriptome engineering purposes, it is important to better
understand the basic biological mechanisms of Casl3b activity and then use this knowledge
to develop useful RNA tools.
Basic Biology to Investigate Related to Casl3b
There are many fundamental properties of Casl3b that are currently unknown and there
are many intriguing questions that could be investigated to better understand Casl3b function.
What are the mechanisms of Casl3b binding, Casl3b targeted cleavage activity, and Casl3b
collateral effect? How specific is Casl3b? How do the small proteins Csx27 and Csx28 modulate
Casl3b interference? Lastly, what is the mechanism of acquisition that occurs in Casl3b loci?
Thoughts on these questions will be elaborated on below.
The mechanism by which Casl3b first binds to its target RNA, cuts its target, and then
undergoes collateral nonspecific RNAse activity is currently unknown. It is possible that the
Casl3b binding mechanism operates similarly to other Cas enzymes. For example, Cas9 has a
known process to first promiscuously bind DNA at the PAM sequence, unwind the DNA from the
seed region next to the PAM to the end of guide RNA sequence, and with sufficient pairing, cleavage
occurs (Sternberg et al., 2014; Wu et al., 2014). Based on biochemical studies shown in Chapter 2,
Casl3b must be able to specifically recognize RNA versus DNA, search the sequence, and initiate
binding to its RNA target. This is specifically supported by the EMSA data: WT Casl3b exhibits non-
specific RNA binding affinity (188-10OOnM KD) and higher affinity binding to RNA with its target
sequence (33nM KD). Analogous to the Cas9 PAM binding, it is possible that the 5' and/or 3' PFS
sequences have roles in the initial Casl3b binding process. Casl3b may recognize 5'D or
3'NAN/NNA first and then direct RNA base pairing between the crRNA and the target RNA proximal
to the 5' or 3' end. However, data from C2c2 studies suggests that the seed region is in the central
part of the guide (Abudayyeh et al., 2016). Based on the number of functional similarities between
72
C2c2 and Casl3b, it is possible that Casl3b may also have a central seed region. If binding
nucleation is independent from the PFS sequences, it is interesting to speculate on the role of the
PFS sequences. It is hypothesized that the non-permissive PFS base across from the DR sequence in
the crRNA is determined by the protospacer flanking base that is able to form a complementary
base-pair with the first base of the DR (see Figure 1). For C2c2, its 3' PFS is H (or not G) and the
protospacer adjacent DR base in the crRNA is C (the crRNA orientation is 5'-DR-spacer-
3')(Abudayyeh et al., 2016) whereas the Casl3b 5'PFS is D (or not C) and the protospacer adjacent
DR base in the crRNA is G. This hypothesis is supported by the C2c2 crystal structure, which
illustrates a kink in this region that would be disrupted if base pairing between the DR and PFS
would occur (Liu et al., 2017). Cas13b has also been identified to have a 3'PFS preference of
NAN/NNA. One hypothesis is that a flexible A in either position 2 or 3 from the protospacer target
participates in a stabilizing protein:RNA interaction with Casl3b to promote cleavage activity.
Casl3b C2c2
5' PFS= D (not C) 3' PFS PFS= H (not G)
RNA Target 5' - AAGGGA - AAGA. .- 3' 5'- . .AAA :.UCGACCUGCAGG..-3'
A UACUCUCGUCAAG-GUUGUGUAC-CCUAAUGGGIiACAGCUUCUGUUGU -5' 3'- UACCUAAUGAACCAUCUUGUCGUUAGAUCAAAAUCAGGGGAAGC
UGGAGGGUAA-UCACAAC 
-3' 5'-CCA--CCCCAAUA
Figure 1: PFS sequence requirements opposite the first DR base. This figure illustrates the
hypothesis, which suggests that the PFS opposite the first DR base is based on the DR base. The
figure shows a RNA target and the protospacer (blue) targeted either by Cas13b or C2c2. The PFS
opposite the DR is shown in orange.
After Cas13b binds to its target RNA, it would be ideal to understand how targeted RNA
cleavage is accomplished and the specificity of Cas13b targeted RNA cleavage. C2c2, like all other
single effector Class 2 Cas proteins, is a bilobed structure in which its HEPN domains associate
together when bound to its target (Liu et al., 2017). It is probable that Cas13b also undergoes a
conformational shift to bring together its HEPN domains, perhaps like joining one's index finger to
thumb. In addition to conformational changes, Cas13b cleavage specificity is key to understand.
Cas9 is known to tolerate some mismatches and indels (Kim et al., 2015; Tsai et al., 2015) and WT
Cas13b may also tolerate mutations. To test specificity, single synthesized targets with specific
mismatches could be tested for biochemical cleavage at a low throughput levels. Although pooled
higher throughput screening assays would be more efficient, the presence of the collateral effect
would likely confound results.
The mechanism by which Cas13b undergoes the collateral effect is worthwhile to
investigate and structural analyses could provide insight. The C2c2 crystal structure illustrates that
its catalytic active site is exposed to the surface, hypothesized to interact with RNAs with a variety
73
of secondary structures (Liu et al., 2017). The openness of the catalytic residues may explain
nonspecific RNAse activity because after targeted cleavage, the targeted RNA may dissociate and
the active catalytic site may remain exposed to other RNAs. Casl3b may have a similar exposed or
otherwise accessible active site. Crystal structures of Casl3b with and without a crRNA and target
RNA could show the structure of the active site before and after targeted RNA cleavage.
Alternatively, protein structure assays such as FRET could be employed to analyze domain
reorganization. Although the phenomenon of the collateral effect itself is interesting, it is just as
interesting to speculate why a targeted RNAse would also have subsequent activity as a nonspecific
RNAse. The collateral effect shows parallels to other bacterial immunity systems such as abortive
infection, which eliminates all RNAs in the system to reduce the chance of bacteriophage survival.
Additionally, the collateral effect parallels the activity of the non-targeted RNase enzymes Csm6 or
Csx1 involved in Type III CRISPR-Cas systems to mediate interference. Perhaps a generalized RNase
activity during viral infection provides a critical survival advantage to the population of bacterial
hosts.
Small proteins Csx27 and Csx28
One key distinction between Casl3b and C2c2 is that the small proteins Csx27 or Csx28 can
modulate Casl3b interference activity, but the mechanism by which these proteins modulate
interference is unknown. One key piece of information that is unknown is whether or not Csx27 or
Csx28 directly interact with either Casl3b or the crRNA. To test for direct interaction between
Casl3b and either Csx27 or Csx28, a protein complex immunoprecipitation assay could be
completed. Since the small Csx27 or Csx28 proteins are predicted to be membrane associated, and
if stable direct interactions occur between Csx27 or Csx28 and Casl3b, these small membrane
proteins may recruit Casl3b to the membrane, which is the first area of entry of bacteriophage
nucleic acid. Alternatively, it is possible that Csx27 and Csx28 are independent interference
mediators or that Csx27 and Csx28 interact with the crRNA only to mediate interference. If Csx27
and Csx28 are independent interference mediators, then expressing these proteins alone should
mediate some level of interference without expressing Casl3b or the crRNA. It is also possible that
Csx27 and Csx28 interact specifically with the crRNA to mediate interference. Binding assays
between these proteins and the crRNA would be critical to illustrate this effect. Lastly, Csx28 may
participate in RNA cleavage if it contains a bonafide active HEPN motif, perhaps dimerizing with a
second Csx28 protein.
74
Casl3b Adaptation
Casl3b was not identified in previous computational searches for novel Class 2 CRISPR-Cas
systems because it is lacking Cast, the only known protein required for adaptation. Without Cast
and Cas2 in Cast3b loci, there are two main possibilities to explain acquisition: acquisition either
occurs using Cast proteins from other CRISPR-Cas loci in the genome or it has a unique form of
adaption without Cast, perhaps involving Csx27 or Csx28. A recent report by Laanto et al. provides
evidence supporting an acquisition mechanism that utilizes Cast proteins in trans expressed by
other non Type VI CRISPR-Cas loci in the genome. In brief, Laanto et al. conducted a long-term field
sampling study with the bacterial species Flavobacterium columnare, which has a Type II locus and
Type VI-B locus. The study illustrated that both CRISPR loci acquired spacers over time matching
sequenced phages in the environment. Notably, all fifteen spacers in the VI-B CRISPR locus targeted
ORFs on the phage coding strand, and the authors provided some, but in my opinion weaker,
evidence of sequence similarity in the predicted PAM sequence for F. columnare Cas9. Overall, this
paper provides evidence that Type VI loci can undergo acquisition specifically for RNA coding viral
sequences and suggests that trans expressed acquisition machinery assist in this process. This data
is also supported by our findings that most type VI loci have other diverse CRISPR-Cas loci in their
genome. Although we found three exceptions of species of bacteria with exclusively a Cast3b Type
VI system, the lack of a second CRISPR-Cas locus could be explained by an incompletely sequenced
genome, misannotation, or loss of acquisition machinery or CRISPR-Cas system at some point in the
species history. Lastly, since the F. columnare VI-B locus does not have either csx27 or csx28 genes
in its locus, this study provides evidence that Csx27 and Csx28 are not involved in the adaptation
process.
Engineering applications of Casl3b
Like other Class 2 single effector Cas enzymes, Cast3b has the potential to direct cellular
perturbations or development of other biological assays. Since Cast3b acts on RNA, it may be
developed for transcriptome modulation of cells through RNA knockdown, RNA editing, or
epitranscriptomics and can be used to develop assays including RNA imaging or RNA diagnostics.
The properties of Cast3b itself could also be modulated through structure-guided mutagenesis to
engineer new or altered functions. This section will further describe the engineering applications of
Cast3b.
75
Targeting cellular RNA for transcriptome modulation represents an unrealized potential in
genome engineering, but utilizing CRISPR-Cas enzymes is not the only proposed method for
targeted RNA modulation. In total, three systems have the potential for targeted transcriptome
engineering and they include the use of Pumilio proteins, argonaute proteins, or CRISPR-Cas
systems. Pumilio/FBF (also known as PUF or Pum), function similarly to TAL effectors or zinc
fingers and rely on protein engineering of modular domains that bind to specific RNA bases
(Crittenden et al., 2002; Zamore et al., 1997). As mentioned before, protein engineering is more
difficult than using a programmable RNA, but Pum systems have an additional key disadvantage.
Reprogramming Pum domains for targeted RNA editing exhibits low target specificity due to
flexible nature of RNA (Wang et al., 2013). Therefore, these systems have not become a widespread
or simple method to modulate RNA. Besides Pum or CRISPR-Cas systems there exists another
system, the Natronobacterium gregoryi Argonaute (NgAgo), which has a potential for targeted RNA
modulation. NgAgo undergoes DNA-dependent RNA cleavage using a DNA guided endonuclease
protein and has no PAM requirement (Sunghyeok et al.). This is a relatively new finding and has
potential, but additional studies are required to better understand the mechanism and specificity.
Additionally, NgAgo has some confounding conflicting studies (Lee et al., 2016) suggesting that
NgAgo cleaves dsDNA (Gao et al., 2016) and further studies are needed to better understand its true
target.
While Casl3b represents a new area of technology with promise for RNA modulation, it is
worthwhile to note existing RNA modulating technologies and where Cas13b can make an impact.
RNA knockdown can already be achieved in mammalian cells using RNAi pathways. However
siRNA, shRNA, or morpholino antisense oligo technologies exhibit a wide variety of specificity with
some that are very specific and some that are less specific (Cullen, 2006). While some shRNAs have
been quite useful, one unmet need is directing RNA knockdown to the nucleus. SiRNAs only
interact with cytoplasmic RNAs and by targeting RNAs in the nucleus, one could directly probe
nuclear lincRNAs, nascent transcripts, or probe allele specific RNA expression. A table comparing
RNAi with Cas13b is shown in Table 1.
76
RNAi Casl3b
Specificity Range of non-specific or Unknown
specific depending on target
Delivery Viruses, liposomes, cationic Likely plasmid based using
polymers, siRNA cojugates, transfection reagents or viral
direct injection(Kanasty et al., vectors, possibly purified
2013) protein and RNA
RNA cleavage mechanism Association of siRNA with RNA- crRNA targeted cleavage by
induced silencing complex and Casl3b enzyme and later
argonaute 2 collateral cleavage
Potential or current Knockdown only Knockdown, splicing
Applications regulation, translational
regulation, localization analysis
Cellular location of RNA Cytoplasm only Cytoplasm or nucleus
modulation
Table 1: Comparison of RNAi and Casl3b with respect to known or predicted specificity,
delivery, RNA cleavage mechanism, applications or cellular location of RNA modulation.
There are many potential applications for targeting mammalian RNA transcripts using
Casl3b. A summary of mammalian RNA targeting applications is illustrated in Figure 2. Casl3b can
meet an unmet need in directing RNA binding proteins to modulate translation, dictate alternative
splicing, assess or alter transcript localization, target transcripts for degradation, serve as a binding
protein to prevent binding with other RNA transcripts, or add epitranscriptomic marks. Each one of
these tools could be promising to develop and used to elucidate the basic biology of different
cellular functions or diseases. To create an RNA binding protein with RNA editing modules,
structure guided engineering of Casl3b can be completed similar to Cas9 engineering used for
transcriptional activation. Specifically, Casl3b can be directly fused to RNA editing proteins or
alternatively, MS2 or PP7 hairpins can be added onto permissive areas of the crRNA and MS2 or
PP7 coat protein fusions with RNA editing proteins can be utilized. Although various regions of
Casl3b and the crRNA can be empirically tested for optimal placement of a fusion protein or stem
loop addition, it is generally easiest to use structure-guided placement, which would require the
Casl3b crystal structure.
77
a5'UTR AG
b
C 1
d U
U'
l
g
-14UG.,.m' UTR(~Split
hV,/RNase
AU
-AUG
Enhance
translation
Repress
translation
Skip an
exon
Include
an exon
Assess
localization
Alter
localization
Degrade
RNA
Degrade
Block cANA
a '~~AUBlock ncRNA
function
Figure 2: Applications for targeted RNA modulation in mammalian cells. The yellow ovals
represent RNA binding proteins, in this case Casl3b. Modular RNA effector domains are illustrated
by various linked shapes. Figure from (Mackay et al., 2011).
Selecting functional and specific modular domains to mediate targeted translation,
alternative splicing, RNA editing, or RNA epitranscriptomics is critical for development of an RNA
editing tool. To direct translation of a specific mRNA, Casl3b with the eukaryotic translation
initiation factor elF4G can be targeted near the start codon within the 5' UTR. For targeted
translation, it is ideal to target an early and rate-limiting step, which for translation is formation of
the initiation complex. The initiation of translation requires many different components. The
protein eIF4G serves as a scaffold protein, which binds to many other proteins including elF3 and
plays a role in recruiting the 40S ribosomal subunit (Bhat et al., 2015). Critically, the elF4G subunit
78
was shown to be sufficient in driving translation of downstream mRNAs in HeLa cells using a direct
fusion of the C-terminal end of eIF4G to the protein IRP-1 (De Gregorio et al., 1999). Casl3b fusions
with eIF4G could be a tool for targeted upregulation of translation.
Casl3b could also modulate alternative splicing by blocking splice sites or targeting splicing
events. For RNA splicing to occur, proteins of the spliceosome interact with critical sequences at the
5' splice site, A branch point, polypyrimidine tract within the intron, and 3' splice site (Matlin et al.,
2005). To target alternative splicing, Casl3b can be fused to an arginine and serine-rich (RS)
protein domain to facilitate binding of the Ul small nuclear ribonucleoprotein to the 5' splice site. A
proof of concept example of this has been illustrated in one study that fused the arginine and
serine-rich domain of the protein SF2 to an RNA targeting PUM protein to enhance RNA-splicing at
a particular site (Wang et al., 2009). Alternatively, Casl3b can be targeted to a splice site and
physically block it from interacting with the spliceosome.
Casl3b could also be used for targeted RNA editing or alteration of the primary RNA
sequence when used in conjunction with RNA base editing enzymes such as ADAR or cytidine
deaminase (Gerber and Keller, 2001). Unlike DNA's repair mechanisms NHEJ or HR that fix cut
DNA, cut RNA does not undergo repair mechanisms and thus RNA base editing enzymes could be
used for targeted single base substitutions. ADAR is an enzyme that mediates A to I transitions on
dsRNA, via adenosine deamination where I can be recognized as G base. Alternatively cytidine
deaminase could be used, which mediates C to U transitions (Nishikura, 2010). If successful, RNA
editing enzymes could be used to better understand RNA editing in a natural context. For example,
C to U transitions in the 3'UTR of an mRNA may alter miRNA targeting specificity (Gu et al., 2012)
and with a targeted editing tool, miRNA binding on endogenous transcripts could be studied.
Alternatively, RNA editors could in theory have therapeutic applications, especially for targeting
single base pair mutation diseases such as hemophilia or tyrosinemias.
Lastly, targeted RNA binding proteins could be useful to direct and better understand RNA
epitranscriptomics. RNA is known to acquire numerous post-transcriptional small chemical
modifications (Frye et al., 2016), for example, the addition of a methyl group on adenosine residues
to create N6-methyladenosine (m 6A), a well-studied mRNA modification. The presence of m 6A in
the 5' UTR of an mRNA has been correlated with cap-independent translation (Meyer et al., 2015).
and this modification is also enriched proximal to stop codons and at large internal exons
(Dominissini et al., 2012; Meyer et al., 2015). Proteins known to deposit (large complex with
METTL3 protein) (Bokar et al., 1997) or remove (FTO and ALKBH5) (Jia et al., 2011; Zheng et al.,
2013) m 6A modifications have also been identified. A tool to target RNA modifications could help
79
link a phenotype or mechanism directly to an RNA modification. Although creating tools for
targeted epigenomic modifications has potential, additional basic biology studies are needed to
better understand the various function or mechanism of less understood RNA modifications and
additional mapping technologies are needed to identify the location of these modifications.
Two major points of concern for any RNA targeting tool that modulates cellular transcripts
include specificity and temporary versus long-lasting effects. Specificity for any genome or
transcriptome engineering tool is critical. It is important that the tool functions on the targeted
transcript with little to no effects on non-targeted transcripts. At this point, the initial targeting
specificity of Casl3b is unknown and will require further studies. With Casl3b tools requiring a
modular domain, it is also important to ensure that these domains are specific as well. Another
concern about RNA targeting tools is the inherent temporary nature of RNA. For example, mRNAs
are transcribed and readily degraded within the cell and thus any edit to RNA would only have a
temporary effect. To make any lasting or permanent effect, RNA would need to be continually
edited. However, editing RNA would not produce any permanent off-target mutations that are
possible with DNA targeting enzymes.
Besides modulating RNA within a mammalian cell, Casl3b could be developed for use in a
RNA diagnostic assay. Casl3b used as a specific RNA detection tool relies on the collateral effect: an
activating target RNA, mixed with quenched-fluorescent RNAs, undergoes targeted RNA cleavage
and then the activated enzyme cleaves the fluorescent quenched RNAs in the sample, resulting in
fluorescence. This proof of concept experiment was described for C2c2 and illustrated in Figure 3
(East-Seletsky et al., 2016). A sensitive RNA diagnostic assay could be used to detect clinically
relevant RNA viruses in patients such as HIV or Zika virus. However, detection would require an
increase in accuracy, sensitivity, and speed (Gootenberg et al., 2017), potentially by developing
rapid paper-based diagnostics (Pardee et al., 2014).
80
trans cleavage of fluorescent RNA oligo
crRNA
Quenched-fluorescent
RNA
Activating
target RNA
Accumulation of fluorescent signal
upon RNA target recognition
Figure 3: RNA diagnostics using single effector CRISPR-Cas RNA targeting enzymes with the
collateral effect. Image from (East-Seletsky et al., 2016).
Cas13b also shows promise for RNA based imaging applications. RNA localization is of
interest to study because cellular localization of mRNA can illustrate spatial specific mRNA
expression or regulation, especially in cells with well-defined subcellular compartments such as
budding yeast, Drosophila embryos, or neurons (Martin and Ephrussi, 2009). Catalytically inactive
Casl3b with a tag such as FLAG, HA, or even GFP can bind to its target and be visualized using
simple traditional immunostaining methods or potentially live cell imaging. If efficacious, RNA
localization could enable simple and rapid mRNA localization studies compared to other techniques
such as fluorescence in situ hybridization (FISH) which are time consuming and require
conjugation of a visible marker to each unique RNA probe (Long et al., 1995). Additionally, Casl3b
localization can work on endogenously expressed mRNAs unlike other studies that transfected
artificial mRNAs containing a MS2 hairpin loop and assayed for its localization using a MS2 coat
protein GFP fusion (Bertrand et al., 1998). Lastly for a Cas13b localization tool, it is critical to
ensure that the targeted protein binding does not alter RNA expression levels, localization, or
stability, all of which would need to be validated through orthogonal assays such as qPCR or FISH.
Conclusion
CRISPR-Cas adaptive immune systems in bacteria and archaea naturally exhibit wide
diversity in protein composition to target foreign nucleic acids for degradation. Class 2 CRISPR-Cas
systems employ a large single effector nuclease to mediate interference and can be harnessed for
genome engineering. The targeting capabilities for engineered single effector Cas proteins is
broadened by identifying, characterizing, harnessing, and analyzing the diverse pool of naturally
81
available enzymes and thus there is an interest to continue to identify uncharacterized CRISPR-Cas
systems. In a large computational sequence database mining approach, two novel Class 2 CRISPR-
Cas systems containing the signature protein Casl3b were discovered. Casl3b is classified as a
Type VI CRISPR-Cas system targeting single stranded RNA for cleavage using its two HEPN domains
and also exhibits cleavage of its own CRISPR array(s) and non-specific collateral cleavage activity.
Casl3b loci are devoid of the acquisition genes casi and cas2 but may harbor either csx27 or csx28,
whose proteins repress or enhance interference activity, respectively. Casl3b represents the
second unique single effector Cas protein with targeted RNA cleavage or binding activity that may
be repurposed as a transcriptome modifying tool or to develop RNA diagnostic or imaging assays.
Harnessing Cas13b to modify the transcriptome would enable targeted modifications to RNA
translation, splicing, or deposition of epitranscriptomic marks, all of which are RNA modifications
that are not easily manipulated using current technologies. Before harnessing Cas13b, it is critical
to elucidate its basic biology including functional mechanisms, specificity, role of small proteins, or
even how it undergoes acquisition. Genome modulating technologies have the promise to
revolutionize biological understanding of endogenous nucleic acids or epigenetic markers in any
organism through targeted perturbation of the genome. CRISPR-Cas systems provide an ideal
platform for genome editing because they are active, specific, easy to target, and modular.
Development of these CRISPR-Cas genome or transcriptome engineering tools could help elucidate
the function of causal genetic variations and even offer a potential therapeutic avenue to fix rare
genetic diseases.
82
References:
Abudayyeh, 0.0., Gootenberg, J.S., Konermann, S., Joung, J., Slaymaker, I.M., Cox, D.B., Shmakov, S.,
Makarova, K.S., Semenova, E., Minakhin, L., et al. (2016). C2c2 is a single-component programmable
RNA-guided RNA-targeting CRISPR effector. Science 353, aaf5573.
Bertrand, E., Chartrand, P., Schaefer, M., Shenoy, S.M., Singer, R.H., and Long, R.M. (1998).
Localization of ASH1 mRNA particles in living yeast. Mol Cell 2, 43 7-445.
Bhat, M., Robichaud, N., Hulea, L., Sonenberg, N., Pelletier, J., and Topisirovic, I. (2015). Targeting the
translation machinery in cancer. Nat Rev Drug Discov 14, 261-278.
Bokar, J.A., Shambaugh, M.E., Polayes, D., Matera, A.G., and Rottman, F.M. (1997). Purification and
cDNA cloning of the AdoMet-binding subunit of the human mRNA (N6-adenosine)-
methyltransferase. RNA 3, 1233-1247.
Crittenden, S.L., Bernstein, D.S., Bachorik, J.L., Thompson, B.E., Gallegos, M., Petcherski, A.G.,
Moulder, G., Barstead, R., Wickens, M., and Kimble, J. (2002). A conserved RNA-binding protein
controls germline stem cells in Caenorhabditis elegans. Nature 417, 660-663.
Cullen, B.R. (2006). Enhancing and confirming the specificity of RNAi experiments. Nat Methods 3,
677-681.
De Gregorio, E., Preiss, T., and Hentze, M.W. (1999). Translation driven by an eIF4G core domain in
vivo. EMBO J 18, 4865-4874.
Dominissini, D., Moshitch-Moshkovitz, S., Schwartz, S., Salmon-Divon, M., Ungar, L., Osenberg, S.,
Cesarkas, K., Jacob-Hirsch, J., Amariglio, N., Kupiec, M., et al. (2012). Topology of the human and
mouse m6A RNA methylomes revealed by m6A-seq. Nature 485, 201-206.
East-Seletsky, A., O'Connell, M.R., Knight, S.C., Burstein, D., Cate, J.H., Tjian, R., and Doudna, J.A.
(2016). Two distinct RNase activities of CRISPR-C2c2 enable guide-RNA processing and RNA
detection. Nature 538, 270-273.
Frye, M., Jaffrey, S.R., Pan, T., Rechavi, G., and Suzuki, T. (2016). RNA modifications: what have we
learned and where are we headed? Nat Rev Genet 17, 365-372.
Gao, F., Shen, X.Z., Jiang, F., Wu, Y., and Han, C. (2016). DNA-guided genome editing using the
Natronobacterium gregoryi Argonaute. Nat Biotechnol 34, 768-773.
Gerber, A.P., and Keller, W. (2001). RNA editing by base deamination: more enzymes, more targets,
new mysteries. Trends Biochem Sci 26, 376-384.
Gootenberg, J.S., Abudayyeh, 0.0., Lee, J.W., Essletzbichler, P., Dy, A.J., Joung, J., Verdine, V., Donghia,
N., Daringer, N.M., Freije, C.A., et al. (2017). Nucleic acid detection with CRISPR-Cas13a/C2c2.
Science.
Gu, T., Buaas, F.W., Simons, A.K., Ackert-Bicknell, C.L., Braun, R.E., and Hibbs, M.A. (2012). Canonical
A-to-I and C-to-U RNA editing is enriched at 3'UTRs and microRNA target sites in multiple mouse
tissues. PLoS One 7, e33720.
Laanto, Elina, et al. "Long-term genomic coevolution of host-parasite interaction in the natural
environment." bioRxiv (2017): 101576.
Jia, G., Fu, Y., Zhao, X., Dai, Q., Zheng, G., Yang, Y., Yi, C., Lindahl, T., Pan, T., Yang, Y.G., et al. (2011).
N6-methyladenosine in nuclear RNA is a major substrate of the obesity-associated FTO. Nat Chem
Biol 7, 885-887.
83
Kanasty, R., Dorkin, J.R., Vegas, A., and Anderson, D. (2013). Delivery materials for siRNA
therapeutics. Nat Mater 12, 967-977.
Kim, D., Bae, S., Park, J., Kim, E., Kim, S., Yu, H.R., Hwang, J., Kim, J.I., and Kim, J.S. (2015). Digenome-
seq: genome-wide profiling of CRISPR-Cas9 off-target effects in human cells. Nat Methods 12, 237-
243, 231 p following 243.
Lee, S.H., Turchiano, G., Ata, H., Nowsheen, S., Romito, M., Lou, Z., Ryu, S.M., Ekker, S.C., Cathomen, T.,
and Kim, J.S. (2016). Failure to detect DNA-guided genome editing using Natronobacterium
gregoryi Argonaute. Nat Biotechnol 35, 17-18.
Liu, L., Li, X., Wang, J., Wang, M., Chen, P., Yin, M., Li, J., Sheng, G., and Wang, Y. (2017). Two Distant
Catalytic Sites Are Responsible for C2c2 RNase Activities. Cell 168, 121-134 e112.
Long, R.M., Elliott, D.J., Stutz, F., Rosbash, M., and Singer, R.H. (1995). Spatial consequences of
defective processing of specific yeast mRNAs revealed by fluorescent in situ hybridization. RNA 1,
1071-1078.
Mackay, J.P., Font, J., and Segal, D.J. (2011). The prospects for designer single-stranded RNA-binding
proteins. Nat Struct Mol Biol 18, 256-261.
Martin, K.C., and Ephrussi, A. (2009). mRNA localization: gene expression in the spatial dimension.
Cell 136, 719-730.
Matlin, A.J., Clark, F., and Smith, C.W. (2005). Understanding alternative splicing: towards a cellular
code. Nat Rev Mol Cell Biol 6, 386-398.
Meyer, K.D., Patil, D.P., Zhou, J., Zinoviev, A., Skabkin, M.A., Elemento, 0., Pestova, T.V., Qian, S.B., and
Jaffrey, S.R. (2015). 5' UTR m(6)A Promotes Cap-Independent Translation. Cell 163, 999-1010.
Nishikura, K. (2010). Functions and regulation of RNA editing by ADAR deaminases. Annu Rev
Biochem 79, 321-349.
Pardee, K., Green, A.A., Ferrante, T., Cameron, D.E., DaleyKeyser, A., Yin, P., and Collins, J.J. (2014).
Paper-based synthetic gene networks. Cell 159, 940-954.
Sternberg, S.H., Redding, S., Jinek, M., Greene, E.C., and Doudna, J.A. (2014). DNA interrogation by
the CRISPR RNA-guided endonuclease Cas9. Nature 507, 62-67.
Sunghyeok, Ye, et al. "DNA-dependent RNA cleavage by the Natronobacterium gregoryi
Argonaute." bioRxiv (2017): 101923.z
Tsai, S.Q., Zheng, Z., Nguyen, N.T., Liebers, M., Topkar, V.V., Thapar, V., Wyvekens, N., Khayter, C.,
lafrate, A.J., Le, L.P., et al. (2015). GUIDE-seq enables genome-wide profiling of off-target cleavage
by CRISPR-Cas nucleases. Nat Biotechnol 33, 187-197.
Wang, Y., Cheong, C.G., Hall, T.M., and Wang, Z. (2009). Engineering splicing factors with designed
specificities. Nat Methods 6, 825-830.
Wang, Y., Wang, Z., and Tanaka Hall, T.M. (2013). Engineered proteins with Pumilio/fem-3 mRNA
binding factor scaffold to manipulate RNA metabolism. FEBS J 280, 3755-3767.
Wu, X., Scott, D.A., Kriz, A.J., Chiu, A.C., Hsu, P.D., Dadon, D.B., Cheng, A.W., Trevino, A.E., Konermann,
S., Chen, S., et al. (2014). Genome-wide binding of the CRISPR endonuclease Cas9 in mammalian
cells. Nat Biotechnol 32, 670-676.
Zamore, P.D., Williamson, J.R., and Lehmann, R. (1997). The Pumilio protein binds RNA through a
conserved domain that defines a new class of RNA-binding proteins. RNA 3, 1421-1433.
84
Zheng, G., Dahl, J.A., Niu, Y., Fedorcsak, P., Huang, C.M., Li, C.J., Vagbo, C.B., Shi, Y., Wang, W.L., Song,
S.H., et al. (2013). ALKBH5 is a mammalian RNA demethylase that impacts RNA metabolism and
mouse fertility. Mol Cell 49, 18-29.
85