Enabling genetic analysis of diverse bacteria with Mobile-CRISPRi

Abstract

The vast majority of bacteria, including human pathogens and microbiome species, lack genetic tools needed to systematically associate genes with phenotypes. This is the major impediment to understanding the fundamental contributions of genes and gene networks to bacterial physiology and human health. Clustered regularly interspaced short palindromic repeats interference (CRISPRi), a versatile method of blocking gene expression using a catalytically inactive Cas9 protein (dCas9) and programmable single guide RNAs, has emerged as a powerful genetic tool to dissect the functions of essential and non-essential genes in species ranging from bacteria to humans^1-6. However, the difficulty of establishing effective CRISPRi systems across bacteria is a major barrier to its widespread use to dissect bacterial gene function. Here, we establish 'Mobile-CRISPRi', a suite of CRISPRi systems that combines modularity, stable genomic integration and ease of transfer to diverse bacteria by conjugation. Focusing predominantly on human pathogens associated with antibiotic resistance, we demonstrate the efficacy of Mobile-CRISPRi in gammaproteobacteria and Bacillales Firmicutes at the individual gene scale, by examining drug-gene synergies, and at the library scale, by systematically phenotyping conditionally essential genes involved in amino acid biosynthesis. Mobile-CRISPRi enables genetic dissection of non-model bacteria, facilitating analyses of microbiome function, antibiotic resistances and sensitivities, and comprehensive screens for host-microorganism interactions.

Fig. 1 |. Mobile-CRISPRi overview.

a, Mechanism of CRISPRi repression. A dCas9-sgRNA complex binds to DNA by base-pairing and sterically blocks progression of RNA polymerase (RNAP), reducing gene expression. b, Mobile-CRISPRi modularity. Individual modules are flanked by unique restriction sites, which can be used for cloning sgRNA libraries or exchanging other components (e.g. antibiotic resistance marker (AB^R), or dcas9 promoter). c, Strain construction using Mobile-CRISPRi. Top: a Tn7 transposon carrying CRISPRi components (shown in b) and a plasmid containing Tn7 transposition genes are transferred to recipient bacteria by tri-parental mating. Donor cells contain a chromosomal copy of the RP4 transfer machinery used to mobilize the Tn7 plasmids. Once inside the recipient cell, Tn7 transposition proteins integrate the CRISPRi DNA (purple) flanked by left and right Tn7 end sequences (green) into the recipient genome downstream of the glmS gene. Selection on antibiotic plates lacking DAP eliminates the E. coli donors and retains recipients with an integrated CRISPRi system. Bottom: An ICE element carrying CRISPRi components is transferred to recipient bacteria by bi-parental mating. Once inside the recipient cell, the ICE integrase inserts ICE into trnS-leu2. Double antibiotic plates that select for ICE and for the intrinsic resistance of the recipient strain (Streptomycin-resistance in this work) are used to identify recipients with an integrated CRISPRi system.

Fig. 2 |. Mobile-CRISPRi stability, transfer, and knockdown efficiency.

a, Mobile-CRISPRi stability after 50 generations of growth in the absence of antibiotic selection. Stability was calculated as the plating efficiency on kanamycin (the marker associated with Mobile-CRISPRi) vs no antibiotic n=4 for B. subtilis and E. coli and n=3 for E. cloacea and K. pneumoniae. b, Mobile-CRISPRi transfer and integration efficiency. ICE or Tn7 containing CRISPRi was transferred to the species listed. Efficiency was calculated as: %AB^R/total recipients n=3 for all strains except V. casei for which n=1. c, Efficiency of Mobile-CRISPRi knockdown in various species. Knockdown was tested using a Mobile-CRISPRi variant containing a constitutively expressed rfp reporter and an sgRNA targeting rfp. RFP expression was normalized to a strain lacking either dcas9 (for P. aeruginosa) or an sgRNA (all others; no sgRNA controls are recommended for future experiments). n=4 for all strains except E. faecalis, P. mirabilis, and V. casei for which n=3 Data are represented as mean ± s.d. The full species names for the strains used in panels b and c and their corresponding rfp fold knockdowns are: Bacillus subtilis (182-fold), Staphylococcus aureus (150-fold), Listeria monocytogenes (105-fold), Enterococcus faecalis (45-fold), Escherichia coli (65-fold), Enterobacter cloacae (32-fold), Enterobacter aerogenes (40-fold), Pseudomonas aeruginosa (8-fold), Klebsiella pneumoniae (34-fold), Vibrio casei (8-fold), Acinetobacter baumannii (10-fold), Salmonella enterica (54-fold), and Proteus mirabilis (35-fold).

Fig. 3 |. CRISPRi knockdown of folA increases sensitivity to trimethoprim in multiple species.

MIC assays for trimethoprim sensitivity in P. aeruginosa and E. aerogenes with or without Mobile-CRISPRi targeting folA (n=3). Data are represented as mean ± s.d.

Fig. 4 |. A Mobile-CRISPRi library targeting auxotrophic genes in E. cloacae.

a, Tn7 Mobile-CRISPRi library construction. sgRNAs were cloned as a pool, transformed, and mated into E. cloacae, or b, sgRNAs were cloned individually, mixed as a pool with equal representation, and mated into E. cloacae as a pool. Representation of individual CRISPRi strains was determined by Illumina sequencing. c, Fitness of CRISPRi strains in glucose minimal media after 6 or 12 doublings with or without CRISPRi induction by IPTG, determined from the library constructed by pooled cloning (n=2). Asterisks indicate strains had no fitness change in pooled screen but decreased fitness in the arrayed screen. Putative auxotrophic or essential gene knockdown strains are indicated next to sgRNA names. d, Correlation between fitness measurements in libraries constructed by pooled cloning or individual cloning (n=2); a linear fit is shown.