Analysis of rare variations reveals roles of amino acid residues in the N-terminal extracellular domain of nicotinic acetylcholine receptor (nAChR) alpha6 subunit in the functional expression of human alpha6*-nAChRs

Background Functional heterologous expression of naturally-expressed and apparently functional mammalian α6*-nicotinic acetylcholine receptors (nAChRs; where ‘*’ indicates presence of additional subunits) has been difficult. Here we wanted to investigate the role of N-terminal domain (NTD) residues of human (h) nAChR α6 subunit in the functional expression of hα6*-nAChRs. To this end, instead of adopting random mutagenesis as a tool, we used 15 NTD rare variations (i.e., Ser43Pro, Asn46Lys, Asp57Asn, Arg87Cys, Asp92Glu, Arg96His, Glu101Lys, Ala112Val, Ser156Arg, Asn171Lys, Ala184Asp, Asp199Tyr, Asn203Thr, Ile226Thr and Ser233Cys) in nAChR hα6 subunit to probe for their effect on the functional expression of hα6*-nAChRs. Results N-terminal α-helix (Asp57); complementary face/inner β-fold (Arg87 or Asp92) and principal face/outer β-fold (Ser156 or Asn171) residues in the hα6 subunit are crucial for functional expression of the hα6*-nAChRs as variations in these residues reduce or abrogate the function of hα6hβ2*-, hα6hβ4- and hα6hβ4hβ3-nAChRs. While variations at residues Ser43 or Asn46 (both in N-terminal α-helix) in hα6 subunit reduce hα6hβ2*-nAChRs function those at residues Arg96 (β2-β3 loop), Asp199 (loop F) or Ser233 (β10-strand) increase hα6hβ2*-nAChR function. Similarly substitution of NTD α-helix (Asn46), loop F (Asp199), loop A (Ala112), loop B (Ala184), or loop C (Ile226) residues in hα6 subunit increase the function of hα6hβ4-nAChRs. All other variations in hα6 subunit do not affect the function of hα6hβ2*- and hα6hβ4*-nAChRs. Incorporation of nAChR hβ3 subunits always increase the function of wild-type or variant hα6hβ4-nAChRs except for those of hα6(D57N, S156R, R87C or N171K)hβ4-nAChRs. It appears Asp57Lys, Ser156Arg or Asn171Lys variations in hα6 subunit drive the hα6hβ4hβ3-nAChRs into a nonfunctional state as at spontaneously open hα6(D57N, S156R or N171K)hβ4hβ3V9’S-nAChRs (V9’S; transmembrane II 9’ valine-to-serine mutation) agonists act as antagonists. Agonist sensitivity of hα6hβ4- and/or hα6hβ4hβ3-nAChRs is nominally increased due to Arg96His, Ala184Asp, Asp199Tyr or Ser233Cys variation in hα6 subunit. Conclusions Hence investigating functional consequences of natural variations in nAChR hα6 subunit we have discovered additional bases for cell surface functional expression of various subtypes of hα6*-nAChRs. Variations (Asp57Asn, Arg87Cys, Asp92Glu, Ser156Arg or Asn171Lys) in hα6 subunit that compromise hα6*-nAChR function are expected to contribute to individual differences in responses to smoked nicotine.


Background
Mammalian neuronal nicotinic acetylcholine receptors (nAChRs) are formed out of six α subunits (α2-α7) and three β subunits (β2-β4) [1] where in five subunits either α (e.g., α7, α9α10; etc.) or α plus β come together (e.g., α4β2, α4β2α5, α6β2, α6β2β3, α6β4, α7β2; etc.) to form a ligand gated ion channel [1]. Each nAChR subunit has a large extracellular N-terminal domain (E1 or NTD), followed by four transmembrane domains (TM I, II, III and IV) and a small C-terminal domain (CTD). TM domains are connected to each other via loops: a small intracellular loop (C1) connects TM I to TM II, an extracellular loop (E2) connects TM II to TM III and a large intracellular loop (C2) connects TM III to TM IV. The NTD of a human nAChR subunit is presumed to contain an inner β-sheet (composed of strands β1-β3, β5, β6 and β8) and outer β-sheet (composed of strands β4, β7, β9 and β10) like those typically seen in the crystal structure of Torpedo muscle nAChR or other eukaryotic and prokaryotic ACh binding protein subunits [2][3][4][5]. The β-strands connect to each other via loops: they are known as loops A, B and C in the principal or positive (+) face and loops D, E and F in the complementary or negative (−) face (Figures 1 and 2). These loops and additional residues in the NTD of participating α and β subunits form the ligand-binding site for nAChRs. The β6-β7 loop has a pair of disulfide-bonded cysteines separated by 13 residues that form the cysteine-loop (i.e., cys-loop) motif and it is essential for nAChR assembly and channel gating [6,7].
Some studies suggest that β3 subunits may facilitate α6*-nAChR trafficking to the cell membrane [35]. However, others indicate that coexpression of wild-type (WT) nAChR β3 subunits has a dominant-negative effect on function of α6*-nAChRs [25,26,28]. The dominant negative effect of nAChR β3 subunits could be overcome by coexpression with mutant β3 V9'S subunits [i.e., valine(V)-to-serine(S) mutation at 9' position in the putative second transmembrane domain] [25,26]. Upon further investigation it has been shown that amino acid (AA) residues in the NTD of α6 subunits influence the effects of the WT or mutant β3 subunits (i.e., β3 V9'S ) and are involved in gain-of-function effects of mutant β3 V9'S subunits on α6*-nAChR function [26]. These findings also suggest that co-assembly of nAChR β3 subunits with α6 plus β2 or β4 subunits to form functional nAChR is determined to some degree by the α6 subunit N-terminal, extracellular region. This is because improved success in functional expression of α6*-nAChRs is obtained when chimeric mouse-human α6 subunits (where the Nterminal domain of mouse α6 subunit is fused to the rest of human α6 subunits) are employed instead of WT hα6 subunit. This has led to the discovery of Asn143 in the putative loop E region of the NTD of hα6 subunit as a key determinant for the heterologous functional expression of α6*-nAChRs [26][27][28].
Therefore it appears that there probably exist additional residues in the NTD of hα6 subunit some of which will be of crucial importance and mutations in them will not be tolerated, and others which could be target for improved functional expression of α6*-nAChRs. In the absence of these information (and not to use random mutagenesis as a tool towards this end) we reasoned to use naturally occurring rare variations in the NTD of hα6 subunit as probes to uncover N-terminal molecular bases important for functional expression of hα6hβ2*-and hα6hβ4*-nAChRs. To this end we evaluated the effect of 15 naturally occurring rare missense mutations, that occur in various regions of the N-terminal extracellular domain of hα6 subunit [N-terminal α-helix: Ser43Pro, Asn46Lys and Asp57Asn; complementary face/inner β-fold: Arg87Cys (β2-strand); Asp92Glu, Arg96His and Glu101Lys (loop D/ β2-β3 loop); Ser156Arg (β6-strand); and Asp199Tyr and Asn203Thr (loop F)] and principal face/outer β-fold: Ala112Val (loop A), Asn171Lys (cysteine loop, potential site for N-glycosylation), Ala184Asp (loop B), Ile226Thr (loop C) and Ser233Cys (β10-strand); see Figures 1 and 2;  Table 1], on the functional expression of hα6*-nAChRs. On occasion we took aid of gain-of-function hβ3 subunits (i.e., hβ3 V9'S = hβ3 V273S ) (to increase receptor sensitivity) and a codon optimized nAChR hβ2 subunit (to increase receptor expression efficiency) to provide additional insight into the effect of these natural variations on the functional expression of hα6*-nAChRs. Our results indicate that some of the rare variations abrogate, decrease, increase or do not affect the functional expression of hα6hβ2*-and hα6hβ4*-nAChRs. There also appear to be a subtype specific effect of some of these variations. By undertaking this study, as anticipated, we were able to uncover some of the N-terminal molecular bases in hα6 subunit that could be taken advantage for modulating functional expression of hα6hβ2*-, hα6hβ4-and hα6hβ4hβ3-nAChRs. These results provide foundations for undertaking more specialized and individual mutation specific studies later on.

Results
Bioinformatic analyses indicate possible functional consequences of rare variations in nAChR α6 subunits Information retrieved from NCBI dbSNP database, NHLBI Grand Opportunity Exome Sequencing Project (ESP), multiple protein sequence alignment of human nAChR α subunits, multiple protein sequence alignment of nAChR α6 subunits from various model organisms and homology model of nAChR hα6 subunit (Figures 1 and 2) are combined to present an overview of the characteristics of the 15 single nucleotide variations (SNVs) evaluated for their effect on function of hα6*-nAChRs (Table 1). African American and European American populations sampled in the ESP indicate that these nucleotide variations are rare and their combined frequency ranges from 0.001 to 0.3% (Table 1).
Full length protein sequence alignments of human nAChR α subunits indicate that fully conserved (Asp 92 and Ser 156 ), strongly conserved (Arg 87 and Asn 171 ), weakly conserved (Asp 57 , Asp 199 and Asn 203 ), and non-conserved (the rest 8: Ser 43 Figure 1B]. Together, these results indicate that nAChR hα6 subunit AAs at positions 87 (Arg), 92 (Asp), 156 (Ser), and 171 (Asn) are conserved in both human nAChR α subunits and nAChR α6 subunits from other organisms. Lacking a priori knowledge, we hypothesize that these variations at conserved AAs would have an effect on the structure and/or function of hα6*-nAChRs.
Positions of the WT or variant AAs in nAChR hα6 subunit are mapped to their secondary structural features such as α-helices, β-strands and loops (Figures 1 and 2; Table 1) by sequence alignment and/or comparison to other nAChR subunits and most prominently to that of the muscle nAChR α subunits of Torpedo marmorata (PDB code: 2BG9.A) [2]. We hypothesized that some of these variations in nAChR hα6 subunit occurring in the loop A  (Figure 1), conservation in nAChR α6 subunit of other species and putative locations in the secondary structure of the nAChR hα6 subunits ( Figure 2) are presented. The WT AAs with the exception of S43 are strongly conserved in nAChR α6 subunits studied from a limited number of other species (Figure 1). Please note that hα6 variations R87C, S156R, D199Y and S233C retrieved from ESP do not have an rs ID number yet. The N-terminal domain of a typical human nAChR subunit is presumed to contain an inner β-sheet (strand β1-β3, β5, β6 and β8) and outer β-sheet (strand β4, β7, β9 and β10) like those of seen in the crystal structure of Torpedo muscle nAChR subunits. '-' indicates lack of indicated information in these positions.
(Ala112Val), loop B (Ala184Asp), loop C (Ile226Thr), loop D/β2-β3 loop (Arg87Cys, Asp92Glu and Arg96His), loop F (Asp199Tyr and Asn203Thr) and cysteine-loop (Asn171Lys) would affect the structure and/or function of hα6*-nAChR as these regions are known to be important in subunit assembly, ligand binding and/or signal transduction of various other subtypes of nAChRs [2,7]. Also change in the electrical properties and other characteristics of the AAs as a result of the variation potentially could impact the intra-and/or inter subunit molecular interactions involving ionic and other types bonds that may affect the structure and/or function of hα6*-nAChRs (Tables 1 and 2).
Current responses are null whether WT nAChR hα6 subunits are expressed in oocytes with WT or codon optimized hβ2 subunits alone or in the additional presence of hβ3 subunits Functional expressions of human α6β2and α6β2β3-nAChRs were not achieved in Xenopus oocytes whether WT or codon optimized human nAChR β2 subunits were expressed with hα6 subunits alone or in the additional presence of hβ3 subunits.
Human α6β2β3 V9'S -nAChRs are functional whether they are expressed using WT or codon optimized nAChR hβ2 subunits but the current responses are higher from oocytes expressing the codon optimized hβ2 subunits Incorporation gain-of-function hβ3 V9'S subunits instead of WT hβ3 subunits lead to expression of functional hα6hβ2hβ3 V9'S -nAChRs but their peak current responses are minimal [25]. In an earlier effort [26], this observation could not be achieved although the use of mouse (m) α6 subunit instead of hα6 subunit led to expression of functional mα6hβ2hβ3 V9'S -nAChRs [26]. Nonetheless here we show that functional hα6hβ2hβ3 V9'S -nAChRs could be expressed in oocytes by injecting higher amount of cRNAs (~23 ng) for each subunit. Furthermore the use of a codon optimized nAChR hβ2 subunit instead of a WT hβ2 subunit increases the current responses of human α6β2β3 V9'S -nAChRs. Results (Figures 3 and 4) indicate that oocytes co-injected with hα6 subunit, codon optimized hβ2 subunit, and hβ3 V9'S subunit cRNAs elicit~31-fold higher (15 ± 2 nA vs. 469 ± 75 nA; p < 0.05) current responses than those injected with hα6, hβ2 and hβ3 V9'S subunit cRNAs in response to activation by 100 μM nicotine. Therefore we decided to measure current responses from oocytes coexpressing variant-hα6 subunits, codon optimized hβ2 subunits and hβ3 V9'S subunits first and then verify these results using WT hβ2 subunits wherever it is feasible.
Human α6β4and α6β4β3-nAChRs are functional and current responses obtained from oocytes expressing α6β4β3-nAChRs are higher than those expressing hα6hβ4-nAChRs While hα6hβ4-and hα6hβ4hβ3-nAChRs are difficult to express [25,26], functional hybrid nAChR consisting of mα6, hβ4 and hβ3 subunits (i.e., mα6hβ4hβ3-nAChR) [26] and others such as human α2β4-, α3β2-, α3β4-, α4β2or α4β4-nAChRs could be easily expressed in oocytes using a relatively lower amount (~1-4 ng) of cRNA for each subunit [36]. In this study we used relatively larger amount of cRNA (~23 ng) for each subunit for functional expression of hα6hβ4-or hα6hβ4hβ3-nAChRs. This approach has been shown to result in functional expression of hα6hβ4*-nAChRs [29]. Like reported previously [29], here we observed hα6hβ4-and hα6hβ4hβ3-nAChRs are functional and the peak current responses of hα6hβ4hβ3-nAChRs exceed those of hα6hβ4-nAChRs (35 ± 2 nA vs. 407 ± 34 nA in response to 100 μM nicotine,~12 -fold increase, p < 0.05; and 25 ± 3 nA vs. 502 ± 60 nA in response to 1000 μM ACh,~20fold increase, p < 0.05), demonstrating a potentiation effect of hβ3 subunits on the peak current responses of hα6hβ4-nAChRs (Figures 5-6 and 7; Table 3 and Additional file 1: Figure S1 and Table S1). The EC 50 values for nicotine and ACh acting at hα6hβ4hβ3-nAChR is determined to be 12 μM and 40 μM respectively (Tables 3 and Additional  file 1: S1). Lack of current responses at lower concentrations of nicotine or ACh precluded our ability to construct CR curves for hα6hβ4-nAChRs. For mutants, the first amino acid (single letter code, numbering begins at the translation start methionine) designates the WT human nAChR hα6 subunit residue that is replaced with the indicated, second amino acid. In the forward primer nucleic acid sequence, capitalization indicates the nucleotide changed from the WT subunit to create the corresponding mutant.
Variations in nAChR hα6 subunit influence the current responses of human α6β2β3 V9'S -nAChRs but they do not alter the null responses observed from oocytes coexpressing nAChR hα6 and hβ2 (WT or codon optimized) subunits in the presence or absence of hβ3 subunits We sought to determine whether coexpression of the variant nAChR hα6 subunits with WT or codon optimized hβ2 subunits in the presence or absence of hβ3 subunits would result in expression of functional nAChR in oocytes. We did not observe detectable current responses from these oocytes except that coexpression of codon optimized hβ2 and hβ3 subunits with either hα6 D199Y or hα6 S233C subunits appeared to result in functional nAChRs but their peak current responses to 100 μM nicotine (in the range of 10 to 20 nA) could not be reliably and consistently measured (data not shown We attempted to reproduce these results using WT hβ2 subunits. Oocytes coexpressing nAChR hα6 D57N , hα6 R87C , hα6 S156R , hα6 N171K , hα6 S43P , hα6 N46K or hα6 D92E subunits along with WT hβ2 subunits and hβ3 V9'S subunits did not yield current responses to nicotine. These results conversely indicate that a null response observed for hα6 (S43P, N46K or D92E) hβ2hβ3 V9'S -nAChRs using WT Figure 3 Variations in nAChR hα6 subunit influence the current responses of human α6β2β3 V9'S -nAChRs expressed in oocytes using codon optimized nAChR hβ2 subunits. Mean (±SEM) peak inward current responses upon exposure to 100 μM nicotine (5 sec exposure; ordinate) are estimated from oocytes (n = 3-7) voltage clamped at −70 mV and heterologously expressing the indicated nAChR subunits. Current responses of hα6hβ2hβ3 V9'S -nAChR are completely abolished (D57N or S156R), partially abolished (S43P, N46K, R87C, D92E or N171K), not changed (E101K, A112V, A184D, N203T or I226T) and increased (R96H, D199Y or S233C) as a result of the indicated variations in nAChR hα6 subunits. Oocytes coexpressing nAChR hα6(D199Y) subunits, codon optimized hβ2 subunits and hβ3 V9'S subunits yield largest current responses to nicotine. Comparisons of peak current responses between control (hα6hβ2hβ3 V9'S -nAChR) and variant nAChR groups were analyzed using one-way ANOVA with Dunnett's multiple comparisons test (*, p < 0.05; and **, p < 0.01).
hβ2 subunits is not truly null as some degree of function is detected using the codon optimized hβ2 subunits. Hence a null function detected using WT hβ2 subunits in fact is a level of current response that is below our limit of detection.
The (hα6, hα6 E101K , hα6 A112V , hα6 A184D , hα6 N203T or hα6 I226T )hβ2hβ3 V9'S -nAChRs expressed in oocytes 3 days after injection yielded~15-30 nA (data not shown) current in responses to 100 μM nicotine. In order to confirm that these current responses are real and due to the expression of functional nAChRs, recordings were done after waiting for additional 2 days. This time peak current responses could be reliably measured [ Figure 4B]. Mean (±SEM) level of current responses for these nAChR ranged from 53 ± 12 nA (for hα6 N203T hβ2hβ3 V9'S -nAChR) to 123 ± 40 nA [for hα6 I226T β2β3 V9'S -nAChR; Figure 4B]. However, they were similar (p > 0.05) to those of the control. Hence coexpression with WT hβ2 subunits instead of codon optimized hβ2 subunits did not alter the outcome. Note that the current responses of hα6hβ2hβ3 V9'S -nAChRs to nicotine were assayed 5 days after injection (but not 3 days after injection as shown for Figure 4A]. Hence there is disparity in peak current responses of hα6hβ2hβ3 V9'S -nAChR in the two panels of Figure 4.

Variations in nAChR hα6 subunits that affect the agonist (nicotine or ACh) sensitivity of hα6hβ4*-nAChRs
Concentration-response (CR) curves for WT or variant hα6hβ4*-nAChR were produced, wherever feasible, in a manner to glean maximum comparative information about them. As we could not produce CR curves for hα6hβ4-nAChRs, for comparative analysis, we have adopted the EC 50 values for nicotine (7.1 μM) and ACh hβ4-nAChRs are lower (~1.5-2.5 fold; p < 0.05) than those of corresponding nAChRs additionally incorporating hβ3 subunits. Furthermore nicotine EC 50 values at hα6 (A184D or S233C) hβ4hβ3-nAChRs and ACh EC 50 values at hα6 (R96H, A184D, D199Y or S233C) hβ4hβ3-nAChRs are also lower (~1.7-2.5 fold; p < 0.05) than those of hα6hβ4hβ3-nAChRs. Hence nicotine or ACh sensitivity of hα6hβ4and/or hα6hβ4hβ3-nAChRs are marginally or significantly increased as a result of Arg96His, Ala184Asp, Asp199Tyr or Ser233Cys variations in hα6 subunit (Figure 8 and Additional file 1: Figure S1, Tables 3 and Additional file 1:  Table 3 for parameters of ACh action. Table S1). These results, in general, indicate that incorporation of nAChR hβ3 subunit into WT or variant hα6hβ4-nAChR complexes result in marginally lower EC 50 values for nicotine or ACh.

Discussion
In discussion of our results we presume that empirical changes in peak current levels and/or sensitivity of nAChRs are indicative of successful incorporation of WT or mutant β3 subunits into functional WT or variant hα6*-nAChRs. However, in the absence of data for the effect of the variations in nAChR hα6 subunit on subunit biogenesis and trafficking; receptor assembly and level of cell surface expression; ligand binding and channel open probability; etc. we have relied on cell surface functional receptors to draw inferences, conclusions and/or propose additional hypotheses about functional expression of WT or variant hα6*-nAChRs.
The inability to express functional hα6hβ2*-nAChRs without use of gain-of-function hβ3 subunits (i.e., hβ3 V9'S ) confounds our ability to make inferences about assembly of the WT subunits but it gives credence to the idea that nAChR hβ3 subunits exert dominant negative effect on the function of α6β2*-nAChRs [25]. Alternatively, there possibly exists an undetectable level of basal function for hα6hβ2hβ3-nAChRs that gets amplified upon substitution of hβ3 V9'S subunits for hβ3 subunits. The lack of function for heterologously expressed hα6hβ2-and hα6hβ2hβ3-nAChRs additionally could be indicative of lack of presence of other nAChR subunits (e.g., α3 or α4), chaperones or cellular components that typically would facilitate assembly and functional expression of α6β2*-nAChRs in neurons or other cells [26,27,29]. Nonetheless in this study we have successfully used hα6hβ2hβ3 V9'S -nAChRs as a model to evaluate the effect of variations in hα6 subunit on function of hα6hβ2*-nAChRs. Our results are greatly enhanced by the use of a codon optimized nAChR hβ2 subunit and advantages of use of such codon-optimized nAChR subunits were demonstrated previously [27,28,37,38].
The current results and experimental approach indicate that hβ3 subunits do not exert a dominant negative effect on the function of hα6hβ4*-nAChRs [25][26][27][28]. Rather they promote its functional expression as is shown here and elsewhere [29]. It appears that hβ3 subunits need a consistent basal level of expression of nAChR hα6 and hβ4 subunits, manifested as consistent functional expression of hα6hβ4-nAChRs, for their integration and promotion of functional expression of the resultant nAChRs. hα6hβ4-nAChRs expressed using 1 ng or similar amount of cRNAs for each subunit, typical for functional expression of many other nAChR subtypes [36], are barely detected functionally on cell surface. Further microinjection of hβ3 subunits in similar amounts or in 20-fold excess of other α and β subunits [25,26] seems not to affect the outcome but upon substitution of hβ3 V9'S subunits for hβ3 subunits there is emergence of highly functional hα6hβ4hβ3 V9'S -nAChRs. Remember large excess of hβ3 subunits might be promoting non-functional, dead-end intermediates [39] exacerbating the already poor expression of nAChR hα6 and hβ4 subunits or poor functional expression of hα6hβ4-nAChRs leading to the notion that hβ3 subunits exert a dominant negative effect on the function of hα6hβ4*-nAChRs. The increased functionality of hα6hβ4hβ3-nAChRs relative to hα6hβ4-nAChRs, rather than a null or decreased functionality [25], is analogous to studies using chimeric (α6/α3) subunits (containing the N-terminal domain of the nAChR α6 subunit substituting for that of the otherwise α3 subunit) instead of WT α6 subunits that shows a potentiating effects of WT β3 subunit on function of (α6/α3) (β2 or β4)*-nAChRs [29].
Lack of consistent and reproducible current responses from oocytes coexpressing hα6 (D57N, R87C, S156R or N171K) subunits with hβ2 and hβ3 V9'S subunits; or with hβ4 subunits probably indicate that these variations, respectively, presumably in the N-terminal α-helix, strand β2, strand β6 and cysteine-loop disrupt assembly of the subunits into a typical functional pentamer in a manner similar to those encountered with hα6 (D57N, R87C, S156R or N171K) hβ4hβ3-nAChRs. The Asn 171 residue in hα6 subunit, mutated to Lys (K), is a potential target for N-glycosylation. Potential de-glycosylation of the nAChR hα6 subunit as a result of the Asn171Lys variation could affect stability and trafficking of hα6(N171K)hβ2hβ3 V9'S -, hα6(N171K)hβ4-, hα6 (N171K)hβ4hβ3-or hα6(N171K)hβ4hβ3 V9'S -nAChRs to the cell surface [40,41]. The outward current responses of hα6 N171K hβ4hβ3 V9'S -nAChR to ACh or nicotine, in contrast to an expected inward current, probably corroborate this point. Also the Asn171Lys variation in hα6 subunit could disrupt the essential role of the cys-loop in coupling agonist (e.g., nicotine) binding to channel gating (opening) in hα6hβ2*-, hα6hβ4-or hα6hβ4hβ3-nAChRs. Note that Asp 57 , Arg 87 , Ser 156 and Asn 171 residues of hα6 subunit are strongly conserved in nAChR subunits from various organisms indicating that a variation in these residues could not be tolerated without negative consequences on the structure-function relationship of hα6*-nAChRs. Additionally it appears elimination of a positively charged residue at AA position 87 (Arg87Cys) or introduction of a positively charged residue at position 171 (Asn171Lys) is having negative functional consequences.
For oocytes expressing hα6 (S43P, N46K or D92E) hβ2hβ3 V9'S -, hα6 D92E hβ4-or hα6 D92E hβ4hβ3-nAChRs to display compromised or reduced current responses could be due to decreased cell surface functional expression of these receptors. Additionally it could be due to change in inherent structure of the receptor molecules because of the indicated AA variations. It appears that the Asp 92 residue, strongly conserved across various nAChR subunits located in loop D (β2-β3 loop) in the complementary face of the hα6 subunit, upon mutation to a glutamate (that results in an increase in AA side chain length) is having a negative effect on the subunit assembly of α6β2β3 V9'S -, α6β4-, or . The reduction in current responses of hα6hβ2hβ3 V9'S -nAChRs but not those of hα6hβ4-or hα6hβ4hβ3-nAChRs as a result of Ser43Pro variation in hα6 subunit could be due to a subtype specific effect of the variation. Similarly a reduction in current responses of hα6hβ2hβ3 V9'S -nAChRs; no change in current responses of hα6hβ4hβ3-nAChRs; and an increase in current responses of hα6hβ4-nAChRs as a result of Asn46Lys variation in hα6 subunit could be a subtype specific effect of the variation. This in turn could be attributed to the introduction of a positively charged residue (i.e., Lys) at AA position 46 in hα6 subunit.
Variations in nAChR hα6 subunit that occur at residue 101 (Glu101Lys: β2-β3 loop/loop D), 112 (Ala112Val: loop A), 184 (Ala184Asp: loop B), 203 (Asn203Thr: loop F) or 226 (Ile226Thr: loop C) do not affect the peak current responses of hα6hβ2hβ3 V9'S -or hα6hβ4hβ3-nAChRs, an indication that natural AA substitutions at these positions do not grossly affect the assembly, cell surface expression and/or structure-function relationship of these nAChRs. However, variations in loops A (Ala112Val), B (Ala184Asp) or C (Ile226Thr) in hα6 subunit substantially increases current responses from minimally functional hα6hβ4-nAChRs signifying the emerging notion that N-terminal loop residues are important in the assembly and functional expression of hα6*-nAChRs [26][27][28]. Also it is of significance that some of these variations in various loop residues are innocuous or beneficial for the functional expression of hα6hβ2hβ3 V9'S -, hα6hβ4-or hα6hβ4hβ3-nAChRs an indication that these substitutions are tolerated in regions that are crucial in ligand binding and/or subunit assembly. Coincidentally hα6 subunits have non-conserved AAs at position 101 (Glu), 112 (Ala), 184 (Ala) or 226 (Ile); and a weakly conserved AA at position 203 (Asn) in an alignment analyses of human nAChR α subunits (see Figure 1). A weak or nonconserved AA residue probably indicates, but not necessarily always, a tolerance for these substitutions. However, such a broader interpretation becomes difficult to generalize as hα6 subunit seems to have strongly conserved AAs at positions 101 (Glu), 112 (Ala), 184 (Ala) and 226 (Ile) and a weakly conserved AA at position 203 (Asn) in an alignment analysis of α6 subunits from a limited number of other species.
Variations that occur in the nAChR hα6 subunit at AA residues 96 (Arg96His: β2-β3 loop/loop D), 199 (Asp199Tyr: loop F) or 233 (Ser233Cys: strand β10) increases the peak current responses from oocytes expressing hα6hβ2hβ3 V9'S -nAChRs. However, these variations except Asp199Tyr do not have any effect on the peak current responses of hα6hβ4-or hα6hβ4hβ3-nAChRs expressed in oocytes again demonstrating a subtype specific effect of these variations in β2-β3 loop/loop D, loop F or β10-strand of hα6 subunit. It appears that the Arg 96 or Asp 199 residue located in the complementary face of the hα6 subunit is exerting a positive effect on the subunit assembly of α6β2β3 V9'S -and . These results are similar to previously described reports that AA residue (e.g., Asn143: loop E) located in the negative face of the hα6 subunit influence the functional expression of hα6hβ2hβ3 V9'S -nAChRs [26].
In the absence of data for WT hα6hβ4-nAChRs comparisons of EC 50 values among WT and variant hα6hβ4-nAChRs remained incomplete. However, it is reported [29] that ACh (18 ± 5 μM) or nicotine (7.1 ± 2.6 μM) EC 50 values at hα6hβ4-nAChRs are marginally lower than those of respective hα6hβ4hβ3-nAChRs (ACh EC 50 : 33 ± 8 μM and nicotine EC 50 :10 ± 3 μM). Our results indicate that this directionality in change in potency is preserved for hα6 (A184D, D199Y or S233C) hβ4-nAChRs indicating that EC 50 values at hα6 (A184D, D199Y or S233C) hβ4-nAChRs is lower than those containing the same subunits but also additionally containing nAChR hβ3 subunits. But note that the increase in agonist sensitivity of nAChRs as a result of the variations in loop B (Ala184Asp), loop F (Asp199Tyr) and β strand 10 (Ser233Cys) (which connect to the trans-membrane domain I) in hα6 subunit are nominal implying that there could be subtle changes in receptor structures.
In the final analyses it is incumbent upon us to know the significance, if any, of these variations in hα6 subunit in the etiology of nicotine dependence and/or other hα6*-nAChR involved diseases. We would not know them until currently available tools, statistical or biotechnological, becomes mainstream. Specifically analyses of rare variations (i.e., Asp57Asn, Arg87Cys, Ser156Arg or Asn171Lys) in nAChR hα6 subunit that compromise the function of hα6*-nAChRs would be of prime interest in epidemiological or in vivo studies. Nonetheless individuals displaying altered hα6*-nAChR pharmacology as a result of rare variation in nAChR hα6 subunit are expected to exhibit differential responses to smoked nicotine.

Conclusions
Our results presented here are in general agreement with the accumulated evidences that changes/mutations in loop residues and other structural residues could affect cell surface expression, assembly, structure and/or function of various nAChRs. Specifically N-terminal αhelix (Asp 57 ); complementary face/inner β-fold (Arg 87 or Asp 92 ) and principal face/outer β-fold (Ser 156 or Asn 171 ) residues in hα6 subunit are crucial for functional expression of hα6hβ2*-, hα6hβ4-and hα6hβ4hβ3-nAChRs and natural variations in them (i.e., Asp57Asn, Arg87Cys, Asp92Glu, Ser156Arg or Asn171Lys) compromises their function. Additionally Ser 43 or Asn 46 (N-terminal α-helix) residues in hα6 subunit are important for functional expression of hα6hβ2*-nAChRs and natural variations in them (i.e., Ser43Pro or Asn46Lys) compromise its function. However, natural variations indicate Arg 96 (β2-β3 loop/loop D), Asp 199 (loop F) or Ser 233 (β10strand) residues in hα6 subunit could be taken advantage for promoting functional expression of hα6hβ2*-nAChRs. Similarly residues in N-terminal α-helix (Asn 46 ), loop A (Ala 112 ), loop B (Ala 184 ), loop F (Asp 199 ) or loop C (Ile 226 ) could be substituted with their respective natural variations (i.e., Asn46Lys, Glu101Lys, Ala112Val, Ala184Asp, Asp199Tyr or Ile226Thr) for increased functional expression of poorly functional/expressed hα6hβ4-nAChRs. Thus, by studying natural variations in nAChR hα6 subunit, we have mapped AA residues in nAChR hα6 subunit important for cell surface functional expression of various subtypes of hα6*-nAChRs. These novel sites in nAChR hα6 subunit could be of promising use for creation of functional cell lines that could be helpful for drug screening; and development of new drug candidates selective for hα6*-nAChRs. This is of increasing importance given the potentially important roles for α6*-nAChRs in movement and movement disorders, mood disorders, and drug dependence [12,19,[42][43][44].

Bioinformatics analyses
Human nAChR α (α1-α7, α9, α10) subunits or nAChR α6 subunits from various organisms were aligned using ClustalW and then edited for the purposes of presentation ( Figure 1). A homology model of the human nAChR α6 subunit modeled on the 3-D coordinates of the muscle nAChR α subunit of Torpedo marmorata (PDB: 2BG9.A) [2] was retrieved using SWISS-MODEL [45] protein modeling server; and subsequently was rendered using UCSF Chimera (http://www.cgl.ucsf.edu/chimera/), a program for interactive visualization and analysis of molecular structures (Figure 2).

Chemicals
All chemicals used in electrophysiology were obtained from Sigma Chemical Co. (St. Louis, MO, USA) except that L-nicotine was obtained from Arcos Organics (New Jersey, USA). Fresh agonist (acetylcholine or nicotine) and antagonist (atropine) stock solutions were made daily or diluted from frozen stock in Ringer's solution (OR2) which consisted of (in mM) 92.5 NaCl, 2.5 KCl, 1 CaCl 2 , 1 MgCl 2 , and 5 HEPES; pH 7.5.

Preparation of cRNA mixture for Xenopus oocyte microinjection
We planned to introduce identical amounts of cRNA, presumably producing equal amounts of each subunit protein, into oocytes largely due to lack of information about the levels of mRNA for each subunits that compose α6*-nAChRs in neurons or other cells. Concentrations of cRNAs for each nAChR α and β subunits (WT, mutant or variant) were adjusted to 500 ng μL −1 . We provisionally assumed that hα6 or its variants in association with hβ2 or hβ4 subunits would form complexes having 2:3 and/or 3:2 ratios of the indicated subunits and that oocytes also injected with WT or mutant form of β3 subunits (i.e., β3 V9'S ) would express nAChR with 2:2:1 ratios of α:β:(β3 or β3 V9'S ) subunits. For expression of binary nAChRs (i.e., two subunit containing nAChRs; α + β but not β3) cRNA mixtures were prepared by mixing 1 μL of cRNA for each subunit and an additional μL of RNAse free water (i.e., total volume 3 μL). Similarly for expression of ternary nAChRs [i.e., three subunit containing nAChRs; (α + β) + (β3 or β3 V9'S )] cRNA mixtures were prepared by mixing 1 μL of cRNA for each subunit. Several preparations of each mixture were prepared and stored at −80°C until further use. Injection of 138 nL of cRNA, out of 3 μL cRNA mixtures, into each oocyte would deliver~23 ng of cRNA for each subunit whether binary or ternary nAChRs are expressed.

Oocyte preparation and cRNA microinjection
All Xenopus laevis (Nasco, Fort Atkinson, WI, USA) procedures were conducted in accordance with the guidelines of the National Institutes of Health (NIH) for the proper use of laboratory animals and approved by the Institutional Animal Care and Use Committee (IACU) of University of Virginia. Female Xenopus laevis (Nasco, Fort Atkinson, WI) were anesthetized using 0.2% tricaine methanesulfonate (MS-222) (Nasco, Fort Atkinson, WI). Ovarian lobes were surgically removed from the frogs and placed in an incubation solution that consisted of (in mM) 82.5 NaCl, 2.5 KCl, 1 MgCl2, 1 CaCl2, 1 Na 2 HPO 4 , 0.6 theophylline, 2.5 sodium pyruvate, 5 HEPES, 50 mg/ml gentamycin, 50 U/ml penicillin, and 50 μg/ml streptomycin; pH 7.5. The lobes were cut into small pieces and digested with Liberase™ (research grade, medium Thermolysin concentration; Roche Applied Science, Indianapolis, IN) with constant stirring at room temperature for 1.5-2 hours. The dispersed oocytes were thoroughly rinsed with incubation solution. Stage VI oocytes were selected and incubated at 16°C before injection. Micropipettes used for injection were pulled from borosilicate glass (Drummond Scientific, Broomall, PA) using a Sutter P1000 horizontal puller, and the tips were broken with forceps to~40 μm in diameter. cRNA was drawn up into the micropipette and injected into oocytes using a Nanoject microinjection system (Drummond Scientific) at a total volume of 138 nL.

Oocyte electrophysiology
Two to 5 days after injection, oocytes were placed in a small-volume chamber and continuously perfused with OR2. The chamber was grounded through an agarose bridge. The oocytes were voltage-clamped at −70 mV (unless otherwise noted) to measure agonist or antagonist induced currents using Axoclamp 900A and pClamp 10.2 software (Molecular Devices, Sunnyvale, CA). The current signal was low-pass filtered at 10 Hz with the built-in low-pass Bessel filter in the Axoclamp 900A and digitized at 20 Hz with Axon Digidata1440A and pClamp10. Electrodes contained 3 M KCl and had a resistance of 1-2 MΩ. Drugs (agonists and antagonists) were prepared daily in OR2. Drug was applied using a Valvelink 8.2 perfusion system (Automate scientific, Berkeley, CA). One micromolar (1 μM) atropine was always coapplied for acetylcholine (ACh)-based recordings to eliminate muscarinic AChR (mAChR) responses. nAChR hα6 constructs were tested individually or in batches as they became available to get an approximate idea about their effect on the function of hα6*-nAChRs. Then for the purpose of comparison electrophysiological recordings were performed in a given day in a given batch of oocytes following the same order of injections. Hence data points in a figure panel were obtained under almost similar experimental conditions. All hβ4*-nAChR recordings were done in similar conditions to facilitate comparisons between hα6hβ4-and hα6hβ4hβ3-nAChRs. All electrophysiological measurements were conducted or checked in at least two batches of oocytes.

Experimental controls
Injection of water or empty vector or of cRNA corresponding to one subunit alone or pairwise combinations of β3 or β3 V9'S subunits with either an α or β2 or β4 subunit did not result in the expression of functional nAChR. Current responses from these oocytes to 100 μM nicotine were less than 5-10 nA (data not shown).