Schistosome Calcium Channel (cid:1) Subunits UNUSUAL MODULATORY EFFECTS AND POTENTIAL ROLE IN THE ACTION OF THE ANTISCHISTOSOMAL DRUG PRAZIQUANTEL*

Schistosomes are parasitic flatworms that cause schistosomiasis, a major tropical disease. The current drug of choice against schistosomiasis is praziquantel (PZQ), which has minimal side effects and is potent against all schistosome species. The mode of action of PZQ is unknown, though the drug clearly affects Ca 2 (cid:2) homeostasis in worms, and there is indirect evidence for interaction of PZQ with schistosome voltage-gated Ca 2 (cid:2) channels. We have cloned and expressed two Ca 2 (cid:2) channel (cid:1) subunits, one from Schistosoma mansoni and one from Schistosoma japonicum . These two subunits ( Sm Ca v (cid:1) A and Sj Ca v (cid:1) ) have structural motifs that differ from those found in other known (cid:1) subunits. Surpris-ingly, coexpression of either Sm Ca v (cid:1) A or Sj Ca v (cid:1) with a cnidarian ( Cy Ca v 1) or mammalian (Ca v 2.3) Ca 2 (cid:2) channel (cid:3) 1 subunit results in a striking reduction in current amplitude. In the case of Ca v 2.3, this current reduction can be partially reversed by addition of 100 n M PZQ, which results in a significant increase in current amplitude. Thus, these unusual schistosome


EXPERIMENTAL PROCEDURES
Isolation of ␤ Subunit cDNAs from Schistosomes-The initial cDNA fragment was amplified by PCR with degenerate primers using either a cDNA pool (RT-PCR) or a cDNA library as template. For RT-PCR, total RNA was extracted from fresh mixed-sex adult S. japonicum by homogenization of the tissue in Trizol reagent (Life Technologies, Inc.). RNA was precipitated in isopropanol and resuspended in diethylpyrocarbonate-treated water. cDNA was synthesized from 10 g of total RNA using 1 pmol of an oligo(dT) adapter primer and 200 units of SuperScript II reverse transcriptase (Life Technologies, Inc.). Incubation was for 2 h at 42°C. Degenerate primers for PCR were designed against highly conserved regions from other Ca 2ϩ channel ␤ subunits. The sense primer (AAYAAYGAYGGTGGAT) was based on the amino acid sequence NNDWWI. The antisense primer (GCYTTYTGCATCATRTC) was based on the amino acid sequence DMMQK. The PCR consisted of 1 l of template in a reaction containing 50 pmol of each degenerate primer and LA Taq (Panvera) as the enzyme. 30 cycles of 94 (1 min), 52 (1 min), and 72°C (1 min) were used. A single, ϳ500-bp band was gel-purified (QIAquick; Qiagen) and ligated into pCR4-TOPO TA cloning vector (Invitrogen). Plasmids were sequenced on an ABI 310 Genetic Analyzer using the BigDye Terminator Cycle Sequencing Kit (ABI). Both the RT and library generated identical clones containing a 564-bp insert encoding sequence similar to known ␤ subunits. Both 5Ј and 3Ј ends of the cDNA were obtained from the cDNA library by PCR, using a combination of a vector-specific primer and a gene-specific primer. Once the entire coding sequence was determined, PCR from a different cDNA pool was used to amplify a full-length copy of the coding sequence, providing a consensus sequence named SjCa v ␤. SmCa v ␤A was cloned using similar methods but using cDNA pools and libraries made from S. mansoni RNA.
Expression in Xenopus Oocytes-RNA was transcribed from cDNA clones containing full-length coding regions of SjCa v ␤, SmCa v ␤A, and other subunits, using the T7 mMessage mMachine in vitro transcription kit (Ambion). The amount of purified, transcribed RNA was estimated on glyoxal denaturing gels. RNA was injected into oocytes by combining the ␣ 1 subunit RNA with the ␤ subunit RNA or water at a 1:3 molar ratio. Stage V-VI oocytes isolated from Xenopus were injected with 10 -20 ng of RNA and incubated in sterile ND96 medium (containing 1.8 mM CaCl 2 , supplemented with 2.5 mM sodium pyruvate, 100 units/ml penicillin, 100 g/ml streptomycin, and 5% horse serum) at 17°C for 3-7 days. In some cases, other amounts of RNA were also tested. Because levels of channel expression vary over time, recordings on any given day were from oocytes in all treatment groups. Oocytes were injected with 40 nl of 100 mM BAPTA dissolved in 10 mM HEPES at least 1 h prior to recordings to eliminate potential artifacts from the endogenous Ca 2ϩ -activated chloride current in oocytes. The bath solution consisted of 40 mM Sr(OH) 2 , 40 mM N-methylglucamine, 10 mM glucose, and 10 mM HEPES, pH 7.4. Sr ϩ2 was used as the charge carrier, because both ␣ 1 subunits used (CyCa v 1 and Ca v 2.3) are most * This work was supported in part by National Institutes of Health Grant AI 40522 (to R. M. G. and P. A. V. A.). The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked "advertisement" in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
These permeable to that ion (6,7). Similar results were obtained using Ba 2ϩ as the charge carrier.
Currents were recorded using a Warner Instruments two-electrode voltage clamp and 3 M KCl-filled borosilicate microelectrodes with impedances of 0.1-1 megohms. Oocytes were clamped at Ϫ90 mV. Data were recorded and analyzed with pClamp6 and pClamp8 software (Axon Instruments), and leakage currents were subtracted on-line. All experiments were conducted at room temperature.
Praziquantel (Sigma) was dissolved as a 10 Ϫ2 M stock solution in Me 2 SO and diluted in bath solution to the final concentration. Effects of PZQ were measured within seconds following application of the drug. Me 2 SO alone had no effect at the dilutions used.

RESULTS
Cloning of Ca 2ϩ Channel ␤ Subunit cDNAs from Schistosomes-We used degenerate oligonucleotide primers and PCR to amplify a 564-bp portion of a cDNA from S. japonicum with similarity to other known Ca 2ϩ channel ␤ subunits. The 2319-bp full-length coding region of the sequence was subse-quently obtained and named SjCa v ␤. Northern blots probed with a portion of this sequence reveal several bands, ranging in size from 2.4 to 4.4 kilobases (data not shown). The cDNA for SjCa v ␤ codes for a predicted protein of ϳ84 kDa, which, as with other invertebrate ␤ subunits, does not cluster with any of the mammalian subtypes. We also cloned a similar (88% identical) ␤ subunit cDNA from S. mansoni (SmCa v ␤A). SmCa v ␤A codes for a slightly larger protein of ϳ87 kDa. As is found for other schistosome sequences, coding regions for both of these sequences are very A/T-rich (ϳ62%).
Although the sequences of SjCa v ␤ and SmCa v ␤A are more similar to other known ␤ subunits than to any other proteins, they both have unusual structural features in comparison to other ␤ subunits (see Fig. 1). For example, both of these schistosome ␤ subunits are larger than other ␤ subunits, containing extended sequence in the C-terminal region. Both sequences also show differences within the ␤-interaction domain (BID), FIG. 1. SjCa v ␤ and SmCa v ␤A sequences. A, alignment of the two schistosome ␤ subunits (SjCa v ␤ and SmCa v ␤A) with the ␤2a subunit from rat (Ca v ␤2a; accession number M80545). Amino acid sequences were aligned with ClustalW. Identities are shaded black. The BID is shown. Note the extended C-terminal sequence in the two schistosome ␤ subunits. B, alignment of regions corresponding to the BID consensus sequence (8,9) from several Ca 2ϩ channel ␤ subunits. Two serine residues that represent potential phosphorylation sites found in nearly every known BID are not conserved in the two schistosome ␤ subunits (black shading). The only other known BID that does not contain both these serine residues is found in a putative ␤ subunit sequence in the Caenorhabditis elegans genome. The serine residue shown to reduce increases in current amplitude when mutated to arginine (8) is denoted by an asterisk. In SjCa v ␤, the final residue in this region is a serine that is also a potential protein kinase C phosphorylation site (box). Accession numbers are as follows: rabbit ␤1a, M25817; rat ␤1b, X61394; human ␤1c, M76560; human ␤2a, AAD33730; rabbit ␤2b, X64298; rat ␤3, M88751; rat ␤4, L02315; Drosophila, AAF21096; C. elegans-1, AAB53056; C. elegans-2 (partial sequence), AAK21500; jellyfish (Cyanea capillata), AAB87751.
Schistosome Calcium Channel ␤ Subunits and Praziquantel 36874 the primary site of ␤ subunit interaction with the ␣ 1 subunit (see Ref. 8 and Fig. 1B). Two conserved serine residues that represent potential phosphorylation sites in the BID are changed to other residues in both SjCa v ␤ and SmCa v ␤A. Interestingly, in SjCa v ␤ the final residue of the BID consensus sequence is a serine, creating a potential protein kinase C phosphorylation site not found in other ␤ subunits.
SjCa v ␤ and SmCa v ␤A Decrease Current Amplitude-Each of the schistosome ␤ subunits was coexpressed with either human Ca v 2.3 or jellyfish CyCa v 1 ␣ 1 subunits. For both ␣ 1 subunits, coexpression of either of these schistosome ␤ subunits results in a dramatic reduction in current amplitude (Fig. 2, A-C).
SmCa v ␤A appears to be somewhat more potent at reducing peak currents than SjCa v ␤, with up to 10 -20-fold reductions in current amplitude. Similar levels of current reduction were found when expression included a mammalian ␣ 2 -␦ subunit (data not shown). Coexpression of CyCa v 1 or Ca v 2.3 ␣ 1 subunits with ␤ subunits from other organisms showed expected large (CyCa v ␤) or minor (␤2a) increases in current amplitude (Fig. 2, B and C).
Other than their unusual effects on current amplitude, both SjCa v ␤ and SmCa v ␤A modulate ␣ 1 subunits as other ␤ subunits do. For example, these ␤ subunits shift the current/ voltage relationship of expressed Ca v 2.3 (see Fig. 2D) and Cy-Ca v 1 (not shown) in a hyperpolarizing direction and have effects on inactivation rates and voltage dependence of inactivation similar to those found for other ␤ subunits. 2 Ca v 2.3 Exhibits Sensitivity to PZQ When Coexpressed with SjCa v ␤ or SmCa v ␤A-Currents were measured in oocytes expressing different combinations of ␣ 1 and ␤ subunits in the presence or absence of 100 nM PZQ (Fig. 3). Oocytes expressing Ca v 2.3 alone or CyCa v 1 alone (not shown) showed no increase in current in the presence of 100 nM PZQ. Similarly, Ca v 2.3 showed no sensitivity to PZQ when coexpressed with a jellyfish ␤ subunit (CyCa v ␤) or with a mammalian ␤ subunit (Ca v ␤2a; not shown). However, oocytes coexpressing mammalian Ca v 2.3 with either SjCa v ␤ or SmCa v ␤A showed nearly a doubling of current amplitude in the presence of 100 nM PZQ. In contrast to Ca v 2.3, the jellyfish CyCa v 1 ␣ 1 subunit remained insensitive to PZQ when coexpressed with either SjCa v ␤ or SmCa v ␤A. 2 Manuscript in preparation.

FIG. 3. Ca v 2.3 exhibits sensitivity to PZQ when coexpressed
with schistosome ␤ subunits. The ratio of the peak current in the presence and absence of 100 nM PZQ is shown for different ␣ 1 /␤ combinations (x axis). If PZQ has no effect, then the ratio will equal one. Results are shown for Ca v 2.3 alone (n ϭ 6), Ca v 2.3 plus a jellyfish ␤ subunit (CyCa v ␤; n ϭ 8), a jellyfish L-type ␣ 1 subunit (CyCa v 1) plus SjCa v ␤ (n ϭ 7), Ca v 2.3 plus SjCa v ␤ (n ϭ 12), and Ca v 2.3 plus SmCa v ␤A (n ϭ 13). Asterisks indicate significant differences from the first column (p Ͻ 0.005, two-tailed t test). In this report, we describe the cloning and functional expression of Ca 2ϩ channel ␤ subunit cDNAs from two schistosome species. Both of these ␤ subunits have unusual structural features compared with other known ␤ subunits. Furthermore, in contrast to other known ␤ subunits, coexpression of these schistosome ␤ subunits with two different ␣ 1 subunits results in a dramatic decrease in current amplitude. Finally, coexpression of the PZQ-insensitive Ca v 2.3 with either of these schistosome ␤ subunits results in an expressed current that is sensitive to PZQ, indicating that these ␤ subunits, in combination with particular ␣ 1 subunits, may be important molecular targets of PZQ action.
␤ subunits modulate high voltage-activated ␣ 1 subunits in several typical ways (reviewed in Refs. 9 and 10). These include increasing peak current amplitude, shifting activation in a hyperpolarizing direction, changing inactivation kinetics, and changing the rate of recovery from inactivation (11). ␤ subunits are also likely required for proper assembly of mature channels in the membrane. These effects are found even when coexpressing ␣ 1 and ␤ subunits from widely divergent phyla (12,13).
The two schistosome ␤ subunits we describe here clearly modulate ␣ 1 subunits in an unusual manner. Instead of increasing current amplitude as other ␤ subunits do, these ␤ subunits dramatically decrease amplitude of currents expressed in Xenopus oocytes. An explanation for this unusual effect may reside within the BID. Both of these schistosome ␤ subunits lack two conserved serines within their BIDs. These serines are consensus protein kinase C phosphorylation sites. Mutation of one of these serines (Fig. 1B, *) to arginine in the ␤1b subunit results in a reduced enhancement of current amplitude with no detectable effect on ␣ 1 /␤ binding (8). Perhaps the loss of both serines in these schistosome ␤ subunits results in an even more dramatic effect on current amplitude. Other modulatory effects of these ␤ subunits on ␣ 1 subunits are similar to those found for other ␤ subunits, indicating that the reduction in current is likely not the result of some nonspecific interaction.
Perhaps most interestingly, both of these schistosome ␤ subunits appear to confer PZQ sensitivity to an otherwise insensitive mammalian channel. Coexpression of either of these ␤ subunits with human Ca v 2.3 results in a significant increase in voltage-gated current amplitude in the presence of the drug. Besides potentially providing insight into the mode of PZQ action, this result also indicates that the reduction in current amplitude induced by schistosome ␤ subunits cannot be explained entirely by a defect in channel assembly. Clearly, channels appear to be present in the membrane but are either not gating currents or are gating reduced currents. In this crossspecies heteromer, PZQ appears to partially reverse this inhibitory effect.
These schistosome ␤ subunits are likely competing with the endogenous ␤ subunit found in Xenopus oocytes (14) for binding to expressed ␣ 1 subunits. Indeed, we have preliminary data from coexpression experiments (not shown) indicating that these schistosome ␤ subunits compete with other expressed ␤ subunits in a concentration-dependent manner to modulate ␣ 1 subunits.
Interestingly, PZQ-sensitive currents do not result when the jellyfish L-type ␣ 1 subunit (CyCa v 1) is coexpressed with SjCa v ␤ or SmCa v ␤A. There are at least two possible explanations for this difference. First, it may reflect some subtle distinction in the interaction of L-type (CyCa v 1) and non L-type (Ca v 2.3) ␣ 1 subunits with these schistosome ␤ subunits or with PZQ. Vertebrate non-L-type Ca v 2.1 and Ca v 2.3 ␣ 1 subunits have been shown to contain a secondary ␤ subunit interaction domain at the C terminus (14,15). Second, it is possible that the difference reflects structural motifs that are common to Ca v 2.3 and, presumably, schistosome ␣ 1 subunits but not found in CyCa v 1. These differences would be independent of whether the channel is L-type or non-L-type. For example, schistosome and other flatworm L-type ␣ 1 subunits have a non-charged residue at a conserved negatively charged site in the Domain I pore region (5). Ca v 2.3, but not CyCa v 1, also has a non-charged residue at that site. These and other hypotheses are currently being tested.
The results presented here suggest a model for PZQ action that is consistent with the observed effects of PZQ on Ca 2ϩ homeostasis in schistosomes. These unusual schistosome ␤ subunits presumably inhibit currents through those schistosome ␣ 1 subunits with which they are associated. We speculate that PZQ may disrupt this ␣ 1 /␤ interaction in these channels, thereby either allowing more channels to open or allowing more current to flow through individual channels. As a consequence, normal Ca 2ϩ homeostasis is disrupted. We are currently testing the effects of PZQ directly on schistosome Ca 2ϩ channels expressed in a heterologous system. Eventually, recording of currents through single channels will refine this model, and coexpression of different combinations of schistosome ␣ 1 and ␤ subunits will be used to test it more rigorously.
Interestingly, we have cloned a third schistosome ␤ subunit (SmCa v ␤B) that more closely resembles other ␤ subunits in its structure. We would expect that these two (or more?) types of schistosome ␤ subunits may have characteristic patterns of expression in the worm and may have distinct modulatory effects on the different schistosome ␣ 1 subunits.