2.1.

Worm Strains and Reagents

Wild type (Bristol N2), CB1370 [daf-2 (e1370)], BX26 [fat-2 (wa17)], and BX30 [fat-3 (wa22)] mutant C. elegans were obtained from the Caenorhabditis Genetics Center. The fat-2(wa17); vit-2p::vit-2::gfp mutant was a gift from Dr. M. A. Miller (University of Alabama, Birmingham, Alabama). Nile Red was purchased from Invitrogen (Grand Island, New York) for detection of lipid droplets. Levamisole was purchased from Sigma (St. Louis, Missouri) and used to paralyze worms for microscopy.

2.2.

Maintenance of C. elegans

All strains used in this study were cultured on nematode growth medium plate seeded with an E. coli OP50 and incubated at 20 to 22°C.¹⁹ Animals were analyzed by using Nile Red staining and CARS/TPEF microscopy.

2.3.

Nile Red Staining and Colocalization Analysis

Approximately 500 to 1000 larva 4 stage C. elegans were collected and washed twice by using phosphate-buffered saline (PBS) buffer. For fixation, worms were transferred to modified Ruvkun's Witches Brew solution (80 mM KCl, 20 mM NaCl, 7 mM ethylene glycol tetraacetic acid, 15 mM 1,4-Piperazinediethanesulfonic acid pH 7.4, 0.5 mM spermidine, 1% formaldehyde, 0.1% mercaptoethanol) in liquid nitrogen to freeze and thawed for 1 h.²⁰ Fixation solution was removed by washing twice in PBS buffer and worms were transferred to 1 ml M9 buffer with or without Nile Red dye ($1\mu \mathrm{g}/\mathrm{ml}$) and incubated at 4°C for 30 min. After washing twice with PBS to remove staining solution, worms were mounted on 5% agar pad for microscopy. The colocalization between CARS and Nile Red staining images was analyzed by using MetaMorph software (version 7.5, Molecular Devices, California).²¹ Both CARS and Nile Red images were thresholded to mark the area of lipid droplets. Colocalization of CARS over Nile Red is defined as the overlapping area of CARS over Nile Red, divided by the area of CARS (%). The colocalization between CARS and VIT-2::GFP fluorescence images was analyzed by the same protocol as described above.

2.4.

CARS/TPEF Microscopy and Lipid Droplet Quantification

Nile Red stained worms were collected as described above. For live worm imaging, $\sim 40$ worms were collected and paralyzed on the agar pad. Lipid droplets and Nile Red or GFP fluorescence were monitored using CARS and TPEF microscopy, respectively. The setup of CARS was described in the previous report.²² Briefly, two Ti-sapphire laser beams (Mira-900P and Mira-900F, Coherent, California) were synchronized by Synchro-Lock system (Coherent, California) and tuned at $\sim 710$ and $\sim 890\mathrm{nm}$ to obtain CARS imaging of lipid C-H stretching mode at $\sim 2845{\mathrm{cm}}^{-1}$. The forward CARS signal (at $\sim 590\mathrm{nm}$) is collected by a condenser ($\mathrm{NA}=0.55$), passing through a set of bandpass filters (FF01-630/92 and FF01-590/10, Semrock, New York) and detected by a photomultiplier tube (PMT) (R7400U-02, Hamamatsu, Japan). Fluorescence signal was collected by a $60\times$ $\mathrm{NA}=1.2$ water immersion objective, passing through a bandpass filter (FF01-624/40 for Nile Red and FF01-512/25 for GFP, Semrock, Rochester, New York), and detected by a PMT (R3896, Hamamatsu, Japan). The background, defined as the average intensity of CARS, was subtracted from the original image for quantification of CARS intensity.¹⁵ The number, size, and CARS intensity of lipid droplets were analyzed by using the MetaMorph software version 7.5.

3.1.

Detecting Lipid Droplets in C. elegans by Nile Red Fluorescence and CARS Microscopy

The distribution of lipid droplets in worms can be revealed by Nile Red staining, a common fluorescence indicator of neutral lipids. To evaluate the capability of CARS imaging on lipids of C. elegans, we compared the CARS and Nile Red fluorescence images of larva 4 stage worms with or without Nile Red staining. Lipid droplets in the hypodermis cell were easily detected by both methods [Fig. 1(a), upper row]. While almost no signal was detected in worms without Nile Red staining by TPEF microscopy, lipid droplets were clearly detected in the same worm by CARS microscopy [Fig. 1(a), lower row]. The colocalization of the lipid droplets between Nile Red fluorescence and CARS images is shown in Fig. 1(a) (right column). To estimate this colocalization, both CARS and Nile Red staining images were thresholded to mark the area of lipid droplets [Fig. 1(b), lower row], followed by the calculation of the colocalization (see Materials and Methods for the detailed description). The result shows a high colocalization of CARS over Nile Red ($62.7\pm 2.3\%$) of worms with Nile Red staining, while almost no colocalization ($0.4\pm 0.1\%$) is present in the sample without Nile Red staining [Fig. 1(c)]. These data indicate that the CARS signal well colocalizes with Nile Red signal in stained C. elegans. Furthermore, the enlarged images [the insets of Figs. 1(a) and 1(b)] clearly show that the resolution and contrast of the CARS image are better than those of the Nile Red fluorescence image, suggesting that CARS microscopy is a better method to visualize small lipid droplets.¹⁸ In contrast to the Nile Red staining method that requires sample fixation, CARS microscopy can be used to visualize lipid droplets directly in live organisms without labeling. This enables CARS microscopy to monitor dynamic change of lipid droplet distribution.

Fig. 1

The comparison of Nile Red and CARS imaging in detecting lipid droplets of C elegans. (a) Nile Red fluorescence and CARS images of lipid droplets in fixed worms with (upper row) or without (lower row) Nile Red staining (the heads of wild-type worms at larva 4 stage). Their merged images are shown in the right column. (b) Nile Red fluorescence and CARS images before (upper row) and after (lower row) thresholding. (c) The colocalization of CARS over Nile Red image. Data are from three independent experiments. Scale bars represent 10 μm. The numbers ($n$) of worms for each condition are as follows: for Nile Red staining, $n=35$; for non-Nile Red staining, $n=30$. Error bars indicate standard error of mean (SEM). ***$P<0.001$.

The daf-2 gene encodes for the insulin-like growth factor 1 receptor in C. elegans and is involved in the regulation of the life span.²³ The daf-2 strain is known as a lipid-enriched mutant in C. elegans and has been studied by using CARS microscopy.^13,14,24 We next examined daf-2 mutant to testify the quantification method of CARS images [Fig. 2(a)]. For quantification, after background subtraction and image thresholding, the lipid droplet regions were marked [Fig. 2(a) middle column], counted (number), and measured (size, intensity) by the software [Fig. 2(a) right column). Compared with previous reports,^13,14,24 similar quantitative results were obtained showing that the number and content of lipid droplets significantly increase in daf-2 mutant [Figs. 2(b) and 2(d)]. In contrast, we found that the lipid droplet size decreases in daf-2 mutant [Fig. 2(c)], while previous reports showed no statistical difference.^13,14,24 Different results of lipid droplet size may be due to the various stages of worms used. Nevertheless, the good specificity and sensitivity of CARS microscopy with label-free imaging, can provide a powerful tool to monitor and characterize lipid droplet distribution in live C. elegans.

Fig. 2

Lipid droplets in the wild type (WT) and daf-2 mutant. (a) The analysis process of CARS images for lipid droplet quantification in the larva 4 stage of the WT and daf-2 mutant. The scale bar represents 10 μm. Quantification of the number (b), size (c), and CARS intensity (d) of their lipid droplets. Data are from three independent experiments. The numbers ($n$) of worms for each strain are as follows: WT, $n=20$; daf-2, $n=20$. Error bars indicate SEM. **$P<0.01$; ***$P<0.001$.

3.2.

Significant Decrease of Lipid Droplets in Fat-2 Mutant

Although the body fat of fat-2 and fat-3 mutants has been examined by Nile Red staining and gas chromatography,^11,12 their lipid droplet patterns have not been characterized in detail. Here we used CARS microscopy to visualize their lipid in live worms and analyzed their lipid droplets. Figure 3(a) shows CARS images of the wild type, fat-2, and fat-3 mutants. Our results show a drastic decrease of lipid droplet number in fat-2 but not in fat-3 mutant [Fig. 3(b)]. Remarkable decreases in lipid droplet size and content were also found in the fat-2 mutant worms, while only a slight increase in lipid droplet size and a decrease in lipid content were observed in the fat-3 mutant [Figs. 3(c) and 3(d)]. Taken together, these data indicate that the loss of fat-2 function, but not fat-3, causes a significant decrease in number, size, and content of lipid droplets in C. elegans. In previous studies, fat-2 was well documented to be the upstream of fat-3 and is the unique enzyme to participate in the first step of PUFA biosynthesis.¹⁰ Consistently, we found more phenotypic changes of lipid droplets in fat-2 than in fat-3 mutant. To our knowledge, this is the first detailed characterization of lipid droplet patterns in fat-2 and fat-3 mutants. Our results suggest that the downregulation of PUFA synthesis in fat-2 mutant can be directly revealed by the change of lipid droplet pattern.

Fig. 3

Lipid droplets in the WT, fat-2, and fat-3 mutants. (a) CARS images of the larva 4 stage worms in the WT, fat-2, and fat-3 mutants. The scale bar represents 10 μm. I, intestine; V, vulva. Quantification of the number (b), size (c), and CARS intensity (d) of their lipid droplets. Data are from three independent experiments. The numbers ($n$) of worms for each strain are as follows: WT, $n=25$; fat-2, $n=16$; fat-3, $n=20$. Error bars indicate SEM. *$P<0.05$; ***$P<0.001$; ns, no significance.

3.3.

NonDroplet-Like Structure with Enhanced CARS Signal in the Uterus of Fat-2 and Fat-3 Mutants

In addition to lipid droplet structure, an enhanced CARS signal located in the uterus of fat-2 and fat-3 mutants was detected in contrast to little or no signal in the uterus of the wild type [arrowheads in Fig. 4(a)]. A higher percentage of worms in the fat-2 mutant contains such nondroplet-like structure in the uterus than that of the fat-3 mutant [55.2% versus 33.3%, Fig. 4(b)]. Additionally, an increase of nondroplet-like structure area was found in both fat-2 and fat-3 mutants as compared with the wild type [Fig. 4(c)]. However, the CARS intensities of nondroplet-like structure in the wild type and fat-2 mutant are stronger than in the fat-3 mutant [Fig. 4(d)]. This nondroplet-like structure with enhanced CARS signal in both fat-2 and fat-3 mutants was not reported before. In previous studies, fat-2 mutation is well known to cause a reduction in the level of PUFA,⁹ which is required for lipid transportation from intestine to oocytes by yolk lipoproteins.²⁵ Edmonds et al. found an obvious accumulation of yolk lipoprotein in the uterus of the fat-2 mutant by using vitellogenin-2::GFP fusion protein as an indicator.²⁶ This suggests that the nondroplet-like structure with enhanced CARS signal in the uterus of the fat-2 mutant may come from the accumulation of yolk lipoproteins.

Fig. 4

The nondroplet-like structure with enhanced CARS signal in the uterus of C. elegans at the adult stage. (a) CARS images show the uterus of WT, fat-2, and fat-3 mutants. Arrowheads and dashed lines indicate the nondroplet-like structure. The scale bar represents 10 μm. E, embryo; I, intestine; V, vulva. (b) Percentage of worms containing nondroplet-like structure in the uterus as indicated. Data are from three independent experiments. The numbers ($n$) of worms for each strain are as follows: WT, $n=89$; fat-2, $n=29$; fat-3, $n=36$. The quantified area (c) and averaged intensity (d) of the nondroplet-like structure in different strains as indicated. The numbers ($n$) of worms for each strain are as follows: WT, $n=14$; fat-2, $n=16$; fat-3, $n=12$. Error bars indicate SEM. **$P<0.01$; ***$P<0.001$; ns, no significance.

To test the hypothesis, we compared the pattern of CARS signal of nondroplet-like structure with the fluorescence signal of yolk lipoprotein reporter (vitellogenin-2::GFP) in the uterus of the fat-2 mutant. The yolk lipoprotein reporter was detected by TPEF microscopy [dashed line in Fig. 5(a)] as well as by CARS microscopy. In contrast, lipid droplets [arrowheads in Fig. 5(a)] were only detectable in CARS but not in a fluorescence channel. In colocalization analysis, Fig. 5(b) demonstrates high colocalization of CARS over VIT-2::GFP ($70.2\pm 2.2\%$) in the VIT-2::GFP positive regions of the fat-2 mutant [dash line in Fig. 5(a)], whereas there is almost no colocalization ($2.5\pm 2.5\%$) in the VIT-2::GFP negative regions [randomly selected and shown in squares in Fig. 5(a)]. Yolk lipoproteins are well characterized as lipid transporters from intestine to gonad and provide the essential nutrients for embryo development.²⁷ They are composed of lipid-binding proteins (vitellogenin-1 to 6) and lipid²⁸ that can be detected by CARS microscopy. In previous studies, fat-2 mutation causes the disorder in biosynthesis of PUFA,⁹ which is required for endocytosis of yolk lipoproteins (by using exogenous vitellogenin-2::GFP reporter) into oocytes via the low-density lipoprotein receptor (RME-2, receptor-mediated endocytosis-2).^25,27 In this study, we demonstrate that PUFA biosynthesis is required for endocytosis of “native” yolk lipoproteins in live animals without labeling, and the first visualize of the accumulation of native yolk lipoproteins in the uterus of the fat-2 mutant is shown by using CARS microscopy. It remains unclear whether the accumulation of yolk lipoproteins is a common phenotype in other mutants impaired in lipid metabolism.

Fig. 5

The colocalization between VIT-2::GFP fluorescence and CARS signal of the nondroplet-like structure in the uterus of fat-2 mutant at the adult stage. (a) VIT-2::GFP fluorescence image (left column), CARS image (middle column), and their merged image (right column) of the uterus in live fat-2 mutant. VIT-2::GFP positive regions are labeled by dash line, and VIT-2::GFP negative regions are randomly selected and labeled in squares. Lipid droplets in CARS image are indicated by arrowheads. Scale bar represents 10 μm. (b) The colocalization of VIT-2::GFP fluorescence over CARS signal in VIT-2::GFP positive and negative regions. The number of worms is 13. Error bars indicate SEM. ***$P<0.001$.