Materials and Methods

Virulence of the PathogenPorphyromonas gingivalisIs Controlled by the CRISPR-Cas Protein Cas3

MATERIALS AND METHODSBacterial growth conditions. Porphyromonas gingivalis ATCC 33277 was cultured anaerobically at 37°C. Cells were maintained on brain heart infusion (BHI)-blood agar plates supplemented with 5 μg/ml hemin and 1 μg/ml menadione (vitamin K). Two different liquid growth media were used in the experiments. For the construction of the mutant and the infection experiments, we used BHI broth supplemented with 1 mg/ml yeast extract, 5 μg/ml hemin, and 1 μg/ml menadione. For the transcriptome experiments, we grew P. gingivalis in 20% heat-inactivated human serum (Sigma-Aldrich H3667-20ML) supplemented with 5 μg/ml hemin and 1 μg/ml menadione as described previously by Grenier et al. (81).Construction of a CRISPR-Cas3 gene knockout strain of Porphyromonas gingivalis. We constructed the fig|431947.7.peg.1929 (PATRIC annotation 82) (see Fig. S1a in the supplemental material) gene knockout strain by replacing the whole gene with an erythromycin resistance cassette. First, we constructed a plasmid (pUC19) carrying the erythromycin resistance cassette (ermF gene from pVA2198) flanked by the 1-kb region upstream and downstream of fig|431947.7.peg.1929. The fig|431947.7.peg.1929 gene was replaced by the erythromycin cassette and was kept in frame with the rest of the genes in the operon. The construct was made in such a way that the complete fig|431947.7.peg.1929 gene was replaced by the erythromycin cassette and was kept in frame with the rest of the genes in the operon. This construct was made using a NEBuilder HiFi DNA assembly kit. The kit was used with 100 ng of P. gingivalis genomic DNA according to the manufacturer's protocol. The NEBuilder Assembly Tool (NEB) was used to design the primers for the NEBuilder HiFi DNA kit protocol. Dh5-alpha chemical competent cells were used for the transformation. The resulting plasmid-transformed Dh5-alpha cells were selected on ampicillin LB plates. The plasmid was extracted using EZNA plasmid DNA minikit II (Omega) and sequenced. After verification by sequencing, the plasmid was named pFLUF001.Primers JS_Cas3KOPCR1-F (5′-GGAAGTGACCGTTATCGAAGAT-3′) and JS_Cas3KOPCR1-R (5′-GCCTTACGAATAGGCCATAAGA-3′) were used to amplify, from pFLUF001, the 2.7-kb DNA fragment containing the erythromycin cassette and its flanking regions. Pfu polymerase (Fermentas) was used according to the manufacturer’s protocol. The amplified fragment was cleaned using an EZNA gel extraction kit (Omega) and used for electroporation of P. gingivalis electrocompetent cells. P. gingivalis ATCC 33277 electrocompetent cells were made by growing the cells on tryptic soy broth (TSB) supplemented with hemin and vitamin K to an optical density at 600 nm (OD600) of 0.6 to 0.7. After centrifugation, the cells were washed twice in ice-cold electroporation buffer (10% glycerol, 1 mM MgCl2) and finally resuspended in a minimal amount of buffer. A 100-μl sample of P. gingivalis electrocompetent cells was electroporated with different amounts of the purified DNA fragment. Cells were plated on BHI-blood agar plates supplemented with hemin and vitamin K and 10 μg/ml of erythromycin. After anaerobic incubation at 37°C for 7 days, the resulting colonies were streaked on new plates, and single colonies were obtained. The fig|431947.7.peg.1929 (Cas3) gene knockout strain was verified by colony PCR using primers from the erythromycin cassette, and the adjacent region to the flanking region was amplified and cloned in pFLUF001. The amplified product was confirmed by sequencing. The correct gene knockout strain was grown on liquid media, and glycerol and dimethyl sulfoxide (DMSO) stocks were prepared and stored in an −80°C freezer.THP-1 cell culture. As described previously (83), human monocyte/macrophage cell line THP-1 (ATCC, TIB-202) was cultured at 37°C in a 5% CO2 incubator in RPMI 1640 supplemented with 10% heat-inactivated fetal bovine serum, l-glutamine (2 mM), penicillin/streptomycin (100 U/100 μg/ml), HEPES (10 mM), sodium pyruvate (1 mM), glucose (4.5 mg/ml), sodium bicarbonate (1.5 mg/ml), and 2-mercaptoethanol (Sigma-Aldrich) (0.05 mM). Cells were adjusted to 5 × 105 viable cells/ml, and, to induce differentiation into a macrophage-like state, THP-1 cells were placed into fresh medium containing 100 ng/ml phorbol 12-myristate 13-acetate (PMA; Sigma-Aldrich) and 1 ml of THP-1 cells was added to each well of 24-well cell culture plates. After 48 h of incubation, the cell culture medium was replaced by antibiotic-free medium, and cells were used in challenge studies.Bacterial infection experiments. P. gingivalis wild-type and Δcas3 mutant cells were harvested from the BHI broth culture by centrifugation, and the bacteria were washed three times in RPMI 1640 medium and adjusted to an OD660 of 1.0 (approximately 1 × 109 CFU/ml) and added to PMA-activated THP-1 cells at a multiplicity of infection of 100, in the antibiotic-free medium, and after 2 h of infection, the THP-1 cells were washed to remove nonadherent bacteria. After 2 h of incubation, the THP-1 cells were treated for 1 h with metronidazole (200 μg/ml)/gentamicin (300 μg/ml) to kill extracellular bacteria (51), were washed to remove antibiotics, and then were lysed with sterile water, followed by addition of an equal volume of 2× phosphate-buffered saline (PBS). Serial 10-fold dilutions of each lysate were plated on blood agar plates. CFU/ml of P. gingivalis was calculated. In separate experiments, THP-1 cells were cultured with the P. gingivalis ATCC 33277 wild-type strain or the Δcas3 mutant, and both cell culture supernatant fluids and RNA were harvested for measurement of cytokine expression and RNA was extracted for dual transcription. Differences in bacterial counts were assessed by performing a Kruskal-Wallis test corrected for multiple comparisons using the function ‘kruskal’ from the ‘agricolae’ package in R, with a false-discovery rate (FDR) value of <0.05.RNA extraction. Total RNA was extracted from those samples using a mirVana RNA isolation kit (Life Technologies). Samples were bead-beaten for 1 min at maximum speed with 300 μl of 0.1-mm diethyl pyrocarbonate (DEPC)-treated zirconia-silica beads (BioSpec Products) in the mirVana lysis buffer. Bacterial rRNA was depleted using Ribo-Zero gold rRNA removal kits (Illumina) following the manufacturer's protocol. In the THP-1 infection experiments, sequencing of total RNA, including human and bacterial mRNA, was performed. Their identification was performed bioinformatically, as described below.G. mellonella RNA was extracted using a mirVana RNA extraction kit (Thermo Fisher Scientific) with minor modifications. Individual larvae were washed with autoclaved PBS and were cut at the distal end. The internal content was snap-frozen in liquid nitrogen, homogenized and pulverized, and immediately incubated at room temperature for 1 h in lysis buffer (mirVana kit; Thermo Fisher Scientific). After the lysis steps, the manufacturer’s instructions were followed.In the experiments performed for analysis of G. mellonella infection, we divided the total RNA into two subsamples. One half was used for transcriptome analysis of G. mellonella, for which we used a Dynabeads mRNA purification kit to isolate eukaryotic mRNA for transcriptome analysis. The other half was depleted of eukaryotic mRNA using a MICROBEnrich kit (Thermo Fisher Scientific, catalog no. AM1901).RT-qPCR quantification of cas3 transcripts. From the initial RNA extraction, possible traces of DNA were removed using Ambion’s Turbo DNA-free kit (Ambion) following the manufacturer’s instructions. The volume of Turbo DNase I (Ambion’s Turbo DNA-free) was increased to 3 μl, and the reaction mixture was incubated at 37°C for 60 min. RNA (100 ng) was reverse transcribed with random hexamer primers and with SuperScript II reverse transcriptase (Invitrogen) following the manufacturer's instructions. Reverse transcription was performed at 42°C for 2 h, after an initial incubation step of 10 min at 25°C. The synthesized cDNA was used in a DyNAmo Flash SYBR green qPCR kit (New England Biolabs, Ipswich, MA) according to manufacturer’s instructions, using the specific primers for the genes of interest. To compare the relative expression levels of genes, we modified the threshold cycle (2−ΔΔCT) method (84), and we used the formula cDNAmutant/cDNAcontrol = (1 + EcDNAcontrol Only two)CtcDNAcontrol/(1 + E cDNAmutant)CtcDNAmutant to take into consideration the different amplification efficiencies in the separate qPCR runs.RNA-seq library construction. Sequencing was performed at the Interdisciplinary Center for Biotechnology Research (ICBR) at the University of Florida using a HiSeq 2500 machine. First, rRNAs were removed from total RNA by the use of an Illumina Ribo-Zero gold rRNA removal kit following the manufacturer’s protocol and eluted into 10 μl EB buffer. The transcriptome sequencing (RNA-seq) library was then processed using a NEBNext ultradirectional RNA library prep kit for Illumina (NEB, USA) following the manufacturer’s recommendations. A 5-μl volume of depleted RNA mix was used together with 5 μl of first-strand synthesis reaction mix (NEBNext first-strand synthesis reaction buffer and NEBNext random primers), fragmented by heating at 94°C for the desired time. This step is followed by first-strand cDNA synthesis performed using reverse transcriptase and oligo(dT) primers. Synthesis of double-stranded (ds) cDNA is performed using the second-strand master mix provided in the kit, followed by end repair and adaptor ligation. Finally, the library is enriched (each library has a unique barcode, and each primer has a common adaptor sequence which was added in the previous adaptor ligation step and a unique overhang index unique to each sample) by a certain number of cycles of amplification and purified by the use of Agencourt AMPure beads (Beckman Coulter, catalog no. A63881). Finally, individual libraries were pooled with equimolar volumes and sequenced by the use of an Illumina HiSeq 3000 system (Illumina Inc., CA, US) and a run of 2 × 100 cycles.Illumina instrument run. In preparation for sequencing, barcoded libraries were sized on an Agilent 2200 TapeStation system. Quantitation was done by the use of both QUBIT and qPCR (Kapa Biosystems, catalog no. KK4824). Individual samples were pooled in equimolar volumes at 2.5 nM. This “working pool” was used as the input in the HiSeq 3000 instrument sample preparation protocol (Illumina material no. 20015630, document no. 15066496 v04, January 2017). Typically, a 250 pM library concentration was used for clustering on a cBOT amplification system. This resulted in an optimum clustering density at which the proportions of clusters passing the filters ranged between 65% and 75%. Six RNA-seq barcoded libraries were pooled for sequencing in multiplex on a single flow cell lane, using a configuration of 2 × 100 (paired-end) cycles. Such an sequencing configuration was achieved by pooling the reagents from 150-cycle and 50-cycle Illumina HiSeq 3000 SBS kits. A typical sequencing run in the HiSeq 3000 instrument produced >300 million reads from each end, per lane, with a Q30 value of ≥90%. For RNA-seq, the use of 50 million reads per end per sample provided sufficient depth for transcriptome analysis. The sequencing run was performed at the NextGen research facility of the Interdisciplinary Center for Biotechnology Research (University of Florida).Assessment of cytokine and chemokine production. Frozen cell culture supernatant fluids were thawed, and the levels of TNF-α, IL-1β, IL-6, IL-8, IL-10, and RANTES were determined by the use of Milliplex multiplex assays (EMD, Millipore). Data were acquired on a Luminex 200 system running xPONENT 3.1 software (Luminex) and analyzed using a 5-parameter logistic spline-curve fitting method and Milliplex Analyst V5.1 software (Vigene Tech). Statistical differences were assessed by two-way analysis of variance (ANOVA) using the ‘emmeans’ package in R, applying an FDR value of <0.05 for multiple-comparison corrections. Experiments were performed in triplicate.Galleria mellonella infection model. For all of the G. mellonella experiments, insects in the final instar larval stage were purchased from Vanderhorst, Inc. (St. Marys, OH). Upon arrival, any dead larvae present were separated from healthy larvae, which were then weighed and placed randomly into groups and kept at room temperature until the next day, when infection was performed. Seven groups of 15 larvae, ranging in weight from 200 to 300 mg, and with no signs of melanization, were randomly chosen and used for subsequent infection. A 25-μl Hamilton syringe was used to inject 5-μl aliquots of bacterial inoculum into each larva's hemocoel via the last left proleg. Three groups received wild-type P. gingivalis (3.85 × 108, 7.7 × 107, and 3.85 × 106 CFU per larvae), and three groups received the cas3 mutant (2.15 × 108, 4.3 × 107, and 2.15 × 106 CFU per larvae). Three control groups included THSB medium alone, tryptic soy broth (TSB) (BD, Becton, Dickinson and Co.) plus P. gingivalis wild-type heat-killed (10 min at 75°C), and THSB plus Δcas3 mutant heat-killed (10 min at 75°C). After injection, larvae were incubated at 37°C, and the appearance of melanization and survival were recorded at 0.5, 1, 1.5, 2, 3, 3.75, 4.25, 6.25, 22, 24, 28, and 42 h. After injection, larvae were incubated in the dark at 37°C, and appearance (signs of melanization) and survival were recorded at selected intervals. Larvae were scored as dead when they displayed no movement in response to touch. Kaplan-Meier killing curves were plotted, and estimations of differences in survival were compared using a log-rank test. A P value of ≤0.05 was considered significant. All data were analyzed with the ‘survival’ and ‘survminer’ packages in R. Experiments were repeated three independent times with similar results.We followed the protocol described above for the infection transcriptome experiments, but the initial concentrations of P. gingivalis injected were 7.0 × 108 CFU/ml for the wild-type strain and 1.0 × 108 CFU/ml for the mutant strain.Host-bacterium transcriptome analysis. We used the PATRIC annotation for genome identifier (ID) 431947.7 of P. gingivalis 33277 (82) for our study. Low-quality sequences were removed from the query files using Trimmomatic (85). Cleaned data were aligned against the P. gingivalis ATCC 33277 genome database using bowtie2 (86). Eukaryotic sequences, human and G. mellonella, were aligned against genome release 33 (GRCh38.p13) in GenCode ( and RefSeq assembly accession no. GCF_003640425.2, respectively. Alignment for eukaryotic sequences was performed using STAR (87). Read counts from the BAM files were obtained using featureCounts (88).In the case of the THP-1 infection experiments, differential expression analysis was performed using NOISeqBio (89). After exploratory analysis with NOISeqBio, we selected reads per kilobase per million (RPKM) normalization for the THP-1 cell data, tmm normalization with length correction for the P. gingivalis intracellular transcriptome, and tmm normalization without length correction for the P. gingivalis planktonic transcriptome. Posterior analyses were used only with significant differential expression and log changes >2.The genome of G. mellonella has been sequenced and contains 14,067 protein-coding genes (90). Time-series analysis of the G. mellonella infection transcriptomes was performed in two steps. First, we identified differentially expressed genes using a DESeq2 spline approach, as recommended by Spies et al. for short-time series data (<8 time points) (91). In the second step, we clustered those genes based on their trajectories during the experimental period. For the following step, we used the Dirichlet process Gaussian process mixture model (DPGP) and DP_GP cluster software with 500 iterations (92).We used Gene Ontology (GO) terms for gene set enrichment analysis (GSEA). In the case of P. gingivalis and G. mellonella, we generated our own GO annotation using the pipeline in OmicsBox (93). The GO annotation was extracted for the human genome directly from its gff3 file. GSEA was performed using GOrilla (94) in the case of the human sequences and OmicsBox for any other enrichment analysis, including enrichment using Fisher’s exact test. In all cases, we consider an FDR value of <0.05. Summaries of GO terms and GO treemaps were obtained using REVIGO (95). GSEA results corresponding to enriched terms from OmixBox, quantified as percentages of sequences in the reference and test sets, were represented as heat maps using the ‘pheatmap’ package in R, after normalization with the function ‘decostand’ from the ‘vegan’ package.Coexpression networks of host and P. gingivalis genes. The integration of host-microbe expression profiles was performed using the R package ‘mixOmics’ (96). We calculated the sparse partial least-square (sPLS) correlations between the differentially expressed genes from eukaryotic cells and the different mutants of P. gingivalis. Relevance networks showing correlations between genes from eukaryotic cells and microorganisms were visualized in Cytoscape (97) with a threshold correlation of 0.90.We used the Cytoscape plugin Diffany (41) to obtain and analyze consensus and differential coexpression networks. On those networks, we visualized enriched GO terms on the G. mellonella genes at the biological process level using the Cytoscape plugin GOlorize (98), which uses the Cytoscape BiNGO (99) plugin to perform GSEA on the nodes of the network. For GSEA performed on BiNGO, we generated the BiNGO annotation based on the GO annotation described above. The tests were performed using the default settings with an FDR significance value of <0.05. The selection of the enriched terms to be visualized was performed by summarizing the results with REVIGO and selecting for the representative GO terms. In the case of P. gingivalis genes present in the networks, we associated them with GO terms, removed the singletons, and summarized the results using REVIGO.Data availability. The data sets used in these analyses were deposited at the Gene Expression Omnibus (GEO) data repository of NCBI. Sequences have been deposited at the GEO database ( with submission number {"type":"entrez-geo","attrs":{"text":"GSE154569","term_id":"154569"}}GSE154569.

Article TitleVirulence of the PathogenPorphyromonas gingivalisIs Controlled by the CRISPR-Cas Protein Cas3


The data sets used in these analyses were deposited at the Gene Expression Omnibus (GEO) data repository of NCBI. Sequences have been deposited at the GEO database ( with submission number{"type":"entrez-geo","attrs":{"text":"GSE154569","term_id":"154569"}}GSE154569.

Login or Signup to leave a comment
Find your community. Ask questions. Science is better when we troubleshoot together.
Find your community. Ask questions. Science is better when we troubleshoot together.

Have a question?

Contact or check out our support page.