Genetic Analysis of Digestive Physiology Using Fluorescent Phospholipid Reporters

See allHide authors and affiliations

Science  18 May 2001:
Vol. 292, Issue 5520, pp. 1385-1388
DOI: 10.1126/science.1060418


Zebrafish are a valuable model for mammalian lipid metabolism; larvae process lipids similarly through the intestine and hepatobiliary system and respond to drugs that block cholesterol synthesis in humans. After ingestion of fluorescently quenched phospholipids, endogenous lipase activity and rapid transport of cleavage products results in intense gall bladder fluorescence. Genetic screening identifies zebrafish mutants, such as fat free, that show normal digestive organ morphology but severely reduced phospholipid and cholesterol processing. Thus, fluorescent lipids provide a sensitive readout of lipid metabolism and are a powerful tool for identifying genes that mediate vertebrate digestive physiology.

To assay lipid metabolism in living larvae, we synthesized fluorescently tagged or quenched phospholipids (Figs. 1A and 2A). Such modified lipids are effective substrates for phospholipase A2 (PLA2) cleavage and sensitive reporters of enzymatic activity in vivo (1, 2). PLA2 was targeted because of its importance in the generation of lipid signaling molecules, host defenses, lipid absorption, and cancer (3–5). Given the shared features of lipid processing in mammals and teleosts (6, 7), zebrafish mutagenesis screens using lipid reporters should identify genes relevant to human lipid metabolism and disease.

Reporters were constructed by covalently linking fluorescent moieties to sites adjoining the cleavage site of phospholipids. Dye-dye or dye-quencher interactions modify fluorescence without impeding enzyme-substrate interactions (8). PLA2 cleavage results in immediate unquenching and detectable fluorescence. Quenched phospholipids [N-((6-(2,4-dinitrophenyl)amino)hexanoyl)-1-palmitoyl-2- BODIPY-FL-pentanoyl-sn-glycero-3-phosphoethanolamine (PED6) (2)] allow subcellular visualization of PLA2 activity in zebrafish embryos (1) and reveal organ-specific activity (9). Larvae [5 days post-fertilization (dpf)] bathed in PED6 show intense gall bladder fluorescence and, shortly thereafter, intestinal luminal fluorescence (Fig. 1C).

Figure 1

PED6, a fluorescent PLA2 reporter. (A) Dinitrophenyl on the phospholipid head group (red arrowhead) quenches emission of the BODIPY-labeled acyl chain (excitation 505 nm) (green arrowhead). PLA2 cleavage (arrow) liberates the BODIPY-acyl chain, resulting in unquenching and green fluorescent emission. (B) Bright-field image of a 5 dpf larva soaked in PED6 (0.3 μg/ml, 2 hours). (C) Corresponding fluorescent image, with intestinal (arrow) and gall bladder (arrowhead) labeling. (D) Larva soaked in BODIPY-C5-PC (0.2 μg/ml). In contrast to (C), unquenched fluorescent lipid labels the pharynx (arrowhead), confirming that lipid is swallowed before gall bladder labeling (arrow). (E) Organ fluorescence intensity determined at specific times was normalized to intensity at 45 min. Data are expressed as means ± SEM. (Fand G) Atorvastatin (ATR) inhibits processing of PED6 (F) but not of BODIPY FL-C5 (Molecular Probes) (G). Larvae were bathed in fluorophore (0.6 μM) in the presence or absence of atorvastatin (Lipitor tablet suspension containing 1 mg/ml) (arrowhead, gall bladder). (H) Mouse digestive organs. (I) Gall bladder fluorescence after processing (t = 30 min) of PED6 (1 μg), administered by gavage. Symbols: gb, gall bladder; d, common bile duct; lv, liver. Scale bars, 1.0 mm (H and I), 200 μm (other images).

Larvae were bathed in an unquenched fluorescent lipid to confirm that lipids were swallowed (Fig. 1D). We determined that PED6 is cleaved within the intestine, because PLA2 activity was not detected in bile (10) and fluorescent PED6 metabolites underwent rapid hepatobiliary transport, labeling the liver before the gall bladder (Fig. 1E). Moreover, PED6 labeling of the gall bladder was completely blocked by atorvastatin (Lipitor, Parke-Davis) (Fig. 1F), a potent inhibitor of cholesterol synthesis in humans (11). Because the addition of exogenous bile reversed this effect (12), and atorvastatin failed to inhibit processing of a water-soluble short-chain fatty acid (BODIPY FL-C5, Molecular Probes Inc.) (7) (Fig. 1G), we hypothesize that atorvastatin blocks the synthesis of cholesterol-derived biliary emulsifiers needed for dietary lipid absorption (13,14). These results, coupled with the rapid appearance of biliary fluorescence in mice fed PED6 (Fig. 1, H and I), demonstrate that PED6 metabolism in zebrafish corresponds to established mechanisms of lipid processing in mammals (15).

To determine the site of PLA2 activity more precisely, we developed a substrate [BODIPY FR-PC (Fig. 2A)] that emits distinct fluorescent profiles before and after cleavage (16). This phospholipid contains two dyes that interact via fluorescence resonance energy transfer (FRET) (17). The substrate and cleavage product have unique spectral signatures, allowing their tissue distribution to be distinguished by fluorescence microscopy. Excitation (505 nm) of micelle-incorporated BODIPY FR-PC produces an orange emission (568 nm; Fig. 2B) from the sn2 fluorophore via the FRET effect (18). However, cleavage of the substrate by purified PLA2 releases the sn2 acyl chain and abolishes FRET (Fig. 2B). Excitation of the resultant lysolipid produces a green emission (515 nm).

Figure 2

BODIPY FR-PC reporter reveals both substrate and cleavage products. (A) Upon integration into cells, excitation (505 nm) results in orange (568-nm) emission due to FRET. After PLA2 cleavage (arrow), 515-nm emission (green) is observed. (B) Fluorescence emission spectrum of BODIPY FR-PC (excitation, 505 nm) with peaks at 515 and 568 nm indicative of FRET. Solid line, before addition of PLA2; dashed line, after addition of Naja. naja PLA2. (C) Fluorescence of a 5 dpf larva after BODIPY FR-PC (5 μg/ml) labeling: Visualization of segment II enterocytes (left of red bar) after pinocytosis of uncleaved BODIPY FR-PC (1 hour after labeling). Arrowhead indicates the segment II domain. Green gall bladder fluorescence is not shown. (D) After 4 hours, BODIPY FR-PC metabolites reside in the gall bladder (green emission, green arrowhead). Uncleaved substrate (orange emission, red arrowhead) is observed in the anterior intestinal epithelium. Scale bars, 200 μm.

Within 1 hour, larvae bathed in BODIPY FR-PC showed a bright green fluorescent gall bladder and orange fluorescence within the apical cytoplasm of posterior intestinal cells (Fig. 2C). These cells are known to absorb luminal macromolecules through pinocytosis (19). Orange fluorescence indicative of uncleaved substrate was also observed in the anterior intestine epithelium, the site of lipid absorption in fish (6, 20). FRET emission (orange) was never detected in the liver even after prolonged incubation (Fig. 2D), supporting the hypothesis that gall bladder fluorescence after ingestion of labeled lipids is due to cleavage by intestinal PLA2.

To use fluorescent lipid assays for genetic screening, we first examined PED6 processing in mutants known to perturb intestinal morphology in a region-specific manner (21). The mutationsslim jim (slj) and piebald(pie) each cause degeneration of anterior intestinal epithelium and exocrine pancreas, whereas meltdown(mlt) results in cystic expansion of the posterior intestine. As shown in Fig. 3, A and B,mlt mutants retain PED6 processing in the anterior intestine and show normal levels of gall bladder fluorescence. In contrast,pie (Fig. 3, C and D) and slj (Fig. 3E) mutants display greatly reduced gall bladder labeling. Hence, fluorescent lipids can be used to identify mutants with abnormal digestive organ morphology.

Figure 3

(left). Lipid processing in intestinal mutants. (A and C) Bright-field images of mlt and pie larvae (5 dpf). (B and D) Fluorescent images corresponding to (A) and (C). Normal lipid processing of PED6 (0.3 μg/ml, 2 hours) occurs in mlt larva (arrowhead marks gall bladder). Abnormal lipid processing is seen in PED6-labeledpie (D) and slj (E) larvae; fluorescence is present in the intestinal lumen (white arrowhead) and reduced in the gall bladder (red arrowhead). Scale bar, 200 μm.

Lipid reporters have also proven effective for identifying mutations that perturb lipid metabolism without causing obvious morphological defects. In a pilot screen, larval progeny of individual F2families derived from an ENU mutagenesis protocol (22) were bathed in PED6 and screened for digestive organ morphology and phospholipid processing. PED6 fluorescent metabolites enhance visibility of digestive organ structure, facilitating scoring of gall bladder development, intestinal folding, and bile duct morphology (23). Of the 190 genomes screened, two mutations were identified and confirmed in the subsequent generation. The recessive lethal mutation canola produces a phenotype closely resembling that of slj and pie, with degeneration of the intestinal epithelium and reduced gall bladder and intestinal fluorescence (24). The fat free recessive lethal mutation causes a more pronounced reduction in fluorescence (Fig. 4B) . However, in contrast tocanola and other intestinal mutants, the fat freedigestive tract has a normal appearance at 5 dpf (Fig. 4A). Therefore, fat free would have escaped detection in a screen based solely on morphological criteria.

Figure 4

(right). fat free larvae lack gall bladder fluorescence. (A) Bright-field images of wild-type (WT) and mutant 5 dpf larvae. (B) Corresponding fluorescent image of the gall bladder after PED6 labeling (0.3 μg/ml) (arrowheads). (C) Biliary secretion of phenol red in fat free larvae. Mutant larvae identified by PED6 labeling were injected with phenol red as described (26) (arrowhead, gall bladder). (D) Reduced lipid processing in fat free mutants. TLC of the pooled lipid fraction from eight larvae is shown after labeling with BODIPY C5-PC (12 hours, 0.16 μg/ml); red arrowhead denotes BODIPY-cholesteryl ester, black arrowhead indicates BODIPY C5-PC (lane 1, WT; lane 2, fat free; lane 3, endogenous fluorescent lipids present in unlabeled WT larvae). TLC: solvent system 1 (chloroform:methanol:water; 35:7:0.7); solvent system 2 (ethyl ether:benzene:ethanol:acetic acid; 40:50:2:0.2). (E) fat freelarvae fail to accumulate 22-[N-(7-nitrobenz-2-oxa-1,3-diazol-4-yl)amino]-23,24-bisnor-5-cholen-3-ol (NBD-cholesterol) (2 hours, 3 μg/ml, solubilized with fish bile). Scale bars, 200 μm.

To determine how the fat free mutation affects larval digestive physiology, we first determined that swallowing was normal in mutants by their ingestion of fluorescent beads (25). Next, we demonstrated that the fat free hepatobiliary system was functional by monitoring biliary excretion of the organic anion phenol red (Fig. 4C) (26). Thin-layer chromatography (TLC) of metabolites produced by fat free larvae labeled with an unquenched BODIPY phosphatidylcholine analog (1) revealed that PLA2 processing was intact (27). However, extracts prepared from mutants contained significantly reduced levels of fluorescent cleavage products (Fig. 4D).

Lipid processing was further examined using a fluorescent cholesterol analog (28) whose absorption is dependent on biliary emulsification. Cholesterol was poorly absorbed by fat free larvae (Fig. 4E). Because uptake of BODIPY FL-C5, a more soluble short-chain fatty acid, was reduced to a lesser extent than uptake of cholesterol and PED6 (24), we propose thatfat free directly regulates bile synthesis or secretion rather than intestinal lipid absorption or transport. Further, the zebrafish fat free mutation provides in vivo evidence supporting the hypothesis, derived from mammalian cell culture experiments, that phospholipid turnover and cholesterol metabolism are coupled (29).

Mutagenesis screens using fluorescent lipids exploit the advantages of zebrafish for combining genetic analyses with imaging of enzymatic function. The existence of related lipid processing mechanisms in mammals and teleosts, and the finding that therapeutic drugs used to modify lipid metabolism in humans are active in zebrafish, suggest that genetic screens can be designed to probe the mechanistic basis of acquired and heritable human disorders. Fluorescent reporters are predicted to identify genes involved in diseases of lipid metabolism, such as atherosclerosis; in disorders of biliary secretion, such as biliary atresia; and in cancer, a disease in which lipid signaling plays an important role.

  • * These authors contributed equally to this work.

  • To whom correspondence should be addressed. E-mail: sfarber{at}, mpack{at}

  • Present address: Department of Microbiology and Immunology, Kimmel Cancer Center, Thomas Jefferson University, Philadelphia, PA 19107, USA.


View Abstract

Navigate This Article