The prototoxin LYPD6B modulates heteromeric α3β4-containing nicotinic acetylcholine receptors, but not α7 homomers

Abstract

Prototoxins are a diverse family of membrane-tethered molecules expressed in the nervous system that modulate nicotinic cholinergic signaling, but their functions and specificity have yet to be completely explored. We tested the selectivity and efficacy of leukocyte antigen, PLAUR (plasminogen activator, urokinase receptor) domain-containing (LYPD)-6B on α3β4-, α3α5β4-, and α7-containing nicotinic acetylcholine receptors (nAChRs). To constrain stoichiometry, fusion proteins encoding concatemers of human α3, β4, and α5 (D and N variants) subunits were expressed in Xenopus laevis oocytes and tested with or without LYPD6B. We used the 2-electrode voltage-clamp method to quantify responses to acetylcholine (ACh): agonist sensitivity (EC₅₀), maximal agonist-induced current (I_max), and time constant (τ) of desensitization. For β4–α3–α3–β4–α3 and β4–α3–β4–α3–α3, LYPD6B decreased EC₅₀ from 631 to 79 μM, reduced I_max by at least 59%, and decreased τ. For β4–α3–α5D–β4–α3 and β4–α3–β4–α–α5D, LYPD6B decreased I_max by 63 and 32%, respectively. Thus, LYPD6B acted only on (α3)₃(β4)₂ and (α3)₂(α5D)(β4)₂ and did not affect the properties of (α3)₂(β4)₃, α7, or (α3)₂(α5N)(β4)₂ nAChRs. Therefore, LYPD6B acts as a mixed modulator that enhances the sensitivity of (α3)₃(β4)₂ nAChRs to ACh while reducing ACh-induced whole-cell currents. LYPD6B also negatively modulates α3β4 nAChRs that include the α5D common human variant, but not the N variant associated with nicotine dependence.—Ochoa, V., George, A. A., Nishi, R., Whiteaker, P. The prototoxin LYPD6B modulates heteromeric α3β4-containing nicotinic acetylcholine receptors, but not α7 homomers.

Because nicotinic cholinergic signaling is known to be involved in an array of human behaviors, such as learning, memory, addiction, and attention (1, 2), the discovery of endogenous modulators of nicotinic acetylcholine receptors (nAChRs) is significant. These modulators are proteins known as prototoxins, so named because they share structural characteristics with 3-fingered snake venom proteins, such as α-bungarotoxin and cobratoxin (3, 4). Prototoxins belong to a larger family known as Ly-6/urokinase plasminogen activator receptor (Ly6/uPAR). At the primary structural level, prototoxin proteins have 8–10 conserved cysteine residues that allow for the formation of disulfide bonds, constraining the protein to the 3-fingered motif secondary structure (5). Several prototoxins have been identified that are expressed in the nervous system and act as allosteric modulators of nAChRs. These include leukocyte antigen (LY)-6/neurotoxin (LYNX)-1, LYNX2 [also known as leukocyte antigen, PLAUR (plasminogen activator, urokinase receptor) domain-containing (LYPD)-2], LY6H, LYPD6, and prostate stem cell antigen (PSCA) (6–10).

When examining the interactions of prototoxins with nAChRs, the complexity of nAChR structure must be taken into consideration. Functional nAChRs are either homomeric or heteromeric pentamers. Functional homomeric nAChRs can be assembled exclusively from α7, α8, or α9 subunits (11, 12), whereas the heteromeric nAChRs are composed of combinations of α2–6 and β2–4 subunits (11, 13). In addition, the α7-nAChR subunit forms heteromeric nAChRs with α8, α5, or β2 subunits, and the α9 forms a heteromeric nAChR with the α10 subunit (14–21). The various subtypes, and even alternate stoichiometries within a given subtype, contribute to different receptor properties, including channel kinetics and conductance, ligand potency, and desensitization rate (11, 20, 22–25). The availability of various combinations of α and β subunits can lead to the expression of multiple nAChR subtypes within the same cell (26–28).

The modulatory effects of prototoxins on nAChR signaling have been studied in a limited number of nAChR subtypes. The first prototoxin to be identified was LYNX1. An increase in induced ACh macroscopic currents is observed when soluble LYNX1 is perfused onto Xenopus laevis oocytes expressing α4β2 or α7 (6); however, when coexpressed with α4β2, LYNX1 increases the rate of desensitization by ACh (29, 30). LYNX1 knockout mice also have greatly enhanced responses to ACh in the habenula (31), and similar effects of LYNX2 have been observed (32).

Because autonomic neurons express a limited number of nAChR subtypes [homomeric α7 and heteromeric α3β4* nAChRs: any combination of α3 with β4, as well as α3 and β4 with α5 (26, 33)], they are a useful system for studying the subtype selectivity of prototoxins. We discovered 3 prototoxins that are expressed in parasympathetic neurons of the Gallus ciliary ganglion. One of them, PSCA, blocks calcium influx through the α7 nAChRs, but not α3-containing nAChR heteromers in ciliary ganglion neurons (10). In addition, expressing PSCA before the period of developmental cell loss rescues neurons from dying (10).

Herein, we report that the other 2 prototoxins in the Gallus ciliary ganglion are LY6E and LYPD6B. Of these 2, LYPD6B is expressed in the nervous system (http://mouse.brain-map.org) (4). Therefore, we hypothesized that, in contrast to PSCA, LYPD6B selectively modulates the function of heteromeric α3β4* nAChRs. Since heteromeric receptors can be composed of pentamers containing 2 or 3 α subunits, we restricted the stoichiometry by expressing pentameric concatemers in X. laevis oocytes (34). These concatemers are composed of all 5 nAChR subunits expressed as a single fusion protein with linkers between each subunit that allow the polypeptide to fold into a functional pentamer (34–36). The linker allows complete control of the subunit stoichiometry and associations to be exercised. Expression studies of nAChR concatemers have faithfully mimicked pharmacological properties of multiple natively expressed nAChR subtypes, including α3β4* nAChRs (34–36). In our study, LYPD6B reduced the EC₅₀ for ACh while also reducing the net whole-cell current induced by ACh on α3β4 nAChRs containing 3 α3 subunits, as opposed to those containing only 2. In addition, LYPD6B reduced the whole-cell current induced by ACh through α3β4* nAChRs containing the α5D subtype, but had no effect on the α5N subtype associated with nicotine dependence.

MATERIALS AND METHODS

Chemicals

All buffer components and pharmacological reagents (ACh and atropine) were purchased from Sigma-Aldrich (St. Louis, MO, USA). Stock drug solutions were made daily and diluted as required.

nAChR concatemers and monomeric α7 plasmids

Pentameric nAChR concatemers were designed and constructed from human nAChR subunits with subunit sequences β4–α3–β4–α3–X (34) or β4–α3–X–β4–α3, where X is β4, α3, α5(D variant), or α5(N variant). These series generated pairs of identical pentamers that varied in the location of the inserted subunit X within the length of the fusion protein. Thus, they served as controls for whether the isoform-selective effects of LYPD6B were consistently observed between the 2 series (e.g., between (α3)₃(β4)₂ generated by β4–α3–β4–α3–α3 or β4–α3–α3–β4–α3). To express the concatemer as a single polypeptide, the Kozac and signal peptide sequences were removed from all subunits, with the exception of the first subunit. In addition, a 40-amino-acid sequence that includes the C-terminal tail of the preceding subunit and alanine-glycine-serine repeats arranged to encode enzyme restriction sites are inserted in the linker region between each subunit. In either series, the first agonist-binding site occurs between the negative face of the first β4 subunit and the positive face of the neighboring α3 subunit (34). Concatemeric receptors and the monomeric α7 subunit were expressed from the pSGEM oocyte high-expression vector (34).

RNA synthesis

Concatemeric plasmids were linearized with NheI for 2 h at 37°C and treated with proteinase K for 30 min at 50°C. The DNA was purified with the Qiaquick PCR purification kit protocol (Qiagen, Limburg, The Netherlands). The synthesis and clean-up of cRNA was achieved by following the protocol from the mMessage mMachine T7 kit (Thermo Scientific–Applied Biosystems, Waltham, MA, USA). Reactions were treated with Turbo DNase (1 U for 15 min at 37°C), and cRNAs were purified by lithium chloride precipitation.

Oocyte preparation and RNA injection

Ready-to-inject X. laevis oocytes were purchased from Ecocyte Bioscience (Austin, TX, USA). The oocytes were stored at 13°C in incubation buffer (82.5 mM NaCl, 2.5 mM KCl, 1 mM MgCl₂·6H₂O, 1 mM CaCl₂·2H₂O, 1 mM Na₂HPO₄, 5 mM 2-[4-(2-hydroxyethyl)piperazin-1-yl]ethanesulfonic acid, 600 μM theophylline, 2.5 mM Na pyruvate, 50 U/ml penicillin, 50 μg/ml streptomycin, 50 μg/ml gentamycin sulfate; pH to 7.5). Injection electrodes were pulled glass micropipettes that were broken to achieve an outer diameter of 40 μm (resistance of 2–6 mΩ) and were used to inject 60 nl containing 20 ng α3β4 or α3β4α5-heteromeric nAChRs cRNA/oocyte, or 10 ng of the monomeric α7 subunit cRNA/oocyte. The monomeric α7 subunit was coinjected with the chaperone protein, resistant to inhibitor of cholinesterase (RIC3), at a 1:50 mass ratio to enhance functional expression (37). Oocyte coinjections with the receptor and the prototoxin LYPD6B were performed at a 1:1 cRNA concentration by weight per oocyte.

Dose and maximal current response recordings of X. laevis oocytes expressing α3β4*-nAChR concatemers and monomeric α7 nAChRs

Recordings were performed 5 d after injection. X. laevis oocytes were voltage clamped at −70 mV with an Axoclamp 900A amplifier (Molecular Devices, Sunnyvale, CA, USA). Recordings were sampled at 10 kHz (low-pass Bessel filter, 40 Hz; high-pass filter, DC), and the resulting traces were saved to disk (Clampex v10.2; Molecular Devices). Data from oocytes with leak currents (I_leak) >50 nA were excluded from the recordings. Agonist (ACh) was applied with a 16-channel, gravity-fed, perfusion system with automated valve control (AutoMate Scientific, Inc., Berkeley, CA, USA). All solutions contained atropine sulfate (1.5 μM) to block muscarinic responses. Oocytes expressing the α7 homomeric or α3β4*-nAChR concatemers were perfused with receptor agonist (ACh) for 5 s, with 60-s washout times between each subsequent application.

Data analysis for logEC₅₀ and maximal current response

LogEC₅₀ and maximal current values (I_max) were determined from individual oocytes. The I_max values were determined from the maximal peak inward current. For logEC₅₀, responses were normalized to the I_max response. Normalized current responses were plotted against ACh concentration, and the logEC₅₀ was determined through nonlinear least-squares curve fitting (Prism 6.0; GraphPad Software, Inc., La Jolla, CA, USA) with unconstrained, monophasic logistic equations used to fit all parameters. LogEC₅₀ values are presented as means ± 95% confidence interval (CI). For determination of LYPD6B’s effect on I_max, responses from oocytes coexpressing the nAChRs and LYPD6B were normalized to control oocytes (oocytes expressing only the nAChRs in question). Normalized logEC₅₀ and I_max responses for oocytes expressing a defined nAChR concatemer averaged with or without coexpression of LYPD6B were averaged, and then tested for significant differences between groups using Student’s t test (Prism 6.0; GraphPad Software, Inc.).

Desensitization rate of α3β4*-nAChR concatemers

Recordings were made 7 d after injection. X. laevis oocytes were voltage clamped, and desensitization kinetics were measured (38). ACh (1.0 mM; corresponding to EC₁₀₀) was applied by using 1 channel from a 16-channel gravity-fed perfusion system (AutoMate Scientific, Inc., Berkeley, CA, USA). Because perfusion rates can influence channel desensitization or recovery from activation, the flow rate of the system (6 ml/min) was continuously monitored and measured between groups. Oocytes expressing concatemers were perfused with receptor agonist ACh for 20–45 min; all α5 containing nAChRs were perfused for ∼20 min, (α3)₃(β4)₂ nAChRs were perfused for ∼25 min, and (α3)₂(β4)₃ nAChRs were perfused for ∼45 min. Perfusion times were determined according to the amount of time required for the entire population of receptors to reach 10% of the maximal response; this allows for an accurate measurement of the desensitization time constant (τ).

Data analysis for current recovery at steady-state desensitization recordings

The rate of desensitization was determined from individual oocytes. Each oocyte’s current responses were observed during prolonged agonist application and normalized to the maximal response in each individual case. Normalized current responses were plotted as a function of time, and desensitization was fitted with a 1- or 2-phase exponential decay (Prism 6.0; GraphPad Software, Inc.). However, there was no significant increase in the quality of fit between the 1- and 2-phase models in any of the measurements taken, either from individual oocytes or for averaged data across multiple oocytes. Therefore, the single-site fit model was chosen to determine τ. The statistical significance of any difference between τ, in the presence or absence of LYPD6B, was determined by Student’s t test (Prism 6.0; GraphPad Software, Inc.).

Oocyte ELISA

After the recordings, oocytes expressing the concatemer alone vs. concatemer+ LYPD6B were incubated on ice for 1 h with a 1:50 dilution of primary antibody (concentrated mAb35 rat anti-AChR (Developmental Studies Hybridoma Bank, Iowa City, IA, USA) in blocking solution (incubation buffer, 10% horse serum, and 0.1% sodium azide). The oocytes were then washed, fixed on ice for an hour with Zamboni’s fixative (0.2 M Na₂HPO₄, 4% paraformaldehyde, 0.2 M NaH₂PO₄, and 10% picric acid; pH 7.2), blocked, and incubated for at least 1 h at room temperature in a 1:250 dilution of biotinylated rat (Vector Laboratories, Burlingame, CA, USA) in modified blocking solution (without sodium azide), washed, and incubated at room temperature for an hour with 1:500 dilution β-galactosidase avidin D (Vector Laboratories). Control oocytes were incubated in modified blocking solution, washed, and incubated at room temperature for 24 h with 100 μM 4-methylumbelliferyl β-d-galactopyranoside (Thermo Scientific–Molecular Probes, Eugene, OR, USA). The reaction was stopped by adding 15 μl of 10 M NaOH. Fluorescence was measured at excitation and emission of 450 and 350 nm, respectively, with a Synergy H4 Hybrid Reader (Biotek, Winooski, VT, USA).

Data analysis for oocyte ELISA

The average relative fluorescence values per group were determined from each oocyte fluorescence value normalized to the average fluorescence of control oocytes (incubated without β-galactosidase avidin D). Student’s t test (Prism 6.0; GraphPad Software, Inc.) was used to test for statistically significant differences between the average of relative fluorescence values of oocytes expressing the concatemer alone vs. concatemer + LYPD6B.

RESULTS

Identifying prototoxins in Gallus parasympathetic neurons

In our previously reported study, 6 expressed sequence tag sequences similar to the structure of Mus lynx1 were identified and 3 (Ch6ly, Ch3ly, and Ch5ly) were found to be expressed in the embryonic ciliary ganglion (10). Ch6ly was identified as PSCA (10). Before undertaking the present study, we determined by BLAST (blast.ncbi.nlm.nih.gov/Blast.cgi) that Ch3ly is LY6E and Ch5ly is LYPD6B. Gallus LYPD6B and LY6E share a 72 and 44% amino acid identity, respectively, with the Homo sapiens LYPD6B and LY6E (Fig. 1A, B). When translated, both sequences have the canonical characteristics of a prototoxin protein: conserved 8–10 cysteine residues with defined disulfide bonding pattern that forms the 3-fingered tertiary structure similar to the toxin α-bungarotoxin (39). Because it is expressed in the brain, we focused our attention on LYPD6B (4). LYPD6B is located on chromosome 2 and is also known as LYPD7 (Fig. 1C). Although LYPD6B and LYPD6 are similar in name, they are encoded by different genes and are not splice variants of each other (Fig. 1C).

Figure 1.

Alignment of amino acid sequences for LYPD6B from Gallus gallus and H. sapiens. A) Mature protein alignment of the G. gallus LYPD6B (GenBank accession number XM_422156), and H. sapiens prototoxin LYPD6B (GenBank accession number XM_006712283). B) Mature protein alignment of the G. gallus LY6E (GenBank accession number NM_204775), and H. sapiens prototoxin LY6E (GenBank accession number NM_002346). Asterisks indicate consensus between the amino acids and the shaded cysteine residues is necessary for a 3-fingered motif secondary structure. C) A cartoon representation of human chromosome 2 highlighting gene location of LYPD6B and LYPD6 (www.ensemble.org).

Enhancement of ACh sensitivity by LYPD6B of specific concatemers

We tested the effects of LYPD6B on 4 different α3β4-nAChR concatemers expressed in X. laevis oocytes using the 2-electrode voltage clamp method to measure responses to ACh: 2 nAChRs with a composition of (α3)₂(β4)₃, having concatemeric sequences of β4–α3–β4–β4–α3 and β4–α3–β4–α3–β4, and 2 nAChRs with a composition of (α3)₃(β4)₂, having concatemeric sequences of β4–α3–α3–β4–α3 and β4–α3–β4–α3–α3. The current responses of X. laevis oocytes expressing each α3β4 nAChR concatemer were compared to those expressing the same concatemer in the presence of LYPD6B. The responses to different concentrations of ACh for the α3β4 heteromers are illustrated in Fig. 2. The EC₅₀ of ACh was not affected by LYPD6B in (α3)₂(β4)₃ concatemers: β4–α3–β4–β4–α3 and β4–α3–β4–α3–β4 (Fig. 2A, B). In contrast, LYPD6B decreased the EC₅₀ of ACh for the (α3)₃(β4)₂ concatemers, β4–α3–α3–β4–α3 and β4–α3–β4–α3–α3 (Fig. 2C, D), from 631 μM [95% CI, 603–832 μM) to 79 μM (95% CI, 66–98 μM) and 631μM (95% CI, 631–708 μM) to 79 μM (95% CI, 85–105 μM) for β4–α3–α3–β4–α3 and β4–α3–β4–α3–α3, respectively]. The differences in EC₅₀ caused by the presence of LYPD6B were similar, regardless of whether the responses to ACh were graphed as raw (I) or normalized (I/I_max) data.

Figure 2.

LYPD6B decreases ACh logEC₅₀ values for (α3)₃(β4)₂ heteromeric nAChRs. Oocytes expressing the indicated concatemers were tested with and without a 1:1 (w/w) RNA concentration ratio of nAChR to LYPD6B and whole-cell currents were monitored with 2-electrode voltage clamp after perfusion of different concentrations of ACh. Responses are normalized to I_max for each concatemer. Open circles are recordings collected from oocytes expressing the receptor alone. Closed circles are recordings collected from oocytes coexpressing the concatemer and LYPD6B. The pie diagrams for each concatemer tested indicate the locations of the amino and carboxy ends of the fusion protein and the orientation of the positive and negative faces of the α and β subunits. Black teardrops indicate ligand-binding sites. For each concatemer, at least 3 groups of independently injected oocytes were tested and all concatemers were injected and tested in each round. A) β4–α3–β4–β4–α3; n = 11 (alone), n = 10 (with LYPD6B); [ACh] tested = 10⁻⁶–10^−2.5 M. B) β4–α3–β4–α3–β4; n = 12 (alone), n = 13 (with LYPD6B); [ACh] = 10⁻⁶–10^−2.5. C) β4–α3–α3–β4–α3; n = 5 (alone; with LYPD6B); [ACh] = 10^−5.5–10⁻² M. D) β4–α3–β4–α3–α3; n = 17 (alone), n = 18 (with LYPD6B); [ACh] tested = 10^−5.5–10⁻² M. Data points represent means ± se, and the drug potency differences between the receptor vs. the receptor and LYPD6B groups were analyzed using a 2-tailed Student’s t test.

In addition to α3β4, autonomic neurons express α3β4α5 nAChRs (28). Genome-wide association studies on humans have identified a rare allele of the gene encoding the subunit cholinergic receptor, nicotinic, α5 (CHRNA5) that is associated with nicotine dependence in cigarette smokers (40, 41). The common allele of CHRNA5 encodes the α5D variant, and the allele associated with nicotine dependence encodes the α5N variant. Therefore, we tested the effects of LYPD6B on 4 concatemers that encode α3β4α5 nAChRs. Two of the α3β4α5 nAChRs have a composition of (α3)₂(β4)₃(α5D) with sequences of β4–α3–α5D–β4–α3 and β4–α3–β4–α3–α5D, and the other 2 are (α3)₂(β4)₃(α5N), with sequences of β4–α3–α5N–β4–α3 and β4–α3–β4–α3–α5N. The responses to ACh of X. laevis oocytes expressing α3β4α5 nAChR concatemers were compared to those coexpressed with LYPD6B (Fig. 3). Regardless of the presence of the α5D variant or the α5N variant, the coexpression of LYPD6B did not alter the EC₅₀ of ACh of the 4 different stoichiometries tested (Fig. 3; Table 1).

Figure 3.

LYPD6B does not affect the ACh logEC₅₀ for α3β4α5 heteromeric nAChRs. Oocytes expressing the indicated concatemers were tested alone and with a 1:1 (w/w) RNA nAChR:LYPD6B concentration ratio. Whole-cell currents were monitored with a 2-electrode voltage clamp after perfusion of different concentrations of ACh. Responses were normalized to I_max for each concatemer. Open circles: recordings collected from oocytes expressing the receptor alone. Closed circles: recordings collected from oocytes coexpressing the concatemer and LYPD6B. Pie diagrams of each concatemer are as previously described in the legend for Fig. 2. A) β4–α3–α5D–β4–α3; n = 9 (with and without LYPD6B); [ACh] = 10^−5.5 to 10⁻² M. B) β4–α3–α5N–β4−α3; n = 9 (alone), n = 8 (with LYPD6B); [ACh] = 10^−5.5 to 10⁻² M. C) β4–α3–β4–α35D; n = 9 (alone and with LYPD6B); [ACh] = 10^−5.5–10⁻² M. D) β4–α3–β4–α3–α5N-; n = 7 (alone), 9 (with LYPD6B); [ACh] = 10^−5.5 to 10⁻² M. Data points represent means ± se, and the drug potency differences between the receptor versus the receptor and LYPD6B groups were analyzed using a 2-tailed Student’s t test.

TABLE 1.

The effects of LYPD6B on functional characteristics of nAChR concatemers

nAChR concatemer	ACh logEC50		Imax, normalized (%)		τ, normalized (s)
	−LYPD6B	+LYPD6B	−LYPD6B	+LYPD6B	−LYPD6B	+LYPD6B
β4α3α3β4α3	−3.2 ± 0.03 (10)	−4.1 ± 0.01 (14)***	100 (17)	59 ± 3.0 (15)***	14 ± 1.1 (12)	10 ± 0.9 (9)*
β4α3β4β4α3	−4.5 ± 0.01 (11)	−4.3 ± 0.01 (10)	100 (10)	100 ± 10.4 (10)	8 ± 2.1 (5)	12 ± 2.6 (6)
β4α3α5Dβ4α3	−3.9 ± 0.01(8)	−4.0 ± 0.01 (9)	100 (15)	63 ± 6.9 (19)**	11 ± 0.6 (8)	13 ± 1.4 (6)
β4α3α5Nβ4α3	−3.8 ± 0.01 (9)	−3.8 ± 0.02 (8)	100 (12)	68 ± 16.5 (14)	9 ± 0.7 (8)	9 ± 0.4 (7)
β4α3β4α3α3	−3.2 ± 0.01 (17)	−4.0 ± 0.02 (18)***	100 (16)	32 ± 7.9 (18)***	27 ± 2.1 (9)	19 ± 1.9 (6)*
β4α3β4α3β4	−4.4 ± 0.01 (12)	−4.2 ± 0.01 (13)	100 (8)	106 ± 5.9 (9)	6 ± 0.7 (8)	8 ± 0.7 (7)
β4α3β4α3α5	−3.9 ± 0.01 (9)	−4.0 ± 0.02 (9)	100 (13)	32 ± 3.3 (12)***	11 ± 1.3 (8)	12 ± 0.6 (11)
β4α3β4α3α5N	−3.9 ± 0.01 (7)	−4.0 ± 0.02 (9)	100 (9)	83 ± 18.8 (15)	11 ± 0.8 (11)	11 ± 0.8 (11)
(α7)	-3.6 ± 0.04 (6)	−3.6 ± 0.03 (6)	100 (6)	116 ± 10.8 (6)	ND

The indicated parameters were measured in X. laevis oocytes expressing the indicated nAChR concatemers, with and without coexpression of LYPD6B, by the 2-electrode voltage-clamp method after applying ACh by perfusion. Data are presented as means ± se, with the number of individual oocytes tested in parentheses (n). The α7 nAChRs were formed from mRNA encoding the CHRNA7 monomer, rather than produced as a concatemer. Student’s t test was performed to determine statistical significant differences between concatemeric nAChR function in the presence or absence of LYPD6B, *P ≤ 0.05; **P ≤ 0.01; ***P ≤ 0.001.

Because autonomic neurons prominently express homomeric α7 nAChRs, in addition to heteromeric α3β4* nAChRs (28, 42), we tested the effects of LYPD6B on α7 nAChRs (Fig. 4). The homomeric α7-nAChR requires the chaperone protein RIC3, to traffic to the cell surface (37, 43); therefore, X. laevis oocytes were coinjected with either mRNAs encoding the chaperone protein RIC3 or the α7-nAChR subunit monomer at a 50:1 ratio, or with RIC3, α7, and LYPD6B at a 50:1:1 concentration ratio. The ACh EC₅₀ for α7-nAChR homomers is not affected by the presence of LYPD6B (Fig. 4; Table 1).

Figure 4.

LYPD6B does not affect the ACh logEC₅₀ for α7 homomeric nAChRs. Oocytes were injected with mRNA encoding the monomeric α7 nAChR subunit, chaperone protein RIC3, and LYPD6B at a 1:50:1 ratio by weight. Open circles: recordings collected from oocytes expressing the receptor alone. Closed circles: recordings collected from oocytes coexpressing α7 and LYPD6B. The pie diagram is described in the legend for Fig. 2. n = 6 oocytes (alone), 6 (with LYPD6B); [ACh] = 10^−5.5–10⁻² M. Data points represent means ± se. Drug potency differences between the receptor vs. the receptor and LYPD6B groups were analyzed with a 2-tailed Student’s t test.

Preferential modulation by LYPD6B of I_max induced by ACh in specific concatemers

The second functional component tested was the I_max response to ACh. LYPD6B significantly decreased the I_max of both (α3)₃(β4)₂ nAChR concatemers (β4–α3–α3–β4–α3 and β4–α3–β4–α3–α3) by ∼30 and 60%, respectively (Fig. 5). LYPD6B did not alter the I_max for either of the 2 (α3)₂(β4)₃ nAChR concatemers or that of the homomeric α7 nAChRs. These differences in I_max cannot be attributed to differences between the number of pentamers expressed on the surface of the oocytes, because we did not detect any differences in cell surface mAb35 binding to oocytes expressing these constructs in the presence or absence of LYPD6B (Supplemental Fig. 1).

Figure 5.

LYPD6B decreases the ACh-induced maximal current response for (α3)₃(β4)₂ concatemers. For each indicated concatemer, averaged recordings from oocytes at each ACh concentration are shown together with a pie diagram of the structure of the concatemer and the aggregated data is plotted as a histogram comparing the responses of the concatemer alone vs. the concatemer coexpressed with LYPD6B. The features of the pie diagram are described in the legend for Fig. 2. A) Recordings from β4–α3–β4–β4–α3 (n = 6 alone and with LYPD6B); [ACh] = 10⁻⁶–10^−2.25 M. A1) Mean normalized currents for β4–α3–β4–β4–α3 alone (n = 10) and with LYPD6B (n = 10). B) Recordings from β4–α3–β4–α3–β4 (n = 6 alone and with LYPD6B); [ACh] = 10⁻⁶–10^−2.25 M. B1) Mean normalized currents for β4–α3–β4–α3–β4 (n = 8 alone and n = 9 with LYPD6B) C) Recordings from β4–α3–α3–β4–α3 (n = 6 alone and with LYPD6B); [ACh] = 10^−5.5 to 10⁻² M. C1) Mean normalized currents for β4–α3–α3–β4–α3 (n = 17 alone and n = 15 with LYPD6B). The oocytes coexpressing the LYPD6B exhibit about a 30% decrease in I_max. ***P < 0.01, 2-tailed Student’s t test. D) Recordings from β4–α3–β4–α3–α3 (n = 6 alone and with LYPD6B); [ACh] = 10^−5.5–10⁻² M. D1) Mean normalized currents for β4–α3–β4–α3–α3 (n = 16 alone and n = 18 with LYPD6B). The oocytes coexpressing the LYPD6B and β4–α3–β4–α3–α3 exhibit about a 60% decrease induced ACh maximal response P < 0.001. E) α7–α7–α7–α7–α7 receptor at ACh doses (10^−5.5–10⁻² M) (n = 6 alone and with LYPD6B). ***P < 0.001. E1) α7 Homomer alone and with LYPD6B (n = 6 alone and with LYPD6B).

LYPD6B decreased the I_max for the β4–α3–α5D–β4–α3 and β4–α3–β4–α3–α5D nAChRs by ∼30 and ∼60%, respectively (Fig. 6). There was no change in I_max for either format of α3β4α5Ν concatemers tested. Thus, LYPD6B selectively reduces the I_max of the α5D variant containing α3β4 nAChRs compared to those containing the α5N variant associated with heavy smoking.Preferential modulation by LYPD6B of desensitization rate to ACh in specific concatemers. The third functional component tested was the desensitization rate of the nAChRs. For each nAChR, the desensitization τ was measured during an ACh perfusion period, long enough for the inward current to reach 10% of its initial response. LYPD6B decreased τ for desensitization of only the (α3)₃(β4)₂ concatemers β4–α3–α3–β4–α3 and β4–α3–β4–α3–α3 (Fig. 7; Table 1), indicating an increase in the desensitization rate of the receptors. LYPD6B did not change the desensitization rates to ACh to (α3)₂(β4)₃ or any of the α5D- or α5N-containing α3β4 heteromers (Fig. 8). Note that our perfusion system is capable of exchanging solutions on a subsecond timescale (38), at least 1 order of magnitude faster than the measured τ. This capability ensures that α3β4*-nAChR desensitization effects can be measured accurately because they (and not solution exchange effects) are by far the predominant contributors to the observed τ. For the same reason, it was not possible to measure accurately the desensitization rate for the homomeric α7 nAChR in our system (α7-nAChR desensitization occurs much more rapidly than solution exchange) (44).

Figure 6.

LYPD6B affect the ACh induced maximal current response for α3β4α5D nAChRs. For each indicated concatemer, averaged recordings from oocytes at each ACh concentration are shown together with a pie diagram of the structure of the concatemer, and the aggregated data are plotted as a histogram comparing the responses of the concatemer alone vs. the concatemer coexpressed with LYPD6B. The features of the pie diagram are described in the legend for Fig. 2. A) β4–α3–α5D–β4–α3 (n = 6 alone and with LYPD6B); [ACh] = 10^−5.5–10⁻² M. A1) Mean normalized currents for β4–α3–α5D–β4–α3. The oocytes coexpressing LYPD6B exhibit about a 30% decrease in I_max over those without LYPD6B; P < 0.01, Student’s t test. B) β4–α3–α5N–β4–α3 (n = 6 alone and with LYPD6B); [ACh] = 10^−5.5–10⁻² M. B1) Mean normalized currents for β4–α3–α5N–β4–α3 (n = 12 alone and n = 14 with LYPD6B). C) β4–α3–β4–α3–α5D (n = 6 alone and with LYPD6B); [ACh] = 10^−5.5–10⁻² M. C1) Mean normalized currents for β4–α3–β4–α3–α5D (n = 13 alone and n = 12 with LYPD6B). The oocytes coexpressing the LYPD6B and β4–α3–β4–α3–α5D exhibit an ∼60% decrease induced ACh maximal response; P < 0.001, 2-tailed Student’s t test. D) β4–α3–β4–α3–α5N (n = 6 alone and with LYPD6B); [ACh] = 10^−5.5–10⁻² M. D1) Mean normalized currents for β4–α3–β4–α3–α5N (n = 9 alone and n = 15 with LYPD6B). **P < 0.1; ***P < 0.001.

Figure 7.

LYPD6B decreases the time constant (τ) for steady-state desensitization for (α3)₃(β4)₂ concatemers. A) Graphs of the averaged normalized time in seconds vs. the averaged normalized ACh current response per α3β4* nAChR subtype with and without LYPD6B. Black background: expression of all data curves of the concatemer subtype alone; gray background: concatemer+LYPD6B; white lines: curve fits. B) Aggregated data for all groups are means ± se. Black bar: concatemer alone; gray bar: concatemer+LYPD6B. The β4–α3–α3–β4–α3 and β4–α3–β4–α3–α3 concatemers show a statistically significant decrease in τ when LYPD6B is present, indicating that LYPD6B enhances the rate of desensitization. The differences between each concatemer vs. the concatemer+LYPD6B was analyzed with a 2-tailed Student’s t test (n = 5–12 oocytes per group; Table 1). *P < 0.05.

Figure 8.

LYPD6B does not affect the time constant (τ) for steady-state desensitization of α3β4α5 concatemers. A) Graphs of the averaged normalized time in seconds vs. the averaged normalized ACh current response per α3β4α5 nAChR subtype with and without LYPD6B. Black background: concatemer alone; gray background: concatemer + LYPD6B; white lines: curve fits. B) Aggregated data for all groups are means ± se. Black bar: concatemer alone; gray bar: concatemer + LYPD6B. The differences between the receptor and the receptor and LYPD6B groups were analyzed with 2-tailed Student’s t test (n = 6–11 oocytes per group; Table 1). There was no statistically significant difference among the time constants.

DISCUSSION

In this study, LYPD6B was identified as a prototoxin that modulates the function of α3β4* heteromeric, but not α7 homomeric, nAChRs. Further, the effects of LYPD6B were dependent on the particular subunit composition and stoichiometry of α3β4 containing nAChRs. For (α3)₃(β4)₂ nAChRs, LYPD6B enhanced the sensitivity to ACh, yet decreased the I_max and increased the rate of desensitization to ACh, whereas these effects are not seen with (α3)₂(β4)₃ nAChRs. When the α5 subunit was introduced, LYPD6B decreased the I_max induced by ACh only when the α5D variant was present, but had no effect when the α5N variant associated with heavy smoking was included. LYPD6B had no apparent effect on ACh responsiveness of α7-subunit–containing homomers. This demonstrates that the effects of LYPD6B on α3β4*-nAChR function are highly subtype and stoichiometry selective and illustrates a previously unappreciated complexity in prototoxin modulation of nAChRs.

In comparison to the 2 best-studied prototoxins to date [LYNX1 and -2 (LYPD1)], in our study, LYPD6B exhibited a much higher selectivity and complexity in modulating nicotinic signaling. LYNX1 and -2 decreased the sensitivity to ACh and enhanced the rate of desensitization of both α7 homomeric and α4β2 heteromeric nAChRs in X. laevis oocytes (31, 32), there by acting as a “brake” on nicotinic signaling. This finding is supported by studies of the LYNX1 knockout mouse where nicotinic responses were enhanced in the habenula (31). LYNX1 also colocalizes and coimmunoprecipitates with homomeric α7 nAChRs in the amygdala, CA3/CA1 hippocampal neurons, and thalamic reticular nuclei (29). In contrast, LYPD6B enhanced ACh sensitivity, yet also decreases the I_max to ACh on only low-sensitivity α3β4 pentamers containing 3 α3 subunits and had no detectable effect on α7-containing homomers. LYNX1 facilitates the assembly and trafficking of the low-sensitivity (α4)₃(β2)₂ nAChRs (45); such trafficking may explain the observed increase in ACh EC₅₀ (ACh decreased the sensitivity) of α4β2 nAChRs when expressed from subunit monomers and coexpressed with LYNX1 (29). However, in our study, differences in nAChR subunit assembly cannot explain the enhanced ACh sensitivity of (α3)₃(β4)₂, because of the fixed stoichiometry of our concatemeric constructs. As a result, changes in EC₅₀ in our system must have been caused by a direct allosteric effect of LYPD6B on the function of (α3)₃(β4)₂ nAChR.

The differential responses of (α3)₃(β4)₂- and (α3)₂(β4)₃-nAChR concatemers to LYPD6B suggest that the additional α3–α3 interface provides a structural site for LYPD6B allosterism. These 2 types of stoichiometries have been identified for both α3β4 and α4β2 heteromeric nAChRs (22, 24) with distinct pharmacologic properties, including a higher sensitivity to ACh for the heteromers that contain 2 α nAChR subunits (38, 46, 47). This concept of α–α interface sites for allosteric modulation is supported by the observation of modulation by divalent cations at an α4–α4 interface (46). When LYPD6B is coexpressed with the less sensitive (α3)₃(β4)₂, the EC₅₀ to ACh becomes comparable to that of the more sensitive (α3)₂(β4)₃. With α4β2 heteromeric nAChRs, the additional α4–α4 interface is thought to act equivalently to a coagonist site by producing a second phase of low-affinity function when it is engaged by high agonist concentration (38, 47). Thus, if the α3–α3 interface parallels the behavior of the α4–α4 interface, then LYPD6B may be influencing the interaction of agonist at the additional α–α interface. Studies similar to ours that constrain stoichiometry of α4β2 nAChRs, while examining interaction with LYNX1 and -2, have yet to be performed. Prototoxin proteins have a protein structure similar to that of α-bungarotoxin; thus, it is assumed that they interact with nAChRs near the agonist-binding site (39, 48); however, this behavior may differ among prototoxins based on structural characteristics such as the length of each of the 3 fingers and according to the nAChR subtypes with which they interact. It is also worth noting that many features of conventionally recognized agonist-binding interfaces are conserved at nonagonist-binding interfaces (38, 46, 47), so prototoxin interaction sites on nAChRs could be located at any subunit pair interface.

Our results showing that LYPD6B decreased the maximum-induced ACh current response only in concatemers containing the α5D variant and not the α5N variant suggest a possible importance in nicotine dependence. Of the many SNPs identified in CHRNA5 through a genome-wide association study of nicotine dependence, only one was a variant in the coding region (40, 41, 49). This SNP replaces the 398th amino acid, aspartic acid (D), with an asparagine (N) amino acid in the intracellular loop of the nAChR subunit (40, 49). The amino acid sequences surrounding the SNP are conserved across 8 species, highlighting the importance of this region (50). The α5 subunit is known to modulate the properties of nAChRs in which it is incorporated (34, 51), and α5 knockout mice and α5 knockdown rats exhibit reduced aversion to nicotine (52, 53). Our study confirmed our previous finding that nAChRs containing the α5N variant exhibit a decreased response to ACh when compared to those with the α5D variant. However, LYPD6B decreased the I_max to ACh only in α3β4 heteromers containing the normal α5D variant, as opposed to the α5N variant associated with nicotine dependence. That such a dramatic difference was seen in the effects of LYPD6B on α3β4 (α5D vs. α5N) suggests that individuals expressing the α5N variant lose an important mechanism for modulation of α3β4α5-nAChR function. It is clear that considerably more studies of nAChRs and their associated prototoxins in pathways regulating nicotine reward and aversion must be pursued.

Because nicotinic signaling has central functions in guiding behavior, nAChRs have adapted a variety of means to fine tune their signaling. First, cells can regulate the subtypes of nAChRs that are expressed and where they are expressed. Second, neurons can express various nAChR subunit stoichiometries within the same overall nAChR subtypes, which further affect receptor properties. Third, prototoxin proteins belonging to the Ly6-uPAR family can modulate nAChR responsiveness. The data presented in this article identify LYPD6B as a prototoxin that modulates a specific stoichiometry of α3β4-nAChR subtype and the responsiveness of nAChRs containing only a specific variant of α5, serving as another means of fine tuning nicotinic signaling. The exquisite selectivity for subtype, and even stoichiometry, of LYPD6B modulation of nAChR activity, together with the diversity of effects observed in this study, indicates that the nAChR-regulatory roles of at least some prototoxins are likely to be complex and highly specialized.

Glossary

ACh

acetylcholine

CHRNA5

cholinergic receptor, nicotinic, α5

CI

confidence interval

I_max

maximal current

LY

leukocyte antigen

Ly6/uPAR

Ly-6/urokinase plasminogen activator receptor

LYPD

leukocyte antigen, PLAUR (plasminogen activator, urokinase receptor) domain-containing

LYNX

Ly-6/neurotoxin

nAChR

nicotinic acetylcholine receptor

PSCA

prostate stem cell antigen

RIC

resistant to inhibitor of cholinesterase

SNP

single-nucleotide polymorphism