Introduction

Top Abstract Introduction Materials & Methods Results Discussion References

The tachykinin SP is a neurotransmitter and neuromodulator in the central and peripheral nervous systems (Nicoll et al., 1980). As a neurotransmitter, SP acts via the NK-1, a member of the seven transmembrane, G protein-coupled receptor superfamily. The binding of SP to NK-1 receptors leads to activation of phospholipase C, resulting in increased inositol trisphosphate levels and the release of calcium from intracellular stores (Mau and Saermark, 1991). In Xenopus oocytes expressing cloned NK-1 receptors, the release of calcium causes the activation of chloride currents (Fong et al., 1992). As a neuromodulator, SP has been shown to inhibit agonist-induced nAChR activation, as Steinacker and Highstein (1976) first demonstrated at the Mauther fiber-giant fiber synapse in the hatchet fish. Since then, SP has been shown to modulate nicotinic responses of both neuronal (Livett et al., 1979; Akasu et al., 1983; Clapham and Neher, 1984; Simasko et al., 1985; Simmons et al., 1990; Stafford et al., 1994) and skeletal muscle (Akasu et al., 1983; Simasko et al., 1985; Min and Weiland, 1992) nAChRs. These studies have shown that noncompetitive inhibition by SP is a general characteristic of nAChRs, most consistent with a direct interaction with the receptor at a unique site. This site has a pharmacology distinct from that of the G protein-coupled NK receptors. The evidence for a physiological role for this direct modulation is strongest in the adrenal gland where SP-containing neurons innervate the chromaffin cells and SP modulates nAChR-mediated catecholamine secretion (Livett and Zhou, 1991). SP may protect the nAChR from agonist-mediated irreversible deactivation (Boyd and Leeman, 1987) and could be involved in maintaining catecholamine secretion during stress (Livett and Zhou, 1991).

Muscle and neuronal nAChRs are pentameric proteins forming ligand-gated ion channels (Changeux, 1990) that mediate signal transmission at the neuromuscular junction and in the central and peripheral nervous systems. Whereas muscle receptors require four different subunits (, , , and ), functional neuronal receptors can be formed from a combination of  and  subunits (Boulter et al., 1987) or, in certain cases, a single type of  subunit (Couturier et al., 1990). The family of neuronal nAChR subunits continues to grow and there are currently eight  and three  neuronal receptor subunits (Papke, 1993; McGehee and Role, 1995). Both types of subunits of neuronal receptors have been shown to be involved in determining the sensitivity of the receptor to agonists and antagonists (Luetje and Patrick, 1991; Figl et al., 1992; Papke et al., 1993; Harvey and Luetje, 1996). Because it has been demonstrated that the  and  subunits of the muscle nAChR play a role in agonist and antagonist binding (Sine and Claudio, 1991; Czajkowski et al., 1993; Sine, 1993), the involvement of both neuronal subunits is not surprising. Moreover, given the heterogeneity of neuronal nAChR responses in vivo, the "mix and match" of various subunits probably provides the molecular basis for diversity of function (Papke, 1993).

Some of the structural determinants for SP modulation of nAChRs are now becoming apparent. Min et al. (1993) found that the  and  subunits of Torpedo nAChRs were affinity labeled with either [³H]SP and a bifunctional cross-linker or the photoaffinity reagent [¹²⁵I]p-benzoylphenylalanine-SP. Blanton et al. (1994) demonstrated that [¹²⁵I]p-benzoylphenylalanine-SP labeled the M2 region of the Torpedo  subunit. Using the oocyte expression system we recently found that the  subunit of the neuronal receptor contributes to the IC₅₀ for SP inhibition, with -4 subunit-containing receptors having a 25-fold higher apparent affinity for SP than -2-containing receptors, whether coexpressed with -3 or -4 (Stafford et al., 1994). These findings suggested that, using molecular biological approaches and the Xenopus oocyte expression system, the structural domains of the nAChR involved in the interaction of SP with the nAChR might be resolved as they had been previously for several agonists and antagonists (Figl et al., 1992; Luetje et al., 1993; Papke et al., 1993).

We undertook to identify the region(s) and amino acid(s) responsible for the difference in the IC₅₀ for -4- vs. -2-containing receptors, taking advantage of the significant sequence similarities between the subunits. A series of chimeric -4/-2 subunits were expressed with the -3 subunit in Xenopus oocytes and by quantitating the inhibition of agonist-induced current by SP, we were able to identify two regions that appeared to be the determinants of the difference between the subunits. Site-directed mutagenesis of the candidate residues individually demonstrated the importance these amino acids; however, mutation of both residues together was not sufficient to completely interconvert each receptor's sensitivity to SP. It is apparent from these results that more global structural and conformational issues are involved in the binding and inhibition of nAChR activity by SP which cannot be duplicated with two single amino acid changes.

	Materials and Methods

Top Abstract Introduction Materials & Methods Results Discussion References

Construct and plasmid preparation. Plasmid containing cDNA coding for rat -3 was kindly supplied by Dr. Roger Papke (University of Florida, Gainesville, FL). Chimeras between rat neuronal nAChR subunits -2 and -4 were generated as previously described (Figl et al., 1992). The cDNA coding for some of the chimeric subunits were subcloned into the oocyte DNA expression vector pOEV (a gift of Dr. William L. Taylor, Vanderbilt University) at the polylinker sequence which is located between a TFIIIA promoter and an SV40 transcription terminator (Pfaff et al., 1990). Plasmids were propagated in the Escherichia coli host (Dh5 strain) and purified using the Wizard miniprep kit (Promega, Madison, WI). mRNA was transcribed using SP6 and T3 Ampliscribe (Epicentre, Madison, WI) and capped by the inclusion of diguanosine triphosphate.

Preparation and injection of Xenopus oocytes. Oocytes were harvested from adult Xenopus laevis (Nasco, Fort Atkinson, WI) under anesthesia (0.15% MS222) and manually dissected into groups of several dozen. The follicle layers were removed by incubation in Ca⁺⁺-free oocyte saline solution (82.5 mM NaCl, 2.5 mM KCl, 1 mM Na₂HPO₄, 15 mM HEPES, 1 mM MgCl₂, pH 7.4) containing collagenase type I (1-2 mg/ml). Oocytes were maintained at 18°C in oocyte saline solution (with 1 mM CaCl₂) containing 5% horse serum, 5 U/ml penicillin, 5 µg/ml streptomycin and 150 µg/ml amikacin, and the medium was changed daily. Four to five days before recording, 10 nl of DNA (~2 ng of plasmid DNA) was injected into the nucleus. Alternatively, 50 nl of RNA (2 to 5 ng/subunit) was injected into the cytoplasm 3 to 5 days before recording. Injections were made using the Nanoject positive displacement oocyte injector (Drummond, Broomall, PA).

Voltage-clamp measurements and analysis. Two electrode voltage-clamp measurements were made at room temperature using a Turbo Tec 01C amplifier (Adams & List, Westbury, NY). The voltage electrode was filled with 3 M KCl and had a resistance of 0.4 to 2 M. The current electrode was filled with 250 mM CsCl, 250 mM CsF and 100 mM EGTA, pH 7.3. The resistance of the current electrode was between 0.5 and 2 M. Cells were routinely voltage-clamped at -70 mV. Bath solution (oocyte saline with 1 mM CaCl₂ and 1 µM atropine to prevent activation of muscarinic acetylcholine receptors) was delivered at ~6 ml/min through a linear perfusion system to oocytes placed in a Delrin chamber with a total volume of 0.45 ml. ACh/peptide solutions were delivered by preloading 2 ml in a loop at the terminus of the perfusion system using a syringe. A Mariotte flask filled with oocyte saline solution was used to maintain constant hydrostatic pressure, and the ACh/peptide application was initiated by a computer-triggered stream-switching valve (Rainin, Emeryville, CA). The time between applications was 6 to 10 min to allow recovery from ACh-induced desensitization. Data were collected on-line with an IBM AT computer using software developed in the laboratory. Current traces were recorded at the same time on a chart recorder. The digitized recordings were transferred from the IBM AT to a Sun 4/330 computer for further analysis using PLOT (Gradient Software, Ithaca, NY).

(1)

(2)

	Results

Top Abstract Introduction Materials & Methods Results Discussion References

Dependence of inhibition by substance P on subunit structure. Because of the significant sequence identity between the beta subunits (63% overall and more than 90% in some regions, see fig. 1), a series of chimeric -4/-2 constructs were coexpressed with -3 in Xenopus oocytes to attempt to identify the regions responsible for the 25-fold difference in the IC₅₀ of SP for -4- vs. -2-containing receptors (fig. 2). The chimeric beta subunits examined focused on two potentially important areas, the extracellular N-terminal domain [-4(214):-2, -4(116):-2, -4(113):-2, -4(111):-2, -4(109):-2 and -4(105):-2] and the first three transmembrane domains [-4(301):-2, -4(214):-2 and -2(299):-4] which include the putative pore-lining M2 region. We expected the N-terminal region to be important since it comprises more than 90% of the extracellular domain and had previously been shown to be important in the interaction of agonists with the receptor (Figl et al., 1992). We were also particularly interested in the M2 region, because biochemical studies had indicated it contributed to the binding site of SP (Min et al., 1993; Blanton et al., 1994). The large intracellular loop and fourth transmembrane region were also examined [-4(301):-2, -4(325):-2, -2(299):-4, -2(323):-4 and -2(421):-4]. Based on the results of these studies, several point-mutated subunits were generated and characterized (fig. 2).

View larger version (64K): [in this window] [in a new window]   Fig. 1.   Aligned amino acid sequences of 2 (top) and 4 (bottom). Dots between sequences indicate identical residues. The four putative transmembrane regions are denoted by horizontal bars.

View larger version (34K): [in this window] [in a new window]   Fig. 2.   Linear maps and IC50 values for substance P inhibition and EC50 values for acetylcholine-induced currents of chimeric and mutated  subunits coexpressed with 3. Black areas represent amino acids from 4, white areas represent amino acids from 2, amino acid residues are numbered at the bottom, and lines above the subunits denote putative transmembrane regions. Inhibition by SP was determined at a concentration of ACh within a factor of 1.7 (0.60-1.43) of its EC50 value and the responses for each oocyte were normalized to the current measured in the presence of ACh alone. IC50 values and standard errors were obtained by nonlinear fitting equation 2 to all the data from each subunit/construct (6-24 points). To determine EC50 values for ACh, peak current was measured in the presence of increasing concentrations of ACh and normalized to 100 µM ACh. The parameters of ACh activation were determined by fitting equation 1 to all the data from each subunit/construct. Hill coefficients ranged from 0.71 ± 0.08 (2) to 2.5 ± 0.09 (4(105):2). Data from each subunit combination were pooled from (n) oocytes. The constructs 2(107):4, 2(109):4, 2(111):4, 4(423):2 and 2(V253F) did not produce measurable ACh-induced currents when coinjected with 3, presumably for technical rather than structural reasons. ND, Not determined.

View larger version (20K): [in this window] [in a new window]   Fig. 3.   Currents induced by application of acetylcholine (solid line) or acetylcholine plus 5 µM substance P (broken line). Drugs were applied for 20 sec, as shown by horizontal bar. Current traces shown in the presence and absence of SP are from the same oocyte. The concentrations of ACh used were 10 µM for 32(F106V;V253F), 25 µM for 34(109):2 and 32(323):4 and 100 µM for the others.

View larger version (21K): [in this window] [in a new window]   Fig. 4.   Concentration-dependence of substance P inhibition of acetylcholine-induced currents for wild-type 4 and 2 and several chimeric 4:2 subunits coexpressed with 3. Peak current responses to ACh were measured in the presence of increasing concentrations of SP. ACh concentrations were 100 µM for 34, 34(301):2, 34(214):2, 34(116):2, and 34(105):2; 50 µM for 32. Data for each subunit combination are from three to six oocytes. For each oocyte the responses were normalized to the current measured in the presence of ACh alone. Each point is the mean of one to four measurements at each concentration. Error bars represent the S.E.M.; if there are no error bars, SEM is smaller than the symbol or only one determination was made at that concentration. The lines are nonlinear fits of equation 2 to all the data from each subunit/construct (6-24 points).

View larger version (14K): [in this window] [in a new window]   Fig. 5.   IC50 values for SP and EC50 values for ACh of receptors containing 3 and chimeric  subunits. Value of zero amino acids on the x-axis is the 2 subunit, 475 is 4. Data are from figure 2. Lines were drawn to connect data points which appear continuous. Discontinuities in the IC50 values between 4 residues 105 and 116 and between 214 and 301 reflect the importance of these regions in the interaction of SP with the receptor.

View larger version (20K): [in this window] [in a new window]   Fig. 6.   Concentration-dependence of substance P inhibition of acetylcholine-induced currents for wild-type 4 and 2 and point-mutations of 4 coexpressed with 3. ACh concentrations were near the EC50 value determined for each mutant as shown in figure 2. Data for each subunit combination are from two to five oocytes. For each oocyte the responses were normalized to the current measured in the presence of ACh alone. Each point is the mean of one to six measurements at each concentration. Error bars represent the S.E.M.; if there are no error bars, S.E.M. is smaller than the symbol or only one determination was made at that concentration. The lines are nonlinear fits of equation 2 to all the data from each subunit/construct (6-24 points).

Acetylcholine dose-responses. The EC₅₀ value of ACh for each subunit/chimera/mutation combined with 3 was determined so that SP inhibition could be investigated using concentrations of ACh that gave comparable relative responses. A 20-sec application of ACh induced activation of inward cationic currents, as shown in figure 3 (solid lines). EC₅₀ values for all the expressed receptor subunit combinations are shown in figure 2. In this study we observed only about a 4-fold difference in EC₅₀ values of ACh for receptors containing 2 or 4 subunits (37 vs. 143 µM, fig. 2). This is in contrast to the near 20-fold difference (10 vs. 210 µM) previously reported (Cohen et al., 1995). The discrepancies in the EC₅₀ values for ACh most likely reflect 1) differences in the method of agonist application (in the current study we used a relatively slow bath perfusion, while the previous study used a more rapid U-tube application), 2) differences in the holding potential (-70 mV in the current study and -50 mV in the previous report) and 3) differences in the composition of perfusion solutions, most notably 1 µM atropine was included in the current study. It is not unexpected that differences in the rate of drug application would cause discrepancies in the quantitative values determined (especially EC₅₀ values of agonists where desensitization can significantly affect the peak current observed). For example, Harvey and Luetje (1996) reported only a 3-fold difference in EC₅₀ values of ACh for 32 vs. 34 receptors (71 vs. 210 µM), using a relatively slow perfusion method and with a holding potential of 70 mV. It should be noted that despite these differences in EC₅₀ values, both these studies and the current one (data not shown) found significant differences in the apparent cooperativity for ACh activation of 2- vs. 4-containing receptors. For all these studies the Hill coefficient for ACh was near 1.0 for 32 and near 2.0 for 34.

	Discussion

Top Abstract Introduction Materials & Methods Results Discussion References

We had previously found that neuronal nAChRs containing 4 subunits have a higher affinity for SP than do 2-containing receptors, whether they are coexpressed with 3 or 4 (Stafford et al., 1994). We identified two separate regions of the  subunit that are important for this difference in the sensitivity of the receptors to SP: between 4(105) and 4(116) and between 4(214) and 4(301). Chimeric subunits containing portions of the two  subunits were used to locate these regions, and point mutations were introduced to attempt to determine the individual amino acids involved. The amino acids V108 and F255 of 4 and the homologous 2 residues (F106 and V255) were identified as important determinants of the affinity for SP, although the double mutations did not convert one subtype to the other, indicating other residues are making significant, but more subtle, contributions to the interaction with the peptide.

The  subunit has previously been shown to affect many properties of neuronal nAChRs, including the single channel characteristics. Papke and Heinemann (1991) have shown that the  subunit affects the rate of ACh dissociation and the rate of channel opening. Receptors containing 4 are much more sensitive to the ganglionic stimulants cytisine and nicotine (Luetje and Patrick, 1991; Figl et al., 1992) but much less sensitive to the neurotransmitter ACh (Cohen et al., 1995) and the antagonists DHE (Harvey and Luetje, 1996) and nBGT (Papke et al., 1993; Harvey and Luetje, 1996) than are 2 containing receptors. Chimeric 4/2 subunits have been used to map the regions responsible for most of these differences. From these studies it is clear that the extracellular N-terminus is the most important region of the  subunit for the interactions of these compounds with the receptor. Amino acids 4(108) and 4(110) can account for much of the relative sensitivity to cytisine, although the difference in nicotine sensitivity could not be localized to a particular region of the  subunit (Figl et al., 1992). The first 121 amino acids of the  subunit determined the kinetics of nBGT block (Papke et al., 1993). The major determinant of DHE and nBGT affinity was shown to be 4(K61)/2(T59) with other minor determinants in the first 100 residues of the N-terminus (Harvey and Luetje, 1996). Chimeras of 4/2 subunits were used by Cohen et al. (1995) to demonstrate the importance of the first 120 residues in determining the EC₅₀ for ACh, with residues between 2(104) and (120) accounting for the relative sensitivity of 32 to cytisine, TMA, and ACh.

Chimeric subunits followed by site-directed mutation have been used to identify amino acids of the  and  subunits involved in curare binding to muscle nAChRs (Sine, 1993) and of the  subunits that contribute to agonist and antagonist sensitivity of neuronal nAChRs (Luetje et al., 1993). Sine (1993) expressed / chimeras to identify two regions in mouse  and  subunits that were determinants of curare affinity. Interestingly, two of the residues identified in the mouse  subunit (I116 and Y117) are homologous to  subunit residues 2(I118) and 2(F119) that are very near the region we found to affect ACh and substance P interactions with the receptor (see fig. 1). Luetje et al. (1993) used chimeric 2/3 subunits expressed with 2 in Xenopus oocytes to identify the determinants of nBGT sensitivity and the relative nicotine/ACh sensitivities between 2 and 3 coexpressed with 2.

We have identified two regions of the  subunit that were responsible for the difference in sensitivity to SP inhibition of ACh-induced current in nAChRs. Receptors containing chimeric subunits with the first 105 amino terminal residues from 4 displayed the same sensitivity as 2-containing receptors. If the amino terminal 4 residues were extended to 109, the resultant receptors displayed SP sensitivity intermediate between 32 and 34. Extending the amino terminal 4 residues to 214, thus including the entire N-terminal extracellular domain, had no additional effect and resulted in receptors with similar intermediate sensitivity. However, when the chimeric  subunits contained 4 amino terminal residues past the third transmembrane domain (~300 amino terminal residues), the receptors displayed SP sensitivity of wild-type 34. Similarly, receptors with chimeric subunits having 2 past M3 had wild-type 32 SP sensitivity. Each of these regions, 4(105-109) and 4(214-301), accounted for approximately half of the difference in sensitivity to SP of the two wild-type receptors (see figs. 2 and 5).

There is a single amino acid in the region of 4(105) and 4(109) that differs between 2 and 4: a phenylalanine at 2(106) is replaced by a valine at 4(108). It was expected that substitution of that residue with the homologous one would result in receptors with intermediate IC₅₀ values for SP. Receptors with the single mutation 34(V108F) were only slightly less sensitive to SP than wild-type 34, and 32(F106V) had an IC₅₀ for SP of about 150 µM, much greater than 32 wild-type. This lack of reciprocity when exchanging residues between subtypes was also seen by Luetje et al. (1993) for the relative sensitivity of nicotine vs. ACh, and probably reflects subtle, but significant, contributions by other residues. Because this region has also been shown to be important in the interaction of agonists (fig. 5; Figl et al., 1992; Cohen et al., 1995) and inhibition by SP is noncompetitive (Stallcup and Patrick, 1980; Simasko et al., 1987; Stafford et al., 1994), it is most likely that this region is not involved directly in the binding of SP. These residues may contribute to the agonist binding site as suggested by Cohen et al. (1995) or may participate in conformational changes involved in agonist-induced activation or desensitization, which could indirectly affect the apparent affinity of SP by altering agonist properties. Even if these residues are within the agonist binding site, they can contribute to the gating properties of the channel as has been shown, for example, by Chen et al. (1995) who found mutation of tyrosine 190 within the ACh binding site of the -subunit affected both agonist binding and activation kinetics (for review see Arias, 1997).

Between residues 4(214) and 4(301) there are five amino acid differences, but only a single nonconservative change. In the M2 region, 2 has a valine at 253 and 4 has a phenylalanine at the comparable position (255). The receptor 34(F255V) had an IC₅₀ for SP about halfway between the two wild-type receptors. We believe that the single residue 4(255) alone can account for the difference in SP sensitivity mapped to between 4(214) and 4(301) because the difference in the IC₅₀ between wild-type 4 and 4(F255V) (3.3 vs. 19 µM) is essentially the same as the difference between 4(301):2 and 4(214):2 (5.5 vs. 20 µM). Because 32(V253F) would not express, the effect of the single mutation in M2 of 2 remains unknown. The amino acid 4(F255) [and the homologous residue, 2(V253)] is located in the middle of the putative second transmembrane domain, which is the believed to line the channel pore (Imoto et al., 1986; Oiki et al., 1988). 4(F255)/2(V253) is four amino acids nearer the extracellular mouth of the receptor than the highly conserved leucine [2(L249)/4(L251), see fig. 1] that is thought to face the lumen and contribute to the narrowest region of the pore (Unwin, 1993). Therefore, it should also face the lumen of the pore, whether M2 is an -helix or -structure. It is not unlikely that SP is binding in the channel near this residue, because Blanton et al. (1994) have cross-linked the affinity label [¹²⁵I]p-benzoylphenylalanine-SP to the M2 region of the  subunit from Torpedo. The greater sensitivity of 4 containing receptors to SP inhibition could then be explained by the stabilization of the positive charges on SP by the aromatic  electrons of the phenylalanine, much as the aromatic amino acids stabilize ACh in the binding pocket of acetylcholinesterase (Dougherty and Stauffer, 1990; Sussman et al., 1991). This could account for the reduction in SP IC₅₀ for 32(F106V;V253F) receptors over 32(F106V) receptors. However, longer range effects cannot be ruled out, as found with mutations in M2 of 7 where agonist and antagonist interactions with the receptor were affected by point mutations in the channel region (Revah et al., 1991; Bertrand et al., 1992; Devillers-Thiéry et al., 1992).

Combining the two single mutations in 4 did not produce an additive effect, and did not result in the predicted conversion of 4 to 2 sensitivity to SP. 34(V108F;F255V) receptors had essentially the same IC₅₀ as 34(F255V). The double mutant in the 2 subunit, 32(F106V;V253F) had an intermediate IC₅₀ for SP, demonstrating that the phenylalanine in the M2 could render the 2 containing receptor more sensitive to SP. However, changing 2(253) from a valine to a phenylalanine in the chimera 4(109):2 had no effect on SP sensitivity. This suggests, as discussed previously, that the differences in affinity for substance P are not due to only these two amino acids, but other residues are involved, making subtle, three-dimensional structural contributions that are not apparent from the results with the chimeras.

Using chimeric  subunits coexpressed with 3 in Xenopus oocytes, it has been possible to define two areas of the  subunit that contribute to the differences in sensitivity to SP of 2- and 4-containing nAChRs. Although the two amino acids identified could not account for all of the difference, 4(108) and 4(255) [and the 2 homologs, 2(106) and 2(253)], clearly play important roles in the inhibition of nAChR activation by SP. Most likely 4(255) is involved in the binding of SP to the receptor while 4(108) may be involved in agonist binding and/or receptor activation and indirectly participate in the inhibitory action of the peptide.