A gene-controlled fluorescent CRISPR interference system for Staphylococcus aureus that works without added inducers

Standard agarose, type LE for routine gel electrophoresis was purchased from BioNordika, Oslo, Norway. DNA Ladder, 1.0 Kb Plus (100 bp to 15,000 bp) was purchased from Invitrogen. Primers were obtained from Sigma Life Science (Oslo, Norway), and sequencing reactions were performed using BigDye Terminator V3.1 (Life Technologies) and analyzed at the Faculty of Health Sciences, UiT - The Arctic University of Norway, Tromsø, Norway. The Wizard Genomic DNA Purification Kit was purchased from Promega. E.Z.N.A. Plasmid DNA Mini Kit I (Q – spin) was purchased from Omega. Phusion High-Fidelity DNA Polymerase and DreamTaq Green PCR Master Mix were purchased from Thermo Scientific. All restriction enzymes were purchased from New England Biolabs and used according to the manufacturer’s instructions. MagExtractor-PCR and Gel Clean up were purchased from Toyobo (Osaka, Japan). Rabbit Plasma (for the detection of Staphylocoagulase) was purchased from Bio-Rad. Roswell Park Memorial Institute (RPMI) 1640 medium and heat-inactivated, sterile-filtered Fetal Bovine Serum (FBS), Dulbecco’s Phosphate-Buffered Saline, PBS (D8537), and DNase I were purchased from Sigma-Aldrich. RNAprotect Bacteria Reagent and RNeasy Mini Kit were purchased from QIAGEN (Oslo, Norway). High-Capacity cDNA Reverse Transcription Kit was purchased from Applied Biosystems. All other materials were purchased from an analytical-grade commercial source and used according to the manufacturer’s instructions.

The bacterial strains and plasmids used in this study are listed in Table 1 (S. aureus strains), S1 (Escherichia coli strains), and Table 2 (plasmids). The media used were lysogeny broth (LB; 244620, BD Difco), LB agar (244520, BD Difco), Tryptic Soy Broth (211825, BD Difco), and Tryptic Soy Agar (TSA; 236950, BD Difco). RPMI-1640 was purchased from Sigma-Aldrich (R8758) and used after supplementation with 10% FBS (F7524, Sigma-Aldrich) and is designated as RPMI+.

E. coli cells were grown in LB with shaking or on an LB agar plate at 37°C. Ampicillin was used at a final concentration of 100 µg/mL in LB or agar plates for the selection of recombinant plasmid-transformed E. coli colonies. S. aureus was routinely grown at 37°C on TSA plates or in TSB with shaking (220 rpm). Chloramphenicol (10 µg/mL) was used for the maintenance of the fluorescent reporter plasmids (pCM29- PsarA P1-GFP, pCM29-Patl-GFP, pCM29-PfnbA-GFP, and pCM29-Pcoa-GFP, see next section) in S. aureus. Erythromycin (5 µg/mL) and chloramphenicol (10 µg/mL) were used for the selection and maintenance of pLOW and pCM29 backbone plasmids, respectively, in S. aureus CRISPRi strains. IPTG was used as 250 µM (final concentration) for lac-promoter-induced dCas9 expression in the CRISPRi system. Plasmid pCM29-PsarA P1-GFP was provided by Dr. Alexander Horswill (University of Colorado School of Medicine, Colorado, USA) (31). Chemically competent E. coli IM08B cells were prepared according to reference (32) and routinely used for the transformation of constructed plasmid according to standard heat shock protocol (33). Chemically competent E. coli cells (TOP 10, NEB 5-alpha, NEBExpress I^q, and NEB 10-beta) were transformed according to the manufacturer’s instructions. S. aureus USA 300 LAC was transformed with dCas9 expression CRISPRi plasmid DNA isolated from S. aureus RN4220 (30) and fluorescent reporter plasmid DNA isolated from IM08B (34) by electroporation. The preparation of electrocompetent cells and electroporation were performed as described before (35).

Restriction enzyme digestion and ligation reactions were performed according to the manufacturer’s instructions (New England Biolabs). Briefly, a prepared restriction digestion reaction mixture containing sample DNA (1 µg), 10× rCutSmart buffer (5 µL), and restriction enzyme (10 units) in a total of 50 µL reaction volume with nuclease-free water was incubated at the optimum temperature for the given restriction enzyme for 2 hours. In this work, restriction digestions with BsmBI were carried out at 55°C, while all other restriction digestions were performed at 37°C. For restriction digestions of destination vectors, 1 µL calf intestinal alkaline phosphatase was added to the reaction after ~2 hours of incubation, followed by a subsequent incubation for ~30 min. Digested DNA was verified using agarose gel electrophoresis and purified either from the agarose gel or directly from the digestion reaction using MagExtractor-PCR and Gel Clean up kit (Toyobo, Osaka, Japan).

Digested vectors and inserts were ligated together using the T4 DNA ligase (New England Biolabs). The components of the reaction were mixed with a molar insert:vector ratio of 3:1. The reaction volume was 20 µL, using 2 µL of the supplied 10× reaction buffer and 1 µL T4 DNA ligase. Ligation reactions were carried out overnight at 16°C. Ligation reactions were stored at −20°C or directly transformed into the desired host.

Bacterial growth (OD₆₀₀) and GFP fluorescence were measured in a microplate assay on a Synergy H1 Hybrid Reader (BioTek). Bacterial cells were cultivated at 37°C overnight with shaking (220 rpm). Cell cultures from E. coli WT (IM08B or NEB10-beta), S. aureus USA 300 WT, and their fluorescent reporter strains, MR 1–11 (Table 1; Table S1), and fluorescently labeled (MR 16, 27–32) and non-labeled (MR 15) CRISPRi strains (Table 1) were diluted to an OD₆₀₀ of 0.1 in their respective fresh medium (TSB or RPMI+ for S. aureus and LB for E. coli strains). Two microliter of these diluted overnight cultures was inoculated into 298 µL of each respective fresh medium in a 96-well flat-bottom polystyrene tissue culture plate (Falcon). The plate was incubated directly in a Synergy H1 Hybrid Reader (BioTek) microtiter plate reader at 37°C with constant double orbital shaking (400 rpm for 5 s) in between measurements. Optical density (OD₆₀₀) and GFP fluorescence (excitation 479 and emission 520) were measured at 1-hour intervals for up to 24 hours. GFP fluorescence was recorded as relative fluorescence units (RFUs, i.e., relative to the internal standard in the instrument). The OD₆₀₀ and GFP fluorescence values from three biological replicates that were each derived from three technical replicates were averaged and corrected from blank wells containing only medium for each strain.

The overnight cultures of coagulase-interfering CRISPRi strains MR 23–26 (Table 1) and their respective non-targeting control strains MR 27, 29, and 31 (Table 1) were diluted to an OD₆₀₀ of 0.1 in fresh TSB or RPMI+ and grown at 37°C for 6 hours. Silencing of the coa gene in the Plac-based strain MR23 was induced by adding IPTG to a final concentration of 250 µM. Cells from each culture were diluted in PBS (D8537, Sigma) and adjusted to McFarland 0.5 (i.e., about 10⁸ cells/mL). Subsequently, cells were treated with 2 × volume of RNAprotect Bacteria Reagent (QIAGEN) for 10 min at room temperature and collected by centrifugating for 10 min at 5,000 × g. Lysostaphin, 1 µg/µL, and lysozyme, 10 µg/µL, were added as final concentration to the suspended solution and incubated at 37°C for 30 minutes for lysis of bacterial cells before RNA isolation. RNA extraction was performed using a RNeasy Mini Kit (QIAGEN). The remaining DNA in the isolated RNA was degraded by DNase I (Sigma-Aldrich). DNA-free RNA was subjected to reverse transcription (RT) with the High-Capacity cDNA Reverse Transcription Kit (Applied Biosystems). RNA extraction, DNase treatment, and RT were performed according to the manufacturer’s direction. Synthesized cDNA solution was 10- and 10,000-fold diluted with RNase-free water, and 2 µL of the diluted cDNA solution was subjected to real-time PCR assay performed with PowerTrack SYBR Green Master Mix (Applied Biosystems) containing 0.5 µM each primer set. After cycling, melt curves analysis was performed between 70°C and 90°C. All quantitative PCR (qPCR) data were analyzed using LightCycler 96 system software version 1.1 (Roche Diagnostics). Relative coagulase expression was calculated according to the 2^−∆∆Ct method after normalization by 16S rRNA (rrsA). The real-time PCR primers are listed in Table S2. qPCR was performed in two technical replicates from their three biological replicates.

The overnight cultures of coagulase-interfering CRISPRi strains MR 23–26 (Table 1) and their respective non-targeting control strains MR 27, 29, and 31 (Table 1) were diluted to an OD600 of 0.1 in fresh TSB or RPMI+ and grown at 37°C for 18 hours. Silencing of the coa gene in the control strain MR23 [USA300 LAC carrying pLOW-Plac-dcas9, pCM29-PsarA-P1 gfp-sgRNA (coa); Table 1] was induced by adding IPTG to a final concentration of 250 µM. Cells from each culture were diluted in PBS (D8537, Sigma-Aldrich) and adjusted to McFarland 0.5 (i.e., about 10⁸ cells/mL). Subsequently, 1 mL of each McFarland-adjusted culture was added to 0.5 mL rabbit plasma (Bio-Rad). The samples were incubated in a water bath at 37°C for 4 hours. The level of coagulation was verified by tipping the tubes to a 45° angle. A negative control (NC) sample contained medium only. A test was considered positive if the plasma in the tube formed a coherent clot. All experiments were repeated three times as independent biological replicates to examine the reproducibility.

In the previously published staphylococcal CRISPRi-system, dcas9 [S. pyogenes (SpydCas9)] expression is controlled by an IPTG-inducible lac promoter in the low copy number pLOW-dcas9 plasmid, and sgRNA is expressed by constitutive P3 promoter in the vector pVL2336-sgRNA (or equivalent vector pCG248-sgRNA) (9, 11, 12). To generate an inducer-independent system for S. aureus, we replaced the lac promoter with S. aureus-specific virulence gene promoters in the pLOW plasmid (Fig. 1A through C). We selected well-known virulence gene promoters controlling the expression of genes encoding autolysin (Patl), fibronectin-binding protein A (PfnbA), and coagulase (Pcoa). Since the activity levels of endogenous virulence gene promoters are dynamic, there is condition-dependent variation as to when the promoter is activated and dcas9 is expressed, which is a precondition for the activity of the CRISPRi system. To address this question, we incorporated a fluorescent reporter gene into the CRISPRi system by placing gfp (into the sgRNA-expressing plasmid) under the control by the same promoter as dcas9.

During the generation of these constructs, we encountered technical challenges caused by the unexpected activity of S. aureus gene promoters in E. coli (Fig. S1) prompting dCas9-related toxicity and plasmid instability in the E. coli cloning host IM08B (Fig. S2), a strain that methylates the plasmids to conveniently allow direct transformation into many S. aureus lineages including USA300 LAC. These pitfalls were bypassed by identifying E. coli NEB10-beta as an alternative cloning host that is less susceptible to dCas9-mediated toxicity (Fig. S2). Isolated plasmids from NEB10-B could then be successfully electroporated into S. aureus RN4220 (to enable correct methylation), repurified, and electroporated into the clinical MRSA-isolate USA300 LAC, providing a viable strategy for generating S. aureus gene-promoter-driven CRISPRi constructs in S. aureus.

We first evaluated dynamic differences in the selected S. aureus promotors under different growth phases and media conditions. We recorded 24-hour-growth curves and fluorescent signals of pCM29-GFP-based fluorescent reporter strains (strains MR 9–11, Table 1) in default bacteriological culture media (TSB) as well as RPMI-1640 supplemented with 10% FBS (RPMI+), a mammalian cell culture medium that mimics the composition of human body fluids (36, 37). Compared to a PsarAP1 reporter, all strains showed low levels of fluorescence in TSB that was increasing throughout the exponential and stationary phase (Fig. 2). Interestingly, the Pcoa- reporter strain showed a greater than sixfold increase in GFP signal during exponential growth in RPMI+ indicating the differential activity of these promoters in different environments (Fig. 2). Since Pcoa showed the highest activity levels across the conditions tested, we considered it a suitable promoter for a proof-of-principle experiments with the vgp-CRISPRi system.

Classical CRISPRi systems are based on the IPTG-inducible dcas9 expression pLOW plasmid and sgRNA-target sequence on pVL2236 plasmid (strain MR15; Table 1). We first tested if the pCM29-PsarA-P1-gfp-based sgRNA plasmids were functional (like the pVL2236 plasmid) in combination with the pLOW plasmid in strain MR16 (Table 1). We chose to target the gene encoding the essential protein PBP1, which has been reported to result in a growth inhibition phenotype upon induction (11). After IPTG induction, a characteristic growth halt for up to 10 hours was observed for both strains, MR15 (Fig. 3A) and MR16 (Fig. 3B). Because strain MR 16 exhibited a similar outcome to MR 15, it indicates that the pCM29-PsarA-P1-gfp-based sgRNA plasmid is also functional, like the pVL2236 plasmid, in the IPTG-inducible system. Interestingly, growth resumption coincided with a decline in GFP signal (controlled by PsarA-P1) after IPTG induction in strain MR16 (Fig. 3B).

Finally, to investigate if dcas9 expression under control of virulence gene promoters leads to similar growth arrest when sgRNA is targeting pbp1, we generated a CRISPRi strain (MR32; Table 1) carrying pLOW-Pcoa-dcas9 and pCM29-Pcoa-gfp-sgRNA(pbp1), in which dcas9 and gfp are controlled by the coa-promoter, whereas sgRNA expression is driven by the constitutive P3 promoter (Fig. 1). Indeed, this strain (MR32) also showed pronounced growth inhibition with declined GFP signal in both TSB (Fig. 3C) and RPMI+ (Fig. 3D) when compared to a non-target control strain (MR31; Table 1). CRISPRi constructs where dcas9 (and gfp) was controlled by the Patl and PfnbA and showed similar temporary growth arrest with declined GFP signal when used in combination with sgRNA (pbp1; Fig. S3). Collectively, these observations can be attributed to the emergence of loss-of-function mutations in the CRISPRi system that are commonly observed when targeting essential genes (6, 8, 38).

Of note, the growth patterns of strains with a non-targeted sgRNA of the classical IPTG-inducible system (MR17 and 18; Table 1) and the vgp-CRISPRi system (MR 27, 29, and 31, Table 1) in both TSB and RPMI+ medium were identical, regardless of whether dCas9 expression was induced with IPTG or not, or which promoter was used to control dCas9 expression (Fig. S4). These data suggest that dCas9 expression per se does not affect the growth of S. aureus. However, compared to the WT, all mentioned CRISPRi strains showed a noticeable growth delay in both TSB and RPMI and a reduction in growth rate in TSB (Fig. S4), which might be attributed to plasmid carriage and growth in the presence of antibiotics selection.

Finally, we aimed to demonstrate that vgp-CRISPRi is suitable to interfere with specific virulence-associated phenotypes. We focused on coagulase, which converts fibrinogen to fibrin thus inducing clotting of blood plasma. We constructed CRISPRi strains with coa as the sgRNA target and dCas9 expression under the control of the Plac, Patl, PfnbA, and Pcoa promoters, respectively (MR23–26; Table 1). We first analyzed the effect on gene expression. qPCR analysis revealed that vgp-CRISPRi strains in which dcas9 expression was driven by either Patl, PfnbA, or Pcoa P showed reduced coa levels that were similar to those achieved with the IPTG inducible system in both TSB and RPMI+ (Fig. 4). Having confirmed the effect on gene expression, we proceeded to test if silencing of the coa gene was sufficient to interfere with the biological function of coagulase. While the addition of WT bacteria to rabbit plasma induced visible clots after 4 hours, no clotting was observed for either of these inducer-free constructs (nor the IPTG-inducible CRISPRi strains with IPTG induction) suggesting successful inhibition of coagulation by knock-down of the coa gene (Fig. 5). Non-target controls showed coagulation levels comparable to the WT. The same outcome was achieved regardless of whether bacteria were originally cultivated in TSB or RPMI+, suggesting that all tested promoters were sufficiently activated for the functioning of vgp-CRISPRi system.