Small regulatory RNA RSaX28 promotes virulence by reinforcing the stability of RNAIII in community-associated ST398 clonotype Staphylococcus aureus

ABSTRACT

Staphylococcus aureus (S. aureus) is a notorious pathogen that cause metastatic or complicated infections. Hypervirulent ST398 clonotype strains, remarkably increased in recent years, dominated Community-associated S. aureus (CA-SA) infections in the past decade in China. Small RNAs like RNAIII have been demonstrated to play important roles in regulating the virulence of S. aureus, however, the regulatory roles played by many of these sRNAs in the ST398 clonotype strains are still unclear. Through transcriptome screening and combined with knockout phenotype analysis, we have identified a highly transcribed sRNA, RSaX28, in the ST398 clonotype strains. Sequence analysis revealed that RSaX28 is highly conserved in the most epidemic clonotypes of S. aureus, but its high transcription level is particularly prominent in the ST398 clonotype strains. Characterization of RSaX28 through RACE and Northern blot revealed its length to be 533nt. RSaX28 is capable of promoting the hemolytic ability, reducing biofilm formation capacity, and enhancing virulence of S. aureus in the in vivo murine infection model. Through IntaRNA prediction and EMSA validation, we found that RSaX28 can specifically interact with RNAIII, promoting its stability and positively regulating the translation of downstream alpha-toxin while inhibiting the translation of Sbi, thereby regulating the virulence and biofilm formation capacity of the ST398 clonotype strains. RSaX28 is an important virulence regulatory factor in the ST398 clonotype S. aureus and represents a potential important target for future treatment and immune intervention against S. aureus infections.

KEYWORDS:

• Staphylococcus aureus
• ST398
• RSaX28
• virulence
• RNAIII

Background

Staphylococcus. aureus (S. aureus) is one of the most notorious bacteria clinically, colonizing approximately 30% of the healthy population, and is well adapted to the healthcare environment and its human and animal hosts [1]. S. aureus can cause a large variety of community- and hospital-acquired infections opportunistically including skin and soft tissue infections, and bone and joint infections [2]. According to the latest study released by the China Antimicrobial Surveillance Network (CHINET, http://www.chinets.com/Data/AntibioticDrugFast), S. aureus ranks third in prevalence among all isolated pathogens, accounting for approximately about 9.47% of clinical infections in China in 2022.

Multi-locus Sequence Typing (MLST) is a widely accepted and standardized method for typing S. aureus strains. It allows for the comparison of strains from different sources and locations, providing valuable information for epidemiological studies and tracking the spread of specific S. aureus lineages [3,4]. MLST helps researchers understand the population structure, genetic relatedness, and evolution of pathogens, and can also identify clonal complexes and determine genetic diversity within lineages [5,6]. ST398 is a globally prevalent subtype of S. aureus associated with pig farming. Its antimicrobial resistance and genomic diversity pose challenges in public health and clinical treatment [7,8]. Our recent work showed that the ST398 clonotype was one of the most epidemic clonotypes in Shanghai in the last decade. It is considered the most representative community-associated S. aureus (CA-SA) clonotype in China and was found to be hypervirulent among other clonotype clinical isolates, indicating that the ST398 clonotype merits further attention [9–15]. Continued monitoring and research are crucial for understanding transmission pathways and disease characteristics, and implementing control measures.

The ability of S. aureus to colonize a wide range of organs and cause acute and chronic disease is most likely due to its production of a large number of virulence factors [16]. Numerous genome-wide studies have demonstrated that the expression of virulence factors in S. aureus is regulated at multiple levels, encompassing transcription, translation, and mRNA degradation. These multilayer regulations involve various trans-acting factors, such as sigma factors, two-component systems (TCSs), metabolic response regulatory proteins, RNA-binding proteins, and regulatory RNAs. This intricate and dynamic network enables S. aureus to swiftly adjust its metabolism and precisely modulate the synthesis of virulence factors in response to internal and external signals, as well as environmental fluctuations [17,18]. Small RNAs (sRNAs), collectively referred to as noncoding RNA, play a regulatory role within bacteria, representing an effective and economical mode of regulation [19]. Multiple studies have revealed that S. aureus harbours numerous small RNAs that play crucial regulatory roles, such as RNAIII, RsaE, teg49, RSaA, SprC and SprD [20–25]. With a high transcription level in each growth phase in S. aureus, RNAIII serves as a crucial intracellular effector of the quorum-sensing system. It is a multifunctional RNA that not only encodes a small peptide but also possesses noncoding regions that function as a regulatory RNA [20]. These noncoding segments play a regulatory role in controlling the translation and/or stability of mRNAs encoding transcriptional regulators, key virulence factors, and enzymes involved in cell wall metabolism [20].

Through transcriptome analysis of the ST398 Clonotype CA-SA, we confirmed a highly expressed RNA named RSaX28 in the research presented herein. We determined that RSaX28 is 533 nucleotides in size and is highly conserved among common clinically isolated S. aureus clonotypes. However, when analyzing the transcription characteristics of RSaX28 in different clones, we found that it is most highly transcribed in the hypervirulent ST398 clonotype S. aureus. The transcription level of RSaX28 is growth phase dependent. Both in vivo and in vitro experiments have demonstrated that RSaX28 enhances the virulence of the ST398 clonotype S. aureus. Further predictions and in vitro experiments have revealed that RSaX28 can bind to RNAIII and promote its stability, thereby facilitating the translation of downstream alpha-toxin and suppressing the translation of Sbi, ultimately affecting the virulence and biofilm formation ability of S. aureus.

Materials and methods

Bacterial strains, plasmids, oligonucleotides, and growth conditions

Clinical S. aureus isolates were all collected and stored as described previously [9,13,26]. Needed ST398, ST59, ST239, and ST5 clonotype isolates were randomly selected from 2008 to 2019 for RNA-seq using a random selection module in Microsoft Office Excel and grown on sheep blood agar plates or in tryptic soy broth (TSB) (Oxoid). Three ST398 strains used for RNA-seq were the strains with intermediate phenotype such as hemolytic ability, of which ST398-1 was previously sequenced using the third-generation sequencing platform (2012–3, Accession Number: NZ_CP021178.1) [26]. The strains and plasmids, and primers used in this research are listed in the Supplementary Materials. S. aureus was grown in TSB or on blood sheep agar plates at 37°C, and Escherichia coli (E. coli) was routinely grown in Luria–Bertani medium (LB; Oxoid). When necessary, antibiotics were used at the following concentrations: ampicillin, 100 mg/L; chloramphenicol, 10 mg/L.

RNA extraction and RNA-seq

Bacterial cells of three ST398 clonotype strains with the same initial OD600 of 0.03 were grown at 37°C with shaking at 200 rpm for 4 h or 8 h and harvested by centrifugation. After washing the cells with sterile PBS, RNA extraction was performed using miRNeasy Kits (Qiagen). Libraries were constructed using a starting amount of 5 μg of total RNA. The rRNA was removed before library construction using the Ribo-Zero Magnetic kit based on magnetic bead separation. The mRNA was fragmented using an ion-based method (TruSeqTM Stranded RNA sample prep Kit). Double-stranded cDNA synthesis was performed, with dUTP replacing dTTP during the synthesis of the second strand. After synthesizing the double-stranded cDNA, index adapters were ligated using the TruSeqTM Stranded RNA sample prep Kit. UNG enzyme was added to degrade the second strand of cDNA. Library enrichment was performed through PCR amplification for 15 cycles. Quantification was performed using TBS380 (Picogreen), and the samples were mixed according to the desired data ratio. Bridge PCR amplification was carried out on the cBot instrument to generate clusters. The RNA-seq libraries were then sequenced on a Novaseq 6000 platform (Illumina), generating 2 × 150 bp reads. The transcription data were analyzed using CLC Genomics Workbench 22.0 (Qiagen). The sRNA reference database SRD was used [27], and Rockhopper was used to identify sRNAs [28]. The raw data are available in the Sequence Read Archive (BioProject accession numbers: PRJNA1018674). Analyzed data of RNA-seq in this research are listed in the Supplementary Table.

Allelic gene replacement by homologous recombination and genetic complementation

The gene deletions were carried out in the representative clinical isolate ST398-1 (2012–3) [26]. The homologous recombination procedure using the plasmid pKOR1 was performed as described [29]. DNA fragments for the upstream and downstream sequences of RSaX28 were PCR amplified from the chromosomal DNA, and overlap PCR was used to obtain a fused PCR product. This product was cloned and inserted into pKOR1 using a Clonase reaction and attB sites, yielding the plasmid pKOR1ΔRSaX28. The plasmid pKOR1ΔRSaX28 was transferred via electroporation into S. aureus RN4220 and then into ST398-1 to construct the deletion mutant ΔRSaX28. Several other gene deletions also refer to this method. The allelic replacement procedure was performed as described by others [30]. For genetic complementation, RSaX28 was amplified by PCR with the primers comp-RSaX28-F and comp-RSaX28-R. The complementation plasmid was generated by cloning RSaX28 into the vector pOS1 and was then transferred into ΔRSaX28. The plasmid control strain was obtained by transferring the pOS1 plasmid into ΔRSaX28.

RACE for 5′ and 3′ ends of RSaX28

RNA end mapping by the rapid amplification of cDNA end (RACE) experiments was performed using a 5′-RACE kit (Sangon Biotech, B605102) and a 3′-RACE kit (Sangon Biotech, B605101) as previously described (38). For 5′RACE, using a specific reverse transcription primer 5RACE-1 that binds to a specific sequence within RSaX28, the first strand cDNA is synthesized using a reverse transcriptase mix (RNase H-). After annealing, 10–15 dC residues are added using the TdT enzyme. The (dC) residues then pair with a 5′ adaptor primer containing an oligonucleotide sequence. The specific primer 5RACE-2 was used as the downstream primer, and the first strand cDNA served as the template for the first round of PCR amplification. Subsequently, a universal primer, 5′ RACE Outer Primer, containing partial adaptor sequences,was used as the upstream primer, and another specific primer, 5RACE-3, was used as the downstream primer to amplify the 5′ end cDNA fragment of RSaX28. For 3′RACE, using the poly(A) tail at the 3′ end of RSaX28 as the binding site, reverse transcription was performed to synthesize standard first-strand cDNA. The specific primer 3RACE-1 was used as the upstream primer, and a universal primer, 3′ RACE Outer Primer, containing partial adaptor sequences, was used as the downstream primer for the first round of PCR amplification. Then, the specific primer 3RACE-2 was used as the upstream primer, and a universal primer, 3′RACE Inner Primer, containing partial adaptor sequences, was used as the downstream primer for the second round of PCR amplification. This allows for the amplification of the DNA fragment at the 3′ end of RSaX28.

In silico analysis of RSaX28 orthologs

RSaX28 orthologs were identified using the BLAST module in CLC Genomics Workbench 22.0 (Qiagen). The top 10 most prevalent clonotypes S. aureus and the accession numbers were listed below: ST5(NC_002745), ST239(NC_017341), ST1(NC_003923), ST7(CP096272), ST398(NZ_CP021178), ST188(AP018922), ST59(CP076823), ST6(CP047021), ST15(ST15), ST630(CP047321). Accession numbers of other staphylococcus species were listed below: Staphylococcus epidermidis (NC_002976), Staphylococcus lugdunensis (NC_013893), Staphylococcus pseudintermedius (NC_014925), Staphylococcus pseudintermedius (NC_017568), Staphylococcus xylosus (NZ_CP008724), Staphylococcus warneri (NC_020164), Staphylococcus pasteuri (NC_022737), Staphylococcus hyicus (NZ_CP008747), Staphylococcus xylosus (NZ_CP007208), Staphylococcus argenteus (NC_016941), Staphylococcus carnosus (NC_012121), Staphylococcus cohnii (NZ_LT963440), Staphylococcus succinus (NZ_CP018199), Staphylococcus agnetis (NZ_CP009623), Staphylococcus piscifermentans (NZ_LT906447), Staphylococcus stepanovicii (NZ_LT906462), Staphylococcus equorum (NZ_CP013980), Staphylococcus nepalensis (NZ_CP017460), Staphylococcus lutrae (NZ_CP020773), Staphylococcus muscae (NZ_LT906464), Staphylococcus simulans (NZ_CP023497). The ClustalW tool was used with the Clustal algorithm to perform multiple sequence alignment, and the alignment results were imported into ESPript3.0 to generate a multiple sequence alignment plot.

Quantitative real-time polymerase reaction (qRT-PCR)

qRT-PCR assays were carried out as described previously [31]. A total of 0.5 μg RNA was reverse-transcribed to cDNA using a QuantiTect® kit (Qiagen). For real-time PCR, which was performed on a 7500 real-time PCR system (Applied Biosystems), each 25 μL reaction contained 12.5 μL SYBR green mix (Roche), 0.1 μL of 100 μM of each primer, 9.8 μL highly purified H2O, and 2.5 μL template. A CT value between 10 and 35 was considered eligible when performing the qRT-PCR assay, and relative mRNA expression was calculated using the 2^−ΔCT method.

Northern blot

A northern blot assay was carried out as described previously [32]. Total RNA was prepared from S. aureus culture samples using miRNeasy Kits (Qiagen). Electrophoresis of total RNA (5-10 μg) was performed onto a polyacrylamide gel (PAGE) containing 7% urea (1.5 h, 120 V, 4°C). After migration, RNAs were vacuum transferred on Hybond®-N+ hybridization membranes (Cytiva). Hybridization with specific biotin-labeled probes complementary to RSaX28 sequences, followed by luminescent detection, was carried out as guidance of the North2South™ Chemiluminescent Hybridization and Detection Kit (Thermo Scientific).

Growth curve

Growth curves of S. aureus strains were generated as described previously [33]. Generally, isolates were grown overnight in 3 ml of TSB with shaking (200 rpm) at 37°C. Overnight cultures were diluted 1:200 in 200 µl of fresh TSB with shaking (200 rpm) at 37°C in a 96-well plate in triplicate and were read with a Micro Enzyme Linked Immunosorbent Assay (Micro-ELISA) Autoreader (Synergy 2) at 600 nm.

Lysis of erythrocytes by culture filtrates

Erythrocyte lysis tests were carried out as described previously [10]. S. aureus isolates were grown on TSB cultures for 15 h, and then the supernatants were collected for later use. Hemolytic activities were identified by incubating supernatant samples with 2% sheep red blood cells in PBS for 20 min at room temperature. Ultimately, hemolysis was determined by measuring the optical density at 540 nm using Micro ELISA Autoreader (Synergy 2).

Semiquantitative biofilm assay

Semiquantitative biofilm assays were performed as described previously [30]. Overnight cultures of S. aureus strains were sub-cultured (1:500) in fresh TSB containing 0.5% glucose and grown for 24 h in 96-well plates. The supernatant and non-adherent cells were discarded, and each well was washed twice with PBS. Subsequently, the cells were fixed with Bouin’s fixative. The fixative was removed after incubation for 1 h, and then washed twice with PBS. Organisms in the wells were later stained with crystal violet, and the floating stain was washed off with slow-running water. After drying, the stained biofilm was read with a Micro-ELISA Autoreader (Synergy 2) at 570 nm.

Western blot analysis

Western blotting assays were carried out as described before [34]. In brief, equal volumes (10 µL) of proteins were separated by SDS-PAGE and transferred onto nitrocellulose membranes (Invitrogen). After blocking, the membranes were incubated with specific antiserum at 4°C overnight and then incubated with a horseradish peroxidase-conjugated secondary antibody at room temperature for 2 h. Images of Western blots were acquired using a Tanon-5200 system. S. aureus α-SrtA was used as the internal reference protein.

In silico prediction of sRNA targets of RSaX28

Sequences of −75 to +75 upstream and downstream of the agrADBC, saeRS, sigB and sarA gene loci were selected because they function as regulatory proteins. The complete sequence of RNAIII was selected because it functions as a regulatory RNA. These sequences, used as Target sequences, along with the sequence of RSaX28 as the Query sequence, were input into the IntaRNA server to predict potential binding interactions. The most likely binding sequences, which correspond to the lowest predicted free energy, were selected. The data were then compiled into Supplementary Materials.

In vitro transcription and EMSA

In vitro transcription and EMSA assays were carried out as described previously [32]. The PCR fragment of RSaX28_WT was amplified using primers T7-RSaX28-F and T7-RSaX28-R, and RSaX28_mut was amplified using T7-RSaX28-F, T7-RSaX28-mut2, T7-RSaX28-mut1 and T7-RSaX28-R, so as to the PCR fragments of RNAIII_WT and RNAIII _mut. Transcription of T7 RNA was performed using RiboMAX Large Scale RNA Production Systems (Promega) with or without Biotin-16-UTP (Roche) as described in the Guidance. The RNAs were then purified by using miRNeasy RNA extraction kits (Qiagen) followed by dissolving in TMN buffer (20 mM Tris-acetate at pH 7.5, 10 mM magnesium acetate, 150 mM sodium acetate, 1 mM DTT). Biotin-labelled purified RSaX28_WT or RSaX28-mut (50 nM/sample) and cold RNAIII_WT or RNAIII_mut (0 nM, 50 nM, 100 nM, 200 nM/sample) were mixed in TMN buffer in a total volume of 10 uL. Binding was performed at 37°C in TMN buffer for 10 min. After incubation, RNA loading buffer was added and electrophoresis of the samples were performed onto a polyacrylamide gel (PAGE) under nondenaturing conditions (1 h, 100 V, 4°C). After migration, RNAs were transferred onto Hybond®-N+ hybridization membranes (Cytiva), followed by luminescent detection were carried out using North2South™ Chemiluminescent Hybridization and Detection Kit (Thermo Scientific).

Rifampicin transcription inhibition assays

Rifampicin was used as the treatment to stop the transcription [35]. S. aureus ST398 strain and its derivatives were cultured overnight, diluted to 1/100, grown for 5 h at 37°C, and incubated with 20 mg/ml rifampicin. About 8 ml of each strain was collected before and at 5, 10, 20, 30 min after adding rifampicin. These samples were centrifuged, the pellets were frozen in liquid nitrogen then stored at −80°C. Total RNA was extracted. cDNAs preparations and qRT-PCR experiments were performed as previously described. The tmRNA was used for normalization.

Reporter strains construction and in vivo ß-galactosidase assays

Construction of Reporter strains was carried out as described previously [36]. The hld promoter (Phld) and the constitutive rpoB promoter (PrpoB) was amplified using primers hld-f, hld-R, rpoB-F and rpoB-R. The leader regions of the, hla, sbi, and spa genes were amplified using the primers hla-f, hla-R, sbi-f, sbi-R, spa-f, and spa-R, respectively, which were cloned downstream of the PrpoB. The resulting construct was cloned and inserted into the pOS1-lacZ plasmid. Subsequently, the plasmid was heat-shocked into top10 and then electroporated into RN4220 cells, followed by electroporation into WT and ΔRSaX28 strains. Beta-galactosidase activity was measured as previously described [23].

Murine systemic bacteremia model

A systemic infection model was established as previously described [33]. BALB/c mice were used in the bacteremia model. All mice were between 4 and 6 weeks of age at the time of use and had access to food and water ad libitum. Overnight cultures of ST398 WT and ΔRSaX28 were sub-cultured (1:100) in fresh TSB and grown for 4 h. Then, cells were harvested by centrifugation at 5000 g for 10 min, washed twice with PBS, and resuspended in PBS at 1 × 10⁸ CFU/ml. A total of 100 μl (1 × 10⁷CFU) of that suspension was then injected intravenously into the animal via the orbital vein. The untreated group received 100 μl of PBS at the same time points instead. At 72 h post-infection, the animals were euthanized, and their kidneys were collected in tubes containing 1 ml PBS. The tissues were homogenized on ice using a manual homogenizer (Tiangen) and appropriate dilutions were prepared and plated for the determination of bacterial loads. Tissue sections were fixed in 4% formalin and processed for routine histopathology. The survival (high-dose) model was performed analogously to the above-described model, but with an inoculum of 5 × 10⁷ CFU of S. aureus. The survival status of the mice was recorded at different intervals up to 72 h.

Statistical analysis

Unpaired two-tailed Student’s t tests, One-way ANOVA tests, Kaplan-Meier tests or Pearson correlation tests were performed to analyze statistical significance. All data were analyzed using GraphPad Prism 8.0. Error bars in all graphs indicate the standard deviation (mean ± SD), and P values < 0.05 were reported as statistically significant.

Ethics approval

This study was approved by the Ethics Committee of Renji Hospital, School of Medicine, Shanghai Jiaotong University, Shanghai, China. All individual patients or their legal guardians provided informed consent. This project is a retrospective study. The S. aureus isolates from patient samples were cultured and identified in routine microbiology laboratories. All animal experiments were performed following the Guide for the Care and Use of Laboratory Animals of the Chinese Association for Laboratory Animal Sciences (CALAS) and approved by the ethics committee of Renji Hospital, School of Medicine, Shanghai Jiao Tong University, Shanghai, China.

Results

The highly transcribed RSaX28 could remarkably regulate the virulence of the ST398 strain

The ST398 clonotype constituted a main part of CA-SA according to our previous study, the isolation rate of which has increased in recent years (Figure S1). Three representatively previous isolated ST398 strains were selected. And bacterial cultures of which were collected during the mid-log phase and stationary growth phases. Total RNA was extracted and subjected to transcriptome sequencing. The transcriptome data were analyzed and a cutoff threshold of transcripts per million (TPM) greater than 10000 was used for filtering, resulting in the identification of 11 sRNAs (Figure 1(A)). Subsequently, the antisense RNAs srn_4190, the well-studied quorum-sensing regulator RNAIII, and four widely sRNAs with extensive nomenclature, sequence, and functional conservation beyond Staphylococcal genomes (4.5S, 6S, tmRNA, and RNase P) were excluded from the analysis [27], leaving behind a set of five sRNAs, namely SprG3, Teg23, ncRNA_1023, RsaX28, and Teg26 (Information on these sRNAs is listed in Figure 1(B)) [37,38]. Knockout strains of these sRNAs were generated, and the hemolytic ability, alteration of which could best reflect virulence of S. aureus clinically, of these strains was evaluated. Our findings revealed a distinct reduction in hemolytic activity in the knockout strains of SprG3, Teg23 and RSaX28. Of these sRNAs, RSaX28 exhibited the most significant effect, resulting in a 50% decrease in hemolytic ability in the deletion isolates (named ΔRSaX28) compared to the wild-type strain (Figure 1(C)). Consequently, RSaX28 was chosen as the subject for subsequent investigations.

Figure 1. The highly transcribed RSaX28 could remarkably regulate the virulence of the ST398 strain. (A) The transcripts per million (TPM) of small RNAs analyzed using transcriptome data in mid-log growth (4 h) and stationary growth phase(8 h) of three ST398 isolates. (B) Information of sRNAs with TPM greater than 10⁴ in both mid-log growth (4 h) and stationary growth phase(8 h). (C) Hemolytic activities of the wild-type ST398 strain and ΔSprG3, ΔTeg23, ΔncRNA_1023, ΔRsaX28, and ΔTeg26 mutants. One-way ANOVA tests; *, P < 0.05; ** P < 0.01.

RSaX28 is a 533nt small RNA with high conservation in S. aureus

RSaX28 was previously reported as SSR42 in laboratory standard strain SH1000, UAMS and USA300 (LAC) [38,39]. The length of RSaX28 varies among different studies, possibly due to the different strains studied. Some researchers measured its length at 891nt, while others proved its length at 1232nt [38,39]. What they have in common is, RSaX28 promoted the hemolytic activity of S. aureus, which was consistent with the clinically ST398 CA-SA strains in our study. We analyzed the position and length of RSaX28 in our ST398 CA-SA clinical strains. In the genome of ST398-1 (NZ_CP021178), RSaX28 is located between SAPIG2433 and SAPIG2434(Figure 2(A)). We subsequently performed the 5′ RACE and 3′RACE experiment in the ST398 CA-SA clinical strain to determine the transcription start and termination sites of RSaX28. As depicted in Figure S2, we identified that the transcript of RSaX28 was start at position 2350468 and end at position 2349936, namely about 533nt in size in ST398 CA-SA strains.

Figure 2. RSaX28 is a 533nt small RNA with high conservation in S. aureus. (A)Location of RSaX28 in the ST398 reference genome AM990992. (B) RSaX20 orthologues in the 10 most prevalent clonotypes of S. aureus and in other Staphylococci (dark blue and wathet blue bars indicate the 100% identity and absence of orthologues, respectively; instructions for other colours are indicated). (C)Hemolytic phenotype, transcripts levels of RSaX28 at (D) mid-log growth phase and (E) stationary growth phase in eight clinical strains in each clonotype from representative clones of CA-SA, including ST398 and ST59 clones, as well as from representative clones of HA-SA, including ST5 and ST239 clones. One-way ANOVA tests; ns, no significant difference; *, P < 0.05; ** P < 0.01; ***, P < 0.001.

ST clonotypes are of great importance for analyzing the molecular epidemiology of S. aureus infections clinically. The sequences of many genes are always somewhat at variance between different ST types. To analyze the conservation of RSaX28 in S. aureus, we chose the 10 most isolated ST types in recent decades on the basis of our recent study: ST5, ST239, ST398, ST1, ST59, ST188, ST15, ST7, ST6 and ST630 [9]. Sequences of RSaX28 were analyzed in these clonotypes. We found that RSaX28 was highly conserved (100% identity and 100% length) in most ST clonotypes, but shared 99% identity in the ST188 clonotype; in addition, sequence of RSaX28 shared 96% identity and 66% length in ST6 comparing to the ST398 clonotype, indicating that it might play an important role in regulating S. aureus (Figure 2(B), Figure S3). We next analyzed the existence of RSaX28 in other Staphylococcus species. We found that RSaX28 was only BLAST in Staphylococcus argenteous, which is one of the closest staphylococcal species to S. aureus. The sequence of RSaX28 in S. argenteous sharing an identity of 90% and 100% length compared to that in S. aureus (Figure 2(B), Figure S3).

Next, we selected eight clinical strains in each clonotype from representative clones of CA-SA, including ST398 and ST59 clones, as well as from representative clones of HA-SA, including ST5 and ST239 clones. We assessed the hemolytic phenotype of these strains and measured the transcripts levels of RSaX28 at 4 and 8 h post-bacterial growth. We found that the CA-SA clones, represented by ST59 and ST398, exhibited stronger hemolytic ability than the ST5 and ST239 clones (Figure 2(C)). Additionally, RSaX28 showed higher transcripts levels in CA-SA as RSaX28 transcripts level was higher in CA-SA than HA-SA in the mid-log growth phase (Figure 2(D)). And we observed that, quite interestingly, the transcripts level of RSaX28 in the ST398 clone was higher than that in the ST59 clonotype during the stationary growth phase, especially in the HA-SA clones (Figure 2(E)). Due to the phenomenon of high transcripts levels, we deem that RSaX28 plays an important regulatory role in the CA-SA ST398 clonotype, hence further investigation was focused on it.

RSaX28 is transcribed with a growth phase-dependent pattern in ST398 clinical isolates

We then utilized Northern blot analysis to characterize RSaX28. As illustrated in Figure 3(A), we detected RSaX28 in the mid-log growth phase (4 h) and stationary growth phase (8 h) of three ST398 strains. The size of RSaX28 was found to be approximately 500 base pairs, consistent with the results obtained from the RACE experiment. Moreover, we collected bacterial samples at different growth stages and observed the transcription of RSaX28, indicating a growth phase-dependent transcription pattern. RSaX28 reached its highest transcripts level during the stationary phase (Figure 3(B)).

Figure 3. RSaX28 transcribed with a growth phase-dependent pattern in ST398 clinical isolates. (A) Northern blot analysis of RSaX28 and the internal control tmRNA in the mid-log growth phase (4 h) and stationary phase (8 h) of three ST398 strains with biotin labelling markers of 200, 300, 500 and 1000nt on the left panel. (B) Northern blot analysis of RSaX28 and the internal control tmRNA in different growth stages of the ST398 strain, and the relative quantity is shown in the down panel. The relative quantity is the ratio of the integrated density of the RSaX28 band to the integrated density of tmRNA band. (C) Northern blot analysis of RSaX28 and the internal control tmRNA in the wide-type ST398 strain, ΔRsaX28, C-RSaX28 and C-pOS1 mutants in the stationary phase (8 h).

We subsequently constructed an RSaX28 complementation strain using shuttle expression plasmid pOS1(named C-RSaX28) and a strain containing empty pOS1 plasmid (named C-pOS1) as control in ΔRSaX28 strain subsequently. Northern blot analysis was carried out in these strains. In the knockout strain, the signal corresponding to RSaX28 was completely abolished. However, in the complementation strain, the RSaX28 signal was restored, confirming the successful construction of these mutants (Figure 3(C)).

RSaX28 enhanced the virulence of ST398 isolates in vitro and in vivo

To eliminate the effect of growth ability to other phenotypes, the growth curves of WT, dRSaX28, C-RSaX28 and C-pOS1 were assayed, and no significant difference was found between the four isolates (Figure 4(A)). While the hemolytic ability of ΔRSaX28 was reduced, it reverted to a certain extent in the C-RSaX28 isolate (Figure 4(B)). Moreover, we found that the biofilm formation ability of ΔRSaX28 was significantly enhanced, indicating that RSaX28 may be involved in the global regulation of S. aureus (Figure 4(C)).

Figure 4. RSaX28 enhanced the virulence of ST398 isolates in vitro and in vivo. Growth curves(A), hemolytic activities(B), and semiquantitative biofilm formation abilities(C) of the wide-type ST398 strain and the ΔRsaX28, C-RSaX28 and C-pOS1 mutants. Kidney body and HE staining (D), and bacterial CFU load (E) from a murine systemic bacteremia model caused by intravenous injection of 1 × 10^7 CFU of the wide-type ST398 strain, RSaX28 and PBS control. (E) Survival curve of the mice from the murine systemic bacteremia model caused by intravenous injection of 5 × 10^7 CFU of the wide-type ST398 strain, RSaX28 and PBS control. One-way ANOVA tests (B and C), Student’s t tests (E)and Kaplan-Meier tests (F); *, P < 0.05; ** P < 0.01; ***, P < 0.001.

To further investigate the impact of RSaX28 on the virulence regulation of S. aureus, we established a murine bloodstream infection model, which represents a systemic infection. Intravenous injection of 1 × 10^7 CFU bacterial load into the mice via the tail vein resulted in systemic infection, with S. aureus colonizing various organs, most notably the kidneys. In WT-infected mice, the kidneys exhibited large and prominent abscesses, accompanied by extensive leukocyte infiltration and tissue damage as visualized by HE staining. However, in mice infected with the ΔRSaX28 strain, the abscess area in the kidneys was smaller, and there was reduced leukocyte infiltration (Figure 4(D)). Furthermore, bacterial quantification in the kidneys revealed a significant decrease in bacterial load in the ΔRSaX28 strain-infected mice (Figure 4(E)).

We next increased the bacterial load to 5 × 10^7 CFU and plotted the survival curve of the mice. We observed that mice infected with the ΔRSaX28 strain exhibited enhanced survival compared to those infected with the wild-type strain (Figure 4(F)). Taken together, these experiments provide evidence that RSaX28 significantly enhances the virulence of the ST398 clonotype S. aureus. This is consistent with previous studies on the virulence regulation role of SSR42 in S. aureus strains such as SH1000 and USA300 [38,39].

RSaX28 binds to RNAIII through a partially complementary sequence

Next, we aimed to identify the target of RSaX28. Based on our phenotype experiments, we observed significant changes in hemolytic activity and biofilm formation ability upon RSaX28 knockout. Therefore, we hypothesized that RSaX28 might interact with some global regulatory factors. To investigate this, we employed IntaRNA to predict the binding capabilities between RSaX28 and RNAIII, agrADBC, saeRS, sigB and sarA. Interestingly, we found a significant binding site (ΔE = −14.95 kcal/mol) between RSaX28 and RNAIII (Information is shown in the Supplementary Materials). Moreover, when we assessed the transcripts levels of RNAIII and RSaX28, we found significant correlation between them in the former eight ST398 clonotype strains (Figure S4A), while no significant correlation was found between saeR and saeS to RSaX28, further suggesting that there might be a mutual relationship between them (Figure S4B and Figure S4C). To validate their interaction, we performed in vitro transcription of RNAIII and RSaX28, followed by an EMSA experiment. The results confirmed the binding between RNAIII and RSaX28 (Figure 5(B)). We also performed in vitro transcription of saeRS, and no binding signal was found in the EMSA assays between saeRS and RSaX28 in Figure S4D.

Figure 5. RSaX28 binds to RNAIII through a partially complementary sequence. (A)Predicted binding sites between RNAIII and RSaX28 by IntaRNA. The mutated sequences of RSaX28_mut and RNAIII_mut are shown on the right. (B)EMSA experiments of in vitro transcribed RNAIII_WT or RNAIII_mut (0 nM, 50 nM, 100 nM, 200 nM/sample from the left to the right lane) and biotin-labelled RSaX28_WT or biotin-labelled RSaX28_mut (50 nM/sample in each lane).

The binding sites between RNAIII and RSaX28 predicted by IntaRNA are shown in Figure 5(A). To further investigate this interaction, the predicted binding sites of RSaX28 and RNAIII were mutated, as depicted in Figure 5(A). Subsequently, we conducted in vitro transcription and EMSA experiments. We found that the mutated RSaX28 failed to bind to wild-type RNAIII, and similarly, the mutated RNAIII could not bind to wild-type RSaX28. Meanwhile, the mutated RSaX28 and mutated RNAIII regained their binding capabilities as their mutated sequences are complementary (Figure 5(B)). This provides evidence that RNAIII and RSaX28 indeed interact with each other through partially complementary sequences.

RSaX28 regulates the translation of alpha-toxin and Sbi by promoting the stability of RNAIII

We further investigated the effects on the binding of RSaX28 to RNAIII. Generally, sRNA binding to target RNA can have three main effects: regulation of transcription levels, influence on stability, and modulation of post-transcriptional translation. We measured the transcription levels of hld by using a lacZ reporter plasmid containing hld promoter in WT, ΔRSaX28 and C-RSaX28 but did not observe any significant changes compared to WT (Figure 6(A)). However, when we assessed the stability of RNAIII through rifampicin transcription inhibition experiments, we found that the stability of RNAIII was significantly reduced in ΔRSaX28, which was mostly restored in C-RSaX28 (Figure 6(B and D)). We also evaluate the transcripts level of RNAIII in the WT and ΔRSaX28. The qRT-PCR analysis demonstrated a two- to three-fold decrease in RNAIII transcripts abundance in the ΔRSaX28 strain relative to the WT strain (Figure 6(C)). Furthermore, the quantitative data from the Northern Blot analysis revealed that the RNAIII transcripts level in the WT were approximately two-fold higher compared to the ΔRSaX28 strain (Figure 6(E)). This suggests that RSaX28 can exert regulatory effects by promoting the stability of RNAIII.

Figure 6. RSaX28 regulates the translation of alpha-toxin and Sbi by promoting the stability of RNAIII. (A)Transcriptional levels of hld in the wild-type ST398 strain and ΔRSaX28 mutant containing lacZ translation reporter plasmids with hld promoter. (B) RNAIII levels in the wild-type ST398 strain and the ΔRSaX28 and C-RSaX28 mutants at 0, 5, 10, 20 and 30 min after rifampicin was added and determined using RT-qPCR (The Y-axis values represent the ratio of relative transcripts abundance at that time point to that at 0 min). (C)The quantitative data of the qRT-PCR for RNAIII in the WT and the ΔRSaX28 strain in time 0. (D) Transcription level of RNAIII and the internal control tmRNA in wild type ST398 strain and ΔRSaX28 mutant in 0, 5, 10, 20 and 30 min after rifampicin added using the Northern blot assay. (E)The quantitative data of the Northern Blot for RNAIII in the WT and the ΔRSaX28 strain in time 0. Translation levels of Hla(F), Sbi(G) and Spa(H) in the wild-type ST398 strain and ΔRSaX28 mutant containing lacZ translation reporter plasmids with the constitutive rpoB promoter and leader regions of hla, sbi and spa. Translation levels of Hla (I) and Sbi (J) in ΔRNAIII strain and ΔRNAIIIΔRSaX28 strain containing lacZ translation reporter plasmids with the constitutive rpoB promoter and leader regions of hla, and sbi. Student’s t tests; ns, no significant difference; *, P < 0.05; ** P < 0.01; ***, P < 0.001.

RNAIII is an important globally regulatory sRNA, so when RSaX28 affects its stability, downstream genes may undergo changes. We selected downstream functional genes, namely hla, sbi, and spa, for further investigation. We constructed lacZ translation reporter plasmids using the constitutive rpoB promoter and leader regions of these three genes in the WT and ΔRSaX28 strains. Through β-galactosidase activity assays, we found that the translation activity of hla was significantly reduced in the ΔRSaX28 strain, while the translation activities of sbi were significantly increased (Figure 6(F and G)). The findings are in line with the regulatory effects of RNAIII.

To demonstrate the regulatory effect of RSaX28 on the translation levels of hla and sbi through its impact on RNAIII stability, we constructed the RNAIII knockout strain (ΔRNAIII) and the RNAIII/RSaX28 double knockout strain (ΔRNAIIIΔRSaX28). Our findings revealed that, in comparison to ΔRNAIII, the β-galactosidase activities of pOS1-Prpob-sbi-lacZ and pOS1-Prpob-hla-lacZ in ΔRNAIIIΔRSaX28 did not exhibit significant changes. This provides evidence that RSaX28 indeed enhances the translation levels of alpha toxin and Sbi by influencing the stability of RNAIII (Figure 6(I and J)).

Surprisingly, we observed an increase in the translation activity of Spa (Figure 6(H)), indicating that RSaX28 may have supplementary regulatory effects through alternative binding pathways. For instance, the SaeRS regulatory system, which was mentioned in a prior study [38], may be involved, as suggested by our IntaRNA prediction results (see Supplementary Materials), despite the lack of apparent correlation between their transcripts (Figure S4). Additionally, we performed Western blot analysis to measure the levels of Hla protein in the WT, ΔRSaX28, C-RSaX28, and C-pOS1 strains. We found that the Hla protein was reduced in ΔRSaX28 but restored in C-RSaX28 (Figure S5). This further confirms that RSaX28 can regulate the translation of downstream hla by affecting RNAIII. Furthermore, the immunoblot results of Spa protein were consistent with the translation reporter plasmid experiments (Figure S5), suggesting that RSaX28 may affect the translation of Spa through other regulatory pathways.

Discussion

The growth, pathogenesis, and various metabolic processes of S. aureus are tightly regulated. Several two-component systems and global regulatory proteins play important roles in these processes [17]. Notably, the regulation by non-coding RNAs is worth mentioning. Due to their ability to function without translation into proteins, small noncoding RNAs (sRNAs) have emerged as an economically efficient regulatory mechanism [19,40]. RNAIII plays a crucial regulatory role in the quorum-sensing system, affecting multiple processes such as invasion, dissemination, cell wall metabolism, and biofilm formation in S. aureus [20]. In our study, we report the noncoding sRNA RSaX28, which exhibits high transcription levels similar to RNAIII in the ST398 clone. Furthermore, our experimental evidence demonstrates that RSaX28 can influence the stability of RNAIII, thereby affecting the translation of downstream alpha-toxin and Sbi and regulating the virulence and biofilm formation ability of S. aureus (Figure 7). This finding holds significant implications for the regulation of ST398.

Figure 7. RSaX28 with a high transcription level in ST398 clonotype CA-SA can influence the stability of RNAIII, thereby affecting the translation of downstream alpha-toxin and Sbi and regulating the virulence and biofilm formation ability of S. aureus.

RSaX28 has also been identified in other studies, where it has been designated Teg27 [39,41], sRNA363 [4,42], or SSR42 [38,39,43]. John et al. described the regulation of RNA SSR42 by S. aureus in their popular study [39]. They found that RNA SSR42 regulates the expression of approximately 80 mRNAs, including several virulence factors, in the UAMS-1 and USA300 (LAC) strains of S. aureus during the stationary growth phase. Their research revealed that SSR42 encodes an 891-nt RNA molecule in LAC, which is essential for hemolysis, resistance to human polymorphonuclear leukocyte killing, and pathogenesis in a mouse model of skin and soft tissue infection. Jessica et al. identified a significant decrease in acute cytotoxicity and hemolytic activity in surface protein inhibitor (Rsp) mutants in previous studies [38,43]. These mutants exhibited reduced transcription of toxin genes and loss of transcription of a 1232-nt long noncoding RNA (ncRNA) called SSR42. Subsequently, they reported that SSR42 acts as an effector in the transcriptional regulation of the α-toxin gene hla by Rsp in S. aureus 6850 (a highly cytotoxic and clinically virulent methicillin-sensitive ST50 strain). Exposure of S. aureus to subinhibitory concentrations of benzylpenicillin resulted in enhanced transcription of SSR42, leading to increased SSR42-dependent hemolysis [38]. The secondary structures of RSaX28 and SSR42 were further predicted using the online RNAfold webserver [44]. As shown in Figure S6, the structure of RSaX28 is similar to SSR42; more importantly, the structure of the binding sites to RNAIII were extremely similar. Above all, we think that RSaX28 is the same of the longer SSR42. Due to the specificity of the clinical strains, the size of the transcript in ST398 strains was 533nt.

Recent epidemiological data have indicated that ST398 dominated CA-SA infections in the past decade in Shanghai; furthermore, we found that there was a significant progressive enhancement of the proportion of ST398 clonotype S. aureus isolates [9,11–13]. Moreover, ST398 clonotype isolates were found to be hypervirulent among other clonotype clinical isolates, which might cause more severe infection symptoms and worse clinical prognosis, indicating that it merits further attention [10]. Thus, our study focused on the important clinical ST398 CA-SA strains. In the transcriptome of ST398, we initially identified RSaX28 with high transcription levels. We further demonstrated the significant correlation between RSaX28 and the virulence of ST398 CA-SA through in vitro hemolytic assays and in vivo murine bloodstream infection experiments. The defined transcription length of RSaX28 is 533 nt in our study, which is also different from that in previous research. We also observed a high degree of conservation in the RSaX28 sequence among the commonly isolated clinical clones, and a remarkable transcriptional upregulation of RSaX28 in the ST398 CA-SA strain. This elevated transcription level may indicate a more crucial regulatory role of RSaX28 in clinical isolates of ST398 CA-SA.

Another highlight of our study is the demonstration of the specific interaction between RSaX28 and RNAIII, which can affect the stability of RNAIII and subsequently influence the downstream effector genes. We initially used IntaRNA to identify the potential interaction between RSaX28 and RNAIII among a series of key global regulatory factors. Subsequently, we confirmed the existence of this interaction through in vitro transcription and EMSA experiments, and determined the binding site of their interaction. In a CLASH (cross-linking, ligation, and sequencing of hybrids) study conducted by McKellar et al, RSaX28's possible role with hld was mentioned [45]. However, the predicted binding site differs from our experimental results, which could be attributed to variations between clinical strains and reference strains or differences in prediction models. The regulation of RNAIII by RSaX28 affects the stability of RNAIII, which in turn reasonably impacts the translation of downstream alpha-toxin. This can explain the contribution of RSaX28 to the virulence of ST398 CA-SA. Additionally, we found that the translation of Sbi downstream was suppressed by RSaX28. This may be the reason for the enhanced biofilm formation ability in the knockout strains. However, during our investigation of spa, we observed a decrease in the translation level of Spa in the knockout strains, which differs from the findings of John et al [39]. This discrepancy could be attributed to the use of different strains. It is worth noting that Spa, as a virulence factor, is also influenced by other regulatory pathways, which further supports the positive impact of RSaX28 on the virulence of ST398 clonotype CA-SA strains [46,47]. Although sRNAs primarily exert posttranscriptional regulatory effects, transcriptomic data still provide valuable insights into the regulatory role of sRNAs. In subsequent transcriptomic analyses, we found that in addition to regulating the virulence and infectivity of S. aureus, RSaX28 also regulates multiple metabolic pathways in the bacterium. This includes several metabolic pathways related to biofilm formation, which further validates the observed significant changes in biofilm formation ability. We demonstrated that RNAIII is a target of RSaX28 regulation, and therefore, we reviewed a study on the transcriptome of RNAIII and the agr system [48]. We discovered certain differences between the RNAIII and RSaX28 knockout strains at the transcriptional level, with major discrepancies observed in changes in metabolic genes. This suggests that future research on RSaX28 needs to focus more on its regulatory role in metabolic genes.

In conclusion, our research has filled the gap regarding the role of RSaX28 in virulence regulation in clinical ST398 clonotype CA-SA. Furthermore, we discovered that RSaX28 exerts its regulatory effects by influencing the stability of RNAIII, thereby expanding the regulatory pathway of RSaX28. We have demonstrated the significant regulatory role of RSaX28 in the virulence and tolerance of ST398 CA-SA, while the prevalence of highly virulent ST398 CA-SA strains has been increasing year by year. We reasonably believe that RSaX28 has the potential to become an important target for the treatment of S. aureus infections and future vaccine research, providing more options for the clinical prevention and control of clinical S. aureus infections. However, it is important to acknowledge the limitations of our study. In addition to RNAIII, there may be other genes that can be bound by RSaX28, thereby exerting subsequent regulatory effects. Furthermore, changes in the translation level of downstream Spa cannot be solely attributed to the influence of RSaX28 on RNAIII. In our follow-up research, we will use multi-omics data such as transcriptomics, proteomics and metabolomics to further elucidate the regulatory pathway of RSaX28. Additionally, we are attempting to use the MS2-Affinity Purification coupled with RNA Sequencing (MAPS) method to directly identify the targets of RSaX28 [49]. We believe that through further research, we will gain a clearer and more comprehensive understanding of the regulatory mechanisms of RSaX28 in various physiological processes of S. aureus.

Author contribution

Ying Jian: Conceptualization, Methodology, Software, Investigation, Formal Analysis, Writing – Original Draft; Tianchi Chen: Conceptualization, Methodology, Software, Investigation, Formal Analysis, Data Curation; Ziyu Yang: Methodology, Software, Visualization, Investigation, Formal Analysis, Data Curation; Guoxiu Xiang: Software, Investigation; Kai Xu: Methodology, Software; Yanan Wang: Funding Acquisition, Methodology, Investigation; Na Zhao: Funding Acquisition, Methodology, Investigation; Lei He: Visualization, Validation, Supervision; Qian Liu: Resources, Supervision; Min Li: Conceptualization, Funding Acquisition, Resources, Supervision, Writing – Review & Editing.

Disclosure statement

No potential conflict of interest was reported by the author(s).

Additional information

Funding

This work was supported by the National Key Research and Development Program (No. 2022YFC2603800); National Natural Science Foundation of China (grant number 82172325, 82102455); Shanghai Sailing Program (21YF1425500, 23YF1423400).