RESEARCH FRONT
CSIRO PUBLISHING
Review
Aust. J. Chem. 2014, 67, 1741–1750
http://dx.doi.org/10.1071/CH14319
RNA and RNA–Protein Complex Crystallography
and its Challenges
Janine K. Flores,A,B James L. Walshe,A,B and Sandro F. AtaideA,C
A
School of Molecular Bioscience, University of Sydney, Sydney, NSW 2006, Australia.
These authors contributed equally to this paper.
C
Corresponding author. Email: [email protected]
B
RNA biology has changed completely in the past decade with the discovery of non-coding RNAs. Unfortunately, obtaining
mechanistic information about these RNAs alone or in cellular complexes with proteins has been a major problem. X-ray
crystallography of RNA and RNA–protein complexes has suffered from the major problems encountered in preparing and
purifying them in large quantity. Here, we review the available techniques and methods in vitro and in vivo used to prepare
and purify RNA and RNA–protein complex for crystallographic studies. We also discuss the future directions necessary to
explore the vast number of RNA species waiting for their atomic-resolution structure to be determined.
Manuscript received: 20 May 2014.
Manuscript accepted: 24 June 2014.
Published online: 19 August 2014.
Introduction
Interactions between nucleic acids (DNA or RNA) and proteins
are well-known regulators of essential cellular functions. DNA–
protein complexes are involved in regulation of chromosomal
maintenance, replication, transcriptional regulation, and DNA
repair.[1] Initially, RNA–protein interactions were thought to be
only directly involved in transcription, translation, RNA
metabolism, and telomere maintenance among others.[2] However, the RNA component of ribonucleoprotein (RNP) complexes has been found to play a major role in their function and
can participate in catalysis as part of the RNP complex or as the
sole contributor (e.g. spliceosome, ribosome, signal recognition
particle, riboswitches).[2–4] On the other hand, RNA–protein
interactions have only recently begun to be further elucidated.
One of the first steps towards the molecular and mechanistic
understanding of these nucleic acid–protein complexes is to
determine their structure. Whereas the number of crystallographic structures of DNA–protein complexes (2785 structures)
is larger than RNA–protein complexes (1364 structures), most of
RNPs remain to be solved using crystallography. The discrepancy in numbers is due to several problems encountered during
the purification and crystallization of RNA and RNP complexes.
Janine Flores received her B.Sc. in biochemistry and psychology from the University of Sydney in 2012. She completed her
B.Sc. (Honours) in biochemistry in 2013 in the Ataide laboratory, where her research focused on determining the crystal
structure of the protein and RNA components of the eukaryotic signal recognition particle (SRP). Currently, Janine is a Ph.D.
candidate at the Ataide laboratory focusing on the structural characterisation of ribonucleoprotein complexes.
James Walshe received his B.Sc. degree from the University of Western Australia (UWA) in 2010. He completed his B.Sc.
(Honours) in biochemistry in 2011, working in Alice Vrielink’s laboratory (UWA) on metal binding sites found in the crystal
structure of N. meningitidis Lipopolysaccharide transferase A. He is currently a Ph.D. candidate at the Ataide laboratory
(University of Sydney) investigating the structural interactions of ribonucleoproteins that associate with non-coding RNAs.
Dr Sandro Ataide received his B.Sc. in chemistry in 1999 from UNICAMP (Brazil) and obtained his M.Sc. in biochemistry from
the University of Sao Paulo (Brazil) in 2001. He was awarded his Ph.D. by The Ohio State University (USA) in 2006. Aiming to
gain better experience in the RNA and crystallography field, he conducted research on the complex signal recognition particle
(SRP) and ribosome while working as a post-doctorate researcher at Jennifer Doudna’s laboratory at UC Berkeley (USA) and
Dr Nenad Ban’s laboratory at ETH Zurich (Switzerland). He joined the University of Sydney as a lecturer in 2012 to continue
his work on structural and molecular characterisation of RNA and RNA-protein complexes.
Journal compilation Ó CSIRO 2014
www.publish.csiro.au/journals/ajc
1742
This review will address the crystallization of RNA and RNP
complexes from sample preparation, complex reconstitution,
and crystallization both in vitro and in vivo and show the path
required to be taken in order to solve the remaining structures.
In Vitro Techniques to Purify RNA for RNP Complexes
In vitro transcription is the most extensively used method
employed to prepare milligram quantities of the desired RNA,
from a template DNA using RNA polymerase, for crystallographic studies (see methodology in Ref. [5]). Chemical synthesis of RNA is another in vitro method widely used to generate
the desired RNA, which relies on obtaining the desired molecule
from a commercial company.[5,6] Although in vitro transcription
is the most common method of RNA production, there are
several pitfalls with this method for RNA synthesis.[6] One
major concern is the heterogeneity caused by the use of RNA
polymerases for transcription. This is because they can include
several untemplated nucleotides at the 30 end of RNA as well
as having the tendency to prematurely terminate, which is
undesirable for use in crystallography.[6–8] Although this may be
the case, there are several methods that have been compiled to
overcome this issue.
One of the main methods of introducing homogeneity into
the RNA sample is the use of cis- or trans-ribozymes on the
50 and 30 end that will self-cleave after the production of the
RNA.[7–10] Although there is only a limited range of RNAs that
can be purified using current methods, there are several ribozymes that cleave either at the 50 or 30 end and can be used to
assist in purification.[7–10] The widely used ribozymes are
hepatitis delta virus (HDV) ribozyme[7,9,11] and the hammerhead (HH) ribozyme, which are used on the 50 and 30 end
respectively.[8,10] Both are commonly used in creating multimilligram amounts of RNA for crystallography as the requirements for cleavage are compatible with most RNA constructs.
However, introduction of ribozymes causes great difficulty in
the purification of in vitro-transcribed RNA, rendering it quite a
lengthy procedure.
Several methods have been developed to obtain homogeneous in vitro-transcribed RNA with the use of both native
and denaturing methods of purification. However, along with
each method of purification, there are several advantages and
disadvantages depending on the length and structure of the RNA
being purified (Table 1). It is also essential that the purification
protocol used can efficiently remove any heterogeneity in the
sample before the formation of RNPs for crystallography. In
fact, the sample must not only be homogeneous in terms of size
and sequence, but also in terms of structure. Above all, it also
must produce a sufficient yield (milligrams) in order to determine the structure using crystallography.
Denaturing Polyacrylamide Gel Electrophoresis (PAGE)
Denaturing polyacrylamide gel electrophoresis (PAGE) is
one of the most commonly used techniques to purify in vitrotranscribed RNA for crystallography.[6,11] This is often
performed after transcription, where the product of the crude
transcription reaction is loaded onto a pre-heated polyacrylamide gel in which the percentage of the acrylamide depends on
the length of RNA.[12] The desired RNA product is then excised
and extracted from the gel before refolding under optimal
refolding conditions.[6,11,13] Advantages of this procedure
include the effective separation of the desired RNA from free
J. K. Flores, J. L. Walshe, and S. F. Ataide
oligonucleotides, proteins, DNA template, cleaved ribozymes,
and abortive transcripts that may have formed during transcription.[5,14,15] This procedure yields milligram amounts of
RNA for the efficient crystallization of RNA and RNP complexes such as the yitJ SAM-I (S box) riboswitch (nucleotides
(nt) 29–146) of Bacillus subtilis (Fig. 1a) (Protein Data Bank
number (PDB): 4KQY).[16]
Although this method is often used for RNA purification of
both long and short RNAs, there are many disadvantages in
using this technique (Table 1). First, the apparatus has a limited
loading capacity, meaning that several gels are required to gain a
significant yield for the crystallographic analysis of RNA and
RNP complexes.[15,17] Second, this procedure can only resolve
RNA that is 20 to 600 nt with agarose gels being preferable for
larger RNA molecules.[14,15] It is possible to separate RNA to
single-nucleotide resolution by using a higher percentage of
acrylamide (.10 %) in the gel for RNAs .60 nt.[12] Last, a
significant amount of RNA is lost through this purification
procedure, decreasing the yield of transcribed RNA.[18]
Denaturing High-Performance Liquid
Chromatography (HPLC)
In an attempt to find an effective method of RNA purification
without the need for the standard laborious PAGE procedure,
anion-exchange HPLC[19] and ion-pair reverse-phase HPLC[20]
can be used to purify homogeneous RNA that was chemically
synthesized.
The anion-exchange HPLC method involves the use of an
anion-exchange HPLC column (Dionex NuceloPacÒ PA-100,
22 250 mm) of medium capacity fitted with a column heater
that will heat the sample at 80–958C. The sample can elute with
high resolution and sharp peaks on the chromatogram owing to
the conformational homogeneity of denatured RNA at high
temperatures.[19] A further advantage of this approach is the
yield, which is ,2–5 mg per injection of sample. Using this
method, it is possible to achieve single-nucleotide resolution,
while high yield and high resolution can be achieved for RNAs
that are ,25 nt.[19]
In comparison with the previous HPLC method, ion-pair
reverse-phase HPLC requires the column (XBridge C18) to be
heated to 508C during RNA purification to resolve RNA
molecules #59 nt.[20] Although it allows the purification of
longer RNA molecules, owing to the decrease in the temperature
of the column, the nucleotide resolution is highly dependent on
the gradient used. To gain single-nucleotide resolution, a shallow gradient is required, which increases the time required to run
the mixture through the column, whereas a steeper gradient will
decrease the resolution of the purification. Alternatively,
reverse-phase HPLC can be used without heat-denaturing the
RNA to produce RNA suitable for crystallography. This was
done by Sheng et al., where they successfully crystallized 4-Seuridine-derivatized RNA (50 -rU-SeU-CGCG-30 ) (PDB: 3HGA)
using a 21.2 250-mm Zorbax RX-C8 column.[21]
Prep Cell RNA Purification: A Denaturing
and Non-Denaturing PAGE Alternative
In order to improve PAGE as a single purification step,
researchers such as Cunningham et al. and Huang et al. have
used PAGE in conjunction with column chromatography to
purify RNA.[15] This method in comparison with the conventional PAGE protocol involves loading the RNA into a polyacrylamide gel that elutes the sample through a dialysis
Native RNA
purification
techniques
Denaturing RNA
purification
techniques
– DNA affinity and DNAzyme
cleavage
– MS2 affinity and glmS ribozyme
cleavage
ARiBo tag
Affinity tags
– Size-exclusion chromatography
(SEC)
– Anion-exchange chromatography
Chromatographic methods
– Denaturing
– Non-denaturing
Prep-cell
– Anion-exchange HPLC
– Ion-pair, reverse-phase HPLC
Denaturing high-performance
liquid chromatography (HPLC)
Denaturing polyacrylamide
gel electrophoresis (PAGE)
Methods
Time-efficient
Easily removes abortive transcripts
from mixture
High affinity for resin (ARiBo)
Is used as initial purification step (anion)
Time efficient
- Separates constituents of RNA in
in vitro reaction mixture
Separates multimeric species of RNA
Separates homogenous RNA in
term of structure and size (native)
Gel is reusable
Conformational homogeneity
of sample
Separates reagents from RNA
transcription reaction effectively
Advantages
Sequence specific requirement for DNAzyme
method
Cannot separate ribozyme and RNA construct
if similar in length (SEC)
Cannot separate RNA with heterogeneous 30 ends
High salt elution conditions can disrupt RNA
folding (anion)
Not feasible to purify RNA with heterogeneous
30 or 50 ends at a large scale for crystallographic
analysis
Complex to set up
Time-consuming
Resolution dependent on depth of gradient used
Not feasible to purify RNA with heterogeneous
30 or 50 ends at a large scale for crystallographic
analysis
Limited loading capacity
Expensive
Time-consuming
Laborious
Loss of RNA
Disadvantages
Yield of RNA dependent on initial
transcription reaction volume
Overall charge of RNA must be significantly
different to ribozyme’s to efficiently purify
(anion)
Anion can purify RNA constructs that are
30–500 nt in length
Can purify RNA constructs that are
from 20 to 600 nt
2–5 mg of RNA/injection of sample
(anion-exchange HPLC)
– ,29 nt for anion-exchange HPLC
– ,59 nt for ion-pair, reverse-phase HPLC
Can only purify small RNA molecules
Yield dependent on number of gels run
Can purify RNA constructs that are #20
to ,600 nt
Limitations
Table 1. Advantages and disadvantages of denaturing and non-denaturing methods of purifying in vitro-transcribed RNA
Purification and Crystallization of RNA and RNPs
1743
1744
J. K. Flores, J. L. Walshe, and S. F. Ataide
(a)
(b)
(c)
(d)
(e)
(f)
Fig. 1. Cartoon representations of RNA and ribonucleoprotein complexes. RNA molecules depicted in blue, DNA in yellow, and
bound metals as grey spheres. (a) The 2.9-Å-resolution crystal structure of the yitJ S-adenosylmethionine (SAM) riboswitch
(nt 29–146) of B. subtilis (blue). SAM molecule coloured bright orange (Protein Data Bank (PDB) number: 4KQY).[16] (b) The
2.65-Å-resolution crystal structure of the distal region of stem I of T box RNA (PDB: 4JRC).[29] (c) The 2.85-Å-resolution crystal
structure of the second glycine-sensing domain from V. cholerae riboswitch (PBD: 3OWZ).[32] (d) The 3.9-Å-resolution crystal
structure of the prokaryotic signal-recognition particle (orange) in complex with the soluble domain of the signal recognition
particle (SRP) membrane-bound receptor, FtsY (grey), and its cognate 4.5S RNA (PDB: 2AAX).[79] (e) The 2.5-Å-resolution
crystal structure of S. pyogenes Cas9 (grey) in complex with single guide RNA and target DNA (PDB: 4OO8).[80] (f) The 2.8-Å
resolution crystal structure of the multisubunit S. cerevisiae exosome–RNA bound complex. Various subunits are coloured either
grey, orange or cyan (PDB: 4CMP).[81]
membrane and that is then sorted by passing through a UV
detector and separated by fractionation.[15,20] Using this
approach, the eluted RNA can be chromatographically identified. The advantages of this approach include the fact that the gel
is reusable and can extract milligram quantities of RNA over
several runs.[15]
The major advantage of this method can be seen in a paper
by Huang et al. last year in which a 59-nt RNA construct that
could generate different structures was able to be separated
(Table 1).[20] The use of a native gel allowed the separation
of RNA strands that have the same length, but different
structures.[20]
Chromatographic Methods of Native RNA Purification
Recently, there has been a range of chromatographic protocols
for native purification of RNA. They are used as an alternative
to PAGE owing to the efficiency and high yield (milligrams)
of RNA produced by this method. There are two types
of chromatography that have been widely used: size-exclusion
chromatography[22,23] and anion-exchange chromatography.[24–26] These columns are used in conjunction with fastperformance liquid chromatography equipment to allow the
large-scale purification of RNA at a much smaller timescale
than PAGE.
Size exclusion chromatography (SEC) was the first nondenaturing chromatographic method developed to purify RNA.
This method involves first partially purifying the crude transcription mixture using phenol/ chloroform extraction and then
applying the resulting product through SEC to obtain the RNA
product[22,23] (full methodology in Ref. [27]). There are several
advantages to this approach, in particular its ability to successfully remove all the constituents of the transcription reaction in a
time-efficient manner, without the need of denaturants.[22] SEC
can separate multimeric species of RNA from the desired RNA
construct.[22,28]
One disadvantage of SEC is its inability to resolve the closely
related species of RNA, in particular RNA with additional
Purification and Crystallization of RNA and RNPs
nucleotides at the 30 end.[27,28] Therefore, the RNA is required to
be flanked with ribozymes, which in turn leads to problems in
separating the RNA from the ribozymes, requiring a considerable difference in size to be able to separate them using SEC.[23]
Although there are limitations to obtaining suitable samples for
crystallography with particular RNA constructs, this method
does not require any other purification steps.[28] Milligram
quantities of the distal region of stem I of T box RNA transcript
were obtained and crystallized (Fig. 1b) by Grigg et al. (PDB:
4JRC).[29]
Anion-exchange chromatography has become a favourable
technique for RNA purification owing to the quick nature of the
protocol (,4 h). The crude RNA lysate can be directly injected
into the column without the need for prior purification.[28]
Initially, procedures used weak anion-exchange chromatography (diethylaminoethanol (DEAE) Sepharose) for RNA
purification.[25,26] However, RNA purification using stronganion exchange chromatography (Mono Q) has allowed largescale preparation of RNA[24] (full methodology in Ref. [26]).
Although both methods are the same, these procedures differ in
terms of the binding affinity to the column that requires high salt
conditions for elution in the latter procedure, which may affect
RNA folding.[24]
Similarly to SEC, anion-exchange chromatography has the
ability to separate multimeric and aggregated species of RNAs
from the RNA sample.[24,25] However, this is only possible if
their overall charge significantly differs from the desired RNA
sample. Also, this method can separate samples that are 30–500
nt long,[24,25] but any RNA molecules that are .500 nt will
likely be contaminated with DNA.[25] Furthermore, this method
like most others cannot resolve RNA with heterologous
30 ends.[25]
Several structures have been determined through X-ray
crystallography using this as a primary method of purification,
such as the human transfer RNA, tRNASec (PDB: 3A3A),[30]
which used strong anion-exchange chromatography (Resource
Q). Others have used this technique in conjunction with denaturing PAGE as a secondary means of purification to solve
structures such as the catalytic domain of B. stearothermophilus
RNase P RNA, Bst P7D RNA (PDB: 3DHS),[31] and the second
glycine-sensing domain from Vibrio cholerae riboswitch in
several different forms (Fig. 1c) (PBD: 3OWZ).[32]
Using Affinity Tags to Assist in Native RNA Purification
Since the emergence of native RNA purification techniques,
there has been a focus on the development of affinity-tagged
RNA molecules that can be in vitro-transcribed and later cleaved
from the desired RNA with its corresponding substrate.[33–38]
This technique involves two primary components, which
include a sequence that binds into a stationary protein and/or
nucleotide sequence and an activatable ribozyme.[33–38] This
method is analogous to well-known affinity-tagged protein
purification protocols. Owing to the number of abortive
transcripts RNA polymerases make during the process of transcription, the tags are designed on the 30 end. This will allow any
abortive transcripts to flow through the column and allow the
full-length RNA molecules to bind.
Although the majority of tags involve a well-known RNA–
protein interaction, there is one affinity tag technique that uses
DNA affinity to capture the RNA and subsequent cleavage with
a DNAzyme (methodology in Ref. [34]). One of the key aims of
this method is to remove the 30 heterogeneity that is created by
1745
the RNA polymerases.[33] Also, owing to the introduction of the
tag on the 30 end, this eliminates any abortive constructs in
the mixture.[33] This procedure can allow the preparation of
milligrams-worth of RNA and with the removal of the 30
heterogeneity, makes this suitable to produce RNA for crystallography. Furthermore, owing to the lack of the conventional
ribozyme used in the 30 end, the use of DNAzyme makes this
procedure quite cost-effective. However, there is a sequencespecific requirement for this procedure, which is the need for a
purine at the 30 end upstream from the reaction site.[33] Although
this is a simple procedure, the long period of several days is a
drawback.[34]
A commonly used affinity tag combines the glmS ribozyme
followed by the MS2 RNA that interact with the MS2-MBP
fused protein.[35] This combination of affinity tags significantly
improves the purification of RNA through the use of selective
buffers and cleavage substrates.[35] This purification scheme can
allow a broad range of RNAs to be purified; for example, Batey
and Kieft tested RNA molecules of 9, 66, and 94 nt in length and
found that these molecules could be easily purified. They also
assessed the quality of the RNA produced using this method
through recrystallizing the SAM-I riboswitch RNA (PDB:
2GIS).
The ARiBo tag is another commonly used tag, replacing the
MS2 affinity approach. The ARiBo tag contains an activatable
ribozyme (glmS), which has been modified to be able to
integrate the lBoxB RNA into its variable P1 stem[38] (methodology in Ref. [36]). One of the advantages of using this affinity
approach is the fact that the RNA–protein affinity is in the
picomolar range.[37] This allows the bound RNA to be washed
several times, removing any impurities in the mixture before
cleavage. The yield for this method is comparable with gelbased procedures, with the advantage of being more timeefficient.[37] These methods purify RNA using purely native
conditions in comparison with PAGE but with the same yield;
the resulting product is more likely to contain the correctly
folded RNA molecule.
In Vivo RNA Transcription
Along with in vitro methods, in vivo methods for both RNA
transcription and subsequent RNP complex formation have
greatly advanced in recent years.
There are a growing number of reviews that compile evidence highlighting the importance of RNA secondary and
tertiary structure in its biological function. This fact coupled
with the known difficulty of refolding RNA, especially long
non-coding RNA, into a homogeneous species,[39] suggest the
standard method of purifying RNA by denaturing PAGE may
not be ideal. As early as 1988, Uhlenbeck questioned the
wisdom of denaturing RNA during the purification step, likening it to purifying enzymes via sodium dodecyl sulfate (SDS)
gels and expecting them to remain active.[13]
The original pitfalls of in vivo transcription of RNA, such as
the heterogeneity of products, degradation by RNases, and
difficultly purifying RNA from cell extracts, have been
overcome by novel methods of purification. These include
utilizing a tRNA scaffold to avoid RNase degradation[40,41]
and co-expression of RNA with protein chaperones.[42]
Transfer RNA ‘Scaffolding’
Previously, the only successfully expressed RNAs in vivo were
native RNAs, such as tRNAs, or were highly structured and had
1746
J. K. Flores, J. L. Walshe, and S. F. Ataide
(a)
(b)
Sephadex aptamer
tRNA scaffold
RNA insert
MS2 binding hairpin
tRNA scaffold
tRNA scaffold
5’
3’
RNA insert
5’ 3’
tRNA scaffold
MS2 coat protein
Sephadex
aptamer
RNA insert
(c)
(d)
RNase P
(e)
Fab antibodies
U1A
Tetraloop
tRNA
TR
HDV ribozyme
RNA ligase ribozyme
RNase P Protein
Fig. 2. Schematic representations of mechanisms for in vivo RNA expression and engineered crystal contact
of RNAs and RNA–chaperone complexes. The RNA of interest is represented in blue or purple, protein
chaperones or binding proteins are represented in grey, tRNA scaffold in green, and affinity tags or engineered
crystal loops or protein binding loops in red. (a) Plasmid diagram and resulting transcribed product of the tRNA
scaffold method for in vivo RNA and purification. Sephadex affinity domain may be inserted to allow rapid
purification.[41] (b) In vivo-transcribed tRNA scaffold containing the MS2 binding hairpin is simultaneously
produced with overexpressed MS2 coat protein. In vivo pseudo-particle formation internalises the RNA of
interest and protects it from nucleosomal degradation.[41] (c) Distinct crystal contacts between tRNA and RNase
P. The inserted GAAA tetraloop caps the P12 loop of RNase P (red) and contacts the tetraloop receptor (red) on
the tRNA (PDB: 3Q1Q).[14] (d) Designed crystal contacts symmetry-related U1A bound hepatitis delta virus
(HDV) ribozyme. The 14-nt U1A binding loop (red) facilitates U1A protein association (PDB: 1DRZ).[63]
(e) Crystal contact mediated between Fab bound to L1 RNA ligase (PDB: 3IVK).[73]
intrinsic RNase stability.[43] Dardel and Ponchon developed an
elegant mechanism to overcome degradation of the product
RNA during in vivo transcription in Escherichia coli. Transfer
RNA’s intrinsic stability and its precise 50 and 30 end, achieved
through processing by intracellular enzymes, make it an excellent scaffold molecule to insert an RNA of interest. They built
recombinant vectors capable of expressing homologous RNA
‘scaffolds’ in vivo that had the natural ability to withstand
degradation by intracellular nucleases. In these, they replaced
the anti-codon stemloop of the tRNA with an RNA of interest
and either a Sephadex or streptavidin-binding region flanking
the RNA to allow efficient purification from the cell
(Fig. 2a).[41]
One drawback to this approach is the method of removing the
desired RNA from the tRNA scaffold. RNase H with a pair of
DNA oligonucleotides complementary to the scaffold on either
side of the RNA insert has been shown to facilitate cleavage.
However, annealing of the DNA oligonucleotides requires
heating and cooling of the purified RNA, which is not ideal if
native RNA folding is desired. Furthermore, the addition of
RNases, for enzymatic cleavage of the tRNA scaffold, to a
purified RNA introduces the possibility for non-specific digestion of the RNA. To circumvent the use of RNases, the RNA
construct can be flanked with cis-acting hammerhead ribozymes.[44] However, for complete separation of the RNA from
the scaffold post cleavage, both these methods require either
denaturing PAGE separation or denaturing anion-exchange
chromatography.[41,45]
Crystallization of RNA
Crystallization of RNAs for X-ray diffraction studies is
undoubtedly far more challenging than for their protein counterparts. The RNAs’ ability to stably fold into multiple secondary structures,[39] along with their surfaces being dominated
by a regular array of negatively charged phosphates lead to a
general inability of RNAs to form the multiple regular crystal
contacts. This, of course, is not favourable for homologous
crystal growth and X-ray diffraction to high resolution
(,2 Å).[46]
RNA crystallization follows the established principles of
protein crystallization.[47,48] However, as secondary structure
information is often already known for the RNA or has been
computationally predicted, there are several sequence variants
(helix lengths, loops sequences, nucleotide modifications, 50 or
30 overhangs) that can be trialled to optimize crystal growth[46]
and improve diffraction resolution. If the RNA has been denatured during purification, an annealing step must also be
performed before crystallization.[49]
The initial crystallization screening process for RNA varies
only slightly from proteins. RNA and RNP sparse matrix screens
have been developed based on their protein counterparts and
Purification and Crystallization of RNA and RNPs
incomplete factorial screens for both hanging and sitting drop
trials.[49–52] Similarly to proteins, RNA will crystallize from a
wide range of solutions including cacodylate buffers, HEPES,
Tris-HCl and MES using typical precipitants such as PEG and
sodium chloride. More common in nucleic acid crystallization
is the use of lithium salts, 2-propanol, and 2-methyl-2,4pentanediol (MPD) as precipitants.[53] The polyanionic nature
of RNA requires the addition of a neutralizing charge in the
crystallization solution provided by one of a variety of divalent
cations (most commonly a magnesium ion), and/or a polyamine
such as spermine or spermidine.[54] Cations of other valency
states may be added depending on the nature of the RNA
molecules being studied. The hydrolysable nature of RNA
has resulted in a relatively ‘fixed’ pH range, between pH 5
and 8, for crystallization. Although RNA crystallization outside
this range is not impossible, the number of successfully solved
crystal structures decreases significantly outside this range.
Temperature variation has a profound effect on RNA crystallization, more so than that for protein. Initial screening should
include room-temperature samples along with 48C and 308C
samples. The 160-nucleotide domain of the group 1 intron
from Tetrahymena thermophila was the first RNA crystallized
at 308C.[49]
Sequence Optimization
RNA crystals form several types of intermolecular interactions:
base-pairing and stacking, minor groove packing, tetraloop–
minor groove interactions, and pseudo-continuous helices. The
most obvious means of introducing variation into an RNA
sequence is base mutations at the solvent-exposed surface of the
molecule. This may readily be done for small RNAs by varying
the length of the 50 or 30 overhangs.[55]
Determining which RNA base may be mutated in an attempt
to increase possible crystal contacts may be done by experimental probing with chemical reagents such a Fe-EDTA,[56]
DMS or nucleases,[57] and further probing experiments such as
selective 20 -hydroxyl acylation analyzed by primer extension
(SHAPE) and inline probing analysis may be performed.[58,59]
The reactive groups identified by probing indicate solventexposed regions of the RNA capable of forming crystal contacts.
Through elimination, probing allows one to gain an appreciation
of which RNA bases are involved in secondary and tertiary
interactions and must not be targeted for mutations. Although
not routinely done, this step does allow targeted optimization of
RNA and is particularly helpful to access the overall secondary
structure for those that lack an in vitro activity assay and for
which post-mutational activity assays cannot be performed.
Replacement of solvent-exposed regions of the RNA with
RNA tetraloops, especially the GNRA/G tetraloops (where N is
A, G, C or U),[51] is a common strategy, because these motifs
provide a compact structure and are known to mediate RNA
packing.[60] In addition to the GNRA tetraloops, the GAAA
tetraloop–receptor interaction has also been used to facilitate
RNA packing and crystallization in multiple structures.[46,61,62]
Perhaps the most famous example is of the structure of the
universal ribozyme RNase P complexed with tRNA. To achieve
diffraction-quality crystals, the P12 loop of RNase P was capped
with a GAAA tetraloop and the tetraloop receptor was added to
the pre-tRNA anticodon stem loop (ASL) (Fig. 2c). This
modification, along with subsequent nucleotide elongations at
the ASL and P12 loop allowed diffraction-quality crystals
(3.8 Å) to be obtained.[14]
1747
Chaperone Proteins, Antibodies, and Pseudo-Particles
Further to the introduction of foreign loops to assist with crystal
packing, RNA protein-binding motifs or antibody-binding
sequences may be introduced. Not only will the associated
protein or antibody provide a ‘rigid’ tertiary structure, which
will facilitate ordered crystal contacts during crystal growth, but
they will also provide a platform for solution of the phase
problem through either molecular replacement or multi-wavelength anomalous dispersion (MAD) phasing.
The first, and now most commonly used, chaperone protein
for RNA crystallization is the RRM-I domain of the human
spliceosomal U1A protein (U1A-RBD). It was used effectively
to solve the long-awaited structure of the HDV ribozyme. The
alteration of the P4 stem of the HDV ribozyme, with the 22-nt
RNA stemloop ‘ligand’, did not adversely affect the selfcleavage activity of the HDV ribozyme, nor did the binding of
U1A-RBD,[63] suggesting RNA tertiary structure was retained.
U1A-RBD was an ideal choice for cocrystallization with RNA,
first because it binds its cognate RNA with very high affinity
(dissociation constant Kd of 2 1011 M),[64] and second, the
high-resolution crystal structure of U1A-RBD bound to SLII
RNA showed the protein–RNA interaction is confined to a
single 10-nt loop of the RNA,[65] leading to the possibility of
as few as 12 nucleotides being required for high-affinity binding
to U1A-RBD.[66] This method has been used to solve the crystal
structure of the Hairpin ribozyme (Fig. 2d),[67] Glm ribozymeriboswitch,[68] and the flexizyme[69] among other structures.
The use of antibodies is an alternative approach to overcome
the inherent instability issues of RNA when performing crystallization trials. Antibody fragments have been used as chaperones, in a similar manner as with membrane proteins.[70] The use
of antibodies may also potentially eliminate conformational
heterogeneity associated with RNA and, in addition, the large
size of the antibody would allow phasing via molecular replacement.[71] To this end, Ye et al. used a phage platform for the
display of libraries of synthetic antigen-binding fragments
(Fabs) and used the P4–P6 domain from the Tetrahymena group
I intron as a ‘proof of concept’ RNA antigen. The selected Fab
demonstrated high binding affinity (Kd 51 nM) with the P4–P6
RNA domain and showed specificity at the tertiary level when
compared with other RNA. The antibody–RNA complex was
successfully crystallized and diffracted to 2.25 Å.[72] Following
a similar methodology, the catalytic core of a class 1 RNA ligase
ribozyme was crystallized and diffracted to 3.0 Å (Fig. 2e);[73]
however, they remain the only two RNA structured crystallized
in complex with an antibody.
Crystallization of RNA Protein Complexes
Complex Reassembly In Vitro
Typically, crystallization of RNA protein complexes is performed using complex reconstituted in vitro from individual
components. RNA, most commonly transcribed by T7 polymerase in vitro transcription, or commercially chemically synthesized if short enough, is incubated with various components
of the RNP complex (usually expressed and purified from
E. coli[74,75]), before the reconstituted complex is purified from
unbound components via gel filtration or ion-exchange chromatography. Molar RNA : protein ratios of 1 : 0.9 are suggested
for initial screening, guided by gel shift assays to ensure complete complex formation.[76] Incubation time and temperature
must be optimized, through trial and error, to ensure maximal
efficiency of complex reassembly. Special care must be taken
1748
during this process to ensure that first, the RNA transcribed is
chemically homogeneous (methods described earlier ensure
this) and second, that purified protein is free of RNase
contamination.
Several RNP complexes have been successfully crystallized
in this way. Ataide et al. reconstituted the prokaryotic signal
recognition particle (SRP) in complex with the soluble domain
of the SRP membrane-bound receptor, FtsY. The prokaryotic
SRP consists of a 4.5S RNA bound to the Ffh protein. To
reassemble the complex, 4.5S was transcribed by T7 polymerase
in vitro, purified by denaturing PAGE, and refolded.[77,78]
Sequential incubations of refolded 4.5S with 2 molar excess
of Ffh and FtsY, followed by complex purification using anionexchange chromatography yielded pure complex, which was
successfully crystallized, and diffraction to 3.9 Å obtained
(Fig. 1d).[79]
Similarly, the crystal structure of Streptococcus pyogenes
Cas9 in complex with single-guide RNA (sgRNA) and its target
DNA was recently solved using in vitro reassembly.[80] Cas9
was expressed and purified from E. coli and incubated with the
98-nt in vitro-transcribed sgRNA, purified by 10 % PAGE,
along with commercially purchased 23-nt target DNA at a molar
ratio of 1 : 1.5 : 2.3. The complex was purified by gel filtration
before crystallization (Fig. 1e).[80]
Makino et al. successfully used a well-designed small commercially synthesized RNA to solve the remarkable structure of
the 440-kDa S. cerevisiae exosome–RNA-bound complex.[81]
The exosome is responsible for 30 –50 degradation of RNA in
eukaryotes and consists of 11 subunits forming a non-specific
RNA-binding channel. Following in vitro reconstitution of the
subunits (expressed and purified from E. coli[82]), RNAse
protection assays were performed to determine the exact length
of solvent-inaccessible RNA bases within the complex. This
allowed the design of a synthetic RNA, which consists of a 30
overhang that spans the RNA-binding pore, and a 50 -CG duplex
closed by a tetraloop, which stalls further entry of the RNA into
the narrow pocket. Incubation of the complex with RNA, in 2
molar excess, was sufficient for complex formation and subsequent crystallization, and diffraction to 2.8 Å (Fig. 1f).[81,82]
RNP Complex Reassembly In Vivo
Until recently, crystallization of overexpressed RNP complexes
reassembled in vivo has been a prospective concept. Methods
for in vivo production of RNA have progressed significantly,
based around the use of a stable tRNA scaffold.[41] This work
was furthered with the design of an RNA–protein expression
system utilizing both single or double plasmid mechanisms.
RNA–protein co-expression assays in E. coli were used to validate the system and co-expression of phi2 prohead RNA
(pRNA) and its binding partner gp10 were successful in vivo.[42]
Furthermore, genomic RNA packaged in the E. coli bacteriophage MS2 is resistant to RNase digestion. The bacteriophage
MS2 adopts a simple RNP structure composed of 180 MS2
molecules. Additionally, non-phage recombinant RNAs containing a 19-nt hairpin operator sequence can also induce
pseudo-particle formation.[83] Co-expression of both MS2 and
transfer–messenger RNA (tmRNA) (a unique 347-nt bifunctional mRNA/tRNA) containing the hairpin operator sequence
in E. coli led to the assembly of pseudo-particles that were
protected against RNA degradation, when compared with
tmRNA in vivo expression alone. Following phenol extraction,
tmRNA was purified in a single Superose 6 SEC step, providing
the first example of in vivo RNA transcription and purification
J. K. Flores, J. L. Walshe, and S. F. Ataide
utilizing pseudo-particle complex assembly for nuclease protection (Fig. 2b).[42]
This approached was extended to RNA–protein complexes
without the tRNA scaffold. The authors theorized that any RNAbinding protein co-expressed with its RNA would have the
similar shielding effect from nuclease degradation as the MS2
pseudo-particles did. They used the broad RNA-binding Hfq
protein and the 227-nt SgrS RNA as a proof of concept. In vivo
expression of SrgS was observed only when it was co-expressed
with Hfq, suggesting Hfq decreases the sensitivity of SgrS to
nucleolysis and confirming in vivo reassembly.[42] Affinity
tagging of the RNA-binding protein would allow rapid purification of the in vivo-reconstituted complex and may prove to be a
useful method for RNP purification that circumvents in vitro
reassembly.
To date, the most well-known high-resolution crystal structure of an RNP complex purified from in vivo is the ribosome.
Decades of experimentation finally led to the structural solution
of the prokaryotic ribosome from T. thermophilus[84,85] and
H. marismortui.[86] The natural abundance of the ribosome, its
large size, and the inherent stability of the incorporated RNA
when complexed with the ribosome allowed sufficient quantities to be purified from these organisms. More recently, the
structure of the eukaryotic 80S ribosome was solved following
direct harvesting from S. cerevisiae cells[87] as well as the 40S
and 60S subunits from T. thermophila.[88,89]
Conclusion
Over the past three decades, a new era in RNA biology has been
rapidly gaining momentum. What started with the discovery of
the first renegade non-coding RNAs and small nuclear RNAs in
the 1980s progressed to micro-RNAs in the 2000s and today,
with the advancements in deep sequencing, thousands of long
non-coding RNAs are being unearthed. Along the way, ribozymes and riboswitches were exposed as remarkable RNAs that
utilize their structure to catalyze reactions in a role previously
thought to be the sole responsibility of protein enzymes.[90]
Overall, RNAs can regulate gene expression at the transcription
level, RNA processing and translation, they can protect against
foreign nucleic acids, guide DNA synthesis, catalyze reactions,
and provide a scaffold for RNP assembly (for a review, see Ref.
[91]). The transcribed non-coding region of human DNA far
eclipses the size of the protein-coding region.[91,92] Additionally, RNA structure and fold variability surpasses that of proteins, in terms of complexity, suggesting a plethora of
structurally specific biological functions of which our current
understanding barely scratches the surface.
Despite the growing appreciation of the importance of
RNAs, the methods used to study them, particularly those used
to determine molecular models by high-resolution X-ray diffraction, remain limited. At present, there are 88532 X-ray
diffraction structures in the PDB. Only 2.33 % (2062 structures)
are of RNA molecules and 1.54 % (1364 structures) are of RNPs,
demonstrating the lack of advancement in the area. Unlike the
novel techniques, such as cubic-phase lipid crystallization[93]
currently being developed for crystallization of membrane
proteins, RNA and RNP crystallography so far has lacked a
similar step forward. Current techniques, as discussed in the
review, are the beginning of tackling the bottleneck surrounding
RNA and RNP crystallization. New techniques, utilizing combinations of the methods discussed and novel techniques, are
required to further advance the field.
Purification and Crystallization of RNA and RNPs
Acknowledgements
The Australian Research Council (ARC) and the University of Sydney
supported this work. A Commonwealth Australian Postgraduate Award and
an ARC-supported scholarship funded James Walshe and Janine Flores
respectively.
References
[1] B. Dey, S. Thukral, S. Krishnan, M. Chakrobarty, S. Gupta,
C. Manghani, V. Rani, Mol. Cell. Biochem. 2012, 365, 279.
doi:10.1007/S11010-012-1269-Z
[2] A. Perederina, A. Krasilnikov, in Bacterial Regulatory RNA, Methods
in Molecular Biology Series (Ed. K. C. Keiler) 2012, Vol. 905,
pp. 123–143 (Humana Press: Totowa, NJ).
[3] E. Obayashi, C. Oubridge, D. Pomeranz Krummel, K. Nagai, Methods
Mol. Biol. 2007, 363, 259. doi:10.1007/978-1-59745-209-0_13
[4] S. Cusack, Curr. Opin. Struct. Biol. 1999, 9, 66. doi:10.1016/S0959440X(99)80009-8
[5] J. L. Linpinsel, G. L. Conn, Methods Mol. Biol. 2013, 941, 43.
doi:10.1007/978-1-62703-113-4_4
[6] J. F. Milligan, D. R. Groebe, G. W. Witherell, O. C. Uhlenbeck,
Nucleic Acids Res. 1987, 15, 8783. doi:10.1093/NAR/15.21.8783
[7] H. Schurer, K. Lang, J. Schuster, M. Morl, Nucleic Acids Res. 2002, 30,
56e. doi:10.1093/NAR/GNF055
[8] S. C. Walker, J. M. Avis, G. L. Conn, Nucleic Acids Res. 2003, 31, 82e.
doi:10.1093/NAR/GNG082
[9] A. R. Ferre-D’Amare, J. A. Doudna, Nucleic Acids Res. 1996, 24, 977.
doi:10.1093/NAR/24.5.977
[10] S. R. Price, N. Ito, C. Oubridge, J. M. Avis, K. Nagai, J. Mol. Biol.
1995, 249, 398. doi:10.1006/JMBI.1995.0305
[11] See pp. 365–370 in: J. A. Doudna, Ribozyme Protocols 1997 (Springer:
Berlin).
[12] A. Petrov, T. Wu, E. V. Puglisi, J. D. Puglisi, Methods Enzymol. 2013,
530, 315. doi:10.1016/B978-0-12-420037-1.00017-8
[13] O. C. Uhlenbeck, RNA 1995, 1, 4.
[14] N. J. Reiter, A. Osterman, A. Torres-Larios, K. K. Swinger, T. Pan,
A. Mondragon, Nature 2010, 468, 784. doi:10.1038/NATURE09516
[15] L. Cunningham, K. Kittikamron, Y. Lu, Nucleic Acids Res. 1996, 24,
3647. doi:10.1093/NAR/24.18.3647
[16] C. Lu, F. Ding, A. Chowdhury, V. Pradhan, J. Tomsic, W. M. Holmes,
T. M. Henkin, A. Ke, J. Mol. Biol. 2010, 404, 803. doi:10.1016/J.JMB.
2010.09.059
[17] F. Wincott, A. DiRenzo, C. Shaffer, S. Grimm, D. Tracz, C. Workman,
D. Sweedler, C. Gonzalez, S. Scaringe, N. Usman, Nucleic Acids Res.
1995, 23, 2677. doi:10.1093/NAR/23.14.2677
[18] C. Wilms, P. Wollenzien, Anal. Biochem. 1994, 221, 204. doi:10.1006/
ABIO.1994.1399
[19] T. P. Shields, E. Mollova, L. Ste. Marie, M. R. Hansen, A. Pardi, RNA
1999, 5, 1259. doi:10.1017/S1355838299990945
[20] Z. Huang, S. Jayaseelan, J. Hebert, H. Seo, L. Niu, Anal. Biochem.
2013, 435, 35. doi:10.1016/J.AB.2012.12.011
[21] J. Sheng, J. Gan, A. S. Soares, J. Salon, Z. Huang, Nucleic Acids Res.
2013, 41, 10476. doi:10.1093/NAR/GKT799
[22] I. Kim, S. A. McKenna, E. Viani Puglisi, J. D. Puglisi, RNA 2006, 13,
289. doi:10.1261/RNA.342607
[23] P. J. Lukavsky, J. D. Puglisi, RNA 2004, 10, 889. doi:10.1261/RNA.
5264804
[24] J. Koubek, K. F. Lin, Y. R. Chen, R. P. Cheng, J. J. T. Huang, RNA
2013, 19, 1449. doi:10.1261/RNA.038117.113
[25] L. E. Easton, Y. Shibata, P. J. Lukavsky, RNA 2010, 16, 647.
doi:10.1261/RNA.1862210
[26] A. Y. Keel, L. E. Easton, P. J. Lukavsky, J. S. Kieft, Methods Enzymol.
2009, 469, 3. doi:10.1016/S0076-6879(09)69001-7
[27] E. P. Booy, H. Meng, S. A. McKenna, Methods Mol. Biol. 2013, 941,
69. doi:10.1007/978-1-62703-113-4_6
[28] R. T. Batey, Curr. Opin. Struct. Biol. 2014, 26, 1. doi:10.1016/J.SBI.
2014.01.014
1749
[29] J. C. Grigg, Y. Chen, F. J. Grundy, T. M. Henkin, L. Pollack, A. Ke,
Proc. Natl. Acad. Sci. USA 2013, 110, 7240. doi:10.1073/PNAS.
1222214110
[30] Y. Itoh, S. Chiba, S.-i. Sekine, S. Yokoyama, Nucleic Acids Res. 2009,
37, 6259. doi:10.1093/NAR/GKP648
[31] A. V. Kazantsev, A. A. Krivenko, N. R. Pace, RNA 2009, 15, 266.
doi:10.1261/RNA.1331809
[32] L. Huang, A. Serganov, D. J. Patel, Mol. Cell 2010, 40, 774.
doi:10.1016/J.MOLCEL.2010.11.026
[33] H.-K. Cheong, E. Hwang, C. Lee, B.-S. Choi, C. Cheong, Nucleic
Acids Res. 2004, 32, e84. doi:10.1093/NAR/GNH081
[34] H.-K. Cheong, E. Hwang, C. Cheong, Methods Mol. Biol. 2013, 941,
113. doi:10.1007/978-1-62703-113-4_9
[35] R. T. Batey, J. S. Kieft, RNA 2007, 13, 1384. doi:10.1261/RNA.528007
[36] G. Di Tomasso, P. Dagenais, A. Desjardins, A. Rompre-Brodeur,
V. Delfosse, P. Legault, Methods Mol. Biol. 2013, 941, 137.
doi:10.1007/978-1-62703-113-4_11
[37] G. Di Tomasso, P. Lampron, P. Dagenais, J. G. Omichinski, P. Legault,
Nucleic Acids Res. 2011, 39, e18. doi:10.1093/NAR/GKQ1084
[38] A. Salvail-Lacoste, G. Di Tomasso, B. L. Piette, P. Legault, RNA 2013,
19, 1003. doi:10.1261/RNA.037432.112
[39] D. H. Turner, N. Sugimoto, S. M. Freier, Annu. Rev. Biophys. Biophys.
Chem. 1988, 17, 167. doi:10.1146/ANNUREV.BB.17.060188.001123
[40] L. Ponchon, G. Beauvais, S. Nonin-Lecomte, F. Dardel, Nat. Protoc.
2009, 4, 947. doi:10.1038/NPROT.2009.67
[41] L. Ponchon, F. Dardel, Nat. Methods 2007, 4, 571. doi:10.1038/
NMETH1058
[42] L. Ponchon, M. Catala, B. Seijo, M. El Khouri, F. Dardel, S. NoninLecomte, C. Tisne, Nucleic Acids Res. 2013, 41, e150. doi:10.1093/
NAR/GKT576
[43] C. Tisné, M. Rigourd, R. Marquet, C. Ehresmann, F. Dardel, RNA
2000, 6, 1403. doi:10.1017/S1355838200000947
[44] F. H. Nelissen, E. H. Leunissen, L. van de Laar, M. Tessari, H. A. Heus,
S. S. Wijmenga, Nucleic Acids Res. 2012, 40, e102. doi:10.1093/NAR/
GKS292
[45] L. Ponchon, F. Dardel, Methods 2011, 54, 267. doi:10.1016/
J.YMETH.2011.02.007
[46] A. R. Ferré-D’Amaré, K. Zhou, J. A. Doudna, J. Mol. Biol. 1998, 279,
621. doi:10.1006/JMBI.1998.1789
[47] A. McPherson, J. Cryst. Growth 1991, 110, 1. doi:10.1016/0022-0248
(91)90859-4
[48] A. McPherson, Crystallization of Biological Macromolecules 1999
(Cold Spring Harbor Laboratory Press: New York, NY).
[49] J. A. Doudna, C. Grosshans, A. Gooding, C. E. Kundrot, Proc. Natl.
Acad. Sci. USA 1993, 90, 7829. doi:10.1073/PNAS.90.16.7829
[50] J. Jancarik, S.-H. Kim, J. Appl. Cryst. 1991, 24, 409. doi:10.1107/
S0021889891004430
[51] W. G. Scott, J. T. Finch, R. Grenfell, J. Fogg, T. Smith, M. J. Gait,
A. Klug, J. Mol. Biol. 1995, 250, 327. doi:10.1006/JMBI.1995.0380
[52] C. Carter, J. Biol. Chem. 1979, 254, 12219.
[53] M. S. Jurica, L. J. Licklider, S. R. Gygi, N. Grigorieff, M. J. Moore,
RNA 2002, 8, 426. doi:10.1017/S1355838202021088
[54] A. R. Ferré-D’Amaré, J. A. Doudna, Curr. Protoc. Nucleic Acid Chem.
2001, 7, 1.
[55] F. E. Reyes, A. D. Garst, R. T. Batey, in Methods in Enzymology
(Ed. H. Daniel) 2009, Vol. 469, pp. 119–139 (Academic Press:
New York, NY).
[56] D. W. Celander, T. R. Cech, Biochemistry 1990, 29, 1355.
doi:10.1021/BI00458A001
[57] C. Ehresmann, F. Baudin, M. Mougel, P. Romby, J.-P. Ebel,
B. Ehresmann, Nucleic Acids Res. 1987, 15, 9109. doi:10.1093/NAR/
15.22.9109
[58] E. J. Merino, K. A. Wilkinson, J. L. Coughlan, K. M. Weeks, J. Am.
Chem. Soc. 2005, 127, 4223. doi:10.1021/JA043822V
[59] See pp. 53-67 in: E. E. Regulski, R. R. Breaker, Post-Transcriptional
Gene Regulation 2008 (Springer: Berlin).
[60] H. W. Pley, K. M. Flaherty, D. B. McKay, Nature 1994, 372, 111.
doi:10.1038/372111A0
1750
[61] L. A. Coonrod, J. R. Lohman, J. A. Berglund, Biochemistry 2012, 51,
8330. doi:10.1021/BI300829W
[62] J. L. Childs-Disney, I. Yildirim, H. Park, J. R. Lohman, L. Guan, T.
Tran, P. Sarkar, G. C. Schatz, M. D. Disney, ACS Chem. Biol. 2014, 9,
538. doi:10.1021/CB4007387
[63] A. R. Ferré-D’Amaré, K. Zhou, J. A. Doudna, Nature 1998, 395, 567.
doi:10.1038/26912
[64] K. B. Hall, W. T. Stump, Nucleic Acids Res. 1992, 20, 4283.
doi:10.1093/NAR/20.16.4283
[65] C. Oubridge, N. Ito, P. R. Evans, C.-H. Teo, K. Nagai, Nature 1994,
372, 432. doi:10.1038/372432A0
[66] A. R. Ferré-D’Amaré, Methods 2010, 52, 159. doi:10.1016/J.YMETH.
2010.06.008
[67] P. B. Rupert, A. R. Ferré-D’Amaré, Nature 2001, 410, 780.
doi:10.1038/35071009
[68] J. C. Cochrane, S. V. Lipchock, S. A. Strobel, Chem. Biol. 2007, 14, 97.
doi:10.1016/J.CHEMBIOL.2006.12.005
[69] H. Xiao, H. Murakami, H. Suga, A. R. Ferre-D’Amare, Nature 2008,
454, 358. doi:10.1038/NATURE07033
[70] S. Uysal, V. Vásquez, V. Tereshko, K. Esaki, F. A. Fellouse, S. S.
Sidhu, S. Koide, E. Perozo, A. Kossiakoff, Proc. Natl. Acad. Sci. USA
2009, 106, 6644.
[71] Y. Koldobskaya, E. M. Duguid, D. M. Shechner, N. B. Suslov, J. Ye,
S. S. Sidhu, D. P. Bartel, S. Koide, A. A. Kossiakoff, J. A. Piccirilli,
Nat. Struct. Mol. Biol. 2011, 18, 100. doi:10.1038/NSMB.1945
[72] J.-D. Ye, V. Tereshko, J. K. Frederiksen, A. Koide, F. A. Fellouse, S. S.
Sidhu, S. Koide, A. A. Kossiakoff, J. A. Piccirilli, Proc. Natl. Acad.
Sci. USA 2008, 105, 82.
[73] D. M. Shechner, R. A. Grant, S. C. Bagby, Y. Koldobskaya, J. A.
Piccirilli, D. P. Bartel, Science 2009, 326, 1271. doi:10.1126/
SCIENCE.1174676
[74] F. Baneyx, Curr. Opin. Biotechnol. 1999, 10, 411. doi:10.1016/S09581669(99)00003-8
[75] J.-C. Janson, Protein Purification: Principles, High Resolution Methods and Applications 2012 (John Wiley & Sons: Hoboken, NJ).
[76] A. Ke, J. A. Doudna, Methods 2004, 34, 408. doi:10.1016/J.YMETH.
2004.03.027
[77] R. T. Batey, M. B. Sagar, J. A. Doudna, J. Mol. Biol. 2001, 307, 229.
doi:10.1006/JMBI.2000.4454
[78] F. Y. Siu, R. J. Spanggord, J. A. Doudna, RNA 2007, 13, 240.
doi:10.1261/RNA.135407
J. K. Flores, J. L. Walshe, and S. F. Ataide
[79] S. F. Ataide, N. Schmitz, K. Shen, A. Ke, S.-o. Shan, J. A. Doudna,
N. Ban, Science 2011, 331, 881. doi:10.1126/SCIENCE.1196473
[80] H. Nishimasu, F. A. Ran, P. D. Hsu, S. Konermann, S. I. Shehata,
N. Dohmae, R. Ishitani, F. Zhang, O. Nureki, Cell 2014, 156, 935.
doi:10.1016/J.CELL.2014.02.001
[81] D. L. Makino, M. Baumgartner, E. Conti, Nature 2013, 495, 70.
doi:10.1038/NATURE11870
[82] J. C. Greimann, C. D. Lima, in Methods in Enzymology (Eds E. M.
Lynne, K. Megerditch) 2008, Vol. 448, pp. 185–210 (Academic Press:
New York, NY).
[83] G. G. Pickett, D. S. Peabody, Nucleic Acids Res. 1993, 21, 4621.
doi:10.1093/NAR/21.19.4621
[84] F. Schluenzen, A. Tocilj, R. Zarivach, J. Harms, M. Gluehmann,
D. Janell, A. Bashan, H. Bartels, I. Agmon, F. Franceschi, A. Yonath,
Cell 2000, 102, 615. doi:10.1016/S0092-8674(00)00084-2
[85] B. T. Wimberly, D. E. Brodersen, W. M. Clemons, R. J. MorganWarren, A. P. Carter, C. Vonrhein, T. Hartsch, V. Ramakrishnan,
Nature 2000, 407, 327. doi:10.1038/35030006
[86] N. Ban, P. Nissen, J. Hansen, P. B. Moore, T. A. Steitz, Science 2000,
289, 905. doi:10.1126/SCIENCE.289.5481.905
[87] A. Ben-Shem, L. Jenner, G. Yusupova, M. Yusupov, Science 2010,
330, 1203. doi:10.1126/SCIENCE.1194294
[88] J. Rabl, M. Leibundgut, S. F. Ataide, A. Haag, N. Ban, Science 2011,
331, 730. doi:10.1126/SCIENCE.1198308
[89] S. Klinge, F. Voigts-Hoffmann, M. Leibundgut, S. Arpagaus, N. Ban,
Science 2011, 334, 941. doi:10.1126/SCIENCE.1211204
[90] K. Kruger, P. J. Grabowski, A. J. Zaug, J. Sands, D. E. Gottschling,
T. R. Cech, Cell 1982, 31, 147. doi:10.1016/0092-8674(82)90414-7
[91] E. Birney, J. A. Stamatoyannopoulos, A. Dutta, R. Guigó, T. R.
Gingeras, E. H. Margulies, et al., Nature 2007, 447, 799.
doi:10.1038/NATURE05874
[92] J. Cheng, P. Kapranov, J. Drenkow, S. Dike, S. Brubaker, S. Patel,
J. Long, D. Stern, H. Tammana, G. Helt, V. Sementchenko,
A. Piccolboni, S. Bekiranov, D. K. Bailey, M. Ganesh, S. Ghosh,
I. Bell, D. S. Gerhard, T. R. Gingeras, Science 2005, 308, 1149.
doi:10.1126/SCIENCE.1108625
[93] E. M. Landau, J. P. Rosenbusch, Proc. Natl. Acad. Sci. USA 1996, 93,
14532. doi:10.1073/PNAS.93.25.14532