A Mutation in the C. elegans EXP-2 Potassium Channel That Alters Feeding Behavior

See allHide authors and affiliations

Science  24 Dec 1999:
Vol. 286, Issue 5449, pp. 2501-2504
DOI: 10.1126/science.286.5449.2501


The nematode pharynx has a potassium channel with unusual properties, which allows the muscles to repolarize quickly and with the proper delay. Here, the Caenorhabditis elegans exp-2 gene is shown to encode this channel. EXP-2 is a Kv-type (voltage-activated) potassium channel that has inward-rectifying properties resembling those of the structurally dissimilar humanether-à-go-go–related gene (HERG) channel. Null and gain-of-function mutations affect pharyngeal muscle excitability in ways that are consistent with the electrophysiological behavior of the channel, and thereby demonstrate a direct link between the kinetics of this unusual channel and behavior.

Byerly and Masuda (1) were the first to describe an unusual K+ current in the pharyngeal muscles of Ascaris lumbricoides. They called this current the negative spike current because it mediates negative-going spikes from depolarized potentials in current-clamped pharyngeal muscles. The negative spike channel has properties reminiscent of the HERG K+ channel that mediates theI Kr current in the heart. Subsequently, current-clamp studies have shown that the C. eleganspharyngeal muscle has similar repolarization kinetics to those ofAscaris (2), suggesting that C. elegans also has this unusual K+ channel.

The original, semidominant exp-2(sa26) allele was isolated in a screen for worms with defective defecation behavior (3). Animals heterozygous for the gain-of-functionexp-2 allele had hyperactive head movements, defective egg laying, and brief, shallow pharyngeal contractions. Homozygotes arrested as newly hatched larvae and eventually died, probably because they do not eat.

To examine the null phenotypes of exp-2, we isolated severalexp-2(sa26) revertant alleles (4). Each of these alleles was recessive, homozygous viable, and caused no obvious defects in defecation (5) or in head movement. Instead of brief pharyngeal muscle contractions, null mutants had long-lasting pharyngeal muscle contractions that could last several seconds (6).

Wild-type pharyngeal action potentials begin with a rapid depolarization to +20 to +35 mV (7) and have a slowly declining plateau phase that lasts around 150 to 250 ms (Fig. 1). At the end of the action potential, there was a rapid hyperpolarization that reached approximately −70 mV followed by a slow depolarization to rest (around −40 mV). Inexp-2 gain-of-function heterozygotes, depolarization was normal. However, the plateau phase was very brief, ending with a rapid hyperpolarization about 70 ms after depolarization. Furthermore, the membrane potential between action potentials was near −70 mV. Theexp-2 loss-of-function pharynxes had normal interpump membrane potentials, and their action potentials had normal depolarization and normal plateau phases for the first 150 ms (Fig. 1) (7). However, instead of a rapid repolarization at the end of this period, the potential slowly oscillated near 0 mV for up to 7 s (Fig. 1). After this, the membrane slowly hyperpolarized with little or no overshoot.

Figure 1

Action potentials from wild-type andexp-2 mutant pharyngeal muscles. (A) Wild-type, (B) gain-of-function exp-2(sa26)/+, and (C) null exp-2 (ad1426) action potentials. (D) The null mutant action potential on a contracted time scale. The bar above the initial part of the action potential indicates the portion of the action potential shown in (C).

We placed exp-2 between cosmids ZC250 and W06H8 by mapping with respect to deficiencies with known molecular break points (8). We found a gene on the cosmid F12F3 (9), which lies between ZC250 and W06H8, that appeared to encode a K+ channel. We rescued the exp-2loss-of-function pharyngeal action-potential defect with a polymerase chain reaction (PCR) product containing only the promoter and coding region of a putative K+ channel from cosmid F12F3, demonstrating that this gene is likely to be exp-2(10). We sequenced the K+ channel coding region from several exp-2 alleles (11) and found that the gain-of-function allele, sa26, was a Cys-to-Tyr mutation in the S6 segment of the K+ channel (12). Most of the loss-of-function alleles are in the highly conserved pore domain (12), and all but ad1559 appear to be null (13). Comparison of the exp-2 coding sequence with those of other K+ channel genes showed thatexp-2 is a member of the six-transmembrane, voltage-activated (Kv-type) family of K+ channels (12).

We fused a 4047–base pair length of the exp-2 promoter and most of the exp-2 coding region to green fluorescent protein (GFP). In transgenic worms, the GFP fusion protein was expressed strongly in pharyngeal muscle (Fig. 2A), in the intestinal muscles (Fig. 2B), and occasionally faintly in the egg-laying muscles (14). Each of these muscle types was affected by the gain-of-function exp-2 mutation. The reporter was also expressed in several neurons of the nerve ring ganglia (Fig. 2A). The amphid neurons are the primary chemosensory neurons of the animal, so disrupting their normal excitability would explain the chemotaxis defect seen in exp-2 loss-of-function animals (15).

Figure 2

Expression pattern of anexp-2::GFP reporter fusion. GFP fused to the COOH-terminus of exp-2 is strongly expressed in (A) the pharyngeal muscle and amphid neurons in the anterior portion of the worm and (B) phasmid neurons and the intestinal muscle in the posterior portion of the worm. Theexp-2::GFP fusions contain the exp-2promoter and coding regions encoded by an Nae I to Sal I fragment from the cosmid F12F3 fused to a Sal I to Apa I fragment from pPD95.77 (gift from A. Fire, J. Ahnn, G. Seydoux, and S. Xu) containing GFP coding sequences and the unc-54 3′ untranslated region.

We expressed in vitro–transcribed EXP-2 complementary RNA (cRNA) in Xenopus laevis oocytes (16). Oocytes were held at −80 mV, and 1-s depolarizing pulses were applied, followed by 200-ms pulses to −120 mV (Fig. 3A). We saw no outward currents, but saw large inward-directed tail currents during the 200-ms pulse to −120 mV following pulses to −30 mV or higher (Fig. 3A). The tail currents did not appear instantaneously but showed a rising phase, indicating time-dependent recovery from inactivation. The magnitude of these currents increased with increasing extracellular K+concentration, and the zero current level was close toE K, indicating that oocyte-expressed EXP-2 channels are potassium selective (17). The peak levels of the tail currents had a midpoint of activation (V 0.5) at −17 ± 2 mV (mean ± SEM,n = 5) (Fig. 3D). The voltage-dependent activation of EXP-2 channels during the test pulse occurred relatively slowly at +20 mV with an activation time constant of 54 ± 5 ms (mean ± SEM, n = 10 oocytes) (17).

Figure 3

Voltage dependence of EXP-2 and S6 mutant C480Y channels. Oocytes were subjected to a two-electrode voltage-clamp protocol in a modified frog Ringer solution containing 100 mM KCl to facilitate the measurement of inward and outward K+currents. (A) Oocytes were held at −80 mV. We applied 1-s depolarizing pulses to potentials from −60 to +60 mV in 10-mV increments, followed by a 200-ms pulse to −120 mV. During a pulse to −30 mV, EXP-2 channels become measurably activated, as seen by the inward tail current during the 200-ms pulse to −120 mV. (B) Oocytes expressing the C480Y mutant show inward K+ currents at potentials as negative as −160 mV, indicating the presence of constitutively open channels. Strong depolarization (positive to ∼−40 mV) appears to open additional channels, as seen by the increasing tail currents at −140 mV. We never saw inward currents of this magnitude, either in uninjected (data not shown) or in oocytes injected with wild-type EXP-2 cRNA (A). (C) Co-injection of wild-type and C480Y cRNA generates putative heterotetrameric channels with novel properties. Most channels are closed at a holding potential of −80 mV but begin to activate at ∼−70 mV. (D) Conductance-voltage relationship for wild-type EXP-2 and the C480Y mutant heterotetramer. The midpoint of activation and the apparent activation threshold are shifted to hyperpolarized potentials by 13 and ∼30 mV, respectively, in the C480Y mutant heterotetramer.

Oocytes expressing the Cys480 → Tyr480 (C480Y) mutant cRNA showed relatively large inward currents that increased at more negative holding potentials (Fig. 3B). Depolarizing 1-s pulses to potentials more positive than −50 mV appeared to activate additional channels because there was a small increase in the subsequent tail currents at −140 mV. Hence, it appeared that the C480Y mutation in S6 prevented the majority of channels from closing, even at potentials as negative as −160 mV.

We co-injected wild-type EXP-2 and C480Y mutant cRNA (at a 1:1 ratio) to model the channel composition in heterozygous worms that show very brief action potentials (Fig. 3C). We saw some inward currents that increased at more negative potentials (from −80 to −140 mV). These inward currents were probably caused by a small fraction of homotetrameric C480Y mutant channels. The first inward tail current appeared around −70 mV, indicating the formation of K+ channels with novel properties (presumably heterotetramers) that open at more negative potentials (V 0.5 = −30 ± 4 mV; mean ± SEM,n = 4) (Fig. 3D).

The C480Y mutation, which apparently uncouples channel opening from voltage-dependent gating in EXP-2, is located in S6 close to a position involved in C-type inactivation in Shaker (A463) (18) and Kv1.3 (A413) (19) and a position in Kv2.1 (C393) involved in stabilizing the open state (20). We used the crystal structure of the KcsA K+ channel (21) as a scaffold to generate a three-dimensional model of the EXP-2 pore, including the transmembrane segments S5 and S6, to understand how the structural change in the C480Y mutant traps the channel in the open state at more negative membrane potentials (22). The large Tyr side chain in the C480Y mutant protrudes from S6 to near S5 (Fig. 4). In contrast, the smaller side chain of Cys is buried close to S6. It has been suggested that S6 needs to rotate counterclockwise (viewed from the outside) for the KcsA channel to open (23). If EXP-2 is similarly gated, the rotation of S6 would move the bulky Tyr side chain away from the S5 helix on opening, relieving the unfavorable steric interaction. Thus, the mutation would tend to shift the equilibrium toward the open state, as observed.

Figure 4

Location of residue 480 in the three-dimensional structure of EXP-2. A top view of (A) a channel heterotetramer (from the outside) and a view straight down S6 (also from the outside) of (B) a wild-type subunit and of a (C) C480Y subunit with a CPK (Corey-Pauling-Koltun) representation of the Cys and Tyr residues at position 480. These views of the putative closed state of the channel suggest how the Tyr side chain could promote the counterclockwise rotation of S6 that may be required for channel opening.

The physiological role of EXP-2 channels in the pharynx is reminiscent of the behavior of HERG channels in the vertebrate heart; both EXP-2 and HERG channels act to repolarize the muscle after a long action potential. The channels have similar kinetics: slow voltage-dependent activation followed by fast inactivation [inactivation in EXP-2 is even faster than in HERG (17)]. However, EXP-2 and HERG do not have similar amino acid sequences. EXP-2 is a member of the Kv family of K+ channels, while HERG is a member of theether-à-go-go (eag) family of K+ channels. Thus, it is likely that these channels convergently evolved to serve similar functions.

Here, we describe a mutation that goes all the way from gene to its effect on behavior. First, we characterized the channel at the molecular level. We identified a novel channel gene, and a point mutation that alters its kinetics. The crystal structure of a K+ channel pore domain (21) and electron paramagnetic resonance spectroscopy studies of K+ channel gating movements (23) allowed us to make a plausible argument for how this sequence change leads to the observed changes in channel kinetics. Second, we characterized the biophysical properties of the wild-type and mutant channel in Xenopus oocytes. This allowed us to determine the kinetics of the channel in isolation and make predictions about its effects on muscle physiology. Finally, our electrophysiological studies in wild-type, null, and gain-of-functionexp-2 animals showed how the properties of the channel contribute to the physiology of the muscle. The exp-2 null mutants had no noticeable defect in the action potential until the point when the EXP-2 channel should open and rapidly repolarize the muscle. When the channel was mutated in a way that changed its kinetics, we saw corresponding changes in muscle physiology and behavior. Thus, exp-2 links channel sequence to channel kinetics, and channel kinetics to muscle physiology. Understanding the connection between genes and behavior in molecular detail is one of the ultimate goals of neurobiology; exp-2 demonstrates that this goal can be reached.

  • * Present address: Department of Biology, University of Utah, Salt Lake City, UT 84112–0840, USA.

  • To whom correspondence should be addressed. E-mail: wdavis{at}

  • Present address: Department of Biology, McGill University, Montreal, PQ H3A 1B1, Canada.


Stay Connected to Science

Navigate This Article