Feeding and food choice are crucial to the survival of an animal. The nematode Caenorhabditis elegans feeds on various microorganisms in nature, and is usually fed Escherichia coli in the laboratory. To elucidate the mechanisms of food/non-food discrimination in C. elegans, we examined the accumulation of various fluorescent polystyrene microspheres in the absence and presence of bacterial food. In the absence of food and on agar plates, C. elegans worms actively accumulated 0.5 and 1 μm diameter microspheres, whereas those microspheres <0.5 μm or >3 μm were rarely accumulated. Carboxylate microspheres were accumulated more than sulfate or amine microspheres. These results of accumulation in the absence of food probably well simulate uptake of or feeding on the microspheres. Presence of food bacteria even at bacteria:nematode ratios of 1:100 or 1:10 significantly reduced accumulation of 0.5 μm microspheres, and accumulation was reduced to approximately one-fourth of that observed in the absence of bacteria at a ratio of 1:1. When accumulation of microspheres was examined with the chemical sense mutants che-2, tax-2, odr-1 and odr-2, or the feeding mutant eat-1, all the mutants showed less accumulation than the wild type in the absence of food. In the presence of food, the che-2 mutant showed more accumulation than the wild type. It is possible that C. elegans discriminates food both physically, based on size, and chemically, based on taste and olfaction.
In animals, various feeding methods are seen, such as absorption and endocytosis of nutrient molecules directly through exterior body surfaces, filter feeding, fluid feeding, and capturing large prey via the mouth (Randall et al., 2002). The nematode Caenorhabditis elegans is a filter feeder, although it does not have an obvious filter (Avery and Shtonda 2003; Fang-Yen et al., 2009). In nature, C. elegans likely feeds on microorganisms (Brenner, 1974), and it is usually fed Escherichia coli bacteria in the laboratory (Sulston and Hodgkin, 1988). It takes liquid containing suspended food particles into the pharynx, traps the bacteria, ejects the liquid and transports the bacteria into the intestine through the so-called grinder at the end of the pharynx (Avery and Shtonda, 2003; Fang-Yen et al., 2009). The mechanisms involved in the selective uptake and transport processes in the wild type have been studied extensively (ibid.), and many genes have been shown to be involved in these processes, such as eat-1, eat-2, eat-4, eat-5, eat-6, eat-12, eat-18 and exp-2 (Avery, 1993; Raizen and Avery, 1994; Avery and Thomas, 1997; Davis et al., 1999; McKay et al., 2004).
Food-seeking or food-choice behavior is important for, and might be closely related to, feeding in nematodes. Food-seeking behavior in C. elegans is known to be based on chemotaxis to water-soluble or volatile chemicals that represent taste or olfaction (Ward, 1973; Dusenbery et al., 1975; Bargmann et al., 1993). Food choice has been studied using various bacteria to examine chemotaxis, food transport and pathogenicity (Andrew and Nicholas, 1976; Avery and Shtonda, 2003; Zhang et al., 2005; Shtonda and Avery, 2006). Artificial particles have also been used to study the effects of starvation and drugs on feeding (Avery and Horvitz, 1990), mechanisms of selective uptake (Avery and Shtonda, 2003; Fang-yen et al., 2009) or incorporation into cells (Pluskota et al., 2009). However, it is not clear to what degree and how a natural food and a non-food are discriminated. In this report, we used fluorescent microspheres of various sizes and chemical natures in the absence and presence of food bacteria to elucidate the mechanisms of food discrimination in the nematode C. elegans.
MATERIALS AND METHODS
Cultivation of C. elegans worms
Caenorhabditis elegans wild-type strain N2, feeding-defective mutant DA531 eat-1 (e2343), and chemotaxis mutants CB1033 che-2 (e1033), FK100 tax-2 (p671), CX2065 odr-1 (n1936) and CX2205 odr-3 (n2150) were obtained from the Caenorhabditis Genetics Center, University of Minnesota.
The nematodes were handled essentially as described by Sulston and Hodgkin (Sulston and Hodgkin, 1988). Three young adult worms were placed, using a platinum wire picker, on a nematode growth medium (NGM) agar plate in a 6 cm plastic Petri dish seeded with E. coli strain OP50, cultured at 20°C for 3–4 days, and maintained by this cycle of cultivation. OP50 was cultured in Luria–Bertani (LB) liquid medium (1% polypeptone, 0.5% Difco Yeast Extract, 1% NaCl) at 37°C.
Assay for accumulation of microspheres
As microspheres, we mainly used Fluoresbrite® Polystyrene Carboxylate Size Range Kits I and II (yellow-green fluorescent; Polysciences Inc., Warrington, PA, USA; Kit I: 0.116±0.005, 0.210±0.013, 0.516±0.011, 0.748±0.021 and 0.968±0.028 μm diameters; Kit II: 1.67±0.032, 1.83±0.061, 2.89±0.15, 4.87±0.25 and 6.60±0.60 μm diameters). In some experiments, Fluoresbrite® Carboxy BB microspheres (blue fluorescent, 0.483±0.010 μm; Polysciences Inc.) were also used; in the experiments to examine discrimination based on the surface chemical nature of the microspheres, carboxylate, amine or sulfate-modified latex (polystyrene) beads (yellow-green fluorescent, φ1.0 μm; Sigma-Aldrich, St Louis, MO, USA) were used. Apart from the original stocks, the microspheres were stored in the refrigerator as 1/10 diluted stocks in 90% ethanol and 1/100 diluted stocks in 9% ethanol.
For the assay of microsphere accumulation, a 100 μl suspension of 1.0×109 ml–1 of 0.5 μm or 1.0×108 ml–1 of 1.0 μm microspheres was mixed with 100 μl S-basal buffer (Sulston and Hodgkin, 1988), and the mixture was put on a 3 cm NGM agar plate (Sulston and Hodgkin, 1988) and spread by tilting the plate on a clean bench. The plate was left for liquid absorption for 30–60 min and stored in a 20°C incubator. For an uptake assay in the presence of food, E. coli OP50 cells were recovered from an NGM culture plate into S-basal, the density was adjusted and the cell suspension was mixed with a microsphere suspension. The density of microspheres or cells was measured with a bacterial counter and a microscope.
For the assay, the worms that had been cultured for 4 days were recovered from a 6 cm NGM plate in 1 ml S-basal, and after the worms sedimented, washing was repeated three times in total by removal of the supernatant, addition of 1 ml of new S-basal and self-sedimentation. Fifteen microliters of the washed worm suspension was put on a 6 cm NGM plate seeded with OP50, and the worms were cultured for 2 h at 20°C. The worms were then recovered as before with two washings, and 5 μl of the worm suspension was put on a 3 cm NGM plate seeded with microspheres prepared as above, and the plate was incubated for uptake of the microspheres. After incubation, three worms in the center of the plate were separated from the microspheres by washing in 50 μl of S-basal on Parafilm for 2 min. The worms were then anesthetized on an agar pad (Sulston and Hodgkin, 1988) containing 200 mmol l–1 Na azide (NaN3) and placed on a slide glass. The worms on the agar pad were examined under a Nikon ECLIPSE 80i microscope (Tokyo, Japan), and black-and-white fluorescent microphotographs of the microspheres were taken with a digital camera (Digital Sight DS-2MBWc-U2, Nikon) with the neutral density filter 4 (excitation light intensity at 25% of the maximum) as the standard condition. The fluorescence of microspheres accumulated in a worm was estimated using the image-analysis software WinROOF (Mitani Corp. Ltd, Tokyo, Japan), and was corrected by adding the fluorescence of the microspheres expelled from the worm during anesthetization. Means ± s.e.m. were derived from the fluorescence of 10 worms. In some experiments without this correction, 50 μl of 50 mmol l–1 Na azide was added to the plate to anesthetize the worms after uptake, and then 15 adult worms were placed into 200 μl of 50 mmol l–1 Na azide. Eight hundred microliters of S-basal was added, the mixture was centrifuged at 1700 g for 1 min, and 15 μl of the sedimented worms was placed inside a 1 cm2 space cut out of a piece of double-sided adhesive tape affixed to a slide glass. The fluorescence was examined as described above. The number of microspheres accumulated by a worm was estimated by dividing the total fluorescence of the worm by the mean fluorescence of a microsphere, which was directly measured for microspheres down to 0.5 μm and estimated for 0.2 and 0.1 μm microspheres by the linear relationship between fluorescence and volume for spheres 0.5 to 2 μm in size.
For the assay of accumulation by first instar (L1) larvae, worms cultured for 4–5 days were treated with an alkaline bleach to prepare nematode eggs (Sulston and Hodgkin, 1988). The eggs were incubated on a 3 cm NGM plate without bacteria, and the resultant L1 larvae were recovered, washed and incubated for uptake for 30 min at 20°C.
Conditions for accumulation of microspheres
To study the uptake of artificial particles, we initially examined accumulation of fluorescent polystyrene microspheres of 1.0 or 0.5 μm diameter by C. elegans worms in a buffer solution, as in the first study on the uptake of artificial particles (5 μm iron particles) by C. elegans (Avery and Horvitz, 1990). However, when we tested accumulation on an agar plate, far more microspheres were accumulated. In addition, the worms are usually fed on an agar plate. Therefore, subsequently we examined accumulation on agar plates instead of in a buffer solution. Fig. 1 shows typical images of a worm that took up 0.5 μm microspheres on an agar plate. Most of the microspheres were found in the intestine, and some in the pharynx. To determine the standard conditions for accumulation, we varied time and temperature for accumulation of the 0.5 μm microspheres at a density of 108 microspheres plate–1 (3 cm diameter) and in the absence of food (Fig. 2A,B). Similar results were obtained with 1.0 μm microspheres (data not shown). Based on the results, we chose 15 min and 20°C as the standard conditions. The results of accumulation at various microsphere densities under the standard conditions are shown in Fig. 2C. Although more microspheres were accumulated at a density of 109 microspheres plate–1, we chose 108 microspheres plate–1 as the standard because this was sufficient for the assay and the lower density allowed us to save on materials.
Effects of the size and chemical nature of the microspheres
We examined accumulation of microspheres of various sizes under the standard time and temperature conditions and using an equal volume of microspheres per plate (Fig. 3A). The highest fluorescence was observed with 1 μm spheres; spheres of 5 μm or <0.5 μm were accumulated much less by C. elegans. The fluorescence intensities of these microspheres are roughly proportional to their volumes for the 0.5 to 2 μm range, and so Fig. 3A simulates accumulation on a volume or mass basis. On a number basis, more 0.5 μm spheres were accumulated than 1 μm spheres (estimated to be 2.5×104vs 3.7×103 microspheres worm–1). Fig. 3B shows the results of accumulation by L1 larvae, which are much smaller than adults used in the experiment shown in Fig. 3A, and most other experiments. The larvae actively accumulated 0.5 μm spheres, but few spheres larger than 0.5 μm were accumulated. It is interesting that even the small larvae took up spheres smaller than 0.5 μm only rarely, as in the case of adults. Fig. 3C,D shows the results of experiments with adult worms that tested whether the presence of 0.5 μm spheres, which are actively accumulated, stimulates accumulation of 0.1 μm or 0.2 μm spheres, which alone are accumulated to a much lesser degree. The 0.5 μm spheres significantly stimulated accumulation of 0.2 μm spheres, but not of 0.1 μm spheres. When 1 μm spheres were used in some of the following experiments, they were used at a density of 107 microspheres plate–1, which nearly corresponds to the sphere density of 108 microspheres of 0.5 μm diameter plate–1.
Accumulation of sulfate- or amine-modified polystyrene microspheres was compared with that of the standard carboxylate-modified spheres (Fig. 4). The three kinds of spheres showed significantly different accumulation, indicating that the surface chemical nature of the spheres has a role in the mechanisms of accumulation.
Effects of food bacteria
We tested whether the presence of food bacteria affected the accumulation of the 0.5 μm spheres (Fig. 5A), and found that accumulation exhibited gradual inhibition depending on the ratio of bacterium/sphere; at a 1:1 ratio the accumulation was reduced to approximately one-fourth of that observed in the absence of bacteria. The experiments with 1.0 μm spheres showed similar results (data not shown). Presence of bacteria did not affect the accumulation of 0.2 μm spheres, but drastically reduced the accumulation of 0.1 μm spheres (Fig. 5B,C).
Examination of genes possibly involved in the discrimination of food and non-food
Many mutants abnormal in the taste for water-soluble substances or in the smell or olfaction for volatile substances have been isolated and characterized (reviewed by Bargmann and Mori, 1997). Some of these, and a feeding-defective mutant (see the Introduction) were selected and examined for accumulation of the spheres in the absence and presence of food bacteria (Fig. 6). In the absence of food, all the mutants tested showed reduced accumulation of the spheres. In the presence of food, all the mutants except eat-1 showed a significant reduction of sphere accumulation compared with those in the absence of food, but the degree of reduction was less in all of the mutants than that in the wild type. Among them, only che-2 showed increased accumulation compared with that of the wild type (P=0.0080).
Accumulation, uptake and defecation
The main objective of our study was to elucidate the mechanisms of food/non-food discrimination in the feeding process. For this purpose, we should examine uptake of or feeding on microspheres, which may be different from their actual accumulation. Accumulation of food in a worm is generally considered to be the result of uptake minus digestion and defecation. For the polystyrene microspheres, digestion does not take place. In the absence of food bacteria, the time course of accumulation shown in Fig. 2A indicates that accumulation clearly slows down after 15 min. Also, defecation is drastically suppressed in the absence of food (Thomas, 1990; Avery and Thomas, 1997), and our preliminary observation shows that defecation of the microspheres rarely takes place (data not shown). Therefore, we think that our results of microsphere accumulation during 15 min in the absence of food (Figs 1, 2, 3, 4, 6) closely represent uptake of the spheres. In the presence of food, defecation may take place and we should carefully consider the relationship between accumulation and uptake, as described later.
Mechanisms of size selection
The results shown in Fig. 3A,B suggest that the adult worms prefer food 0.5 to 1 μm in size, whereas L1 larvae cannot ingest food 1 μm in size. The difference in the results between adults and larvae is clear and must be related to the difference in their sizes (approximately 1.4 mm long and 4 nl in volume for an adult vs 0.4 mm and 0.1 nl for an L1 larva). For the food bacterium E. coli strain OP50, we obtained mean (±s.d.) sizes of 0.76±0.18 μm, 0.53±0.07 μm and 0.23±0.06 μm2 for the length, diameter and sectional area, respectively, in our laboratory (So et al., 2011) (S. So, K.M. and Y.O., unpublished results). The shape of OP50 cell is nearly spherical and its volume estimated from these data is 0.11 μm3, between that of 0.5 and 1 μm spheres (0.072 and 0.48 μm3, respectively). Active uptake of 0.5 μm and 1 μm spheres by adults may be reasonable as the size and shape of the E. coli cells are close to those of these spheres. Also, the results are consistent with the fact that L1 larvae can take up E. coli cells, the size of which is close to that of a 0.5 μm sphere.
To determine whether spheres of different sizes were discriminated even when they were present together, we tested whether uptake of the small spheres of 0.1 or 0.2 μm diameter was stimulated by the presence of 0.5 μm spheres that were actively taken up (Fig. 3C,D). For 0.1 μm spheres, the presence of 0.5 μm spheres did not result in a significant difference in the uptake, whereas it clearly stimulated uptake of 0.2 μm spheres. The results indicate that the worms have some difficulty in discriminating 0.2 μm spheres from 0.5 μm spheres when they are mixed, whereas 0.2 μm spheres are taken up less than 0.1 μm spheres when they are present alone. Uptake of few 5 μm spheres (Fig. 3A) is consistent with the results of studies by Fang-Yen et al. (Fang-Yen et al., 2009), in which virtually no 4.5 μm polystyrene spheres were taken up, and Avery and Horvitz (1990), in which less than one 5 μm iron particle per worm was taken up under similar conditions (well-fed worms in the absence of food).
The results shown in Fig. 3 are mostly novel, although dependence of uptake on the particle size of Fig. 3A is partially similar to the data reported by Fang-Yen et al. (Fang-Yen et al., 2009), which were based on the fraction of the worms carrying each of various spheres and not on quantitation of the spheres. However, they distinguished the location of the spheres within a worm and did extensive kinetic studies on uptake and transport, enabling them to reveal the mechanisms, which we have not done.
On the mechanisms of selective uptake and size selection of food, Avery and Shtonda (Avery and Shtonda, 2003) found that the contraction–relaxation cycle of the middle section of the pharynx (isthmus) is delayed relative to that of the anterior section (corpus), and proposed that this delay causes net particle transport. They also proposed that the particles (0.8 μm latex beads) in the center of the pharyngeal lumen move faster than the fluid on average. Fang-Yen et al. (Fang-Yen et al., 2009) reported that the stoma or buccal cavity at the beginning of the pharynx and the relaxation of the anterior tip of the corpus (metastomal flaps) exclude excessively large particles from entering, and that the latter also works as a valve to prevent food-sized particles from getting out. They also showed radial filtering of food bacteria and polystyrene particles, by which bacteria and particles 0.5 μm or larger are restricted to the center of the pharyngeal lumen while 0.03 and 0.1 μm particles diffuse into peripheral channels of the lumen. A constriction of 0.1–0.2 μm separating the channels from the central lumen, which was observed in electron micrographs, was proposed to function as the radial filter (Fang-Yen et al., 2009). Exclusion of most microspheres larger than 3 μm in adults and those larger than 0.5 μm in L1 larvae in the present study can be explained by the exclusion by the stoma and metastomal flaps described by Fang-Yen et al. (Fang-Yen et al., 2009). Based on these results, the diameter of the stoma of an adult seems to be between 3 and 5 μm and that of an L1 larva between 0.5 and 0.75 μm. Fang-Yen et al. (Fang-Yen et al., 2009) did not examine the behavior of 0.2 μm spheres, and we showed that 0.2 μm spheres, as well as 0.1 μm spheres, were not taken up significantly (Fig. 3A). Based on their radial filtering mechanism, this result suggests that the constriction separating the central lumen and the channels is larger than 0.2 μm in vivo. It is interesting that even very small L1 larvae do not take up 0.1 and 0.2 μm spheres. For nematodes, which mostly live in soil, excluding small soil particles may be important. If the explanation for the difference in the behavior of 0.1 and 0.2 μm spheres in the presence of 0.5 μm spheres (Fig. 3C,D) is related to this radial filtering mechanism, then 0.5 μm particles in the central lumen may form a kind of barrier, preventing the 0.2 μm particles from diffusing freely into the peripheral channels, but preventing diffusion only slightly for the smaller 0.1 μm spheres.
The results shown in Fig. 4 suggest that chemical modification on the surface of the spheres affects their uptake. Reduction in the accumulation of 0.5 μm spheres by the presence of E. coli cells (Fig. 5A) indicates that E. coli cells are moderately selected for accumulation against the carboxylate polystyrene spheres of a similar size. This result suggests that chemical discrimination occurs based on the nature of the surfaces in the uptake. Alternatively, chemical stimulation in the uptake may not have actually occurred, instead, stimulation of defecation by the presence of food may have led to the present results. We think that the effects of E. coli cells on the accumulation of 0.2 and 0.1 μm spheres (Fig. 5B,C) favor the former possibility because, if E. coli cells significantly stimulate defecation of spheres, the accumulation of both 0.1 and 0.2 μm spheres is likely to be affected similarly. Also, the results shown in Fig. 5B,C differ from the effects of 0.5 μm spheres shown in Fig. 3C,D, and must be related to a difference in the chemical nature of the bacteria and the spheres. An explanation for this may be that 0.1 μm spheres are easier to be excluded from uptake in the presence of E. coli cells based both on physical and chemical selection, whereas uptake of 0.2 μm spheres is stimulated physically by the addition of 0.5 μm spheres (Fig. 3D) and is compensated by a possible reduction based on chemical discrimination.
Genes possibly involved in the discrimination of food
All the mutations tested reduced accumulation of the spheres in the absence of food (Fig. 6), suggesting that the corresponding genes function in the uptake of the spheres. In the presence of food, accumulation of the spheres was further reduced in all the mutants except eat-1, but the rate of reduction was significantly lower than that in the wild-type N2. Among the mutants tested, only the che-2 mutant showed a significant difference (increase) in the uptake of the spheres in the presence of food from that of the wild type, suggesting a unique and important role of the che-2 gene in the negative discrimination of the spheres from food. Because the common function of the che-2, tax-2, odr-1 and odr-3 genes is the chemical sense (taste or olfaction) and because no abnormal phenotypes in pumping rate or defecation are reported (WormBase, www.wormbase.org), we suggest that they function to distinguish chemically between the spheres and the food bacteria in the uptake. The che-2 mutant is defective in chemotaxis to water-soluble substances (taste) and in the structure of sensory cilia, and the gene is known to encode a WD40 repeat protein (Fujiwara et al., 1999). The tax-2 mutant is defective in chemotaxis and thermotaxis, and the gene encodes a β-subunit of the cyclic nucleotide-gated ion channel expressed in sensory neurons (Coburn et al., 1996; Komatsu et al., 1996). The odr-1 and odr-3 mutants are defective in chemotaxis to volatile substances (olfaction), and the genes encode a guanylyl cyclase and an α-subunit of a tripartite G-protein complex, respectively (Bargmann et al., 1993; L’Etoile and Bargmann, 2000; Roayaie et al., 1998). The eat-1 mutant showed the lowest uptake of the spheres among the mutants tested in the absence of food (Fig. 6), and the uptake changed little in the presence of food, suggesting that the eat-1 gene has an important role both in the uptake and discrimination of food. The eat-1 mutant is known to be severely defective in food uptake and to show irregular pharyngeal pumping and sluggish movement (Avery, 1993). The eat-1 gene encodes an Alp-Enigma family protein (McKeown and Beckerle, 2001).
In conclusion, our results form a basis to elucidate the mechanisms of discrimination between food and non-food in the nematode. Further studies will be needed to understand more detailed mechanisms in the future.
We thank Y. Kaku for his participation in the initial stage of this study, and the Caenorhabditis Genetics Center for the C. elegans strains.
This research was funded by a grant from Sojo University.