Starvation-induced collective behavior in C. elegans

Abstract

We describe a new type of collective behavior in C. elegans nematodes, aggregation of starved L1 larvae. Shortly after hatching in the absence of food, L1 larvae arrest their development and disperse in search for food. In contrast, after two or more days without food, the worms change their behavior--they start to aggregate. The aggregation requires a small amount of ethanol or acetate in the environment. In the case of ethanol, it has to be metabolized, which requires functional alcohol dehydrogenase sodh-1. The resulting acetate is used in de novo fatty acid synthesis, and some of the newly made fatty acids are then derivatized to glycerophosphoethanolamides and released into the surrounding medium. We examined several other Caenorhabditis species and found an apparent correlation between propensity of starved L1s to aggregate and density dependence of their survival in starvation. Aggregation locally concentrates worms and may help the larvae to survive long starvation. This work demonstrates how presence of ethanol or acetate, relatively abundant small molecules in the environment, induces collective behavior in C. elegans associated with different survival strategies.

Figure 1. Time course of C. elegans L1 aggregation.

One million 3 day starved wild-type L1 larvae were thoroughly washed, dispersed in 100 μl of M9 buffer, and pipetted into the center of a clean 6 cm NGMSR plate at time zero (dark circular area in the lower half of the images).

Figure 2. Even-chain alcohol is required for L1 aggregation.

Images obtained 23 h after pipetting 3 day starved L1s to agar plates supplemented with 17 mM of various alcohols (10⁶ L1 larvae per plate). The leftmost image is a control without alcohol.

Figure 3. Addition of ethanol quickly induces L1 aggregation.

5 μl of ethanol was added to one side of a 6 cm plate containing L1s (red asterisk).

Figure 4. Short-chain fatty acids support L1 aggregation.

L1s aggregate in the presence of 17 mM potassium acetate, butyric, or hexanoic acids, but not glucose.

Figure 5. sodh-1 is required for aggregation in the presence of ethanol.

sodh-1(ok2799) L1s fail to aggregate on ethanol-containing plates but aggregate normally in the presence of 17 mM potassium acetate. The other sodh-1 alleles, bet20 and tm2829, showed the same result.

Figure 6. Glycerophosphoethanolamides are produced from ethanol.

A general formula of ethanol/acetate specific metabolites detected in L1 medium. Ethanol-derived carbons are shown in red.

Figure 7. sodh-1 L1s participate in wild-type aggregates.

sodh-1(tm2829) L1s expressing tdtomato fluorescent protein do not aggregate (a) but a 1:1 mixture with wild-type (N2) worms does (b,c). (d) A fluorescence image of the same area as shown in (c) illustrates that sodh-1 larvae are mostly in the aggregates. We obtained similar results with other alleles of sodh-1 and with different ratios of sodh-1:N2, varying from 1:4 to 3:1.

Figure 8. C. briggsae L1s do not aggregate.

(a) Starved C. briggsae AF16 L1 worms show no sign of aggregation 15.5 h after being put on a NGMSR plate. (b) A bright field image of a mixture of C. elegans L1s expressing GFP (PD4792) and C. briggsae L1s expressing dsRed (JU1018). Fluorescence images (c, d) allow the two species to be distinguished and demonstrate that C. briggsae worms do not participate in the aggregates. (e) Overlay of two fluorescence channels. Images b-e correspond to the same area.