Changes in zebrafish gene expression upon pneumococcal infection

In our previous studies we reported that the intravenous injection of pneumococcus induces a common inflammatory response in zebrafish larvae and that the circulating bacteria are cleared by the phagocytosing leukocytes [28]. In the present study, we further characterized the innate immune response to pneumococcus in zebrafish larvae through a whole-genome level transcriptome analysis. For the analysis, 2 dpf (days post fertilization) zebrafish larvae were infected with 466–542 colony forming units (cfu) of TIGR4 (T4) strain of S. pneumoniae, and the total RNA was collected at 18 hours post infection (hpi) from the pools of ten challenged or unchallenged larvae. According to our previous studies in zebrafish larvae [28], a time point of 18 hpi represents a stage of pneumococcal infection at which the bacteria are replicating rapidly, the host innate immune responses have been activated and the signs of infection (e.g. lack of movement) appear, and was therefore chosen for the analysis. The transcriptome analysis was carried out from three biological replicates by RNA sequencing. The RNA sequencing yielded 21–28 million reads per library and approximately 90% were successfully mapped to the transcript database based on the GRCz10 reference genome. To analyze the differentially expressed genes between S. pneumoniae challenged and unchallenged larvae, a fold change of ≥3 was considered significant induction or reduction. In addition, only genes with a mean normalized read count of ≥20 (calculated from the normalized read counts of three biological replicates) after infection were included in the analysis. Finally, two genes with significant variation in their read counts (normalized read counts of 0, 0, and 67 for gene si:ch73-44m9.5 and 0, 0, and 167 for gene BX640576.2) were excluded from the analysis.

With the cutoffs described above, the transcriptome analysis revealed 132 differentially expressed genes in the larvae infected with S. pneumoniae with 81 of them upregulated and 51 downregulated (Table 1 and S1–S4; the complete expression data are available at Gene Expression Omnibus [46] under accession number GSE112433). The differentially expressed protein coding genes were subjected to gene ontology (GO) analysis using the DAVID 6.7 Functional annotation tool [47,48]. According to the analysis, the most enriched biological processes reaching the p-value of <0.05 were proteolysis, defense response to bacterium, complement activation, chitin metabolic process, and defense response to Gram-negative bacterium. However, the DAVID, as well as other tested annotation tools (GOrilla and the Gene Ontology Consortium enrichment analysis [49,50]), only recognized approximately 30% of the submitted genes, and thus gave a distorted view of the enriched processes. Therefore, we manually divided the upregulated protein coding genes into the biological functions according to the available GO data (of zebrafish genes and their mouse and human orthologs) from the Ensembl genome browser versions 91 and 95 [51], The Zebrafish Information Network (ZFIN) [52], the NCBI Gene database [53], and literature. Of the 61 protein coding genes induced upon pneumococcal infection, 29 genes have previously been associated with an immune response to bacterial infection, emphasizing the activation of the zebrafish defense mechanisms against pneumococcus (Fig 1 and Table 1). The other induced protein coding genes were loosely categorized into metabolic processes (12 genes), other processes (7 genes), and unknown processes (13 genes) (Fig 1 and S1 Table). Among the downregulated protein coding genes with available annotation data, the most enriched functional group was reproduction and development (8 genes), while one gene was associated with metabolic processes and six genes were categorized in the group of other processes (S2 Table). In contrast, for a total of 18 downregulated protein coding genes no functional prediction was found. In addition to the protein coding genes, 20 noncoding RNAs were upregulated and 18 downregulated upon pneumococcal infection (S3 and S4 Tables).

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Fig 1. Functional categories of the pneumococcus-responsive genes in zebrafish larvae.

The genes were divided into functional groups based on the available gene ontology data obtained from the Ensembl genome browser versions 91 and 95, The Zebrafish Information Network (ZFIN) and literature (pie chart on the left). Immune response genes were further divided into subcategories (pie chart on the right). The numbers in the sectors indicate the number of genes in each functional group.

https://doi.org/10.1371/journal.pgen.1010586.g001

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Table 1. Immune system–related genes induced in pneumococcal infection in zebrafish.

https://doi.org/10.1371/journal.pgen.1010586.t001

The transcriptome analysis indicates an innate immune response to pneumococcus in zebrafish, with conserved pathways

Our transcriptome data provide evidence for the activation of an innate immune response to pneumococcal infection at the level of gene expression. To characterize this early response in more detail, the pneumococcus-responsive genes associated with innate immunity were further divided into the functional groups of complement system (10 genes), cell migration and chemotaxis (4 genes), antimicrobial peptides (4 genes), immune signaling and regulation (3 genes), acute phase response (3 genes), and macrophage function (3 genes) (Fig 1 and Table 1). Of note, the genes associated with the complement system were particularly enriched in our transcriptome data, suggesting an important role for complement-mediated immunity in the defense against pneumococcus in zebrafish larvae. Among the enriched complement-related genes were those coding for four variants of C3 (c3a.1 c3a.2, c3a.3, c3a.6), a putative counterpart for human C2 (si:ch1073-280e3.1), C4b (c4b), C1r (c1r), complement factor B (cfb), and two negative regulators, complement factor H like 5 (cfhl5) and C1 inhibitor (serping1) (Fig 2). In addition, two zebrafish homologs for the human gene encoding C-reactive protein (crp2 and crp3, ENSDARG00000056498 and ENSDARG00000042613, respectively) were induced upon infection in zebrafish larvae. CRP in mammals is known to recognize S. pneumoniae and to have multiple functions in antipneumococcal response including the activation of the complement system and phagocytosis [54].

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Fig 2. The induced complement -related genes in pneumococcal infection at 18 hpi.

Normalized read counts in wild type zebrafish larvae injected with KCl (unchallenged AB) or ~500 cfu of S. pneumoniae (challenged AB). The genes which had normalized read counts of ≥20 after infection and were induced by ≥3-fold are shown. Each dot represents the normalized read count of a sample of 10 larvae. The differential expressions did not reach statistical significance.

https://doi.org/10.1371/journal.pgen.1010586.g002

Besides the induction of Crp-encoding genes and complement genes, the early response to pneumococcus in zebrafish was characterized by the expression of other genes previously associated with innate immune response to pathogens (Table 1). These included the genes for antimicrobial peptide hepcidin (hamp) [55] and for one of the Ribonuclease A superfamily members (ribonuclease-like 3, rnasel3) which have been shown to have antibacterial, angiogenic and neurotrophic activity in zebrafish [56,57]. Similarly, the expression of two zebrafish chitinase genes chia.1 and chia.2, shown to have antimicrobial activity but also other immunoregulatory functions in fish and mammals [58–62], were upregulated. The zebrafish itln3, coding for a member of the well-conserved lectin family which serves multiple functions in pathogen recognition and the regulation of inflammation [63–66], was also among the most induced genes at 18 hpi. The category of the immune signaling and regulation, on the other hand, included the lipid transport gene apoa4b.3, the G-protein coupled receptor gene gpr84 [67] and a member of lymphocyte antigen 6 family (ly6m2). In addition, lrrc66, an as yet functionally uncharacterized gene coding for a protein with a leucine-rich repeats (LRR)-containing domain, which in many pattern-recognition receptors mediates the receptor-pathogen interaction [68], was induced.

The upregulation of the genes associated with leukocyte activation, migration and phagocytic functions were also detected from the transcriptome data, supporting their important role in the clearance of pneumococcus (Table 1). These included genes for two chemokines (ccl34a.4 and ccl19b), a predicted chemokine (CABZ01001434.1) and a selectin P ligand (selpgl) which in mammals promotes leukocyte recruitment and which shows conserved selectin binding in zebrafish [69]. Also, a macrophage-specific gene aconitate decarboxylase 1, acod1 (also known as immunoresponsive 1 homolog, irg1), which is known to be highly expressed under bacterial infections and to be important for the production of reactive oxygen species by macrophages [70,71], was among the most induced genes in pneumococcal infection. Moreover, consistent with the previous studies on Staphylococcus epidermis-responsive genes in zebrafish, two duplicates of the gene microfibril associated protein 4 (mfap4.5 and mfap4.2), a well-known macrophage marker shown to promote macrophage differentiation during hematopoiesis in zebrafish [72,73], were also upregulated upon infection. Of note, while not quite reaching the cutoff levels set for the induced genes, the induction of the genes for common pro-inflammatory cytokines Il1b (4.4-fold induction) and Il6 (9.9-fold induction) was also evident, with mean normalized read counts of 14 and 4.6 in challenged larvae, and 3 and 0.3 in unchallenged larvae, respectively.

Identification of novel pneumococcus-responsive genes

Besides the genes related to the immune response, upregulation of a set of protein coding genes associated with metabolic processes (12 genes) and other functions (7 genes), was detected in the transcriptome analysis (S1 Table). While their relation to immune response is yet to be unraveled, these genes represent potential novel factors with a role in the early response to pneumococcus. For example, the functional group of metabolic process contained 6 (predicted) serine proteases (ctrl, ctrb1, ela3l, prss59.1, prss59.2, ctrb.3), a group of proteolytic enzymes participating in various biological processes including the immune response [74,75]. In addition, the analysis also revealed multiple upregulated protein coding genes and non-coding RNAs (S1 and S3 Tables) with unknown function, also serving as candidates for further functional characterization.

The forward genetic screen identifies a fish line with compromised resistance to pneumococcal infection

To identify novel genes affecting the innate immune response to pneumococcal infection, we carried out a forward genetic screen in zebrafish utilizing a transposon-based mutagenesis method created by Clark et al. (2011) [76]. In the screen, we generated 126 insertional mutant zebrafish lines, and screened them in F3-F4 generations for altered susceptibility to pneumococcal infection by survival assays. All the mutant lines chosen for the assays survived normally in the laboratory conditions and did not show any defects in morphology, development, or well-being. Zebrafish larvae showing GFP expression as a sign of the presence of one or more randomly located mutations in their genome, were selected for the assays. In the survival assays, 2 dpf zebrafish larvae from each mutant line were infected with S. pneumoniae and their survival was compared to the AB wild type line and the other mutant lines in the same experiment. A total of 58 survival experiments were conducted (bacterial doses ranging from 30–570 cfu between experiments). In each experiment, a wild type control and at least three mutant lines were included to distinguish hypersusceptible lines. In the initial screening, about 10% of the lines appeared to have altered susceptibility to pneumococcal infection, and one of the lines in particular, named as mutant94, showed a notably decreased survival of 28% after S. pneumoniae challenge, compared to the final survival of 74% (p<0.0001) in challenged wild type (AB) controls (Fig 3A). For a comparison, Fig 3A shows the survival rates also for the other two hypersusceptible mutant lines, mutant450 (38% survival, p<0.0001), and mutant14 (45% survival, p<0.0001), as well as for two mutant lines with comparable survival to AB larvae, mutant445 (83% survival) and mutant12 (79% survival).

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Fig 3. Mutant94 larvae are hypersusceptible to pneumococcal infection.

A) Survival of three mutant zebrafish lines, mutant94, mutant450, and mutant14, is compromised in pneumococcal infection compared to wild type (AB) fish and other mutant lines (mutant445 and mutant12). In the experiment, mutant and AB embryos were infected with 35–440 cfu of S. pneumoniae at 2 dpf. The graphs represent the results from three (mutant94, mutant450, mutant14, AB) or one survival experiment (mutant445 and mutant12) (n = 24–48 in each). B) Bacterial burden in mutant94 larvae is elevated after pneumococcal challenge compared to AB and other hypersusceptible mutant lines. Embryos were infected with 505 cfu of S. pneumoniae at 2 dpf and bacterial counts were determined at 18 hpi. The circles represent bacterial counts in a single larva from one experiment (n = 8). Median bacterial count is depicted with a line. The statistical comparison of difference was calculated with log-rank (Mantel-Cox) test in A, and with Kruskal-Wallis test with Dunn’s multiple comparisons test in B. cfu = colony forming units.

https://doi.org/10.1371/journal.pgen.1010586.g003

To determine whether the hypersusceptible mutants had defects in resistance or in tolerance, we measured the bacterial burden in the infected larvae at 18 hpi. As shown in Fig 3B, the mutant94 larvae had elevated bacterial burden compared to AB larvae (p<0.001) and thus showed decreased ability to control bacterial growth. In contrast, bacterial burden in the other hypersusceptible lines did not differ from the bacterial burden in the wild type larvae. These results indicate that the poor survival of mutant94 larvae after pneumococcal challenge is mainly due to compromised resistance against pneumococcal infection while the mutant450 and mutant14 larvae rather have poorer tolerance.

Poor crp2 expression is associated with severe pneumococcal infection in mutant94 larvae

To obtain insight into the underlying defect in the innate defense mechanisms to pneumococcus in the hypersusceptible mutant line, we determined the differentially expressed genes in mutant94 and AB larvae at 18 hours post S. pneumoniae challenge by RNA sequencing. Of the 61 above-mentioned protein-coding genes induced in pneumococcal infection, 9 genes were differentially expressed in mutant94 larvae compared to the AB larvae (S5 Table). All the genes were downregulated by at least 3-fold and represented the functional categories of immune response (4 genes), metabolic processes (2 genes), and unknown processes (3 genes). The biggest difference was observed for the expression of crp2 (ENSDARG00000056498), with an 83-fold reduction compared to the AB larvae. The second and third greatest reductions in response among protein-coding genes were seen with BX548011.1 (43-fold decrease in expression), and si:dkey-9c18.3 (29-fold decrease in expression), for which putative human homologs or functional predictions were not found. In addition, there were more subtle changes (from 3.5- to 7.1-fold reduction) in the expression levels of some other genes involved in immune response (chia.1, ly6m2, chia.2) and for genes with other functions (cyp7a1, zgc:173443, ctrb.3). Of the observed differences, the poor induction of crp2 expression in mutant94 appeared to be a plausible explanation for compromised resistance and thus was selected for further analysis. As there are two genes named as crp2 in CRCz11 reference genome, we will hereafter refer to the one with reduced expression in mutant94 larvae (with the gene ID of ENSDARG00000056498) as crp2-1, and the other (with the gene ID of ENSDARG00000056462) as crp2-2.

In human hepatocytes, the expression of CRP is induced as a part of an acute response to inflammation by the pro-inflammatory cytokines IL1B, IL6 and TNF [77–80]. To examine whether the reduction of crp2-1 expression in mutant94 larvae was due to a defect in the expression of these mediators, we looked at their expression levels in the mutant94 larvae. As seen in the Fig 4, according to the RNA sequencing data, il1b, il6, and tnfa are upregulated during pneumococcal infection in AB (fold-change of 4.4, 9.9, and 2.7 for il1b, il6, and tnfa, respectively, compared to unchallenged larvae), and even more so in mutant94 larvae (fold-change of 3.3, 2.0, and 3.3 for il1b, il6, and tnfa, respectively, compared to challenged AB larvae). Similar induction in il1b and tnfa expression in pneumococcal infection in AB and mutant94 larvae, as well as reduction in crp2-1 expression in mutan94 larvae was also seen with quantitative PCR (qPCR), validating the RNA sequencing results (S1 Fig). From these results, we concluded that the lack of crp2-1 expression in the mutant larvae does not result from the general defect in the inflammatory response, and the increase in the expression of il1b, il6, and tnfa in the mutant larvae may be due to a compensatory response to the hampered crp2-1 expression, or due to more severe infection and higher bacterial burden in mutant94 larvae compared to AB larvae.

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Fig 4. Mutant94 larvae show hampered induction of crp2-1 upon pneumococcal infection compared to wild type larvae.

The figure shows normalized read counts for crp2-1, il1b, il6, and tnfa in wild type larvae injected with KCl (unchallenged AB), wild type larvae infected with ~500 cfu of S. pneumoniae (challenged AB), and mutant94 larvae infected with ~500 cfu of S. pneumoniae (challenged mutant94) at 18 hpi. Each dot represents the normalized read count in a sample of 10 larvae. The differential expressions did not reach statistical significance.

https://doi.org/10.1371/journal.pgen.1010586.g004

Knockout of crp2-1 causes a moderate phenotype during pneumococcal infection, and the effect is augmented by concomitant knockdown of crp3

According to the current reference genome GRCz11, the zebrafish has six CRP homologs, crp1 (ENSDARG00000071454), crp2-1 (crp2, ENSDARG00000056498), crp2-2 (crp2, ENSDARG00000056462) crp3 (ENSDARG00000042613), crp6 (ENSDARG00000071457), and crp7 (ENSDARG00000071456) [51]. Of these, the two crp2 genes and crp3 are highly similar in sequence (88–89% similarity between crp2-1, crp2-2 and crp3), while crp1, crp6, and crp7 are somewhat more distinct (81–83% similarity with crp2-1, crp2-2 and crp3) (analyzed with the BLAST tool of Ensembl Genome Browser [51]). According to the mRNA sequencing data, crp2-1 and crp3 have similar expression patterns in our infection model, with induction in challenged AB larvae (fold increase of 14.8 and 3.4 for crp2-1 and crp3, respectively, compared to unchallenged AB) but reduced expression in challenged mutant94 larvae (fold reduction of 83.2 and 2.5 for crp2-1 and crp3, respectively, compared to challenged AB) (S2 Fig). crp2-2, on the other hand, exhibits opposite expression pattern with no expression in challenged AB larvae but high expression in challenged mutant94 (fold increase of 7.2 compared to challenged AB). According to the data, crp1 is expressed independent of the treatment, while crp6 and crp7 do not seem to be expressed in zebrafish larvae at 18 hpi. Therefore, it seems that crp2-1, crp2-2 and crp3 may be relevant in pneumococcal infection in zebrafish larvae, and due to the sequence homology and clearly related expression patterns of these three Crp-encoding genes, it is possible, that their products also have overlapping functions.

To gain more insight into the importance of zebrafish Crps in the defense against pneumococcal infection in zebrafish, we produced a crp2-1 knockout zebrafish line with CRISPR-Cas9 mutagenesis (Fig 5). During the past few years, the engineered type II CRISPR/Cas system has become a popular tool for reverse genetics and has been widely used in various animals, including zebrafish, to model human genetic disorders [81,82]. To ensure efficient mutagenesis of crp2-1 in zebrafish, the sequence of the potential target site for the mutagenesis was verified by sequencing and found to be slightly different from the reference sequence in GRCz11 (in-house sequenced wild type sequence presented in Fig 5A). By injecting Cas9 protein and a single guide RNA targeting the second exon of crp2-1, we induced insertion/deletion mutations in the founder (F0) generation (Fig 5B). By sequencing the mutated loci in outcrossed F1 larvae (F0 x AB), we identified several germline-transmitted mutations, one of which, an insertion of 23 nucleotides (loss of A and gain of TGGTCTCCACACAGATGGTCGTGT) at the target locus was chosen for the production of F2 and further generations. This mutation leads to a disrupted reading frame and a premature stop codon after 97 amino acids (aa) in the Crp2-1 protein normally with total length of 225 aa (Fig 5C). Of note, homozygous mutants appeared phenotypically normal, with no defects in morphology or well-being in standard laboratory conditions (Fig 5D). This knockout mutant line is hereafter called crp2^tpu6.

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Fig 5. crp2-1 knockout zebrafish produce truncated form of Crp2-1 and show compensatory expression of crp3.

A) A schematic presentation of the zebrafish crp2-1 (crp2, ENSDARG00000056498) gene structure and the gRNA target site for CRISPR-Cas9 mutagenesis. The target site sequence with depicted gRNA target sequence and protospacer adjacent motif (PAM) site were verified with Sanger sequencing. B) Successful mutagenesis in the embryos injected with target specific gRNA and Cas9 protein was verified with heteroduplex mobility assay. In the heteroduplex mobility assay, multiple bands on polyacrylamide gel after heteroduplex formation indicate the presence of insertion/deletion mutations at the target site. C) Schematic representation of the structure of wild type (WT) Crp2-1 protein and the structure of Crp2^tpu6 mutant protein after introduction of +23 nucleotide insertion and premature stop codon by CRIPSR-Cas9 mutagenesis. D) Homozygous crp2^tpu6/tpu6 larvae have normal morphology under standard laboratory conditions. The panel shows an overview and a representative image of a single larvae of 2 dpf wild type (WT (crp2^tpu6)) and homozygous (crp2^tpu6/tpu6) larvae of F3 generation. Images were taken with Lumar V.12 fluorescence stereomicroscope with an exposure time of 100 ms. E) crp2-1 expression is induced in WT (crp2^tpu6) larvae at 24 hpi. F) crp3 expression is induced in homozygous crp2^tpu6/tpu6 larvae at 24 hpi. In E and F, the expression relative to eef1a1l1 expression was measured with the 2^-ΔCt method from the pools of five WT or mutant larvae infected with 241 cfu of S. pneumoniae. The line depicts the median expression level, and the statistical analyses were conducted with Kruskal-Wallis test with Dunn’s multiple comparisons test. Comparisons were only done to 0 h time point of the same gene.

https://doi.org/10.1371/journal.pgen.1010586.g005

In addition to the changes in protein structure and function, an indel mutation introduced by CRISPR-Cas9 may also affect the transcript levels of the target gene [83,84]. Multiple recent studies have also demonstrated changes in the expression levels of other, closely related genes in knockout models due to transcriptional adaptation (reviewed in El-Brolosy & Stainier, 2017) [85]. To see whether the mutation affects the expression of crp2-1 and the related crp2-2 and crp3, we measured the relative expression levels of these genes in homozygous and WT larvae. For the analysis, crp2^tpu6/tpu6 and WT (crp2^tpu6) larvae of F3 generation were infected with 241 cfu of S. pneumoniae at 2 dpf and the infected larvae were collected at 0h, 6h, 24h, 46h, and 68h for the analysis of gene expression by qPCR. Due to the high level of homology between the three Crp-encoding genes, the specificity of the amplification in selected crp2^tpu6/tpu6 and WT (crp2^tpu6) samples was evaluated by Sanger sequencing. Consistent with the analysis of crp2-1 expression in challenged AB larvae, crp2-1 expression was induced in challenged WT (crp2^tpu6) larvae after 24 hpi by 3.9-fold (p<0.01) (Fig 5E). crp2-2 and crp3, on the other hand, only showed minor expression in the same larvae and no induction upon pneumococcal challenge (Fig 5E). We were unable to verify the presence or absence of crp2-1 transcript in homozygote mutants by qPCR due to the amplification of transcripts aligning equally well with crp2-1, crp2-2 and crp3 (S3 Fig). Sequencing of the amplified cDNA of crp2^tpu6/tpu6 revealed overlapping spectra, but through manual interpretation, we identified one of the sequences as crp2-1 with the +23 nt mutation. Of note, no such amplification of multiple targets was detected in WT (crp2^tpu6) larvae. The expression levels of crp2-2 and crp3 in the challenged crp2^tpu6/tpu6 larvae showed only minor expression and no induction of crp2-2 upon pneumococcal infection, but clear, 5-fold induction of crp3 after 24 hpi (p<0.01) (Fig 5F). Overall, while the unspecific amplification in homozygote samples but not in WT samples might suggest a decrease in the level of crp2-1 transcript in the mutants, this could not be verified with the qPCR analysis. However, these results show that whether present at equal or lower levels than in WTs, the crp2-1 transcripts in crp2^tpu6/tpu6 mutants carry the specific frame-shifting insertion and the premature stop codon leading to a truncated Crp2-1 protein in the mutant larvae. In addition, the increased crp3 transcript levels in crp2^tpu6/tpu6 larvae suggest activation of compensating mechanisms for the loss of functional crp2-1 transcript or protein.

Next, to see whether the phenotype observed in mutant94 larvae could be explained by Crp2-1 deficiency, we infected homozygous, heterozygous and WT crp2^tpu6 embryos at 2 dpf with 1,646 cfu of S. pneumoniae and monitored their survival for three days. The experimental infections and the survival assay were carried out in a blinded and randomized setting as described in Materials and Methods. As shown in Fig 6, both the homozygous and heterozygous mutants showed decreased survival compared to their WT siblings with survival percents of 76%, 77% and 84% for crp2^tpu6/tpu6, crp2^tpu6/+, and WT (crp2^tpu6), respectively (p<0.01). While the modestly compromised survival in the mutants indicates that Crp2-1 does participate in the defense response to pneumococcus in this model, the milder phenotype of the knockout mutant compared to the mutant94 indicates that other factors also contribute to the hypersusceptibility of mutant94 observed in the forward genetic screen. Alternatively, as was suggested by the increased expression of crp3 in crp2^tpu6/tpu6 larvae, it is plausible that the observed transcriptional adaptation to the loss of function of Crp2-1 results in a milder phenotype in pneumococcal infection model.

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Fig 6. crp2^tpu6 mutants show modest impairment in survival in systemic pneumococcal infection.

The figure shows the survival of wild type (WT (crp2^tpu6)), heterozygous (crp2^tpu6/+) and homozygous (crp2^tpu6/tpu6) larvae after pneumococcal infection. In the experiment, 1,646 cfu of S. pneumoniae was injected into the bloodstream of 2 dpf larvae and the survival was monitored until 5 dpf. Data were collected from a single experiment with the group sizes (n) indicated in the figure. The statistical comparison of difference was calculated with log-rank (Mantel-Cox) test.

https://doi.org/10.1371/journal.pgen.1010586.g006

To test whether crp3 compensates Crp2-1 deficiency, we used morpholino-oligonucleotides to knockdown crp3 expression in crp2^tpu6/tpu6 larvae. A splice-blocking (SB) morpholino targeting the crp3 5’ upstream sequence and tentatively causing the skipping of exon 1 where the start codon lies, was chosen for the study. Using a dose titration test, dose of 4 ng per embryo was chosen as this was the maximum dose that showed no developmental effects within the first 5 dpf in unchallenged AB larvae (Fig 7A). To address the knockdown efficiency of crp3 SB morpholino, the amount of WT crp3 transcript in morpholino injected larvae was determined by qPCR utilizing primers which tentatively bind only to WT transcript. As a sign of knockdown, the crp3 SB morphants showed lower levels of WT crp3 transcript at 3–4 dpi, the most important time considering S. pneumoniae infection kinetics in zebrafish larvae, with residual expression of 24% at 3 dpi (p<0.01) and 48% at 4 dpi (p<0.05) compared to larvae injected with the same dose of random control (RC) morpholino (Fig 7B). To test the synergistic effect of crp2-1 knockout and crp3 knockdown, crp3 SB morpholino and RC morpholino injected crp2^tpu6/tpu6 larvae were infected with 296 cfu of S. pneumoniae at 2 dpf and their survival was followed for three days. During the infection experiment, no signs of off-target effects affecting development could be detected in either of the groups. The survival of crp3 SB injected larvae from pneumococcal infection, on the other hand, was significantly lower (53% survival) compared to the RC injected larvae (75% survival) (p<0.05). We therefore conclude, that while the knockout of crp2-1 alone seems to have only a moderate effect on zebrafish survival from systemic pneumococcal infection, together with crp3 knockdown it results in significantly compromised survival of the larvae. However, this phenotype still does not fully recapitulate the phenotype of mutant94 larvae, indicating that other genes, such as BX548011.1 which was the second most downregulated gene in mutant94 larvae, or crp2-2, may affect the resistance against S. pneumoniae in zebrafish larvae. This however, remained to be experimentally investigated as we failed to knockdown the expression of BX548011.1 and crp2-2 in zebrafish larvae by specific SB morpholino oligonucleotides.

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Fig 7. Partial knockdown of crp3 further impairs the survival of crp2^tpu6/tpu6 larvae from systemic pneumococcal infection.

A) Percentage of surviving and normally developing uninjected zebrafish larvae, larvae injected with 4 ng of RC morpholino, and larvae injected with different doses (1 ng, 2 ng or 4 ng) of crp3 SB morpholino. In the experiment, RC or crp3 SB morpholino was microinjected into the yolk sac of AB zebrafish embryos at the 1–4 cell stage and their survival and development was followed for 5 days. The percentage represents the live (black) and morphologically normal (gray) larvae in each group (n = 31–50) at 5 dpi. B) crp3 transcript levels in RC morpholino and crp3 SB morpholino injected AB zebrafish larvae at 0–5 dpi. In the experiment, 4 ng of RC or crp3 SB morpholino was microinjected into the yolk sac of AB zebrafish embryos at 1–4 cell stage and pools of 3–5 larvae (n = 7–8) were collected at specific time points. crp3 expression levels were measured with qPCR from technical duplicates and normalized with eef1a1l1 expression. A dot/circle represents the relative gene expression in a single pool of embryos/larvae, which was calculated using the 2^-ΔCt method, and the line depicts the median expression level. C) The survival of RC or crp3 SB morpholino injected homozygous (crp2^tpu6/tpu6) larvae after pneumococcal infection. In the experiment, 4 ng of RC or crp3 SB morpholino was microinjected into the yolk sac of crp2^tpu6/tpu6 embryos of F3 generation at 1–4 cell stage, and at 2 dpi the larvae were infected with 296 cfu of S. pneumoniae. The survival of infected larvae was monitored until 5 dpf. Data were collected from a single experiment, with the group sizes (n) indicated in the figure. The statistical comparisons of difference were calculated with Kruskal-Wallis test with Dunn’s multiple comparison in B, and with log-rank (Mantel-Cox) test in C. dpi = days post injection.

https://doi.org/10.1371/journal.pgen.1010586.g007

Zebrafish and ethics statement

The wild type zebrafish (Danio rerio) lines were obtained from and maintained by the Tampere Zebrafish Core Facility (Tampere university, Finland). The AB wild type zebrafish line was used as a reference line in the infection experiments and in the production of the mutant lines. The TL (Tupfel long fin, leo^{t1 -/-}, lof^{dt2 -/-}) line was used in the outcrosses in the production of the mutant lines for the forward genetic screen. Zebrafish were maintained according to the standard protocols [159] and the ethical guidelines set by the Ethical Board in Finland. Briefly, zebrafish embryos and larvae were kept in E3-medium (5 mM NaCl, 0.17 mM KCl, 0.33 mM CaCl₂, 0.33 mM MgSO₄, 0.0003 g/l Methylene Blue) in an incubator at 28°C, with a light/dark cycle of 14/10 hours until 6 dpf after which they were moved to a filtered flow-through system with the same light/dark cycle and fed once a day with GEMMA Micro 75–500 (Skretting, Stavanger, Norway) or twice a day with SDS 100–400 (Special Diets Services, Essex, UK). The well-being of the fish was monitored daily, and moribund fish were immediately euthanized with an overdose (0.08%) of an anesthetic Tricaine (3-aminobenzoic acid ethyl ester, pH 7.0) (Sigma-Aldrich, St. Louis, Missouri, USA). During the experiments, embryos and larvae were maintained in corresponding conditions. Prior to the treatments, zebrafish larvae were anesthetized with 0.02% Tricaine. The maintenance and the experiments were done in accordance with the Finnish Act on the Protection of Animals Used for Scientific or Educational Purposes (497/2013) and the EU Directive 2010/63/EU. The ethical permissions for the generation of the mutant zebrafish lines as well as the maintenance of the fish were granted by the Animal Experiment Board in Finland (the licenses ESAVI/10079/04.10.06/2015, ESAVI/6407/04.10.2012, ESAVI/10366/04.10.07/2016, ESAVI/2235/04.10.07/2015, ESAVI/11133/04.10.07/2017, ESAVI/2464/04.10.07/2017, ESAVI/2776 /2019 and ESAVI/7251/2021). All the zebrafish experiments were conducted prior to the independently feeding larval stage and therefore, according to the EU directive, no permission for experimentation was required.

The gene-breaking tol2 transposon-based mutagenesis and the generation of mutant fish lines for the forward genetic screen

Generation of the zebrafish mutant lines for the forward genetic screen was carried out with the method published by Balciunas et al (2006). and Clark et al. (2011) [76,160], and as described in detail in Harjula et al. (2020) [45]. Briefly, the gene-breaking transposon pGBT-RP2.1 (RP2) vector (Addgene plasmid # 31828) and the pT3TS-Tol2 vector coding for the tol2 transposase (Addgene plasmid # 31831) [76] were transformed into E. coli One Shot TOP10 cells (Invitrogen, Thermo Fisher Scientific, Waltham, Massachusetts, USA) and purified from overnight cultures with the QIAGEN Plasmid Plus Maxi Kit (Qiagen, Hilden, Germany). To produce tol2 mRNA, the linearized pT3TS-Tol2 plasmid was used as a template in in vitro transcription conducted with the mMESSAGE mMACHINE T3 Kit, (Invitrogen, Thermo Fisher Scientific) according to the manufacturer’s instructions. To carry out the mutagenesis, 12.5 pg RP2 and 12.5 pg tol2 mRNA in phosphate buffered saline (PBS) and with 0.6% phenol red (Sigma-Aldrich) were microinjected into the 1-cell stage AB zebrafish embryos and the mutation carrying embryos (designated as F0) were selected based on the expression of green fluorescent protein (GFP) under the Lumar V.12 fluorescence stereomicroscope (Carl Zeiss MicroImaging GmbH, Göttingen, Germany). To produce the mutant zebrafish lines, F0 founder fish were outcrossed with TL, resulting in the F1 generation of fish potentially carrying multiple insertional mutations. These fish were named after a running number and, again, were crossed to TL to obtain F2 generation fish. The F3 and the following generations (until F5) were generated by incrosses. In some cases, in order to maintain the line, the fish were again crossed to TL. In each generation, mutation-carrying progeny were selected based on the expression of GFP.

CRISPR-Cas9 mutagenesis

CRISPR-Cas9 mutagenesis in zebrafish was carried out based on the methodology by Hruscha and Schmid (2015) [161]. An optimal target sequence for a guide RNA (gRNA) targeting crp2-1 was selected with an online CRISPR design tool CHOPCHOP v2 [162], after which the sequence was manually adjusted to match with the target sequence in our AB variant. NCBI BLAST (https://blast.ncbi.nlm.nih.gov/Blast.cgi) analysis against the zebrafish reference genome GRCz11 was used to verify the specificity of the chosen target sequence 5’- AATGTGTGGAGAGAAAAAGATGG-3’. The production of gRNA and the mutagenesis were carried out as previously described in Uusi-Mäkelä et al. (2018) [163] with a few adjustments. Briefly, a gRNA targeting the second exons of crp2-1 was produced from a DNA oligo template (Sigma-Aldrich) with the MEGAshortscript T7 Transcription Kit (Invitrogen, Thermo Fisher Scientific) according to the manufacturer’s instructions. For the mutagenesis, a mixture of 230 pg of the gRNA, 330 pg of in-house produced Cas9 protein (Protein Service core facility, Tampere University, Finland) and 1% Rhodamine dextran tracer (Invitrogen, Thermo Fisher Scientific) was injected into the cytoplasm of 1-cell stage AB zebrafish embryos and the mutagenesis efficiency was evaluated by heteroduplex mobility assay (HMA) [163] from isolated DNA of 2 dpf embryos. Germline-transmitted mutations at the target site were screened in the progeny of outcrossed F0 zebrafish using HMA and Sanger sequencing, and F1 zebrafish with a desired nonsense mutation were incrossed to generate a stable knockout zebrafish line.

Genotyping of CRISPR mutants

At all stages of the mutant line production, genotyping of the adult fish was carried out from the tailfin DNA by Sanger sequencing. Specifically, fish were anesthetized with 0.02% Tricaine, and a small piece of tail fin was cut off with a scalpel blade. Tail fin tissue was lyzed for 4–24 h at 55°C in lysis buffer (10mM Tris pH 8,2, 10mM EDTA, 200mM NaCl, 0.5% SDS, 200μg/ml Proteinase K). DNA was precipitated for 1h-24h at -20°C using two volumes of ethanol after which DNA was pelleted by centrifuging at 16,000g for 10min. The pellet was washed with 70% ethanol and resuspended in water. The genomic site containing the mutagenesis target site was amplified using DreamTaq Hot Start DNA Polymerase (Thermo Scientific) and sequenced using the BigDye Terminator v3.1 Sanger sequencing Kit (Applied Biosystems, Life technologies, Carslbad, USA). Primers used in the genotyping are listed in S7 Table. To distinguish homozygous and wild type fish, the obtained sequence was compared to the wild type sequence using Clustal Omega [149,150]. To analyze the mutation site in the heterozygote fish, CRISP-ID v1.1 web application [164] for detecting CRISPR induced indels was used. Alternatively, the two overlapping sequences (one from the WT allele and the other from the mutant allele) were manually separated from the sequencing chromatogram.

Generation of zebrafish morphants and quantification of knockdown efficiency

A splice-blocking morpholino against the zebrafish crp3 was designed by and ordered from GeneTools LLC (Philomath, OR, USA) and its specificity for crp3 was checked with NCBI BLAST tool (https://blast.ncbi.nlm.nih.gov/Blast.cgi). The crp3 SB morpholino oligo sequence was 5’-TGCTCCAAGTCTGAAAAGGAAGAAA-3’. To determine the maximum dose of the morpholino with no off-target effects, three doses (4 ng, 2 ng and 1 ng) of the crp3 SB morpholino were microinjected into the yolk sac of 1–4 cell stage wild type AB zebrafish embryos. Prior to microinjection, morpholinos were heated at 65°C for 10 min to ensure complete suspension. In microinjection, injection volume of 1 nl per embryo was used and the injection solution was dyed with 1% Rhodamine dextran tracer (Invitrogen, Thermo Fisher Scientific) to verify the success of the injection. After injection, the development and well-being of injected larvae were monitored daily for five days, with abnormal and dead embryos/larvae recorded.

To quantify the effect of crp3 SB morpholino on the levels of WT crp3 transcript, AB zebrafish embryos were microinjected as above with 4 ng of crp3 SB morpholino or random control (RC) morpholino (GeneTools LLC) and pools of 3–5 larvae were collected daily from 0 to 5 dpi for the extraction of total RNA.

Experimental S. pneumoniae infection and survival assay

The Streptococcus pneumoniae wild type strain TIGR4 (T4) of serotype 4 was used throughout the study [165] and the cultivation and the harvesting of the bacteria was carried out as previously reported [28]. Briefly, bacteria were grown overnight on 5% lamb blood agar plates (Tammer BioLab Ltd, Tampere, Finland) at 37°C and 5% CO₂ and after that, in liquid culture in Todd Hewitt broth (Becton, Dickinson and Company, New Jersey, USA) supplemented with 0.5% yeast extract (Becton, Dickinson and Company) until mid-logarithmic growth phase determined by optical density at a wavelength of 620 nm. Bacterial cells were harvested by centrifuging for 10 min at 4000 rpm (rounds per minute). Using the theoretical bacterial density in the liquid culture (10⁸ cells/ml), the bacterial pellet was suspended in desired volumes of 0.2 M KCl (infection experiments related to the genetic screen and mutant94) or PBS (infection experiments related to crp2^tpu6) to obtain the desired bacterial dose. The actual number of bacteria in the injection solution was verified by plating samples of single injections on 5% lamb blood agar plates (Tammer BioLab Ltd). Due to variation in bacterial growth rate, accuracy of absorbance measurement, slightly varying volumes of bacterial solution released by the injector, accuracy of quantitative plating etc., relatively big variation in bacterial doses between experiments (30–570 cfu in infection experiments related to the genetic screen and mutant94 and 296–1,646 cfu in infection experiments related to crp2^tpu6) was seen. To avoid the influence of varying doses, all groups within a single experiment were infected with the same bacterial solution and only groups from the same assay were compared to find differences in survival or bacterial count.

Experimental S. pneumoniae infections and the subsequent survival assays were carried out by microinjecting a specific dose of S. pneumoniae into the blood circulation valley of 2 dpf zebrafish larvae and following the survival for 72–96 hpi as described in detail in Rounioja et al. (2012) [28]. The unchallenged control groups were similarly treated with either 0.2 M KCl (infection experiments related to the genetic screen and mutant94) or PBS (infection experiments related to crp2^tpu6). The survival assay for the crp2^tpu6 mutant line was carried out in F2 larvae (the progeny of incrossed F1 heterozygotes) showing Mendelian ratios of homozygous, heterozygous, and wild type larvae. To achieve a blinded and randomized experimental set up, the larvae were infected prior to the genotyping. In the experiment, the survival of the infected larvae was followed for 72 hours, during which the larvae were checked twice a day. Terminally diseased and dead larvae were collected, euthanized if necessary, and frozen for later genotyping. For the genotyping, the genomic DNA extraction from whole larvae, target site amplification, and sequencing, were done as described for adult fish genotyping in the section Genotyping of CRISPR mutants.

Determining the bacterial load from zebrafish larvae

The bacterial loads were measured from infected zebrafish larvae at desired time points by quantitative plating of homogenates. More specifically, infected larvae were euthanized with 0.08% Tricaine (Sigma-Aldrich) and washed three times with PBS. By using a pestle, single larvae were manually homogenized in 200 μl of PBS with 1% Triton-X (Sigma-Aldrich). To obtain the bacterial counts, the homogenates were serially diluted and grown on 5% lamb blood agar plates (containing 1 μg/ml erythromycin) (Tammer BioLab Ltd) overnight at 37°C and 5% CO₂.

RNA extraction and the quality control

For the RNA sequencing, total RNA was extracted from the pools of five larvae using the Qiagen RNeasy Mini Kit (Qiagen) according to the manufacturer’s instructions. The genomic DNA was removed from the samples using the RapidOut DNA removal Kit (Thermo Scientific, Thermo Fisher Scientific) and the purity and the concentration of the samples were measured with Qubit RNA BR Assay Kit (Invitrogen, Thermo Fisher Scientific). The RNA integrity was analyzed with the Fragment Analyzer (Advanced Analytical Technologies, Iowa, USA) using the Standard Sensitivity RNA Analysis Kit (Advanced Analytical Technologies) and the PROSize® 2.0 Data Analysis Software (Advanced Analytical Technologies). The RNA samples with an integrity number (RIN) >9 were used in the transcriptome analysis. Finally, two RNA samples were combined to achieve sufficient concentration for the transcriptome analysis.

For the qPCR analysis of crp2-1, crp2-2 and crp3 expression in WT and mutant crp2^tpu6 larvae and of crp3 expression in crp3 SB morpholino and RC morpholino injected larvae, total RNA was extracted from the pools of 3–5 larvae using the Qiagen RNeasy Plus Mini Kit with genomic DNA elimination step (Qiagen) according to the manufacturer’s instructions. The purity and the concentration of the samples were measured with NanoDrop 2000 Spectrophotometer (Thermo Scientific, Thermo Fisher Scientific).

Quantitation of the relative gene expression levels by qPCR

The expression levels of il1b, il6, tnfa and crp2-1 in the RNA sequencing samples were measured with qPCR using the iScript One-Step RT-PCR Kit with SYBR® Green (Bio-Rad Laboratories, Hercules, California, USA) according to the manufacturer’s instructions. Thermal cycling was performed using the Bio-Rad CFX 96 instrument (Bio-Rad Laboratories) and the CFX Software version 3.1 (Bio-Rad Laboratories) was used to analyze the results. The target gene expression levels were normalized to the eef1a1l1 expression [166], and the double delta Ct method (2^-ΔΔCt) was used to calculate the fold change in expression in challenged larvae compared to unchallenged larvae.

For the measurement of the expression levels of crp2-1, crp2-2 and crp3 in mutant and wild type crp2^tpu6 larvae, and the expression levels of crp3 in crp3 SB morpholino and RC morpholino injected larvae, the reverse transcription of the RNA samples was done with the SensiFast cDNA synthesis kit (BioLine, London, UK) and the qPCR with the PowerUp SYBR® master mix (Applied Biosystems, Thermo Fisher Scientific). Thermal cycling was performed using the Bio-Rad CFX Opus 96 instrument (Bio-Rad Laboratories) and the CFX Maestro Software version 2.2 (Bio-Rad Laboratories) was used to analyze the results. The specificity of the amplified qPCR products was evaluated by melt curve analysis and Sanger sequencing of selected samples. Delta Ct method (2^−ΔCt) was used for calculating the expression levels of target genes relative to the expression of eef1a1l1. The primers used for the qPCR are listed in S7 Table.

RNA sequencing and data analysis

The cDNA library preparation and the RNA sequencing were conducted at Novogene, Hong Kong. The 250–300 bp cDNA library was sequenced with the 150 bp paired-end sequencing on the Illumina platform and with a sequencing depth of >20 million reads/sample.

The quality of the reads was inspected using the FastQC and the reads were aligned against the GRCz10 reference genome with STAR using the default parameters (http://www.bioinformatics.babraham.ac.uk/projects/fastqc) [167]. To improve the alignment accuracy, Ensembl GRCz10.91 was used as the reference set for known transcripts. FeatureCounts was used to quantify the expressions of the genes belonging to the same reference gene set as that used during the alignment step [168]. Normalization of raw expressions and differential gene expression analysis were both conducted using R-package DESeq2 [169].

Statistical analyses

The ClinCalc web tool (https://clincalc.com/Stats/SampleSize.aspx) was used to carry out the calculations for sample sizes in survival experiments. Based on our previous experiments in S. pneumoniae challenged zebrafish larvae, the difference in end-point mortality between wild type and mutant lines was estimated to be 40%. With an 80% statistical power, a minimum group size of 16 was determined for survival experiments.

All the statistical analyses were carried out with the GraphPad Prism software version 5.02 or 9.0.0 (GraphPad Software, Inc, California, USA). The log-rank (Mantel-Cox) test was used to analyze the differences in the survival rates between the zebrafish lines. The differences in the gene expression levels measured by qPCR and the bacterial amounts were analyzed with Kruskal-Wallis with Dunn’s multiple comparisons test in Figs 3B, 5E–5F and 7B and with two-tailed Mann-Whitney test in S1 Fig. In each analysis, a p-value of <0.05 was considered as statistically significant and only p-values <0.05 are shown in the figures.