University of Groningen
Structural basis for CRISPR RNA-guided DNA recognition by Cascade
Jore, Matthijs M.; Lundgren, Magnus; van Duijn, Esther; Bultema, Jelle B.; Westra, Edze R.;
Waghmare, Sakharam P.; Wiedenheft, Blake; Pul, Uemit; Wurm, Reinhild; Wagner, Rolf
Published in:
Nature Structural & Molecular Biology
DOI:
10.1038/nsmb.2019
IMPORTANT NOTE: You are advised to consult the publisher's version (publisher's PDF) if you wish to cite from
it. Please check the document version below.
Document Version
Publisher's PDF, also known as Version of record
Publication date:
2011
Link to publication in University of Groningen/UMCG research database
Citation for published version (APA):
Jore, M. M., Lundgren, M., van Duijn, E., Bultema, J. B., Westra, E. R., Waghmare, S. P., Wiedenheft, B.,
Pul, U., Wurm, R., Wagner, R., Beijer, M. R., Barendregt, A., Zhou, K., Snijders, A. P. L., Dickman, M. J.,
Doudna, J. A., Boekema, E. J., Heck, A. J. R., van der Oost, J., ... Pul, Ü. (2011). Structural basis for
CRISPR RNA-guided DNA recognition by Cascade.
Nature Structural & Molecular Biology
,
18
(5), 529-
U141. https://doi.org/10.1038/nsmb.2019
Copyright
Other than for strictly personal use, it is not permitted to download or to forward/distribute the text or part of it without the consent of the
author(s) and/or copyright holder(s), unless the work is under an open content license (like Creative Commons).
The publication may also be distributed here under the terms of Article 25fa of the Dutch Copyright Act, indicated by the “Taverne” license.
More information can be found on the University of Groningen website: https://www.rug.nl/library/open-access/self-archiving-pure/taverne-
amendment.
Take-down policy
If you believe that this document breaches copyright please contact us providing details, and we will remove access to the work immediately
and investigate your claim.
Downloaded from the University of Groningen/UMCG research database (Pure): http://www.rug.nl/research/portal. For technical reasons the
number of authors shown on this cover page is limited to 10 maximum.
Download date: 10-08-2022
A r t i c l e s
nAture structurAl & moleculAr biology VOLUME 18 NUMBER 5 MAY 2011 529
The constant pressure of invading viruses and conjugative plasmids
has shaped the evolution of host defense systems in prokaryotes. The
widely distributed CRISPR immune system represents a recently
discovered adaptive and inheritable defense strategy
1–5
. The system
consists of repeats that are interspersed with unique sequences called
spacers, which are derived from viral and plasmid DNA
6–8
.
The Cas protein machinery is encoded by gene clusters that are located
close to the CRISPR locus
9
. Multiple types of cas gene sets have been rec-
ognized
10,11
that correlate with specific families of repeat sequences
12
.
The mechanism by which CRISPR and Cas induce immunity has been
divided into three stages. In the first stage—CRISPR adaptation—the
host encounters an invader and integrates a random fragment of foreign
DNA nondirectionally into the CRISPR locus as a new spacer, resulting
in resistance to foreign genetic elements carrying this sequence
13–15
. The
metal-dependent DNase Cas1 might be involved in generating small
DNA fragments that are used as precursors for CRISPR adaptation
16
.
The second stage—CRISPR expression—involves transcription and
translation of the cas genes and transcription of the CRISPR, yielding a
precursor CRISPR RNA (pre-crRNA). The pre-crRNA is cleaved in the
repeat regions
17–19
by specific Cas endoRNases: CasE from E. coli
20
, Cas6
from Pyrococcus furiosus
21
and Csy4 from Pseudomonas aeruginosa
22
.
CasE-generated mature crRNAs remain bound to a protein complex
called Cascade (CRISPR-associated complex for antiviral defense),
which consists of five Cas proteins of the Cse-type (CasABCDE; Fig. 1a).
Cas6-generated crRNAs end up in the Cmr complex (Cmr1–Cmr6) and
are further trimmed at the 3′ end from around 67 to 39 or 45 nucleotides.
The guide RNA-loaded Cmr complex cleaves single-stranded target RNA
sequence-specifically
23
. In E. coli, the third stage—CRISPR interference—
requires not only Cascade loaded with anti-invader crRNA, but also the
participation of the predicted nuclease-helicase Cas3. Because crRNAs
that are complementary to either strand of the phage DNA provide resist-
ance, it has been proposed that Cascade is a crRNA-guided complex that
targets DNA rather than mRNA
20
. In vivo experiments in Staphylococcus
epidermidis involving a conjugative plasmid proved that the Csm-type
CRISPR-Cas system targets DNA
24
. In Streptococcus thermophilus,
bacteriophage resistance and plasmid loss result from site-specific
cleavage of viral and plasmid DNA within the spacer-targeted sequence,
and this process requires the csn1-like gene
15
.
Although viral or plasmid DNA seems to be the target molecule
in these three bacterial systems, there is no biochemical evidence
1
Laboratory of Microbiology, Department of Agrotechnology and Food Sciences, Wageningen University, Wageningen, The Netherlands.
2
Biomolecular Mass
Spectrometry and Proteomics Group, Bijvoet Center for Biomolecular Research and the Utrecht Institute for Pharmaceutical Sciences, Utrecht University, Utrecht,
The Netherlands.
3
The Netherlands Proteomics Center, Utrecht, The Netherlands.
4
Department of Biophysical Chemistry, Groningen Biomolecular Sciences and
Biotechnology Institute, University of Groningen, Groningen, The Netherlands.
5
ChELSI Institute, Department of Chemical and Process Engineering, University
of Sheffield, Sheffield, UK.
6
Howard Hughes Medical Institute, University of California, Berkeley, Berkeley, California, USA.
7
Department of Molecular and Cell
Biology, University of California, Berkeley, Berkeley, California.
8
Physical Biosciences Division, Lawrence Berkeley National Laboratory, Berkeley, California, USA.
9
Physikalische Biologie, Heinrich-Heine-Universität Düsseldorf, Düsseldorf, Germany.
10
Present addresses: Department of Cell and Molecular Biology, Uppsala
University, Uppsala, Sweden (M.L.); MRC Clinical Sciences Centre, Imperial College London, London, UK (A.P.L.S.).
11
These authors contributed equally to this work.
Correspondence should be addressed to J.v.d.O. (john.vanderoost@wur.nl).
Received 6 May 2010; accepted 24 January 2011; published online 3 April 2011; doi:10.1038/nsmb.2019
Structural basis for CRISPR RNA-guided DNA
recognition by Cascade
Matthijs M Jore
1,11
, Magnus Lundgren
1,10,11
, Esther van Duijn
2,3,11
, Jelle B Bultema
4,11
, Edze R Westra
1
,
Sakharam P Waghmare
5
, Blake Wiedenheft
6–8
, Ümit Pul
9
, Reinhild Wurm
9
, Rolf Wagner
9
, Marieke R Beijer
1
,
Arjan Barendregt
2,3
, Kaihong Zhou
6–8
, Ambrosius P L Snijders
5,10
, Mark J Dickman
5
, Jennifer A Doudna
6–8
,
Egbert J Boekema
4
, Albert J R Heck
2,3
, John van der Oost
1
& Stan J J Brouns
1
The CRISPR (clustered regularly interspaced short palindromic repeats) immune system in prokaryotes uses small guide RNAs to
neutralize invading viruses and plasmids. In Escherichia coli, immunity depends on a ribonucleoprotein complex called Cascade.
Here we present the composition and low-resolution structure of Cascade and show how it recognizes double-stranded DNA
(dsDNA) targets in a sequence-specific manner. Cascade is a 405-kDa complex comprising five functionally essential CRISPR-
associated (Cas) proteins (CasA
1
B
2
C
6
D
1
E
1
) and a 61-nucleotide CRISPR RNA (crRNA) with 5′-hydroxyl and 2′,3′-cyclic phosphate
termini. The crRNA guides Cascade to dsDNA target sequences by forming base pairs with the complementary DNA strand while
displacing the noncomplementary strand to form an R-loop. Cascade recognizes target DNA without consuming ATP, which
suggests that continuous invader DNA surveillance takes place without energy investment. The structure of Cascade shows an
unusual seahorse shape that undergoes conformational changes when it binds target DNA.
© 2011 Nature America, Inc. All rights reserved.
530 VOLUME 18 NUMBER 5 MAY 2011 nAture structur Al & moleculAr biology
A r t i c l e s
to explain how Cas protein complexes recognize their target DNA
molecules. Here we show that the molecular basis of the specificity
of CRISPR immune system relies on the formation of R-loops: that is,
the formation of Watson-Crick base pairs between the crRNA-spacer
sequence and the complementary strand of double-stranded target
DNA. We also present a structural model of Cascade that provides
insight into the architecture of Cas protein complexes.
RESULTS
Core subcomplexes of Cascade
The E. coli K12 CRISPR-Cas system (Cse-subtype) consists of a gene
cluster that includes cas3, the Cascade genes (casABCDE), cas1, cas2
and a downstream CRISPR locus
20
(Fig. 1a). To investigate the role of
the individual Cascade subunits, we first tested whether each subunit
was required for antiviral defense. Viral plaque assays with Cas3 and
Cascade lacking one type of protein subunit showed that all protein
components of Cascade are indispensible for the virus-resistant pheno-
type of E. coli (Supplementary Fig. 1).
We systematically overproduced and purified different combinations
of Cascade subunits and checked for the presence of mature crRNA.
We found that CasA or CasAB could be omitted without affecting the
apparent stoichiometry of the remaining subunits or the mature crRNA
(Fig. 1b,c). Cascade, unlike CasBCDE and CasCDE, always copurified
with large nucleic acid molecules (>300 nucleotides; Fig. 1c). Removal
of the Cas proteins followed by nuclease treat-
ments showed that RNase A only hydrolyzed
the crRNA, whereas DNase I removed the long
nucleic acids, thereby identifying the copurified
nucleic acid as DNA (Fig. 1d). Size exclusion
chromatography of the three types of complex
revealed that the majority of CasBCDE and
CasCDE were present in a single form, whereas
Cascade showed a substantial void peak in
addition to a discrete peak at ~11 ml (Fig. 1e).
DNase I treatment before gel filtration elimi-
nated the void peak without disrupting the
discrete Cascade peak, again indicating the
presence of Cascade-bound DNA (Fig. 1e).
Architecture of crRNA
The characteristics of the mature crRNA species
have been accurately determined by subjecting
mature crRNAs isolated from Cascade to dena-
turing RNA chromatography
25,26
and electro-
spray ionization mass spectrometry (ESI-MS).
To simplify the analysis, we obtained a uniform crRNA preparation by
coexpressing Cascade with a designed CRISPR containing eight repeats
and seven identical spacers (denoted R44 CRISPR (Supplementary
Fig. 2)). This setup resulted in a Cascade preparation in which each
molecule was loaded with the same crRNA. We used chromatography
to show the purity and homogeneity of the crRNA in Cascade (Fig. 2a),
and the retention time was consistent with an approximate length of
60 nucleotides. The ESI-MS spectra indicated that the crRNA had a
molecular weight of 19,660.80 Da (Fig. 2b), which corresponds well to
an expected molecular weight of 19,660.82 Da for a 61-nucleotide crRNA
resulting from single CasE endoRNase cleavage events in each repeat.
ESI-MS/MS analysis of crRNA treated with RNase T1 and RNase A
resulted in a number of oligoribonucleotide digests that could be
assigned to the mature crRNA sequence (Supplementary Fig. 3)
and were consistent with a cleavage site 5′ of the terminal base of the
hairpin
20
. The molecular weight analysis of the crRNA indicated a 5′-
hydroxyl group and a 2′,3′-cyclic phosphate terminus. We confirmed
the latter by acid treatment of the crRNA, which showed a mass shift
of 18 Da, corresponding to the hydrolysis of the cyclic phosphate to a
2′ or 3′ phosphate (Fig. 2c). Mature crRNA is 61 nucleotides long and
contains the 32-nucleotide spacer sequence, flanked by repeat-derived
sequences on either end: 8 bases at the 5′ terminus (5′ handle) and 21
bases forming a hairpin with a tetranucleotide loop at the 3′ terminus
(3′ handle; Fig. 2d).
0 2 4 6 8 10
Elution volume (ml)
12 14
1
1 2 12
1
0
100
CasA
300
150
80
50
crRNA
CasC*
CasC
CasD
CasE
CasB*
M
a e
b c d
cas3 casA casB casC
Cascade
CRISPR
casD cas1 cas2casE
M
Cascade
CasBCDE
CasCDE
Cascade
Cascade
CasBCDE
+ RNase
A
+ DNase l
CasCDE
75
50
25
20
15
37
CasCDE
CasBCDE
Cascade
Cascade +
DNase l
1
0
Normalized absorbance (A
280
)
1
0
0
Figure 1 Core complexes of Cascade retain
crRNA. (a) Schematic diagram of the CRISPR-
Cas locus in E. coli K12 containing cas3 (ygcB),
casA (cse1, ygcL), casB (cse2, ygcK), casC (cse4,
ygcJ), casD (cas5e, ygcI), casE (cse3, ygcH),
cas1 (ygbT) and cas2 (ygbF)
9,10
. (b) Coomassie
blue–stained SDS-polyacrylamide gel showing
StrepTactin-purified Cascade, CasBCDE and
CasCDE. Protein marker sizes in kDa. Asterisk,
Strep-tagged subunits. (c) Ethidium bromide–
stained denaturing PAA-gel showing nucleic acids
isolated from purified Cascade (sub)complexes.
RNA marker sizes in nucleotides. (d) RNase A
or DNase I treatment of Cascade-bound nucleic
acids. (e) Size exclusion elution profiles of
CasCDE, CasBCDE and Cascade before and after
DNase I treatment.
200
a c
d
b
Absorbance (mV)
0
100
–28
–27
–26
–25
–24
–23
–22
–21
–21
–20
–19
–18
–17
–16
–15
1,091.269
Relative intensity (%)
0
700 800 900 1,000 1,100 1,200 m/z
m/z
0 10
Retention time (min)
20
1,091
5′ handle
5′OH
2′ , 3′P
3′ handle
Spacer
1,091.2
1,092.3
crRNA
Acid-treated crRNA
1,093.5
1,093.4
1,092.2
Observed M
w
= 19,660.80 Da
Observed M
w
= 19,678.81 Da
Theoretical M
w
= 19,678.83 Da
Theoretical M
w
= 19,660.82 Da
crRNA
1,092 1,093 1,094 1,095 1,096
Figure 2 Architecture of crRNA. (a) Ion-pair reversed-phase HPLC purification of mature R44
crRNA at 75 °C. (b) Multiple-charged ESI-MS spectrum of the purified mature crRNA. (c) Enhanced
view of the −18 charged species before (top) and after (bottom) acid treatment indicating hydrolysis
of the 2′,3′-cyclic phosphate. (d) Diagram of mature crRNA derived from the R44 CRISPR.
© 2011 Nature America, Inc. All rights reserved.
nAture structurAl & moleculAr biology VOLUME 18 NUMBER 5 MAY 2011 531
A r t i c l e s
Target DNA recognition
The observation that DNA copurified with Cascade prompted us
to analyze the DNA-binding behavior of Cascade in detail (Fig. 3).
Electrophoretic mobility shift assays (EMSAs) showed that Cascade
could bind single-stranded (ss) DNA containing a sequence complemen-
tary to the spacer sequence of the crRNA (Fig. 3a,c and Supplementary
Fig. 4). Cascade also bound double-stranded target DNA (Fig. 3b,d)
without the need for, or enhancement by, additional cofactors such as
divalent metal ions or ATP (data not shown). The dissociation constants
(K
d
) of Cascade for single- and double-stranded target DNA were 8 and
790 nM, respectively. Cascade also bound weakly to nontarget DNA
(Fig. 3c,d). Competitor DNA had little effect on preformed Cascade–
target DNA complexes, indicating that the interaction between Cascade
and complementary DNA substrates is stable. At high competitor con-
centrations a proportion of Cascade lacking CasA bound target DNA,
as was evident from the faster migration rate of the CasBCDE–DNA
complex (Fig. 3a,c and Supplementary Fig. 4).
Cascade subcomplexes that lacked CasA (CasBCDE and CasCDE)
showed only sequence-specific binding to ssDNA and dsDNA targets,
and were not affected by the addition of competitor DNA (Fig. 3e–h).
The addition of purified CasA to CasBCDE preparations restored
Cascade-like nonspecific DNA binding (Supplementary Fig. 5), but
neither CasA alone (Supplementary Fig. 5)
nor a combination of CasA and CasCDE (not
shown) showed nonspecific DNA binding.
Competition assays between Cascade and
CasBCDE showed that a ratio of 1:25 was
required to distribute the ssDNA target
equally between the types of complexes
(Fig. 4a). This difference was more pro-
nounced when we used dsDNA target probes
(data not shown). Furthermore, much less
dsDNA target was shifted by CasBCDE than
by Cascade under equivalent conditions
(Fig. 3d,f), indicating again that CasA
enhanced target DNA localization.
It seems that Cascade recognizes 32-
nucleotide target sequences (referred to
as protospacers) in dsDNA molecules by
formation of base pairs between the crRNA spacer sequence and the
complementary DNA strand. Analysis of long dsDNA targets (proto-
spacer with 27-base-pair (bp) flanks) showed that both strands shifted
(Fig. 4b), probably owing to base-paired flanking regions. Only the
complementary strand shifted when we used short dsDNA targets
(corresponding to the protospacer), which suggests that the formation
of base pairs between crRNA and the complementary strand displaced
the noncomplementary strand (Fig. 4c).
We demonstrated that the noncomplementary strand had been dis-
placed by performing enzymatic and chemical footprint analyses specific
for ssDNA
27
using Cascade loaded with targeting (R44) and nontarget-
ing (K12) crRNA. Endonuclease P1 footprints showed that an 18-base
region (G27–C44) of the nontargeted strand corresponding to about half
of the protospacer was susceptible to cleavage after binding by Cascade
(Fig. 4d). In line with this observation, six thymines and one adenine
in the same region of the nontargeted strand were also sensitive to per-
manganate modification and subsequent piperidine cleavage (Fig. 4e).
These results indicate that only the first half of the nontargeted strand
of the protospacer was exposed in the Cascade–DNA complex and the
second half of the protospacer was shielded. The targeted strand was
sensitive to P1 in a region that overlapped the protospacer adjacent
motif (PAM)
28
(Fig. 4d), which may be indicative of a distortion in
Targeting Cascade
and ssDNA
a
c
e
g
i j
h
f
d
b
– –
– – – –
– –
Competitor
concentration
Cascade–
DNA complex
CasBCDE–
DNA complex
Unbound DNA
Competitor
concentration
Cascade–
DNA complex
CasBCDE–
DNA complex
Cascade
concentration
Cascade–
RNA complex
CasBCDE–
DNA complex
CasCDE–
DNA complex
CasCDE
Cascade
CasBCDE
Cascade
Cascade
Unbound DNA
Unbound DNA
Unbound RNA
Unbound DNA
Nontargeting
Cascade and ssDNA
Target ssDNA Nontarget ssDNA Target dsDNA Nontarget dsDNA
Target ssRNANontarget ssRNA
Target dsRNANontarget dsRNA
Targeting Cascade
and dsDNA
Nontargeting
Cascade and dsDNA
Figure 3 Target recognition by Cascade.
(a,b) Effect of the type of crRNA bound.
Cascade was loaded with either targeting
crRNA (R44 CRISPR, Supplementary Fig. 2)
or nontargeting crRNA (K12 CRISPR).
The binding of these two types of Cascade
complex to one type of probe is shown. DNA
probes are 86-nucleotide ssDNA or dsDNA
sequences containing the R44 protospacer
(32 nucleotides) flanked by 27 nucleotides
on either end. (c–h) Effect of uniform crRNA-
loaded (sub)complexes (R44 CRISPR) on the
binding of single- or double-stranded target
and nontarget DNA. Nontarget DNA probes
contain a scrambled R44 protospacer sequence.
(i,j) Effect of uniform crRNA-loaded Cascade
(R44 CRISPR) on the binding of target and
nontarget ssRNA and dsRNA. (a–h) Labeled
probe concentration 1 nM; DNA competitor
concentration 2,500, 500, 50, 5 and
0.5 ng µl
−1
(highest concentration not used for
CasCDE); protein concentration 200–300 nM
except in i,j where the Cascade concentrations
were 200, 100, 50, 25 and 12.5 nM.
© 2011 Nature America, Inc. All rights reserved.
532 VOLUME 18 NUMBER 5 MAY 2011 nAture structur Al & moleculAr biology
A r t i c l e s
DNA conformation resulting from Cascade-induced strand opening
further downstream. Permanganate assays also indicated that a single
base of the targeted strand of the protospacer (T33) was sensitive to
modification (Fig. 4e), suggesting that this base is unpaired in the
hybrid crRNA–DNA duplex. Exonuclease III footprints showed that
the sequences that flank the protospacer are double-stranded and that
Cascade protects a region of around 9 bases including the PAM on one
flank and around 14 bases on the non-PAM flank (Fig. 4f).
To show that crRNA and DNA form base pairs, we used a 3-kb plasmid
containing the target sequence. Specific binding of R44-Cascade to the
supercoiled plasmid resulted in a mobility shift of the plasmid (Fig. 4g).
The retardation of the plasmid was partially relieved by proteolytic
a
Target ssDNA
without competitor
CasBCDE conc.
Cascade conc.
Cascade
conc.
Cascade–
DNA complex
Cascade-R44
– –+
– +–
– –+
– +–
– –+
– +–
– –+
– +–
– –+
– +–
– –+
– +–
Seq
Seq
Seq
Nontarget strand
G59
T21
T43
T28
T73
T20
T28
C59
G59
A28
T33
T38
T34
T32
A28
T28
C44
C59
C59
G59
G31
A28
G27
Target plasmid
3.5 kb OC
Target
strand
Exo. lll
3′
5′
Nontarget
strand
Cascade
PAM
Spacer/protospacer
Exo. lll
5′OH
5′
2′, 3′P
3′
∧
10
∧
20
∧
30
###### ## ####### # # #
######
**** ** *
*
∧
40
∧
50
∧
60
∧
70
∧
80
L
SC
3.0 kb
2.5 kb
M
+ ++++++ +
– –––+++ –
– ++–––– +
– +––++– +
– –––+–– +
Nontarget strandTarget strand
Seq
Target strand
Seq
Nontarget strand
Seq
Target strand
Cascade–
DNA complex
Cascade
conc.
Cascade–
DNA complex
CasBCDE–
DNA complex
Unbound DNA
Unbound DNA
Unbound DNA
Nuclease P1
KMnO
4
Exonuclease III
Target ssDNA
with competitor
Long dsDNA
complementary
strand labeled
Short dsDNA
complementary
strand labeled
Long dsDNA
noncomplementary
strand labeled
Short dsDNA
noncomplementary
strand labeled
b c
d
g h
e f
Cascade-R44
Cascade-K12
Proteinase K
RNaseH
Cascade-K12
– –
– –
– –
––
Figure 4 R-loop formation by Cascade. (a) Competition assay between R44-crRNA–loaded Cascade and CasBCDE for R44 ssDNA target. Total protein
concentration was 500 nM in each reaction, and the Cascade:CasBCDE ratio was 1:0, 100:1, 10:1, 1:1, 1:10, 1:100 and 0:1. DNA competitor
concentration was 1 µg µl
−1
. (b,c) Effect of labeling the complementary or noncomplementary strand of a target dsDNA with 27-bp flanks (b) or without
protospacer-flanking sequences (c). Cascade concentrations were 1,500, 300, 60 and 12.5 nM. (d,e) Mapping of ssDNA regions in the Cascade–target
DNA complex using nuclease P1 and KMnO
4
. Sensitive regions are indicated by dashed lines and the protospacer by a solid line according to the G+A
sequencing lanes of each strand. Cascade loaded with K12-derived crRNA was used as a control. (f) Exonuclease III mapping of accessible dsDNA
regions upstream and downstream of the Cascade–DNA complex. The borders of the Cascade-protected regions are indicated by arrows. (g) Detection of
the R-loop in a target plasmid. Agarose gel indicating the mobility of the different plasmid forms (SC, supercoiled; L, linear; OC, open circular) and the
mobility shifts caused by R44-Cascade and R44-crRNA binding. (h) Schematic diagram of the R-loop formed in crRNA-guided dsDNA recognition by
Cascade. Regions sensitive to nuclease P1 and KMnO
4
are indicated by hash and asterisk signs, respectively.
© 2011 Nature America, Inc. All rights reserved.