Review Article

Anti-CRISPR Discovery: Using Magnets

to Find Needles in Haystacks

Kevin J. Forsberg

Department of Microbiology, University of Texas Southwestern Medical Center, Dallas, TX 75390, USA

Correspondence to : kevin.forsberg@utsouthwestern.edu, @kvnforsberg,

https://doi.org/10.1016/j.jmb.2023.167952
Edited by Karen Maxwell

Abstract
CRISPR-Cas immune systems in bacteria and archaea protect against viral infection, which has spurred
viruses to develop dedicated inhibitors of these systems called anti-CRISPRs (Acrs). Like most host-virus
arms races, many diverse examples of these immune and counter-immune proteins are encoded by the
genomes of bacteria, archaea, and their viruses. For the case of Acrs, it is almost certain that just a small
minority of nature’s true diversity has been described. In this review, I discuss the various approaches
used to identify these Acrs and speculate on the future for Acr discovery. Because Acrs can determine
infection outcomes in nature and regulate CRISPR-Cas activities in applied settings, they have a dual
importance to both host-virus conflicts and emerging biotechnologies. Thus, revealing the largely hidden
world of Acrs should provide important lessons in microbiology that have the potential to ripple far beyond
the field.
Ó 2023 The Author. Published by Elsevier Ltd. This is an open access article under the CC BY-NC-ND license (http://creati-
vecommons.org/licenses/by-nc-nd/4.0/).

Introduction CRISPR-Cas systems enable this defense by incor-

porating short (
30 – 40 bp) spacer sequences
Bacteria and archaea have been locked in a tit- from invading genomes into a prescribed immunity
for-tat evolutionary arms race with their respective locus in their host genome (called the ‘CRISPR’
viruses for billions of years, as host and virus locus). These spacers are then transcribed and pro-
must constantly innovate to keep pace with their cessed into CRISPR RNAs (crRNAs) that, together
competitor’s latest inventions. This has produced with Cas proteins, cleave invader genomes in a
a wealth of antiviral defense systems in these sequence-specific manner.5 In response to this
hosts and a similarly diverse set of counter- interference, viruses and other mobile genetic ele-
defenses in their viruses, the composition of which ments have evolved dedicated CRISPR-Cas inhibi-
varies tremendously from genome to genome.1–4 tors, called anti-CRISPRs, which allow these
These defense systems include the CRISPR (clus- invaders to overcome CRISPR-Cas immunity.9
tered, regularly interspersed palindromic repeats)- Most known anti-CRISPRs bind directly to their cog-
Cas (CRISPR-associated) proteins, which are the nate Cas proteins to antagonize function, though
only known adaptive immune systems in bacteria additional modes of CRISPR-Cas inhibition have
or archaea. been observed.10
Approximately 40% of sequenced bacteria and The molecular machinery responsible for
85% of sequenced archaea encode for CRISPR- incorporating new spacer sequences into a
Cas systems that protect against viruses, CRISPR locus is largely conserved across
plasmids, and other mobile genetic elements.5–8 CRISPR-Cas systems.11,12 In contrast, the effector

0022-2836/Ó 2023 The Author. Published by Elsevier Ltd.This is an open access article under the CC BY-NC-ND license (http://creativecommons.org/
licenses/by-nc-nd/4.0/). Journal of Molecular Biology xxx (xxxx) xxx

Please cite tts article as: K.J. Forsberg, Anti-CRISPR Discovery: Using Magnets to Find Needles in HaystacksJournal of Molecular Biology, https://doi.org/10.1016/j.jmb.2023.167952
K.J. Forsberg Journal of Molecular Biology xxx (xxxx) xxx

proteins that interfere with virus replication vary sig- CRISPR-Cas activity.16–18 As is often the case,
nificantly from system to system and can have one of most promising strategies emerged from nat-
entirely distinct evolutionary histories.13 These var- ure, with the discovery of anti-CRISPR proteins
ious effectors are used to define and categorize encoded by bacterial viruses called bacterio-
CRISPR-Cas systems into two broad classes (1 phages, or ‘phages’ for short (Figure 1(A)).19
and 2) that encompass six different types (I – VI) The discovery of anti-CRISPRs ten years ago set
and, at last count, include 33 sub-types (indicated off an enthusiastic effort from numerous groups to
by trailing letters).5 These categories are defined identify and characterize the many inhibitors of
by phylogeny, locus organization, and mode of CRISPR-Cas systems suspected to exist in
action. The highest-level distinction, classes 1 and nature. The focus of this review will be to recap
2, separate CRISPR-Cas systems based on the these discovery efforts, from the early days of
number of Cas effector proteins used to mediate anti-CRISPR biology to today and will include
interference. Class 1 systems (which include types speculations on the future of this fast-paced field.
I, III, and IV) use multi-protein complexes to dock
with a crRNA, hybridize to a target locus, and initiate Discovery of the first Anti-CRISPRs
interference. Class 2 systems incorporate each of
these activities into a single, large, multi-domain The initial discovery of phage anti-CRISPR genes
protein and encompass type II, V, and VI systems. was the result of a targeted search by Joseph
Because CRISPR-Cas systems enable Bondy-Denomy, April Pawluk, Karen Maxwell, and
programmable DNA binding and nuclease activity, Alan Davidson at the University of Toronto.19 The
they have spurred a technological revolution to authors hypothesized that inhibitors of CRISPR-
edit, eliminate, modify, or detect specific Cas systems would be encoded by at least some
sequences in almost any genome or sample phage genomes. To search for anti-CRISPRs, they
type.14,15 The therapeutic and economic implica- screened a set of isolates derived from Pseu-
tions for these technologies are staggering but domonas aeruginosa PA14, which each contained
come with significant ethical and safety concerns a different prophage integrated into an otherwise
that have fueled research into ways of regulating identical genomic background. The term prophage

Figure 1. Phage genomes encode Acrs. (A) Infectivity screens can reveal phage genomes with potential acr genes,
for example, based on phage plaques appearing on an otherwise CRISPR-immune bacterial host. (B) Phages often
cluster acr genes together in discrete genomic loci. For this reason, genes nearby known acr genes are good
additional acr candidates. This feature can be used in an Acr discovery approach called guilt by association (GBA)
that often spans several phage genomes to reveal diverse Acrs.
2
K.J. Forsberg Journal of Molecular Biology xxx (xxxx) xxx

refers to a phage that is integrated into the genome identity.20,22 This strong conservation and tight
of a bacterial host, which is then called a lysogen to association with acr loci prompted Pawluk et al. to
indicate its prophage-positive status. The authors use aca1 as a marker gene to search for previously
infected their lysogen collection with additional unknown anti-CRISPRs.23 This approach led to the
phages that are typically incapable of forming pla- discovery of two new anti-CRISPRs in P. aerugi-
ques on wild-type P. aeruginosa PA14, due to its nosa (AcrIF6 and AcrIF7), one of which, AcrIF6,
native type I-F CRISPR-Cas system. Surprisingly, could inhibit both type I-F and I-E CRISPR-Cas sys-
in three of these experiments, the authors observed tems. Homologs of acrIF6 were the first acr genes
a complete restoration of phage infectivity, to the found outside the genus Pseudomonas and
level of a control strain that lacked CRISPR-Cas included one example adjacent to another HTH
immunity. This indicated that genes encoded by transcription factor gene, named aca2.24 A second
these prophages were responsible for inhibiting search with Aca2 revealed three additional Acrs
the activity of this type I-F CRISPR-Cas system. (AcrIF8 – AcrIF10) in diverse proteobacteria.23 Of
The authors identified a candidate anti-CRISPR note, the loci encoding these Acrs shared no homol-
locus by comparing the genomes of the CRISPR- ogy with Aca1 or AcrIF6, the initial search queries.
inhibiting prophages to those of related Instead, they were found via a daisy chain search
Pseudomonas phages in the same Mu-like family. strategy that relies on the propensity of phages to
This locus was nestled between conserved phage group genes with similar functions into discrete
structural genes and contained an assortment of loci.25 This approach uses homologs of an initial
small genes of unknown function. The small query to find novel Acr loci, which provides new
genes often lacked homology to one another, queries for additional homology searches, revealing
even in otherwise closely related genomes. yet more Acr loci, and so on and so forth. Because
Because many of these genes were absent from co-localization in the same genomic locus is key for
the CRISPR-sensitive phages, the authors tested linking dissimilar sequences in the daisy chain, this
17 examples from their collection for CRISPR-Cas search strategy is referred to generally as ‘guilt by
inhibition, finding eight genes from five protein association’ (GBA) and has been foundational for
families could restore infectivity to CRISPR- the identification of many novel Acrs (Figure 1
sensitive phages. These five proteins were (B)).26–30
ultimately named AcrIF1 through AcrIF5, denoting For instance, the first genetically encoded
their anti-CRISPR (‘Acr’) activity against type I-F inhibitors of CRISPR-Cas9 were discovered via
(‘IF’) CRISPR-Cas systems. With their discovery, guilt-by-association searches, which revealed
the field of anti-CRISPR biology was born.19 three unrelated inhibitors of the type II-C CRISPR-
A year later, a second publication from this group Cas system from Neisseria meningitidis.26 A
completed the characterization of this anti-CRISPR BLAST search with Aca2 uncovered an acrIIC1
locus, revealing that many of the previously gene in Brackiella oedipodis, which had a homolog
undescribed genes could inhibit a type I-E in N. meningitidis that was adjacent to yet another
CRISPR-Cas system of P. aeruginosa.20 These aca gene (named aca3). Additional searches with
genes represented four protein families, now Aca3 revealed two more anti-CRISPRs in putative
named AcrIE1 – AcrIE4, and, like the type IF Acrs, N. meningitidis prophages, ultimately named
licensed a previously CRISPR-sensitive phage to acrIIC2 and acrIIC3. All three Acrs protected a plas-
infect an otherwise immune P. aeruginosa strain, mid from Cas9 attack, inhibited nuclease activity
this time expressing an active type I-E CRISPR- in vitro, bound directly to Cas9, and prevented its
Cas system. This analysis also revealed a similar genome editing in human cell lines.26 Subsequent
promoter upstream of these acr loci and a con- searches with Aca2 revealed two more type II-C
served gene with a helix-turn-helix (HTH) DNA bind- Acrs (AcrIIC4 and AcrIIC5) while an independent
ing domain at the loci’s 30 end. This acr-associated query with Aca3 uncovered AcrIIC6.27,28
gene (aca1) was later revealed to encode an auto- The success of any guilt-by-association search is
regulatory transcriptional repressor of acr loci, necessarily dependent on the natural gene
restricting Acr expression to the early stages of distribution of the starting query. Some genes are
phage infection where CRISPR-Cas inhibition is narrowly restricted to a single phylogenetic group,
most needed and fitness costs are minimal.21 This like Aca1 is restricted to Pseudomonas.20,23 Others,
core regulatory function probably explains the like AcrIF11 and Aca5, are found across highly
strong degree of conservation in Aca1 homologs diverged taxa and are associated with diverse Acr
within the phages of P. aeruginosa. effectors and Aca proteins.29,30 AcrIF11 was dis-
covered via its association with Aca1 in Pseu-
domonas but is also found outside that genus with
Guilt By Association: finding anti- various Acrs that inhibit multiple CRISPR-Cas types
CRISPRs in diverse bacteria (described below) and with four other Aca proteins
(Aca4 – Aca7).29 Homologs of Aca5 are found in
Within Pseudomonas, aca1 homologs are found dozens of bacterial genera and led to the discovery
exclusively near acr genes and their encoded of 11 new Acrs (AcrIE8, active against type I-E
proteins share, on average, 95% amino acid
3
K.J. Forsberg Journal of Molecular Biology xxx (xxxx) xxx

CRISPR-Cas systems and AcrIF15 – AcrIF24, mechanisms, including the first examples of Acrs
active against type I-F systems).30 Interestingly, that inactivate Cas proteins enzymatically34–37 (for
many of these acr genes co-localize with genes pre- further details, see the review on type V Acrs by
dicted to encode inhibitors of other bacterial Marino et al. elsewhere in this special issue).
defense systems. This suggests that a GBA search Though STS-based approaches can identify
strategy could extend beyond acrs, enabling the genomes that are likely to encode for Acrs, a
discovery of new phage inhibitors that potentially secondary strategy is often required to pinpoint
act against many of the novel bacterial defense sys- the specific loci and genes responsible for
tems described in recent years (these defense sys- CRISPR-Cas inhibition. One way to reduce the
tems are reviewed elsewhere, for instance in ref).31 genomic search space is to consider only the
AcrIF11 was first discovered by Marino et al. in an prophages and mobile genetic elements present,
Aca1-based GBA search that revealed two type I-F as these are the most common sources of Acrs in
Acrs (AcrIF11 and AcrIF12), three type I-E Acrs nature. However, for the average lysogen, over
(AcrIE5 – AcrIE7), and a dual-acting AcrIE4- 100 Kb of DNA is attributable to prophages, which
AcrIF7 fusion protein.29 Some of the genomes with leaves a long list of putative acr candidates for
acrIF11 homologs also encoded for type V follow-up testing, especially considering the high
CRISPR-Cas systems, which use the single effec- coding density typical of phage genomes.25,38 For
tor protein Cas12 to process crRNAs and cleave instance, to find three type V acr genes, Watters
complementary target DNA substrates.29,32 et al. screened more than 500 Kb of DNA from Mor-
Because no Acrs against Cas12 had yet been iden- axella genomes with self-targeting CRISPR spac-
tified, the authors wondered whether acrIF11 may ers.33 To achieve this, they used a cell-free
occur alongside type V acrs in these bacteria.29 transcription-translation (TXTL) assay that employs
To choose examples for further study, they priori- CRISPR-Cas targeting to repress the expression of
tized genomes with type V CRISPR loci that had a fluorescent reporter, which can be inhibited in the
perfect matches in their spacer regions to other presence of an Acr, thus restoring fluorescence
sequences within the same genome, oriented in a (Figure 2(C), Figure 2(D)).39,40 In this scheme, they
manner expected to trigger CRISPR-Cas activa- amplified 57 DNA fragments that spanned their tar-
tion. Under normal circumstances, these self- get search space, added these amplicons to a TXTL
targeting spacers (STSs) would cause Cas12 to assay with Cas12, crRNA, and GFP-encoding frag-
cleave its own host’s genome, triggering a form of ments, and identified four genomic fragments that
autoimmunity that can result in cell death (Figure 2 reproducibly restored fluorescence. By individually
(A)). Thus, the authors reasoned, the presence of a testing 14 genes from these four fragments, they
stable STS in a bacterial genome is evidence for an ultimately identified acrVA1, acrVA4, and acrVA5
inactive CRISPR-Cas system in that host, perhaps (described above).
due to an undiscovered acr gene (Figure 2(B)). This Because STS and GBA searches work at
logic proved fruitful, as it correctly identified a geno- different scales, they are highly complementary.
mic locus in Moraxella catarrhalis with both acrIF11 STS approaches can be automated, so they
and a gene encoding for an inhibitor of the type V-A efficiently identify candidate genomes that may
effector nuclease, Cas12a. A closer examination of encode desired Acrs. However, this strategy
this locus and its close relatives led the authors to struggles to predict specific Acr loci, which is a
discover some of the first type V Acrs (AcrVA1 – strength of hand-curated, GBA-based searches.
AcrVA3), along with the first inhibitor of a type I-C Demonstrating this complementarity, Leon et al.
CRISPR-Cas system, AcrIC1.29 integrated these two search strategies in their
quest to identify inhibitors of type I-C CRISPR-Cas
systems.41 The authors first identified a strain of
Self-Targeting Spacers: an indicator of P. aeruginosa with a self-targeting type I-C CRISPR
CRISPR-Cas inhibition spacer and then predicted an Acr locus within this
genome based on the presence of an Aca1 homo-
A self-targeting CRISPR spacer suggests that an log. This analysis revealed a dual-acting inhibitor
acr gene may also be present in that genome to of type I-F and I-C CRISPR-Cas systems with
prevent autoimmunity (Figure 2(A), Figure 2(B)). homology to AcrIF2 (also named AcrIF2/AcrIC2).
With this logic, some Acr discovery efforts have Subsequent GBA searches with this locus and
used an STS-based strategy to find new CRISPR- Aca4 revealed an additional six type I-C Acrs
Cas inhibitors, without relying on GBA. For (AcrIC3 – AcrIC8) and one type I-E inhibitor
instance, Watters et al. developed a systematic (AcrIE9). Many of these acr genes were found nat-
search tool for scanning bacterial genomes to find urally within mobile genetic elements of Pseu-
STSs, which helped the authors identify three acrs domonas that are targeted by type I-C CRISPR
in Moraxella that enabled Cas12 inhibition.33 Pub- spacers in this genus.41 Evidence of CRISPR-
lished simultaneously with the discovery of AcrVA1 targeting suggests an ecological need for Acr activ-
– AcrVA3,29 this study co-discovered AcrVA1 while ity and has been used since the earliest Acr discov-
also revealing two new type V Acrs, AcrVA4 and eries to underscore the real-world importance of
AcrVA5.33 These five Acrs inhibit Cas12 via diverse
4
K.J. Forsberg Journal of Molecular Biology xxx (xxxx) xxx

Figure 2. Self-targeting spacers can reveal Acr-encoding genomes. (A) A self-targeting spacer (STS) in a CRISPR-
Cas competent host should trigger double-strand DNA breaks in the host genome, resulting in an unstable
autoimmune state that can lead to cell death. (B) The presence of a CRISPR-Cas inhibitor in an STS-encoding
genome can prevent autoimmunity and enable the stable inheritance of the STS across generations. For this reason,
the presence of an STS in any CRISPR-Cas containing genome has been used to infer the presence of a putative Acr
elsewhere in the genome. (C) One strategy for finding Acrs in STS-encoding genomes is to use a cell-free
transcription-translation (TXTL) assay. Processed cell extracts capable of in vitro TXTL are mixed with DNA
fragments that encode for Cas machinery, GFP, and a GFP-targeting crRNA. In the absence of an Acr, GFP signal is
suppressed due to active CRISPR-Cas targeting. However, if a candidate DNA fragment encodes an Acr, this will
block CRISPR-Cas activity and restore GFP fluorescence. (D) Hypothetical time course fluorescence data for an Acr-
encoding and Acr-negative DNA fragment.

Acrs in natural phage infections.19 For a detailed many CRISPR-Cas systems from many bacteria
discussion of type I Acr functions and mechanisms, without known antagonists. Among these was the
see the review by Feng et al. elsewhere in this spe- type II-A CRISPR-Cas9 system, which includes
cial issue. Streptococcus pyogenes Cas9 (SpyCas9), the vari-
ant used most frequently for genome-editing appli-
cations.42 Given the importance of SpyCas9 in
Terra Nova: independent Acr microbiology and biotechnology, multiple groups
discovery reveals new gene-sharing sought to identify natural inhibitors of this phage-
networks defense enzyme.
The first phage inhibitors of type II-A CRISPR-
Of the 53 Acr families mentioned thus far, 51 can Cas9 systems were discovered experimentally,
be linked via guilt-by-association to the original acr guided by an STS search to identify bacterial and
locus in P. aeruginosa identified by Bondy- prophage genomes with putative acr genes.43
Denomy et al. (the exceptions are AcrVA4 and Rauch et al. noticed that 20% of the Listeria mono-
AcrVA5).19,33 Put another way, the discovery of a cytogenes genomes with a type II-A CRISPR-Cas9
single Acr locus presaged the eventual identification system also possessed a self-targeting CRISPR
of 51 Acrs active against five different CRISPR-Cas spacer. They next performed plasmid transforma-
subtypes. This observation emphasizes the tion assays in different L. monocytogenes strains
extreme mobility of acr and aca genes – they are and observed that Cas9 was inactive in self-
readily exchanged between genomes and habitats, targeting genomes, suggesting that some mecha-
over short evolutionary timescales. This produces nism of Cas9 inhibition was present in these bacte-
gene-sharing networks with extremely high flux, ria. Because many self-targets mapped to resident
but which are often restricted to certain taxonomic prophages, the authors reasoned that these phages
groups. For instance, GBA searches that originated were a likely source of the putative acr genes. They
from the original Pseudomonas Acr locus could used one of these phages to lysogenize a L. mono-
identify only Acrs from other Proteobacteria, leaving cytogenes strain that previously showed Cas9
5
K.J. Forsberg Journal of Molecular Biology xxx (xxxx) xxx

activity and demonstrated that this new prophage Interestingly, all four of these Acrs manipulate the
could block Cas9 function. By comparing the Cas9 ribonucleoprotein complex, either by prevent-
Cas9-inhibiting prophage to a close relative that ing its formation, reducing guide RNA levels, or by
was incapable of inactivating the enzyme, the truncating the guide RNA enzymatically.45,48 These
authors identified ten potential Acr-encoding loci, guide RNA-centric mechanisms could help to
based on their differential presence across these explain the broad-spectrum activity of these Acrs,
phage genomes. Two genes from one of these loci given the relatively conserved nature of this impor-
were individually sufficient for Cas9 inhibition and tant Cas9 co-factor.45,49
were named acrIIA1 and acrIIA2.43 Shortly following the initial discovery of type II-A
AcrIIA1 is widely distributed across Firmicutes Acrs, Hynes et al. identified two additional
and has proven to be an effective query for GBA examples, AcrIIA5 and AcrIIA6. These Acrs were
searches, via its direct associations or indirect discovered using empirical phage screens, based
links to eight distinct Acr families.43–45 AcrIIA1’s on the capacity of their host phages to overcome
prevalence across Acr loci may be related to its dual Cas9 immunity.50,51 To find the acrIIA5 gene
Aca-Acr function. It has a conserved N-terminal, responsible for Cas9 inhibition, the authors
HTH domain that regulates transcription in an expressed individual genes from its CRISPR-
Aca-like manner and has a variable C-terminal resistant phage and tested whether each gene
domain that mediates Cas9 inhibition.44,46 In Liste- could restore infectivity of a second, CRISPR-
ria and Streptococcus phages, AcrIIA1 is associ- sensitive phage. Only one gene mediated this
ated with AcrIIA2 (as previously discussed) along effect, which was named acrIIA5.50 To find acrIIA6,
with AcrIIA3 and AcrIIA4, which were identified in the authors used GBA, as the gene immediately
the same study via GBA searches in Listeria and upstream of acrIIA5 was homologous to the gene
Streptococcus prophages.43 Three of these four upstream of acrIIA6 in its CRISPR-resistant
Acrs were also capable of inhibiting SpyCas9, as phage.51 Though both Acrs were discovered for
determined by this study or successive efforts.43,44 their ability to inhibit Streptococcus thermophilus
A subsequent GBA analysis later identified Cas9 (denoted St1Cas9, to distinguish it from
AcrIIA12 in Listeria phages, via its association with another Cas9 ortholog in this bacterium), they dif-
AcrIIA1.44 With this discovery, all known acrIIA1 fered in the spectrum of Cas9 orthologs they antag-
genes in Listeria phages were observed to co- onized. AcrIIA6 exhibited a limited inhibitory range,
occur with at least one other acr gene, corroborating acting against only St1Cas9, whereas AcrIIA5 could
the Aca-like nature of AcrIIA1 and pushing the block the activity of type II-A, II-B, and II-C Cas9
search for additional type II-A Acrs outside of these orthologs.51–53 Despite inhibiting fewer Cas9 ortho-
genomes. logs than AcrIIA5, AcrIIA6 is more prevalent in
Watters et al. used an STS-based search strategy Streptococcus phages.51,54 This prevalence may
to identify two Staphylococcus aureus genomes be related to a secondary role as Aca-like transcrip-
with potential Acrs active against the type II-A tional regulator, based on its structural similarity to
CRISPR-Cas9 system in this bacterium.47 From HTH domain proteins and its demonstrated ability
one STS-encoding genome, AcrIIA13 was identi- to bind double-stranded DNA.51
fied using the fragment based TXTL screen The phages that encode for AcrIIA5 and AcrIIA6
described earlier. With the other genome, no were known to impede the acquisition of CRISPR
CRISPR-Cas inhibition was observed using this spacers in S. thermophilus before the discovery of
assay, presumably because the causal acr was these Acrs, based on extensive prior study of S.
not expressed or was located outside the set of thermophilus and its phages, which are of critical
genomic fragments chosen for examination. To find importance to the dairy industry.50,51 In fact, this
additional Acrs, the authors then used a classic observation helped to guide Hynes et al. to these
GBA search with an unnamed, acrIIA13-adjacent phages in their initial Acr search, as they reasoned
aca gene, which revealed two more Acrs, AcrIIA14 that the diminished CRISPR acquisition rate could
and AcrIIA15.47 be due to the presence of a phage-encoded Acr.50
To find type II-A Acrs in diverse taxa, Mahendra This ability of Acrs to inhibit CRISPR acquisition
et al. looked near acrIIA1 homologs in plasmids has been observed for multiple Acrs across both
and conjugative elements.45 This revealed AcrIIA16 class 1 and class 2 systems, though it was initially
in a Listeria plasmid and AcrIIA17 in an Enterococ- presumed to be a secondary consequence of block-
cus conjugative element, whose homologs uncov- ing CRISPR interference.50,51,55 While this view is
ered AcrIIA18 and AcrIIA19 in mobile genetic probably true for many Acrs, the recent discovery
elements of Streptococcus and Staphylococcus, of a naturally occurring AcrIIA6 variant with anti-
respectively. AcrIIA16 and AcrIIA17 exhibited a acquisition but not anti-interference properties sug-
remarkably broad capacity to inhibit both II-A and gests that some phages may benefit from solely
II-C Cas9 orthologs when expressed in human impeding the adaptive phase of CRISPR immu-
cells, perhaps reflective of their presence on plas- nity.56 It remains to be seen whether this finding is
mids and conjugative elements, which can have an outlier or the first of many anti-adaptation Acrs
wider phylogenetic distributions than most phages. to be discovered. Nonetheless, this intriguing dis-
6
K.J. Forsberg Journal of Molecular Biology xxx (xxxx) xxx

covery warrants additional study into the potential of tion.19 Thus, the authors reasoned, plasmids with
other Acr candidates beyond AcrIIA6 to inhibit metagenomic DNA inserts that encode for Acrs
spacer acquisition, especially if they cannot inhibit are likely to express them in Escherichia coli and
CRISPR interference. be protected from SpyCas9 in this lab-friendly
As was the case for AcrIIA1, AcrIIA13, and many screening host. Other plasmids would be eliminated
of the Acas discussed previously, GBA searches during SpyCas9 attack, enabling only plasmids with
with AcrIIA6 enabled the identification of many true Acrs to survive, which could then be recovered
additional Acrs.54 Homologs of the acrIIA6 gene via antibiotic resistance markers that they encode
were found adjacent to acrIIA24 and acrIIA25, the (Figure 3(A)).57,58
latter of which was found next to an aca gene Using this logic, Uribe et al. screened for
(aca11) associated with five new Acrs (AcrIIA26 – SpyCas9 inhibitors from human fecal, soil, and
AcrIIA30). These discoveries led to the identifica- animal-associated metagenomic libraries.57 Before
tion of two additional Acas (Aca12, Aca13) and Acrs performing their selections, the authors empirically
(AcrIIA31, AcrIIA32) in subsequent GBA determined that 1 out of every
104 plasmids in
searches.54 Yet again, these findings reinforce an their library could withstand SpyCas9 attack without
important lesson: new Acr discoveries can be multi- encoding for an Acr, a false-positive rate that was
plicative, especially if the primary finding reveals a later attributed to the frequent occurrence of
novel Acr found outside of a known GBA network spyCas9-inactivating mutations.57,58 Though their
of shared genes. For a detailed discussion of type assay provided strong selection for Acr activity,
II Acrs, please see the review by Hwang and Max- the size of their libraries (
106 clones) meant that
well elsewhere in this special issue. potentially hundreds of the unique DNA inserts in
GBA searches have produced more Acr their selected plasmids lacked acr genes. To find
discoveries than any other search strategy, but the true Acrs among these inserts, the authors pre-
are limited by detectable homology, which dicted which sequences were most likely to come
degrades quickly across phage genomes.22 Practi- from bacteriophage genomes or other mobile
cally speaking, this can limit the search space genetic elements using sequence homology. They
accessible to GBA-based queries to shallow evolu- selected 39 of these sequences for further study,
tionary depths and restricted taxonomic groups. ultimately revealing four new Acrs (AcrIIA7 –
Indeed, the precipitating event for each new wave AcrIIA10). These Acrs show a wide phylogenetic
of GBA discovery has been the experimental identi- distribution, protect a plasmid from SpyCas9
fication of a primary Acr, which often involves an in vivo, and inhibit SpyCas9’s activity in vitro. One
empirical genetic screen (e.g., AcrIF1, AcrIIA1, example, AcrIIA7, appears to inhibit SpyCas9 with-
AcrIIA5, AcrIIA13).19,43,47,50 Turning this logic out directly binding the Cas protein.57
around, functional approaches for finding Acrs Our group reported a similar plasmid protection
should reveal them in loci, taxa, and environments assay to identify Acrs in human fecal, human oral,
that are inaccessible to homology-based efforts and soil microbiomes that inhibit SpyCas9.58 Like
and, in so doing, reveal previously hidden worlds Uribe and colleagues, we also noticed that Spy-
of Acr biology. Cas9 loss-of-function mutants occurred in approxi-
mately 10-4 to 10-5 clones, so added a second
round of SpyCas9 selection to enable high-
Functional selections reveal new Acrs confidence Acr identification (Figure 3(B)). This
entailed purifying plasmids that withstood SpyCas9
In 2019, two groups devised functional selections attack, re-transforming them into fresh E. coli
to identify SpyCas9 inhibitors in both human- strains, and subjecting them once more to SpyCas9
associated and environmental microbiomes, interference. In this second round of selection, the
without relying on sequence homology or the rate of SpyCas9 inactivation remained the same
presence of STSs.57,58 In both cases, the authors as before, meaning acr-encoding plasmids were
screened enormously diverse plasmid libraries for yet again enriched by a factor of
104. In theory,
Acrs, with each plasmid expressing a different 0.5 only two iterations of SpyCas9 selection are suffi-
– 5 Kb DNA insert from one of several bacteria- cient to identify just a single Acr encoded by our
rich habitats (Figure 3(A)). This metagenomic metagenomic libraries, which contained approxi-
DNA was derived from the bacteria that reside in mately 107 unique plasmids (Figure 3(B)). In this
each habitat, along with their phages, plasmids, case, theory matched reality, as >70% of clones
and other mobile genetic elements – natural targets that survived two rounds of selection were con-
of SpyCas9. Their approach exploited two key fea- firmed to inhibit SpyCas9 in validation
tures of Acrs, enabling their identification from experiments.58
libraries with more than 107 unique clones. First, The two most abundant clones from these
since Acrs are encoded by single genes, which selections were examined in detail, revealing two
move readily by horizontal gene transfer, they are new Acrs, AcrIIA11 and AcrIIA22 (the Acrs were
less likely to depend on host-specific factors.39 Sec- named at their first mechanistic description, which
ond, Acrs can confer strong fitness advantages by occurred at different times and explains their non-
protecting their hosts from CRISPR-Cas destruc-
7
K.J. Forsberg Journal of Molecular Biology xxx (xxxx) xxx

Figure 3. Functional selections can reveal Acrs. (A) Plasmid protection assays can be used to identify Acrs in
enormously diverse DNA expression libraries (for example, from metagenomes). CRISPR-Cas activity will eliminate
plasmids that lack Acrs, while Acr-containing plasmids can be revealed via their antibiotic resistance markers. Two
iterations of Acr selection can reduce false positives to low levels. (B) Hypothetical data from an Acr selection using
either an empty plasmid or a metagenomic DNA library. The first exposure of either the empty plasmid or complete
library to CRISPR-Cas attack reduces the proportion of plasmid-containing cells by
104 compared to an unexposed
control (gray bars). The retained plasmids survive primarily because of chance inactivating mutations in Cas proteins
(LOF, loss of function). In a subsequent iteration of the same selection with the same plasmids (black bars), a similar
survival rate is observed for the empty vector control. However, Acrs encoded by the metagenomic library now
dominate the plasmids in this sample, and so cells retain plasmids following CRISPR-Cas attack at a much higher
rate.

consecutive numbering).58,59 AcrIIA11 binds Cas9 Cas9-based efforts to manipulate microbiome com-
and sequesters it at non-target DNA loci to prevent position.65–69
target cleavage.58,60 This inhibitory mechanism is
also used by AcrIF9 and AcrIF24 to inhibit type I
CRISPR-Cas systems but has not been observed
Approaches for targeted Acr discovery
for any other class II Acr.60–63 The second Acr, The Clostridia are just one example where Acr
AcrIIA22, is unique because it does not bind its cog- discovery will be important for manipulating high-
nate Cas proteins, which is a rare quality among interest bacteria in high-value ecosystems.70 Future
Acrs.59 Instead of binding SpyCas9, AcrIIA22 alters discovery efforts will likely include tailored searches
the DNA topology of a target plasmid, impairing in these taxa and habitats of interest, particularly if
SpyCas9’s ability to cleave its target, likely by gaps in Acr knowledge overlap with potential techni-
exploiting SpyCas9’s preference for negatively cal applications.71 In some cases, Acrs may already
supercoiled DNA substrates.59,64 Both AcrIIA11 be known that can seed subsequent GBA searches.
and AcrIIA22 are native to prophages of Clostridia, For instance, AcrIIA11 is common in gut Clostridia
which are important bacteria in human gut micro- and is adjacent to at least ten predicted Aca pro-
biomes and potential targets for phage or teins, which could reveal additional Acrs in these
8
K.J. Forsberg Journal of Molecular Biology xxx (xxxx) xxx

bacteria.58,72 In other cases, a combination of STS- The first discovery of a bona-fide type VI Acr
based searches, empirical phage screens, and occurred in 2020, via a phage-based screen
functional selections may be sufficient to uncover performed in Listeria by Meeske et al..83 The
Acrs of interest, as was seen for the discovery of authors generated a panel of lysogens in Listeria
many Cas9 inhibitors.73–75 However, each of these seeligeri by infecting this bacterium with a series
approaches has drawbacks which can limit Acr of environmentally derived Listeria phages. They
identification. STS searches do not have the resolu- then tested whether these prophages could block
tion to reveal individual Acr loci and phage screens the activity of L. seeligeri’s type VI-A CRISPR-Cas
require a fortuitous combination of phage and system by conjugating a plasmid into each lysogen
host.19,33 Functional selections may struggle to find that contained a target of the native CRISPR locus.
scarce Acrs, which could be absent in the screening Like wild-type L. seeligeri, most lysogens restricted
library, and can only identify Acrs that express in the conjugation due to the natural CRISPR-Cas activity
selection host.57,58 Thus, alternative approaches for of this host. However, one lysogen enabled conju-
Acr discovery are needed. gation at similar rates to a Cas13-knockout strain,
To meet this need, several groups have indicating that its prophage may contain a type VI-
developed machine and deep learning strategies A Acr. Comparative genomics then revealed a puta-
for Acr prediction, revealing Acrs that inhibit type I- tive Acr locus in this phage, based on the presence
C, II-A, and VI-B CRISPR-Cas systems.76–78 As of an HTH-containing protein and a similar genetic
input, these algorithms take pre-defined Acr- organization to a different Acr-locus in a related lis-
encoding or Acr-absent datasets, which train mod- teriophage. One gene in this locus, ultimately
els to associate sequence features with the pres- named acrVIA1Lse, was necessary and sufficient
ence of Acrs. These predictive models can then for blocking Cas13-induced immunity against both
scan future queries to identify putative CRISPR- plasmids and phages. AcrVIA1Lse binds tightly to a
Cas inhibitors. An advantage of this strategy is that crRNA-loaded form of Cas13, making direct con-
it can efficiently search gigantic databases for Acrs tacts with the protein and crRNA to fully enable
that would otherwise be very difficult to predict. phage escape from type VI-A CRISPR-Cas
These lists can then be pruned for Acrs with desired systems.83
phylogenetic or environmental associations. A dis- Of note, the superscript in the name of AcrVIA1Lse
advantage of this approach is that the true accuracy was included to distinguish it from other Acrs
of these predictions is hard to define. The failure of reported to inhibit Cas13, which were described in
an Acr to validate in follow-up experiments could be a near-contemporaneous publication to that of
attributable to poor expression in the indicator host, Meeske et al.83,84 Subsequent work cast significant
the use of a non-cognate Cas protein in validation doubt on whether the Acrs from this second study
testing, or an inaccurate initial Acr call. To learn were genuine.85 Due to this uncertainty, these
more about bioinformatic approaches for Acr dis- questionable Acrs will not be discussed further in
covery, please see the review by Makarova et al. this review.
elsewhere in this special issue. Wandera et al. also identified a type VI Acr, which
was the first example of a type VI-B CRISPR-Cas
inhibitor.78 The authors used a deep learning
Type VI Acrs: a case study for targeted approach to predict Acrs in draft and complete bac-
discovery terial genomes. Compared to other machine learn-
ing prediction algorithms, their scheme identified a
Recently, multiple groups have performed higher number of type IV and type VI Acrs, of which
targeted searches to identify Acrs that inhibit Type few or no examples are known, respectively. The
VI CRISPR-Cas systems. These systems encode overrepresentation of type IV and VI Acrs may
a single effector protein, Cas13, that uses RNA result from the authors’ omission of Acr-adjacent
guides to cleave RNA targets, eliciting phage genome features when training their model, which
interference through both this target cleavage have been mostly characterized for type I and II
event and a promiscuous, secondary, trans-acting Acrs and may thus reduce the confidence of type
RNAse activity.79 Type VI CRISPR-Cas systems IV and VI predictions. Among their predictions,
hold tremendous biotechnological promise due to Wandera et al. chose to experimentally pursue sub-
Cas13’s capacity to recognize RNA in a type VI-B Acrs and tested 77 of their putative inhibi-
sequence-specific manner and then amplify this tors against three re-constituted VI-B CRISPR-Cas
signal via its secondary RNAse activity.80,81 As nat- systems using a TXTL-based assay. Of these 231
ural inhibitors of these systems, type VI Acrs have tests, just one showed strong Acr activity, though
similarly broad applications, akin to the Acrs of the degree of inhibition was nearly complete. This
Cas9 and Cas12, which are other single-effector Acr, named AcrVIB1, was found in predicted pro-
Cas nucleases and used heavily in genome edit- phages of Riemerella anatipestifer and blocked
ing.71,82 For a detailed discussion of the emerging the activity of Cas13b from Prevotella buccae. Both
applications for Acrs in therapy and technology, bacteria belong to the phylum Bacteroidota. This
please see the review by Sontheimer et al. else- singular success demonstrates the feasibility of this
where in this special issue.
9
K.J. Forsberg Journal of Molecular Biology xxx (xxxx) xxx

deep learning approach for finding novel Acrs and rate of cA4 degradation are key determinants of
the challenges associated with validating these whether an AcrIII-1 homolog acts an anti-CRISPR
predictions. or host regulatory factor.86,91 Because these data
Despite these challenges, AcrVIB1, along with are hard to infer without experimentation, many
AcrVIA1Lse, represent excellent candidates for AcrIII-1 homologs cannot be confidently classified
further Acr discovery via GBA searches. In both as true Acrs.
cases, these Acrs reside in canonical loci that Another type III Acr, AcrIIIB1, also prevents
contain putative Aca proteins. These loci are also secondary RNAse activation.87 Unlike AcrIII-1,
replete with small genes of unknown function that AcrIIIB1 binds directly to the type III CRISPR-Cas
vary across related phages. Because AcrVA1Lse effector complex to block cyclic oligonucleotide sig-
and AcrVB1 come from different bacterial phyla naling and is unlikely to be an enzyme.87,92 It was
(Firmicutes and Bacteroidota, respectively), they found in the virus SIRV2, which infects the
almost certainly participate in unique gene-sharing archaeon Sulfolobus islandicus and appears
networks. Thus, each Acr might represent a restricted to the genomes of archaea and their
separate key for unlocking many future discoveries. viruses. Bhoobalan-Chitty et al. first discovered
AcrIIIB1 because its host virus, SIRV2, could infect
a S. islandicus host with type III-B CRISPR immu-
Underrepresented Acrs nity against this virus. The gene encoding AcrIIIB1
was chosen as a candidate because it appears near
Though only two type VI Acrs are known, this a putative aca gene in a locus with a similar geno-
mirrors their presumed scarcity in nature, as type mic configuration and transcriptional profile to a
VI CRISPR-Cas systems are also rare, occurring related Acr-encoding locus in a similar virus. This
in less than 1% of bacteria and no archaea.5 In con- GBA-based prediction turned out to be accurate,
trast, type III CRISPR-Cas systems are much more as this gene, named acrIIIB1, was both necessary
abundant, appearing in approximately 12% of bac- and sufficient for type III CRISPR-Cas inhibition in
teria and 27% of archaea.5 Yet, just two type III Acrs S. islandicus.87 For a detailed discussion of type
have been discovered.86,87 As a result, very little is III Acr functions and mechanisms, see the review
known about how phages and other mobile genetic by Peng et al. elsewhere in this special issue.
elements inhibit these CRISPR-Cas systems. CRISPR-Cas systems are encoded by
Type III systems uniquely target transcribed loci, approximately 85% of archaeal genomes and yet,
as the crRNA-Cas complex of these systems only three archaeal Acrs have been identified: the
hybridizes to a nascently transcribed RNA two type III Acrs previously described and AcrID1,
product.88 After binding, Cas nucleases cleave this which was found in SIRV3, a related virus to
RNA and the non-template DNA strand, which stim- SIRV2 that also infects S. islandicus.5,93 AcrID1
ulates the production of cyclic oligonucleotide sig- was the first Acr identified in archaea and, as such,
nals that activate other, non-specific RNAses.89,90 required empirical discovery. He et al. noticed that a
These secondary RNAses trigger cell dormancy spontaneous 4 Kb deletion in SIRV3 correlated with
and, as a result, can thwart phage infection.90 One the inability of this mutant to replicate on a CRISPR-
widespread type III Acr degrades the oligonu- immune S. islandicus host. One gene in this 4 kb
cleotide signal that activates these RNAses, cyclic deletion completely explained CRISPR inhibition
tetra-adenylate (cA4), blocking some type III phenotype, which was ultimately named acrID1.
CRISPR-Cas systems.86 This enzyme was discov- Infection assays established that AcrID1 could inhi-
ered in the archaeal virus STIV and was first identi- bit the type I-D, but not type I-A, CRISPR-Cas sys-
fied as a putative Acr based on its structural tem in this archaeon. This discovery enabled the
features and capacity to promote viral replication later identification of AcrIIIB1 via GBA, based a
in some hosts. Because the enzyme’s substrate, resemblance between the AcrID1 and AcrIIIB1-
cA4, is used as a signaling molecule in multiple sub- encoding loci. AcrID1 binds directly to the Cas10
types of type III CRISPR-Cas systems, a subtype nuclease of type I-D CRISPR-Cas systems to block
signifier is omitted from this Acr’s name, AcrIII-1, their activity.93
to indicate broad activity. Other cyclic oligonu- Few archaeal and type III Acrs have been
cleotide signals, like cA6, are not degraded by identified relative to the diversity of these
AcrIII-1.86 CRISPR-Cas systems in nature, which represents
AcrIII-1 has a widespread but patchy distribution a major knowledge gap in the field of Acr biology.
across bacteria, archaea, their viruses, and their It is almost certain that many new Acr genes and
other mobile genetic elements.86 These acrIII-1 mechanisms exist that inhibit these CRISPR-Cas
genes rarely cluster with other acr or aca genes, systems. In archaea, GBA-based search
suggesting they may not always encode for dedi- strategies may be useful for finding additional
cated antagonists of CRISPR-Cas immunity. Acrs, much like the first Acrs found in bacteria set
Indeed, some have been proposed to act in concert off a wave of additional Acr discovery.19,20,23
with type III CRISPR-Cas systems, restoring cA4 to Indeed, many homologs of AcrID1 and AcrIIIB1
homeostatic levels following CRISPR-Cas activa- exist across numerous archaeal families and are
tion.86 The timing of AcrIII-1’s expression and its
10
K.J. Forsberg Journal of Molecular Biology xxx (xxxx) xxx

almost always found at the termini of viral genomes AcrIIA7, AcrIIA22, and AcrIII-1.57,59,86 Beyond these
within clusters of small, hypervariable genes of three named examples, phages can also chemically
unknown function.92 These hypervariable loci likely modify target DNA, form nucleus-like protective pro-
influence host tropism and at least one virus has tein shells, or stimulate recombination and DNA
been identified with putative type I-A Acr activity.94 break repair to overcome or block CRISPR activ-
It would not be surprising if a systematic examina- ity.97–101 The genes that mediate these varied forms
tion of the small genes in these genomic termini of CRISPR-Cas resistance are often found outside
revealed many new Acrs in archaea. canonical Acr loci but nonetheless produce an anal-
Some of these new archaeal Acrs will likely inhibit ogous effect. Almost certainly, many other non-
type III CRISPR-Cas systems, which are relatively canonical forms of CRISPR-Cas resistance exist
common in archaea.5 Identifying type III Acrs in among the world’s phages, plasmids and mobile
bacteria may be more challenging, however, as genetic elements that have yet to be discovered.
the only type III Acr known in bacteria is AcrIII-1.86 Another potential untapped reservoir of CRISPR-
Though homologs of AcrIII-1 have been found in Cas inhibitors exists outside phages, plasmids, and
many bacterial genomes, they are often host regu- mobile genetic elements, in the genomes of their
latory factors rather than Acrs.90 Furthermore, host bacteria and archaea. Indeed, there is
these genes do not co-localize with other acr or substantial benefit to bacteria and archaea to
aca genes, like has been observed for many other regulate CRISPR-Cas activity, which maximizes
Acr loci.86 Thus, although AcrIII-1 is a bona-fide their utility during infection while reducing the
Acr, it is an unattractive candidate for seeding costs associated with constitutive expression.102
GBA searches for future Acr discovery. To find As a result, these hosts have developed diverse
additional type III Acrs in bacteria, new functional strategies for CRISPR-Cas regulation, which
screens, selections, or bioinformatic approaches include controlling its activity via quorum sensing,
are needed, like occurred for the identification of the detection of membrane perturbation, the action
many type II and type VI Acrs.58,78,83 The same is of nucleoid-like proteins, via metabolic sensors or
true for type IV Acrs, which are predicted to exist other indicators of infection, and by employing ded-
but lack known examples in nature.78 The absence icated regulators.103–109 Except for lysogenic
of type IV Acrs reflects the uncertain role of type IV phages where the inhibitor clearly serves a phage
CRISPR-Cas systems in its hosts, as these sys- benefit, no Acrs are known to moonlight as
tems have been proposed to execute anti- CRISPR-Cas regulators, though this scenario has
plasmid, anti-phage, and host regulatory been proposed for AcrIII-1.44,86 Importantly, bacte-
activities.5,95 ria have co-opted Aca proteins on numerous occa-
sions, to repress Acr expression and protect their
CRISPR-Cas immune systems.10,46,110 It would
The future of Acr discovery not be surprising, therefore, if Acrs have been
repurposed by host bacteria or archaea to enable
Over the past ten years, a picture of a the post-translational regulation of CRISPR-Cas
conventional Acr has emerged: it is a small systems, though these examples may be difficult
protein, encoded by a gene in a hypervariable to detect if the regulatory Acrs lack homology to pre-
region of a phage genome, that binds to its viously described inhibitors.
cognate Cas protein to exert an inhibitory effect.74 In most host-virus conflicts, the number of viral
This picture has emerged primarily from the study inhibitors of host immunity far exceeds the
of type I, II, and V Acrs that were found based on diversity of immune defenses.111 Though the known
their similarities to the Acrs discovered before them. Acrs represent only a minute fraction of what exists
How frequently alternate strategies are used to inhi- in nature, their diversity relative to CRISPR-Cas
bit CRISPR-Cas systems remains an open ques- immune systems indicates that they are no excep-
tion, especially for underrepresented Acrs like tion to this rule.112 Thus, any weak point in
those that inhibit type III, IV, or VI systems, or those CRISPR-Cas immunity has probably been
that come from non-model microbes (e.g. archaea). exploited by some phage, somewhere. In some
Fortunately, sophisticated approaches for Acr dis- cases, the same exploit may allow phages to over-
covery now exist that may reveal these hidden Acrs come both CRISPR-Cas immunity and additional
and with them, perhaps, new paradigms.33,57,58,76– defense systems, as has been observed for Ocr,
78
a dual inhibitor of both restriction-modification and
One open frontier is the world of Acr mechanisms. BREX defenses encoded by phage T7, and for mul-
Though more than 90 unrelated Acr families have tiple counter-immune proteins in eukaryotic
been identified, multiple new modes of action are viruses.113–116 Because Acrs are the most compre-
reported each year, with many Acrs yet to be hensively studied class of phage counter-defense
characterized and untold larger numbers awaiting gene, they are perhaps the best set to use in a sys-
discovery.96 Clearly, many new molecular mecha- tematic search for dual-acting inhibitors. This dual
nisms remain. Among them are probably multiple activity may be especially useful given the proposed
ways of overcoming CRISPR-Cas immunity without synergism of unrelated defense systems for thwart-
direct Cas binding, as has already been reported for
11
K.J. Forsberg Journal of Molecular Biology xxx (xxxx) xxx

ing phage infection. Dual-acting inhibitors may be 2. Payne, L.J., Todeschini, T.C., Wu, Y., Perry, B.J.,
particularly likely for CRISPR-Cas systems Ronson, C.W., Fineran, P.C., et al., (2021). Identification
because they share the same DNA and RNA sub- and classification of antiviral defence systems in bacteria
strates as many other defenses.117–119 For a more and archaea with PADLOC reveals new system types.
detailed discussion of the ecology and evolution of Nucleic Acids Res. 49 (19), 10868–10878.
Acrs, please see the review by van Houte and Wes- 3. Tesson, F., Hervé, A., Mordret, E., Touchon, M.,
tra elsewhere in this special issue. d’Humières, C., Cury, J., et al., (2022). Systematic and
quantitative view of the antiviral arsenal of prokaryotes.
As the field of anti-CRISPR biology steps into its
Nat. Commun. 13 (1), 2561.
second decade, many fundamental questions
4. Srikant, S., Guegler, C.K., Laub, M.T., (2022). The
remain. Efforts to catalog underrepresented Acrs,
evolution of a counter-defense mechanism in a virus
like those that inhibit type III, IV, or VI systems are constrains its host range. Elife 11, e79549.
just beginning. We have much to learn about the 5. Makarova, K.S., Wolf, Y.I., Iranzo, J., Shmakov, S.A.,
role Acrs play in real-world infections, their Alkhnbashi, O.S., Brouns, S.J.J., et al., (2020).
distributions in natural communities, and their Evolutionary classification of CRISPR-Cas systems: a
diversity of mechanisms. New and improved burst of class 2 and derived variants. Nat. Rev. Microbiol.
methods of Acr discovery will help fill in these 18 (2), 67–83.
knowledge gaps, complementing the tried-and- 6. Barrangou, R., Fremaux, C., Deveau, H., Richards, M.,
true approaches that have served the field in its Boyaval, P., Moineau, S., et al., (2007). CRISPR provides
first ten years. acquired resistance against viruses in prokaryotes.
Science 315 (5819), 1709–1712.
CRediT authorship contribution statement 7. Marraffini, L.A., Sontheimer, E.J., (2008). CRISPR
interference limits horizontal gene transfer in
Kevin J. Forsberg: Conceptualization, Writing – staphylococci by targeting DNA. Science 322 (5909),
original draft, Writing – review & editing, 1843–1845.
Visualization, Supervision, Project administration, 8. Zhang, Q., Rho, M., Tang, H., Doak, T.G., Ye, Y., (2013).
Funding acquisition. CRISPR-Cas systems target a diverse collection of
invasive mobile genetic elements in human
microbiomes. Genome Biol. 14 (4), R40.
9. van Houte, S., Ekroth, A.K., Broniewski, J.M., Chabas, H.,
Ashby, B., Bondy-Denomy, J., et al., (2016). The diversity-
Acknowledgements generating benefits of a prokaryotic adaptive immune
KJF was supported by the Searle Scholars system. Nature 532 (7599), 385–388.
Program, the HHMI Emerging Pathogens 10. Li, Y., Bondy-Denomy, J., (2021). Anti-CRISPRs go viral:
The infection biology of CRISPR-Cas inhibitors. Cell Host
Initiative, the Endowed Scholars Program at the
Microbe 29 (5), 704–714.
University of Texas Southwestern Medical Center,
11. Makarova, K.S., Wolf, Y.I., Alkhnbashi, O.S., Costa, F.,
and by NIH grant DP2-AI154402.
Shah, S.A., Saunders, S.J., et al., (2015). An updated
evolutionary classification of CRISPR-Cas systems. Nat.
Declaration of Competing Interest Rev. Microbiol. 13 (11), 722–736.
12. Makarova, K.S., Haft, D.H., Barrangou, R., Brouns, S.J.J.,
The authors declare that they have no known Charpentier, E., Horvath, P., et al., (2011). Evolution and
competing financial interests or personal classification of the CRISPR–Cas systems. Nat. Rev.
relationships that could have appeared to Microbiol. 9 (6), 467–477.
influence the work reported in this paper. 13. Koonin, E.V., Makarova, K.S., (2019). Origins and
evolution of CRISPR-Cas systems. Philos. Trans. R.
Received 1 November 2022; Soc. Lond. B Biol. Sci. 374 (1772), 20180087.
Accepted 3 January 2023; 14. Wang, J.Y., Pausch, P., Doudna, J.A., (2022). Structural
Available online xxxx biology of CRISPR–Cas immunity and genome editing
enzymes. Nat. Rev. Microbiol. 20 (11), 641–656.
Keywords: 15. Pickar-Oliver, A., Gersbach, C.A., (2019). The next
anti-CRISPRs; generation of CRISPR–Cas technologies and
CRISPR-Cas; applications. Nat. Rev. Mol. Cell Biol. 20 (8), 490–507.
bacteriophage; 16. Rath, J., (2018). Safety and Security Risks of CRISPR/
bacterial defense systems; Cas9. In: Schroeder, D., Cook, J., Hirsch, F., Fenet, S.,
Muthuswamy, V. (Eds.), Ethics Dumping: Case Studies
microbial arms race
from North-South Research Collaborations. Springer
International Publishing, Cham, pp. 107–113.
17. Shivram, H., Cress, B.F., Knott, G.J., Doudna, J.A.,
(2021). Controlling and enhancing CRISPR systems.
References Nat. Chem. Biol. 17 (1), 10–19.
18. Noble, C., Adlam, B., Church, G.M., Esvelt, K.M., Nowak,
1. Rostol, J.T., Marraffini, L., (2019). (Ph)ighting Phages: M.A., (2018). Current CRISPR gene drive systems are
How Bacteria Resist Their Parasites. Cell Host Microbe 25 likely to be highly invasive in wild populations. Elife 7
(2), 184–194.
12
K.J. Forsberg Journal of Molecular Biology xxx (xxxx) xxx

19. Bondy-Denomy, J., Pawluk, A., Maxwell, K.L., Davidson, enzymatic inhibition of CRISPR-Cas12a. Nat. Struct. Mol.
A.R., (2013). Bacteriophage genes that inactivate the Biol. 26 (4), 315–321.
CRISPR/Cas bacterial immune system. Nature 493 37. Dong, L., Guan, X., Li, N., Zhang, F., Zhu, Y., Ren, K.,
(7432), 429–432. et al., (2019). An anti-CRISPR protein disables type V
20. Pawluk, A., Bondy-Denomy, J., Cheung, V.H., Maxwell, K. Cas12a by acetylation. Nat. Struct. Mol. Biol. 26 (4), 308–
L., Davidson, A.R., (2014). A new group of phage anti- 314.
CRISPR genes inhibits the type I-E CRISPR-Cas system 38. Touchon, M., Bernheim, A., Rocha, E.P., (2016).
of Pseudomonas aeruginosa. MBio 5 (2), e00896. Genetic and life-history traits associated with the
21. Stanley, S.Y., Borges, A.L., Chen, K.-H., Swaney, D.L., distribution of prophages in bacteria. ISME J. 10 (11),
Krogan, N.J., Bondy-Denomy, J., et al., (2019). Anti- 2744–2754.
CRISPR-Associated Proteins Are Crucial Repressors of 39. Marshall, R., Maxwell, C.S., Collins, S.P., Jacobsen, T.,
Anti-CRISPR Transcription. Cell 178 (6) 1452-64.e13. Luo, M.L., Begemann, M.B., et al., (2018). Rapid and
22. Borges, A.L., Davidson, A.R., Bondy-Denomy, J., (2017). Scalable Characterization of CRISPR Technologies Using
The Discovery, Mechanisms, and Evolutionary Impact of an E. coli Cell-Free Transcription-Translation System.
Anti-CRISPRs. Annu. Rev. Virol. 4 (1), 37–59. Mol. Cell 69 (1) 146-57 e3.
23. Pawluk, A., Staals, R.H., Taylor, C., Watson, B.N., Saha, 40. Wandera, K.G., Collins, S.P., Wimmer, F., Marshall, R.,
S., Fineran, P.C., et al., (2016). Inactivation of CRISPR- Noireaux, V., Beisel, C.L., (2020). An enhanced assay to
Cas systems by anti-CRISPR proteins in diverse bacterial characterize anti-CRISPR proteins using a cell-free
species. Nat. Microbiol. 1 (8), 16085. transcription-translation system. Methods 172, 42–50.
24. Birkholz, N., Fagerlund, R.D., Smith, L.M., Jackson, S.A., 41. León, L.M., Park, A.E., Borges, A.L., Zhang, J.Y., Bondy-
Fineran, P.C., (2019). The autoregulator Aca2 mediates Denomy, J., (2021). Mobile element warfare via CRISPR
anti-CRISPR repression. Nucleic Acids Res. 47 (18), and anti-CRISPR in Pseudomonas aeruginosa. Nucleic
9658–9665. Acids Res. 49 (4), 2114–2125.
25. Hatfull, G.F., Hendrix, R.W., (2011). Bacteriophages and 42. Knott, G.J., Doudna, J.A., (2018). CRISPR-Cas guides
their genomes. Curr. Opin. Virol. 1 (4), 298–303. the future of genetic engineering. Science 361 (6405),
26. Pawluk, A., Amrani, N., Zhang, Y., Garcia, B., Hidalgo- 866–869.
Reyes, Y., Lee, J., et al., (2016). Naturally Occurring Off- 43. Rauch, B.J., Silvis, M.R., Hultquist, J.F., Waters, C.S.,
Switches for CRISPR-Cas9. Cell 167 (7) 1829-38 e9. McGregor, M.J., Krogan, N.J., et al., (2017). Inhibition of
27. Lee, J., Mir, A., Edraki, A., Garcia, B., Amrani, N., Lou, H. CRISPR-Cas9 with Bacteriophage Proteins. Cell 168 (1–
E., et al., (2018). Potent Cas9 Inhibition in Bacterial and 2) 150-8 e10.
Human Cells by AcrIIC4 and AcrIIC5 Anti-CRISPR 44. Osuna, B.A., Karambelkar, S., Mahendra, C., Christie, K.
Proteins. MBio 9 (6) A., Garcia, B., Davidson, A.R., et al., (2020). Listeria
28. Khan, A.N., (2021). Characterizing a Novel Type II-C Anti- Phages Induce Cas9 Degradation to Protect Lysogenic
CRISPR, AcrIIC6. University of Toronto, Toronto, ON, CA. Genomes. Cell Host Microbe 28 (1) 31-40 e9.
29. Marino, N.D., Zhang, J.Y., Borges, A.L., Sousa, A.A., 45. Mahendra, C., Christie, K.A., Osuna, B.A., Pinilla-
Leon, L.M., Rauch, B.J., et al., (2018). Discovery of Redondo, R., Kleinstiver, B.P., Bondy-Denomy, J.,
widespread Type I and Type V CRISPR-Cas inhibitors. (2020). Broad-spectrum anti-CRISPR proteins facilitate
Science. horizontal gene transfer. Nat. Microbiol. 5 (4), 620–629.
30. Pinilla-Redondo, R., Shehreen, S., Marino, N.D., 46. Osuna, B.A., Karambelkar, S., Mahendra, C., Sarbach,
Fagerlund, R.D., Brown, C.M., Sørensen, S.J., et al., A., Johnson, M.C., Kilcher, S., et al., (2020). Critical Anti-
(2020). Discovery of multiple anti-CRISPRs highlights CRISPR Locus Repression by a Bi-functional Cas9
anti-defense gene clustering in mobile genetic elements. Inhibitor. Cell Host Microbe 28 (1) 23-30.e5.
Nat. Commun. 11 (1), 5652. 47. Watters, K.E., Shivram, H., Fellmann, C., Lew, R.J.,
31. Bernheim, A., Sorek, R., (2020). The pan-immune system McMahon, B., Doudna, J.A., (2020). Potent CRISPR-
of bacteria: antiviral defence as a community resource. Cas9 inhibitors from Staphylococcus genomes. PNAS
Nat. Rev. Microbiol. 18 (2), 113–119. 117 (12), 6531–6539.
32. Zetsche, B., Gootenberg, J.S., Abudayyeh, O.O., 48. Wang, X., Li, X., Ma, Y., He, J., Liu, X., Yu, G., et al.,
Slaymaker, I.M., Makarova, K.S., Essletzbichler, P., (2022). Inhibition mechanisms of CRISPR-Cas9 by
et al., (2015). Cpf1 is a single RNA-guided AcrIIA17 and AcrIIA18. Nucleic Acids Res. 50 (1), 512–
endonuclease of a class 2 CRISPR-Cas system. Cell 521.
163 (3), 759–771. 49. Karvelis, T., Gasiunas, G., Miksys, A., Barrangou, R.,
33. Watters, K.E., Fellmann, C., Bai, H.B., Ren, S.M., Horvath, P., Siksnys, V., (2013). crRNA and tracrRNA
Doudna, J.A., (2018). Systematic discovery of natural guide Cas9-mediated DNA interference in Streptococcus
CRISPR-Cas12a inhibitors. Science 362 (6411), 236– thermophilus. RNA Biol. 10 (5), 841–851.
239. 50. Hynes, A.P., Rousseau, G.M., Lemay, M.L., Horvath, P.,
34. Zhang, H., Li, Z., Daczkowski, C.M., Gabel, C., Mesecar, Romero, D.A., Fremaux, C., et al., (2017). An anti-
A.D., Chang, L., (2019). Structural Basis for the Inhibition CRISPR from a virulent streptococcal phage inhibits
of CRISPR-Cas12a by Anti-CRISPR Proteins. Cell Host Streptococcus pyogenes Cas9. Nat. Microbiol..
Microbe 25 (6) 815-26 e4. 51. Hynes, A.P., Rousseau, G.M., Agudelo, D., Goulet, A.,
35. Knott, G.J., Cress, B.F., Liu, J.J., Thornton, B.W., Lew, R. Amigues, B., Loehr, J., et al., (2018). Widespread anti-
J., Al-Shayeb, B., et al., (2019). Structural basis for CRISPR proteins in virulent bacteriophages inhibit a
AcrVA4 inhibition of specific CRISPR-Cas12a. Elife 8 range of Cas9 proteins. Nat. Commun. 9 (1), 2919.
36. Knott, G.J., Thornton, B.W., Lobba, M.J., Liu, J.J., Al- 52. Garcia, B., Lee, J., Edraki, A., Hidalgo-Reyes, Y., Erwood,
Shayeb, B., Watters, K.E., et al., (2019). Broad-spectrum S., Mir, A., et al., (2019). Anti-CRISPR AcrIIA5 Potently

13
K.J. Forsberg Journal of Molecular Biology xxx (xxxx) xxx

Inhibits All Cas9 Homologs Used for Genome Editing. Cell 67. Sorbara, M.T., Littmann, E.R., Fontana, E., Moody, T.U.,
Rep. 29 (7) 1739-46 e5. Kohout, C.E., Gjonbalaj, M., et al., (2020). Functional and
53. Song, G., Zhang, F., Zhang, X., Gao, X., Zhu, X., Fan, D., Genomic Variation between Human-Derived Isolates of
et al., (2019). AcrIIA5 Inhibits a Broad Range of Cas9 Lachnospiraceae Reveals Inter- and Intra-Species
Orthologs by Preventing DNA Target Cleavage. Cell Rep. Diversity. Cell Host Microbe 28 (1) 134-46.e4.
29 (9) 2579-89 e4. 68. Sheth, R.U., Cabral, V., Chen, S.P., Wang, H.H., (2016).
54. Song, G., Zhang, F., Tian, C., Gao, X., Zhu, X., Fan, D., Manipulating Bacterial Communities by in situ Microbiome
et al., (2022). Discovery of potent and versatile CRISPR– Engineering. Trends Genet. 32 (4), 189–200.
Cas9 inhibitors engineered for chemically controllable 69. Selle, K., Fletcher Joshua, R., Tuson, H., Schmitt Daniel,
genome editing. Nucleic Acids Res. 50 (5), 2836–2853. S., McMillan, L., Vridhambal Gowrinarayani, S., et al.,
55. Vorontsova, D., Datsenko, K.A., Medvedeva, S., Bondy- (2020). In Vivo Targeting of Clostridioides difficile Using
Denomy, J., Savitskaya, E.E., Pougach, K., et al., (2015). Phage-Delivered CRISPR-Cas3 Antimicrobials. MBio 11
Foreign DNA acquisition by the I-F CRISPR-Cas system (2) e00019-20.
requires all components of the interference machinery. 70. Romero, D.A., Magill, D., Millen, A., Horvath, P., Fremaux,
Nucleic Acids Res. 43 (22), 10848–10860. C., (2020). Dairy lactococcal and streptococcal phage–
56. Philippe, C., Morency, C., Plante, P.-L., Zufferey, E., host interactions: an industrial perspective in an
Achigar, R., Tremblay, D.M., et al., (2022). A truncated evolving phage landscape. FEMS Microbiol. Rev. 44 (6),
anti-CRISPR protein prevents spacer acquisition but not 909–932.
interference. Nat. Commun. 13 (1), 2802. 71. Marino, N.D., Pinilla-Redondo, R., Csorgo, B., Bondy-
57. Uribe, R.V., van der Helm, E., Misiakou, M.A., Lee, S.W., Denomy, J., (2020). Anti-CRISPR protein applications:
Kol, S., Sommer, M.O.A., (2019). Discovery and natural brakes for CRISPR-Cas technologies. Nat.
Characterization of Cas9 Inhibitors Disseminated across Methods 17 (5), 471–479.
Seven Bacterial Phyla. Cell Host Microbe 25 (2) 233-41 72. Yin, Y., Yang, B., Entwistle, S., (2019). Bioinformatics
e5. Identification of Anti-CRISPR Loci by Using Homology,
58. Forsberg, K.J., Bhatt, I.V., Schmidtke, D.T., Javanmardi, Guilt-by-Association, and CRISPR Self-Targeting Spacer
K., Dillard, K.E., Stoddard, B.L., et al., (2019). Functional Approaches. mSystems. 4 (5) e00455-19.
metagenomics-guided discovery of potent Cas9 inhibitors 73. Pawluk, A., Davidson, A.R., Maxwell, K.L., (2018). Anti-
in the human microbiome. Elife 8 CRISPR: discovery, mechanism and function. Nat. Rev.
59. Forsberg, K.J., Schmidtke, D.T., Werther, R., Uribe, R.V., Microbiol. 16 (1), 12–17.
Hausman, D., Sommer, M.O.A., et al., (2021). The novel 74. Davidson, A.R., Lu, W.-T., Stanley, S.Y., Wang, J.,
anti-CRISPR AcrIIA22 relieves DNA torsion in target Mejdani, M., Trost, C.N., et al., (2020). Anti-CRISPRs:
plasmids and impairs SpyCas9 activity. PLoS Biol. 19 Protein Inhibitors of CRISPR-Cas Systems. Annu. Rev.
(10), e3001428. Biochem. 89 (1), 309–332.
60. K.E. Dillard, C. Terrace, K. Javanmardi, W. Kim, K.J. 75. Varble, A., Campisi, E., Euler, C.W., Maguin, P., Kozlova,
Forsberg, I.J. Finkelstein, Mechanism of broad-spectrum A., Fyodorova, J., et al., (2021). Prophage integration into
Cas9 inhibition by AcrIIA11. bioRxiv. CRISPR loci enables evasion of antiviral immunity in
2021:2021.09.15.460536. Streptococcus pyogenes. Nat. Microbiol. 6 (12), 1516–
61. Hirschi, M., Lu, W.T., Santiago-Frangos, A., Wilkinson, R., 1525.
Golden, S.M., Davidson, A.R., et al., (2020). AcrIF9 76. Eitzinger, S., Asif, A., Watters, K.E., Iavarone, A.T., Knott,
tethers non-sequence specific dsDNA to the CRISPR G.J., Doudna, J.A., et al., (2020). Machine learning
RNA-guided surveillance complex. Nat. Commun. 11 (1), predicts new anti-CRISPR proteins. Nucleic Acids Res.
2730. 48 (9), 4698–4708.
62. Yang, L., Zhang, L., Yin, P., Ding, H., Xiao, Y., Zeng, J., 77. Gussow, A.B., Park, A.E., Borges, A.L., Shmakov, S.A.,
et al., (2022). Insights into the inhibition of type I-F Makarova, K.S., Wolf, Y.I., et al., (2020). Machine-
CRISPR-Cas system by a multifunctional anti-CRISPR learning approach expands the repertoire of anti-
protein AcrIF24. Nat. Commun. 13 (1), 1931. CRISPR protein families. Nat. Commun. 11 (1), 3784.
63. Mukherjee, I.A., Gabel, C., Noinaj, N., Bondy-Denomy, J., 78. Wandera, K.G., Alkhnbashi, O.S., Bassett, H.v.I.,
Chang, L., (2022). Structural basis of AcrIF24 as an anti- Mitrofanov, A., Hauns, S., Migur, A., et al., (2022). Anti-
CRISPR protein and transcriptional suppressor. Nat. CRISPR prediction using deep learning reveals an
Chem. Biol.. inhibitor of Cas13b nucleases. Mol. Cell 82 (14) 2714-
64. Szczelkun, M.D., Tikhomirova, M.S., Sinkunas, T., 26.e4.
Gasiunas, G., Karvelis, T., Pschera, P., et al., (2014). 79. O’Connell, M.R., (2019). Molecular Mechanisms of RNA
Direct observation of R-loop formation by single RNA- Targeting by Cas13-containing Type VI CRISPR–Cas
guided Cas9 and Cascade effector complexes. PNAS 111 Systems. J. Mol. Biol. 431 (1), 66–87.
(27), 9798–9803. 80. Cox, D.B.T., Gootenberg, J.S., Abudayyeh, O.O.,
65. Džunková, M., Low, S.J., Daly, J.N., Deng, L., Rinke, C., Franklin, B., Kellner, M.J., Joung, J., et al., (2017). RNA
Hugenholtz, P., (2019). Defining the human gut host– editing with CRISPR-Cas13. Science 358 (6366), 1019–
phage network through single-cell viral tagging. Nat. 1027.
Microbiol. 4 (12), 2192–2203. 81. Abudayyeh, O.O., Gootenberg, J.S., (2021). CRISPR
66. Heuler, J., Fortier, L.-C., Sun, X., (2021). Clostridioides diagnostics. Science 372 (6545), 914–915.
difficile phage biology and application. FEMS Microbiol. 82. J. Guan, A. Oromı́-Bosch, S.D. Mendoza, S.
Rev. 45 (5), fuab012. Karambelkar, J. Berry, J. Bondy-Denomy, RNA targeting

14
K.J. Forsberg Journal of Molecular Biology xxx (xxxx) xxx

with CRISPR-Cas13a facilitates bacteriophage genome 98. Malone, L.M., Warring, S.L., Jackson, S.A., Warnecke,
engineering. bioRxiv. 2022:2022.02.14.480438. C., Gardner, P.P., Gumy, L.F., et al., (2020). A jumbo
83. Meeske, A.J., Jia, N., Cassel, A.K., Kozlova, A., Liao, J., phage that forms a nucleus-like structure evades
Wiedmann, M., et al., (2020). A phage-encoded anti- CRISPR–Cas DNA targeting but is vulnerable to type III
CRISPR enables complete evasion of type VI-A CRISPR- RNA-based immunity. Nat. Microbiol. 5 (1), 48–55.
Cas immunity. Science 369 (6499), 54–59. 99. Bryson, A.L., Hwang, Y., Sherrill-Mix, S., Wu, G.D., Lewis,
84. Lin, P., Qin, S., Pu, Q., Wang, Z., Wu, Q., Gao, P., et al., J.D., Black, L., et al., (2015). Covalent Modification of
(2020). CRISPR-Cas13 Inhibitors Block RNA Editing in Bacteriophage T4 DNA Inhibits CRISPR-Cas9. mBio. 6
Bacteria and Mammalian Cells. Mol. Cell 78 (5) 850-61 (3), e00648.
e5. 100. Roy, D., Huguet, K.T., Grenier, F., Burrus, V., (2020).
85. Johnson, M.C., Hille, L.T., Kleinstiver, B.P., Meeske, A.J., IncC conjugative plasmids and SXT/R391 elements repair
Bondy-Denomy, J., (2022). Lack of Cas13a inhibition by double-strand breaks caused by CRISPR–Cas during
anti-CRISPR proteins from Leptotrichia prophages. Mol. conjugation. Nucleic Acids Res..
Cell 82 (11) 2161-6.e3. 101. Hossain, A.A., McGinn, J., Meeske, A.J., Modell, J.W.,
86. Athukoralage, J.S., McMahon, S.A., Zhang, C., Marraffini, L.A., (2021). Viral recombination systems limit
Gruschow, S., Graham, S., Krupovic, M., et al., (2020). CRISPR-Cas targeting through the generation of escape
An anti-CRISPR viral ring nuclease subverts type III mutations. Cell Host Microbe 29 (10) 1482-95.e12.
CRISPR immunity. Nature 577 (7791), 572–575. 102. Westra Edze, R., van Houte, S., Oyesiku-Blakemore, S.,
87. Bhoobalan-Chitty, Y., Johansen, T.B., Di Cianni, N., Makin, B., Broniewski Jenny, M., Best, A., et al., (2015).
Peng, X., (2019). Inhibition of Type III CRISPR-Cas Parasite Exposure Drives Selective Evolution of
Immunity by an Archaeal Virus-Encoded Anti-CRISPR Constitutive versus Inducible Defense. Curr. Biol. 25 (8),
Protein. Cell 179 (2) 448-58 e11. 1043–1049.
88. Kolesnik, M.V., Fedorova, I., Karneyeva, K.A., 103. Patterson, A.G., Jackson, S.A., Taylor, C., Evans, G.B.,
Artamonova, D.N., Severinov, K.V., (2021). Type III Salmond, G.P.C., Przybilski, R., et al., (2016). Quorum
CRISPR-Cas Systems: Deciphering the Most Complex Sensing Controls Adaptive Immunity through the
Prokaryotic Immune System. Biochem. Mosc. 86 (10), Regulation of Multiple CRISPR-Cas Systems. Mol. Cell
1301–1314. 64 (6), 1102–1108.
89. Tamulaitis, G., Kazlauskiene, M., Manakova, E., 104. Hoyland-Kroghsbo, N.M., Paczkowski, J., Mukherjee, S.,
Venclovas, Č., Nwokeoji Alison, O., Dickman Mark, J., Broniewski, J., Westra, E., Bondy-Denomy, J., et al.,
et al., (2014). Programmable RNA Shredding by the Type (2017). Quorum sensing controls the Pseudomonas
III-A CRISPR-Cas System of Streptococcus thermophilus. aeruginosa CRISPR-Cas adaptive immune system.
Mol. Cell 56 (4), 506–517. PNAS 114 (1), 131–135.
90. Athukoralage, J.-A.-O., White, M.-A.-O., (2021). Cyclic 105. Perez-Rodriguez, R., Haitjema, C., Huang, Q., Nam, K.H.,
oligoadenylate signalling and regulation by ring nucleases Bernardis, S., Ke, A., et al., (2011). Envelope stress is a
during type III CRISPR defence. RNA 27, 855–867. trigger of CRISPR RNA-mediated DNA silencing in
91. Athukoralage, J.S., White, M.F., (2022). Cyclic Nucleotide Escherichia coli. Mol. Microbiol. 79 (3), 584–599.
Signaling in Phage Defense and Counter-Defense. Ann. 106. Pul, Ü., Wurm, R., Arslan, Z., Geißen, R., Hofmann, N.,
Rev. Virol. 9 (1), 451–468. Wagner, R., (2010). Identification and characterization of
92. Peng, X., Mayo-Muñoz, D., Bhoobalan-Chitty, Y., E. coli CRISPR-cas promoters and their silencing by H-
Martı́nez-Álvarez, L., (2020). Anti-CRISPR Proteins in NS. Mol. Microbiol. 75 (6), 1495–1512.
Archaea. Trends Microbiol. 28 (11), 913–921. 107. Patterson, A.G., Chang, J.T., Taylor, C., Fineran, P.C.,
93. He, F., Bhoobalan-Chitty, Y., Van, L.B., Kjeldsen, A.L., (2015). Regulation of the Type I-F CRISPR-Cas
Dedola, M., Makarova, K.S., et al., (2018). Anti-CRISPR system by CRP-cAMP and GalM controls spacer
proteins encoded by archaeal lytic viruses inhibit subtype acquisition and interference. Nucleic Acids Res. 43 (12),
I-D immunity. Nat. Microbiol. 3 (4), 461–469. 6038–6048.
94. Guo, T., Han, W., She, Q., (2019). Tolerance of 108. Smargon, A.A., Cox, D.B., Pyzocha, N.K., Zheng, K.,
Sulfolobus SMV1 virus to the immunity of I-A and III-B Slaymaker, I.M., Gootenberg, J.S., et al., (2017). Cas13b
CRISPR-Cas systems in Sulfolobus islandicus. RNA Biol. Is a Type VI-B CRISPR-Associated RNA-Guided RNase
16 (4), 549–556. Differentially Regulated by Accessory Proteins Csx27 and
95. Pinilla-Redondo, R., Mayo-Muñoz, D., Russel, J., Garrett, Csx28. Mol. Cell 65 (4) 618-30 e7.
R.A., Randau, L., Sørensen, S.J., et al., (2020). Type IV 109. Patterson, A.G., Yevstigneyeva, M.S., Fineran, P.C.,
CRISPR–Cas systems are highly diverse and involved in (2017). Regulation of CRISPR–Cas adaptive immune
competition between plasmids. Nucleic Acids Res. 48 (4), systems. Curr. Opin. Microbiol. 37, 1–7.
2000–2012. 110. Shehreen, S., Birkholz, N., Fineran Peter, C., Brown, C.
96. Jia, N., Patel, D.J., (2021). Structure-based functional M., (2022). Widespread repression of anti-CRISPR
mechanisms and biotechnology applications of anti- production by anti-CRISPR-associated proteins. Nucleic
CRISPR proteins. Nat. Rev. Mol. Cell Biol. 22 (8), 563– Acids Res. 50 (15), 8615–8625.
579. 111. Daugherty, M.D., Malik, H.S., (2012). Rules of
97. Mendoza, S.D., Nieweglowska, E.S., Govindarajan, S., engagement: molecular insights from host-virus arms
Leon, L.M., Berry, J.D., Tiwari, A., et al., (2020). A races. Annu. Rev. Genet. 46, 677–700.
bacteriophage nucleus-like compartment shields DNA 112. Bondy-Denomy, J., Davidson, A.R., Doudna, J.A.,
from CRISPR nucleases. Nature 577 (7789), 244–248. Fineran, P.C., Maxwell, K.L., Moineau, S., et al., (2018).

15
K.J. Forsberg Journal of Molecular Biology xxx (xxxx) xxx

A Unified Resource for Tracking Anti-CRISPR Names. apoptosis by human cytomegalovirus protein UL37x1
CRISPR J. 1 (5), 304–305. enables efficient virus replication. Nat. Microbiol. 7 (7),
113. Roberts, G.A., Stephanou, A.S., Kanwar, N., Dawson, A., 1041–1053.
Cooper, L.P., Chen, K., et al., (2012). Exploring the DNA 117. Ofir, G., Melamed, S., Sberro, H., Mukamel, Z.,
mimicry of the Ocr protein of phage T7. Nucleic Acids Silverman, S., Yaakov, G., et al., (2018). DISARM is a
Res. 40 (16), 8129–8143. widespread bacterial defence system with broad anti-
114. Isaev, A., Drobiazko, A., Sierro, N., Gordeeva, J., Yosef, phage activities. Nat. Microbiol. 3 (1), 90–98.
I., Qimron, U., et al., (2020). Phage T7 DNA mimic protein 118. Egido, J.E., Costa, A.R., Aparicio-Maldonado, C., Haas,
Ocr is a potent inhibitor of BREX defence. Nucleic Acids P.-J., Brouns, S.J.J., (2022). Mechanisms and clinical
Res. 48 (10), 5397–5406. importance of bacteriophage resistance. FEMS Microbiol.
115. Chatterjee, S., Basler, C.F., Amarasinghe, G.K., Leung, Rev. 46 (1), fuab048.
D.W., (2016). Molecular Mechanisms of Innate Immune 119. M.C. Williams, A.E. Reker, S.R. Margolis, J. Liao, M.
Inhibition by Non-Segmented Negative-Sense RNA Wiedmann, E.R. Rojas, et al. Phage genome cleavage
Viruses. J. Mol. Biol. 428 (17), 3467–3482. enables resuscitation from Cas13-induced bacterial
116. Ren, Y., Wang, A., Wu, D., Wang, C., Huang, M., Xiong, dormancy. bioRxiv. 2022:2022.07.05.498905.
X., et al., (2022). Dual inhibition of innate immunity and

16