A Genetic Screen for Dihydropyridine (DHP)-Resistant Worms Reveals New Residues Required for DHP-Blockage of Mammalian Calcium Channels

Dihydropyridines (DHPs) are L-type calcium channel (Cav1) blockers prescribed to treat several diseases including hypertension. Cav1 channels normally exist in three states: a resting closed state, an open state that is triggered by membrane depolarization, followed by a non-conducting inactivated state that is triggered by the influx of calcium ions, and a rapid change in voltage. DHP binding is thought to alter the conformation of the channel, possibly by engaging a mechanism similar to voltage dependent inactivation, and locking a calcium ion in the pore, thereby blocking channel conductance. As a Cav1 channel crystal structure is lacking, the current model of DHP action has largely been achieved by investigating the role of candidate Cav1 residues in mediating DHP-sensitivity. To better understand DHP-block and identify additional Cav1 residues important for DHP-sensitivity, we screened 440,000 randomly mutated Caenorhabditis elegans genomes for worms resistant to DHP-induced growth defects. We identified 30 missense mutations in the worm Cav1 pore-forming (α1) subunit, including eleven in conserved residues known to be necessary for DHP-binding. The remaining polymorphisms are in eight conserved residues not previously associated with DHP-sensitivity. Intriguingly, all of the worm mutants that we analyzed phenotypically exhibited increased channel activity. We also created orthologous mutations in the rat α1C subunit and examined the DHP-block of current through the mutant channels in culture. Six of the seven mutant channels examined either decreased the DHP-sensitivity of the channel and/or exhibited significant residual current at DHP concentrations sufficient to block wild-type channels. Our results further support the idea that DHP-block is intimately associated with voltage dependent inactivation and underscores the utility of C. elegans as a screening tool to identify residues important for DHP interaction with mammalian Cav1 channels.


Introduction
Calcium influx into the cell can make both an immediate and a long-term impact on many cellular processes, including membrane excitability, neurotransmitter release, muscle contraction, transcription, and proliferation [1]. It is therefore not surprising that calcium current through calcium-permeable channels is tightly regulated. Calcium channels of the Ca v family, for example, regulate current through voltage dependent mechanisms and exist in three major biophysical states: a resting (non-conducting) state, an activated (conducting) state that is triggered by a rapid change of voltage, and an inactivated (non-conducting) state that follows the conducting state but is refractory to activation. Ca v channels are composed of multiple subunits that regulate calcium current through the pore-forming and voltage-sensing a 1 subunit, which has four repeating domains (called I-IV), each with six transmembrane segments designated S1-S6 [2,3] (Figure 1). The most widely-expressed Ca v channels are the L-type, otherwise known as Ca v 1 channels. In humans, alterations in Ca v 1 abundance or function result in a variety of diseases, including hypertension [4], Timothy Syndrome [5], and congenital stationary night blindness [6]. A better understanding of Ca v 1 function and how this function can be modulated by drug-like molecules may therefore lead to better treatments of Ca v 1-related diseases.
A distinguishing feature of Ca v 1 channels is their sensitivity to 1,4-dihydropyridines (DHPs), which bind to the a 1 subunit and are commonly prescribed to reduce hypertension in humans [7]. Exactly how DHPs physically interact with the a 1 subunit is incompletely understood, in large part because the crystal structure of any Ca v 1 a 1 subunit remains elusive. However, Ca v 1 residues that facilitate DHP interaction have been identified through the analyses of candidate domains and residues. For example, swapping domains IIIS5, IIIS6, and IVS5-IVS6 from an L-type to a non-L-type channel is sufficient to confer DHPsensitivity to otherwise insensitive Ca v channels [8,9,10]. Individual Ca v 1 residues present in DHP-sensitive, but not insensitive channels, have also been extensively investigated to determine if they are necessary and sufficient for DHP-sensitivity [11,12,13,14]. In this way, four residues in domain IS6, two residues in IIIS5, eight residues in IIIS6, and six residues in IVS6 were shown to be important for DHP binding to the a 1 subunit of the Ca v 1 channel in culture [11,14,15,16,17,18] (summarized in Table 1).
DHPs are thought to block calcium current by inducing allosteric structural changes in the channel that engages the channel in the inactivated state [18,19] and promotes high affinity interaction with a single calcium ion within the pore [20,21,22], thereby preventing further ion flow. Calcium ions are bound by the channel through four negatively charged glutamate residues (called the calcium selectivity filter), which reside within four membrane-embedded intersegmental loops between S5 and S6 (called SS1-SS2) of each domain (Table 1). Together, the four SS1-SS2 domains form the extracellular-facing 'outer pore' of the channel (see Figure 1). The relationship between the DHP and the selectivity filter is bidirectional: DHP-binding facilitates a high affinity interaction between the selectivity filter and a calcium ion, and the coupling of the selectivity filter to calcium promotes a high-affinity interaction between the DHP and the channel [21,23,24]. A candidate analysis of residues surrounding the glutamate selectivity filter revealed several residues that are also important for high-affinity DHP interaction with the channel in culture [21,24,25,26] (Table 1).
Clearly, the candidate approach has been fruitful in identifying residues important for DHP antagonism of Ca v 1 channels. However, as the exact molecular mechanism of DHP block remains incompletely understood, the search for additional residues involved in DHP-sensitivity might add to our understanding of DHP-Ca v 1 interaction. We therefore screened random mutants of the tiny nematode worm C. elegans for those animals that are resistant to the effects of a novel DHP analog called nemadipine in the hopes of identifying new Ca v 1 residues important for DHP-sensitivity.
We recently discovered nemadipine through a screen for new biologically active small molecules [27,28]. Through a variety of genetic approaches, we demonstrated that the major target of nemadipine is EGL-19 [27], which is the sole Ca v 1 channel a 1 subunit encoded by the C. elegans genome [29]. We also found that nemadipine can antagonize vertebrate Ca v 1 channels with a similar potency as an FDA-approved DHP in chick ciliary neurons [27]. However, none of the FDA-approved Ca v 1 antagonists examined, including nifedipine, nicardipine, felodipine, and nimodipine, elicit robust phenotypes in whole worms. Worm insensitivity to these DHPs is not likely due to Ca v 1 divergence since nifedipine can antagonize EGL-19 in dissected worms [29]. Instead, we discovered that most FDA-approved DHPs fail to accumulate in whole worms [27], consistent with the idea that C. elegans has extensive xenobiotic defenses. Thus, nemadipine is a unique reagent to genetically probe interactions between DHPs and Ca v 1 channels in vivo.
Here, we present our genetic screen of 440,000 mutant C. elegans genomes for animals that are resistant to nemadipine. Strikingly, 30 of the 35 mutants that we characterized have missense polymorphisms in the egl-19 gene. Only one of these mutations is in a non-conserved residue. Eleven of the mutations correspond to residues of the Ca v 1 channel that are known to be required for DHP interaction with mammalian channels. The remaining mutations are in nine residues not previously associated with DHP interaction, eight of which are conserved. Behavioral and developmental analyses of 12 of our worm mutants revealed that all exhibit phenotypes consistent with increased channel activity. For seven of these worm mutations, we created orthologous mutations in the rat brain Ca v 1.2 channel and examined the ability of DHPs to block ion current through these mutant channels using whole-cell voltage clamp analysis in culture. Two of the mutant channels have decreased sensitivity to the DHP block, and six disrupt the ability of the DHP to block ion current completely. Thus, screening for worm mutants that are resistant to nemadipine is a viable and novel approach to identify mammalian Ca v 1 residues that are important for DHP sensitivity.

A Forward Genetic Screen for DHP-Resistant C. elegans Mutants Yields 30 Polymorphisms in EGL-19
Previous work demonstrated the utility of using C. elegans calcium channels as a model to better understand mammalian channel function [30]. We therefore performed a forward genetic screen for C. elegans mutants that are resistant to nemadipineinduced growth retardation to identify residues in the Ca v 1 calcium channel required for DHP sensitivity. A total of 440,000 haploid mutagenized wild type genomes were screened and 55 candidate mutants were isolated. Homozygous lines were established for 35 of these mutants and polymorphisms were identified in egl-19 for 30 ( Figure 1 and Table 2). Four of the remaining mutants are genetically linked to egl-19 on chromosome IV, but no polymorphism in egl-19 could be found (data not shown). The remaining mutant is not linked to egl-19 and is not discussed further here. Upon out-crossing select mutants, we found that all of them exhibited increased resistance to nemadipine in two separate dose-response assays (Table 2, Figure S1), confirming that they are indeed DHP-resistant mutants.
The DHP binding site on the Ca v 1 channel pore-forming a 1csubunit of vertebrate channels is defined by at least nine residues that are both necessary and sufficient to confer DHP-sensitivity to Author Summary L-type calcium channels are important drug targets because they regulate many physiological processes throughout the body. For example, L-type calcium channels regulate cardiac myocytes and vascular smooth muscle contraction. Antagonists are therefore commonly used to lower blood pressure and treat other related ailments. Despite their medical importance, the mechanism by which L-type antagonists inactivate calcium channels is not fully understood, due in large part to the lack of a channel crystal structure. Here, we present the first large-scale genetic screen for L-type calcium channel residues that are important for sensitivity to a new drug analog that we discovered called nemadipine. We performed the screen using nematodes, and then recreated similar mutations in a mammalian channel to investigate how the mutant residues alter interactions with the antagonists using electrophysiological techniques. Together, our analyses revealed eight new L-type calcium channel residues that are important for DHPsensitivity and highlight the utility of using a simple animal model system for understanding how drugs interact with their targets.
a non-L-type calcium channel ( Figure 1, Table 1). Two of these residues reside in helix S5 of domain III (called IIIS5), three reside in IIIS6 and four in IVS6 [12,13,14]. Other studies have shown the involvement of additional residues required for DHP interactions (Table 1) [15,18]. From our screen for nemadipineresistant mutants, we found ten mutants with polymorphisms within the IIIS6 domain of EGL-19. Six of these mutate a single EGL-19 residue, M1056 (rat a 1c residue M1161), critical for high- Ca v 1 channel a 1 subunit EGL-19 with residues of interest shown. Residues known to be involved in DHP interaction are in red. Blue circles are the previously characterized hypermorphic residues in EGL-19. Yellow circles are the residues that form the calcium ion filter. Black circles are residues involved in voltage dependent slow inactivation. Green circles are the EGL-19 polymorphisms isolated in our forward genetic screen and labeled with the corresponding residue number (worm residue/rat residue) and a star if analyzed further through phenotypic and physiological analyses. See Tables 1 and 2 for more information about DHP-sensitive residues. (B) Sequence alignment between EGL-19 (GenBank accession number: NM_001027908) and the orthologous rat brain Ca v 1.2 a 1C subunit (GenBank accession number: M67515). The predicted positions of the transmembrane segments (IS1-IVS6), channel pore (SS1-SS2) and the EF hand are shown and are based on the EGL-19 sequence [29]. Identical residues between the two sequences are highlighted in black. An open box beneath the amino acid shows the site of the corresponding coloured circle from the topology diagram in (A). A black box below the amino acid indicates the sites of the mutations we isolated in EGL-19. The alignment is shown only slightly past the EF hand as sequence conservation for the C-terminal tail is low after this point. doi:10.1371/journal.pgen.1000067.g001 affinity DHP binding to the vertebrate channel [11,14]. We found a seventh mutation in EGL-19 residue M1055 (rat a 1c residue M1160) that is also necessary for DHP binding to the vertebrate channel [11]. An eighth polymorphism was detected in the EGL-19 residue A1041 (rat a 1c residue S1146). Previous work demonstrated that a rat a 1c S1146A mutation does not disrupt DHP binding [25], which is perhaps not surprising given that the alanine is conserved in some Ca v 1 channels [25]. However, no one has yet tested if other residues in position 1146, such as the valine substitution found in our mutant, would disrupt interactions with DHPs. Finally, two polymorphisms were identified in the EGL-19 residue V1060 (rat a 1c residue V1165) within IIIS6 that has yet to be tested for its role in DHP block. Our finding that seven of the ten mutations that we identified within the IIIS6 region are required for DHP-interaction in mammalian Ca v 1.2 channels validates our genetic approach for identifying residues required for DHP-sensitivity, and suggests that the new residues we uncovered are also important determinants of DHP sensitivity. High-affinity DHP binding is coupled to the interaction between the selectivity filter in the outer pore of the a 1 subunit and calcium ions (see introduction). Mutating the glutamate residues and several of the surrounding residues of the selectivity filter not only dramatically reduces the binding of calcium ions, but disrupts DHPs calcium-dependent affinity for the channel .50 ..50 .50 ..50 Residues numbered as in Table 1 [21]. For example, the S1115 residue within SS1-SS2 of the rat a 1c channel that precedes the key glutamate in domain III has been shown to be necessary for DHP-sensitivity ( Figure 1, Table 1) [25,26]. We found that two of the nemadipine-resistant mutants are polymorphic within the EGL-19 residue S1010 that is equivalent to the rat a 1c residue S1115 ( Figure 1; Table 2). We also isolated a mutant with a polymorphism in a similar position in domain II (EGL-19 residue I652; rat a 1c residue I702). However, this mutant had a second polymorphism in V679 (rat residue V729), confounding the attribution of any mutant property to either polymorphism. Nevertheless, these results suggest that disrupting residues within the outer pore can reduce DHPsensitivity in vivo. DHP antagonists have high affinity for Ca v 1 channels in the inactive state [31]. After voltage dependent opening of the channel, the channel is coordinately inactivated by both calcium-dependent and voltage dependent mechanisms. Previous work has shown that voltage dependent slow inactivation of Ca v 1.2 channels is mediated by hydrophobic residues in the cytoplasmic end of the S6 transmembrane helices of domains I-IV [32,33,34]. Together, these residues are thought to form a hydrophobic voltage-sensitive annulus, the disruption of which leads to prolonged activation of the channel. We found 8 nemadipine-resistant mutants with a polymorphism in EGL-19 A702 within IIS6, which is orthologous to the residue in vertebrates (rat a 1c residue A752) that is most important in slow inactivation [32]. We also found two mutants with a polymorphism in the hydrophobic residue V1060 in the IIIS3 domain of EGL-19, which precedes an orthologous residue that contributes to slow inactivation by three positions (rat a 1c residue V1168) [33] ( Figure 1, Table 1). We also found three mutants with a polymorphism in the hydrophobic residue G365 (rat a 1c residue G402) of domain IS6 that is also required for voltage dependent slow inactivation [5,35]. The same G365R mutation that we isolated was found independently by Michael Hengartner and shown to prolong the activation phase of the mutant EGL-19 channel by Raymond Lee and colleagues in vivo [29]. Importantly, the isolation of at least eleven mutants within the voltage-sensitive annulus suggests that DHP-blockage is linked to voltage dependent slow inactivation in vivo.
Finally, we identified seven other mutants that have polymorphisms in six residues within regions of the Ca v 1 channel that have little functional annotation (A236, A615, V679, T1066, E1079, and G1594) ( Figure 1, Table 2). Of note, two of these mutant residues are in analogous positions in IS5 and IIS5 (A236 and A615, respectively). Five of the six mutant residues are in highly conserved regions of the channel. The sixth mutant residue (G1594D) is in the carboxy-terminal tail, whose sequence is evolutionarily divergent (Figure 1). We have further investigated three of these mutants through both phenotypic and electrophysiological analysis as described below.

DHP-Resistant Channels Have Increased Activity in vivo
In contrast to traditional electrophysiological approaches that rely on transgenics to investigate DHP interactions, our genomic mutations provided us with a unique opportunity to investigate the consequences of these polymorphisms on channel activity within the context of a whole animal. We reasoned that phenotypic analysis could be used to approximate the level of Ca v 1 mutant channel activity in vivo because the level of Ca v 1 EGL-19 channel activity has dramatic consequences on the development and behavior of the animals [27,29]. For example, reduction-offunction mutations or hypomorphs, such as egl-19(n582), are long and flaccid. By contrast, mutant channels that have a prolonged activation phase in electrophysiological studies, such as the egl-19(n2368) hypermorphic mutant, are myotonic and consequently shorter than wild type (Figures 2 and 3) [29]. The defects in body size and rigidity in egl-19 mutants are likely due to altered EGL-19 function in the 95 striated muscles that line the body wall of the worm. The more active the channel, the more pronounced the body size phenotypes become [27,29]. Thus, if the nemadipineresistant mutants have unaltered behaviors compared to the wildtype, we might infer that these residues might directly alter DHP binding without altering channel activity. Alternatively, if the nemadipine-resistant mutants behaved more like hypermorphs, then the mutant residues might prolong the activation phase of the channel and confer DHP-resistance indirectly. Hypomorphic mutations were neither expected nor observed, since even weak egl-19 reduction-of-function mutations dramatically decrease the viability of animals on nemadipine and would not be recovered in our screen [27].
We began our phenotypic analysis by examining the body length of twelve of our mutant strains (Table 2, Figure 3). We chose these 12 mutants because some represent residues that are known to be required for DHP interaction in other systems (worm alleles tr92, tr69, tr89, and tr71), while the remaining ones represent potentially new residues not previously implicated in DHPsensitivity. We measured the length of 20 young adults for each strain and compared their lengths to wild type worms and previously isolated egl-19 hypermorphs and hypomorphs. None of our twelve mutants were as short as either the weak (ad695) or the strong hypermorph (n2368) (Figure 3). However, two of our mutants were significantly shorter than wild type animals (p,0.001), suggesting that they might be hypermorphic.
Next, we examined behaviors that have a greater dynamic range between wild type animals and previously characterized egl-19 mutants, as these may be more sensitive to changes in EGL-19 activity. For example, egl-19 hypomorphs accumulate two-fold or more embryos than the wild-type. Conversely, egl-19 hypermorphs retain approximately two-fold fewer embryos than the wild-type control (Figures 2 and 3). egl-19 plays important roles in several cells of the egg-laying circuit, including the eight vulval muscle cells that directly control vulva opening during embryo deposition [27,29,36]. Two thirds of our nemadipine-resistant mutants behaved similar to the egl-19 hypermorphs, retaining significantly fewer embryos than the wild-type (p,0.001), consistent with the idea that these mutants are also hypermorphic ( Figure 3). Although all of our 12 egl-19 alleles that we examined exhibit the same trends in the length and egg-retention assays, it is unclear to us why tr81 and tr86 behave slightly differently. It may be that background mutations have a mild influence on these two behaviors in these two mutants. As expected, we found a positive correlation (0.64) between the strength of the egg-laying phenotype and the degree of resistance to nemadipine (Table S1).
Finally, we examined how our mutants affect the contraction of a specialized set of muscles in the male tail that regulate spicule protrusion. The spicules are a pair of sclerotized cuticle structures that are housed in the male tail and used during the transfer of sperm to hermaphrodites [37] (Figure 2). The protrusion of each spicule is controlled by a pair of protraction and retraction muscles. egl-19 is a key regulator of spicule-associated muscles and functions autonomously to regulate their contraction [38]. Nearly all males of a previously characterized egl-19 hypermorph have protruded spicules, called a Prc phenotype, while egl-19 hypomorphic males cannot protrude their spicules (Figure 2 and 3) [29,38]. We found that all of our nemadipine-resistant mutants were obviously Prc (Figure 3). The Prc phenotype of all egl-19 alleles examined was suppressed by nemadipine ( Figure 3).
Nemadipine similarly suppressed the shortened length and constitutive egg-laying phenotypes of our egl-19 mutants (Figure 3). This demonstrates that the observed spicule protrusion, constitutive egg-laying and shortened length phenotypes are all mediated by increased egl-19 activity. It is noteworthy that the egl-19 mutations that correspond to the rat a 1c residues M1160, M1161, and V1165, which are necessary for high-affinity DHP binding [12,13,14], also show increased EGL-19 channel activity. Together, our phenotypic analyses suggest that our nemadipineresistant mutants have increased channel activity in vivo.

EGL-19 Polymorphic Residues Are also Required for Complete Current Block of Mammalian Ca v 1.2 Channels by DHPs
We tested if the mutant residues of EGL-19, which confers DHP-resistance and increased channel activity in worms, also alter DHP-sensitivity of a mammalian Ca v 1.2 channel. We created orthologous mutations in the a 1C subunit of the rat Ca v 1.2 channel and transiently expressed these in tsA201 cells along with auxiliary Ca v 1.2 channel components b 2a and a 2 -d [39,40]. Ion currents through both wild-type and mutant channels were then measured in the presence of increasing concentrations of nifedipine, a popular DHP, using whole-cell voltage-clamp electrophysiological techniques (see methods for details). In this way, we could determine if the orthologous rat mutations affect the ability of DHPs to block Ca v 1.2 channel-mediated ionic current.
We first examined polymorphisms in domain IIIS6 residues previously shown to be required for DHP binding. As a positive control, we created a rat a 1C M1161A mutation that is known to disrupt the high-affinity interaction between DHPs and the Ca v 1.2 channel [11,14]. The M1161A mutation increased the IC 50 of nifedipine block ,100-fold compared to the wild type channel, confirming the importance of this residue in DHP sensitivity ( Figure 4). The nearby V1165L mutation also increases the IC 50 of nifedipine block by about 30-fold, consistent with its role in highaffinity DHP interactions [18]. Curiously, a leucine substitution in residue M1160, which is also required for DHP interaction [11], did not increase the IC 50 of nifedipine block (Figure 4). T1171I, which is cytoplasmic and follows the IIIS6 transmembrane domain, also did not exhibit a shift in the IC 50 . However, both M1160L and T1171I exhibited large residual currents at 100 mM nifedipine, which is sufficient to completely block the ion current through wild type channels (Figure 4). The validity of the residual current was further supported by a comparison to the complete block of current through the respective wild-type and mutant channels by 10 mM cadmium, which is a non-selective pore blocker commonly used to block all current through Ca v 1 channels [41]. Next, we examined two mutant residues positioned analogously in IS5 (A283V), and IIS5 (A/S666V). Although no significant shift was observed in the IC 50 response to nifedipine for either channel, the S666V mutant left a significant residual current at saturating concentrations of nifedipine. It is currently unclear why the A283V polymorphic channel behaves similar to the wild-type channel in our assay. Finally, we examined a pair of mutations in a single residue in IIS6 (A752T/V) that contributes to the hydrophobic annulus required for voltage dependent slow inactivation [32]. Consistent with these previous observations, we found that both A752T and A752V yield a significant residual current of about 6-8%, likely reflecting the importance of this site for slow inactivation. Surprisingly, we also found that the IC 50 of nifedipine block for A752V, but not A752T, increases about fivefold ( Figure 4). As expected, we found positive correlations between the IC50 values of the mutant rat channels on nifedipine and the degree of resistance by the respective worm mutants on nemadipine (0.83), and between the amount of residual current of the mutant rat channels and the degree of resistance of the worm mutants (0.63) (Table S1). In total, two of the seven mutant channels exhibit decreased DHP-sensitivity as indicated by an increased IC 50 for the DHPblock, and six of the seven mutant channels resist a completely blocked state. These electrophysiology results therefore demonstrate that the identification of Ca v 1.2 residues that are required for DHP-sensitivity in the worm can reveal residues that are also important for mammalian DHP-sensitivity.

Discussion
Here, we set out to identify Ca v 1 residues that are important for DHP sensitivity in vivo through a forward genetic screen. Our approach was made possible through our discovery of nemadipine, which is the only DHP known to robustly antagonize Ca v 1 channels within the context of undissected C. elegans worms [27]. By screening for mutants that are resistant to the effects of nemadipine, we identified 30 independent Ca v 1 polymorphisms that reduce DHP-sensitivity. Upon further investigating 12 of these, we found that all likely increase channel activity in whole worms. Further investigation of how these mutant residues alter the response of a mammalian ortholog to DHP-treatment not only led us to new residues required for DHP sensitivity, but provides in vivo support for a previously established idea that the DHP block likely requires voltage dependent inactivation of the channel (see below).
Evidence suggests that at least five of the 14 Ca v 1 polymorphic residues that we identified are required for physical interactions with DHPs. Four of the residues we identified (corresponding to rat a 1c residues S1115, M1160, M1161, and V1165) were previously shown to decrease DHP affinity for the channel when mutated [14,18,25]. Additionally, two mutant residues (corresponding to rat a 1c residues A752V and V1165L) showed a significant increase in the IC 50 response to nifedipine in our electrophysiology assay. The observed shift in the IC 50 is consistent with a decrease in the affinity of the DHP for the mutant channels.
Previous biochemical analyses identified residues within IS6 [15], IIIS5, IIIS6, and IVS6 [11,16] that are important for DHP binding. However, our genetic screen did not reveal polymorphisms in many of these residues. One reason for this is that our screen was not done to saturation; only six of the 14 residues identified are represented by multiple alleles. It is therefore likely that several residues required for DHP sensitivity in vivo remain unidentified, which is proven by the absence of the previously identified A906V nemadipine-resistant mutation [27,29] from our screen. Second, many of the residues not identified in the genetic screen play only ancillary or weak roles in DHP binding, including those of IS6 [15] and IVS6 [11] and may not confer DHPresistance by themselves in vivo. Third, polymorphisms in each of several IVS6 residues involved in DHP-binding significantly The nifedipine concentrations ranged 0, 0.1, 1, 10, 30, 100 mM. The current was elicited by an 80 ms +10 mV depolarizing pulse from a holding potential of 2100 mV. In all traces, the baseline was determined by block by 10 mM cadmium (trace with ,0 pA current). The nifedipine effect was normalized to the unblocked current (I max ) obtained in the absence of the drug. The concentration-effect curves were fit to a modified Hill equation taking into account the possibility of incomplete current block, as described in the methods. The mean concentration-effect curve for the M1161A mutant, known to be important for DHP block [11,14], serves as a positive control. Summaries of the IC 50 values (G), and the residual unblocked current measured at 100 mM nifedipine (H), obtained from the same data presented in Figure B, D, and F. The fraction of current remaining at 100 mM nifedipine (F unblock (I 100 /I max )) was used to estimate the completeness of drug block. All the data are presented as mean6s.e.m.; numbers in the parentheses in G reflects the number of independent experiments. * indicates statistical significance relative to the wildtype (p,0.05, one-way ANOVA, post-hoc Holm-Sidak). doi:10.1371/journal.pgen.1000067.g004 disrupt ion flux through the channel [11] and these types of loss-offunction mutations would not be recovered in our forward genetic screen. Fourth, nemadipine may interact with EGL-19 in slightly different ways compared to how canonical DHPs interact with mammalian channels. If true, this difference might prevent the recovery of mutations of some residues required for high affinity DHP interactions with mammalian channels in our worm screen. However, it remains possible that many of the Ca v 1.2 residues shown to be required for high-affinity DHP interactions in culture play little or no role in DHP interactions in vivo. Regardless, our finding that polymorphisms in S1115, M1160, M1161 and V1165 do confer DHP-resistance in vivo adds further weight to existing biochemical and electrophysiological evidence that these residues are critical for DHP interactions.
Several lines of evidence from our work support the previously established idea that DHP-sensitivity of the channel is intimately associated with voltage dependent inactivation [18,19]. First, our behavioral analyses of mutant worms reveals that polymorphisms within or near the predicted hydrophobic annulus (corresponding to rat a 1C polymorphisms G402R, A752T, A752V, V1165L, and T1171I), which are normally required to block the channel through voltage dependent slow inactivation, decrease DHPsensitivity and increase channel activity in vivo. Second, electrophysiological analysis on four of these polymorphisms (A752T, A752V, V1165L, and T1171I) reveal significant residual current in the presence of DHP concentrations that is sufficient to block the wild-type channel in culture and two of these polymorphisms (A752T and T1171I) do not show any change in the IC 50 for the DHP-block. This suggests that these residues are required to engage the DHP-block without being involved in DHP-sensitivity. Because the A752T mutation is known to disrupt voltage-gated slow inactivation [32] (and T1171 is very close to V1168 and I1169 that are also required for voltage-gated slow inactivation [33]), our observations provide additional support for the previously established idea that the DHP block may be mediated through the same mechanism employed by voltage-gated slow inactivation [18,19]. Together, our work demonstrates that a forward genetic screen for worms resistant to DHPs is a viable approach for the discovery of residues that are important for mammalian DHP-sensitivity.

Screening, Mapping, and Phenotypic Analysis of Nemadipine-Resistant Mutants
Candidate mutants were isolated and mapped through their resistance to nemadipine-induced population growth defect as described previously [27,28]. For this work, we used only the nemadipine-A analog, simply referred to as nemadipine here. To determine the IC 50 and IC 10 of the mutants with respect to the egg-laying-retention (Egl) phenotype and the variable abnormal (Vab) morphological defects induced by the compound, a population enriched for adults for each mutant strain was chunked onto MYOB plates containing either nemadipine or only the DMSO solvent as the negative control [28]. Approximately 24 hours later, young adults were picked on to a new nemadipine or DMSO plates. For the Egl assay, adults were scored as either being Egl or non-Egl approximately 24 hours later. For the Vab assay, the L1 and L2 larvae laid by the adults were scored for morphological defects approximately 24 hours later.
To measure length and count embryos in utero, roughly 100 L4s of each strain were picked onto plates containing either DMSO or 5 mM nemadipine. Approximately 24 hours later, resulting young adults were photographed at 206 using a Leica MZFLIII microscope with a Retiga 1300 digital camera (Q Imaging). The lengths were measured using Openlab software (Improvision, Inc.). Embryos in utero were counted approximately 24 hours after the L4s were place on the plates using a Leica MZFLIII microscope at approximately 806 magnification. For the protruding spicule counts, him-5(e1490) doubles were created for each strain. L4 males were picked and put onto either DMSO or 5 mM plates. Approximately 24 hours later, the males were scored as either having protruding spicules (Prc) or non-Prc.
Creating the Homologous Mutation in the Rat Brain Ca v 1.

Channel
The rat brain L-type a 1C subunit in pMT2 and b 2a and a 2 d were generous gifts of Dr. Terry Snutch (University of British Columbia, Vancouver, Canada). The a 1C subunit was divided into 3 fragments based on endogenous restriction sites: KpnI in pMT2, SalI in a 1C , SpeI in a 1C , and a second SpeI in the 39 untranslated region of a 1C . The first fragment of size 2.5 kB from KpnI (pMT2)-SalI spanned the N-terminus up to domain II S6 of the channel. The last fragment of size 4 kB from SpeI-SpeI (in the 39 untranslated region) spanned domain IIIS1 to the end of the Cterminus. Both of these fragments were separately subcloned into pBluescript SK+ (Stratagene). Specific mutagenic primers (Table  S2) were designed and used in PCR-directed mutagenesis on the subcloned fragments using Stratagene's Site-directed Mutagenesis Kit. M1160L and M1161A were created using the GeneEditor Site Directed Mutagenesis Kit (Promega) with the following primers: M1160L (59-CATTGCCTTCTTCCTGATGAA-CATC-39) and M1161A (59-CATTGCCTTCTTCATGGC-CAACATCTTCGTGGG-39). Fragments were subcloned back into the parent a 1C subunit in pMT2, and the mutations were confirmed by DNA sequencing.
Cell Culture, Transfection, Electrophysiology, and Data Analysis Cell culture and transfection using tsA201 cells are presented in the SI materials and were done as previously described [39,40]. Whole-cell (ruptured) recordings were performed on an Axopatch 700A amplifier linked to a personal computer equipped with pClamp9 (Axon Instruments, Foster City, CA). Patch pipettes (Sutter, BF 150-86-15) were pulled using a Sutter P-87 microelectrode puller and polished with a Narashige microforge [39,40]. Pipettes (2-4 MV) were filled with (in mM) 140 Csmethanesulfonate, 4 MgCl 2 , 9 EGTA, 9 HEPES (pH 7.2 adjusted with CsOH). Bath recording solution comprised of (in mM) 20 BaCl 2 , 85 CsCl, 40 TEA-Cl, 1 MgCl 2 , 10 HEPES, 10 glucose (pH 7.3 adjusted with CsOH). Perfusions consisted of bath solution with various concentrations of DHP from 0.01 mM to 100 mM or 10 mM CdCl 2 , and applied using a gravity-driven system. In some cases, 0.1% DMSO was added to the control bath solution (no drug). Data were filtered at 1 kHz (23 dB, 4-pole Bessel) and digitized at 2 kHz. Currents were elicited by stepping from 2100 mV to the indicated test potentials. Drug-response curve were determined by depolarizing to elicit maximal inward Ba 2+ currents (usually 0 to +10 mV) every 5 seconds, first perfusing with control solution and then with various concentrations of DHP compounds until the current amplitude reached a steady-state level. Recordings were done with Ba 2+ as Ca 2+ may greatly affect current through mechanisms such as Ca 2+dependant inactivation or Ca 2+ -dependant facilitation [42]. Channel behavior was otherwise normal in the mutants as Ba 2+ current through the mutant channels were elicited at the range of applied voltages from 230 to +40 mV, which is similar to that of the wild-type channel. Dose-response curves were fitted to the equation: Where I Ba is the current recorded either in the absence (I Ba(control) ) or the presence of drug; F residual is the fraction residual current at saturated concentrations of DHP; C is the concentration of DHP; IC 50 is the DHP concentration required to elicit half the maximal inhibition; and n is the Hill coefficient (see Table S3 for more information). As DHPs do not physically occlude the pore [43] F residual was included in the analyses so that the estimated IC 50 accurately represents the interaction with the channel and not confounded by the efficacy of the block. Fits were performed in Origin 7 (OriginLab Corp., Northampton, MA). All experiments were performed at room temperature (,22uC). Data are presented as mean6sem. Statistical analysis was done using one-way analysis of variance (ANOVA, Holm-Sidak post-hoc). Differences were considered significant if p,0.05.

Chemicals and Solutions
All chemicals used in the cell culture were purchased from GIBCO (Invitrogen). Chemicals used for physiological recordings were purchased from Sigma. Nifedipine was purchased from Alamone Labs (Jerusalem, Israel) and from Sigma, and nemadipine was purchased from Chembridge Corp. (Chembridge ID#: 5619779). Figure S1 Animals with egg laying (Egl) defects and variable abnormal (Vab) defects in response to increasing concentrations of nemadipine. (a) Dose response curves for the tr89, the domain I mutant, ad695 and n2368, the egl-19 hypermorphs, and N2, the wild-type control. (b) Dose response curves for domain II mutants. (c) Dose response curves for domain III mutants. All the mutants analyzed in the Egl and Vab assays have increased resistance to nemadipine compared to N2 showing that these mutants are DHP-resistant. The Egl and Vab assays were performed as outlined in Materials and Methods. Error bars, 6s.e.m. Found at: doi:10.1371/journal.pgen.1000067.s001 (2.23 MB EPS)