Fig 1.
Deep sequencing reveals high mutagenesis rates for acute multi-sgRNA CRISPR/Cas9 KO of osgep and tprkb.
(a) Mutation rates per sgRNA (i.e. likelihood of generating at least one mutation on each allele), given hypothetical mutation rates for 3 different sgRNAs. Note that the achieved likelihood for occurrence of at least one mutation per allele is high (92.16%), even though mutation rates for each sgRNA are moderate (50–80%). (b-e) Deep sequencing reveals high mutagenesis rates for most sgRNAs and sgRNA dependent frameshift rates. For each gene, deep sequencing data of 96 larvae at 48 hpf were analyzed individually using the tool CRISPResso [24]. (b) For osgep, median mutagenesis rate was 69.3% for sgRNA1, 93.0% for sgRNA2, and 79.6% for sgRNA3. (c) For all mutated osgep alleles, the median fractions of frame shifts were 56.7% for sgRNA1, 44.7% for sgRNA2, and 61.0% for sgRNA3. (d) For tprkb, median mutagenesis rate was 100% for sgRNA1, 92.3% for sgRNA2, and 41.0% for sgRNA5. (e) For all mutated tprkb alleles, the median fractions of frame shifts were 44.3% for sgRNA1, 81% for sgRNA2, and 52.8% for sgRNA5.
Fig 2.
Likelihoods for at least one mutation on each allele and at least one frameshift mutation on each allele are clearly improved in acute multi-sgRNA CRISPR/Cas9 KO of osgep and tprkb.
(a-b) The likelihood is shown of generating at least one mutation on each allele P(M) (a) and of generating at least one frameshift mutation on each allele P(F) (b), where M = at least one mutation on each allele, F = at least one frameshift mutation on each allele, q = probability of no mutation, r = specific sgRNA, im = fraction of in-frame mutations of all mutations, nm = fraction of non-coding mutations of all mutations. (c-h) The likelihoods of at least 1 mutation on each allele and at least 1 frameshift mutation on each allele were calculated based on observed mutagenesis and frameshift rates for 96 individual fish per gene according to the equation in (a) and (b). (c) The individual analysis for osgep revealed a median P(M) of 48.1% for sgRNA1, 86.6% for sgRNA2, and 63.4% for sgRNA3. (d) The median P(F) was 15.1% for sgRNA1, 14.6% for sgRNA2, and 18.5% for sgRNA3. (e) For the pooled osgep sgRNAs, the median P(M) was 99.3%, and 64.2% for P(F). (f) The individual analysis for tprkb revealed a median P(M) of 100% for sgRNA1, 84.7% for sgRNA2, and 16.9% for sgRNA5. (g) The median P(F) was 15.5% for sgRNA1, 52.2% for sgRNA2, and 4.1% for sgRNA5. (h) For the pooled tprkb sgRNAs, the median P(M) was 100%, and 78.8% for P(F).
Fig 3.
Deep sequencing reveals sgRNA dependent indel profiles.
For each gene, deep sequencing data of 96 larvae at 48 hpf were analyzed individually using the tool CRISPResso [24]. (a-f) On the X-axis, the graphs show the numbers of base pairs (bp) deleted (negative numbers) or inserted (positive numbers). The Y-axis displays the fraction of the particular indel of all mutated reads per each one of the 96 individual larvae. As many larvae had more than 1 indel, all fractions give the mean frequency in 96 individual larvae. (a) For osgep sgRNA1, the most frequent deletion was 6 bp in length (23.5%, arrow head), followed by 8 bp (11.1%). The most frequent insertion was 1 bp (6.8%). (b) For osgep sgRNA2, the most frequent deletion was 10 bp (18.2%, arrow head), followed by 2 bp (16.3%). The most frequent insertion was 1 bp (6.3%). (c) For osgep sgRNA3, the most frequent deletion was 10 bp (17.1%, arrow head), followed by 12 bp (10.9%). The most frequent insertion was 1 bp (4.7%). osgep sgRNA2 and sgRNA3 additionally create larger deletions up to 65 bp together (b-c) since they target the same exon. (d) For tprkb sgRNA1, the most frequent deletion was 3 bp in length (22.9%, arrow head), followed by 6 bp (13.5%). The most frequent insertion had a length of 1 bp (3.9%). (e) For tprkb sgRNA2, the most frequent deletion was 2 bp (23.8%, arrow head), followed by 1 bp (15.3%). The most frequent insertion was 1 bp (3.1%). (f) For tprkb sgRNA5, the most frequent deletion was 3 bp (26.9%, arrow head), followed by 2 bp (10.0%). The most frequent insertion was 7 bp (2.6%). (g) For osgep, deletions accounted for 77.9% of all mutations for sgRNA1, 82.8%for sgRNA2 and 83.6%. for sgRNA3, compared to 22.1% insertion for sgRNA1, 17.2% for sgRNA2 and 16.4% for sgRNA3. (h) For tprkb, deletions accounted for 73.8% of all mutations for sgRNA1, 83.4%for sgRNA2 and 82.5%. for sgRNA5, compared to 26.2% insertion for sgRNA1, 16.6% for sgRNA2 and 17.5% for sgRNA5.
Fig 4.
Acute multi-sgRNA CRISPR/Cas9 KO and stable KO lines for osgep and tprkb demonstrate reduced survival.
In order to generate survival curves, larvae were monitored twice a day for 22 days and transferred to rotifer feeding solution at 8 dpf. Tracing started at 24 hpf. (a) The acute multi-sgRNA KO for osgep shows an increased death rate as of 10 dpf with a survival of 42.9% at 22 dpf compared to scrambled and uninjected control with both survival of 73.4% at 22 dpf (P<0.0001). (b) In the stable osgep KO line with the truncating mutation c.102_105del, homozygous larvae show complete mortality by 5 dpf, whereas 71% of the heterozygous and 80% of the wildtype larvae survive until the end of the observation at 21 dpf. No significant difference in survival was found between heterozygous and wildtype larvae (P = 0.3269). (c) For osgep, homozygous stable KO larvae show a significantly increased and earlier mortality compared to acute KO larvae (P<0.0001). (d) The acute multi-sgRNA KO for tprkb shows an increased death rate as of 12 dpf with a survival of 19.4% at 22 dpf compared to scrambled (58.0%) and uninjected control (63.8%) at 22 dpf (P<0.0001).
Fig 5.
Acute multi-sgRNA CRISPR/Cas9 knockout reproduces the microcephaly phenotype of stable osgep and tprkb KO lines.
Larvae were imaged from a dorsal view. Head diameter to total body length was calculated. This ratio was normalized by the mean of the uninjected/wildtype control and defined as “microcephaly index” (MI). MI was determined at 6 dpf except for the stable KO line osgep c.102_105del at 4 dpf due to early lethality (see Fig 4a). (a) Representative phenotypes of a larva that was injected with the osgep targeting multi-sgRNA pool compared to a scrambled control larva. The red lines display the typical axes used for the measurements. (b) For osgep, the acute multi-sgRNA KO showed a significant microcephaly compared to uninjected and scrambled control. (c) The microcephaly phenotype is recapitulated in the homozygous larvae of the stable KO line osgep c.102_105del compared to wildtype and heterozygous clutch mates. (d) For osgep, MI is significantly higher in acute KO larvae compared to homozygous stable KO larvae. (e) For tprkb, the acute multi-sgRNA KO showed a significant microcephaly compared to uninjected and scrambled control. (f) The microcephaly is recapitulated in the homozygous larvae of the stable KO line tprkb c.370_376delinsAA compared to wildtype and heterozygous clutch mates. (g) For tprkb, no significant difference in MI was found for acute KO larvae compared to homozygous stable KO larvae.