The H-NS homologues MvaT and MvaU repress CRISPR-Cas in Pseudomonas aeruginosa

Abstract

CRISPR-Cas is an adaptive immune system of bacteria and archaea that protects against foreign genetic elements. In Escherichia coli and Salmonella, CRISPR-Cas is inhibited by the conserved global repressor the histone-like nucleoid structuring protein (H-NS), which blocks the expression of AT-rich horizontally acquired genes. While the opportunistic pathogen Pseudomonas aeruginosa harbours two partially redundant H-NS homologues, MvaT and MvaU, their role in CRISPR-Cas regulation in this bacterium remains unexplored. Here, we demonstrate that in the absence of both MvaT and MvaU, CRISPR-Cas activity increases more than tenfold, as measured by a reduction in the transformation efficiency of a CRISPR-targeted plasmid. Importantly, we find that in the absence of MvaT and MvaU, Cas proteins are already produced at low cell density prior to the onset of quorum sensing-mediated activation of CRISPR-Cas, which occurs at high cell density. Moreover, the ∆mvaT ∆mvaU mutant has a significantly reduced growth rate, known to independently increase CRISPR-Cas activity. In addition to regulating CRISPR-Cas, the absence of MvaT and MvaU affects phage–host interactions, including enhancing the adsorption of the LPS-binding phage JBD44, highlighting their broader role in coordinating bacterial defence mechanisms.

This article is part of the discussion meeting issue ‘The ecology and evolution of bacterial immune systems’.

1. Introduction

Bacteria and bacteriophage (phage) viruses that specifically prey on bacteria engage in an evolutionary arms race [1]. Bacteria have evolved an arsenal of defence mechanisms to protect themselves against phages, while phages have adapted strategies to overcome bacterial defences [2]. This dynamic interplay drives evolution. It significantly shapes microbial community composition and diversity, with profound ecological implications, and presents challenges for biotechnological applications [3,4]. Bacterial anti-phage defences vary in function from blocking phage adsorption, to degrading phage DNA, to abortive infection in which the bacterium self-destructs in order to prevent phage progeny from spreading [2,5–7].

One of the most sophisticated and well studied bacterial anti-phage defence systems is the clustered regularly interspaced short palindromic repeats (CRISPR) with their CRISPR-associated (Cas) proteins. CRISPR-Cas systems are adaptive immune systems of bacteria and archaea that protect against foreign genetic material, including viruses and plasmids [6]. CRISPR arrays store the genetic memory of previous genetic invaders as short nucleotide sequences called spacers, interspaced by repeat regions. The CRISPR arrays are transcribed as pre-CRISPR RNAs and processed into mature CRISPR RNAs (crRNAs) containing the spacer sequence and portions of the repeats. Eventually, crRNAs guide the Cas-mediated recognition and subsequent cleavage of protospacer sequences within foreign genetic elements [6].

Expression of CRISPR-cas is costly [8], yet about 40% of all bacteria harbour CRISPR-cas [9]. The bacterium Pseudomonas aeruginosa, which is abundant in diverse human-associated environments, is no exception, with 40–50% of all P. aeruginosa strains harbouring a functional CRISPR-cas system [10]. In particular, P. aeruginosa UCBPP-PA14 (PA14) is an important CRISPR-Cas model strain [11]. PA14 harbours the type I-F CRISPR-Cas system. In PA14, cas1, cas3, cas8f, cas5f, cas7f and cas6f are collectively flanked by two CRISPR arrays [11]. The Cas protein Cas6f is the endoribonuclease that recognizes the 16 nt hairpin sequence of the repeats within pre-crRNAs and processes them into mature crRNAs [11–13]. Cas5fto 8f forms a complex with mature crRNA, essential for target recognition [12,14]. Ultimately, the Cas3 nuclease cleaves targeted DNA [15].

Despite extensive research on CRISPR-Cas function, our understanding of the environmental factors that regulate its activity is limited. Previous studies have found nutrient availability [8,16], envelope stress [17] and metabolic stress [18] as regulators of CRISPR-cas expression in bacteria. Furthermore, we have shown that P. aeruginosa PA14 employs the cell–cell signalling system called quorum sensing (QS) to activate CRISPR-cas expression at high cell density (HCD) when the risk of phage infection is increased [19]. QS depends on the production, release and group-wide detection of small signalling molecules. At HCD, accumulation of QS molecules induces bacterial populations to synchronously activate large regulons of genes that underpin collective behaviours [20]. Pseudomonas aeruginosa harbours multiple hierarchical and intertwined QS systems: the lasIR system, the rhlIR system and the Pseudomonas quinolone signal (PQS) system [21–23]. At HCD, the QS molecules N-3-oxo-dodecanoyl-l-homoserine lactone (3OC12-HSL), N-butyryl-l-homoserine lactone (C4-HSL) and 2-heptyl-3-hydroxy-4-quinolone (PQS) accumulate and activate their respective QS receptors LasR, RhlR and PqsR, which turn on transcription of virulence genes and anti-phage defences such as CRISPR-cas [21–24].

In addition to activation by QS, a reduced growth rate increases CRISPR-Cas activity in P. aeruginosa [25,26]. Specifically, lower temperatures, low oxygen [25] and bacteriostatic antibiotics [26] slow P. aeruginosa growth, leading to a modest increase in Cas protein abundance and a high increase in spacer acquisition. The slow growth likely provides additional time for CRISPR-mediated protection against foreign invaders. CRISPR-Cas is also regulated by specific proteins. The cyclic AMP receptor protein (CRP) acts as an activator in Thermus thermophilus and Pectobacterium atrosepticum [16,27,28], but as a repressor in Escherichia coli [18,29]. In P. aeruginosa, the CRP homologue Vfr also activates CRISPR-cas [30].

Apart from CRP, the histone-like nucleoid structuring protein (H-NS) represses CRISPR-cas in E. coli, Salmonella enterica serovar Typhi and Acinetobacter baumannii [31–33]. H-NS reduces the fitness cost of horizontally acquired DNA, including CRISPR-cas [31,32,34–37], by binding and suppressing AT-rich foreign genetic material [36]. It forms filaments that bridge AT-rich DNA stretches, repressing transcription and structuring the nucleoid [35]. Homologues of H-NS are widespread across Pseudomonadota (Proteobacteria) [38]. Pseudomonas spp. harbour the H-NS homologues MvaT and MvaU, which are 51% identical at the amino acid level and form homo- and heterodimers as well as oligomers [39,40]. Similar to other H-NS proteins, MvaT forms higher-order oligomers to condense and repress AT-rich DNA [41–43]. MvaT and MvaU are partially functionally redundant and act as global regulators of approximately 350 genes, including metabolic and virulence genes [39,40,43–46]. Here, we report that the H-NS homologues MvaT and MvaU additionally repress CRISPR-Cas activity and affect bacteria–phage interactions beyond CRISPR-Cas regulation in P. aeruginosa.

2. Results

(a). MvaT and MvaU repress CRISPR-Cas activity

We sought to test if the H-NS homologues MvaT and MvaU regulate CRISPR-Cas in P. aeruginosa. Most studies on MvaT and MvaU have been conducted in P. aeruginosa PAO1, which lacks CRISPR-cas. Therefore, we studied the P. aeruginosa strain PA14, which is genetically similar to PAO1, with MvaT and MvaU being identical at the amino acid level. We deleted mvaT and mvaU in a ∆pf5 mutant of PA14 [47] to avoid Pf5 prophage activation in the absence of MvaT and MvaU, similar to Pf4 prophage induction in PAO1 [43,45,48].

We then quantified CRISPR-Cas activity at OD₆₀₀ = 0.05, representing low cell density (LCD), and OD₆₀₀ = 2.0, representing HCD, in the ∆pf5 strain, in the ∆pf5 ∆mvaT ∆mvaU strain, and in a ∆pf5 ∆CRISPR ∆cas strain as a negative control, as described previously [19]. At the given cell density, we transformed the strains with either the control plasmid pHERD30T or the pHERD30T-based pCR2SP1 plasmid, which is targeted by the CRISPR 2 spacer 1 crRNA [19,49]. We assessed CRISPR-Cas activity by comparing the relative transformation efficiency of the CRISPR-targeted plasmid pCR2SP1 to that of the pHERD30T control plasmid, with 100% relative transformation efficiency representing no CRISPR-Cas activity. Figure 1 shows the relative transformation efficiency of pCR2SP1 normalized to that of pHERD30T, while electronic supplementary material, figure S1, presents the data plotted as the number of transformants (CFU per μg plasmid DNA).

Figure 1.

MvaT and MvaU repress CRISPR-Cas activity. Transformation efficiency of the PA14 strains ∆pf5 ∆CRISPR ∆cas, ∆pf5 and ∆pf5 ∆mvaT ∆mvaU was determined at low cell density (LCD, OD₆₀₀ = 0.05) and high cell density (HCD, OD₆₀₀ = 2.0), quantified as the percentage transformation of the CRISPR-targeted pCR2SP1 plasmid normalized to the control pHERD30T plasmid. Error bars denote the standard deviation of n = 3 biological replicates. Statistical significance was determined using a two-way repeated measures ANOVA, with post hoc Tukey’s HSD test for multiple comparisons (LCD: ∆pf5 ∆CRISPR ∆cas versus ∆pf5 ∆mvaT ∆mvaU, **p = 0.0020; ∆pf5 versus ∆pf5 ∆mvaT ∆mvaU, **p = 0.0064. HCD: ∆pf5 ∆CRISPR ∆cas versus ∆pf5, *p = 0.0266; ∆pf5 ∆CRISPR ∆cas versus ∆pf5 ∆mvaT ∆mvaU, *p = 0.0189). For values below the detection limit, we assumed a value at the detection limit to determine significance.

The ∆pf5 ∆CRISPR ∆cas strain had similar transformant counts for both plasmids (figure 1; electronic supplementary material, figure S1), confirming the expected absence of CRISPR-Cas activity and equal basal transformation frequencies of both plasmids. Likewise, the Δpf5 strain showed a relative transformation efficiency of 100% at LCD, indicating no CRISPR-Cas activity against the CRISPR-targeted plasmid. Conversely, at HCD, the relative transformation efficiency dropped to 6% in the Δpf5 strain (figure 1), demonstrating frequent destruction of the CRISPR-targeted plasmid, consistent with the previously reported CRISPR-Cas activation at HCD [19]. The ∆pf5 ∆mvaT ∆mvaU mutant had substantially lower relative transformation efficiencies of the CRISPR-targeted plasmid, with 10% at LCD and 0.1% at HCD. Thus, in the absence of MvaT and MvaU, CRISPR-Cas is activated already at LCD and reaches overall higher activity at HCD. Notably, at HCD, the recovery of pCR2SP1 transformed colonies in the ∆pf5 ∆mvaT ∆mvaU strain decreased to 0–2 colonies per ml, reaching the detection limit of the assay. These data collectively show that MvaT and MvaU inhibit CRISPR-Cas activity in P. aeruginosa.

As MvaT and MvaU repress hundreds of genes [33,50], their activity may influence phage–host interactions beyond CRISPR-Cas, including the expression of phage receptors. Therefore, we first quantified the adsorption of phage JBD44 ∆cI, which depends on LPS for infection [51] and phage DMS3^vir, which uses type IV pili [52] for infection, to ∆pf5 and ∆pf5 ∆mvaT ∆mvaU strains at both LCD and HCD. Electronic supplementary material, figure S2A,B shows that 15 min after JBD44 ∆cI addition, approximately tenfold more unadsorbed phages remain in the supernatant of ∆pf5 cultures compared with those of ∆pf5 ∆mvaT ∆mvaU, regardless of cell density, suggesting that the absence of MvaT and MvaU increases JBD44 ∆cI adsorption, likely by increasing LPS receptor abundance or availability. In contrast, electronic supplementary material, figure S2C,D shows no difference in DMS3^vir adsorption between the strains at LCD and HCD, indicating similar type IV pili abundance under the tested conditions. However, DMS3^vir formed fourfold more plaques on the ∆pf5 ∆mvaT ∆mvaU strain compared with ∆pf5 at LCD but not at HCD (electronic supplementary material, figure S2E,F). Thus, JBD44 ∆cI and DMS3^vir exhibit increased adsorption and plaquing efficiency, respectively, in the absence of MvaT and MvaU. This highlights the broader role of MvaT and MvaU in shaping phage–host interactions beyond CRISPR-Cas regulation and underscores the complexity of interpreting data from CRISPR-targeted and non-targeted versions of these phages.

(b). Absence of MvaT and MvaU decreases growth rate and increases Cas protein abundance

Since the absence of MvaT and MvaU derepresses hundreds of genes [33,50], which burdens the cells, we set out to investigate the extent to which the absence of the global regulators MvaT and MvaU impacts bacterial growth, and thereby indirectly CRISPR-Cas activity [25,26]. We detected similar doubling times of the PA14 ∆pf5 strain (33 ± 2.9 min) and the mutants lacking either MvaT (32 ± 3.6 min) or MvaU (30 ± 4.9 min) (figure 2A). This was expected since MvaT and MvaU are partially functionally redundant [39,43,45]. However, in the absence of both MvaT and MvaU, the doubling time increased significantly (47 ± 5.6 min), consistent with previous observations in P. aeruginosa PAO1 [53]. Thus, the reduced growth rate caused by the absence of MvaT and MvaU probably contributes to the increased CRISPR-Cas activity.

Figure 2.

Cas6f protein abundance and bacterial doubling time increase in the absence of MvaT and MvaU. (A) Growth curves and doubling times of the PA14 strains ∆pf5, ∆pf5 ∆mvaT, ∆pf5 ∆mvaU and ∆pf5 ∆mvaT ∆mvaU, showing OD₆₀₀ over time. (B) Western blot showing the relative abundance of Cas6f in PA14 strains ∆pf5 and ∆pf5 ∆mvaT ∆mvaU over time from LCD, OD₆₀₀ = 0.05 to HCD, OD₆₀₀ = 2.0. RpoB was used as an internal control. (C) Relative Cas6f levels normalized to RpoB are displayed at a given OD₆₀₀. Error bars denote the standard deviation of n = 5 replicates (electronic supplementary material). Statistical significance was determined using Dunnett’s multiple comparisons test (∆pf5 versus ∆pf5 ∆mvaT ∆mvaU, ***, p = 0.0001) (A) and Welch’s t-tests (LCD: ∆pf5 versus ∆pf5 ∆mvaT ∆mvaU, p = 0.0337; HCD: ∆pf5 versus ∆pf5 ∆mvaT ∆mvaU, p = 0.0009) (B). ns, not significant.

Given that H-NS represses CRISPR-Cas in E. coli and S. enterica, we asked if MvaT and MvaU affected Cas protein levels. Because P. aeruginosa CRISPR-Cas activity is controlled by both QS and bacterial growth rate, and the ∆pf5 ∆mvaT ∆mvaU mutant is growth inhibited, we sought to determine the relative Cas6f levels during the entire growth curve (figure 2B,C). Using a FLAG-tagged version of Cas6f that does not interfere with protein function [25], we quantified relative Cas6f levels using western blotting (figure 2B,C). At LCD, Cas6f was barely detectable in the ∆pf5 strain and 6.5-fold more abundant in the ∆pf5 ∆mvaT ∆mvaU mutant (figure 2B). We found that in both strains, the relative abundance of Cas6f increased with cell density, as expected for a QS-regulated trait [19,25]. Furthermore, at HCD, the relative Cas6f levels in the ∆pf5 ∆mvaT ∆mvaU mutant were also significantly higher, showing a 3.8-fold increase compared with the ∆pf5 strain. These data highlight that the ∆pf5 ∆mvaT ∆mvaU mutant already accumulates significantly more Cas proteins at LCD, which, likely in combination with the reduced growth rate, contributes to the early-onset and overall increased CRISPR-Cas activity in this mutant (figure 1).

(c) MvaT and MvaU repress PQS and pyocyanin but not 3OC12-HSL or C4-HSL.

CRISPR-Cas activity is upregulated via 3OC12-HSL and C4-HSL-dependent QS signalling [19]. To test the possibility that the observed increase in CRISPR-Cas activity in the ∆pf5 ∆mvaT ∆mvaU strain is due to changes in QS regulation, we quantified the QS molecules 3OC12-HSL, C4-HSL and PQS, as well as the PQS-activated virulence factor pyocyanin [54], (figure 3). Specifically, we measured the accumulation of the QS molecules in supernatants of the ∆pf5 and ∆pf5 ∆mvaT ∆mvaU strains at OD₆₀₀ = 0.05, 1.0 and 2.0 using three established E. coli reporter strains that produce luminescence in the presence of the respective QS molecules [55,56]. At LCD, we could not detect 3OC12-HSL and C4-HSL in either strain, suggesting that these molecules were not produced at LCD, as expected of the ∆pf5 strain (Figure 3A,B). 3OC12-HSL accumulated with cell density in both strains to similar levels at HCD of 280 and 291 nM in the ∆pf5 and ∆pf5 ∆mvaT ∆mvaU strains, respectively. Although measurements for C4-HSL were below the lower limit of reliable quantification for the assay, there was a similar C4-HSL reporter induction by ∆pf5 and ∆pf5 ∆mvaT ∆mvaU supernatants. This is in contrast to quantifications of both PQS and pyocyanin levels. PQS levels at HCD were 51% higher in the ∆pf5 ∆mvaT ∆mvaU strain (324 nM) compared with the ∆pf5 strain (214 nM; figure 3C), while pyocyanin abundance was 9.2-fold higher in the absence of MvaT and MvaU at HCD (figure 3D), suggesting that MvaT and MvaU suppress PQS signalling and thereby pyocyanin production. Thus, 3OC12-HSL and C4-HSL levels are similar in the ∆pf5 and the ∆pf5 ∆mvaT ∆mvaU strains, while PQS signalling is increased (figure 3A–C). Importantly, in the ∆pf5 ∆mvaT ∆mvaU strain, CRISPR-Cas is activated at LCD before the production of 3OC12-HSL and C4-HSL, the QS signals known to induce CRISPR-Cas (figures 1 and 3A,B) [19]. This strongly indicates that the elevated CRISPR-Cas activity in the absence of MvaT and MvaU cannot be explained by differential QS molecule accumulation in the strains.

Figure 3.

The absence of MvaT and MvaU leads to differential PQS quorum sensing (QS). (A–C) The levels of the QS molecules 3OC12-HSL, C4-HSL and PQS produced by PA14 Δpf5 and Δpf5 ΔmvaT ΔmvaU were measured using three established E. coli QS reporter strains [55,56]. The relative QS molecule levels are shown as relative light units (RLU), defined as luminescence per OD₆₀₀, of the reporter strains specific for (A) 3OC12-HSL, (B) C4-HSL, (C) PQS, in response to sterile supernatants from Δpf5 and Δpf5 ΔmvaT ΔmvaU sampled at OD₆₀₀ 0.05, 1.0 and 2.0. Synthetic QS molecules were used as standards for determining concentrations. ND, not detected. (D) The relative levels of the PQS-regulated virulence factor pyocyanin are shown as absorbance at OD₆₉₅ of supernatants sampled as in (A–C). Error bars denote the standard deviation of n = 3 replicates. Statistical significance was determined using Welch’s t-tests (3OC12-HSL, HCD: ∆pf5 versus ∆pf5 ∆mvaT ∆mvaU, p = 0.1505; PQS, HCD: ∆pf5 versus ∆pf5 ∆mvaT ∆mvaU, p = 0.0003; pyocyanin, HCD: ∆pf5 versus ∆pf5 ∆mvaT ∆mvaU, p = 0.0004).

3. Discussion

This study reveals that the global transcriptional regulators MvaT and MvaU act as repressors of the CRISPR-Cas system in P. aeruginosa and that their absence leads to premature and enhanced CRISPR-Cas activation. Specifically, we report increased CRISPR-Cas activity in the absence of MvaT and MvaU, as evidenced by the reduced transformation efficiency of a CRISPR-targeted plasmid compared with that of the non-targeted control plasmid (figure 1). Here, we detected CRISPR-Cas activity in the absence of MvaT and MvaU already at LCD, a condition in which CRISPR-Cas is otherwise inactive [19]. The leader sequences of CRISPR arrays 1 and 2, as well as putative promoters of cas8f and cas1 (100 bp upstream regions), exhibit AT contents of 45, 59, 51 and 46%, respectively, which are substantially higher than the average AT content of 34% in P. aeruginosa PA14 (GenBank: CP000438.1). This indicates that MvaT and MvaU could potentially inhibit CRISPR-cas expression by directly repressing the CRISPR-cas promoters. The increased CRISPR-Cas activity in the ∆pf5 ∆mvaT ∆mvaU strain is in accordance with our finding that the Cas protein Cas6f accumulates at LCDs and reaches higher levels at HCDs in the ∆pf5 ∆mvaT ∆mvaU strain compared with the ∆pf5 strain (figure 2B,C). Additionally, we report that the key QS molecules 3OC12-HSL and C4-HSL, known to induce CRISPR-Cas in P. aeruginosa, are not increased in the ∆pf5 ∆mvaT ∆mvaU mutant (figure 3A,B). Furthermore, the CRISPR-Cas activation in this mutant occurs at a lower cell density than the production of these QS molecules (figure 1). Thus, the increased CRISPR-Cas activity is not driven by an altered QS response in the ∆pf5 ∆mvaT ∆mvaU strain. While the cellular concentration of CRISPR-Cas complexes plays an important role in the overall CRISPR-Cas activity, the significantly impaired growth of the ∆pf5 ∆mvaT ∆mvaU strain (figure 2A) probably contributes to the increased CRISPR-Cas activity (figure 1), as CRISPR-Cas activity was previously shown to be increased by slow growth [25,26].

In Vibrio harveyi, the QS receptor and regulator LuxR displaces H-NS complexes, alleviating the repression of QS-activated genes [57]. Interestingly, a mutant lacking H-NS no longer requires LuxR to activate the otherwise QS-dependent bioluminescence gene expression, suggesting that LuxR acts as a counter-silencer to activate gene expression [57]. We show that P. aeruginosa lacking the H-NS homologues MvaT and MvaU exhibits a premature and increased activation of CRISPR-Cas (figure 1), which is otherwise activated by QS via the LuxR homologues LasR and RhlR [57]. Thus, similar to observations in V. harveyi, the loss of H-NS homologues in P. aeruginosa leads to the activation of a QS-regulated trait. We envisage that LasR and or RhlR could activate CRISPR-cas expression by displacing MvaT and MvaU in P. aeruginosa. This would be consistent with our finding that CRISPR-Cas is activated at LCD in the ∆pf5 ∆mvaT ∆mvaU strain, as QS would no longer be required to trigger CRISPR-cas expression (figures 1 and 2). Thus, our data support the possibility that QS regulators may activate target gene expression by displacing H-NS-like proteins beyond Vibrio [57]. Potentially, QS-mediated activation of CRISPR-Cas in Serratia may also work through H-NS displacement [58].

In addition to regulating CRISPR-Cas, we found that MvaT and MvaU strongly modulate phage–host interactions. Specifically, the increased adsorption of phage JBD44 ∆cI to the ∆pf5 ∆mvaT ∆mvaU strain suggests elevated LPS receptor availability (electronic supplementary material, figure S2A,B), whereas adsorption of the type IV pilus-dependent phage DMS3^vir remains unchanged (electronic supplementary material, figure S2C,D). Notably, DMS3^vir exhibited increased plaquing efficiency at LCD in the ∆pf5 ∆mvaT ∆mvaU strain (electronic supplementary material, figure S2E). This suggests that MvaT and MvaU may play a role in activating additional non-CRISPR phage defences, underscoring the complex interplay between global regulators and phage susceptibility.

Phages manipulate cellular processes in the bacterial host to ensure optimal phage reproduction [59]. Interestingly, some phages encode H-NS homologues or MvaT-inhibiting proteins, which may serve as a potential way to avoid transcriptional silencing by MvaT and MvaU or to manipulate the overall host transcriptional landscape [60–63]. Additionally, plasmid-encoded H-NS homologues enable plasmid replication without draining the cellular H-NS pool. These homologues are thought to facilitate integration into the host cell regulatory network and lower the fitness cost of acquiring horizontally transferred genetic material [35,64,65]. Plasmid-encoded H-NS homologues can also inhibit host H-NS function as well as regulate a different subset of genes compared with the host H-NS [65–69]. We imagine that phages expressing H-NS homologues could similarly infect bacteria without sequestering cellular MvaT/MvaU, potentially preventing activation of CRISPR-Cas activity. This aligns with a bioinformatics analysis that revealed a phage-encoded H-NS homologue with putative binding sites in CRISPR-cas promoters in the genome of Candidatus Accumulibacter phosphatis. The authors proposed that the phage-encoded H-NS homologue could repress host CRISPR-cas [61]. Apart from encoding H-NS homologues, phages can avoid transcriptional silencing by MvaT/MvaU by inhibiting their function, a strategy relevant for virulent phages. The virulent phage LUZ24 expresses the 46 aa peptide Mip that inhibits MvaT, presumably protecting its relatively high AT content genome (47.8%) from MvaT silencing [59,63].

Given their capacity to silence horizontally acquired genes, H-NS homologues may play a broader regulatory role of other defence systems. CRISPR-cas systems can be carried within mobile genetic elements and acquired through horizontal gene transfer [70]. Anti-phage defences are often clustered in defence islands that are transferred horizontally with a relatively high frequency [71]. Thus, phage defences may be collectively repressed by H-NS homologues at LCD and subsequently de-repressed through QS at HCD, when the risk of phage infection is increased.

Despite the mechanism remaining elusive, phage infection also triggers CRISPR-Cas activation, potentially serving as a backup mechanism in case infection occurs when bacterial CRISPR-Cas is inactive [8,27]. Based on the data we present here, coupled with previous studies, we imagine the following scenario: during phage infection, replicating AT-rich phage DNA could likely sequester MvaT and MvaU. This, in turn, may cause infection-induced global de-repression of MvaT and MvaU, thereby activating hundreds of otherwise silenced genes, including CRISPR-cas. This would slow down bacterial growth, causing an additional increase in CRISPR-Cas activity [25,26].

In summary, we show that MvaT and MvaU repress CRISPR-Cas activity in P. aeruginosa and that without this repression, the cells prematurely activate CRISPR-Cas at LCD. Additionally, we find that MvaT and MvaU modulate phage–host dynamics, by, for example, limiting the adsorption of an LPS-binding phage, which emphasizes their broader role in regulating bacterial defences. To fully understand the outcomes of phage–host interactions, it is key to further dissect the interplay between major regulators, such as H-NS homologues and QS, in shaping phage defences and, conversely, to explore the strategies phages evolve to exploit or circumvent these barriers for their evolutionary benefit.

4. Material and methods

(a). Bacterial strains, plasmids and phages

Bacterial strains, plasmids and phages used in this study are listed in table 1. To construct chromosomal deletions in P. aeruginosa PA14, DNA fragments flanking the gene of interest were amplified using KOD polymerase (Novagen) and cloned into pEXG2 [72] using NEBuilder HiFi DNA Assembly Master Mix (New England BioLabs). The resulting plasmids were transformed into E. coli SM10 λpir and subsequently transferred into PA14 via mating, followed by selection on lysogeny broth (LB) agar containing gentamicin (30 μg ml⁻¹) and irgasan (100 μg ml⁻¹), followed by counter-selection on LB agar medium containing 15% (w/v) sucrose and lacking NaCl. To generate an obligate lytic variant of temperate phage JBD44, the cI repressor gene (gp77) was deleted from a PA14 JBD44 lysogen, as described above. Selected colonies were cultured in LB at 37°C for 2−5 h, and the supernatant was sterile-filtered and streaked onto lawns of PA14, following overnight incubation at 37°C. Plaques displaying clear centres were re-streaked three times. The 3×flag-tagged strains were generated as in [25]. Mutants (bacteria, plasmids and phages) were confirmed by polymerase chaing reaction (PCR) and sequencing (Eurofins).

Table 1.

Bacterial strains, plasmids, and phages used in this study.

strain, plasmid or phage	description	source
UCBPP-PA14	WT PA14 strain	laboratory stock
NMHK267	PA14 ∆pf5	[47]
NMHK427	PA14 ∆pf5 ∆mvaU	this study
NMHK428	PA14 ∆pf5 ∆mvaT	this study
NMHK430	PA14 ∆pf5 ∆mvaT ∆mvaU	this study
JJM57	PA14 ∆pf5 ∆CRISPR ∆cas	this study
NMHK436	PA14 ∆pf5 cas6f 3×flag	this study
NMHK439	PA14 ∆pf5 ∆mvaT ∆mvaU cas6f 3×flag	this study
E. coli SM10λpir	thi thr leu tonA lacY supE recA::RP4-2-Tc::Mu	laboratory stock
E. coli	pCS26 lasB::luxCDABE pBAD lasR, kanamycin and ampicillin resistance	[55]
E. coli	pCS26 rhlA::luxCDABE pBAD rhlR, kanamycin and ampicillin resistance	[55]
E. coli	pCS26 pqsA::luxCDABE pBAD pqsR, kanamycin and ampicillin resistance	[56]
pEXG2	allelic exchange suicide vector, gentamicin resistance, sacB marker	[72]
pHERD30T	shuttle vector, gentamicin resistance	[73]
pCR2SP1	pHERD30T containing the protospacer to CRISPR2 spacer 1	[49]
JBD44	temperate phage	[74]
JBD44 ∆cI	obligate lytic variant of temperate phage JBD44	this study
DMS3vir	obligate lytic variant of temperate phage DMS3	[49]

(b). Growth conditions

PA14, PA14 mutants and E. coli QS reporter strains were grown overnight at 37°C with 300 r.p.m. shaking in LB broth (10 g l⁻¹ tryptone, 5 g l⁻¹ yeast extract and 10 g l⁻¹ NaCl). Cultures were back-diluted to an appropriate OD₆₀₀ and subsequently grown to OD₆₀₀ = 0.05 (LCD) and OD₆₀₀ = 2.0 (HCD).

(c). Growth curves

PA14 overnight cultures were back-diluted to OD₆₀₀ = 0.001 in LB. The OD₆₀₀ was measured in a BioTek H1M multimode plate reader or an Implen™ OD₆₀₀ DiluPhotometer, as indicated.

(d). Transformation assay

PA14 ∆pf5, ∆pf5 ∆mvaT ∆mvaU and ∆pf5 ΔCRISPR Δcas mutants were made electrocompetent through repeated washes in a 300 mM sucrose solution. Electrocompetent cells were electroporated with the CRISPR-targeted pCR2SP1 plasmid or the empty control pHERD30T plasmid, lacking the CRISPR-targeted protospacer sequence. Following electroporation, the cells were grown in LB at 37°C with 300 r.p.m. shaking for 75 min. CFUs were enumerated on LB agar with gentamicin (50 μg ml⁻¹). Transformation efficiency was calculated as pCR2SP1 CFU relative to pHERD30T CFU.

(e). Phage proliferation

Phage JBD44 ∆cI was amplified using plate lysates prepared on PA14 lawns. JBD44 ∆cI was mixed with 50 µl PA14 overnight culture and 5 ml molten soft agar (0.75% agar) poured onto LB agar plates, and incubated overnight at 37°C. To harvest phages, 5 ml SM buffer (100 mM NaCl, 8 mM MgSO₄, 50 mM Tris HCl, pH 7.5, 0.01% gelatin) was added to the lawns and incubated for 3−5 h at room temperature. The lysates were sterile-filtered and titred via standard plaque assay on PA14 lawns.

(f). Adsorption assay

PA14 ∆pf5 and ∆pf5 ∆mvaT ∆mvaU were grown overnight, back-diluted and grown until reaching OD₆₀₀ = 0.05 (LCD) and OD₆₀₀ = 2.0 (HCD). Phage JBD44 ∆cI was added to the cultures and to LB-only controls at final concentrations of 5 × 10⁵ and 2 × 10⁹ plaque forming units ml⁻¹, respectively, and incubated at 37°C with 300 r.p.m. shaking. Samples were filter-sterilized using 0.22 µm filters, serially diluted and spotted on lawns of PA14 ∆pf5, followed by incubation overnight. Plaques were counted after 15−24 h to determine the concentration of non-attached phages.

(g). Efficiency of plaquing

PA14 ∆pf5 and ∆pf5 ∆mvaT ∆mvaU were grown overnight, back-diluted and grown until reaching OD₆₀₀ = 0.05 (LCD) and OD₆₀₀ = 2.0 (HCD). Upon reaching the respective cell densities, 50 µl HCD cultures or the equivalent amount of cells from the LCD cultures were mixed with molten soft agar (0.75% agar) and poured onto LB agar plates. DMS3^vir was serially diluted in SM buffer and spotted on the lawns. The plates were incubated at 37°C. Plaques were counted after 15 h to determine the efficiency of plaquing.

(h). Western blot

The PA14 cas6f-3×flag-tagged strains, Δpf5 cas6f-3×flag and Δpf5 ΔmvaT ΔmvaU cas6f-3×flag, were grown in LB medium and harvested at the indicated time points. Cell pellets were resuspended in Laemmli SDS sample buffer and denatured at 98°C for 10 min. Samples were separated by SDS-PAGE on a 4–20% Mini-PROTEAN TGX stain-free polyacrylamide gel (Bio-Rad) and blotted onto an Amersham Protran nitrocellulose 0.45 μm membrane. After blocking in Tris-buffered saline with Tween 20 (TBST) and 5% skimmed milk, the membrane was cut and incubated with monoclonal Anti-FLAG M2-peroxidase (HRP) antibody (catalogue no. A8592; Sigma) or monoclonal Anti-RpoB antibody (catalogue no. ab191598; Abcam), respectively, both for 1 h at a 1 : 3000 dilution, in TBST and 5% skimmed milk. After three washes with TBST, the Anti-rabbit IgG HRP conjugate antibody (catalogue no. W4011; Promega) was used as the secondary antibody for detection of the Anti-RpoB antibody at a 1 : 10 000 dilution. The membranes were washed in TBST and developed using SuperSignal West Femto maximum sensitivity substrate (catalogue no. 34095; Thermo Scientific). The relative band intensities were quantified in Image Studio.

(i). Assessment of quorum-sensing molecule production

PA14 Δpf5 and Δpf5 ΔmvaT ΔmvaU were grown overnight. Cultures were back-diluted and grown to either OD₆₀₀ = 0.05, 1.0 or 2.0. Supernatants were filter-sterilized using 0.22 µm filters. QS molecules 3OC12-HSL, C4-HSL and PQS were quantified using three established E. coli reporter strains. The reporter strains were grown overnight with ampicillin (100 µg ml⁻¹) and kanamycin (50 µg ml⁻¹), back-diluted to 1: 100 and supplemented with ampicillin (100 µg ml⁻¹), kanamycin (50 µg ml⁻¹) and arabinose (0.1%). Filter-sterilized PA14 Δpf5 and Δpf5 ΔmvaT ΔmvaU supernatants were added at appropriate dilutions to stay within the linear range of detection. The reporter strains were grown at 30°C with 200 r.p.m. shaking in a BioTek H1M multimode plate reader, and OD₆₀₀ and luminescence were measured. The levels of QS molecules were assessed based on standard curves prepared with synthetic 3OC12-HSL (O9139), C4-HSL (SML3427) and PQS (94398), respectively (all Sigma-Aldrich). The maximum reporter luminescence relative to OD₆₀₀ obtained at 15, 16 and 5 h was used to determine QS signal levels of 3O-C12-HSL, C4-HSL and PQS levels, respectively. Pyocyanin abundance was measured as absorbance at OD₆₉₅ of sterile supernatants.

(j). Statistics

GraphPad Prism (v. 10.4.2) was used to generate graphs and run statistical analyses. GraphPad Software, Boston, Massachusetts USA (www.graphpad.com). Additional statistical analyses were conducted in R (v. 4.1.2; R Core Team 2021).

Ethics

This work did not require ethical approval from a human subject or animal welfare committee.

Data accessibility

The datasets supporting this article have been uploaded as part of the online electronic supplementary material [75].

Declaration of AI use

We have not used AI-assisted technologies in creating this article.

Authors’ contributions

K.C.K.: data curation, formal analysis, investigation, methodology, resources, software, visualization, writing—original draft, writing—review and editing; J.J.M.: data curation, formal analysis, investigation, methodology, resources, software, supervision, validation, visualization, writing—review and editing; I.F.H.: data curation, formal analysis, investigation, methodology, visualization, writing—review and editing; E.F.V.: formal analysis, investigation; E.H.M.P.: formal analysis, investigation; A.S.W.: formal analysis, investigation; F.H.L.: formal analysis, investigation; N.M.H.-K.: conceptualization, funding acquisition, investigation, methodology, project administration, resources, supervision, writing—original draft, writing—review and editing.

All authors gave final approval for publication and agreed to be held accountable for the work performed herein.

Conflict of interest declaration

We declare we have no competing interests.

Funding

N.M.H.-K. was supported by the Lundbeck Foundation grant R264-2017-3936 and the Independent Research Fund Denmark grant 1054-00099B.