An rRNA fragment in extracellular vesicles secreted by human airway epithelial cells increases the fluoroquinolone sensitivity of P. aeruginosa

Lung infections caused by antibiotic resistant strains of Pseudomonas aeruginosa are difficult to eradicate in immunocompromised hosts such as those with cystic fibrosis. We previously demonstrated that extracellular vesicles (EVs) secreted by primary human airway epithelial cells (AEC) delivermiRNA let-7b-5p to P. aeruginosa to suppress biofilm formation and increase sensitivity to beta-lactam antibiotics. In this study, we show that EVs secreted by AEC transfer multiple distinct sRNA fragments to P. aeruginosa that are predicted to target the three subunits of the fluoroquinolone efflux pump MexHI-OpmD, thus increasing antibiotic sensitivity. Exposure of P. aeruginosa to EVs resulted in a significant reduction in the protein levels of MexH (-48%), MexI (-50%) and OpmD (-35%). Moreover, EVs reduced planktonic growth of P. aeruginosa in the presence of the fluoroquinolone antibiotic ciprofloxacin by 20%. A mexGHI-opmD deletion mutant of P. aeruginosa phenocopied this increased sensitivity to ciprofloxacin. Finally, we found that a fragment of an 18S rRNA external transcribed spacer that was transferred to P. aeruginosa by EVs reduced planktonic growth of P. aeruginosa in the presence of ciprofloxacin, reduced the minimum inhibitory concentration (MIC) of P. aeruginosa for ciprofloxacin by over 50%, and significantly reduced protein levels of both MexH and OpmD. In conclusion, an rRNA fragment secreted by AEC in EVs that targets the fluoroquinolone efflux pump MexHI-OpmD down-regulated these proteins and increased the ciprofloxacin sensitivity of P. aeruginosa. A combination of rRNA fragments and ciprofloxacin packaged in nanoparticles or EVs may benefit patients with antibiotic-resistant P. aeruginosa infections.

EVs increase the fluoroquinolone sensitivity of P. aeruginosa 119 In a previous study we demonstrated that EVs secreted by AEC increased the sensitivity of P. 120 aeruginosa to beta lactam antibiotics by delivering the miRNA let-7b to P. aeruginosa (19). To 121 test the hypothesis that EVs may also alter the sensitivity to other antibiotics typically used to 122 was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission.
The copyright holder for this preprint (which this version posted April 21, 2022. ; https://doi.org/10.1101/2022.04.19.488721 doi: bioRxiv preprint treat infections by P. aeruginosa (48), we examined the ability of EVs to alter the sensitivity of P. 123 aeruginosa to the fluoroquinolone antibiotic ciprofloxacin. To that end, the planktonic growth of 124 P. aeruginosa strain PA14 as well as the mucoid P. aeruginosa CF clinical isolate SMC1585 125 (49) were measured in the presence and absence of EVs and in the presence and absence of 126 eight different concentrations of ciprofloxacin to determine the minimal inhibitory concentration 127 (MIC). EVs alone had no significant effect on planktonic growth of PA14 in the absence of 128 ciprofloxacin ( Fig 1A) but reduced the MIC of ciprofloxacin by 25% (Fig 1B). EVs reduced 129 planktonic growth of PA14 in the presence of 0.02 µg/ml ciprofloxacin (corresponding to about 130 one third of the MIC) by 26%, compared to 0.02 µg/ml ciprofloxacin alone ( Fig 1C). Likewise,

131
EVs alone had no significant effect on planktonic growth of clinical isolate SMC1585 in the 132 absence of ciprofloxacin ( Fig 1D) but reduced the MIC of ciprofloxacin by 48% (Fig 1E). EVs 133 reduced planktonic growth of SMC1585 in the presence of 0.02 µg/ml ciprofloxacin by 27%, 134 compared to 0.02 µg/ml ciprofloxacin alone (Fig 1F). Full 20-hour growth curves for EV-exposed 135 and vehicle exposed PA14 and SMC1585 in the presence and absence of 0.02 µg/ml 136 ciprofloxacin are shown in S1 Fig  To begin to elucidate the mechanism whereby EVs increased the sensitivity of P. aeruginosa to 145 ciprofloxacin we analyzed the proteome of P. aeruginosa that had been exposed to EVs or 146 vehicle control. EVs reduced the protein levels of all three subunits of the fluoroquinolone efflux 147 was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission.

167
We previously found that EVs secreted by AEC deliver the miRNA let-7b-5p to P. aeruginosa,

168
where it reduced biofilm formation and increased sensitivity to beta-lactam antibiotics by 169 targeting key genes involved in biofilm formation and beta-lactam resistance (19). However, let-170 7b-5p alone did not reduce protein levels of MexHI-OpmD. To determine whether a different 171 sRNA was responsible for the EV-mediated reduction in MexHI-OpmD protein levels and the 172 resulting increased sensitivity to fluoroquinolone antibiotics, we re-analyzed our existing RNA-173 seq data (GSE174690) to identify sRNAs in EVs that are predicted to target MexHI-OpmD. The 174 alignment-based characterization of EV sRNA content showed that tRNA and tRNA like 175 fragments were the most abundant sequences in EVs, accounting for 65% of all aligned reads 176 ( Table 2). To gain a better understanding of the sRNA content of EVs at the sequence level, we 177 generated count tables for unique sequences and filtered them for the most abundant unique 178 sequences with at least 20 nucleotide length and at least 100 counts in each sample. The 179 resulting 1346 sequences were aligned to homo sapiens using BLAST+ to obtain annotations 180 for the 909 most abundant unique sequences with 100% identity and coverage (S2 Table).

181
Collectively, these 909 unique sequences accounted for 60% of total reads detected in all 3 182 replicate samples. The sequences with the highest total count (listed at the top of S2 Table) 183 include many tRNA and lncRNA fragments.  was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission.

193
Transfer of EV sRNAs to P. aeruginosa 194 We previously described the transfer of miRNAs secreted in AEC EVs to P. aeruginosa (19). To 195 assess the transfer of other sRNAs besides miRNAs to P. aeruginosa, we re-analyzed RNA-seq 196 data of EV-exposed P. aeruginosa and unexposed controls (GSE174710). We generated count 197 tables of unique sequences from trimmed reads for individual samples -three unexposed 198 controls and three samples of P. aeruginosa exposed to EVs secreted by different AEC donors.

199
Reads were filtered to include only sequences with a length of at least 20 nucleotides that were 200 detected in all three EV-exposed P. aeruginosa samples and none of the unexposed control  was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission.
The copyright holder for this preprint (which this version posted April 21, 2022. ; https://doi.org/10.1101/2022.04.19.488721 doi: bioRxiv preprint chosen because shorter sequences tend to be less specific and map to many different regions 204 in the genome. The 15 human EV sRNA sequences listed in Table 3 were detected in EV-205 exposed P. aeruginosa, but not in any of the unexposed control samples, suggesting that they 206 are delivered to P. aeruginosa by EVs. Subsequently, we used IntaRNA to predict target genes 207 of these 15 human EV sRNAs in P. aeruginosa.

211
Listed are 15 human EV RNAs that were exclusively detected in EV-exposed P. aeruginosa.  Table 3). As described next, of

250
To validate that the rRNA fragment increased the ciprofloxacin sensitivity of P. aeruginosa by 251 targeting the fluoroquinolone efflux pump MexHI-OpmD, we measured protein levels of MexHI-

252
OpmD in the presence and absence of the rRNA fragment and found that compared to the 253 empty vector control, the rRNA fragment decreased MexH by 53%, MexI by 26% and OpmD by 254 52% (Table 4). The decrease in the protein levels of MexH and OpmD was statistically 255 significant, and there was a trend for the rRNA fragment to decrease MexI protein levels that did 256 not quite reach statistical significance. Normalized peak intensities for each sample set as well 257 as average log2 fold changes and p-values for all 1911 detected proteins in P. aeruginosa are 258 was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission.

276
The finding that AEC EVs contain rRNA fragments is consistent with prior characterizations of 277 the sRNA content of EVs secreted by airway epithelial cells (40,41). It has been previously 278 described that rRNA fragments are produced in a controlled manner that is population-, sex-279 was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission.    Thus, to our knowledge, our data represent the first report of an rRNA fragment secreted by a 303 eukaryotic cell that regulates protein expression and alters the phenotype of a prokaryotic 304 organism. The exact mechanism of action and potential protein binding partners of rRNA 305 was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission.

307
We found that some, but not all, of the most abundant sequences in EVs can be detected in P. 308 aeruginosa following EV exposure. However, some human EV sRNA sequences detected in 309 EV-exposed P. aeruginosa are not among the most abundant EV sequences, suggesting 310 differential stability in P. aeruginosa. While RNA contained within EVs is protected from 311 degradation by extracellular RNases, once the EV RNA cargo is delivered to P. aeruginosa,

312
RNA that is not protected by binding to protein or P. aeruginosa mRNA targets may be more 313 prone to degradation, explaining the discrepancy between abundance in EVs compared to the 314 abundance after delivery to P. aeruginosa for some of the sRNAs.

315
One limitation of our study is that we did not assess other AEC EV cargo such as DNA, protein, 316 lipids, or metabolites, which could contribute to EV induced sensitivity to ciprofloxacin. However,

354
To assess whether EV sRNA targeting of efflux pumps is limited to P. aeruginosa or extends to 355 efflux pump orthologs in other common lung pathogens we utilized IntaRNA to predict targeting 356 of Burkholderia cenocepacia (which has orthologs for mexH, mexI and opmD) and 357 was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission.         , -qcov_hsp_perc 100 (requiring 100% query coverage) and -task blastn-short (to adjust 474 for a short input sequence). Genbank accessions returned by BLAST were annotated using the 475 R package rentrez (78). Subsequently, RNA-seq samples from EV-exposed P. aeruginosa were 476 used to assess whether human sRNAs are transferred to P. aeruginosa by EVs.

RNA-seq to detect transfer of EV RNA to P. aeruginosa 479
Existing RNA-seq data of P. aeruginosa exposed to PBS vehicle control in triplicate or to EVs 480 from 3 AEC donors (19) was re-analyzed to identify EV-mediated transfer of sRNA other than 481 was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission.

482
Omnibus (GEO Series accession number GSE174710). For each sample, trimmed reads were 483 compiled into count tables of unique sequences and samples were filtered for unique 484 sequences of at least 20 nucleotides in length that were contained in all three EV-exposed P.   was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission.

565
(A) 20-hour planktonic growth curves (OD600) of PA14 exposed to EVs or vehicle control (ctrl) 566 in the absence of ciprofloxacin. There was no significant difference between the growth curves 567 for the last 4 hours of growth. (B) 20-hour planktonic growth curves (OD600) of PA14 exposed 568 to 0.02 µg/ml ciprofloxacin and EVs or 0.02 µg/ml ciprofloxacin and vehicle control (ctrl). There 569 was a significant difference between the growth curves for the last 4 hours of growth. (C) 20-570 hour planktonic growth curves (OD600) of clinical isolate SMC1585 exposed to EVs or vehicle 571 control (ctrl) in the absence of ciprofloxacin. There was no significant difference between the 572 growth curves for the last 4 hours of growth. (D) 20-hour planktonic growth curves (OD600) of 573 clinical isolate SMC1585 exposed to 0.02 µg/ml ciprofloxacin and EVs or 0.02 µg/ml 574 was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission.
The copyright holder for this preprint (which this version posted April 21, 2022. ; https://doi.org/10.1101/2022.04.19.488721 doi: bioRxiv preprint ciprofloxacin and vehicle control (ctrl). There was a significant difference between the growth 575 curves for the last 4 hours of growth. Each data point shows the mean +/-SEM of EVs isolated 576 from four AEC donors (or matching controls), averaged across three technical replicates.

578
was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission.

600
was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission.
The copyright holder for this preprint (which this version posted April 21, 2022. ; https://doi.org/10.1101/2022.04.19.488721 doi: bioRxiv preprint S1 Table. Proteomics data of EV-exposed P. aeruginosa and unexposed controls.  was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission.