Optical clarity of larvae makes the zebrafish ideal for real-time analyses of vertebrate organ function through the use of fluorescent reporters of enzymatic activities. A key function of digestive organs is to couple the generation of enzymes with mechanical processes that enable nutrient availability and absorption. However, it has been extremely difficult, and in many cases not possible, to directly observe digestive processes in a live vertebrate. Here we describe a new method to visualize intestinal protein and lipid processing simultaneously in live zebrafish larvae using a quenched fluorescent protein (EnzChek) and phospholipid (PED6). By employing these reagents, we found that wild-type larvae exhibit significant variation in intestinal phospholipase and protease activities within a group but display a strong correlation between the activities within individuals. Furthermore, we found that pancreas function is essential for larval digestive protease activity but not for larval intestinal phospholipase activity. Although fat-free (ffr) mutant larvae were previously described to exhibit impaired lipid processes, we found they also had significantly reduced protease activity. Finally, we selected and evaluated compounds that were previously suggested to have altered phospholipase activity and are known or suspected to have inflammatory effects in the intestinal tract including nonsteroidal anti-inflammatory drugs, and identified a compound that significantly increases intestinal phospholipid processing. Thus the multiple fluorescent reporter-based methodology facilitates the rapid analysis of digestive organ function in live zebrafish larvae.
- exocrine pancreas
dietary proteins and lipids that enter the lumen of the intestine are digested by various extracellular enzymes and mainly absorbed by the specialized, highly polarized enterocytes that are the main constituent of the intestinal epithelium (3, 25). In mammals, the secretion of digestive enzymes from multiple organs is tightly regulated by a combination of neural, hormonal, and paracrine signaling pathways (10). To capture the complexity of signals that regulate digestive processes, a model system must maintain the normal physiological state of the intestine. Thus in vivo screening systems that are amenable to both genetic and high-throughput pharmacological approaches are feasible for identifying the genes or chemicals that exert crucial roles on digestive physiology. However, it has been extremely difficult, and in many cases not possible, to directly observe digestive processes in a live vertebrate.
Larval zebrafish are an excellent model vertebrate system for studying both protein and lipid metabolism since many features of its digestive physiology are similar to those of mammals (2, 19–21, 23) and its optical clarity makes it ideal for the use of fluorescent reporters. Previously we utilized a quenched fluorescent reporter (PED6) of phospholipase activity for mutagenesis screens and succeeded in isolating several mutants with altered intestinal lipid processing (2). Furthermore, we were able to identify a new gene (fat-free) that regulates Golgi structure and is required for efficient intestinal lipid processing (7). These results demonstrate the usefulness of the fluorescent reporter-based assays to investigate digestive physiology in live zebrafish larvae. Furthermore, our approach is quite novel: to date, of >10,000 zebrafish publications there are fewer than 10 studies of digestive organ function.
Although PED6 has since been used to assay intestinal lipid processing in a number of zebrafish mutants (16, 19), there remains significant variance of fluorescence observed in the larval digestive tract of wild-type siblings. This variability between larvae of the same environmental and genetic background significantly limits the usefulness of PED6 in both chemical and genetic screens. We hypothesized that by combining PED6 with other reporters of digestive function we could significantly enhance the utility of this screening tool for lipid metabolism. Previously, a protease assay employing a fluorescently quenched protein was developed and utilized to monitor protease activity in a variety of biological contexts (5, 8, 12, 17). Here we report that an intramolecularly quenched fluorescent protein (EnzChek, Invitrogen) can be used to assay intestinal protease activity in live zebrafish larvae. Furthermore, intestinal protease activity was found to correlate strongly with intestinal phospholipase activity, enabling us to use the ratio of digestive activities as opposed to the absolute values for screening. These experiments indicate that automated detection and quantification of the ratio of fluorescent digestive reporters is feasible. The simultaneous screening of multiple physiological and developmental processes results in a powerful in vivo readout of digestive processes.
MATERIALS AND METHODS
Fish husbandry and animal use protocol.
Fish were maintained under standard conditions (26). Embryos were collected from natural spawning and raised up in embryo medium (26). All zebrafish care and experimental procedures were carried out as specified in our independently reviewed and approved Institutional Animal Care and Use Committee Protocol (no. 647A).
Preparation of fluorescent reporters.
PED6 [N-((6-(2,4-dinitrophenyl)amino)hexanoyl)-1-palmitoyl-2-BODIPY-FL-pentanoyl-sn-glycero- 3-phosphoethanolamine] (D23739, Invitrogen, Carlsbad, CA) was dissolved in chloroform and purified by preparative TLC on silica gel (Whatman LK5D, P.J. Cobert Associates, St. Louis, MO) developed with CHCl3/MeOH/dH2O (50/10/1). The single spot of pure PED6 was scraped from the silica gel with a razor blade and eluted with CHCl3/MeOH (2:1). Purified PED6 was dissolved with CHCl3/MeOH (2:1) and stored at −80°C. Quenched BODIPY TR-X casein (EnzChek, E6639, Invitrogen) was dissolved in 0.1 M sodium bicarbonate (pH 8.3) at a concentration of 1 mg/ml and stored at −20°C.
Fish labeling and quantification of fluorescence intensity.
Immediately before use, aliquots of PED6 were dehydrated completely (SpeedVac, Thermo Savant, Holbrook, NY) and reconstituted to a concentration of 60 μg/ml in 10% EtOH/dH2O (vol/vol). Aliquots of EnzChek were diluted to 400 μg/ml in dH2O. 300 μl of 2% fluorescent microsphere slurry (F8783, Invitrogen) was transferred to a microcentrifuge tube, washed twice with dH2O containing 0.1% bovine serum albumin (A6003, Sigma), reconstituted with 25 μl of 0.1% BSA/dH2O, and sonicated at an output level of 4.0 for 3 s × 3 times (Sonicator 3000, Misonix, Farmingdale, NY). Larvae (n = 5 to 15) were transferred to a microcentrifuge tube, washed three times, and placed in 200 μl of embryo medium. Reconstituted PED6, EnzChek, and fluorescent microsphere solution were added into the tube with larvae at a final concentration of 3 μg/ml, 20 μg/ml, and 2.4%, respectively, and incubated for 3 h at room temperature. After soaking, larvae were placed on ice, washed three times with cold embryo medium for anesthetization, and arrayed in 3% methylcellulose on glass slides. Larvae with any developmental abnormality (screened by absence of gall bladder or short body length) were removed. Images were captured by using a fluorescence stereomicroscope (SMZ1500, Nikon, Tokyo, Japan) with a charge-coupled device camera (AxioCam HR, Carl Zeiss AG, Oberkochen, Germany). Images were taken with two different exposure times for each fluorescent reporter (0.3 and 1 s for PED6; 1 and 3 s for EnzChek and microsphere); the figures show the images taken at the longer exposure time whereas quantification was performed using the images taken at the shorter time to optimize intensity and remove pixel saturation. The fluorescence intensity of whole individual larvae was calculated within a 25,000 pixel rectangular region as shown in Fig. 1D by use of Axiovision v4.6 software (Carl Zeiss AG).
Embryos at 1- to 4-cell stage were injected (1 nl) with morpholino (MO) or water (control) via glass needles fitted to a gas pressure injector (PL1–100, Harvard Apparatus). Glass needles were pulled (1 × 90 mm, Narishige; P-97, Flaming/Brown) and filled with stock solutions of MO (5′-CCAACACAGTGTCCATTTTTTGTGC-3′) targeted to the ptf1a gene (4 ng/nl, Gene Tools). Phenol red solution (0.2% final concentration) was added to all injection solutions to visualize injected embryos.
Quantitative PCR analysis.
Total RNA was extracted from groups of embryos by the TRIzol method (Invitrogen), and subsequently reverse transcribed using oligo-dT priming and SuperScript II (Invitrogen). Synthesized cDNA was quantified using the SYBRgreen method (Applied Biosystems, Foster City, CA). Primer sequences were as follows: zebrafish phospholipase A2 (PLA2) group1B, 5′-CCCGGTGGATGAACTGGAC-3′ and 5′-ATT TCAGTGTAGGGGTTGTCCAAG-3′; PLA2 group III, 5′-GGTGTTTGGAGGACAGCTCTTTC-3′ and 5′-TTATTAACACCCAGGAGACAGCG-3′; and PLA2 group XII, 5′-TACACACCGGTTCCTCGTCCT-3′ and 5′-GCTCATTACAGCATCTGGTCATAGAC-3′.
Genotyping of fat-free mutants.
Following analysis of digestive function, individual larvae obtained from a cross of fat-free −/+ parents were genotyped. Genomic DNA was extracted using standard protocols (26) and the fat-free genotype was determined by PCR using primers 5′-AGGCATTCCTTGTTGTGGAGAATAT-3′ and 5′-GTATTCGGATTTCCCACAGGCAG-3′.
Larvae were exposed to L364,718 (2034, Tocris Bioscience, Ellisville, MI) in embryo medium for 12 h (L364,718; 10 μM) prior to reporter exposure. For assay validation, 10 candidate compounds were selected for 1) altered PED6 activity in preliminary screens including that of a 640-compound library (Prestwick Chemical, Illkirch, France) and 2) a known or suspected inflammatory effect on the intestinal tract. The 10 compounds were added to embryo medium for 14–16 h (5, 10, 50, or 100 μM) before labeling with the reporters. Control larvae were immersed in embryo medium containing 0.01 and 0.1% DMSO for the L364,718 and validation experiments, respectively.
Statistical analysis was performed by ANOVA followed by Tukey's test to evaluate the effect of drugs or the fat-free mutation. The F-test followed by the Student's t-test was used for the evaluation of ptf1a MO and CCK receptor antagonist treatment.
To simultaneously visualize intestinal lipase and protease activity in live zebrafish, we fed larvae [6 days postfertilization (dpf)] with two intramolecularly quenched fluorescent reporters [PED6 (2, 6) and EnzChek (5, 8)] and nondigestible fluorescent microspheres. Each of the three fluorophores in the triple screening cocktail (PED6, EnzChek, and microspheres) exhibits a sufficiently distinct emission wavelength (Fig. 1A). We confirmed that fluorescence emission of each reporter did not overlap with emission from the other reporters, allowing us to simultaneously assay lipase, protease, and swallowing activities (Fig. 1B). We hypothesized that the variance in the fluorescence intensity of PED6 following feeding reflects the natural variance in the quantity of PED6 ingested by each larvae and/or the amount of larval digestive enzyme secretion. Thus, by being fed the cocktail of reporters, animals that consumed more PED6 would also be expected to consume more EnzChek. Consistent with this hypothesis, a strong correlation (r2 = 0.92) was observed between the intestinal fluorescence resulting from PED6 cleavage and from EnzChek proteolysis (Fig. 1C). To further test the hypothesis, larvae were treated with tricaine methanesulfonate, the most commonly used agent to immobilize zebrafish (1). Although tricaine-treated larvae exhibited a significant dose-dependent reduction of fluorescence intensity in all three reporters (P < 0.05; Fig. 1, D and E), the ratio of PED6 to EnzChek fluorescence was not significantly altered (P > 0.05; Fig. 1E). These results demonstrate the utility of the multiple fluorescent reporter-based assay to detect alterations in digestive enzyme activities while simultaneously correcting for the amount of reporter ingested by each larva.
The function of the exocrine pancreas can be selectively evaluated.
To explore the physiological processes that modify PED6 and EnzChek and ultimately lead to their fluorescent emissions, we utilized targeted antisense reagents to alter digestive organ development. We chose to initially examine the role of the exocrine pancreas because it is known in mammals to secrete pancreatic juice containing various digestive enzymes including lipases and proteases critical for intestinal nutrient absorption (10). Previous studies have established that exocrine pancreas development can be efficiently eliminated by targeting the ptf1a transcription factor with an antisense MO injected at the 1–8 cell stage (11). This treatment leaves the endocrine pancreas unaffected as evidenced by continued insulin expression in injected larvae (11).
Analysis of ptf1a morphant larvae (5 dpf) incubated with the multiple fluorescent reporters indicated that protease activity was markedly reduced (Fig. 2, A and E, P < 0.01). However, these larvae retained their ability to swallow fluorescent microspheres and efficiently process exogenous phospholipids, as indicated by wild-type levels of PED6 fluorescence (Fig. 2, A and E), resulting in a significant increase in the ratio of PED6/EnzChek in the ptf1a morphant (Fig. 2E, P < 0.05). The EnzChek/PED6 slope was significantly altered in ptf1a morphants (means ± SE: 62 ± 1.4° and 6.1 ± 3.4° for control and ptf1a morphants, respectively; n = 3, P < 0.01) (Fig. 2C). However, ptf1a morphants assayed 1 day older (6 dpf) exhibited reductions in both PED6 and EnzChek signals (Fig. 2, B and E, P < 0.01) but not in the amount of swallowed microspheres (Fig. 2E), resulting in no change in the EnzChek/PED6 slope (means ± SE: 41 ± 4.5° and 33 ± 3.8° for control and ptf1a morphant, respectively; n = 3, P > 0.05) (Fig. 2D). When comparing the reporter data between morphants and wild-type larvae, we observed no overlap (Fig. 2D).
To test the hypothesis that the observed increase in PED6 signal between 5 and 6 dpf was due to increased expression of PED6 metabolizing enzymes, quantitative PCR was performed. There are three main types of phospholipase A2 enzymes (groups IB, III, and XII) (18) that likely cleave PED6, resulting in an increase in fluorescence. We found that between 5 and 6 dpf there is a significant increase in expression of both group IB (also known as pancreatic secretory phospholipase A2) and group III phospholipase A2 (Fig. 2F). These results suggest that lipase derived from exocrine pancreas is the main source of intestinal phospholipase activity at 6 dpf, but not at 5 dpf.
Evaluating the role of cholecystokinin signaling.
The selective reduction of protease activity in the 5 dpf ptf1a morphants drove us to further explore the hormonal signals that can regulate phospholipase and protease secretion. We applied a pharmacological strategy that initially focused on roles of cholecystokinin (CCK), a peptide hormone that can stimulate the secretion of digestive enzymes and induce satiety (15). When mammals ingest dietary lipid, CCK is secreted by duodenal enteroendocrine cells into the systemic circulation and subsequently activates the CCK receptor A (CCK-RA) in exocrine pancreas (15). In mammals, CCK-RA activation results in the release of digestive enzymes and bile acids into the digestive tract (15). To examine the contribution of CCK in the digestive physiology of zebrafish, larvae (5 dpf or 6 dpf) were treated with a CCK-RA antagonist and analyzed using the cocktail of fluorescent reporters.
CCK-RA antagonist treatment of 5 dpf zebrafish reduced protease activity (Fig. 3A, C, P < 0.05) but had no effect on phospholipase activity or the amount of swallowed microspheres (Fig. 3, A and C, P > 0.05) In contrast, 6 dpf larvae treated with the CCK-RA antagonist exhibited significant reductions in both phospholipase (Fig. 3, B and C, P < 0.01) and protease activities (Fig. 3, B and C, P < 0.01) but not in the amount of swallowed microspheres (Fig. 3, B and C, P > 0.05). The observation that the CCK-RA antagonist significantly affects intestinal phospholipase at the same developmental stage (6 dpf) as the ptf1a MO supports the hypothesis that CCK exerts its primary effect on the exocrine pancreas. Consistent with this hypothesis, the effects of CCK-RA antagonist were abolished in ptf1a morphants (Fig. 3D). Taken together, these data suggest that CCK signaling mediates exocrine pancreas-derived intestinal protease activity at 5 dpf and phospholipase activity at 6 dpf.
Analysis of mutants: fat-free revisited.
Previously, in our PED6 based genetic screen, we identified the fat-free (ffr) mutation that was characterized by impaired intestinal phospholipase activity (2). Subsequent work identified the molecular nature of this mutation as a truncation of a protein with similarity to conserved oligomeric Golgi (COG) complex 8 (COG8) that is ubiquitously expressed in zebrafish larvae and localizes to the Golgi (7). Normally, zymogen granules containing pancreatic juice are formed within the exocrine pancreas. The ffr mutants exhibited significantly smaller zymogen granules (7), suggesting that the amount of secreted digestive enzymes would be attenuated. As expected, not only phospholipase but also protease activities (Fig. 4C, P < 0.01) are attenuated in intestines of ffr mutants, whereas microsphere ingestion is not, as demonstrated by the unchanged slope of the PED6 vs. EnzChek plot (means ± SE: 34 ± 7.5°, 54 ± 11° for control and ffr mutant, respectively; n = 3, P > 0.05) (Fig. 4). These results suggest that ffr alters the activity of multiple digestive enzymes, not just phospholipase, by reducing the formation of pancreatic granules.
Validating the effect of small molecules on digestive enzyme activities.
Finally, we examined the pharmacological potential of the multiple fluorescent reporter-based assay in validating the effect of small molecules on intestinal digestive conditions. We tested 10 compounds (nimesulide, meloxicam, piroxicam, glafenine, mefenamic acid, etodrac, zoxazolamine, gemfibrozil, fenbendazole, and doxorubicine) including six nonsteroidal anti-inflammatory drugs (NSAIDs), which are known to cause inflammation in the intestine. One molecule, glafenine (G7020, Sigma, St. Louis, MO), was found to selectively enhance intestinal phospholipase activity. Glafenine was previously reported to inhibit cyclooxygenase and act as a NSAID but was withdrawn from clinical use because of severe adverse reactions including hepatotoxicity, nephrotoxicity, and gastrointestinal disturbances (22, 24). Glafenine-treated larvae exhibited significantly enhanced PED6 intensity without significant effect on protease activity or amount of swallowed microspheres, thus demonstrating a significant increase in the ratio of PED6 to EnzChek fluorescence (Fig. 5, A and B). These results evidence the usefulness of multiple fluorescent reporters to analyze small molecules for an effect on intestinal digestive activity.
In this study we utilized multiple fluorescent reporters with different emission wavelengths to simultaneously monitor digestive phospholipase, protease, and swallowing activities. We observed a strong correlation between the fluorescence intensity of the reporters PED6 and EnzChek that suggests a coupling between the protease and lipase signals. This coupling likely reflects the natural variance in the amounts of digestive enzyme substrates consumed by individual larvae.
The approach of using multiple fluorescent reporters in live zebrafish larvae enables the study of many aspects of the vertebrate digestive process while assaying relevant physiology. In the present study, we demonstrated that CCK signaling regulates the intestinal digestive activity in zebrafish, as in mammals (Fig. 3). Furthermore, we focused on the digestive role of the exocrine pancreas, which is the main source of intestinal digestive enzymes in mammals and found that, unlike protease activity, exocrine pancreas-derived phospholipase activity dramatically increases between 5 and 6 dpf (Fig. 2E), raising several possibilities: 1) phospholipase expression level is elevated between 5 and 6 dpf in exocrine pancreas, 2) secretory machinery specific to phospholipase only becomes fully functional at 6 dpf, or 3) cofactors which enhance the activity of intestinal phospholipase are not produced until 6 dpf.
To address these hypotheses, quantitative PCR analysis of the major PED6 metabolizing phospholipase enzymes was performed. We found that zebrafish larvae significantly increase their expression of a low molecular weight secreted pancreatic phospholipase A2 enzyme (group IB) during this developmental period (Fig. 2F). Taken together, these results suggest that one source of the variance observed by investigators using only PED6 could have been differences in developmental staging. Future studies should carefully monitor developmental stage (e.g., by measuring larval length) and if possible perform digestive assays at the later stage (6 dpf).
The finding that both lipase and protease activities were coupled in our assay is consistent with the hypothesis that larvae that consume more PED6 would also consume more EnzChek. Thus we originally hoped to normalize the fluorescence intensity of the reporters by the fluorescent signal resulting from ingested microspheres. However, when the lipase-to-microsphere and protease-to-microsphere fluorescence ratios were determined for each individual larvae from a collection of similarly treated wild-type larvae, the variance increased from that observed with just the fluorescence intensity from a single enzyme reporter (this was not the case when lipase-to-protease ratio was determined). These data suggested a less than optimal correlation between the amount of ingested microspheres and ingested reporters of enzymatic activity. Most likely this is due to a nonlinearity in the fluorescence microsphere signal that occurs when these particles clump together (Figs. 1D, 4, and 5). However, we did find that average microsphere signal from a group of larvae, as well as PED6 and EnzChek signals, decrease in a dose-dependent manner by anesthetic treatment. Thus our method uses microsphere ingestion data as an assessment for whether an experimental group (an average of individual fluorescent signals) consumes significantly less or more than the control group.
An in vivo assay with multiple fluorescent reporters has great potential to identify novel genes or small molecules involved in the digestive process. In the present study, we showed that both phospholipase and protease activities are significantly reduced in the intestine of 6 dpf ffr mutant larvae. Intriguingly, phospholipase activity was also lowered at 5 dpf in contrast to the ptf1a morphants and CCK-RA-treated fish (Figs. 2 and 3) (14). Considering that the ffr gene is ubiquitously expressed, the secretory machinery of digestive enzymes in ffr fish might be affected not only in exocrine pancreas but also in other digestive organs such as intestine.
NSAIDs are known to irritate the gastrointestinal tract by inhibiting cyclooxygenases, which produce the protective prostaglandins that maintain the gastrointestinal mucosa (14). One group of phospholipase A2 enzymes (group II) is known to play a role in the pathogenesis of various inflammatory diseases (9) and is elevated in the inflamed colonic mucosa of inflammatory bowel disease patients including those with Crohn's disease and ulcerative colitis (4, 13). Thus the enhanced intestinal phospholipase activity in glafenine treated larvae might reflect an enhanced intestinal inflammatory response. Ongoing studies are addressing the consequences of glafenine treatment and its mechanism of action. Of note, we examined the effect of other NSAIDs (data not shown);only glafenine enhanced PED6 fluorescence, suggesting a specific effect on the intestinal mucosa.
In conclusion, we report a novel assay to monitor digestive processes in vivo by monitoring both intestinal phospholipase and protease activities. This method is amenable to both genetic and small compound screens.
K. Hama is supported by the Japan Society for the Promotion of Science (Postdoctoral Fellowship for Research Abroad) and the Astellas Foundation for Research on Metabolic Disorders (Fellowship for Research Abroad). This work was supported in part by National Institutes of Health Grants DK56211 (S. D. Leach), DK077480 (E. Provost), GM63904 (S. A. Farber), and R44 DK064472 (A. L. Rubinstein, T. C. Baranowski).
The authors thank Cara Jacob for technical assistance.
The costs of publication of this article were defrayed in part by the payment of page charges. The article must therefore be hereby marked “advertisement” in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
- Copyright © 2009 the American Physiological Society