Mutations in dgat2 cause yolk opacity due to aberrant lipid droplet accumulation.A, in situ hybridization for dgat2 expression at 3 and 6 days post fertilization (dpf) in wild-type AB embryos; dgat2 is expressed in the yolk syncytial layer (YSL), liver and intestine; images are representative of all embryos from three experiments at each stage with n = 10 embryos per probe per experiment; Scale bars = 200 μm. B, representative images of a wild-type embryo and a homozygous mutant dgat2^sa13945 embryo with a fully opaque yolk at 3 dpf; Scale = 500 μm. C, Examples of the heterogeneity in yolk opacity in dgat2^sa13945 mutant embryos at 3 dpf; Scale = 500 μm. D, higher magnification images of the yolk in wild-type and dgat2^sa13945 mutant embryos at 3 dpf highlighting visible accumulation of droplets; left: anterior yolk, right: yolk extension (YE), ∗ melanocyte; Scale = 50 μm. E, dgat2^sa13945 larvae at 6 dpf sometimes retain yolk and/or large droplets; right shows magnification of noted regions; Scale = 200 μm. F, images of wild-type and dgat2^sa13945 mutant embryos stained with Oil Red O to visualize neutral lipid. Arrows point to lipid droplets; images are representative of embryos from three experiments at each stage with n = 8 to 10 embryos per experiment; Scale = 500 μm. G, quantitative RT-PCR for dgat2 expression in wild-type, dgat2^sa13945/+, and dgat2^sa13945 embryos at 3 dpf (N = 6; 10 pooled fish per sample/genotype, one-way ANOVA, p = 0.0844). H, depiction of the dgat2 gene structure highlighting the nature and locations of the c747 and c765 CRISPR/Cas9 mutations in addition to the sa13945 mutation in dgat2 [GRCz11; ENSDARG00000018846, transcript 201 (ENSDART00000066793.7)], for more detail, see Fig. S2. I, representative images of yolk opacity in homozygous mutant dgat2^c747 and dgat2^c765 embryos at 3 dpf; scale = 200 μm.

Cytoplasmic lipid droplets and swollen, electron-dense ER are present in the yolk syncytial layer of dgat2 mutants.A, depiction of the cross-sectional view of a 4 dpf zebrafish illustrating the yolk syncytial layer (YSL) surrounding the yolk mass. The dashed box indicates the location of the images in panel B. B, representative transmission electron micrographs of the yolk and YSL from wild-type and dgat2^sa13945 mutants; dashed lines delineate the YSL region, n = nucleus, YP = yolk platelet, LD and ∗ = lipid droplet, scale = 10 μm. C, higher magnification images of the YSL region. Arrows indicate the endoplasmic reticulum which often encircles the mitochondria (m) in both wild-type and dgat2^sa13945 mutants; scale = 1 μm. D, examples of ER morphology in wild-type and dgat2^sa13945 embryos; scale = 0.5 μm. E, quantification of ER width, N = 4 fish per genotype (30–40 measurements per fish from 3 to 4 images, mean of each fish is shown by the blue and magenta points); bars indicate overall mean ± SD, unpaired t test, ∗∗p = 0.0022. F, representative confocal z-sections of the YSL in a wild-type fish and dgat2^sa13945 mutant carrying the Fus(EGFP-plin2)/+ reporter; scale = 20 μm, 3 dpf, n = 18 dgat2^sa13945 & n = 15 WT or dgat2^sa13945/+ fish from two independent experiments.

dgat2 mutant embryos are still capable of triacylglycerol synthesis in the YSL.A, TopFluor C11 fatty acid was mixed with canola oil and microinjected into the yolk mass of dgat2^sa13945 mutants and siblings at 3 dpf. B, representative image of the TopFluor C11 signal in a dgat2^sa13945 zebrafish yolk sac 4 h after injection; scale = 200 μm. C and D, lipids were extracted from TopFluor C11 injected embryos and separated by thin layer chromatography (20 pooled embryos per sample, wild-type and dgat2^sa13945/+ embryos were combined). C, lipid classes were identified using TopFluor-labeled lipid standards. Arrow indicates TAG synthesized from TopFlour C11 in vivo. (CE: cholesterol ester, TAG: triacylglycerol, DAG: diacylglycerol, MAG: monoacylglycerol; PL: phospholipid; ∗ denote autofluorescent lipids in the fish lysate). D, the quantity of lipids in each class was quantified and expressed as a percent of total fluorescent lipids; N = 6 experiments, mean ± SD, Two-way ANOVA, not significantly different, p = 0.9929. Note, it appears that the TopFluor C11 is unable to be incorporated into cholesterol esters. E, large lipid droplets were manually dissected out of dgat2^sa13945 larvae at 6 dpf (40–50 LDs from 30 – 40 larvae). Lipids in LDs or an equivalent volume of media surrounding dissected larvae were separated by multistep gradient HPLC and detected with a charged aerosol detector (CAD) in picoamperes (pA) as described in (75). The lipid droplet samples contain large quantities of triacylglycerols and cholesterol esters. Traces are representative of three independent experiments.

Dgat2 mutants produce fewer, abnormally smallApoB-containing lipoproteins.A, LipoGlo fish express ApoBb.1 with a C-terminal Nanoluciferase enzyme fusion (11)[Lipoprotein image re-used from (30) under the CC BY 4.0 license]. B, LipoGlo luminescence (RLU = relative luminescence units) in WT, dgat2^sa13945/+ and dgat2^sa13945 fish throughout embryonic development (2–6 dpf). Results represent pooled data from three independent experiments, n = 21 to 38 (WT), n = 65 to 75 (dgat2^sa13945/+), n = 32 to 43 (dgat2^sa13945) fish/genotype/time-point; mean ± SD. Significance was determined with a two-way ANOVA (overall p < 0.0001 for genotype) and Bonferroni’s multiple comparison test was performed to compare genotypes at each day of development (∗∗∗∗p < 0.0001 sa13945 versus WT and sa13945/+, ∗∗∗p < 0.001 sa13945 versus WT). C, representative LipoGlo PAGE gel of wild-type and dgat2^sa13945 mutants shows ApoB-containing lipoprotein (B-lp) size distribution from whole embryo lysates during development. B-lps are divided into four classes based on mobility, including zero mobility (ZM), very low density lipoproteins (VLDL), intermediate density lipoproteins (IDL) and low density lipoprotein (LDL). For heterozygote data, see original gels in Fig. S6. D, graphs show B-lp subclass abundance for WT, dgat2^sa13945/+ and dgat2^sa13945 fish at each day of embryonic development, analyzed from the gels in Fig. S6, as described in (11). Results represent pooled data from n = 9 fish/genotype/time-point; mean ± SD. For each particle class, significance was determined with a two-way ANOVA (overall p < 0.0280 ZM, p < 0.0001 VLDL, p < 0.0004 IDL, p < 0.0001 LDL for genotype) and Tukey’s multiple comparison test was performed to compare genotypes at each day of development (∗p < 0.05, ∗∗p < 0.01, ∗∗∗p < 0.001, ∗∗∗∗p < 0.0001 sa13945 versus WT are shown).

Dgat1 is not responsible for the residual triacylglycerol synthesis in the dgat2 mutants.A, in situ hybridization for dgat1a and dgat1b expression at 3 and 6 dpf in wild-type AB embryos. dgat1a and dgat1b are not detected in the YSL, but are expressed in the intestine; images are representative of all embryos from three experiments at each stage (n = 4–six embryos per probe per experiment); Scale bars = 200 μm. B, quantitative RT-PCR for dgat1a and dgat1b expression in wild-type, dgat2^sa13945/+, and dgat2^sa13945 embryos at three dpf (N = 5–6 clutches; 10 pooled fish per sample/genotype, mean ± SD, One-way ANOVA, p = 0.0887 dgat1a, p = 0.8773 dgat1b). C, depiction of the dgat1a and dgat1b gene structure highlighting the nature and locations of the c770 and c773 CRISPR/Cas9 mutations (GRCz11; dgat1a ENSDARG00000103503, transcript 201 (ENSDART00000158946.2), dgat1b ENSDARG00000054914, transcript 201 (ENSDART00000077185.5)). D, quantitative RT-PCR for dgat1a and dgat1b expression in dgat2^sa13945versus dgat1a^c770;dgat1b^c773;dgat2^sa13945 triple mutant embryos at 3 dpf (N = 3 independent experiments; n = 9 fish per genotype; mean ± SD, unpaired t test, ∗∗∗p < 0.001). E, examples of yolk opacity phenotypes in dgat1a^c770;dgat1b^c773;dgat2^sa13945 triple mutant embryos at 3 dpf; Scale = 500 μm. F, embryos from in-crosses of dgat1a^c770/+;dgat1b^c773/+;dgat2^sa13945 parents were imaged and scored at three dpf for the degree of yolk opacity, binned into the four noted categories prior to genotyping and expressed as a percent of total embryos per genotype (N = 4 independent experiments, n = 22–89 fish per genotype). G, dgat2^sa13945 embryos were co-injected at the 1-cell stage with CMV:dgat1a-FLAG, CMV:dgat1b-FLAG and CMV:EGFP-CAAX plasmids, or CMV:EGFP-CAAX alone as a control. Bright-field images were obtained of all the embryos that expressed EGFP-CAAX in the YSL at 3 dpf; representative images of embryos from the two treatment groups (top, Scale = 500 μm). Images were assessed for the degree of yolk opacity, binned into the four noted categories of yolk opacity as noted in F and expressed as a percent of total EGFP-positive embryos/treatment group (n = 77 EGFP-CAAX and n = 77 dgat1-FLAG embryos pooled from three independent experiments, Chi-square test p < 0.0001).

Mogat3b is not responsible for the residual triacylglycerol synthesis in the dgat2 mutants.A, in situ hybridization for mogat2, mogat3a and mogat3b expression at 3 and 6 dpf in wild-type AB embryos. mogat mRNA is not detected in the YSL, but is expressed in the intestine; images are representative of all embryos from three independent experiments at each stage (n = 4–6 embryos per probe per experiment); Scale bars = 200 μm. B, quantitative RT-PCR for mogat2, mogat3a and mogat3b expression in wild-type, dgat2^sa13945/+, and dgat2^sa13945 embryos at three dpf (N = 6 clutches; 10 pooled fish per sample/genotype, One-way ANOVA, p = 0.8820 mogat1, p = 0.5285 mogat2, p = 0.6988 mogat3b). C, depiction of the mogat3b gene structure highlighting the nature and locations of the c858 and c862 CRISPR/Cas9 mutations (GRCz11; mogat3b ENSDARG00000003635, transcript 201 (ENSDART00000015136.10)). D, quantitative RT-PCR for mogat3b expression in dgat2^sa13945versus mogat3b^c858;dgat2^sa13945 or mogat3b^c862;dgat2^sa13945 double mutant embryos at 3 dpf (N = 3 independent experiments; 8–9 fish per genotype; unpaired t test, p = 0.6778 c858, ∗∗∗p < 0.001 c862). E, representative images of mogat3b^c858 and mogat3b^c862 homozygous mutant embryos at 3dpf; Scale = 200 μm. F, representative images of mogat3b^c858;dgat2^sa13945 and mogat3b^c862;dgat3^sa13945 double mutant embryos at 3 dpf; Scale = 200 μm. G, embryos from in-crosses of mogat3b^c858/+;dgat2^sa13945 parents were imaged and scored at 3 dpf for the degree of yolk opacity, binned into the four noted categories prior to genotyping and expressed as a percent of total embryos per genotype (N = 3 independent experiments, n = 62–130 fish per genotype, Chi-square test p < 0.0001). H, dgat2^sa13945 embryos were co-injected at the 1-cell stage with CMV:mogat3b-FLAG and CMV:EGFP-CAAX plasmids, or CMV:EGFP-CAAX alone as a control. Bright-field images were obtained of all the embryos that expressed EGFP-CAAX in the YSL at 3 dpf; representative images of embryos from the two treatment groups (left, Scale = 200 μm). Images were assessed for the degree of yolk opacity, binned into the four noted categories of yolk opacity as noted in Fig. 6G and expressed as a percent of total EGFP-positive embryos/treatment group (n = 75 EGFP-CAAX and n = 119 mogat3b-FLAG embryos pooled from three independent experiments, Chi-square test p = 0.0923).

Loss of triacylglycerol synthesis enzymes sometimes alters yolk sac morphology.A,B, representative images of mogat3b^c858;dgat2^sa13945 mutants (A) and mogat3b^c862;dgat2^sa13945 mutants (B) with short (top) or no (bottom) yolk extensions (YE) at 3 dpf. Arrows indicate where the yolk extension should be located. Scale = 200 μm. C, all mogat3b^c858;dgat2^sa13945 mutants initially have a YE, but the YE can be lost as the embryo develops (arrows). However, the gut develops normally (6 dpf), and sometimes large residual lipid droplets are noted (arrowheads). Scale = 200 μm. D, The percentage of embryos per clutch exhibiting abnormal yolk extension morphology (lack of YE, short YE, broken YE) on 3 dpf from in-crosses of the noted parental genotypes (N = 4–9 clutches per genotype (clutch size ranged from 20 to 300 embryos)).

Embryo health and survival is significantly affected by concurrent loss of four enzymes in the triacylglycerol synthesis pathway.A, representative images of embryos (3 dpf) and larvae (6 dpf) mutant for mogat3b^c858;dgat1b^c770;dgat1b^c773;dgat2^sa13945; Scale = 500 μm. The dashed white box in the bottom panel indicates the location of the inset; Scale = 100 μm. B, the percentage of embryos per clutch exhibiting abnormal yolk extension morphology (lack of YE, short YE, broken YE) on 3 dpf from in-crosses of the noted parental genotypes (N = 3–7 clutches per genotype (clutch size ranged from 24 to 600 embryos). C–H, standard length, mass and body mass index (BMI) data (C, E, G) for progeny from in-crosses of the noted parental genotypes at 14 weeks of age followed by genotyping. Graphs include data from both sexes; N = 2 to 3 clutches per genotype (n = 112 fish (C), n = 120 fish (E), n = 223 fish (G)); mean ± SD, One-way ANOVA, p < 0.0001 for all sets, Tukey’s multiple comparisons test ∗∗p < 0.01, ∗∗∗∗p < 0.0001. D, F, H, representative images of fish at 15 weeks of age; Scale = 1 cm. I, representative images of dgat1b^c770;dgat1b^c773;dgat2^sa13945/c765 and dgat1b^c770;dgat1b^c773;dgat2^c765 triple mutant fish at 3 and 6 dpf; Scale = 500 μm. The dashed white box indicates the location of the image in panel J. J, lipid droplets are still visible in the YSL of dgat1b^c770;dgat1b^c773;dgat2^c765 mutants, ∗ melanocyte; Scale = 100 μm. K, representative images of dgat1b^c770;dgat1b^c773;dgat2^sa13945/c765 and dgat1b^c770;dgat1b^c773;dgat2^c765 triple mutant fish at 15 weeks of age; Scale = 1 cm. L–N, standard length, mass and body mass index (BMI) data at 14 weeks of age of progeny from in-crosses of mogat3b^c858/+;dgat1a^c770/+;dgat1b^c773/+;dgat2^sa13945/c765 and mogat3b^c858/+;dgat1a^c770/+;dgat1b^c773/+;dgat2^c765 or mogat3b^c858/+;dgat1a^c770;dgat1b^c773;dgat2^sa13945/c765 x mogat3b^c858/+;dgat1a^c770/+;dgat1b^c773/+;dgat2^c765 parents. Graphs include data from both sexes, for clarity only selected genotypes are shown; N = 16 clutches (n = 77 of the shown genotypes, 431 total fish); mean ± SD, One-way ANOVA for shown genotypes, p = 0.0722 (L), p = 0.0843 (M), p = 0.1665 (N). O, representative images of mogat3b^c858/+;dgat1a^c770;dgat1b^c773;dgat2^c765 and mogat3b^c858;dgat1a^c770;dgat1b^c773;dgat2^c765 quadruple mutants at 3 dpf; Scale = 500 μm.