CRISPR Antiphage Defence Mediated by The Cyclic Nucleotide-Binding Membrane Protein Csx23

Nucleic Acids Research, 2024, 52, 2761–2775
https://doi.org/10.1093/nar/gkae167
Advance access publication date: 13 March 2024
NAR Breakthrough Article NAR Breakthrough Article
CRISPR antiphage defence mediated by the cyclic

nucleotide-binding membrane protein Csx23
Sabine Grüschow1 , Stuart McQuarrie1 , Katrin Ackermann2 , Stephen McMahon1 , Bela E. Bode2 ,
Downloaded from https://academic.oup.com/nar/article/52/6/2761/7627476 by Universidade Federal de Vicosa (UFV) user on 10 May 2024
Tracey M. Gloster 1 ,* and Malcolm F. White 1 ,*
1
Biomedical Sciences Research Complex, School of Biology, University of St Andrews, St Andrews, Fife KY16 9ST, UK
2
Biomedical Sciences Research Complex, School of Chemistry, Centre of Magnetic Resonance, University of St Andrews, St Andrews, Fife,
KY16 9ST, UK
*
To whom correspondence should be addressed. Tel: +44 1334 463432; Email: [email protected]
Correspondence may also be addressed to Tracey M. Gloster. Email: [email protected]
Abstract
CRISPR-Cas provides adaptive immunity in prokaryotes. Type III CRISPR systems detect invading RNA and activate the catalytic Cas10 subunit,
which generates a range of nucleotide second messengers to signal infection. These molecules bind and activate a diverse range of effector
proteins that provide immunity by degrading viral components and/or by disturbing key aspects of cellular metabolism to slow down viral
replication. Here, we focus on the uncharacterised effector Csx23, which is widespread in Vibrio cholerae. Csx23 provides immunity against
plasmids and phage when expressed in Escherichia coli along with its cognate type III CRISPR system. The Csx23 protein localises in the
membrane using an N-terminal transmembrane α-helical domain and has a cytoplasmic C-terminal domain that binds cyclic tetra-adenylate
(cA4 ), activating its defence function. Structural studies reveal a tetrameric structure with a novel fold that binds cA4 specifically. Using pulse
EPR, we demonstrate that cA4 binding to the cytoplasmic domain of Csx23 results in a major perturbation of the transmembrane domain,
consistent with the opening of a pore and/or disruption of membrane integrity. This work reveals a new class of cyclic nucleotide binding protein
and provides key mechanistic detail on a membrane-associated CRISPR effector.
Graphical abstract
Introduction ase domain for ssDNA degradation (2–6) and a cyclase do-
Type III (Csm and Cmr) CRISPR systems are multisubunit ri- main for generation of a cyclic oligonucleotide (cOA) sig-
bonucleoprotein complexes that are programmed by CRISPR nal (7,8). cOA molecules, which include cyclic tri-, tetra- and
RNA (crRNA) to detect invading RNA species (1). Target hexa-adenylate (cA3 , cA4 , cA6 ), act as messengers of infec-
RNA binding results in the activation of the catalytic Cas10 tion, activating ancillary effector proteins. Characterised nu-
subunit, which has two possible active sites: an HD nucle- clease effectors include the Csm6/Csx1 family ribonucleases
Received: December 1, 2023. Revised: February 21, 2024. Editorial Decision: February 22, 2024. Accepted: February 26, 2024
© The Author(s) 2024. Published by Oxford University Press on behalf of Nucleic Acids Research.
This is an Open Access article distributed under the terms of the Creative Commons Attribution License (http://creativecommons.org/licenses/by/4.0/),
which permits unrestricted reuse, distribution, and reproduction in any medium, provided the original work is properly cited.
2762 Nucleic Acids Research, 2024, Vol. 52, No. 6
(7–12), the Can1 and Can2/Card1 nucleases and the exonu- domain of Csx23 (Csx23CTD ) were obtained using the
clease NucC (13–16). The activation of these effectors can same strategy. For expression with an N-terminal His-
lead to non-specific degradation of key biomolecules, result- tag, 5 -acttccatgGACGAGATCACTGTCGTCCTG and

ing in cell dormancy or programmed cell death (17,18). Type 5 -GGAGCTCGAATTCGGATCCCT were used as primers
III CRISPR systems have been used as the basis for new diag- and the digested PCR product was inserted into NcoI/XhoI
nostic applications due to their ability to directly detect any sites of pEV5HisTEV. For expression with a C-terminal His-
desired RNA and generate an amplified cOA signal that in turn tag, 5 -actctcatatgAATGACGAGATCACTGTCGTCCTG

activates a reporter nuclease (19–21). In some situations, cellu- and 5 -atatctcgagTGCATTGGAGCCGCTCTTGG were used
lar enzymes known as ring nucleases degrade the second mes- as primers and the digested PCR product was inserted into the
sengers to revert cells to a non-infected ground state (22,23). NdeI/XhoI sites of pEV5HisTEV. Csx23 variants for protein
Viruses also encode ring nucleases to subvert type III CRISPR isolation were obtained using primer-directed mutagenesis
immunity (24). of the expression constructs described above. Expression
A wide range of putative type III CRISPR ancillary effec- constructs for EPR were obtained through successive rounds
tor proteins have been predicted from bioinformatic stud- of primer-directed mutagenesis.
ies (25,26) and these are now being characterised, revealing
new mechanisms for antiviral defence. For example, the cA4 - Constructs for plasmid challenge and phage
binding protease CalpL has been shown to specifically degrade immunity assays
an anti-sigma factor, releasing an extra-cytoplasmic function
The VmeCmr1-6 expression construct (pVmeCmr1-6) con-
(ECF)-family sigma factor to direct a transcriptional response
tained a ColE1 origin of replication and ampicillin resis-
to viral infection (27). Many of the uncharacterised effectors
tance gene, and the expression of cmr1-3 and cmr4-6 were
appear to be membrane associated (25,26). For example, in
driven by individual T7 promoters. The cmr2 (cas10) gene
the Bacteroidetes, an effector related to the magnesium trans-
included a sequence encoding a TEV-cleavable, N-terminal
porter CorA can provide immunity when activated by a novel
His8 -tag. The construction of this plasmid has been previ-
signalling molecule synthesised by the conjugation of ATP to
ously described (21). Production of crRNA was achieved by
S-adenosyl methionine (28).
placing a minimal V. metoecus CRISPR array and V. metoe-
Previously, we described a prophage-encoded type III-B
cus cas6f into pCDFDuet™-1 (Novagen, Merck Millipore) as
(Cmr) CRISPR system from Vibrio metoecus, VmeCmr (21).
previously described (21). The CRISPR array contained two
The VmeCmr Cas10 subunit lacks an HD nuclease domain
repeat sequences flanking two oppositely directed BpiI recog-
and thus relies on cyclic nucleotide signalling for its function.
nition sites (5 -gtgtcttcgtaccttgaagacca) to allow later insertion
On activation by cognate target RNA, VmeCmr generates pre-
of the target/spacer sequence of choice. Spacer sequences were
dominantly cA3 on specific target RNA binding, resulting in
obtained as synthetic oligonucleotides with a 5 -overhang se-
the activation of the NucC effector nuclease for non-specific
quence of 5 -GAAA for the sense strand and 5 -GAAC for the
dsDNA degradation (16). This phage-encoded system is a hy-
antisense strand. After the two strands were annealed, they
brid, using a type I-F Cas6 enzyme and associated CRISPR
were ligated into BpiI-digested pCDFDuet-derivative. vmeRe-
array for crRNA generation and a NucC or Csx23 effector
peat and spacer sequences are listed in Supplementary Table
(29). Here, we demonstrate that Csx23 is a tetrameric mem-
S1. The pCDF derivatives are named according to the gene
brane protein with a novel cytoplasmic cA4 recognition do-
targeted by the spacer sequence as pCRISPRTarget .
main. Csx23 is activated by cA4 to prevent successful phage
Effector proteins were cloned into pRATDuet (31). This
infection, most likely by disruption of the host membrane
plasmid carries the pRSF1030 replicon from RSFDuet™ (No-
integrity.
vagen, Merck Millipore), araC and the araBAD promoter
from pBAD/His (Invitrogen), the tetracycline resistance gene
from pACE2 (Geneva Biotech, Genève, CH) and the two
Materials and methods MCSs from pACYCDuet™-1 (Novagen, Merck Millipore).
Sub-cloning and site-directed mutagenesis Csx23 was inserted into MCS-1 of pRATDuet (NcoI/SalI)
from the NcoI/XhoI-digested gBlock™. NucC was inserted
Enzymes were purchased from Thermo Scientific or New
into MCS-2 of pRATDuet (NdeI/XhoI) from the NdeI/XhoI-
England Biolabs and used according to manufacturer’s
digested gBlock™ and subsequently two nucleotides between
instructions. Oligonucleotides and synthetic genes were ob-
the RBS and translational start codon were removed by
tained from Integrated DNA Technologies (IDT, Coralville,
primer-directed mutagenesis to reduce the amount of NucC
Iowa, USA). Synthetic genes were codon-optimized for E.
being produced.
coli expression and restriction sites for cloning incorporated
Csx23 variants used in plasmid challenge assays were con-
where necessary. All final constructs were verified by se-
structed by single or successive rounds of primer-directed
quencing (GATC Biotech, Eurofins Genomics, DE). Vibrio
mutagenesis using the pRATDuet construct as template and
cholerae HE-45 csx23 (Csx23) was obtained as a G-Block
the same primers as used for mutagenesis of expression
with flanking restriction sites for cloning. After digestion
constructs.
with NcoI and BamHI, csx23 was ligated into linearised
pEV5HisTEV (30). The expression construct for full length
Csx23 with a non-cleavable C-terminal His-tag was ob- Protein production and purification
tained by PCR-amplifying Csx23 from the above construct E. coli C43(DE3) was used as the expression host for full
using primers 5 -ATGGCACATATGAATACTTTCAAGCG length Csx23 and Csx23CTD . Overnight cultures were diluted
and 5 -atatctcgagTGCATTGGAGCCGCTCTTGG. After 100-fold into LB containing 50 μg ml−1 kanamycin, and in-
digestion with NdeI and XhoI, the gene was ligated into cubated at 37◦ C, 220 rpm until the OD600 reached 0.6–0.8.
pEV5HisTEV. Expression constructs for the C-terminal After induction with 200 μM IPTG, incubation was contin-
Nucleic Acids Research, 2024, Vol. 52, No. 6 2763
ued at 37◦ C for 4 h. Cells were harvested by centrifugation vitrogen) to 0.005% final concentration. Samples were im-
and pellets stored at -20◦ C. mediately loaded onto a pre-run 6% acrylamide gel (29:1
For Csx23CTD , cells were resuspended in lysis buffer (50 acrylamide:bis-acrylamide). Radiolabelled material was sep-
mM Tris–HCl, 500 mM NaCl, 20 mM imidazole, 10% glyc- arated for 2 h at 200 V in 1× TBE buffer and visualised by
erol, pH 7.6) and lysed by sonication. The lysate was cleared exposure to a phosphor storage screen (Cytiva).
by ultracentrifugation (40 000 rpm, 40 min, 4◦ C) and loaded
onto a HisTrap FF 5 ml column (Cytiva), washed with 5 CVs Protein crosslinking
lysis buffer, then with 6 CVs 4% elution buffer (50 mM Tris, Csx23 was dialysed against 20 mM HEPES, 150 mM NaCl,
0.5 M NaCl, 0.5 M imidazole, 10% glycerol, pH 7.6) and 0.1% DDM if required, 10% glycerol, pH 7.6 for 2 h at room
protein was eluted in a gradient over 10 CVs leading to 100% temperature. The reaction was carried out in dialysis buffer
elution buffer. Csx23CTD -containing fractions were dialysed at with 40 μM Csx23 in the absence or presence of 80 μM
room temperature overnight in the presence of TEV protease cOA and 0 – 3.3 mM BS-3 (bis(sulfosuccinimidyl)suberate-
against lysis buffer. The protein solution was passed through d0 , Thermo Scientific) crosslinker and allowed to proceed with
the HisTrapp FF column a second time, and the flow-through gentle agitation for 30 min at 20◦ C. SDS-PAGE sample load-
was concentrated using an Amicon Ultracentrifugal filter (3 ing buffer was added to each reaction and the samples were
kDa MWCO, Merck-Millipore) and further purified by gel heated to 95◦ C for 2 min before analysis by SDS-PAGE.
filtration (HiLoad™ 16/600 Superdex™ 200 gp, Cytiva) us-
ing 20 mM Tris–HCl, 250 mM NaCl, 10% glycerol, pH 7.5
as mobile phase. Csx23CTD -containing fractions were pooled
Plasmid challenge assay
and concentrated. E. coli BL21 Star™ (DE3) (Invitrogen) was transformed with
For full length Csx23, cells were resuspended in lysis buffer pVmeCmr1-6 and pCRISPRTetR (tetracycline resistance gene-
containing 1% (w/v) DDM (n-dodecyl-beta-maltoside) and targeting crRNA) or pCRISPRpUC MCS (pUC19 MCS-targeting
lysed by sonication. After stirring at room temperature for 1 crRNA, non-targeting control). Competent cells were pre-
h, the lysate was cleared by ultracentrifugation (40000 rpm, pared from fresh individual transformants as follows: a cul-
40 min, 4◦ C) and loaded onto a HisTrap FF 5 ml column (Cy- ture obtained from 50- to 100-fold dilution of an overnight
tiva), washed with 5 CVs lysis buffer containing 0.1% DDM, culture was grown at 37◦ C, 220 rpm to an OD600 of 0.4–0.6.
then with 7 CVs 4% elution buffer containing 0.1% DDM. All subsequent steps were carried out at 4◦ C with pre-chilled
The protein was eluted in a 10 CV gradient to 100% elu- buffers. Cells were collected by centrifugation (10 min, 3000
tion buffer containing 0.1% DDM. Csx23-containing frac- rpm, Eppendorf centrifuge 5810) and the pellet resuspended
tions were concentrated using an Amicon Ultracentrifugal in an equal volume of 60 mM CaCl2 , 25 mM K-MES, pH 5.8,
filter (30 kDa MWCO, Merck-Millipore) and further pu- 5 mM MgCl2 , 5 mM MnCl2 . After incubation on ice for 1
rified by gel filtration (HiLoad™ 16/600 Superdex™ 200 h, cells were harvested again and resuspended in 1/10th vol-
gp, Cytiva) using 20 mM Tris–HCl, 250 mM NaCl, 10% ume of the same buffer containing 10% glycerol. Single-use
glycerol, 0.06% DDM, pH 7.5 as mobile phase. Csx23- aliquots were stored at –70◦ C. For the plasmid challenge as-
containing fractions were pooled and concentrated. Csx23 say, 50 ng (1 μl) target plasmid containing the effector was
variants were purified as described for wild type Csx23. The added to 50 μl of competent cells. After incubation on ice for
proteins were stored at 4◦ C for short term storage or flash- 30 min, cells were heat-shocked at 42◦ C for 1 min, then placed
frozen as single-use aliquots and stored at –80◦ C. The con- on ice for 2 min. After the addition of 0.5 ml LB medium, the
centration of full length Csx23 and Csx23CTD was deter- transformation mixture was incubated for 2–2.5 h at 37◦ C,
mined spectrophotometrically by absorbance at 280 nm and 220 rpm. A 10-fold serial dilution (2.5 μl/spot) was applied
using the calculated extinction coefficients of 9970 and 1490 onto selective LB agar plates containing 0.1% d-lactose and
M−1 cm−1 , respectively. Concentrations are stated for the 0.2% l-arabinose for induction of the Cmr proteins and ef-
monomer. fector, respectively, and the plates were incubated at 37◦ C for
20 h.
Dynamic light scattering
Phage immunity assay
DLS measurements were performed with the Zetasizer Nano
E. coli BL21 Star™ (DE3) (Invitrogen) was co-transformed
S90 (Malvern) instrument. Samples contained 60–100 μM
with three plasmids: one encoding the VmeCmr complex
protein in 20 mM Tris–HCl, 75 mM NaCl, 10 mM MgCl2 and
(pVmeCmr1-6), one encoding the crRNA (pCRISPRLpa for
0.1% DDM as required. After centrifugation at 12 000 × g for
phage P1 lpa gene-targeting crRNA or pCRISPRpUC MCS as
10 min at room temperature, the sample was filtered (0.22 μm
non-targeting control), and one encoding the effector (pRAT-
PES membrane) and loaded into a quartz cuvette (ZMV1012).
Csx23, pRAT-NucC or pRATDuet with no insert). Single
Measurements were performed at 25◦ C with three measure-
colonies were grown overnight in LB medium containing 50
ments of thirteen runs.
μg ml−1 ampicillin, 25 μg ml−1 spectinomycin and 6.5 μg
ml−1 tetracycline at 25 – 28◦ C with shaking. The overnight
Electrophoretic mobility shift assay cultures were either used for the phage immunity assay (see
[α-32 P]-Radiolabelled cA3 and cA4 were prepared using the below) or aliquots (0.3 ml) were collected, mixed with 50%
VmeCmr or SsoCsm complex, respectively, as previously de- glycerol and stored at –70◦ C for future use. Aliquots were re-
scribed (21,22). A typical EMSA assay contained 50 nM [α- vived by inoculating them into 5 ml LB containing the appro-
32
P]-cA4 and 0–1600 nM Csx23 in 20 mM Tris, pH 8.0, priate antibiotics followed by a 2–3 h outgrowth period at
100 mM NaCl, 10% glycerol, 0.5 mM TCEP. The mix- 37◦ C with shaking.
ture was incubated at room temperature for 10 min before The relationship between OD600 and cell density for BL21
addition of G-250 Native Gel Sample Loading Buffer (In- Star carrying the three plasmids was determined by two inde-
pendently performed microdilution assays. An OD600 of 1.0 pended in dEPR-ND buffer to 4 mg ml−1 and sonicated for 5
yielded around 2 × 108 cfu ml−1 for E. coli BL21 Star™ (DE3) min. The membrane scaffold protein MSP1D1 (32) was used
carrying the three plasmids. as a scaffolding protein as described previously (33). Briefly,
For the phage immunity assay, the OD600 for each culture MTSL-labelled Csx23 (3.6 μM), MSP1D1, and DMPC were
was determined and each culture was diluted to an OD600 of mixed at a molar ratio of 1 : 2 : 130 in 1 ml volume in
0.05 (∼1 × 107 cfu ml−1 ) with LB medium supplemented with dEPR-ND buffer and Triton X-100 was added to 1.5 mM
10 mM MgSO4 , 0.2% w/v l-arabinose and the appropriate final concentration. The mixture was allowed to stand at
antibiotics. It had been found that the presence or absence of room temperature in the dark for 30 min before the deter-
antibiotics made no difference to the assay in the first ∼ 5 h, gent was removed by successive addition of BioBeads SM-2
hence antibiotics were omitted for shorter time courses. 160 (BioRad). Nanodiscs were concentrated using Vivaspin® 500
μl E. coli was added to each well (96-well propylene clear flat- centrifugal filters (10 kDa MWCO PES membrane).
bottom plate, Greiner) alongside supplemented LB without E. PDS samples were prepared at a final monomer concen-
coli as control and background. Phage P1 was diluted from tration of 50–75 μM for micellar Csx23, or 80–100 μM for
a >1 × 1010 pfu ml−1 stock into the supplemented LB broth Csx23 reconstituted into nanodiscs, in the absence or pres-
and 40 μl was added to each well to give final MOIs quoted in ence of a two-fold molar excess of cA4 (30 min incubation
the individual experiments. Growth curves were monitored by time). 50% (v/v) deuterated glycerol (Deutero) was used for
OD600 readings at 37◦ C (FluoStar Omega plate reader, BMG cryoprotection. The samples with a final volume of 65 μl were
Labtech). The absorbance at 600 nm was recorded in matrix transferred to 3 mm quartz EPR tubes which were immedi-
scan mode (2 × 2 matrix, 2 mm scan width, 0.5 s settling time, ately frozen in liquid nitrogen.
10 flashes per well and cycle). The plate was shaken for 10 s
at 200 rpm in double orbital mode before each cycle with a Room-temperature CW EPR
typical cycle time of 15 min.
Phage titre was determined by microdilution drop assay of Room-temperature CW EPR measurements to assess labelling
culture aliquots that were collected 24 h post infection and efficiency were performed using a Bruker EMX 10/12 spec-
trometer equipped with an ELEXSYS Super Hi-Q resonator at
centrifuged for 3 min at 2260 × g. The supernatant was care-
an operating frequency of ∼9.9 GHz (X-band) with 100 kHz
fully transferred to a fresh tube, and a serial 10-fold dilution
was prepared using LB containing 10 mM MgSO4 . 2.5 μl modulation. Samples were recorded using a 120 G field sweep
phage dilution was applied onto soft agar containing 10 mM centred at 3505 G, a time constant of 20.48 ms, a conversion
time of 18.67 ms, and 1714 points resolution. An attenuation
MgSO4 and E. coli BL21 Star™ (DE3) at OD600 0.2 with a
of 20.0 dB (2 mW power) and a modulation amplitude of 0.7
layer of LB agar/MgSO4 underneath. Plates were incubated at
37◦ C overnight before counting plaques. Results were plotted G were used. Csx23 samples were measured in 20 μl capillar-
using Graphpad Prism. Mann–Whitney test (Prism 10.0) was ies at ∼30 μM monomer concentration and double integrals
used to determine statistical significance. were compared to 4-hydroxy-TEMPO (4-hydroxy-2,2,6,6-
tetramethylpiperidine 1-oxyl; Acros) as a standard. Labelling
efficiency was ∼63% for the N62R1 mutant and ∼73% for
Sample preparation for pulse dipolar electron both V52R1 and N59R1 mutants; samples showed negligible
paramagnetic resonance spectroscopy (PDS) free spin label contribution and the shape of the spectra sug-
gested low mobility of the label (Supplementary Figure S9).
For preparation of PDS samples, native cysteine residues were
mutated as follows: C28A, C85A, C105L, C141A. The result-
ing Csx23 AALA mutant was used for site-specific mutagene- Pulse dipolar EPR spectroscopy (PDS)
sis and site-directed spin labelling, yielding the three Csx23 PDS experiments were performed on a Bruker ELEXSYS E580
AALA constructs V52C, N59C and N62C, with the intro- spectrometer with an overcoupled 3 mm cylindrical resonator
duced cysteines located in the transmembrane region. All mu- (ER 5106QT-2w), operating at Q-band frequency (34 GHz).
tants were tested for activity by the plasmid challenge assay Pulses were amplified by a pulse travelling wave tube (TWT)
described above and purified in the same manner as wild type amplifier (Applied Systems Engineering) with nominal output
Csx23. Protein samples were reduced with DTT (5 mM) for of 150 W. Temperature was controlled using a cryogen-free
2 h at 4◦ C. DTT was removed by passing the sample through variable temperature cryostat (Cryogenic Ltd) operating in the
a PD MiniTrap G-25 column (Cytiva) pre-equilibrated with 3.5–300 K temperature range.
20 mM Tris, 200 mM NaCl, pH 8.0, 0.06% DDM in Pulse electron-electron double resonance (PELDOR) exper-
D2 O. Csx23 variants were labelled with a 10-fold mo- iments were performed at 50 K with the 4-pulse DEER (34,35)
lar excess of MTSL ((1-oxyl-2,2,5,5-tetramethyl-3-pyrroline- pulse sequence (π/2(νA ) – τ1 – π(νA ) – (τ1 + t) – π(νB ) – (τ2
3-methyl)methanethiosulfonate; Santa Cruz Biotechnology) - t) – π(νA ) – τ2 – echo) as described previously (36), with a
overnight at 4◦ C. The label was removed in the same manner frequency offset (pump – detection frequency) of + 80 MHz
as DTT. MTSL-labelled protein was concentrated (Vivaspin® (∼3 mT). Shot repetition times (SRT) were set to 1.5 ms; τ1
500 centrifugal filter, 10 kDa MWCO PES membrane, Sarto- was set to 380 ns, and τ2 was set to 4000 ns for the samples
rius). Labelling efficiencies for each Csx23 mutant were as- in detergent and to 2700 ns for those reconstituted into nan-
sessed by continuous wave (CW) EPR measurements. odiscs. The echo decays as function of available dipolar evolu-
For reconstitution into nanodiscs, deuterated 20 mM tion time were assessed from refocused echo decays by incre-
bisTris, 200 mM NaCl, pH 7.0 (dEPR-ND buffer) was menting τ2 in the 4 pulse DEER sequence from a start value of
used throughout. DMPC (1,2-myristoyl-sn-glycero-3- 760 ns and omitting the νB inversion pulse. Pulse lengths were
phosphocholine, Avanti Lipids, Inc.) was dried under nitrogen 16 and 32 ns for π/2 and π detection. Measurements were
to a thin film from a chloroform solution. Residual solvent performed with a reduced inversion efficiency of the pump
was removed by high vacuum. The DMPC film was resus- pulse by approximately 50% to minimise multispin effects, us-
ing rectangular pulses from an arbitrary waveform generator olution. Data were automatically processed using autoPROC
(AWG, Bruker) with a 12 ns ELDOR pump pulse width and (51) and STARANISO (https://staraniso.globalphasing.org/
a 16-step phase cycle (37–40). The pump pulse was placed on cgi-bin/staraniso.cgi). The data were phased by PhaserMR
the resonance frequency of the resonator and applied to the (52) in the CCP4 suite (53) using a model generated by Al-
maximum of the nitroxide field-swept spectrum. An 8-step nu- phaFold2 (54) implemented in Colab, with initial B-factors
clear modulation averaging with 16 ns increments was used modelled in Phenix (55). Model refinement was achieved by
for all experiments. Experiments ran for typically 3–4 h (mi- iterative cycles of REFMAC5 (56) with manual model manip-
cellar Csx23 samples) or 24 h (nanodisc samples). ulation in COOT (57). Electron density for cA4 was clearly
PDS experiments were analyzed using DeerAnalysis2022 visible in the maximum likelihood/σA weighted Fobs – Fcalc
(41). PDS data were first background-corrected using a 3- electron density map at 3σ. The coordinates for cA4 were
dimensional homogeneous background function and ghost generated in ChemDraw (Perkin Elmer) and the library was
suppression (power-scaling) for a four-spin system (42), be- generated using acedrg (58), before fitting of the molecule in
fore Tikhonov regularization followed by statistical analysis COOT. Model quality was monitored throughout using Mol-
using the validation tool in DeerAnalysis2022, varying back- probity (59) (score 1.48; centile 94). Ramachandran statis-
ground start from 5 to 80% of the trace length in 16 trials. tics are 96.3% favoured, 0% disallowed. Data and refinement
Resulting background start time for the best fit was then used statistics are shown in Supplementary Table S3. The coordi-
as starting point for a second round of Tikhonov regulariza- nates and data have been deposited in the Protein Data Bank
tion followed by a second round of statistical analysis, this with deposition code 8QJK.
time including the addition of 50% random noise in 50 trials,
resulting in a total of 800 trials. The regularization parameter
α was chosen according to the GCV (43) or L-curve corner
criterion (44) and the goodness-of-fit.
Results
For comparison, raw PDS data were subjected to the Com- Csx23 functions as a type III CRISPR immune
parativeDEERAnalyzer (CDA) version 2.0 within DeerAnal- effector in vivo
ysis2022 (DEERNet Spinach SVN Rev 5662 (45) and Deer- Some Vibrio genomes host prophage-encoded type III CRISPR
Lab 0.9.1 (46) Tikhonov regularization) for user-independent systems with an uncharacterised gene known as csx23 (Fig-
data processing and analysis, in line with current recom- ure 1A) (29). The Csx23 family of proteins are commonly
mendations (47). CDA reports are provided as shown in found in Vibrio species, where they appear as an alterna-
Supplementary Table S2. The EPR research data underpinning tive to NucC, and have a sporadic distribution associated
this publication can be accessed at https://doi.org/10.17630/ with type III CRISPR systems in other bacterial phyla (26)
e8334069- fc1a- 4329- a07f- 9d908515b7c0. (Supplementary Figure S1).
We used AlphaFold 2 (AF2) (54), implemented on the Co-
Modelling for PDS measurements labfold server (60), to predict the structure of Csx23 from
Distance distributions were modelled based on the Al- V. cholerae HE-45. This yielded a model with a 68 residue
phaFold2 structure obtained for the Csx23 full-length N-terminal α-helical domain, predicted by InterPro (61) to
tetramer (PDB rank_1_model_3_ptm_seed_0_relaxed pro- be membrane spanning, and a 91 residue C-terminal domain
vided in the underpinning data). R1 moieties were introduced (CTD) predicted to be a soluble, cytoplasmic domain (Figure
at residues 52, 59, or 62 of all four chains of the tetramer us- 1B). The AF2 prediction was less confident for the arrange-
ing both mtsslWizard (48) within the mtsslSuite server-based ment of the domains relative to each other (Supplementary
modelling software (49) and ‘tight’ settings, and Multiscale Figure S2A). We reasoned that the CTD would likely be the
Modeling of Macromolecules (MMM) (50) using ‘ambient cOA-binding domain. The AF2 and InterPro predictions were
temperature’ settings. Cartoon structural representations of further strengthened by analysis of the Csx23 sequence us-
spin-labelled Csx23 constructs were generated using Pymol ing DeepTMHMM (62), which strongly predicted the pres-
(Schrödinger Inc.). ence of two transmembrane helices in the N-terminal domain
(Supplementary Figure S2B).
As a first step to elucidate whether Csx23 was indeed a
Crystallisation of Csx23
cOA-dependent effector involved in adaptive immunity, we
Csx23 tetramer at 10 mg ml−1 was mixed in a 1:2 molar tested its activity in a plasmid challenge assay, making use of
ratio with cA4 and incubated at room temperature for 30 our established VmeCmr expression system (21). In the assay,
min, before centrifugation at 13 000 rpm prior to crystalli- cells expressing the VmeCmr complex are transformed with a
sation. Sitting drop vapour diffusion experiments were set up plasmid carrying an effector gene, csx23 or nucC, alongside a
at the nanoliter scale using commercially available crystallisa- target sequence for activation of the Cas10 cyclase. We used
tion screens and incubated at 293 K. Crystals used for data a portion of the tetracycline resistance gene as the target se-
collection were evident after 2 days and grew from a reser- quence and selected for tetracycline resistance after transfor-
voir solution of 42.5% PEG 400, 0.2M LiSO4 , 0.1 M sodium mation of the cells with the target/effector plasmid. For active
acetate pH 5, which also acted as an intrinsic cryoprotectant. effectors, we expected to see fewer transformants compared to
As such crystals were harvested and immediately cryo-cooled inactive effectors due to target depletion and/or programmed
prior to data collection. cell death (Figure 2A) (31).
In the absence of an effector gene, the same number of trans-
X-ray data processing, structure solution and formants was observed in the presence or absence of targeting
refinement crRNA, indicating that the VmeCmr system on its own does
X-ray data were collected at a wavelength of 0.9537 Å, on not confer any protection against invading nucleic acid, for ex-
beamline I04 at Diamond Light Source, at 100 K to 1.76 Å res- ample by Cas7-mediated target RNA knockdown (Figure 2B).
Figure 1. (A) Organisation of Vibrio type III-B CRISPR (Cmr) loci. (B) Predicted structure of V. cholerae HE-45 Csx23 (AF2) coloured by local distance
difference test (LDDT) scaled from blue (high prediction confidence) to red (low).
Figure 2. Plasmid challenge assay. (A) Schematic representation of the basis for the assay. (B) A serial dilution of the transformation mixture of E. coli
cells expressing the VmeCmr complex with plasmids carrying a target sequence and varying effector genes was spotted onto agar plates containing
antibiotics to select for all plasmids. VmeCmr complexes loaded with crRNA targeting the tetracycline resistance gene of the incoming plasmid
(targeting crRNA, VmeCmr[TetR]) or loaded with crRNA targeting a sequence that is not present in the host genome or plasmids (non-targeting crRNA,
VmeCmr[pUC]) were used.
In the presence of either NucC or Csx23, however, the number major reaction product in vitro, while cA5 and cA6 predomi-
of transformants was significantly reduced in the presence of nate in vivo (63).
targeting crRNA. Using the AF2 model as a guide, we removed
the membrane domain of Csx23 to only leave the CTD. Csx23
CTD was inactive in the plasmid challenge assay. These data Crystal structure of the tetrameric Csx23 soluble
confirmed that Csx23 can function as a Cmr-linked effector domain bound to cA4
in vivo, that Csx23 function is dependent upon the presence A range of CRISPR effector proteins bind cA4 , but they
of target and hence likely to be cOA-dependent, and that its tend to utilize a conserved CARF/SAVED/Csx3 domain
membrane domain is essential for activity. (10,12,13,15,23), which is a member of the Rossman fold su-
perfamily (64) or the unrelated Crn2/AcrIII-1 domain (24).
The cA4 binding to CTD of Csx23 was predicted to be com-
Csx23 is a cA4 -dependent, oligomeric effector pletely unrelated to either family and could thus represent a
VmeCmr generates predominantly cA3 with only a minor new class of cA4 recognition domain. To investigate cA4 bind-
amount of cA4 in vitro, and its effector nuclease NucC is ing by Csx23 at an atomic level, we co-crystallised the soluble
activated by cA3 (21). As the Csx23-associated VchCas10 was CTD of Csx23 with cA4 . Unfortunately, no crystals were ob-
99% identical at the amino acid level to VmeCas10, we rea- tained in the absence of cA4 . Diffraction data were collected
soned that Csx23 would be activated by the same cOA species on the crystals to a resolution of 1.76 Å, and the structure was
as NucC. To investigate cOA-binding by Csx23, we expressed solved using molecular replacement with the monomeric AF2
and purified two versions of the protein: a full-length (FL) ver- model as the search model.
sion, which required purification in the presence of detergent, The asymmetric unit contains one protomer of Csx23 CTD,
and a truncated version encoding only the soluble CTD. The comprising two beta-strands, linked via two alpha-helices
latter purified as a monomer of around 12 kDa (calculated to a third beta-strand, which together form a mixed anti-
MW 10 kDa) as determined by dynamic light scattering (DLS, parallel/parallel beta-sheet, followed by a short alpha-helix at
Supplementary Figure S3). FL Csx23 was isolated as a much the C-terminus (Supplementary Figure S5A). The crystal struc-
larger entity of around 150 kDa (calculated MW 18 kDa for ture of the protomer is consistent with the AF2 prediction, and
monomer). Membrane proteins purified in the presence of de- they superimpose with a RMSD of 0.9 Å. Interestingly, the 4-
tergent form lipid-protein conglomerates which pose prob- fold crystallographic symmetry creates a tetrameric arrange-
lems for standard methods to determine the oligomeric state ment of Csx23 CTD (Figure 4A–C; Supplementary Figure
of a protein such as analytical gel filtration or DLS. We there- S5B, C), which is consistent with the behaviour of the pro-
fore explored the quaternary structure of the Csx23 protein tein in the presence of cA4 observed by cross-linking. Electron
by cross-linking with bis(sulfosuccinimidyl)suberate (BS3) fol- density in the Fobs – Fcalc map clearly showed the presence of
lowed by SDS-PAGE analysis. The FL Csx23 protein could a molecule of cA4 bound to Csx23 CTD. The cA4 molecule,
be cross-linked to form dimers, trimers and tetramers (Fig- and a coordinating sodium ion, are positioned at the centre
ure 3A). At the highest concentration of BS3, the tetrameric of rotation for the 4-fold crystallographic symmetry, meaning
species was by far the dominant species, consistent with FL there is effectively one adenylate moiety bound to one Csx23
Csx23 existing as a tetramer. In contrast, the CTD did not CTD protomer in the asymmetric unit; cA4 and the sodium ion
cross-link in solution, suggesting a monomeric composition were modelled with 0.25 occupancy to account for this. Fur-
(Figure 3A). ther discussion will be based on the tetrameric structure bound
We investigated the cOA-binding specificity of Csx23 by to the whole molecule of cA4 as, due to the crystallographic
electrophoretic mobility shift assay (EMSA) using radioac- symmetry, all interactions between each Csx23 CTD subunit
tively labelled cOA produced by the Sulfolobus solfataricus and adenylate in cA4 are, by definition, identical. There are
Csm complex, which predominantly produces cA4 and a small surprisingly few interactions directly between subunits in the
amount of cA5 (11) or by the VmeCmr complex that pro- Csx23 CTD tetramer; the guanidinium group of R95 forms
duces mainly cA3 with a trace of cA4 (21). Unexpectedly, both two electrostatic interactions with E97 in the neighbouring
FL and CTD Csx23 bound to cA4 but not cA3 (Figure 3B, subunit, and R110 forms two hydrogen bonds with the back-
Supplementary Figure S4), which strongly suggested cA4 as bone carbonyl groups of G107 and A108 in the neighbour-
the relevant activator for Csx23. Cross-linking of Csx23 CTD ing subunit. Given Csx23 CTD adopts a monomeric state in
in the presence of cA4 shifted the oligomeric state in SDS- the absence of cA4 , it is likely that binding to cA4 contributes
PAGE from monomer to tetramer (Figure 3A). A 10 times significantly to the tetramerization. FL Csx23 is tetrameric in
higher concentration of Csx23 CTD was required to observe the absence of cA4 , suggesting there are likely to be more pro-
cA4 binding compared to FL Csx23, suggestive of weaker ductive interactions at the interface between subunits in the
binding affinity (Figure 3B). N-terminal membrane spanning domain.
Although the specificity of Csx23 for cA4 rather than cA3 The cA4 molecule is enclosed within the Csx23 CTD
was contrary to our initial assumptions, it is frequently ob- tetramer (Figure 4A–C; Supplementary Figure S5D, E), but
served that the physiologically relevant cOA is not the most surprisingly makes few direct interactions with the protein.
abundant species observed in vitro. For example, the Csm6 R95 of Csx23 CTD makes electrostatic interactions between
ribonucleases of both Streptococcus thermophilus and My- both terminal nitrogen atoms and two different oxygen atoms
cobacterium tuberculosis are activated by cA6 , which is only in the phosphate group of cA4 (Figure 4D). R110 and F111
a minor component of the cOA mix produced in vitro (7,31). sandwich the adenine base in cA4 through π–π interactions.
Furthermore, the distribution of cOA species produced by type Both terminal nitrogen atoms of R110 also form hydrogen
III systems can differ in vivo compared to in vitro, as observed bonds with a water molecule, which in turn mediates hydro-
for the S. thermophilus Csm complex, which makes cA3 as the gen bonds with the C2-hydroxyl group of the ribose and a ni-
B
Figure 3. Csx23 is a tetrameric protein and binds cA4 . (A) Cross-linking of FL or CTD Csx23 in the presence and absence of cA4 or cA3 and increasing
amounts of BS3 cross-linking reagent. The FL protein formed tetramers in detergent regardless of cOA presence; the CTD could form tetramers in
solution only in the presence of cA4 . Blue dots show the predicted quaternary structure corresponding to each band on the SDS-PAGE. (B) EMSA
showing binding of radioactive cA4 by the FL and CTD Csx23 proteins. Protein concentrations were 25, 50, 100, 200, 400, 800 and 1600 nM for FL Csx3
and 0.4, 1, 4, 10, 40 and 100 μM for CTD Csx23; the cA4 concentration is indicated beneath the gel.
trogen atom in the adjacent adenine. Both the backbone amide with high LDDT scores and predicted a transmembrane do-
and carbonyl groups of F111 form hydrogen bonds with dif- main consisting of 8 α-helices (Figure 4E).
ferent nitrogen atoms in the adenine base. A single sodium ion The cA4 forms a ‘cup-like’ structure, with the phosphodi-
is positioned at the centre of the cA4 , which interacts with a ester backbone forming the base of the ‘cup’ which is located
water molecule, but appears to make no direct or indirect in- deep in the binding cavity of Csx23 CTD (Supplementary
teractions with cA4 or Csx23 CTD. Whilst R95 is positioned Figure S5E, F). The adenine bases are close to perpendicular
deep into the cA4 binding cavity, both R110 and F111 are to the backbone, thus forming the sides of the ‘cup’ (Figure
located on a loop (comprising residues 107–113) on the sur- 5F), and ensuring they are in a position near to the surface
face of the Csx23 CTD (Figure 4A–C). This loop must dis- and thus accessible to residues R110 and F111 in the loop
play flexibility to allow cA4 access to the binding cavity, and upon closing. The ribose moieties facilitate the formation of
then changes conformation to close over cA4 once bound. It this distorted, and presumably high energy, conformation of
is therefore likely that R110 and F111 are crucial for ‘lock- cA4 , by adopting an unusual 2 3 T twist conformation with
ing’ the cA4 into position. The dynamic movement of loops C2’-endo/C3’-exo pucker. The conformation of cA4 bound to
has been observed previously with other structurally distinct Csx23 CTD is distinct to that observed in complex with other
domains that bind cOA (10,14,24,65,66). These conforma- cA4 binding proteins such as AcrIII-1 (24), Crn3 (67), Can1
tional changes are often accompanied by movement in other (13), Can2/Card1 (14,15) and Csx1 (10). The angle between
parts of the protein to elicit allosteric regulation. We used each C2’-hydroxyl group on the ribose and adjacent oxygen
AF2 to model the tetrameric structure of FL Csx23 includ- and phosphate atoms in cA4 is 156◦ . An angle close to 180◦
ing the membrane spanning domain. This generated a model between these atoms is required for in-line nucleophilic attack
A B E
C D
Figure 4. Structure of the tetrameric Csx23 CTD bound to cA4 . Structure of tetrameric Csx23 CTD (green cartoon, with exception of flexible loop
(residues 107–113) shown in pink) in complex with cA4 (sticks coloured by element, with carbon in yellow) from (A) ‘side’, (B) ‘top’ and (C) ‘bottom’
views. The sodium ion is shown as a blue sphere. R95 (sticks coloured by element, with carbon in green) and R110 and F111 (sticks coloured by
element, with carbon in pink) are also shown. (D) Interactions formed between cA4 (yellow) and residues R95 (green), R110 and F111 (both pink) in
Csx23 CTD. Colouring as in panels A–C. Residues from just one subunit are shown for clarity. Black dotted lines represent electrostatic interactions; grey
dotted lines represent hydrogen bonds; purple dotted lines represent water-mediated hydrogen bonds. (E) AF2 model of the FL Csx23 tetramer coloured
by LDDT scaled from high (blue) to low (red) prediction confidence (top), and the same model with one subunit shown in light blue and the other three in
light green (bottom).
A B
Figure 5. Investigation of roles of conserved residues. (A) Plasmid challenge assay for wild-type and variant Csx23 proteins. The R95A and R110A
variants did not provide immunity, suggesting Csx23 function was disrupted. (B) EMSA showing that the R95A and R110A variants are unable to bind the
cA4 activator in vitro. Weak binding of cA5 was observed for the R95A variant. Each binding reaction contained a two-fold molar excess of Csx23
tetramer over cOA. The amount of [32 P]-cOA was kept constant, unlabelled cA4 was added to give final cA4 concentrations of 0.25 μM (lanes 2, 5, 8),
2.5 μM (lanes, 3, 6, 9) and 25 μM (lanes, 1, 4, 7, 10).
to break the phosphodiester bond. Therefore, it is unlikely that to form a disulphide bond in reducing conditions, although
Csx23 could support substrate-assisted ring nuclease activity, this bond was not present in the crystallised protein (sulfur
which is a feature of some self-limiting Csm6 family ribonu- atoms from each cysteine were 4.4 Å apart). All four cys-
cleases (12,63,65,66,68). Consistent with this, we observed no teines were removed by primer-directed mutagenesis to give
evidence for Csx23 ring nuclease activity in vitro (Figure 3B). Csx23 AALA. Individual Cys residues were then introduced
The structural fold observed for Csx23 CTD is novel com- into the alpha-helices in the membrane domain to give Csx23
pared to other structures reported to bind cA4 (and other AALA V52C, N59C or N62C. All variants were tested for ac-
cOAs), demonstrating both the structural and functional di- tivity using the plasmid challenge assay and showed similar
versity in the proteins that have evolved to interact with results to the WT protein (Supplementary Figure S8A). Sub-
cyclic nucleotides. A DALI search (69) to identify any struc- sequently the recombinant variants were purified in detergent
tural homologues of Csx23 CTD revealed intriguing matches (Supplementary Figure S8B) and spin-labelled; labelling effi-
to proteins containing PB1 and ubiquitin-like domains. The ciency was determined by CW-EPR (Supplementary Figure S9)
top hit, the PB1 domain of protein kinase C zeta type from The PDS data showed the presence of two distances, as ex-
rat (70), shares just 6% sequence identity with Csx23 CTD, pected for a rotationally symmetric homotetramer. The spin
but structurally overlaps with an RMSD of 3.1 Å over 66 labels form a square geometry with two equal shorter dis-
alpha-carbon atoms (Csx23 CTD comprises 88 residues; PDB tances to the√ adjacent monomer/protomer and a longer (by
4MJS; Supplementary Figure S6A, B). Superimposition of the a factor of 2) distance diagonally across the tetramer. Ex-
two structures (Supplementary Figure S6C, D) demonstrates pected distance distributions for PDS measurements can be
a strong likeness between the secondary structure elements modelled if a structure (experimental or predicted) is avail-
in the proteins, with just an additional beta-strand and loop able. Here, we have used the AF2 predicted tetramer structure
present in the PB1 domain that is absent in Csx23 CTD. of Csx23 and two different approaches for modelling of the
PB1 domains have a ubiquitin-like beta-grasp fold and are spin label rotamers; one based on energy weighted rotamers
involved in protein-protein interactions in a host of biologi- (MMM, 51) and one based on excluded volume (mtsslWizard,
cal processes (71). PB1 domains have not previously been re- 49). For the Csx23 AALA V52R1 (R1 refers to the spin label)
ported to exist in prokaryotes or viruses and the relevance of variant, two sharp and narrow distance peaks were predicted,
this structural relationship is unclear at present. with good agreement between MMM and mtsslWizard. For
comparison, experimental distance distributions for this vari-
Conserved residues important for cA4 binding ant were obtained in detergent as well as following reconstitu-
tion into nanodiscs, with the latter agreeing significantly better
Based on the crystal structure, R95, R110 and F111 were im-
with the modelled distribution (Supplementary Figure S10).
plicated in cA4 binding. R95 and R110 are conserved in Csx23
The full set of PDS data (including the raw and background-
homologues; F111 is well conserved but sometimes replaced
corrected trace of the dipolar oscillation) of the nanodisc re-
by tyrosine, which can facilitate the same π–π interactions
constituted Csx23 AALA V52R1 protein in the presence or
with the adenine base of cA4 (Supplementary Figure S7). FL
absence of cA4 are shown in Figure 6. Notably, a dramatic
csx23 mutants were first screened in the plasmid challenge as-
change in distance distributions could be observed upon addi-
say (Figure 5A). Surprisingly, the F111A mutant showed wild
tion of cA4 , suggesting a strong effect of the cyclic nucleotide
type activity, but the R95A and R110A mutants allowed plas-
on Csx23 conformation. Specifically, the disappearance of the
mid transformation, consistent with inactivation of Csx23. To
two separate distance peaks and instead the presence of a
explore this further, we purified FL Csx23 R95A and R110A
monomodal, substantially broadened distribution (note that
and tested their ability to bind cOA species (Figure 5B). As
distance probabilities in the orange coloured region of Fig-
expected, neither variant bound cA4 , although Csx23 R95A
ure 6C) are less reliable and might represent an artefact) indi-
unexpectedly showed a weak affinity for cA5 .
cates the sampling of a larger conformational ensemble. Sim-
ilar results were obtained for the other two variants, with
Structure of the transmembrane domain and its a broadening/shortening of the overall distribution for the
perturbation upon cA4 binding N59R1 construct and a broadening leading to the loss of the
Our working assumption for the activation of Csx23 was resolution of the bimodality of the distance distribution for
that binding of cA4 in the CTD results in a change in con- N62R1 (Supplementary Figures S11 and S12). These obser-
formation that is communicated to the membrane spanning vations indicate an increased conformational flexibility upon
domain. Structural studies of cOA binding CARF family ef- cA4 binding leading to structural heterogeneity in the trans-
fectors have revealed that a ‘tightening’ of the CARF domains membrane region. Note that modelled distance distributions
around the bound ligand causes allosteric changes in the as- agree best the deeper the residue sits in the membrane (very
sociated effector domains (14,65,66) and a similar scenario high agreement for V52R1), while the variant with the label
can be postulated for Csx23. As crystallisation of FL Csx23 closest to the cytoplasmic region (N62R1) showed largest de-
was unsuccessful, we turned to Pulse Dipolar Electron Para- viations between the two models, which could indicate an am-
magnetic Resonance Spectroscopy (PDS) to obtain informa- biguous labelling site with more than one possible label con-
tion about conformational changes in the membrane span- formation (72).
ning domain of Csx23 upon cA4 binding. PDS requires site-
specific labelling of the protein of interest with a spin-label
(to yield the paramagnetic side-chain R1); we chose to in- Csx23 and NucC confer immunity to phage
troduce cysteine residues at key locations for MTSL spin- infection in E. coli
label attachment. Wild type Csx23 contains 4 native cysteine A limitation of the plasmid challenge assay is that it does not
residues, three in the soluble domain and one in the TM do- allow discrimination between a mechanism invoking effector-
main. Two of these, C85 and C105, are suitably positioned mediated programmed cell death and selective removal of the
A B
C D
Figure 6. Pulse EPR demonstrates cA4 mediated perturbation of the transmembrane domain. Pulse electron-electron double resonance (PELDOR) data
of the Csx23 AALA V52R1 variant reconstituted in nanodiscs (ND) in the presence (red) or absence (black) of cA4 . (A) Raw PELDOR data. (B)
Background-corrected traces with fits. (C) Overlay of corresponding distance distributions shown as 95% confidence bands with predicted distributions
from MMM and mtsslWizard based on the AF2 predicted tetrameric structure; colour bars indicate reliability ranges (green: shape reliable; yellow: mean
and width reliable; orange: mean reliable; red: no quantification possible). (D) Cartoon representation of AF2 predicted tetrameric structure of the
spin-labelled Csx23 V52R1 variant; each subunit is shown in a different colour.
plasmid-encoded target, as both will lead to the same phe- markedly different way. Cell growth stopped earlier (30 vs 60
notype. To extend these studies, we proceeded to investigate min), without a pronounced crash in OD600 and cell growth
the ability of the VmeCmr system to protect against infec- recommenced much more quickly than in the absence of effec-
tion by bacteriophage P1. E. coli cells expressing VmeCmr tor. Similar behaviour was observed for cells expressing NucC
with a crRNA targeting the lpa gene of phage P1 (73) along- (Supplementary Figure S13). The early growth arrest observed
side the target- and effector-containing plasmid were infected here could fit with a programmed cell death (abortive infec-
with phage P1 at varying multiplicities of infection (MOIs) tion) mechanism, as suggested previously for NucC (16), but
and the growth curves were recorded (Figure 7). As observed the growth behaviour overall appeared more consistent with
for the plasmid challenge assay, phage immunity by Csx23 a mechanism involving cell dormancy rather than death.
and NucC was dependent on the presence of targeting crRNA Irrespective of the presence of any effector protein in the
and hence cOA production (Supplementary Figure S13). In the host, the E. coli cultures grew to a similar density given enough
absence of any effector, culture collapse occurred in an MOI- time (>18 h). We therefore tested whether there were any dif-
dependent manner 1 – 3 h post infection. The culture recov- ferences in the amount of viable phage P1 between the strains
ered approximately 10 h post infection (Figure 7A), either due 24 h post infection. Strains containing phage P1-targeting
to the establishment of stable prophages over time or due to VmeCmr expressed with Csx23, NucC or no effector were
the presence and subsequent proliferation of persister cells. infected at an MOI of 15. A serial dilution of the cleared cul-
However, when either Csx23 or NucC were expressed, the ture supernatant 24 h post infection was then applied to agar
growth curves of cells infected with phage P1 at low to mod- plates containing an indicator strain and the phage titre was
erate MOIs were almost indistinguishable from those of un- determined from the number of plaques after overnight incu-
infected cells, suggesting efficient anti-phage defence. At high bation (Figure 7C). In the presence of either NucC or Csx23,
MOI (MOI = 15), where almost all cells should be infected a reduction in viable phage particles of at least 2 orders of
by phage, culture collapse occurred approximately 1 h post magnitude was observed compared to cultures without any ef-
infection in the absence of Csx23, as expected (Figure 7B). fector, confirming interference of phage P1 proliferation. The
However, in the presence of Csx23 the cells responded in a significance of these observations is discussed below.
B C
Figure 7. Phage P1 challenge of E. coli expressing VmeCmr and varying effector genes. (A) Growth curves of cells with no effector, Csx23 or NucC with
increasing amounts of phage P1. MOI: multiplicity of infection. The growth curve in the absence of phage infection is shown for comparison. (B) Growth
curves from cells carrying Csx23 or no effector after phage P1 infection at MOI 15 and expanded view to show early time points. (C) The number of
viable phage particles was determined by applying a dilution series of cleared supernatant from infected cultures to agar plates overlaid with BL21(DE3)
Star cells, the same E. coli strain as used in all in vivo assays. The plaques from ≥3 independent experiments consisting of two biological replicates
each were counted and plotted using Graphpad Prism. Mann–Whitney test (Prism 10.0) was used to determine statistical difference (not significant (ns):
P-value ≥0.05, ****P-value < 0.0001). Cultures of strains containing either Csx23 or NucC contained 100 times fewer viable phage compared to those
that did not carry any effector.
Discussion tion remains unclear (74). Very recently, a CRISPR associated

Type III CRISPR systems utilize a wide range of cyclic membrane protein (Cam1), which is activated by cA4 binding
nucleotide-activated effector proteins for defence against mo- to a dimeric CARF domain and depolarises membranes, has
bile genetic elements (25,26). The system studied here is en- been described (75). More broadly, membrane associated ef-
coded on a prophage that is found in many V. cholerae and fectors are found in many other prokaryotic defence pathways
V. meteocus genomes (29). As such, it may function in inter- that are presumed to function via programmed cell death. One
MGE conflict within the host cell, rather than providing anti- example is the Aga2 protein, which depolarises membranes
phage defence per se. Previously, we showed that this sys- when an associated Argonaute protein is activated during vi-
tem remains activated for an extended period, generates large ral infection in Saccharolobus islandicus (76). Furthermore,
quantities of cOA and is associated with the robust endonu- the CBASS and Pycsar anti-phage defence systems, which also
clease effector NucC (21) – all features that may reflect the function via cyclic nucleotide signalling, frequently encode
priorities of the prophage to prevent superinfection by com- effectors that function via membrane disruption (77,78). To
petitors, at the expense of the host cell if necessary. date, only the CBASS Cap15 effector has been investigated
Membrane-associated CRISPR effectors have been detected in detail. It is activated by binding a cyclic dinucleotide and
bioinformatically but characterised examples are still scarce. oligomerizes to disrupt and depolarise cell membranes, lead-
In addition to the aforementioned B. fragilis CorA effector ing to cell death (79).
(28), some type VI CRISPR systems have an associated Csx28 Csx23 utilises a small tetrameric cytoplasmic domain to
protein that forms an octameric membrane pore, resulting in bind cA4 —representing a new class of cyclic nucleotide recog-
membrane depolarisation, although the mechanism of activa- nition domain. The crystal structure of the CTD of Csx23 re-
vealed that mobile loops harbouring key interacting residues
for cA4 enclose the cyclic nucleotide binding cavity, reminis- Engineering and Physical Sciences Research Council
cent of the CARF and Crn2 families (80). EPR data show that [EP/X016455/1 to K.A., B.E.B., M.F.W.]; BBSRC equip-
cA4 binding to the CTD of Csx23 results in the structural ment grants [BB/R013780/1, BB/T017740/1 to B.E.B.].
perturbation of the trans-membrane helical domain, consis- Funding for open access charge: University of St Andrews
tent with the hypothesis that Csx23 functions via membrane block grant.
disruption, possibly as a type of ligand-gated channel. We
observed effective Csx23-mediated anti-phage immunity but
not a classical abortive infection phenotype, as infection at a Conflict of interest statement
very high MOI did not lead to complete culture collapse. This None declared.
should be caveated by the fact that we overexpressed the de-
fence system in E. coli, and by the understanding that phage P1
is a temperate phage. Nonetheless, these data do not provide References
strong support for a simple model of immunity by altruistic 1. Hale,C.R., Zhao,P., Olson,S., Duff,M.O., Graveley,B.R., Wells,L.,
suicide and are perhaps more consistent with the induction of Terns,R.M. and Terns,M.P. (2009) RNA-guided RNA cleavage by
dormancy. a CRISPR RNA-Cas protein complex. Cell, 139, 945–956.
In conclusion, our work reveals a new class of membrane- 2. Elmore,J.R., Sheppard,N.F., Ramia,N., Deighan,T., Li,H.,
associated CRISPR effector that functions via cyclic oligoad- Terns,R.M. and Terns,M.P. (2016) Bipartite recognition of target
enylate signalling to provide anti-viral immunity. The cA4 RNAs activates DNA cleavage by the type III-B CRISPR-Cas
binding domain of Csx23 is unrelated to the near-ubiquitous system. Genes Dev., 30, 447–459.
3. Estrella,M.A., Kuo,F.T. and Bailey,S. (2016) RNA-activated DNA
CARF/SAVED domain, highlighting the potential for further
cleavage by the type III-B CRISPR-Cas effector complex. Genes
discoveries in the mechanistically diverse type III CRISPR
Dev., 30, 460–470.
and CBASS defence pathways. Although we observe a cA4 - 4. Kazlauskiene,M., Tamulaitis,G., Kostiuk,G., Venclovas,C. and
activated allosteric modulation of the Csx23 trans-membrane Siksnys,V. (2016) Spatiotemporal control of type III-A
domain structure we have not yet observed changes in mem- CRISPR-Cas immunity: coupling DNA degradation with the
brane permeability—despite significant efforts to do so. This target RNA recognition. Mol. Cell, 62, 295–306.
is a clear area for future study. 5. Han,W., Li,Y., Deng,L., Feng,M., Peng,W., Hallstrom,S., Zhang,J.,
Peng,N., Liang,Y.X., White,M.F., et al. (2017) A type III-B
CRISPR-Cas effector complex mediating massive target DNA
Data availability destruction. Nucleic Acids Res., 45, 1983–1993.
6. Liu,T.Y., Iavarone,A.T. and Doudna,J.A. (2017) RNA and DNA
The coordinates and data have been deposited in the Protein
targeting by a reconstituted thermus thermophilus type III-A
Data Bank with deposition code 8QJK. The EPR research data CRISPR-Cas system. PLoS One, 12, e0170552.
underpinning this publication can be accessed at https://doi. 7. Kazlauskiene,M., Kostiuk,G., Venclovas,C., Tamulaitis,G. and
org/10.17630/e8334069- fc1a- 4329- a07f- 9d908515b7c0. Siksnys,V. (2017) A cyclic oligonucleotide signaling pathway in
type III CRISPR-Cas systems. Science, 357, 605–609.
8. Niewoehner,O., Garcia-Doval,C., Rostol,J.T., Berk,C., Schwede,F.,
Supplementary data Bigler,L., Hall,J., Marraffini,L.A. and Jinek,M. (2017) Type III
Supplementary Data are available at NAR Online. CRISPR-Cas systems produce cyclic oligoadenylate second
messengers. Nature, 548, 543–548.
9. Foster,K., Gruschow,S., Bailey,S., White,M.F. and Terns,M.P.
Acknowledgements (2020) Regulation of the RNA and DNA nuclease activities
required for pyrococcus furiosus type III-B CRISPR-Cas immunity.
Dr Laura Remmel is thanked for purification of MSP1D1, and Nucleic Acids Res., 48, 4418–4434.
Dr Samantha Pitt for helpful discussions. The Graphical Ab- 10. Molina,R., Stella,S., Feng,M., Sofos,N., Jauniskis,V.,
stract was produced with BioRender.com. Pozdnyakova,I., Lopez-Mendez,B., She,Q. and Montoya,G. (2019)
Author contributions: Sabine Grüschow, Investigation, Structure of Csx1-cOA4 complex reveals the basis of RNA decay
Methodology, Formal analysis, Visualisation, Writing— in type III-B CRISPR-Cas. Nat. Commun., 10, 4302.
original draft preparation, Writing—review and editing; Stu- 11. Rouillon,C., Athukoralage,J.S., Graham,S., Gruschow,S. and
art McQuarrie, Investigation, Methodology, Formal Anal- White,M.F. (2018) Control of cyclic oligoadenylate synthesis in a
type III CRISPR system. eLife, 7, e36734.
ysis, Writing—review and editing; Stephen McMahon, In-
12. Jia,N., Jones,R., Yang,G., Ouerfelli,O. and Patel,D.J. (2019)
vestigation, Methodology, Formal Analysis, Writing—review
CRISPR-Cas III-A Csm6 CARF domain is a ring nuclease
and editing; Katrin Ackermann, Investigation, Methodology, triggering stepwise cA4 cleavage with ApA>p formation
Formal Analysis, Writing-review and editing; Bela Bode, In- terminating RNase activity. Mol. Cell, 75, 944–956.
vestigation, Methodology, Formal Analysis, Writing-review 13. McMahon,S.A., Zhu,W., Graham,S., Rambo,R., White,M.F. and
and editing; Tracey M. Gloster, Funding acquisition, Su- Gloster,T.M. (2020) Structure and mechanism of a type III
pervision, Visualisation, Writing—review and editing; Mal- CRISPR defence DNA nuclease activated by cyclic oligoadenylate.
colm F. White, Conceptualisation, Formal analysis, Supervi- Nat. Commun., 11, 500.
sion, Project administration, Funding acquisition, Writing— 14. Rostol,J.T., Xie,W., Kuryavyi,V., Maguin,P., Kao,K., Froom,R.,
original draft preparation, Writing-review and editing. Patel,D.J. and Marraffini,L.A. (2021) The Card1 nuclease provides
defence during type-III CRISPR immunity. Nature, 590, 614–629.
15. Zhu,W., McQuarrie,S., Gruschow,S., McMahon,S.A., Graham,S.,
Funding Gloster,T.M. and White,M.F. (2021) The CRISPR ancillary
effector Can2 is a dual-specificity nuclease potentiating type III
Biotechnology and Biological Sciences Research Coun- CRISPR defence. Nucleic Acids Res., 49, 2777–2789.
cil [BB/T004789/1 to M.F.W., T.M.G.]; European Re- 16. Lau,R.K., Ye,Q., Birkholz,E.A., Berg,K.R., Patel,L., Mathews,I.T.,
search Council Advanced Grant [101018608 to M.F.W.]; Watrous,J.D., Ego,K., Whiteley,A.T., Lowey,B., et al. (2020)
Structure and mechanism of a cyclic trinucleotide-activated 33. Kapsalis,C., Wang,B., El Mkami,H., Pitt,S.J., Schnell,J.R.,
bacterial endonuclease mediating bacteriophage immunity. Mol. Smith,T.K., Lippiat,J.D., Bode,B.E. and Pliotas,C. (2019) Allosteric
Cell, 77, 723–733. activation of an ion channel triggered by modification of
17. Mayo-Muñoz,D., Smith,L.M., Garcia-Doval,C., Malone,L.M., mechanosensitive nano-pockets. Nat. Commun., 10, 4619.
Harding,K.R., Jackson,S.A., Hampton,H.G., Fagerlund,R.D., 34. Milov,A.D., Salikhov,K.M. and Shirov,M.D. (1981) Application of
Gumy,L.F. and Fineran,P.C. (2022) Type III CRISPR–Cas provides Eldor in electron-spin echo for paramagnetic center space
resistance against nucleus-forming jumbo phages via abortive distribution in solids. Fiz. Tverd Tela+, 23, 975–982.
infection. Mol. Cell, 82, 4471–4486. 35. Pannier,M., Veit,S., Godt,A., Jeschke,G. and Spiess,H.W. (2000)
18. Rostol,J.T. and Marraffini,L.A. (2019) Non-specific degradation of Dead-time free measurement of dipole-dipole interactions between
transcripts promotes plasmid clearance during type III-A electron spins. J. Magn. Reson., 142, 331–340.
CRISPR-Cas immunity. Nat. Microbiol., 4, 656–662. 36. Kerry,P.S., Turkington,H.L., Ackermann,K., Jameison,S.A. and
19. Santiago-Frangos,A., Hall,L.N., Nemudraia,A., Nemudryi,A., Bode,B.E. (2014) Analysis of influenza A virus NS1 dimer
Krishna,P., Wiegand,T., Wilkinson,R.A., Snyder,D.T., Hedges,J.F., interfaces in solution by pulse EPR distance measurements. J. Phys.
Cicha,C., et al. (2021) Intrinsic signal amplification by type III Chem. B, 118, 10882–10888.
CRISPR-Cas systems provides a sequence-specific SARS-CoV-2 37. Jeschke,G., Sajid,M., Schulte,M. and Godt,A. (2009) Three-spin
diagnostic. Cell Rep Med, 2, 100319. correlations in double electron-electron resonance. Phys. Chem.
20. Steens,J.A., Zhu,Y., Taylor,D.W., Bravo,J.P.K., Prinsen,S.H.P., Chem. Phys., 11, 6580–6591.
Schoen,C.D., Keijser,B.J.F., Ossendrijver,M., Hofstra,L.M., 38. Ackermann,K., Pliotas,C., Valera,S., Naismith,J.H. and Bode,B.E.
Brouns,S.J.J., et al. (2021) SCOPE enables type III CRISPR-Cas (2017) Sparse labeling PELDOR spectroscopy on multimeric
diagnostics using flexible targeting and stringent CARF mechanosensitive membrane channels. Biophys. J., 113,
ribonuclease activation. Nat. Commun., 12, 5033. 1968–1978.
21. Grüschow,S., Adamson,C.S. and White,M.F. (2021) Specificity and 39. Valera,S., Ackermann,K., Pliotas,C., Huang,H., Naismith,J.H. and
sensitivity of an RNA targeting type III CRISPR complex coupled Bode,B.E. (2016) Accurate extraction of nanometer distances in
with a NucC endonuclease effector. Nucleic Acids Res., 49, multimers by pulse EPR. Chem., 22, 4700–4703.
13122–13134. 40. Tait,C.E. and Stoll,S. (2016) Coherent pump pulses in double
22. Athukoralage,J.S., Rouillon,C., Graham,S., Grüschow,S. and electron electron resonance spectroscopy. Phys. Chem. Chem.
White,M.F. (2018) Ring nucleases deactivate type III CRISPR Phys., 18, 18470–18485.
ribonucleases by degrading cyclic oligoadenylate. Nature, 562, 41. Jeschke,G., Chechik,V., Ionita,P., Godt,A., Zimmerman,H.,
277–280. Banham,J.E., Timmel,C.R., Hilger,D. and Jung,D. (2006)
23. Athukoralage,J.S., McQuarrie,S., Grüschow,S., Graham,S., DeerAnalysis2006 - a comprehensive software package for
Gloster,T.M. and White,M.F. (2020) Tetramerisation of the analyzing pulsed ELDOR data. Appl. Magn. Reson., 30, 473–498.
CRISPR ring nuclease Crn3/Csx3 facilitates cyclic oligoadenylate 42. von Hagens,T., Polyhach,Y., Sajid,M., Godt,A. and Jeschke,G.
cleavage. eLife, 9, e57627. (2013) Suppression of ghost distances in multiple-spin double
24. Athukoralage,J.S., McMahon,S.A., Zhang,C., Gruschow,S., electron-electron resonance. Phys. Chem. Chem. Phys., 15,
Graham,S., Krupovic,M., Whitaker,R.J., Gloster,T.M. and 5854–5866.
White,M.F. (2020) An anti-CRISPR viral ring nuclease subverts 43. Edwards,T.H. and Stoll,S. (2018) Optimal Tikhonov regularization
type III CRISPR immunity. Nature, 577, 572–575. for DEER spectroscopy. J. Magn. Reson., 288, 58–68.
25. Shmakov,S.A., Makarova,K.S., Wolf,Y.I., Severinov,K.V. and 44. Chiang,Y.W., Borbat,P.P. and Freed,J.H. (2005) The determination
Koonin,E.V. (2018) Systematic prediction of genes functionally of pair distance distributions by pulsed ESR using Tikhonov
linked to CRISPR-Cas systems by gene neighborhood analysis. regularization. J. Magn. Reson., 172, 279–295.
Proc. Natl. Acad. Sci. U.S.A., 115, E5307–E5316. 45. Worswick,S.G., Spencer,J.A., Jeschke,G. and Kuprov,I. (2018)
26. Shah,S.A., Alkhnbashi,O.S., Behler,J., Han,W., She,Q., Hess,W.R., Deep neural network processing of DEER data. Sci. Adv., 4,
Garrett,R.A. and Backofen,R. (2019) Comprehensive search for eaat5218.
accessory proteins encoded with archaeal and bacterial type III 46. Fabregas Ibanez,L., Jeschke,G. and Stoll,S. (2020) DeerLab: a
CRISPR-cas gene cassettes reveals 39 new cas gene families. RNA comprehensive software package for analyzing dipolar electron
Biol, 16, 530–542. paramagnetic resonance spectroscopy data. Magn. Reson. (Gott.),
27. Rouillon,C., Schneberger,N., Chi,H., Blumenstock,K., Da Vela,S., 1, 209–224.
Ackermann,K., Moecking,J., Peter,M.F., Boenigk,W., Seifert,R., 47. Schiemann,O., Heubach,C.A., Abdullin,D., Ackermann,K.,
et al. (2023) Antiviral signalling by a cyclic nucleotide activated Azarkh,M., Bagryanskaya,E.G., Drescher,M., Endeward,B.,
CRISPR protease. Nature, 614, 168–174. Freed,J.H., Galazzo,L., et al. (2021) Benchmark test and guidelines
28. Chi,H., Hoikkala,V., Gruschow,S., Graham,S., Shirran,S. and for DEER/PELDOR experiments on nitroxide-labeled
White,M.F. (2023) Antiviral type III CRISPR signalling via biomolecules. J. Am. Chem. Soc., 143, 17875–17890.
conjugation of ATP and SAM. Nature, 622, 826–833. 48. Hagelueken,G., Ward,R., Naismith,J.H. and Schiemann,O. (2012)
29. McDonald,N.D., Regmi,A., Morreale,D.P., Borowski,J.D. and MtsslWizard: in Silico spin-labeling and generation of distance
Boyd,E.F. (2019) CRISPR-Cas systems are present predominantly distributions in PyMOL. Appl. Magn. Reson., 42, 377–391.
on mobile genetic elements in Vibrio species. Bmc Genomics 49. Hagelueken,G., Abdullin,D. and Schiemann,O. (2015) mtsslSuite:
(Electronic Resource), 20, 105. probing biomolecular conformation by spin-labeling studies.
30. Rouillon,C., Athukoralage,J.S., Graham,S., Grüschow,S. and Methods Enzymol., 563, 595–622.
White,M.F. (2019) Investigation of the cyclic oligoadenylate 50. Jeschke,G. (2018) MMM: a toolbox for integrative structure
signalling pathway of type III CRISPR systems. Methods modeling. Protein Sci., 27, 76–85.
Enzymol., 616, 191–218. 51. Vonrhein,C., Flensburg,C., Keller,P., Sharff,A., Smart,O.,
31. Grüschow,S., Athukoralage,J.S., Graham,S., Hoogeboom,T. and Paciorek,W., Womack,T. and Bricogne,G. (2011) Data processing
White,M.F. (2019) Cyclic oligoadenylate signalling mediates and analysis with the autoPROC toolbox. Acta. Crystallogr. D
mycobacterium tuberculosis CRISPR defence. Nucl. Acids Res., Biol. Crystallogr., 67, 293–302.
47, 9259–9270. 52. McCoy,A.J., Grosse-Kunstleve,R.W., Adams,P.D., Winn,M.D.,
32. Denisov,I.G., Grinkova,Y.V., Lazarides,A.A. and Sligar,S.G. (2004) Storoni,L.C. and Read,R.J. (2007) Phaser crystallographic
Directed self-assembly of monodisperse phospholipid bilayer software. J. Appl. Crystallogr., 40, 658–674.
nanodiscs with controlled size. J. Am. Chem. Soc., 126, 53. Winn,M.D., Ballard,C.C., Cowtan,K.D., Dodson,E.J., Emsley,P.,
3477–3487. Evans,P.R., Keegan,R.M., Krissinel,E.B., Leslie,A.G., McCoy,A.,
et al. (2011) Overview of the CCP4 suite and current 66. McQuarrie,S., Athukoralage,J.S., McMahon,S.A., Graham,S.,
developments. Acta. Crystallogr. D Biol. Crystallogr., 67, 235–242. Ackermann,K., Bode,B.E., White,M.F. and Gloster,T.M. (2023)
54. Jumper,J., Evans,R., Pritzel,A., Green,T., Figurnov,M., Activation of Csm6 ribonuclease by cyclic nucleotide binding: in
Ronneberger,O., Tunyasuvunakool,K., Bates,R., Zidek,A., an emergency, twist to open. Nucleic Acids Res., 51, 10590–10605.
Potapenko,A., et al. (2021) Highly accurate protein structure 67. Athukoralage,J.S., McQuarrie,S., Gruschow,S., Graham,S.,
prediction with AlphaFold. Nature, 596, 583–589. Gloster,T.M. and White,M.F. (2020) Tetramerisation of the
55. Liebschner,D., Afonine,P.V., Baker,M.L., Bunkoczi,G., Chen,V.B., CRISPR ring nuclease Crn3/Csx3 facilitates cyclic oligoadenylate
Croll,T.I., Hintze,B., Hung,L.W., Jain,S., McCoy,A.J., et al. (2019) cleavage. eLife, 9, e57627.
Macromolecular structure determination using X-rays, neutrons 68. Athukoralage,J.S., Graham,S., Gruschow,S., Rouillon,C. and
and electrons: recent developments in Phenix. Acta Crystallogr. D White,M.F. (2019) A type III CRISPR ancillary ribonuclease
Struct. Biol., 75, 861–877. degrades its cyclic oligoadenylate activator. J. Mol. Biol., 431,
56. Murshudov,G.N., Vagin,A.A. and Dodson,E.J. (1997) Refinement 2894–2899.
of macromolecular structures by the maximum-likelihood method. 69. Holm,L. (2020) DALI and the persistence of protein shape. Protein
Acta. Crystallogr. D Biol. Crystallogr., 53, 240–255. Sci., 29, 128–140.
57. Emsley,P., Lohkamp,B., Scott,W.G. and Cowtan,K. (2010) Features 70. Ren,J., Wang,J., Wang,Z. and Wu,J. (2014) Structural and
and development of Coot. Acta. Crystallogr. D Biol. Crystallogr., biochemical insights into the homotypic PB1-PB1 complex
66, 486–501. between PKCzeta and p62. Sci. China. Life Sci., 57, 69–80.
58. Long,F., Nicholls,R.A., Emsley,P., Graaeulis,S., Merkys,A., 71. Sumimoto,H., Kamakura,S. and Ito,T. (2007) Structure and
Vaitkus,A. and Murshudov,G.N. (2017) AceDRG: a function of the PB1 domain, a protein interaction module
stereochemical description generator for ligands. Acta Crystallogr. conserved in animals, fungi, amoebas, and plants. Sci. STKE, 2007,
D Struct. Biol., 73, 112–122. re6.
59. Chen,V.B., Arendall,W.B., Headd,J.J., Keedy,D.A., 72. Vitali,V., Ackermann,K., Hagelueken,G. and Bode,B.E. (2024)
Immormino,R.M., Kapral,G.J., Murray,L.W., Richardson,J.S. and Spectroscopically orthogonal labelling to disentangle site-specific
Richardson,D.C. (2010) MolProbity: all-atom structure validation nitroxide label distributions. Appl. Magn. Reson., 55, 187–205.
for macromolecular crystallography. Acta. Crystallogr. D Biol. 73. Lobocka,M.B., Rose,D.J., Plunkett,G. 3rd, Rusin,M.,
Crystallogr., 66, 12–21. Samojedny,A., Lehnherr,H., Yarmolinsky,M.B. and Blattner,F.R.
60. Mirdita,M., Schutze,K., Moriwaki,Y., Heo,L., Ovchinnikov,S. and (2004) Genome of bacteriophage P1. J. Bacteriol., 186,
Steinegger,M. (2022) ColabFold: making protein folding accessible 7032–7068.
to all. Nat. Methods, 19, 679–682. 74. VanderWal,A.R., Park,J.U., Polevoda,B., Nicosia,J.K., Molina
61. Paysan-Lafosse,T., Blum,M., Chuguransky,S., Grego,T., Pinto,B.L., Vargas,A.M., Kellogg,E.H. and O’Connell,M.R. (2023) Csx28 is a
Salazar,G.A., Bileschi,M.L., Bork,P., Bridge,A., Colwell,L., et al. membrane pore that enhances CRISPR-Cas13b-dependent
(2023) InterPro in 2022. Nucleic Acids Res., 51, D418–D427. antiphage defense. Science, 380, 410–415.
62. Hallgren,J., Tsirigos,K.D., Pedersen,M.D., Armenteros,J.J.A., 75. Baca,C.F., Yu,Y., Rostol,J.T., Majumder,P., Patel,D.J. and
Marcatili,P., Nielsen,H., Krogh,A. and Winther,O. (2022) Marraffini,L.A. (2024) The CRISPR effector Cam1 mediates
DeepTMHMM predicts alpha and beta transmembrane proteins membrane depolarization for phage defence. Nature, 625,
using deep neural networks. bioRxiv doi: 797–804.
https://doi.org/10.1101/2022.04.08.487609, 10 April 2022, 76. Zeng,Z., Chen,Y., Pinilla-Redondo,R., Shah,S.A., Zhao,F.,
preprint: not peer reviewed. Wang,C., Hu,Z., Wu,C., Zhang,C., Whitaker,R.J., et al. (2022) A
63. Smalakyte,D., Kazlauskiene,M., J,F.H., Ruksenaite,A., Rimaite,A., short prokaryotic Argonaute activates membrane effector to
Tamulaitiene,G., Faergeman,N.J., Tamulaitis,G. and Siksnys,V. confer antiviral defense. Cell Host Microbe, 30, 930–943.
(2020) Type III-A CRISPR-associated protein Csm6 degrades 77. Millman,A., Melamed,S., Amitai,G. and Sorek,R. (2020) Diversity
cyclic hexa-adenylate activator using both CARF and HEPN and classification of cyclic-oligonucleotide-based anti-phage
domains. Nucleic Acids Res., 48, 9204–9217. signalling systems. Nat. Microbiol., 5, 1608–1615.
64. Makarova,K.S., Timinskas,A., Wolf,Y.I., Gussow,A.B., Siksnys,V., 78. Tal,N., Morehouse,B.R., Millman,A., Stokar-Avihail,A.,
Venclovas,C. and Koonin,E.V. (2020) Evolutionary and functional Avraham,C., Fedorenko,T., Yirmiya,E., Herbst,E., Brandis,A.,
classification of the CARF domain superfamily, key sensors in Mehlman,T., et al. (2021) Cyclic CMP and cyclic UMP mediate
prokaryotic antivirus defense. Nucleic Acids Res., 48, 8828–8847. bacterial immunity against phages. Cell, 184, 5728–5739.
65. Garcia-Doval,C., Schwede,F., Berk,C., Rostol,J.T., Niewoehner,O., 79. Duncan-Lowey,B., McNamara-Bordewick,N.K., Tal,N., Sorek,R.
Tejero,O., Hall,J., Marraffini,L.A. and Jinek,M. (2020) Activation and Kranzusch,P.J. (2021) Effector-mediated membrane disruption
and self-inactivation mechanisms of the cyclic controls cell death in CBASS antiphage defense. Mol. Cell, 81,
oligoadenylate-dependent CRISPR ribonuclease Csm6. Nat. 5039–5051.
Commun., 11, 1596. 80. Athukoralage,J.S. and White,M.F. (2021) Cyclic oligoadenylate
signalling and regulation by ring nucleases during type III CRISPR
defence. RNA, 27, 855–867.
Received: December 1, 2023. Revised: February 21, 2024. Editorial Decision: February 22, 2024. Accepted: February 26, 2024
© The Author(s) 2024. Published by Oxford University Press on behalf of Nucleic Acids Research.
This is an Open Access article distributed under the terms of the Creative Commons Attribution License (http://creativecommons.org/licenses/by/4.0/), which permits unrestricted reuse,
distribution, and reproduction in any medium, provided the original work is properly cited.

CRISPR Antiphage Defence Mediated by The Cyclic Nucleotide-Binding Membrane Protein Csx23

Uploaded by

Copyright:

Available Formats

CRISPR Antiphage Defence Mediated by The Cyclic Nucleotide-Binding Membrane Protein Csx23

Uploaded by

Document Information

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

CRISPR Antiphage Defence Mediated by The Cyclic Nucleotide-Binding Membrane Protein Csx23

Uploaded by

Copyright:

Available Formats

Nucleic Acids Research, 2024, 52, 2761–2775

CRISPR antiphage defence mediated by the cyclic

Discussion tion remains unclear (74). Very recently, a CRISPR associated

You might also like