Polyamine regulation of ion channel assembly and implications for nicotinic acetylcholine receptor pharmacology

Small molecule polyamines are abundant in all life forms and participate in diverse aspects of cell growth and differentiation. Spermidine/spermine acetyltransferase (SAT1) is the rate-limiting enzyme in polyamine catabolism and a primary genetic risk factor for suicidality. Here, using genome-wide screening, we find that SAT1 selectively controls nicotinic acetylcholine receptor (nAChR) biogenesis. SAT1 specifically augments assembly of nAChRs containing α7 or α4β2, but not α6 subunits. Polyamines are classically studied as regulators of ion channel gating that engage the nAChR channel pore. In contrast, we find polyamine effects on assembly involve the nAChR cytosolic loop. Neurological studies link brain polyamines with neurodegenerative conditions. Our pharmacological and transgenic animal studies find that reducing polyamines enhances cortical neuron nAChR expression and augments nicotine-mediated neuroprotection. Taken together, we describe a most unexpected role for polyamines in regulating ion channel assembly, which provides a new avenue for nAChR neuropharmacology.

N euronal nicotinic acetylcholine receptor (nAChR) ion channels mediate behavioral, cognitive and autonomic effects of acetylcholine and addictive properties of nicotine [1][2][3][4][5][6] . The nAChR family comprises nine alpha (α2-10) and three beta (β2-4) subunits that combine to form an array of pentameric cation channels throughout the brain and peripheral nervous systems. The major nACh receptors in brain are α7 homomers, α4β2 heteromers, and α6-containing heteromers. These receptors are targets of numerous approved and experimental medicines for diverse conditions including Alzheimer's disease, Parkinson's disease, and neuropathic pain 7,8 .
Whereas nAChRs are compelling drug targets, progress is hampered because most receptor subtypes do not functionally express in the non-neuronal cell lines used for screening 9,10 . We previously identified NACHO, a four-pass transmembrane protein that serves as a client-specific chaperone for assembly of most neuronal nAChRs [11][12][13] . In brain, NACHO is essential for function of α7 receptors, and additional proteins, Ric-3 14 and certain Bcl-2 family members 15 , synergize with NACHO to enhance α7 assembly and surface trafficking. NACHO also promotes assembly of α4β2 receptors, though Ric-3 and Bcl-2 proteins have minimal impact on α4β2.
Here, we sought factors that conspire with NACHO to enhance function of α4β2 receptors, the most abundant nAChR subtype in brain. These neuronal α4β2 receptors mediate diverse physiological actions of acetylcholine and underlie nicotine dependence 16 . Genome-wide cDNA screening for protein enhancers of α4β2 function identified, quite surprisingly, a single standout clone that encodes spermidine/spermine acetyltransferase (SAT1), the rate limiting enzyme for polyamine catabolism. By degrading polyamines, SAT1 promotes α4β2 function to an even greater extent than does NACHO.
Polyamines are abundant polycations, including spermidine and spermine, that play multiple roles in cell growth, differentiation and survival 17 . The interplay between their synthesis by ornithine decarboxylase-1 (ODC1) and their degradation by SAT1 controls polyamine levels. ODC1 is amongst the most dynamically regulated of all human proteins, and ODC1 is a drug target in oncology and infectious disease 17,18 . SAT1 transcription is also highly-regulated, and its acetylation of polyamines promotes their cellular export 18 . Interestingly, numerous large genomic studies link polymorphisms in SAT1 with suicidal behavior 19 . In neurons, polyamines play important roles in synaptic transmission by conferring inward rectification to certain potassium channels, AMPA receptors and nACh receptors [20][21][22] . Polyamines also participate in the pathogenesis of neurodegenerative disorders 17 and the excitotoxicity associated with cerebral ischemia 23 .
We now find that polyamines also control assembly of neuronal α4β2 and α7 receptors. By contrast, polyamines do not modulate assembly of α6β4 nAChRs, AMPA receptors or any other ion channel tested. Whereas polyamines classically regulate channel gating by occluding the ion pore, polyamine regulation of nAChR assembly instead relies on negatively charged residues within the α4 or α7 cytosolic loop. Neuropharmacology studies using wild-type and NACHO knockout mice show that lowering polyamine levels selectively upregulates cerebrocortical α4β2 and α7 levels and enhances the neuroprotective properties of nicotine. These studies identify an unexpected role for polyamines in controlling ion channel biogenesis and suggest new strategies in neuropharmacology.

Results
Genome-wide screening identifies SAT1 as an enhancer of nAChR. To identify novel regulators of α4β2 receptors, we co-transfected human embryonic kidney 293T cells (HEK293T) with plasmids encoding α4 and β2 subunits along with individual constructs from a 5943-cDNA clone library from the Broad Institute 24 . Transfected cells were stimulated with nicotine (100 µM) and intracellular Ca 2+ was quantified with a fluorescence imaging plate reader (FLIPR) (Fig. 1a). As previously shown 11 , α4β2 alone produced a small nicotine-evoked Ca 2+ signal, and co-transfection with NACHO significantly enhanced this (Fig. 1a). High throughput screening identified a single clone that profoundly augmented the nicotine-induced Ca 2+ response-to levels much higher than NACHO; this clone encoded spermidine/ spermine N1-acetyltransferase (SAT1) (Fig. 1a).
We evaluated the effect of SAT1 on other nAChRs and other Cys-loop receptors both by FLIPR and electrophysiology. Consistent with our screening results, SAT1 dramatically increased ACh-evoked currents from α4β2 (Fig. 1c, d). SAT1 also synergized with NACHO to further enhance the α4β2mediated Ca 2+ influx (Fig. 1b, d), suggesting that SAT1 and NACHO employ different mechanisms. Whereas SAT1 alone did not rescue homomeric α7 function, SAT1 powerfully synergized with NACHO to increase α7 mediated currents (Fig. 1b-d). In contrast, SAT1 had no significant effect on α6β4 or 5-HT3A receptor function (Fig. 1b-d).
SAT1, a small cytosolic protein, is the rate-limiting enzyme for polyamine catabolism ( Supplementary Fig. 1). Together with polyamine oxidase, SAT1 acetylates higher-order polyamines converting them to inactive forms that are transported out of cells 18 . By lowering polyamine levels, SAT1 effects on nAChRs could reflect disinhibition, as polyamines can negatively affect nAChR gating 22 . This seemed unlikely for two reasons, First, α6containing nAChRs are more sensitive to polyamine inhibition of gating than are α7 25 whereas we find the opposite sensitivity to SAT1 in our functional assays ( Fig. 1b-d). Second, polyamines do not block nAChRs at the hyperpolarized membrane potentials we used for patch clamp studies 22 . As an alternative mechanism, we asked whether SAT1 increases nAChR surface expression. To assess this, we utilized extracellular HA-tagged subunits that allow detection of surface receptors without disrupting channel function 11 . Strikingly, SAT1 boosted surface levels of both α4β2 and α7 (Fig. 2a, b). SAT1 also augmented surface trafficking atop the effects of the α7 protein chaperones Ric-3, Bcl-X L , and Mcl-1, as well as the α7 chemical chaperone/orthosteric antagonist, methyllycaconitine (MLA) ( Supplementary Fig. 2a, b) implying that receptor upregulation by SAT1 involves a mechanism distinct from any previously described.

SAT1 promotes assembly of nAChRs by catalyzing polyamines.
To determine whether effects on nAChRs involve the catalytic activity of SAT1, we constructed a Tyr140Phe mutant that abolishes SAT1 enzyme activity 26 . This mutant SAT1 (mutSAT1) did not change surface expression of α4β2 or α7 receptors in the presence or absence of NACHO (Fig. 2a, b). Ornithine decarboxylase 1 (ODC1) is the rate-limiting enzyme for polyamine synthesis ( Supplementary Fig. 1), and difluoromethylornithine (DFMO) is an ODC1 active site inhibitor 27 that depletes cellular polyamines 18 . Like SAT1 co-transfection, DFMO pretreatment augmented surface expression of α4β2 and synergized with NACHO to further enhance α4β2 and α7 surface receptors (Fig. 2a, b). Pursuing the opposite tact, we found that co-transfection with the polyamine biosynthetic enzymes ODC1 and adenosylmethionine decarboxylase 1 (AMD1) diminished nAChR surface staining (Supplementary Fig. 3a, b). Neither catabolic (SAT1 co-transfection or DFMO pretreatment) nor anabolic (ODC1 and AMD1 co-transfection) polyamine manipulations affected surface expression of the 5-HT3A receptor (Fig. 2a, b; Supplementary Fig. 3a, b).
We asked next whether polyamines influence nAChR assembly, which can be probed with orthosteric ligands, such as [ 3 H] epibatidine, that only bind at the interface between folded subunits 28 . Remarkably, either SAT1 or DFMO enhanced [ 3 H] epibatidine binding to α4β2 or α7 receptors (Fig. 2c). Furthermore, ODC1 and AMD1 reduced ligand binding to these receptors ( Supplementary Fig. 3c).
N1,N11-bis-(ethyl)-norspermine (BenSpm) is a stable polyamine analog and SAT1 inhibitor 26,29 , and treatment of cells with 10 µM BenSpm specifically reversed the effects of SAT1 on nAChR surface expression (Fig. 2d). By contrast, incubating cells with spermine, which does not penetrate cells and is rapidly degraded by SAT1, did not affect nAChR surface expression (Fig. 2d). Taken together, we find that polyamines, independent of their established role in ion channel gating, enhance nAChR function by promoting surface expression and augmenting receptor assembly.
Polyamines regulate nAChRs assembly via their cytosolic loop. Certain neuronal nAChRs display blunted ion flow at depolarized potentials. This inward rectification is mediated by polyamine binding to pore-lining glutamate residues in the nAChR transmembrane domain 2 (TM2) corresponding to Glu247 in α4 22,30 . It was previously shown that mutating this acidic residue to alanine (α4 E247A ) relieves polyamine block of α4β2 channels in Xenopus laevis oocytes 22 . Loss of negative charge in the α4 E247A also reduces calcium permeability through the mutant receptor. Accordingly, we found that a E247A α4 mutant (Fig. 3a) cotransfected with β2 evinced minimal nicotine-evoked Ca 2+ influx in HEK293T cells (Fig. 3b), and this was unaffected by preincubation with DFMO or co-transfection with SAT1. By contrast, DFMO or SAT1 enhanced surface expression of α4 E247A /β2 similar to wildtype α4β2 (Fig. 3c, d). These data establish distinct mechanisms for polyamine regulation of nAChR gating and trafficking.
We next mutated Trp156 (Fig. 3e) in the α4 ligand-binding domain 30 . As expected, when co-transfected with β2, this α4 W156A mutant was functionally inactive and did not bind to [ 3 H]epibatidine (Fig. 3f, g). By contrast, this mutant showed typical surface staining enhancement upon co-transfection with SAT1 cDNA (Fig. 3h, i) indicating that polyamine regulation is independent of agonist binding.
To identify nAChR regions responsible for regulation, we generated chimeras of α4 with α6, as surface expression of the latter is not regulated by polyamines (Fig. 4a). We co-transfected High throughput cDNA screening for enhancers of α4β2 function identifies SAT1. a HEK293T cells in 384-well plates were co-transfected with α4-expressing and β2-expressing plasmids along with an individual cDNA (black traces) from a human ORF collection (Broad Institute). Cells were stimulated with 100 μM nicotine (Nic). The highest response occurred in the well containing SAT1 (red trace) and was twice the α4β2 + NACHO response (gray trace), which served as positive control. b Quantification (mean ± SD) of maximum Ca 2+ signal upon agonist stimulation from HEK293T cells transfected as indicated (n = 6). Co-transfection of SAT1 significantly enhances Ca 2+ signal in α4β2 (p < 1e −4 ), α4β2+NACHO (p < 1e −4 ) and α7+NACHO (p = 0.002) cells. For all α6β4 conditions, cells were co-transfected with plasmids encoding accessories BARP and SULT2B1 13 .
c Representative whole-cell current responses elicited from HEK cells co-transfected with GFP and cDNAs as indicated. d Summary graphs of agonistevoked peak currents (mean ± SEM) from HEK293T cells transfected with indicated cDNA combinations. Similar to maximum Ca 2+ signal, co-transfection of SAT1 boosts evoked currents mediated by α4β2 (p < 1e −4 ), α4β2+NACHO (p < 1e −4 ), α7+NACHO (p < 1e −4 ) but not α6β4 (p = 0.98) or 5-HT 3 A (p = 0.73). Numbers indicate number of transfected cells that were analyzed and were pooled from three independent cultures. **p < 0.01, ***p < 0.001, Oneway ANOVA between the groups for α4β2 and α7. Mann-Whitney U test versus control for α6β4 and 5-HT 3 A. Source data for panel b and d are provided as a Source Data file. these α4/α6 chimeras with β4 and quantified both receptor surface expression and nicotine-evoked Ca 2+ signaling (Fig. 4). SAT1 co-transfection enhanced a chimera containing the α6 Nterminal and α4 C-terminal region but did not affect the converse chimera ( Fig. 4a-c), which implies that polyamine regulation involves the channel or the large cytosolic loop, which occurs between TM3 and TM4. We next constructed a pair of chimeras that swapped only this cytosolic loop. Quantification of receptor surface expression and function showed that the α4 TM3-TM4 cytosolic loop is both necessary and sufficient for SAT1dependent regulation of these nAChRs ( Supplementary Fig. 4).
Similarly, exchanging the TM3-TM4 loop of homomeric α7 receptor with that of α6 ( Supplementary Fig. 4d) abolished SAT1mediated upregulation of alpha-bungarotoxin (α-Bgt) binding sites on α7 ( Supplementary Fig. 4e, f), which demonstrates that the intracellular loop determines polyamine-regulation of both α4 and α7 nAChR assembly. Previous studies noted that, within the T3-TM4 loop, a subdomain proximal to TM4 contains conserved positively and negatively charged residues 31 . Interestingly, we observed certain negatively charged residues (Glu and Asp) in this subdomain that are conserved between α7 and α4 but diverge in α6 to uncharged Asn ( Supplementary Fig. 5a, b) 31 . Mutating those Glu and Asp to Ala in α7 (E 438 A or D 451 A) and α4 (E 569 A or D 582 A) blocked SAT1-mediated upregulation of α7 assembly ( Supplementary  Fig. 5c, d) and α4β2 surface expression ( Supplementary Fig. 5 e, f). By contrast, these mutations did not blunt effects of Ric-3 on α7 or NACHO on α4β2. Replacing the corresponding Asn residues (N 434 or N 447 ) to Glu within the α6 cytosolic loop conferred moderate but significant SAT1 and DFMO sensitivity to α6β4 receptors ( Supplementary Fig. 5g, h). These experiments identify negatively charged residues within the TM3-TM4 cytosolic loop that can mediate polyamine regulation of nAChR assembly and surface expression. Future studies are needed to fully elucidate these mechanisms.
Polyamines regulate nAChR gating and assembly independently. To further distinguish mechanisms for polyamine control of nAChR assembly and gating, we compared effects of the membrane-permeant polyamine analog BenSpm with philanthotoxin-343 (PhTx-343), which cannot enter cells and engages the channel pore via a long polyamine tail 25,32 . Neither BenSpm nor PhTx-343 preincubation changed basal surface levels of α4β2. Whereas BenSpm abrogated SAT1-mediated enhancement of surface α4β2, PhTx-343 did not (Fig. 5a, b). By contrast, both BenSpm and PhTx-343 preincubation abolished nicotineevoked Ca 2+ with or without SAT1 co-transfection (Fig. 5c). These data demonstrate that only cell-permeant polyamines can control of nAChR surface trafficking.
application abolished nicotine-evoked Ca 2+ signaling from either α4β4 or α6β4 (Fig. 5f). We next evaluated effects of BenSpm and PhTx-343 on the α6/ α4 chimeric constructs, which swap the extracellular N-terminal or TM3-TM4 cytosolic loop domains. As predicted, the α4 cytosolic loop was necessary and sufficient for BenSpm pre-treatment to reverse SAT1-mediated enhancement of receptor surface expression, and PhTx-343 did not alter surface expression for any chimera ( Fig. 5e and Supplementary Fig. 6d, e). By contrast, acute application of either BenSpm or PhTx-343 blocked nicotine-evoked Ca 2+ -signaling from all receptor chimeras in a concentrationdependent manner (Fig. 5f, Supplementary Fig. 6a-c). These molecular biological and pharmacological studies establish distinct mechanisms and protein domains for polyamine control of nAChR trafficking and gating.
Polyamines regulate assembly of neuronal nAChRs. We next explored whether polyamines upregulate endogenous nAChRs in neurons. Accordingly, we treated cultured rat cortical neurons with DFMO and quantified α7 nAChR surface expression with fluorescent α-Bgt. As a positive control, we incubated neurons with 100 μM nicotine 33 and found a significant increase in surface α-Bgt binding sites (Fig. 6a, b). Similarly, treating neurons with DFMO robustly increased surface α-Bgt binding (Fig. 6a, b). By contrast, neither DFMO nor nicotine altered surface expression of the AMPA-type glutamate receptor subunit GluA1 (Fig. 6a, b). Preincubating neurons with DFMO also increased the nicotineevoked Ca 2+ signal (Fig. 6c), an effect that was comparable to the enhancement seen following nicotine pretreatment. These nicotine-evoked Ca 2+ responses reflect α7 activity, as they require the α7-specific positive allosteric modulator (PAM) PNU-12059615 34 and were abolished by the α7-specific inhibitor α-Bgt (Fig. 6c).
[ 3 H]Epibatidine binding in brain largely reflects α4β2 receptors and was absent in cerebral cortex of α4 knockout mice 35 . Importantly, incubating cortical neurons with DFMO or transducing them with a lentivirus expressing SAT1 enhanced levels of [ 3 H]epibatidine binding (Fig. 6d)   epibatidine binding to the transduced neuronal membranes werẽ 20 times higher than untransduced neurons, and this was further increased with DFMO pretreatment (Fig. 6d). Importantly, neither DFMO preincubation nor SAT1 transduction altered binding sites for the GABA receptor ligand [ 3 H] flunitrazepam (Fig. 6d). Taken together, these data show that polyamines regulate assembly of neuronal nAChR but not AMPA or GABA receptors.
Polyamines modulate nicotine-mediated neuroprotection. Nicotine has well-established neuroprotective properties and can mitigate excitotoxicity in vivo 36 and in vitro 37,38 . This neuroprotection is mediated both by α7 39 and α4β2 nAChRs 38,40 . To assess the functional consequence of polyamine upregulation of receptor assembly, we asked whether this nicotine-mediated neuroprotection could be enhanced by reducing cellular polyamine levels. We induced excitotoxicity by challenging cultured neurons for one hour with glutamate (30 µM) and quantified cell survival and cytosolic cytochrome C (CytC), which is a trigger for apoptosis 41,42 . We also surface-labeled α7 receptors with fluorescent α-Bgt. In line with previous reports, glutamate challenge induced excitotoxity and increased cytosolic CytC (Supplementary Fig. 7a-c). Co-application of 100 μM nicotine reduced excitotoxicity, but glutamate did not alter surface expression of α7 ( Supplementary Fig. 7b, d). On the other hand, neurons pretreated with either DFMO or nicotine showed a three-fold increase in surface α-Bgt-binding sites ( Supplementary Fig. 7a, d) and a concordant reduction of glutamate-induced CytC release ( Fig. 6e; Supplementary Fig. 7c). Nicotine or DFMO preincubation also augmented nicotine-mediated protection from glutamate-mediated excitotoxic cell death ( Fig. 6e and Supplementary Fig. 7b).
To more directly link the neuroprotective effects of DFMO with nAChRs, we studied NACHO KO mice, which have dramatically reduced nAChR function in brain 12 . We found that nicotine blunted glutamate-induced cell death in neurons from wild-type but not NACHO KO mice (Fig. 7a-c). Fitting with results from our rat cortical neuron experiments, we found that pre-treating wild-type mouse neurons with DFMO increased surface α-Bgt staining (Fig. 7e) and mitigated glutamate-induced increases in cytosolic CytC (Fig. 7d). By contrast, DFMO pretreatment did not augment α-Bgt binding levels in NACHO KO neurons. Accordingly, DFMO preincubation did not confer subsequent nicotine-induced protection from excitotoxic neuronal death in NACHO KO neurons (Fig. 7c). Collectively, these experiments show that polyamines regulate functional assembly of nAChRs, which promotes neuroprotection in wild-type, but not NACHO KO neurons.

Discussion
This study identifies an unexpected role for polyamines in controlling assembly of neuronal α4β2 and α7 receptors. Our   genome-wide analysis found that SAT1 increases assembly of α4β2 receptors to a much greater extent than any other protein screened, including the nicotinic receptor chaperone NACHO. Polyamine-regulated assembly is specific for certain nAChRs, is distinct from other nAChR chaperone mechanisms, and requires negatively charged amino acids within the α4 or α7 cytosolic loop. Blocking polyamine synthesis with DFMO upregulates neuronal α4β2 and α7 surface levels and promotes nicotinemediated neuroprotection, which provides a new angle for drug discovery.
The biogenesis of pentameric nAChRs is a complex and tightly-regulated process 43 . Seminal studies in the 1980s found that nicotine and other orthosteric ligands upregulate brain nACh protein levels, and this likely participates in nicotine dependence 44,45 . In C. elegans, Ric-3 is required for efficient assembly of worm nAChRs 46 ; in mammalian brain, NACHO is required for α7 11 and many other nAChRs 12 . Additional proteins synergize with NACHO for assembly of specific mammalian nAChRs. Ric-3 14 and certain Bcl-2 family proteins 15 work with NACHO to promote function of α7, whereas BARP, LAMP5, and +SAT1+BenSpm +SAT1+PhTx +Vector f +SAT1 +SAT1+BenSpm +SAT1+PhTx +Vector Fig. 5 Polyamine analogs regulate nAChR gating and assembly by distinct mechanisms. a Fluorescent anti-HA labeling of non-permeabilized HEK293T cells transfected with cDNAs encoding extracellular HA-tagged β2 and α4 with or without SAT1. Scale bar = 50 μm. b Quantification of surface staining for α4β2 receptors with or without SAT1 and BenSpm or PhTx343 pre-treatment (n = 6). BenSpm occluded the SAT1-enhanced receptor surface expression (p < 1e −4 ) but PhTx343 had no effect (p = 0.86). c Nicotine-evoked Ca 2+ signal was fully blocked by both BenSpm or PhTx343 (p < 1e −4 ) (n = 6). d Schematics of chimeric constructs of α4-α6. e Normalized anti-HA surface expression from cells transfected and treated with BenSpm or PhTx343 as indicated (n = 5). Only BenSpm occluded SAT1-mediated enhancement of surface receptor expression of α4β4 (p < 1e −4 ), α6NT/α4β4 (p < 1e −4 ) and α6/ α4-loopβ4 (p < 1e −4 ). Both SAT1 and BenSpm effects on receptor surface expression required the α4 cytosolic loop. f Both BenSpm and PhTx343 fully inhibited nicotine-evoked Ca 2+ in all receptor combinations. All cDNA combinations included BARP and SULT2B1. Quantifications are displayed as mean ± SD. ***p < 0.001, One-way ANOVA between the groups was used for panels b, c and e. Comprehensive dose-response studies are in Supplementary Fig. 5. Source data for panel b, c and e are provided as a Source Data file.
These multiple layers of nAChR regulation utilize distinct mechanisms. In the case of α7, NACHO mediates subunit oligomerization 11 . Subsequent regulation by nicotine, Ric-3, and Bcl-2 proteins synergizes with NACHO to promote protein folding, receptor surface expression, and channel function 14,15 . Nicotine and other orthosteric ligands promote assembly through α7's ACh binding domain 45 whereas Bcl-2 proteins 15 and polyamines engage the receptor cytosolic loop. However, polyamine regulation involves a distinct mechanism, as we find SAT1 augments receptor assembly atop NACHO, nicotinic ligand, Ric-3, or Bcl-2 family proteins.
It is intriguing to ask why nAChRs-but not other receptors in the Cys-loop superfamily including 5-HT3 and GABA A receptors -require accessories for functional expression. One possibility is that multiple assembly mechanisms provide regulatory nodes for controlling nAChR function. NACHO transcription is upregulated by physiological stimuli 47 , and Bcl-2 family protein levels are dynamically-induced during developmental and pathological processes 48 . Furthermore, protein accessories may also determine nAChR cellular localization. The α6-containing nAChRs specifically concentrate at presynaptic terminals of specific monoaminergic neurons, and this may be enabled by the lysosomal protein, LAMP5, which displays a similar restricted distribution 49 .
Polyamines are ubiquitous in biology and play multiple roles in cell growth, survival, and differentiation 17 . In neurons, polyamines control gating of several important ion channels. Cytosolic polyamines confer inward rectification to certain potassium channels, AMPA receptors, and nAChRs [20][21][22] . Elegant biophysical studies showed that neuronal depolarization draws cytosolic polyamines into the channel pore, which precludes ion flow 20 . This polyamine site is accessible to extracellular  . One-way ANOVA compared to control. Neurons transduced with lentivirus expressing α4 and β2 showed much higher levels of [ 3 H]epibatidine binding (hatched bars) (p < 1e −4 , compared to un-transduced control), and these were further enhanced (p < 1e −4 ) with DFMO treatment (n = 8). e Acute application of nicotine (gray bars) partially reduced glutamate-mediated cell death (left panel) and mobilization of CytC (right panel). Pre-treating neurons for six days with DFMO (green, 5 mM) enhanced nicotine-mediated cell survival (p = 0.006) and decreased CytC mobilization (p = 0.001). Nicotine (orange, 100 μM) pre-treatment similarly enhanced nicotine-mediated neuroprotection. Data averaged from 3 independent experiments each containing 6 independent samples. Quantifications are displayed as mean ± SD. **p < 0.01, ***p < 0.001, One-way ANOVA between the groups for panels b-e. Source data for panels b-e are provided as a Source Data file. philanthotoxins 50 , which have long polyamine tails that can engage the channel.
Our studies decisively establish that regulation of α7 and α4β2 assembly by polyamines is distinct from their classical role in controlling ion channel gating. First, we find that SAT1 increases α4β2 and α7 function even when recorded at hyperpolarized potentials that preclude nAChR gating control by cellular polyamines 22 (Fig. 1c, d). Second, SAT1 promotes assembly of nonfunctional α4β2 receptors that cannot bind ACh (Fig. 3e-i).
Third, extracellular philanthotoxin acutely blocks α4β2 receptor function but philanthotoxin does not affect receptor assembly (Fig. 5a-c). Fourth, chronic but not acute application of a cellpermeable polyamine analog, BenSpm, reverses the effects of SAT1 on α7 and α4β2 receptor assembly and function (Fig. 2d). Interestingly, we find that SAT1 does not augment function of α6-containing nAChRs. Taking advantage of this α-subunit specificity, our α6/α4 chimeras determine that polyamine regulation of channel assembly involves the α4 cytosolic loop. Again, this is White squares indicate regions that are magnified. c-e Graphs quantify neuronal survival (c), CytC (d) and surface α-Bgt647 (e) (n = 6). Acute nicotine increases cell survival (p < 1e −4 ) and reduces CytC mobilization (p = 0.0002) during glutamate toxicity in wild-type neurons, but not in NACHO KO neurons (p = 0.9). DFMO pre-treatment further promotes nicotine-mediated cell survival (p < 1e −4 ) and reduces CytC mobilization (p = 0.0002) in wildtype neurons. DFMO significantly enhances surface α-Bgt647 labeling of wild-type (p < 1e −4 ) but not of NACHO KO neurons (p = 0.9). **p < 0.01, ***p < 0.001, n.s = not significant, one-way ANOVA between the groups. Data displayed as mean ± SD. Source data for panel c-e are provided as a Source Data file.
NATURE COMMUNICATIONS | https://doi.org/10.1038/s41467-020-16629-3 ARTICLE distinct from the pore region that determines polyamine control of rectification. As α7 and α4β2 are the most abundant nAChRs in human brain and control diverse aspects synaptic signaling and plasticity, polyamine regulation of their assembly has important physiological and pathophysiological implications. Polyamine levels in neurons are dynamically regulated over both short-time and long-time scales. Acutely, synaptic transmission increases synthesis of polyamines, which modulate integrative neuronal properties by reducing AMPA receptor currents 51 and provide an excitability buffer by negatively regulating Na + channels 52 . Longlasting changes in synaptic transmission associated with pathological processes such as epilepsy dramatically upregulate ODC1 53 , which our study predicts would downregulate nAChR assembly. Indeed, α4-containing nAChRs levels are decreased in piriform cortex of kindled mice 54 .
Polyamines and nAChRs share compelling links to neuropsychiatric disorders. Numerous genetic, genomic, and biochemical studies identify alterations in the polyamine pathway in major depressive disorder and suicide 55 . An SNP in the promoter region of SAT1 that reduces expression is associated with suicide in a French-Canadian founder population 55 . Postmortem studies find reduced SAT1 protein and mRNA levels in precentral gyrus and cortical frontal lobe 56 . Also, polyamine levels increase during anxiety episodes, and this polyamine stress response pathway 19 could suppress nAChR assembly. Fitting with this, preclinical and early clinical studies suggest that α7 or α4β2 agonists can improve depressive behavior in animal models and in depressed patients 57 .
Numerous observations also link polyamine and nAChR alterations with cognitive and neurodegenerative processes. Spermine synthase mutations cause Snyder-Robinson syndrome characterized by mental retardation and juvenile myoclonic epilepsy 58 , and these central symptoms are also found in certain patients with mutations in α4 nAChR 59 . Interestingly, both SAT1 overexpression 60 and nicotine 36 protect animals from kainateinduced toxicity. Polyamine levels are elevated 17 and nAChR levels are reduced 61 in Alzheimer's disease, which is best treated today with cholinesterase inhibitors that augment nAChR activity. Furthermore, DFMO 23 and nicotinic agonists 40 both mitigate neurotoxicity from neuronal ischemia. Our results that blocking ODC1 increases neuronal α4β2 and α7 and promotes nicotinemediated neuroprotection provide a mechanistic link for these observations. As such, targeting this polyamine/nAChR assembly pathway provides a new neuropharmacological strategy. Chimeric constructs used in this study, with residues in the parenthesis are: α6NT/α4 (α6 1-239 /α4 243-628 ), α4NT/α6 (α4 1-242 /α6 240-495 ), α6/α4loop (α6 328-465 exchanged to α4 331-600 ), α4/α6loop (α4 331-600 exchanged to α6 328-465 ), α7/α6loop (α7 328-457 exchanged to α6 339-453 ). In brief, for α6NT/α4 chimera, α4 243-628 region and α6 1-239 was linearized and amplified with PCR. The primers for α6 1-239 contained complementary overhangs immediately upstream and downstream of amplified region. PCR products were ligated using In-Fusion HD Cloning Plus kit (Takara, 683910) in accordance with manufacturer's protocol. All other chimeras were generated following similar strategy and were confirmed by sequencing. Sitedirected mutagenesis was done with two complementary primer reactions (primer details provided in Supplementary Table 1), and all mutations were confirmed by sequencing. The α7-HA, β2-HA, and β4-HA constructs contained a PSGA linker and HA tag immediately following the C-terminal residue.

Methods
HEK293T cells (ATCC® CRL-3216™) were cultured in DMEM medium supplemented with 10% FBS and 1 mM sodium pyruvate. Cells were seeded at 80-90% confluence and were transiently transfected using FuGENE®6 transfection reagents (Promega Corporation). All functional and expression assays in HEK293T cells were performed after 48-72 h incubation at 37°C following transfection, unless stated otherwise.
All animal experiments reported here were overseen and approved by an AAALAC accredited institutional review board. Primary cortical neurons were prepared from E18 rat cortex (supplied by BrainBits®) and E18 mice cortices from wild-type or NACHO knockout mice (TMEM35a tm1(KOMP)Vlcg ) 11 (obtained from the laboratory's vivarium). Cortices were dissociated using 10 U/mL papain (Worthington) digestion for 10 min, followed by trituration with a 10 mL glass pipette. Neurons were seeded at 15,000 cells per well (384-well plates) and maintained in NbActiv4® media (BrainBits®). Immunostaining or FLIPR assays were performed at DIV 20 or DIV 13, respectively, following 6 days incubation in nicotine (100 µM) or DFMO (5 mM) as indicated. For radioligand binding assays, neurons were transduced with SAT1 or α4 and β2 encoding lentiviruses (DIV 7) or treated with DFMO (DIV 14) and harvested on DIV 20. The lentiviral vectors encoding SAT1, α4 and β2 used a PGK promoter for transgene expression. The SAT1-expressing viral particles were packaged by Vigene Biosciences™, while the α4-expressing and β2-expressing viruses were packaged by VectorBuilder.
Broad cDNA library screen. The cDNA library used for α4β2 screening was from the Broad Institute and contains 5943 cDNA clones. This genome-wide screen used a FLIPR assay on transfected HEK293T cells in 384-well plates. Each well was transfected with 60 ng of total cDNA with the following composition: α4:β2:single FLIPR assay for agonist-evoked Ca 2+ influx. High-throughput FLIPR assays were conducted using 384-well BioCoat (Corning) plates. Cells were washed briefly in assay buffer (HyClone™ HEPES-buffered saline [GE Life Sciences] comprising 149 mM NaCl, 4 mM KCl, 10 mM HEPES, and 5 mM glucose at pH 7.4 and 300 mOsm osmolality, supplemented with 2 mM CaCl 2 and 1 mM MgCl 2 ) prior to Calcium5 dye (Molecular Devices) loading for 1 h at RT. Following removal of excess dye, plates were placed in the FLIPR Tetra (Molecular Devices) chamber, and fluorescent Ca 2+ signal was captured using ScreenWorks 4.0™ software (Molecular Devices).
To obtain neuronal α7 mediated FLIPR responses, we used 5 μM PNU-12059615-a selective drug that attenuates receptor desensitization. Tetrodotoxin, or TTX (500 nM) was included in the assay buffer to inhibit spontaneous action potentials.
For neurons, cells were fixed with 4% PFA for 1 h and then stained with fluorescent α-Bgt conjugates (1 μg/mL, AlexaFluor 647, Invitrogen™, B35450) or primary antibodies for 1 h. Then cells were simultaneous permeabilized and blocked in 0.2% Triton X-100 and 10% normal goat serum for 30 min. To label intracellular components, neurons were incubated in primary and secondary antibodies sequentially for 1 h each. Nuclei were stained with DAPI or NucBlue® reagent (Invitrogen™) prior to imaging.
For quantification, images were acquired using Harmony™ high-content imaging software on an Opera Phenix™ screening instrument (PerkinElmer) with a ×20 objective. Data were further analyzed with the Columbus™ data storage and analysis system (PerkinElmer). Viable cells were identified from the DAPI signal (area >30 μm 2 for HEK293T; area >10 μm 2 for neurons). Cell perimeter was determined from the nuclear staining, and area of the cytoplasm was calculated based on the cytoplasmic region of interest. Average fluorescence intensity of wells was determined after subtracting the background signal, which was defined as anti-HA labeling of untransfected HEK293T cells or as α-Bgt647 labeling not displaced by 10 μM epibatidine.
Radioligand binding assays. HEK293T cells or cortical neurons were harvested in 50 mM ice-cold TrisHCl buffer (pH 7.4). Cells were homogenized for 30 s using the T-25 Ultra-Turrax homogenizer (Ika) and total protein concentration of the homogenate was determined using the Pierce™ BCA Protein Assay Kit (Thermo Scientific). Cell homogenates were incubated with 10 nM [ 3 H]epibatidine (for nAChR) or with 30 nM [ 3 H]flunitrazepam (for GABA) in 96-well plates for 3 h at room temperature. Nonspecific binding was determined by co-incubation of the cell samples with 10 μM unlabeled epibatidine or 100 μM unlabeled flunitrazepam. Assays were terminated by filtration through polyethylenimine-treated 96-well Unifilter GF/B plates (PerkinElmer). Filter plates were washed with 500 mL TrisHCL buffer and then desiccated at 65°C for 30 min. MicroScint-0 scintillant cocktail (50 μL; PerkinElmer) was added to each well and plates read with a TopCount NXT scintillation counter (PerkinElmer).
Electrophysiology. HEK293T cells were seeded (1 million/well) on uncoated sixwell plates and transfected with cDNA combinations (total 2 μg/well) using FuGENE®6 transfection reagent (Promega Corporation). eGFP plasmid (10% of total cDNA) identified transfected cells. After 24 h, cells were dissociated using CellStripper™ dissociation reagent (Corning) and re-seeded on 12 mm glass coverslips (40,000/well). Electrophysiological recordings were done 48 h after transfection using external solution composed of HyClone™ HEPES-buffered saline (149 mM NaCl, 4 mM KCl, 10 mM HEPES, 2 mM CaCl 2 , 1 mM MgCl 2 and 5 mM Glucose at pH 7.4; 300 mOsm osmolality). Intracellular solution contained (mM): 140 potassium gluconate, 10 HEPES, 4 Mg-ATP, 0.4 Na-GTP, and 0.6 EGTA (pH 7.3). To study most receptors, fast perfusion of compounds was achieved with the Perfusion Fast-Step system (Warner Instruments). To study α7 nAChRs, ultrafast perfusion of compounds was achieved with a piezo-driven perfusion system and theta glass (Siskiyou) on to eGFP-expressing cells. The membrane holding potential was −70 mV. All recordings were performed at room temperature using an Axopatch 200B amplifier (Axon Instruments) and signals were filtered at 2 kHz and digitized at 10 kHz. For α7 nAChRs, signals were filtered at 10 kHz with a digitization rate of 50 kHz. Data acquisition and subsequent analysis were done with pClamp9 software (Axon Instruments).
Statistics. Results are represented as mean ± SD unless stated otherwise. All FLIPR assay, immunostaining experiments and radioligand binding assay in HEK293T cells and rat neurons were replicated thrice. Significance analyses between two datasets were performed with nonparametric Mann-Whitney U test, while statistical analyses between three or more datasets used one-way ANOVA (GraphPad Prism, Carlsbad, CA). Significance level α = 0.05 was set.
Reporting summary. Further information on research design is available in the Nature Research Reporting Summary linked to this article.

Data availability
Data supporting the findings of this manuscript are available from the corresponding author upon reasonable request. A reporting summary for this Article is available as a Supplementary Information file. The source data underlying Figs. 1b, d; 2b-d; 3b, c, g, i; 5b, c, e; 6b-e; 7c-e and Supplementary Figs. 2b; 3b, c; 4b, c, f; 5d, e; 6b-d are provided in the Source Data File. Source data are provided with this paper.