ABSTRACT Bacterial small noncoding RNAs (sRNAs) play posttranscriptional regulatory roles in cellular responses to changing environmental cues and in adaptation to harsh conditions. Generally, the RNA-binding protein Hfq helps sRNAs associate with target mRNAs to modulate their translation and to modify global RNA pools depending on physiological state. Here, a combination of in vivo UV cross-linking immunoprecipitation followed by high-throughput sequencing (CLIP-seq) and total RNA-seq showed that Hfq interacts with different regions of the Pseudomonas aeruginosa transcriptome under planktonic versus biofilm conditions. In the present approach, P. aeruginosa Hfq preferentially interacted with repeats of the AAN triplet motif at mRNA 5′ untranslated regions (UTRs) and sRNAs and U-rich sequences at rho-independent terminators. Further transcriptome analysis suggested that the association of sRNAs with Hfq is primarily a function of their expression levels, strongly supporting the notion that the pool of Hfq-associated RNAs is equilibrated by RNA concentration-driven cycling on and off Hfq. Overall, our combinatorial CLIP-seq and total RNA-seq approach highlights conditional sRNA associations with Hfq as a novel aspect of posttranscriptional regulation in P. aeruginosa. IMPORTANCE The Gram-negative bacterium P. aeruginosa is ubiquitously distributed in diverse environments and can cause severe biofilm-related infections in at-risk individuals. Although the presence of a large number of putative sRNAs and widely conserved RNA chaperones in this bacterium implies the importance of posttranscriptional regulatory networks for environmental fluctuations, limited information is available regarding the global role of RNA chaperones such as Hfq in the P. aeruginosa transcriptome, especially under different environmental conditions. Here, we characterize Hfq-dependent differences in gene expression and biological processes in two physiological states: the planktonic and biofilm forms. A combinatorial comparative CLIP-seq and total RNA-seq approach uncovered condition-dependent association of RNAs with Hfq in vivo and expands the potential direct regulatory targets of Hfq in the P. aeruginosa transcriptome. INTRODUCTION To thrive in fluctuating environments, bacteria must adapt to environmental threats such as temperature fluctuations, nutrient/oxygen limitations, and antibiotic exposure. In this context, regulatory noncoding RNAs (ncRNAs) have recently been implicated in posttranscriptional regulation of diverse cellular processes, including metabolism, stress response, and virulence ([37]1). Among the ncRNAs, small noncoding RNAs (sRNAs) are transcribed distal to their target RNAs. Regulation is generally accomplished through incomplete base pair formation, owing to which sRNAs often need RNA-binding proteins (RBPs) to facilitate base pairing with their target RNAs ([38]2). Hfq is one of the most extensively studied RBPs among Gram-negative bacteria ([39]3). The primary mode of action of Hfq is through acceleration of sRNA-mRNA annealing ([40]4, [41]5) and subsequent RNA stabilization or degradation ([42]6[43]–[44]8) though alternative regulatory mechanisms have also been described ([45]9[46]–[47]11). Pseudomonas aeruginosa is a notorious bacterium as an opportunistic biofilm-forming pathogen of burn wounds, medical devices, and the lungs of immunocompromised individuals ([48]12). This bacterium can grow in a wide range of environments other than the human body, its high adaptability potentially resulting from its large genome and unusually high proportion of transcriptional and posttranscriptional regulators ([49]13[50]–[51]15). Although 680 ncRNAs have been putatively detected thus far in P. aeruginosa PAO1 and PA14 genomes ([52]16[53]–[54]18), only a few sRNAs have been experimentally validated ([55]19). For example, PrrF1/PrrF2 (PrrF1/2) are expressed in an iron acquisition regulatory factor Fur-dependent manner ([56]20) and contribute to translational regulation of iron-dependent proteins, serving a similar functional role to the Enterobacteriaceae sRNA RyhB ([57]21). P. aeruginosa expresses the RNA chaperone Hfq; however, its C-terminal domain (CTD) is truncated in comparison to CTDs of the model Hfq proteins of Escherichia coli and Salmonella ([58]22). P. aeruginosa Hfq exerts pleiotropic effects: the production of virulence factors, quorum sensing, and motility are impaired in the Δhfq strain ([59]23, [60]24). Additionally, the catabolite repression control protein Crc represses the function involved in utilization of less preferred carbon sources, and the Hfq/Crc-binding sRNA CrcZ binds reciprocally and is cross-regulated with Hfq in P. aeruginosa ([61]25[62]–[63]27). Recently, Kambara et al. have shown that Hfq binds hundreds of nascent transcripts cotranscriptionally, often in concert with Crc ([64]28). New technologies based on high-throughput sequencing are increasingly providing insight into the functions of RBPs and their associated sRNAs ([65]29, [66]30). In particular, in vivo UV cross-linking immunoprecipitation followed by high-throughput sequencing (CLIP-seq) can detect transcriptome-wide binding partners of RBPs and identifies common binding sites down to single-nucleotide resolution ([67]31[68]–[69]34). Moreover, in vivo preferential ligation of RNAs has begun to unravel the sRNA interactome ([70]35[71]–[72]38). While these large-scale approaches have provided global insights into individual RBP-binding RNAs, the mechanism and functions of Hfq-mediated regulation of the P. aeruginosa transcriptome remain unclear, especially under different environmental conditions. In this study, we performed simultaneous CLIP-seq ([73]32, [74]34) and total RNA-seq to understand the molecular mode of action and physiological effects of Hfq under two medically and scientifically relevant conditions: planktonic and biofilm growth. Our comparative approach highlights competitive sRNA regulation depending on expression and allows us to reassess the key functions of Hfq in P. aeruginosa. RESULTS Identification of different RNA interactions between planktonic and biofilm forms. To investigate the interactions of Hfq with target RNAs under two pervasive conditions, i.e., planktonic and biofilm forms, we employed comparative CLIP-seq of the P. aeruginosa PAO1 hfq::3×FLAG strain. In this strain, growth rate, colony morphology, and pigment production remained unimpaired (see [75]Fig. S1 in the supplemental material). UV irradiation to induce RNA-protein cross-linking was carried out on early stationary planktonic cultures (optical density at 600 nm [OD[600]] of 2.0), and suspensions of colony biofilms formed on a cellulose membrane on Luria-Bertani (LB) agar for 48 h. Autoradiography and Western blot analyses indicated that UV cross-linking and anti-FLAG coimmunoprecipitation with stringent washing successfully enriched Hfq-RNA complex under both physiological conditions ([76]Fig. 1A and [77]Fig. S2). FIG 1. [78]FIG 1 [79]Open in a new tab Overview of cross-linking immunoprecipitation with high-throughput sequencing (CLIP-seq) analysis of Pseudomonas aeruginosa Hfq under planktonic and biofilm conditions. (A) Autoradiogram and Western blotting of the CLIP-enriched Hfq-RNA complex under two physiological conditions. XL+, cross-linking; XL-, non-cross-linking; P, planktonic; B, biofilm. Biological replicates are shown in [80]Fig. S1 in the supplemental material. (B) The distribution of Hfq peaks throughout the P. aeruginosa PAO1 genome. Peaks from planktonic and biofilm conditions are highlighted in blue and red, respectively. (C) The Venn diagram shows the number of detected peaks under planktonic and biofilm conditions. Two peaks from both conditions wherein both the start and stop positions are within 40 nt are regarded as the same peak. (D) Classification of Hfq peaks into RNA classes (5′ UTR, CDS, 3′ UTR, sRNA, tRNA, and intergenic peaks). The 5′ UTRs and 3′ UTRs were annotated from TSSs validated by differential RNA-seq ([81]39) and terminators predicted by TransTermHP ([82]40), as well as manual curation of sRNAs from size selection sRNA-seq conducted by previous researches ([83]16, [84]17). (E) DAVID enrichment analysis of Hfq peaks from 502 and 180 mRNAs except for intergenic regions under planktonic and biofilm conditions, respectively. The results of KEGG pathway enrichment analysis are presented. Overall results are shown in [85]Table S2. FIG S1 3×FLAG insertion downstream of hfq did not impair bacterial physiology. (A) Growth rate of wild-type, hfq::3×FLAG, and Δhfq strains. (B) Representative figure of colony morphologies of wild-type, hfq::3×FLAG, and Δhfq strains incubated on 1% tryptone agar with 20 μg/ml Coomassie brilliant blue and 40 μg/ml Congo red for 5 days at 25°C. (C) Overnight planktonic cultures grown in LB medium. Pigmentation was upregulated only in a Δhfq strain. Download [86]FIG S1, TIF file, 1.3 MB^ (1.3MB, tif) . Copyright © 2019 Chihara et al. This content is distributed under the terms of the [87]Creative Commons Attribution 4.0 International license. FIG S2 The Hfq-RNA complex was successfully detected in the Pseudomonas aeruginosa PAO1 hfq::3×FLAG strain under both planktonic and biofilm conditions. (A) Cell extracts from both the PAO1 wild-type and hfq::3×FLAG strain were confirmed via Western blot analysis using anti-FLAG antibody. Hfq::3×FLAG protein was only detected from PAO1 hfq::3×FLAG strain. Detection was carried out under planktonic (OD[600] = 1.0, 2.0, and 3.0) and colony biofilm (24-h-old) conditions. (B) Autoradiogram and Western blots indicate that cross-linking immunoprecipitation enriched the Hfq-RNA complex in both physiological conditions in biological replicates, as shown in [88]Fig. 1A. Download [89]FIG S2, TIF file, 2.6 MB^ (2.7MB, tif) . Copyright © 2019 Chihara et al. This content is distributed under the terms of the [90]Creative Commons Attribution 4.0 International license. We performed next-generation sequencing for both cross-linked and non-cross-linked samples in three biological replicates, subsequently seeing good correlations within each experimental condition ([91]Fig. S3). Peak calling using the tool PEAKachu ([92]https://github.com/tbischler/PEAKachu; see Materials and Methods) identified 991 putative Hfq-binding sites as peaks (average peak length ± standard deviation, 44.2 ± 14.9 nucleotides [nt]) with significant enrichment in cross-linked samples throughout the P. aeruginosa PAO1 genome ([93]Fig. 1B; also [94]Fig. S4 and [95]Table S1). We identified 187 overlapping peaks between planktonic and biofilm conditions, where overlapping peaks were defined as having both the start and stop positions within 40 nt of each other ([96]Fig. 1C). Significant peaks were classified on the basis of RNA classes ([97]Fig. 1D). We generated an untranslated region (UTR) annotation in accordance with previously reported transcription start site (TSS) data from differential RNA-seq ([98]39) and terminators predicted via TransTermHP ([99]40) via the pipeline ANNOgesic ([100]41), along with manual curation of sRNAs from size selection sRNA-seq conducted previously ([101]16, [102]17). Under both conditions, the majority of the peaks were classified into mRNAs (5′ UTR, coding DNA sequence [CDS], and 3′ UTR) and sRNAs, with fewer peaks for remaining unannotated intergenic regions. FIG S3 Heat map of correlation coefficients upon cross-linking immunoprecipitation with high-throughput sequencing. Correlation coefficient ρ was calculated from cross-linking/non-cross-linking (XL+ or XL−), forward/reverse strands (F or R), planktonic/biofilm (P or B), and biological replicates (1 to 3). In both strands, ρ was low between cross-linking and non-cross-linking samples, implying that UV treatment effectively cross-linked the RNA-Hfq complex. Download [103]FIG S3, TIF file, 1.7 MB^ (1.7MB, tif) . Copyright © 2019 Chihara et al. This content is distributed under the terms of the [104]Creative Commons Attribution 4.0 International license. FIG S4 The results of exploratory analysis. Frequency plots of matched cross-linked and background samples. The axes indicate read counts, and the colored regions indicate the frequency of each x-y pair. Download [105]FIG S4, TIF file, 1.5 MB^ (1.5MB, tif) . Copyright © 2019 Chihara et al. This content is distributed under the terms of the [106]Creative Commons Attribution 4.0 International license. TABLE S1 Hfq peaks detected by CLIP-seq. Download [107]Table S1, XLSX file, 0.2 MB^ (234.9KB, xlsx) . Copyright © 2019 Chihara et al. This content is distributed under the terms of the [108]Creative Commons Attribution 4.0 International license. To determine the metabolic pathways in which Hfq-binding RNAs are enriched, DAVID enrichment analysis was performed for the peaks, with the exception of the sRNAs and intergenic regions, under planktonic and biofilm conditions with a modified Fisher’s exact P value threshold of <0.1 ([109]Table S2) ([110]42). In the planktonic growth, genes related to carbon metabolism, such as the tricarboxylic acid (TCA) cycle, glycolysis, and gluconeogenesis, were enriched, consistent with the interaction between carbon catabolite repression control protein Crc and Hfq ([111]Fig. 1E, left) ([112]27, [113]28). In contrast, genes related to carbon metabolism and two-component systems were enriched under the biofilm condition ([114]Fig. 1E, right). Intriguingly, aminoacyl-tRNA biosynthesis was specifically highly enriched under the biofilm condition. Together, comparative CLIP-seq analysis between planktonic and biofilm forms identified different RNAs associated with Hfq and dynamic regulations of biological processes. TABLE S2 DAVID enrichment analysis data for genes identified by CLIP-seq. Download [115]Table S2, XLSX file, 0.05 MB^ (52.3KB, xlsx) . Copyright © 2019 Chihara et al. This content is distributed under the terms of the [116]Creative Commons Attribution 4.0 International license. Sequence and structural motifs from binding sites on mRNAs. Although P. aeruginosa Hfq is a functional homologue of E. coli Hfq, only the N terminus is highly conserved ([117]22). Initially, to investigate whether P. aeruginosa Hfq, whose CTD is truncated compared to that of E. coli Hfq, also interacts with mRNAs in similar regions to those described for enterobacteria ([118]31, [119]32), the peak density of Hfq peaks along all detected mRNAs was determined via meta-gene analysis using start or stop codons as reference points. Strong peak densities were observed around both start and stop codons, showing that P. aeruginosa Hfq preferentially binds the 5′ UTRs and 3′ UTRs ([120]Fig. 2A and [121]B). As examples, Hfq binds to the 5′ UTR of rhlI, which is translationally upregulated in late exponential phase ([122]24), or the 3′ UTR of katA, which putatively functions as a sponge for PrrF1 sRNA ([123]37) ([124]Fig. 2C and [125]D). FIG 2. [126]FIG 2 [127]Open in a new tab Pseudomonas aeruginosa Hfq binds to AAN triplet repeats on 5′ UTRs and rho-independent terminators. (A and B) Meta-gene analysis along mRNAs with start (A) and stop (B) codons as the reference points. (C and D) Read coverage at the rhlI (C) and katA (D) loci as representatives of Hfq peaks at the 5′ UTR and 3′ UTR, respectively. TSS (black arrows) and terminator (open circle) annotations were derived from references