CRISPR-Cas9-based approaches for genetic analysis and epistatic interaction studies in Coxiella burnetii

Abstract

Coxiella burnetii is an obligate intracellular bacterial pathogen that replicates to high numbers in an acidified lysosome-derived vacuole. Intracellular replication requires the Dot/Icm type IVB secretion system, which translocates over 100 different effector proteins into the host cell. Screens employing random transposon mutagenesis have identified several C. burnetii effectors that play an important role in intracellular replication; however, the difficulty in conducting directed mutagenesis has been a barrier to the systematic analysis of effector mutants and to the construction of double mutants to assess epistatic interactions between effectors. Here, two CRISPR-Cas9 technology-based approaches were developed to study C. burnetii phenotypes resulting from targeted gene disruptions. CRISPRi was used to silence gene expression and demonstrated that silencing of effectors or Dot/Icm system components resulted in phenotypes similar to those of transposon insertion mutants. A CRISPR-Cas9-mediated cytosine base editing protocol was developed to generate targeted loss-of-function mutants through the introduction of premature stop codons into C. burnetii genes. Cytosine base editing successfully generated double mutants in a single step. A double mutant deficient in both cig57 and cig2 had a robust and additive intracellular replication defect when compared to either single mutant, which is consistent with Cig57 and Cig2 functioning in independent pathways that both contribute to a vacuole that supports C. burnetii replication. Thus, CRISPR-Cas9-based technologies expand the genetic toolbox for C. burnetii and will facilitate genetic studies aimed at investigating the mechanisms this pathogen uses to replicate inside host cells.

Importance: Understanding the genetic mechanisms that enable C. burnetii to replicate in mammalian host cells has been hampered by the difficulty in making directed mutations. Here, a reliable and efficient system for generating targeted loss-of-function mutations in C. burnetii using a CRISPR-Cas9-assisted base editing approach is described. This technology was applied to make double mutants in C. burnetii that enabled the genetic analysis of two genes that play independent roles in promoting the formation of vacuoles that support intracellular replication. This advance will accelerate the discovery of mechanisms important for C. burnetii host infection and disease.

Keywords: CRISPR; CRISPRi; Coxiella burnetii; base editing; intracellular infection.

Fig 1

A plasmid-based CRISPRi system silences gene expression in C. burnetii, resulting in expected single gene and dual gene knockdown phenotypes in axenic medium. (A) Cartoon representation of CRISPRi-mediated interference of gene expression. The nuclease-deficient dCas9 (light blue) bound to a single guide RNA (sgRNA; red) targeting the non-template strand of a target gene (promoter region or early protein coding sequence) sterically precludes RNA polymerase (RNAP; brown) by acting as a roadblock for transcription initiation and/or elongation. The protospacer and NGG protospacer adjacent motif (PAM) are indicated. (B) Representative brightfield and mCherry fluorescence micrographs of the mCherry-expressing intergenic transposon mutant strain (ig::Tn) harboring individual pCB-CRISPRi plasmid derivatives expressing sgRNAs targeting indicated genes. Scale bars, 10 µm. nt, non-targeting. (C) Representative immunoblot analysis of protein levels of mCherry, Cig2, 3xF-dCas9, and GroEL (loading control) in ig::Tn strain harboring individual pCB-CRISPRi plasmid derivatives expressing sgRNAs targeting indicated genes (single gene or dual gene knockdown). nt, non-targeting. (D) mRNA expression levels (log₂ expression fold change) in axenic medium of genes indicated on the left in ig::Tn strain harboring individual pCB-CRISPRi plasmid derivatives expressing sgRNAs targeting genes indicated by different colors in legend (single gene or dual gene knockdown). Means with SEs from three independent experiments are shown. A strain harboring a pCB-CRISPRi plasmid encoding either a non-targeting sgRNA or a sgRNA targeting a neutral gene was used as control sample, and the 16S rRNA gene served as endogenous reference gene for qRT-PCR and ∆∆Ct analysis. **P < 0.01; ***P < 0.001; ****P < 0.0001 by individual unpaired, two-tailed t tests (in comparison to corresponding control sample each).

Fig 2

CRISPRi-mediated silencing of genes encoding for Dot/Icm secretion system components or effectors alters CCV biogenesis. (A) Representative immunofluorescence micrographs of HeLa cells infected with NMII wild-type harboring individual pCB-CRISPRi plasmid derivatives expressing sgRNAs targeting indicated genes for 5 days at an MOI of 300. Cells were fixed and stained with anti-LAMP-1 antibody (green), anti-C. burnetii antibody (red), and DAPI (blue). Scale bars, 10 µm. nt, non-targeting. (B) Intracellular replication of NMII wild-type harboring individual pCB-CRISPRi plasmid derivatives expressing sgRNAs targeting indicated genes is shown as fold replication (change in genome equivalents (GE) on day 7 post-infection relative to GE on the day of infection). THP-1 cells were infected at an MOI of 0.1. A representative data set is shown. Error bars are SDs. −1 and −2 indicate two separate sgRNAs targeting the same gene. no, pCB-CRISPRi plasmid without sgRNA construct; nt, non-targeting. *P < 0.05; ***P < 0.001; ****P < 0.0001; ns, not significant by one-way ANOVA with Bonferroni’s post hoc test. (C) mRNA expression levels (log₂ expression fold change) in axenic medium of genes indicated on the left in NMII wild-type harboring individual pCB-CRISPRi plasmid derivatives expressing sgRNAs targeting genes indicated on the left. Means with SEs from three independent experiments are shown. Strains harboring either a pCB-CRISPRi plasmid without a sgRNA construct or a pCB-CRISPRi plasmid encoding a non-targeting sgRNA were used as control samples, and the 16S rRNA gene served as endogenous reference gene for qRT-PCR and ∆∆Ct analysis. −1 and −2 indicate two separate sgRNAs targeting the same gene. **P < 0.01; ***P < 0.001; ****P < 0.0001 by individual unpaired, two-tailed t tests (in comparison to corresponding control sample each).

Fig 3

Using CRISPR-Cas9-mediated cytosine base editing to generate C. burnetii loss-of-function mutants by introducing premature stop codons into ORFs. (A) Cartoon representation of CRISPR-Cas9-mediated cytosine base editing. The CBE fusion protein consists of an N-terminal cytidine deaminase (APOBEC1; green), a central Cas9 nickase (Cas9n; light blue), and a C-terminal uracil DNA glycosylase inhibitor domain (UGI; light violet). The 5’ 20 bp sequence of a single guide RNA (sgRNA; red) bound to Cas9n directs the CBE to a complementary target sequence in the genome that is followed by an NGG protospacer adjacent motif (PAM). The cytidine deaminase gains access to the exposed DNA single strand and converts cytosines (C) that are positioned within the editing window to uracils (U). The editing window for the CBE HF-BE3 (76) used here typically is a five-nucleotide window that spans positions 4–8 of the protospacer (highest editing efficiency; counting the PAM as positions 21–23) (50). The UGI domain prevents repair of the deaminated cytidine by the endogenous DNA repair machinery. Aided by the activity of Cas9n that cleaves the non-edited strand, the U:G heteroduplex is permanently converted into a T:A base pair during subsequent DNA replication. The editing window, protospacer, and NGG protospacer adjacent motif (PAM) are indicated. The four different codons that can be converted into premature stop codons based on C to T (= G to A) substitutions mediated by CBEs are listed. The first three codons involve editing of the sense/coding strand; the last codon requires editing of the antisense/non-coding strand. (B and C) In silico genome analysis of ORFs annotated in C. burnetii Nine Mile RSA493 phase I (77) reveals the scope of CBE-mediated introduction of premature stop codons across the genome; 1,833 annotated protein-coding genes were analyzed, including essential genes, but excluding pseudogenes. (B) Histogram showing the total number of genes grouped by the relative position within each gene of their first option for premature stop codon introduction. Percentages shown above bars indicate fraction of total number of genes present in each group; 85% of C. burnetii NMI genes contain at least one option to introduce a premature stop codon. (C) Graph showing the relative position of the first option for premature stop codon introduction within each gene, as a function of gene length; 269 non-editable genes are shown as a function of gene length on top. (D) Plasmid maps of pHelper-CBE-sgRNA (left) and pCB-CBEv1 (right) used for CRISPR-Cas9-mediated cytosine base editing in C. burnetii. pHelper-CBE-sgRNA is a helper plasmid for CBE sgRNA cloning that harbors a sgRNA construct (targeting sequence followed by sgRNA scaffold with Cas9 handle; red) downstream of the synthetic, constitutive bacterial promoter J23119(SpeI) (Registry of Standard Biological Parts). Two BbsI sites allow the insertion of the desired 20 nt targeting sequence as a phosphorylated and annealed oligo duplex. Unique XhoI and NotI sites flank the sgRNA construct for restriction/ligation-mediated subcloning of the sgRNA-encoding region to pCB-CBEv1 (alternatively, the sgRNA-encoding region is moved to pCB-CBEv1 via PCR amplification and e.g. SLIC cloning). pCB-CBEv1 is the C. burnetii CBE plasmid (version 1) without sgRNA construct that encodes the CBE protein 3xF-rAPOBEC1-Cas9n(HF)-UGI (HF-BE3) downstream of the IPTG-inducible Ptac promoter (however, basal expression is sufficient for base editing in C. burnetii using this plasmid). Plasmid contains unique XhoI, NotI, and XbaI sites for the introduction of sgRNA-encoding region(s) from pHelper-CBE-sgRNA helper plasmid derivatives (e.g. via restriction/ligation or SLIC cloning). Within the CBE-encoding gene, the cytidine deaminase-encoding sequence is highlighted in green, and the Cas9n- and UGI-encoding sequences are shown in light blue and light violet, respectively.

Fig 4

C. burnetii premature stop codon mutants constructed using CRISPR-Cas9-mediated cytosine base editing display distinct CCV biogenesis phenotypes. (A) Sanger sequencing chromatograms confirm successful base editing in the C. burnetii genome. Wild-type and edited loci of dotA (top) and cig2 (bottom) are shown. Editable codons in wild-type sequences are shown in green, PAM sites are highlighted in blue, and thymines that were converted from cytosines are shown in red (C to T substitutions). Red horizontal double lines indicate introduced premature stop codons, resulting in Q151* in dotA, and Q45* in cig2. wt, wild-type. (B) Representative immunoblot analysis of protein levels of DotA, Cig2, and GroEL (loading control) in wild-type and indicated base-edited strains. wt, wild-type. (C) Intracellular replication of indicated strains is shown as fold replication (change in genome equivalents (GE) on day 6 post-infection relative to GE on the day of infection). THP-1 cells were infected at an MOI of 0.1. A representative data set is shown. Error bars are SDs. ****P < 0.0001; ns, not significant by one-way ANOVA with Tukey’s post hoc test. wt, wild-type. (D) Representative immunofluorescence micrographs of HeLa cells infected with indicated strains for 6 days at an MOI of 300. Cells were fixed and stained with anti-LAMP-1 antibody (green), anti-C. burnetii antibody (red), and DAPI (blue). Scale bars, 10 µm. wt, wild-type. (E) Quantification of the multivacuole phenotype (see [D]). Graph shows fraction of infected cells with multivacuole phenotype (more than two CCVs/cell) for indicated strains as means with SEs from three independent experiments. HeLa cells were infected for 5–6 days at a high MOI of 100–300. ****P < 0.0001; ns, not significant by one-way ANOVA with Tukey’s post hoc test. wt, wild-type.

Fig 5

Combining premature stop codon mutations in cig57 (Q245*) and in cig2 (Q45*) leads to a severe additive intracellular replication defect that can be complemented. (A and B) Intracellular replication of indicated strains is shown as fold replication (change in genome equivalents (GE) on day 6 post-infection relative to GE on the day of infection). THP-1 cells were infected at an MOI of 0.1. Representative data sets are shown. Error bars are SDs. **P < 0.01; ***P < 0.001; ****P < 0.0001; ns, not significant by one-way ANOVA with Tukey’s post hoc test. wt, wild-type.

Fig 6

Overview of proposed workflow for CRISPR-Cas9-mediated cytosine base editing in C. burnetii using either pCB-CBEv1 (HF-BE3) or pCB-CBEv2 [BE4-PpAPOBEC1(H122A)]. In step 1, divide electroporated cells into two to three parts if independently generated base-edited strains are desired. In step 1, induction of CBE expression with 0.1 mM IPTG is required if using pCB-CBEv2, but not if using pCB-CBEv1 (e.g. for 48 h in liquid ACCM-2 before plating, or for an extended time in liquid ACCM-2 before plating and during growth on ACCM-2 agarose). See the Discussion and Materials and Methods for more details. Cm, chloramphenicol.