An efficient in vivo-inducible CRISPR interference system for group A Streptococcus genetic analysis and pathogenesis studies

Abstract

While genome-wide transposon mutagenesis screens have identified numerous essential genes in the significant human pathogen Streptococcus pyogenes (group A Streptococcus or GAS), many of their functions remain elusive. This knowledge gap is attributed in part to the limited molecular toolbox for controlling GAS gene expression and the bacterium's poor genetic transformability. CRISPR interference (CRISPRi), using catalytically inactive GAS Cas9 (dCas9), is a powerful approach to specifically repress gene expression in both bacteria and eukaryotes, but ironically, it has never been harnessed for controlled gene expression in GAS. In this study, we present a highly transformable and fully virulent serotype M1T1 GAS strain and introduce a doxycycline-inducible CRISPRi system for efficient repression of bacterial gene expression. We demonstrate highly efficient, oligo-based single guide RNA cloning directly to GAS, enabling the construction of a gene knockdown strain in just 2 days, in contrast to the several weeks typically required. The system is shown to be titratable and functional both in vitro and in vivo using a murine model of GAS infection. Furthermore, we provide direct in vivo evidence that the expression of the conserved cell division gene ftsZ is essential for GAS virulence, highlighting its promise as a target for emerging FtsZ inhibitors. Finally, we introduce SpyBrowse (https://veeninglab.com/SpyBrowse), a comprehensive and user-friendly online resource for visually inspecting and exploring GAS genetic features. The tools and methodologies described in this work are poised to facilitate fundamental research in GAS, contribute to vaccine development, and aid in the discovery of antibiotic targets.

Importance: While group A Streptococcus (GAS) remains a predominant cause of bacterial infections worldwide, there are limited genetic tools available to study its basic cell biology. Here, we bridge this gap by creating a highly transformable, fully virulent M1T1 GAS strain. In addition, we established a tight and titratable doxycycline-inducible system and developed CRISPR interference (CRISPRi) for controlled gene expression in GAS. We show that CRISPRi is functional in vivo in a mouse infection model. Additionally, we present SpyBrowse, an intuitive and accessible genome browser (https://veeninglab.com/SpyBrowse). Overall, this work overcomes significant technical challenges of working with GAS and, together with SpyBrowse, represents a valuable resource for researchers in the GAS field.

Keywords: CRISPRi; SpyBrowse; Streptococcus pyogenes; genetic toolbox; group A Streptococcus; infectious disease.

Fig 1

Golden Gate assembly and recombineering to create a highly transformable hsdR mutant in M1T1 GAS strain 5448. (A) Schematic representation of the genomic organization of the hsdRSM locus in WT GAS strain M1 5448 (NV1). Gene annotations were made by Prokka (25). (B) Three PCRs were performed each containing unique BsaI restriction sites that allow for scarless Golden Gate assembly of a linear fragment in which hsdR is replaced by an erythromycin resistance cassette (ery). PCR 1 contains 1,392 bp upstream of hsdR (including the garK gene and excluding the start codon of hsdR), PCR 2 contains the 785 bp promoterless and terminatorless ery cassette (23) flanked by BsaI sites, while PCR 3 contains a 1,238 bp downstream region of hsdR (including its stop codon and hsdS). After Golden Gate assembly with BsaI and T4 ligase (see Materials and Methods), the resulting 3,383-bp ligation product was purified and directly used to transform electrocompetent NV2 (NV1 + pAV488) cells that were grown with 1 mM of β-D-1-thiogalactopyranoside to induce the recombineering system present on plasmid pAV488 resulting in strain NV3 (NV2, hsdR::ery). Recombineering of GAS using pAV488 is described in more detail elsewhere (Andrew Varble, unpublished data). Strain NV3 was plasmid cured resulting in strain NV28 (NV1, hsdR::ery). The genomic organization of the hsdRSM locus in strain NV28 (hsdR::ery) is shown. (C) Methylation motifs identified in strains NV1, NV28, and NV6 using SMRT sequencing (see Materials and Methods for details). (D) Transformation efficiencies of WT GAS 5448 (NV1) and the hsdR::ery mutant (NV28) are shown. Left: strains were transformed with 140 ng (~50 fmol) of plasmid pDC123 (24) that does not contain an HsdR-motif. Right: strains were transformed with 247 ng (~95 fmol) of plasmid pDC-sgRNA (see Fig. 2C) that contains a single HsdR-motif. Each dot represents a replicate, and a Kolmogorov-Smirnov test was used to calculate statistically significant differences between the two strains (**P value < 0.005).

Fig 2

Development of a doxycycline-inducible CRISPRi system for use in group A Streptococcus. (A) Schematic representation of CRISPRi. An sgRNA encoding a 20 nt spacer region targets dCas9 to the non-template strand containing a protospacer adjacent motif (PAM) on the GAS ftsZ locus. Transcription elongation by RNA polymerase (RNAP) is consequently hampered leading to reduced levels of FtsZ in the cell when dCas9 is induced. (B) Schematic representation of the genomic organization of the cas9 locus in wild-type GAS strain 5448 (NV1), in strain NV4 (cas9::kan) and strain NV6 (cas9::tetM, tetR, Ptet-dcas9). (C) Schematic representation of sgRNA cloning vector pDC-sgRNA. (D) Strain NV9 (NV1, pDC-sgRNA) was grown in Todd Hewitt broth containing 2 µg/mL of chloramphenicol at 37°C, and exponentially growing cells were imaged by differential interference contrast (DIC) and fluorescence microscopy. Scale bar: 2 µm. (E) Oligo-based sgRNA cloning in pDC-sgRNA. Plasmid pDC-sgRNA is cut with Esp3I (or its isoschizomer BsmBI), and its resulting sticky ends are shown. Two complementary oligos of each 24 nt long that include a 20-bp spacer sequence are annealed, phosphorylated, ligated, and electroporated to competent GAS (see Materials and Methods). As an example, the two oligos used to clone the sgRNA targeting GAS ftsZ are shown (for oligo design, see Table S1). Successful clones lost the mCherry cassette and will be white on the plate instead of pink.

Fig 3

SpyBrowse is available at https://veeninglab.com/SpyBrowse. (A) A screenshot of the emm locus as shown in SpyBrowse. In the left pane, tracks can be turned on/off. In the right pane, the genome can be browsed by dragging the mouse to the left or right, zooming in and out, or searched on gene name and/or locus tags (e.g., Spy1719 or U2W45_08500 for emm). Annotated features such as genes, rRNAs, and tRNAs are displayed. The designed spacer targeting the non-template strand of emm (JFONFGPK_01746_gR1311f) is shown via the sgRNA track. (B) For each coding sequence, a context menu provides links to external resources, such as Uniprot (with Alphafold prediction), NCBI (with Blast function), and PaperBlast. (C) The predicted Alphafold structure of the M1 protein through the Uniprot link in SpyBrowse is shown. (D) The NCBI entry for the M1 protein is shown. (E) The Paperblast hits for the M1 protein are shown.

Fig 4

Targeted repression of dnaA and ftsZ gene expression by CRISPRi in GAS. (A) Strains NV26 (NV6 + pDC-sgRNA-control), NV22 (NV6 + pDC-sgRNA-dnaA), and NV19 (NV6 + pDC-sgRNA-ftsZ) were grown in THB containing 2 µg/mL of chloramphenicol at 37°C. Exponentially growing cells were diluted to a start OD of 0.004 in a microtiter plate containing fresh THB with varying concentrations of doxycycline (doxy). The optical density at 595 nm was measured every 10 min. Each line is an average of three replicates with the standard deviation shown. For clarity, the optical density is plotted on a linear scale. (B) Strains NV22 and NV19 were grown in the presence or absence of 20 ng/mL of doxy, and after 3 h, cells were imaged by fluorescence microscopy. DAPI was used to stain the nucleoids and Nile red to stain the membrane. Scale bar: 2 µm. Arrows point to anucleate cells (sgRNA-dnaA) or cells with a block in division (sgRNA-ftsZ).

Fig 5

Efficient depletion of M protein and in vivo CRISPRi in GAS. (A) Strain NV25 (NV6 + pDC-sgRNA-emm) was grown in THB containing 2 µg/mL of chloramphenicol at 37°C in microtiter plates in the presence of various concentrations of doxycycline. (B) Strains NV25 (sgRNA-emm), NV26 (sgRNA-control), and M protein knockout GAS M1T1 5448 were grown in THB to mid-logarithmic growth in the presence or absence of 20 ng/mL doxycycline. Bacteria were immunostained using M protein antisera and analyzed by flow cytometry as described in Materials and Methods. (C) Strains NV1 (wild-type GAS 5448) and NV6 (cas9::tetM, tetR, Ptet-dcas9, hsdR::ery) were grown in THB at 37°C, and 1–3 × 10⁸ colony-forming units (CFU) were used to infect CD-1 mice IP (10 mice per group). Disease score and survival were followed for 14 days. There was no statistically significant (n.s.) difference in virulence between both strains (Mantel-Cox Log Rank test). (D) Strains NV26 (NV6 + pDC-sgRNA-control) and NV19 (NV6 + pDC-sgRNA-ftsZ) were grown in THB containing 2 µg/mL of chloramphenicol at 37°C, and 1–3 × 10⁸ CFU were used to infect CD-1 mice IP (10 mice per group). Disease score and survival were followed for 14 days. Doxy-induced mice infected with NV19 showed a statistically significant increased survival (**P = 0.0027, Mantel-Cox Log Rank test).