CdpR Inhibits CRISPR-Cas Adaptive Immunity to Lower Anti-viral Defense while Avoiding Self-Reactivity.

CRISPR-Cas systems as adaptive immunity in bacteria and archaea battle against bacteriophages. However, little is known how CRISPR-Cas systems are precisely regulated to effectively eliminate intruders while not inducing self-reactivity. Here, we identify intrinsic negative modulator of CRISPR-Cas that influences interference and adaptation functions. LasI/RhlI-derived autoinducers activate cas operon by enhancing the binding of virulence factor regulator (Vfr) cis-response elements to cas1 promoter, whereas CdpR represses this intracellular signaling and blocks transcription of cas operon. Importantly, inhibition of Vfr reduces cas1 expression and impairs immunization and immune memory mediated by CRISPR-Cas, leading to more severe phage infection but lower self-targeting activities. In addition, CdpR-mediated LasI/RhlI/Vfr intracellular signaling represses cleavage of bacterial endogenous sequences by impeding Cas3 RNA cleavage activity. Thus, CdpR renders important inhibitory effects on CRISPR-Cas systems to avoid possible self-reactivity but potentially heightening infection risk. Our study provides insight into fine regulation of CRISPR-Cas systems for maintaining homeostasis.

Introduction

Bacteria have evolved multiple defense strategies to resist bacteriophage infection (Koonin et al., 2017, Labrie et al., 2010, Mohanraju et al., 2016). The clustered regularly interspaced short palindromic repeats (CRISPR) and their CRISPR-associated (Cas) systems are the first identified and only adaptive immunity against the foreign invaders and mobile genetic elements (MGEs) via cas genes and CRISPR arrays (Barrangou et al., 2007, Marraffini, 2015, Marraffini and Sontheimer, 2008). The CRISPR arrays consist of DNA remnants from foreign invaders (mostly from phages) to generate CRISPR RNAs (crRNAs) that target nucleic acids in a sequence-specific manner (Garneau et al., 2010). Cas proteins play a critical role in mediating the acquisition of foreign sequences into a CRISPR array (adaptation or immunization) (Heler et al., 2015, McGinn and Marraffini, 2016), facilitating the maturation of crRNAs (Deltcheva et al., 2011), and counteracting invasion of MGEs, DNA (Fonfara et al., 2016), or RNA (East-Seletsky et al., 2016). Both immunization and immunity processes require activation of CRISPR-Cas systems. Currently, two distinct classes of CRISPR-Cas systems have been identified, which are further divided into a series of subtypes based on their distinct Cas effector machineries with substantial differences in targeting mechanisms (Lewis and Ke, 2017, Makarova et al., 2015). New CRISPR-Cas systems have been continuously discovered (Burstein et al., 2017, Smargon et al., 2017). The current understanding of the adaptive immunity is that CRISPR-Cas systems enable bacteria to distinguish nucleic acids between self and foreign sources, relying on the recognition of spacers and protein-mediated protospacer adjacent motif (PAM) to avoid autoimmunity (Hayes et al., 2016, Rollins et al., 2015, Westra et al., 2012, Westra et al., 2013). CRISPR-Cas systems are important for adaptive immunity for bacteria or archaea to survive in adverse environments by combatting numerous phages; however, many intriguing questions remain to be answered (Ledford, 2017). For instance, how do bacteria regulate CRISPR-Cas systems to shape and balance host defense and homeostasis?

To effectively defend against phages or MGEs, bacterial CRISPR-Cas systems rapidly evolved through horizontal transfer of complete loci or individual modules, resulting in functional diversity (Mohanraju et al., 2016). To promote invasive potency, phages also produce inhibitors to enhance the ability to lyse host bacterium or effectively integrate into bacterial genomes (Mohanraju et al., 2016, Samson et al., 2013). Studies revealed that phages encode proteins to inhibit or directly interact with different Cas proteins to prevent the functionality of CRISPR-Cas systems (Bondy-Denomy et al., 2015, Rauch et al., 2017, Sontheimer and Davidson, 2017). However, little is presently known about whether CRISPR-Cas systems can be regulated by bacterial own genes.

Quorum sensing (QS) is known not only to govern bacterial virulence but also to regulate communication between bacterial cells and organize collective behaviors in bacterial populations (Papenfort and Bassler, 2016). Recently, QS signaling was found to mediate the expression and activity of multiple CRISPR-Cas systems (Høyland-Kroghsbo et al., 2017, Patterson et al., 2016). These QS effects on prokaryotic adaptive immune systems are strongly associated with cell density, because increased diversity of CRISPR spacers within communities restricts the success of phage escape mutants (van Houte et al., 2016). Modulating CRISPR-Cas immunity regulated by QS opens up a question of how bacterial signaling controls the CRISPR-Cas system, but how bacterial genes finely regulate CRISPR-Cas system at the molecular levels remains uncertain (Hofer, 2017, Marraffini, 2017, Semenova and Severinov, 2016). We recently identified a novel QS regulator, CdpR (ClpAP-degradation and pathogenicity regulator), which negatively modulates the Pseudomonas quinolone signal (PQS) system in PAO1 strain (Zhao et al., 2016). PQS plays a role in the regulation of multiple genes involved in bacterial QS (Bredenbruch et al., 2006, Hassett et al., 1999). PQS and QS along with a group of transcriptional regulators form a complex regulatory network (Coggan and Wolfgang, 2012). However, whether CdpR can directly alter QS levels and function remains elusive. Furthermore, whether CdpR can influence the expression, activity, and immunity of CRISPR-Cas is completely unknown.

Here, we explored the role of CdpR in type I-F CRISPR-Cas system with Pseudomonas aeruginosa UCBPP-PA14 strain (denoted PA14) and reveal that CdpR represses the immunization and immunity potency of CRISPR-Cas via QS to impede the expression, activity, and spacer acquisition of the CRISPR-Cas system. The CdpR-mediated regulation of CRISPR-Cas influences phage infection by Vfr-mediated cas1 promoter binding and expression. Hence, we propose that CdpR may prevent bacterial self-reactivity via blockade of CRISPR-mediated endogenous cleavage. These findings enlist CdpR as the first endogenous negative regulator of CRISPR-Cas systems to maintain the balance between host defense and self-targeting of CRISPR-Cas systems. Together, our studies highlight the role of precise regulation of CRISPR-Cas in the co-evolution of bacteria with their invaders, phages, to maintain an active host defense without harming their own genes.

Results

CdpR Represses the Activity of CRISPR-Cas Interference and Spacer Acquisition

Recent studies imply that both bacterial genes and intracellular signals may regulate the expression and function of CRISPR-Cas. Since CdpR is a newly discovered regulator of QS genes and the QS systems can regulate CRISPR-Cas, we hypothesize that CdpR may modulate type I-F CRISPR-Cas systems in the PA14 strain. As expected, compared with the PA14-WT strain, the signal-deficient cdpR mutant (PA14-ΔcdpR) exhibited increased expression of Cas surveillance complex (Figure 1A). In agreement, expression of the Cas surveillance complex by complementation with CdpR (PA14-ΔcdpR/p-cdpR) was restored to the PA14-WT level. This result demonstrated that the expression of CRISPR-Cas loci is downregulated by CdpR.

Figure 1

CdpR Represses the Activity of CRISPR-Cas System against Phage Infection

(A) Heatmap for csy1-4, cas1, and cas3-related mRNA transcripts in P. aeruginosa PA14-WT, PA14-ΔcdpR, and PA14-ΔcdpR/p-cdpR with the same cell density quantified by qRT-PCR. Data were normalized with 16sRNA expression as an internal control.

(B) The type I-F CRISPR-Cas locus in PA14. Experiments utilizing a non-targeted plasmid and two CRISPR-targeted plasmids (denoted CR1-sp1 and CR2-sp1) that contain a protospacer matching spacer 1 in CRISPR array 1 and 2, respectively.

(C) Transformation efficiency of CRISPR-targeted plasmids in PA14-WT or PA14-ΔcdpR mutant.

(D) New spacer acquisition (CRISPR expansion) in CRISPR array 1 or 2 locus evaluated by PCR in PA14-WT and mutant strains. Strains harbored the primed plasmid containing a seed mutation to promote adaptation. Naïve represents the native CRISPR arrays in the PA14 genome.

(E) Diagram of crRNACR2-sp1 in P. aeruginosa PA14 type I-F CRISPR-Cas systems interacting with DMS3-T255C and DMS3100% sequences. Lines denote Watson-Crick base pairing between crRNACR2-sp1 and its phage target sequences in the DMS3-T255C and DMS3100%. PAM is shown within a shadowed box.

(F) DMS3100% and DMS3-T255C phages grew on bacterial lawns of PA14-WT, PA14-ΔcdpR, PA14-ΔcdpR/p-cdpR, and PA14-ΔTCR.

(G) Acquisition of new spacer sequences with phage DMS3-T255C infection in PA14-WT and mutant strains analyzed by qPCR.

Data shown are the means ±SEM (n = 3) (one-way ANOVA plus Tukey test, **p < 0.01; *p < 0.05).

To determine whether the function of CRISPR-Cas systems is also modulated by CdpR, we measured the effect of CRISPR-Cas on eliminating CRISPR-targeted or CRISPR-untargeted plasmid in the PA14-ΔcdpR strain. We generated two CRISPR-targeted plasmids, CR1-sp1 and CR2-sp1 (Cady et al., 2012), that possess a targeted protospacer (a sequence complementary to a spacer in CRISPR array 1 or 2, respectively) flanked by a cognate PAM (Figure 1B). We quantified the retention of plasmids in the PA14-WT and PA14-ΔcdpR strains with shaking for 5 h according to a previous report (Høyland-Kroghsbo et al., 2017). There was no loss of untargeted plasmid in all strains (Figure S1A), whereas loss of CR1-sp1 and CR2-sp1 plasmids occurred in PA14-WT, PA14-ΔcdpR, and PA14-ΔcdpR/p-cdpR compared with PA14-ΔTCR lacking cas genes (Figure S1B), indicating that plasmid loss resulted from CRISPR-Cas interference, consistent with the previous reports that CRISPR-Cas systems are required for foreign DNA interference (Barrangou et al., 2007, Marraffini and Sontheimer, 2008). Meanwhile, this loss was significantly influenced by CdpR, showing that CdpR inhibits the efficiency of CRISPR-Cas-mediated interference (Figure S1B). We investigated whether CdpR influences CRISPR-Cas-mediated elimination of an incoming genetic element that resembles a phage attack, which can be assessed by plasmid transformation efficiency of the PA14-WT and mutant strains. The transformation inhibition in the PA14-ΔcdpR strain was more efficient than that in the PA14-WT strain. In addition, complementation of cdpR restored the repression of CRISPR-Cas activity (Figure 1C). Collectively, these results affirm that CdpR represses CRISPR-Cas immunity, thereby hampering host defense against invasive elements.

Next, we questioned whether spacer acquisition is also regulated by CdpR. As the frequency of spacer acquisition can be increased by challenging bacteria with protospacer containing elements or primed process (Datsenko et al., 2012), we constructed a primed plasmid by inserting a protospacer that targets CRISPR array 2 spacer 1 containing a single base mutation (Figure 1B). We detected new spacer acquisition in the CRISPR array locus and found that, although spacer acquisition occurred in both the PA14-WT and PA14-ΔcdpR strains, adaptation in the PA14-ΔcdpR mutant was increased vs. the PA14-WT strain, which was reversed to the WT level in the complemented strain, PA14-ΔcdpR/p-cdpR (Figures 1D and S1C). This result suggests that CdpR represses CRISPR-Cas activity and blocks spacer acquisition. In addition, we found that the adaptation frequency on the CRISPR array 2 locus was higher than that on the CRISPR array 1 locus (Figures 1D and S1C), consistent with a previous study (Westra et al., 2015). Collectively, these findings clearly establish that CdpR is a negative regulator for inhibiting the expression, interference, and adaptation of P. aeruginosa CRISPR-Cas systems.

CdpR-Deletion Mutation Enhances CRISPR-Cas-Mediated Immunity against Phage Invasion

CRISPR-Cas systems are important for bacteria to defend against phage invasion. To assess whether CdpR plays a role in phage infection, we measured plaque-forming efficiency of CRISPR-sensitive phage DMS3-T255C and DMS3100% bearing a protospacer that is partially or completely (100%) complementary to the spacer portion of crRNACR2-sp1 in PA14 (Figure 1E), whereas wild-type DMS3 phage is CRISPR-insensitive (Cady et al., 2012). Phage DMS3100% failed to replicate on the PA14-WT, PA14-ΔcdpR, and PA14-ΔcdpR/p-cdpR strains because of the adaptive immunity of CRISPR-Cas systems but was able to replicate on the CRISPR-Cas deficient PA14-ΔTCR strain (Figure 1F). The DMS3-T255C phage on the PA14-WT strain could form plaques robustly, whereas the DMS3100% displayed at least a 10,000-fold lower plaque efficiency than the DMS3-T255C (Cady et al., 2012). Notably, we observed a low plaquing efficiency of DMS3-T255C on the PA14-ΔcdpR compared with that on the PA14-WT strain (Figure 1F). Remarkably, expression of the cdpR gene led to a marked increase in the plaquing efficiency of DMS3-T255C. Furthermore, new spacer acquisition was increased in the PA14-ΔcdpR strain compared with the PA14-WT strain (Figure 1G). Taken together, these data indicate that cdpR-deficiency mutation attenuates plaquing efficiency owing to its production of more potent CRISPR-Cas activity and that CdpR also negatively regulated CRISPR-Cas function during phage infection.

CdpR Represses CRISPR-Cas Regulation via QS Signals

CdpR was identified as a virulence regulator of the PQS in P. aeruginosa PAO1 strain missing CRISPR-Cas systems (Zhao et al., 2016). Bacteria communicate through QS systems to coordinate cooperative behaviors, which is essential for population fitness and invasion of hosts. We speculated that the QS circuit is associated with CdpR's effects on CRISPR-Cas systems in PA14 strains. Previous reports indicate that bacterial strains use chemical communication via QS systems to modulate CRISPR-Cas (Høyland-Kroghsbo et al., 2017, Patterson et al., 2016). In agreement, the QS circuit regulates the type I-F CRISPR-Cas systems by adjusting the expression of multiple Cas surveillance complexes at a high cell density in PA14 strains (Figures S2A and S2B). Furthermore, there were significantly increased QS-dependent transcripts of lasI/rhlI and other regulators (bfiS, bfiR, bfmS, bfmR, exsA, gacS, gacA, hptB, rpoS, and sagS, whose accumulation is associated with QS systems at a high cell density) in the PA14-ΔcdpR mutant compared with the PA14-WT strain, but these were restored to the WT levels in the P14-ΔcdpR/p-cdpR strain (Figure S2C). These responsive genes are diverse ranging from QS regulation to T3SS and small RNA regulators. We postulate that CdpR may repress adaptive immunity of CRISPR-Cas loci through QS signaling. To test this notion, we used the QS inhibitor baicalein (Luo et al., 2016) to treat the PA14-ΔcdpR background strain, not QS inhibitor meta-bromo-thiolactone (mBLT) owing to binding and inhibiting LasR and RhlR that showed no change in the cdpR mutant strain compared with the WT strain (Figure S2C). Interestingly, baicalein abolished the positive effect of QS signals on cas modules and altered csy expression (Figures 2A and S3A). Furthermore, the interference capability of these two target plasmids on the CdpR-dependent CRISPR-Cas activity was also inhibited by baicalein (Figures S3B and S3C). Moreover, the QS inhibitor affected CdpR-mediated transformation inhibition (Figure 2B). Finally, to address whether QS systems regulate spacer acquisition by CRISPR-Cas systems, we found that baicalein inhibited the efficiency of CRISPR adaptation to acquire new spacers via CdpR (Figures 2C and S3D), indicating that CdpR-mediated reduction of CRISPR-Cas activity represses the generation of immune memory by halting spacer acquisition through inhibition of QS systems. Collectively, these results imply that CdpR-mediated QS signaling is required for the control of CRISPR-Cas activity in PA14.

Figure 2

CdpR Mediates Repression of CRISPR-Cas System via Quorum Sensing

(A) Heatmap for CRISPR-Cas expression in PA14-ΔcdpR background with or without 100 μM QS inhibitor baicalein.

(B) Transformation efficiency of PA14-ΔcdpR background at high cell density with or without baicalein.

(C) Integration of new spacers into CRISPR loci was detected in PA14-ΔcdpR strains with or without baicalein.

(D) Retention assay of transformation efficiency of phage DMS3-T255C infection with PA14-ΔcdpR background with baicalein.

(E) Ability of phage DMS3-T255C to infect cdpR-deficiency mutant strains with or without baicalein.

(F) Integration of new spacers into a CRISPR array locus of PA14-ΔcdpR strain detected by PCR in the presence or absence of baicalein.

(G) Schematic of CdpR represses CRISPR-Cas immune response, including immunization and immunity, via QS.

Data shown are the means ± SEM (n = 3) (one-way ANOVA plus Tukey test, **p < 0.01; *p < 0.05).

We further investigated whether CdpR-mediated regulation of CRISPR-Cas systems is dependent on QS regulators LasI/RhlI. Double mutations of ΔlasI/ΔrhlI negatively affected the expression of cas genes (Figure S4A), consistent with the previous report (Høyland-Kroghsbo et al., 2017). We generated two ΔcdpR/ΔlasI, ΔcdpR/ΔrhlI double mutants and one ΔcdpR/ΔlasI/ΔrhlI triple mutant in the PA14 background. The increase of cas1, cas3, and csy1-4 expression in PA14-ΔcdpR was abolished by double deletion of lasI/rhlI (Figure S4B). Furthermore, compared with that in the PA14-ΔcdpR strain, expansion of CRISPR arrays (Figure S4C) and interference of CRISPR-Cas systems (Figure S4D) were decreased in the PA14-ΔcdpR/ΔlasI/ΔrhlI strain. Collectively, these findings imply that CdpR-mediated QS signaling, especially LasI/RhlI, is required for the control of CRISPR-Cas activity in PA14.

To assess the relationship between QS systems and the consequences of CdpR on CRISPR-Cas function in phage infection, we detected the effect of CdpR on CRISPR-Cas function in the presence or absence of the QS inhibitor. Colony forming units (CFUs) of PA14-ΔcdpR treated with baicalein showed a 1.654-fold reduction of transformation efficiency compared with DMSO-treated controls (Figure 2D). Moreover, the plaquing efficiency of DMS3-T255C on the baicalein-treated PA14-ΔcdpR mutant was higher than that of DMSO-treated controls (Figure 2E). In addition, inhibiting QS signaling reduced the spacer acquisition in the PA14-ΔcdpR strain following DMS3-T255C infection (Figure 2F). These results indicate that QS participates in the inhibition of CdpR in CRISPR-Cas function against phage infection. Taken together, our findings suggest that CdpR broadly represses CRISPR-Cas immune responses, including immunization and immunity, via QS signaling to resist phage infection (Figure 2G).

Vfr Is Required for Regulating CRISPR-Cas Systems by CdpR

P. aeruginosa possesses one of the most sophisticated QS systems of all bacterial species, which coordinate a group of transcriptional regulators (such as VqsR, QscR, VqsM, Vfr, and RpoN) to form a complex regulatory network (Coggan and Wolfgang, 2012). To explore the mechanism of how CdpR controls CRISPR-Cas adaptive immunity via the QS signaling, we performed bioinformatic analysis of the consensus-binding motif of these QS transcription regulators in the promoter region of cas operon and found a putative virulence factor regulator (Vfr)-binding cis-response elements (CREs) in the cas1 promoter that are homologous to the Vfr consensus-binding site (5′-ANWWTGNGAWNYAGWTCACAT-3′) (Fuchs et al., 2010) (Figure 3A). We identified that Vfr is essential for modulating adaptation and interference (Figures 3B–3D). Both phages DMS100% and DMS3-T255C showed lower plaquing efficiency on the PA14-WT strain than on the PA14-Δvfr strain (Figure 3E), indicating that Vfr regulated CRISPR-Cas immunity during phage infection. Vfr was previously implicated in the regulation of a wide range of promoters (Fuchs et al., 2010). To further test whether Vfr regulates the cas operon, we investigated cas1 promoter activity in the PA14-WT and PA14-Δvfr strains in the entire growth period. Remarkably, the cas1 promoter activity was significantly reduced in the PA14-Δvfr strain (Figure 3F) and complementation of vfr to the PA14-Δvfr mutant restored to the levels of the WT strain (Figure 3G), indicating that Vfr can activate cas1 promoter. To determine whether Vfr directly controls cas1 promoter activity, we designed and generated a cas1 promoter-fragment to perform electrophoretic mobility shift assay (EMSA) with the recombinant His-Vfr proteins. Incubation of the cas1 promoter with increasing amounts Vfr proteins resulted in the dose-dependent formation of the Vfr/cas1 promoter complex (Figure 3H), demonstrating that Vfr directly binds to the cas1 promoter. Thus, we designed oligonucleotide probes for these sites or mutated the binding motifs (Figure 3A). EMSA analysis revealed that Vfr bound to the WT Vfr CRE but not to the mutant Vfr CRE (Figures 3G and 3I). To further evaluate whether the Vfr CRE is required for the activation of the cas1 promoter, we mutated the binding sites in the cas1 promoter region. Our results showed that activation of the cas1 promoter was abolished with the mutant Vfr CRE compared with a lacZ reporter possessing the intact Vfr CRE in the PA14-WT strain (Figure 3G). Taken together, these results demonstrate that Vfr requires the specific Vfr CRE to activate the cas operon, which is responsible for the functionality of type I-F CRISPR-Cas in PA14.

Figure 3

Vfr Promotes CRISPR-Cas Activities by Binding CRISPR-Cas Promoter Operon

(A) The PA14 cas1 promoter contains a cis-response element (CRE) similar to Vfr-binding consensus site in P. aeruginosa PAO1. A mutated CRE shown in red for investigating Vfr binding (Vfr CRE-Δ). Star represents bases matching the consensus.

(B) Heatmap for relative transcripts of cas1, cas3, and csy1-4 in PA14-WT, PA14-Δvfr, and PA14-Δvfr/p-vfr quantified by qRT-PCR. Data were normalized with 16sRNA expression as an internal control.

(C) Transformation efficiency of CRISPR-targeted plasmids in the PA14-WT or PA14-Δvfr mutant.

(D) New spacer acquisition (CRISPR expansion) in CRISPR array 1 or 2 locus was quantified in PA14-WT, PA14-Δvfr, and PA14-Δvfr/p-vfr mutant strains by PCR-based analysis.

(E) DMS3100% and DMS3-T255C phages grew on bacterial lawns of PA14-WT, PA14-Δvfr, PA14-Δvfr/p-vfr, and PA14-ΔTCR.

(F) Expression of the integrative cas1-p-lacZ for cas operon reporter in PA14-WT and PA14-Δvfr mutant. Dashed lines indicate growth in lysogeny broth (LB); solid lines represent cas1 promoter activity.

(G) Expression of the cas1 promoter or the cas1 promoter containing mutated Vfr-binding sites (cas1-Δ-p) in the PA14-WT, PA14-Δvfr, or PA14-Δvfr/p-vfr mutant measured at 24 h.

(H) EMSA for binding of Vfr to the cas operon. Left, interaction between Vfr and cas1 promoter; right, mutation analysis of the Vfr-binding site in cas1 promoter binding to Vfr.

(I) EMSA for Vfr binds to the region of Vfr CRE or CRE-Δ probe.

Data shown are the means ± SEM (n = 3) (one-way ANOVA plus Tukey test, **p < 0.01; *p < 0.05).

Deletion of CdpR in PA14 exhibited a pronounced increase in the expression of vfr compared with the WT strain (Figure 4A). To further investigate the influence of Vfr in the CdpR-mediated CRISPR-Cas function, PA14-ΔcdpR, PA14-Δvfr single mutant, and PA14-ΔcdpR/Δvfr double mutant strains were generated and the expression of the CRISPR-Cas system was quantified in these strains. The increase of cas1, cas3, csy1-4 expression and cas1 promoter activity in PA14-ΔcdpR were abolished by deletion of vfr, but complementation of vfr to the PA14-ΔcdpR/Δvfr mutant restored to the WT-PA14 or PA14-ΔcdpR levels (Figures 4B and 4C), indicating that Vfr profoundly influences the CdpR-mediated regulation of the CRISPR-Cas system. Furthermore, transformation efficiency analysis demonstrated that CRISPR-Cas interference was decreased in the PA14-ΔcdpR/Δvfr strain compared with the PA14-ΔcdpR strain but restored to the control level by vfr complementation (PA14-ΔcdpR/Δvfr/p-vfr strain, Figure 4D). Moreover, expansion of the CRISPR array was not detectable in the PA14-ΔcdpR/Δvfr strain compared with the PA14-WT and PA14-ΔcdpR strains (Figure 4E). Taken together, these data elucidate that CdpR connects Vfr to regulate the interference and spacer acquisition by type I-F CRISPR-Cas systems (Figure 4F).

Figure 4

Vfr along with CdpR Modulates Activity of CRISPR-Cas Loci through Binding Cas Promoter

(A) qRT-PCR analysis of vfr in PA14-WT, PA14-ΔcdpR, and PA14-ΔcdpR/p-cdpR.

(B) Heatmap for cas-related transcripts in PA14-WT, PA14-ΔcdpR, PA14-Δvfr, PA14-ΔcdpR/Δvfr, and PA14-ΔcdpR/Δvfr/p-vfr.

(C) cas1-p-lacZ and cas1-Δ-p-lacZ activity in the PA14-WT, PA14-ΔcdpR, PA14-Δvfr, PA14-ΔcdpR/Δvfr, and PA14-ΔcdpR/Δvfr/p-vfr mutant strains were measured at 24 h post inoculation.

(D) Transformation efficiency of CRISPR-targeted plasmids in the PA14-WT or mutant strains.

(E) Integration of new spacers into CRISPR array loci detected in PA14-ΔcdpR strains with or without the deletion of vfr.

(F) Schematic of the CdpR cooperating with Vfr to control CRISPR-Cas system functionality.

Bars, means ± SEM; n = 3; **p < 0.01; *p < 0.05 (one-way ANOVA plus Tukey test).

QS LasI/RhlI Participate in the CdpR/Vfr-Mediated Regulation of CRISPR-Cas Functionality

We investigated whether Vfr affects the CdpR-mediated regulation of CRISPR-Cas system via QS (LasI/RhlI) and found that the ΔlasI/ΔrhlI double mutant negatively impacted cas1 promoter activity (Figure 5A). The enhanced activity of cas1 promoter in the PA14-ΔcdpR was abolished by double deletion of lasI/rhlI (Figure 5B). In addition, the expression of vfr was markedly increased by adding QS autoinducers: 3OC12-HSL and C4-HSL (Figure 5C), whereas baicalein blocked this effect of QS signals, resulting in a modest reduction in vfr expression compared with the WT strain (Figure 5C). Similarly, the cas1 promoter activity was enhanced by the QS autoinducers but decreased by the QS inhibitor baicalein (Figure 5D). Furthermore, the increase of cas1 promoter activity was abolished by the autoinducers when Vfr CRE was mutated (Figure 5D). These results strongly indicate that QS signals regulate Vfr to positively influence the cas operon. Since LasI and RhlI synthesize 3OC12-HSL and C4-HSL, respectively (Papenfort and Bassler, 2016), we delved into the relationship of LasI/RhlI with Vfr. As expected, disruption of lasI/rhlI attenuated vfr expression (Figure 5E). Addition of 3OC12-HSL and C4-HSL to the cultured PA14-ΔlasI/ΔrhlI strain increased vfr expression (Figure 5F), which is consistent with reduction of the cas operon activity in the absence of LasI/RhlI (Figures 5A and 5G). These findings proved that LasI and RhlI help in the production of QS autoinducers to activate vfr. Collectively, our data indicate that CdpR represses QS regulators to achieve the modulation of CRISPR-Cas functionality in a Vfr-dependent manner (Figure 5H).

Figure 5

QS LasI/RhlI Participate in CdpR-Mediated Regulation of CRISPR-Cas System via Vfr

(A) Expression of cas1-p-lacZ reporter in PA14-WT, PA14-ΔlasI, PA14-ΔrhlI, and PA14-ΔlasI/ΔrhlI at 24 h post inoculation.

(B) cas1-p-lacZ activity in PA14-WT, PA14-ΔcdpR, and PA14-cdpR/ΔlasI/ΔrhlI backgrounds at 24 h post inoculation.

(C) qRT-PCR analysis of vfr in PA14-WT with or without QS autoinducers (2 μM 3OC12-HSL and 10 μM C4-HSL) or inhibitor (100 μM baicalein).

(D) Expression of the integrative cas1-p-lacZ and cas1-Δ-p-lacZ for cas operon reporter in PA14-WT in the presence or absence of QS autoinducers or inhibitor.

(E) Relative transcripts of vfr in PA14-WT and PA14-ΔlasI/ΔrhlI quantified by qRT-PCR.

(F) qRT-PCR analysis of vfr in the PA14-ΔlasI/ΔrhlI background with or without QS autoinducers.

(G) Expression of the integrative cas1-p-lacZ for cas operon reporter in PA14-ΔlasI/ΔrhlI mutant in the absence or presence of QS autoinducers.

(H) Schematic of CdpR repressing QS regulators LasI/RhlI to inhibit CRISPR-Cas system functionality via Vfr.

Bars, means ± SEM; n = 3; **p < 0.01; *p < 0.05 (one-way ANOVA plus Tukey test).

CdpR Inhibits CRISPR-Cas Systems to Regulate Endogenous Transcription

Running a constantly active CRISPR-Cas system imposes a risk of cleaving the bacterial own mRNA or DNA as a CRISPR spacer may happen to be partially complementary to their own sequences, which may lead to autoimmunity. Based on the mechanism of RNA binding for PA14 CRISPR-Cas systems as recently described (Li et al., 2016, Müller-Esparza and Randau, 2017), the crRNAs of type I-F CRISPR-Cas system in PA14 may potentially target 189 endogenous transcripts (Figure 6A and Table S1). To investigate whether CdpR-mediated alterations of CRISPR-Cas affect endogenous genes at the transcription level, we probed crRNA-guided recognition of glpF and cysT mRNA based on 5′-GGN-3′ of PAM near its 5′-end but not 3′-end among these 189 candidate endogenous transcripts, which may be potentially recognized by PA14 CRISPR-Cas (Figures 6B and S5). The PA14-ΔTCR mutant strain lacking cas genes showed increased transcripts of glpF or cysT compared with the PA14-WT strain, but this was restored similarly to the WT levels in the complemented strain PA14-ΔTCR/p-TCR. These results argue that CRISPR-Cas systems are indeed involved in the targeted regulation of endogenous genes. Furthermore, the expression of glpF or cysT transcripts was markedly repressed in PA14-ΔcdpR compared with the PA14-WT strain but was not changed in PAO1-ΔcdpR compared with the PAO1-WT strain (Figures 6B and S5A) that does not possess CRISPR-Cas systems and serves as another negative control. We then examined the expression of phzM, which is not a target for the crRNAs, as additional control, and found that it had not been altered in the different strains (Figure S5B).

Figure 6

CdpR Inhibits Endogenous RNA Cleavage Mediated by CRISPR-Cas

(A) Graphical representation of the targeted position of PA14 crRNA spacers in the genome (orange). Rectangular columns show the number of spacers matching the mRNA of endogenous genes.

(B) Homology comparison between the mRNA sequences of endogenous genes (glpF) and CRISPR array 1 spacer 1 in PA14-WT. Transcripts of endogenous genes in PA14-WT and mutant strains at the same cell density quantified by qPCR. Data were normalized with 16sRNA or housekeeping gene pheS expression as an internal control.

(C) Overview of Co-IP with anti-His tag antibody combined with northern blot to identify endogenous transcripts binding to the CRISPR-Cas system. Northern blot analysis of glpF mRNA in the indicated strains (1: PA14-WT; 2: PA14-ΔcdpR; 3: PA14-ΔcdpR/p-cdpR; 4: PA14-WT/cas3-6xHis; 5: PA14-ΔcdpR/cas3-6xHis; 6 PA14-ΔcdpR/p-cdpR/cas3-6xHis; 7: PA14-WT; 8: PA14-ΔcdpR; 9: PA14-ΔcdpR/p-cdpR; 10: PA14-WT/csy3-6xHis; 11: PA14-ΔcdpR/csy3-6xHis; 12 PA14-ΔcdpR/p-cdpR/csy3-6xHis). RNAs determined with radiolabeled oligonucleotide probes.

(D) Single-strand glpF RNA subjected to in vitro digestion by purified, recombinant Cas3 and Csy complex.

(E) Nuclease-dead Cas3 has no effect on glpF mRNA cleavage.

(F) Effect of crRNA-target RNA mismatch on the seed-region base-pairing sitting with glpF mRNA subjected to in vitro digestion by Cas3-Csy complexes.

Data shown are the means ± SEM (n = 3) (one-way ANOVA plus Tukey test; **p < 0.01; *p < 0.05).

To precisely gauge the capacity for endogenous RNA targeting by CRISPR-Cas, we employed a co-immunoprecipitation (Co-IP) approach combined with northern blot analysis (Figure 6C). The cas3 gene was tagged with 6xHis in PA14-WT, PA14-ΔcdpR, and PA14-ΔcdpR/p-ΔcdpR strains. We performed a Co-IP on the cas3-6xHis strains (Figures 6C and S5C, lanes 4–6) and the untagged strains as a control (Figures 6C and S5C, lanes 1–3) to obtain RNA for northern blot. We identified glpF and cysT mRNA fragments with enrichment in Cas3-6xHis Co-IP (Figures 6C, left and S5C, lane 4) but found no change in phzM serving as a negative control (Figure 6C, right), indicating that Cas3 directly binds to endogenous RNA. Moreover, northern blot showed that endogenous transcripts of glpF or cysT were more abundant in the PA14-ΔcdpR strain than in the WT strain (Figures 6C-left and S5C, lanes 4–6). Moreover, similar results for the detection of crRNA binding to Cascade complex were observed via Csy3-6xHis Co-IP (Figure 6C). In addition, cleavage assay showed that a significant amount of glpF mRNAs were cleaved in vitro by the CRISPR-Cas complex (Figure 6D). However, glpF mRNA substrates were not cleaved by nuclease-dead Cas3 ( K427A or D576A mutants in DExD/H domain) (Figure 6E). We also found that the seed-region base-pairing between crRNA and RNA substrates is critical for RNA cleavage, as mutation of these nucleotides results in reduced glpF mRNA cleavage (Figure 6F). Taken together, these studies suggest that CdpR maintains the endogenous transcripts stabilization by inhibiting the activity of CRISPR-Cas. Overall, these findings demonstrate that CRISPR-mediated repression of endogenous transcripts may be neutralized by CdpR, which may be critical for regulating the abundance of individual mRNA and shaping bacterial transcriptomes.

Discussion

To date, knowledge about the microbial CRISPR-mediated adaptive immunity is rapidly evolving, particularly its primary function in preventing phage infection (Marraffini, 2015). Invasion and expansion of phages are likely to occur with the increase of bacterial cell density and can be monitored by QS surveying (Knowles et al., 2016). We uncover that CdpR facilitates the repression of CRISPR-Cas loci via regulation of QS systems. Consistent with this finding, the cdpR-deficiency mutant strain exhibits enhanced CRISPR-Cas immunity via QS signaling against phages' or foreign MGEs' invasion, indicating that CdpR-repressed QS signals, especially combination with LasI/RhlI axis, modulate the anti-phage mechanism during infection. In short, CdpR together with QS signals adds another layer of organization to bacterial anti-phage intracellular signaling. These analyses also suggest that bacterial intrinsic anti-QS components, such as CdpR, may amplify the risk of viral infection, which reflects the critical virulence ability for QS systems in invasion of hosts, consistent with that self-targeting may be harmful to bacteria (Briner and Barrangou, 2016).

Vfr functions as a global regulator of virulence factors in response to environmental cues (Coggan and Wolfgang, 2012). Vfr positively regulates the production of type IV pili (Tfp), type III secretion system (T3SS), and LasR QS system that control the expression of hundreds of additional genes (Albus et al., 1997, Sadikot et al., 2005). In addition, Vfr negatively regulates flagellar gene expression (Coggan and Wolfgang, 2012). We noticed that a consensus Vfr binding sequence located in the cas operon region interacts with Vfr to alter expression levels of CRISPR-Cas systems. Furthermore, Vfr is required for the CdpR-mediated regulation of CRISPR-Cas immune function. Moreover, the function of Vfr was activated by QS autoinducers and repressed by QS inhibitors. These data illustrate that CdpR represses QS regulators to inhibit CRISPR-Cas immunity through the Vfr signaling; however, the detailed mechanism remains to be defined.

CRISPR-Cas adaptive immunity widely exists in the bacterial world because of the everlasting viruses-host arm race and/or collaboration (Mohanraju et al., 2016). However, the expression of CRISPR-Cas loci is costly because of the possibility of self-targeting between the spacer and portion of the endogenous genes in the genome that is not part of a CRISPR array (Dugar et al., 2018, Li et al., 2016). P. aeruginosa type I-F CRISPR-Cas systems (Cas3) are reported to cleave the lasR RNA, resulting in an impaired immune response by the host (Li et al., 2016). In accordance, suppression of both P. aeruginosa biofilm formation and swarming motility by its type I-F CRISPR-Cas system requires the crRNA (Heussler et al., 2015). Furthermore, mutation of the Myxococcus xanthus type I-C CRISPR-Cas system leads to reduced expression of the FruA response regulator, resulting in markedly impaired sporulation (Boysen et al., 2002, Viswanathan et al., 2007). Listeria monocytogenes type I-A CRISPR-Cas systems enhance virulence by promoting the expression of a ferrous iron transporter (Mandin et al., 2007, Toledo-Arana et al., 2009). Moreover, a constantly active CRISPR-Cas system increases chances of accidental incorporation of nucleic acids from the cell's own genome to incur self-reactivity and even death (Stern et al., 2010). The burden of CRISPR-Cas systems, such as targeting endogenous RNA/DNA through imperfect complementarity with crRNA guides and cleavage by Cas nucleases, might provide selective pressure to co-evolution of bacteria against CRISPR-Cas adaptive immunity. Since CdpR is a newly discovered repressor that provides inhibitory effects on CRISPR-Cas function, especially inhibition of CRISPR-mediated endogenous mRNA target, it is highly likely that bacteria need to finely tune CRISPR-Cas activity to provide sufficient host defense while minimizing risk of self-targeting.

Discriminating self from non-self to effectively block invaders is a universal requirement of immune systems to function normally without self-destruction. CRISPR-Cas immunity requires a sequence match between invasive nucleic acids and spacers for cleavage of foreign DNA (Marraffini, 2015, van Houte et al., 2016), and recognition of PAM serves as a mechanism for self- and non-self-discrimination during type I-F CRISPR-Cas interference (Hayes et al., 2016, Kieper et al., 2018, Sashital et al., 2012, Westra et al., 2012, Westra et al., 2013). Hence, the PAM sequence of CRISPR motifs is important for new spacer acquisition (Wang et al., 2015). The sequence of PAM such as Cas3 5′-GG-3′ is widespread in P. aeruginosa chromosomes (Rollins et al., 2015). However, the fact is that only about one bacterium in 10 million will gain a spacer from bacterial chromosomal DNA incorporated into CRISPR loci to defend itself (Stern et al., 2010). This suggests that there is an unknown mechanism to repress the efficiency of CRISPR-Cas systems to acquire spacers from self-genome or exert cell signaling to mediate appropriate CRISPR-Cas function. The CRISPR-Cas immunity is tightly controlled, especially limiting spacer integration, providing one approach to decreasing self-targeting (Marraffini, 2017, Stern et al., 2010). Our results demonstrate that CdpR represses immunization and immunity of CRISPR-Cas systems, suggesting that bacteria may have evolved a variety of mechanisms to reduce the risk of spacer acquisition from bacterial chromosomal DNA. Inhibiting self-targeting resembles negative immune-regulation or immune tolerance in mammals as a surveillance mechanism to prevent severe tissue destruction or chronic diseases. However, how the self/non-self-discrimination is regulated remains to be fully studied (Ledford, 2017, Mohanraju et al., 2016).

In conclusion, our results demonstrate the importance of bacterial intracellular signaling in coordinating adaptive immunity in prokaryotes. This study puts forward a previously unrecognized mechanism for the regulation of CRISPR-Cas defense systems by CdpR, where an internal negative mediator has not been identified. Our proposed model delineates a series of events that are associated with CdpR action. In this model, CdpR, as a repressor, inhibits the expression and function of CRISPR-Cas systems by hampering the stimulation of QS and Vfr signaling during bacterial defense against MGEs or phages. Furthermore, CdpR inhibits CRISPR-mediated bacterial endogenous cleavage to reduce the risk of self-targeting, which needs to be further investigated. Nevertheless, our study provides the first account on how bacteria utilize virulence regulators to down-regulate CRISPR immune capacity thereby maintaining homeostasis. The negative regulatory mechanism of CRISPR-Cas systems helps balance effective host defense and self-repression by CRISPR-Cas activities. Hence, this fine-tuning of CRISPR-Cas prevents self-targeting to avoid potential autoimmunity and even mortality, while maintaining a robust CRISPR-Cas adaptive immunity.

Limitations of the Study

In this study, we identify the first intrinsic negative regulator for CRISPR-Cas that directly impacts functional activities of interference and adaptation, which keeps homeostasis while efficiently counteracting ruthless invasion by bacteriophage. However, we also made a number of observations that are intriguing but need to be probed further experimentally: whether CdpR interacts with a protein or binds to nucleic acids involved in CRISPR-Cas immunity; if so, how does it work with Vfr to control the CRISPR-Cas system. Furthermore, it remains unclear how CdpR controls QS signals and which domain is required for this mechanism. Moreover, it is currently unknown if other regulators can promote or repress the activity of CRISPR-Cas adaptive immunity. Finally, whether the possible mechanistic model of Cas3/Csy complexes-mediated RNA cleavage executes new potential function remains to be defined.

Methods

All methods can be found in the accompanying Transparent Methods supplemental file.