Identification of a novel nuclear localization signal common to 69- and 82-kDa human choline acetyltransferase.

We demonstrated previously that 69- and 82-kDa human choline acetyltransferase are localized predominantly to the cytoplasm and the nucleus, respectively. We have now identified a nuclear localization signal common to both forms of enzyme using confocal microscopy to study the subcellular compartmentalization of choline acetyltransferase tagged with green fluorescent protein in living HEK 293 cells. To identify functional nuclear localization and export signals, portions of full-length 69-kDa choline acetyltransferase were cloned into the vector peGFP-N1 and the cellular distribution patterns of the fusion proteins observed. Of the nine constructs studied, one yielded a protein with nuclear localization and another produced a protein with cytoplasmic localization. Mutation of the critical amino acids in this novel putative nuclear localization signal in the 69- and 82-kDa enzymes demonstrated that it is functional in both proteins. Moreover, 69-kDa choline acetyltransferase but not the 82-kDa enzyme is transported out of the nucleus by the leptomycin B-sensitive Crm-1 export pathway. By using bikaryon cells expressing both 82-kDa choline acetyltransferase and the nuclear protein heterogeneous nuclear ribonucleoprotein with green and red fluorescent tags, respectively, we found that the 82-kDa enzyme does not shuttle out of the nucleus in measurable amounts. These data suggest that 69-kDa choline acetyltransferase is a nucleocytoplasmic shuttling protein with a predominantly cytoplasmic localization determined by a functional nuclear localization signal and unidentified putative nuclear export signal. For 82-kDa choline acetyltransferase, the presence of the unique amino-terminal nuclear localization signal plus the newly identified nuclear localization signal may be involved in a process leading to predominantly nuclear accumulation of this enzyme, or alternatively, the two nuclear localization signals may be sufficient to overcome the force(s) driving nuclear export.

We demonstrated previously that 69-and 82-kDa human choline acetyltransferase are localized predominantly to the cytoplasm and the nucleus, respectively. We have now identified a nuclear localization signal common to both forms of enzyme using confocal microscopy to study the subcellular compartmentalization of choline acetyltransferase tagged with green fluorescent protein in living HEK 293 cells. To identify functional nuclear localization and export signals, portions of fulllength 69-kDa choline acetyltransferase were cloned into the vector peGFP-N1 and the cellular distribution patterns of the fusion proteins observed. Of the nine constructs studied, one yielded a protein with nuclear localization and another produced a protein with cytoplasmic localization. Mutation of the critical amino acids in this novel putative nuclear localization signal in the 69-and 82-kDa enzymes demonstrated that it is functional in both proteins. Moreover, 69-kDa choline acetyltransferase but not the 82-kDa enzyme is transported out of the nucleus by the leptomycin B-sensitive Crm-1 export pathway. By using bikaryon cells expressing both 82-kDa choline acetyltransferase and the nuclear protein heterogeneous nuclear ribonucleoprotein with green and red fluorescent tags, respectively, we found that the 82-kDa enzyme does not shuttle out of the nucleus in measurable amounts. These data suggest that 69-kDa choline acetyltransferase is a nucleocytoplasmic shuttling protein with a predominantly cytoplasmic localization determined by a functional nuclear localization signal and unidentified putative nuclear export signal. For 82-kDa choline acetyltransferase, the presence of the unique amino-terminal nuclear localization signal plus the newly identified nuclear localization signal may be involved in a process leading to predominantly nuclear accumulation of this enzyme, or alternatively, the two nuclear localization signals may be sufficient to overcome the force(s) driving nuclear export.
Choline acetyltransferase (ChAT, 1 acetyl-CoA:choline Oacetyltransferase, EC 2.3.1.6) catalyzes the reaction between choline and acetyl coenzyme A to produce the neurotransmitter acetylcholine (ACh) which mediates cholinergic neuron communication. Cholinergic neurons are distributed throughout the central and peripheral nervous systems, and dysfunction of these neurons underlies aspects of clinical symptoms found in neurological and psychiatric disorders such as Alzheimer's disease, Huntington's disease, and Rett's syndrome (1)(2)(3). ChAT is encoded by a single gene located in a gene locus that also contains the vesicular ACh transporter gene; together these form the cholinergic gene locus (4,5). In recent years, it has been found that there is polymorphism in ChAT mRNA resulting from alternative splicing and differential utilization of at least five different exons within the non-coding region of the ChAT gene promoter (6). Currently, six isoforms of ChAT mRNA, termed R, N1, N2, H, S, and M, have been identified. All six ChAT mRNAs translate into a 69-kDa form of the enzyme, with the M-and S-transcripts also encoding 82-and 74-kDa proteins, respectively, because of the presence of two putative translation initiation sites in these transcripts (6,7). The M-ChAT transcript that can produce both 69-and 82-ChAT appears to be expressed only in primates in the brain and the spinal cord (6).
ChAT is a single-strand globular protein containing histidine, arginine, and cysteine residues at or near the active site of the protein (8). The 82-kDa form of ChAT differs from the 69-kDa enzyme by having a 118-amino acid extension on the amino terminus of the protein. This region of the protein contains a functional nuclear localization signal (NLS) resulting in 82-kDa ChAT having a predominantly nuclear subcellular distribution, whereas 69-kDa ChAT is located largely in the cytoplasm or associated with the plasma membrane (9). The significance of the differential subcellular distribution of ChAT protein isoforms is unknown at present but suggests different cellular or regulatory mechanisms.
Nuclear localization signals (NLS) are responsible for the ability of proteins to translocate from the cytoplasmic compartment of the cell to the nucleus. Recently, the field of study of nuclear transport has grown dramatically resulting in identification of a number of NLS sequences (10 -13). Classical NLS sequences are generally rich in the basic amino acids lysine or arginine (14,15). Nuclear export signals (NES), on the other hand, govern protein translocation from the nucleus to the cytoplasm along nuclear export receptor-mediated pathways (16 -21). NES sites have traditionally been described as being rich in leucine, isoleucine, and valine residues (16).
The combination of enhanced green fluorescent protein (eGFP)-tagged proteins and scanning confocal laser microscopy of living cells provides a powerful technique for investigating subcellular compartmentalization and the molecular trafficking of proteins (22)(23)(24). In the present study, we investigated the dynamics of subcellular distribution of 69-kDa ChAT tagged with eGFP. Although this protein is found predominantly in the cytoplasm of HEK 293 cells, 69-kDa ChAT-eGFP can also be detected in the nucleus of these cells. We hypothesized that the ChAT sequence contains NLS and NES motifs that function in movement of the protein into and out of the nucleus. We report here for the first time that 69-kDa ChAT has features of a nucleocytoplasmic shuttling protein driven into the nucleus by a functional NLS and that it moves out of the nucleus along the Crm-1 nuclear export pathway. We also provide evidence that this NLS is functional in, and necessary for, the nuclear transport of 82-kDa ChAT.

EXPERIMENTAL PROCEDURES
Cell Culture and Transfection-Human embryonic kidney (HEK 293) cells were grown in minimal essential medium supplemented with 10% fetal bovine serum and 50 g/ml each of penicillin and streptomycin. For some studies, cells were seeded at a density of 2 ϫ 10 6 cells on 35-mm culture dishes and then transiently transfected with plasmid DNA using LipofectAMINE 2000 (Invitrogen). Prior to use for confocal microscopy, cells were replated after transfection onto 35-mm glassbottomed dishes (MakTek) and grown in medium containing 20 mM HEPES buffer, pH 7.4. For other studies, G418-resistant stable transformants of HEK 293 cells expressing native 69-or 82-kDa human ChAT were maintained in medium containing 0.1% geneticin (G418); cells were transfected with either 69-or 82-kDa ChAT subcloned into pcDNA3.1 thus giving expression of the native protein without eGFP tag.
Confocal Microscopy-Scanning confocal laser microscopy was performed using a Zeiss LSM510 microscope and a 63ϫ oil immersion objective. Images were acquired using excitation (488 nm) and emission (515 nm) wavelengths for eGFP or 543 and 590 nm for DsRed1-C1, respectively. Images were captured digitally and imported into Adobe Photoshop 5.0 for formatting.
Protein Mapping/Bioinformatics Data-The bioinformatics programs PROSITE (www.expasy.org/prosite) and ScanSite (www.scansite.mit. edu) were utilized in conjunction with published NLS/NES sequences to identify potential NLS and NES sequence domains in ChAT.
Leptomycin B (LMB) Treatment-Cells were co-transfected with fulllength human 69-kDa ChAT cDNA subcloned into the peGFP-N1 vector (Clontech) and human RalGDS cDNA in the DsRed1-C1 vector (Clontech). At 24 -48 h after transfection, cells were treated with 20 ng/ml LMB (kindly provided by Dr. Minoru Yoshida, Department of Biotechnology, University of Tokyo). As RalGDS-DsRed1 is a cytoplasmic protein (25, 26) whose subcellular distribution is not affected by LMB, it was used as a positive control to delineate the cytoplasm and to contrast the nucleus, in order to better assess nuclear accumulation of eGFPtagged proteins in some experiments.
cDNA Constructs for NLS/NES Domain Mapping-Plasmid DNA containing full-length 69-kDa human ChAT cDNA in peGFP-N1 was used as a template to prepare constructs for analysis of novel, functional NLS/NES domains in the enzyme. For this purpose, nine fragments of 69-kDa ChAT cDNA were made by traditional PCR and subcloned into peGFP-N1 using restriction enzymes NheI and SacII. These constructs (called Construct-1 to Construct-9) were transfected into HEK 293 cells, and their cellular distribution was assessed in living cells by confocal microscopy. Mutations were performed using the PCR-based QuikChange site-directed mutagenesis kit (Stratagene) using full-length peGFP-69-kDa or peGFP-82-kDa ChAT cDNA or peGFP-Construct-6 as templates. All constructs were subsequently sequenced to verify the integrity of the cloned DNA.
Fluorescence Recovery after Photobleaching (FRAP) Assay-This procedure was performed using a modification of methodology described previously by Howell and Truant (27). Briefly, HEK 293 cells were transfected with peGFP-82-kDa ChAT and DsRed1-hnRNP; heterogeneous nuclear ribonucleoprotein (hnRNP) is a nuclear shuttling protein (28) and was obtained as a gift from Dr. Ray Truant, Department of Biochemistry, McMaster University. Transfected cells were pretreated with 50 g/ml cycloheximide and then fused together with 50% polyethylene glycol treatment to form bikaryon cells. By using confocal microscopy, bikaryon cells that contained both 82-kDa ChAT-eGFP and hnRNP-DsRed1 exhibit a yellow fluorescence (overlay of red and green fluorescence). One of the two nuclei of the fused cells was targeted and photobleached by 50 iterations with 100% laser excitation. Bikaryon cells were allowed a recovery period of 30 min to 1 h and were examined for evidence of fluorescence recovery in the photobleached nucleus. If the protein of interest is a shuttling protein, it should translocate from the unbleached nucleus of the cell to the bleached nucleus of the adjacent cell, resulting in recovery of fluorescence. As new protein synthesis was blocked with cycloheximide, any fluorescently tagged proteins that moved into the photobleached nucleus had to originate from the unbleached nucleus.
Isolation of Nuclei and Determination of ChAT Activity-Intact nuclei were isolated from HEK 293 cells stably expressing native 69-or 82-kDa ChAT or from untransfected wild-type cells. Cells were removed from flasks with balanced salt solution containing trypsin (0.05%) and recovered by centrifugation (900 rpm for 5 min) in PBS. The pellets were then resuspended and re-pelleted in PBS at 900 rpm for 5 min. The cell pellets were then lysed by resuspending in 5 volumes of lysis buffer (10 mM Tris-HCl, pH 7.5, 0.05% Nonidet P-40, 3 mM MgCl 2 , 10 mM NaCl, 5 mM EGTA, 1 mM 4-(2-aminoethyl)benzenesulfonyl fluoride, 10 g/ml leupeptin, 25 g/ml aprotinin, 10 g/ml pepstatin), incubated on ice for 15 min, and then centrifuged at 900 rpm for 5 min. Following one additional wash with lysis buffer, nuclei were recovered by centrifugation, washed 3 times with wash buffer (10 mM HEPES, pH 6.8, 300 mM sucrose, 3 mM MgCl 2 , 25 mM NaCl, 1 mM EGTA, and 1 mM 4-(2aminoethyl)benzenesulfonyl fluoride), and centrifuged at 900 rpm for 5 min. The final nuclear pellets were resuspended in a small volume of wash buffer and maintained on ice until analysis. Activity of ChAT was determined by incubating nuclei in the presence of substrates choline and [ 3 H]acetyl coenzyme A followed by extraction of [ 3 H]ACh using a modification of the liquid cation exchange method of Fonnum (29).
Immunocytochemical Localization of Native ChAT-HEK 293 cells stably expressing native 69-or 82-kDa ChAT were immunostained using an anti-ChAT antibody to establish subcellular compartmentalization of the enzyme without the eGFP tag. Cells plated on 35-mm glass-bottom culture dishes were rinsed with PBS followed by fixation with cold acidified ethanol (absolute ethanol containing 1% acetic acid) for 20 min at Ϫ20°C. Post-fixed cells were rinsed and blocked with 1% bovine serum albumin in PBS for 30 min at room temperature and then incubated with primary anti-ChAT antibody CTab (1:250) for 1 h at room temperature. Following rinsing with PBS (3 washes at 5 min each), cells were incubated with secondary antibody (rhodamine-conjugated donkey anti-rabbit, 1:100) with 1% bovine serum albumin in PBS for 1 h at room temperature. Cells were rinsed with PBS (3 times for 5 min) and viewed by confocal microscopy.

RESULTS AND DISCUSSION
We demonstrated previously that human 69-and 82-kDa ChAT tagged with eGFP is found predominantly in the cytoplasm and the nucleus of HEK 293 cells, respectively (9). However, our data indicate for the first time that 69-kDa ChAT-eGFP is also present in the nucleus. As the steady-state distribution of this enzyme is largely cytosolic, we hypothesized that it is a nucleocytoplasmic shuttling protein displaying functional NLS and NES moieties. In the present study, we demonstrate the ability of 69-kDa ChAT to redistribute in the cell between the cytoplasm and the nucleus. Furthermore, we show that a novel NLS identified in 69-kDa ChAT also plays a role in nuclear import of 82-kDa ChAT. To establish that native ChAT without potential influences of the eGFP tag assumes the same subcellular localization as the ChAT-eGFP fusion proteins, we immunostained HEK 293 cells stably expressing native 69-and 82-kDa ChAT. As illustrated in Fig. 1, 69-kDa ChAT is found predominantly in the cytoplasm with a small proportion of the enzyme protein located in the nucleus, and 82-kDa ChAT is located almost entirely in the nucleus. Interestingly, ChAT in the nucleus assumes a punctate pattern rather than being distributed homogeneously.
Leptomycin B (LMB) Studies Indicate Nuclear Export Activity-To investigate the ability of 69-kDa ChAT to be transported across the nuclear envelope, we conducted nuclear ex-port inhibition studies employing LMB. LMB, which was originally identified as an antifungal agent, inhibits facilitated movement of proteins out of the nucleus along the Crm1-nuclear export pathway directly by blocking its interaction with proteins carrying a functional leucine-rich NES (30 -32). In these experiments, time-dependent accumulation of a protein in the nucleus of LMB-treated cells indicates that the protein likely contains functional NLS(s) to facilitate the movement into the nucleus, as well as an NES(s) that binds to Crm-1 export receptors that move the protein back into the cytoplasm. HEK 293 cells were transiently transfected with peGFP-69-kDa ChAT cDNA and treated in the presence or absence of LMB over a time course of 24 h. To increase the contrast between the nuclear and the cytoplasmic compartments, cells were co-transfected with pDsRed1-RalGDS. RalGDS protein is localized to the cytoplasm (25,26), and the subcellular distribution of this protein is not affected by LMB (data not shown). As illustrated in Fig. 2, A and C, in untreated control cells 69-kDa ChAT-eGFP was found localized predominantly in the cytoplasm along with its clear presence in the nucleus. At 4and 12-h treatment with LMB, we observed increased accumulation of 69-kDa ChAT in the nucleus of cells by confocal microscopy ( Fig. 1, D, F, G, and I). While accumulation of ChAT in the nucleus continued over the 24-h time course, it did not become localized exclusively to the nucleus. The DsRed1-Ral-GDS protein labeled the cytoplasmic compartment, and its distribution did not alter over the observation time period (Fig.  2, B, E, and H), thereby allowing us to clearly delineate accumulation of 69-kDa ChAT-eGFP in the nucleus (Fig. 2, C, F, and I).
These experiments suggest that 69-kDa ChAT contains one or more sequence motifs that impart functional NLS activity to the protein. However, as the enzyme has a predominantly cytoplasmic distribution in the cell (Figs. 1A and 2A), this protein must shuttle between the cytoplasm and the nucleus achieving a steady-state between the driving forces for nuclear import and nuclear export. To generate this subcellular distribution pattern, 69-kDa ChAT must also contain one or more NESs or be shuttled out of the nucleus by binding to a carrier protein containing an NES. Although LMB is a specific inhib-itor of the Crm1-nuclear export pathway, there are a number of other pathways that move proteins out of the nucleus (16,19,(33)(34)(35). Therefore, in addition to utilizing the Crm1 pathway, 69-kDa ChAT could also be exported from the nucleus by other pathway(s) that are not inhibited by LMB. In our experiments, this could account for the relatively prolonged time required for accumulation of ChAT in the nucleus of LMB-treated cells (Fig.  2G).
Protein Mapping and Identification of Functional NLS/NES Domains-The nuclear pore complex is composed of at least 30 distinct proteins and acts like a molecular sieve allowing passive diffusion of proteins up to a size limit of ϳ60 kDa (14,36). Since the molecular mass of 69-kDa ChAT is above the diffusion limit and we found that the protein accumulates in the nucleus over time, the active nuclear translocation would therefore require functional interaction between an NLS(s) in ChAT and import receptors in the nucleus. Furthermore, subcellular distribution of the enzyme favors a cytoplasmic localization indicating that the 69-kDa enzyme likely contains a functional NES(s) and is transported out of the nucleus by export receptors.
By using the bioinformatics programs PROSITE and Scan-Site and published NLS (10 -15) and NES (16 -21) sequences, we identified putative NLS and NES sites in the primary amino acid sequence of 69-kDa human ChAT ( Fig. 3 and Table  I). Based on the presence of basic lysine-and arginine-rich stretches between amino acids 204 -221 and 357-368, we hypothesized that these regions of the protein may constitute potential NLS sequences. Analysis of the primary sequence also revealed regions between amino acids 81-92, 227-235, and 416 -425 resembling the canonical leucine-rich NES sequence LXLXXLXL (18) that could potentially impart this function to the enzyme; isoleucine or valine may substitute for leucine in this sequence in some circumstances (16,18). Based on this analysis, we hypothesized that ChAT could shuttle into the nucleus by one or more NLSs and then translocate back to the cytoplasm driven by a functional NES. Moreover, we hypothesized that these putative NLS(s) that are also contained in the sequence of 82-kDa ChAT could contribute to the nuclear translocation of that enzyme.
To determine whether the potential NLS and NES sequences were functional, nine different segments of the full-length 69-kDa ChAT cDNA were generated by PCR and subcloned into the vector peGFP-N1 (Construct-1 to Construct-9; Fig. 3A). These nine eGFP-tagged constructs were designed to contain an approximately equal number of ChAT amino acid residues and have molecular masses less than 60 kDa. Based on a nuclear passive diffusion limit of ϳ60 kDa (14,36), we predicted that these recombinant proteins would move freely between the nucleus and the cytoplasm and distribute equally between these two subcellular compartments; this type of cellular distribution is shown in Fig. 3B for the 29-kDa protein eGFP (Fig. 3B). However, protein products that carried a functional NLS would be located predominantly in the nucleus, and those with a functional NES would be localized predominantly in the cytoplasm. Based on analysis of the amino acid sequence of the full-length protein, we predicted that Constructs-4 and -6 would contain putative NLS sequences, and Constructs-2, -4, and -7 would contain potential NES sequences (Fig. 3A).
Of the nine constructs generated, only Construct-6 (amino acids 345-410 of 69-kDa human ChAT) assumed a differential nuclear localization pattern in living cells when viewed by confocal microscopy (Fig. 3B). This construct carries one of the potential NLS motifs identified by our sequence analysis (Fig.  3A). In addition, eGFP-tagged Construct-9 (encoding amino acids 584 -630) had a cytoplasmic distribution suggesting that this region of the ChAT protein likely contains a functional NES (Fig. 3B). This portion of the ChAT protein does not, however, contain an identifiable NES motif based on sequences that have currently been reported as having NES activity. A third construct, Construct-7 (amino acids 411-490) which contains a putative NES, yielded a protein product that appeared to be localized to the Golgi and endoplasmic reticulum (data not shown). To study this potential NES sequence further, we determined whether it was functional within the context of the full-length protein by mutating 2 leucine residues to alanine residues (L421A and L423A). Expression of this new mutant construct in cells yielded a subcellular distribution similar to that observed for the wild-type enzyme ( Fig. 2A), suggesting that this putative NES was not involved in nuclear export of the protein.
Construct-2 and Construct-4 (amino acid residues 72-110 and 186 -254, respectively) contain sequences that could function as putative NLS or NES. The subcellular distribution of these eGFP fusion proteins was the same as that obtained for eGFP alone (Fig. 3), indicating that these diffused passively into and out of the nucleus (data not shown). Construct-4 contains sequences identified by PROSITE as being consistent with that of a classical bipartite NLS and also a potential NES motif (Fig. 3). One possibility is that if these two sites are both functional, in combination they could result in a loss of differential subcellular distribution of the fusion protein. To test this, we expressed the portion of the sequence in Construct-4 that encodes the putative bipartite NLS alone (amino acids 199 -226) with eGFP in HEK 293 cells. The subcellular distribution of this protein did not differ from eGFP alone indicating FIG. 3. A, schematic illustrating potential NLS and NES sites identified in the 69-kDa ChAT protein. To isolate novel NLS and NES sites, regions of 69-kDa ChAT cDNA were generated by PCR and subcloned into peGFP-N1. The subcellular distribution of these proteins was examined in transiently transfected HEK 293 cells by confocal microscopy. Of the nine constructs made, only Construct-6 (amino acids 345-410) showed nuclear localization, and Construct-9 (amino acids 584 -630) showed cytoplasmic localization. Cyto denotes cytoplasmic localization, and Nuc denotes nuclear localization, with plus and minus symbols signifying the presence or absence in the subcellular compartment, respectively. B, confocal images of living HEK 293 cells illustrating the cellular distribution of Construct-6 and Construct-9; Construct-6 contains a potential NLS sequence, whereas a potential NES was not identified in Construct-9 based on known NES sequences. The asterisks indicate critical amino acid residues in the NLS sequence. eGFP alone is distributed evenly throughout the cells as shown. Representative data are shown from at least 8 separate experiments. (1) MGLRTAKKR (9) a Numbers included in parentheses denote the location of amino acids in the 82-kDa ChAT protein, with numbers not in parentheses identifying the location of amino acids in 69-kDa ChAT.
b This NLS is found only in the amino terminus of 82-kDa ChAT.
that this NLS was likely not functional in this context (data not shown). Subcellular distributions similar to that of eGFP alone were obtained with eGFP-tagged Constructs-1, -3, -5, and -8; these fusion proteins were not predicted to contain potential NLS or NES domains.

Mutational Analysis of Possible NLS and NES Activity in 69and 82-kDa
ChAT-Construct-6 has functional NLS activity and also contains a region that encodes an NLS sequence motif. To probe further the functionality of the putative NLS within Construct-6, we employed site-directed mutagenesis. Three lysine residues positioned at amino acids 363, 364, and 366 and an arginine residue positioned at 368 in the wild-type 69-kDa ChAT sequence were mutated to alanine residues. The mutant construct in peGFP-N1 was transiently transfected into HEK 293 cells, and the localization patterns of the proteins were visualized. Importantly, as shown in Fig. 4A, upper panels, the mutant Construct-6 protein (345-410 R363A,R364A,R366A,K368A -eGFP) now exhibited a considerably reduced nuclear localization and was distributed essentially equally between the cytoplasm and the nucleus. To establish whether this NLS was capable of influencing subcellular compartmentalization of the full-length 69-kDa ChAT, we mutated these same four amino acids in the cDNA of the wild-type enzyme. As a confirmation of the functional role of this NLS, mutant full-length 69-kDa ChAT R363A,R364A,R366A,K368A -eGFP has an entirely cytoplasmic distribution with the nuclei devoid of the protein, in contrast to distribution of the wild-type enzyme where the eGFP-tagged ChAT is present in the nucleus (Fig. 4B, upper panels). Quantitative evaluation of the subcellular distribution of the eGFP-tagged wild-type and the mutant proteins was conducted by analyzing fluorescence intensities across the cells in confocal images and plotting these as ratios of fluorescence intensity in the nucleus compared with the cytoplasm. As illustrated in the lower panels of Fig. 4, mutation of the putative NLS identified in both Construct-6 and the wild-type enzyme led to redistribution of the eGFP-tagged proteins in the cells with significantly less fluorescence in the nucleus (p ϭ 0.02, n ϭ 4). Analysis of the ratio of fluorescence in the nucleus to the cytoplasm revealed a value of 0.90 Ϯ 0.01 for mutant Construct-6 compared with 2.54 Ϯ 0.48 for wild-type Construct-6 (mean Ϯ S.E., n ϭ 4). This provides further evidence of the roughly equal distribution of the mutant protein between the two cellular compartments and indicates that the NLS was abolished by this mutation. Similar comparisons made with the full-length enzyme yielded nuclear/cytoplasm fluorescence ratios of 0.52 Ϯ 0.02 and 0.06 Ϯ 0.004 for the wild-type and mutant proteins, respectively (p ϭ 0.005, n ϭ 4), indicating significantly reduced nuclear accumulation of the mutant fulllength protein. These data provide evidence that the NLS sequence of 358 ELPAPRRLRWK 368 is responsible for the nuclear distribution seen for the wild-type Construct-6 (encompassing amino acids 345-410) and for wild-type 69-kDa ChAT.
A predominantly cytoplasmic localization was observed for Construct-9 (amino acids 584 -630) indicating the presence of potential NES activity in this portion of the ChAT protein (Fig.  3B). To address further the functional nature of this potential NES in the context of the full-length ChAT protein, a deletion mutant was prepared in which the carboxyl terminus of the FIG. 4. A, site-directed mutagenesis of critical amino acid residues in Construct-6 (amino acids 385-410) leads to the loss of functional NLS activity. Four critical arginine and lysine residues were mutated to alanine residues: R363A, R364A, R366A, and K368A. When the mutant construct was transfected into HEK 293 cells, the fusion protein expressed assumed a more homogeneous subcellular distribution indicating that the NLS activity had been abolished. When fluorescence intensity in the nucleus and the cytoplasm were calculated and expressed as a ratio (fluorescence intensity of the nucleus/fluorescence intensity of the cytoplasm), a statistically significant difference was found between distribution of the wild-type Construct-6 protein compared with the mutant Construct-6 protein (p ϭ 0.02). C denotes cytoplasm, and N denotes nucleus. B, the same mutations were performed on the full-length 69-kDa ChAT cDNA in peGFP-N1. Visualization of the mutant fusion protein by confocal microscopy revealed that this putative NLS directs the full-length enzyme into the nucleus because the mutant enzyme was found entirely in the cytoplasm with the nuclei devoid of eGFP fluorescence. Fluorescence intensity using profile imaging of eGFP fluorescence indicates that the mutated full-length 69-kDa ChAT protein is essentially absent in the nucleus of cells when compared with the wild-type 69-kDa ChAT (p ϭ 0.005). These data are representative of at least 4 separate experiments; quantitative fluorescence profiles were calculated from 4 independent experiments. enzyme was removed beyond amino acid residue 583. We predicted that if this carboxyl-terminal portion of ChAT was responsible for export of the protein from the nucleus, then we would observe an increased presence of the eGFP-tagged mutant ChAT in the nucleus. However, when cells were transfected with this construct, it yielded a subcellular distribution pattern for the mutant protein that was essentially similar to that obtained for the wild-type enzyme (data not shown). The predominantly cytoplasmic localization of this ChAT mutant suggests that this carboxyl-terminal portion of the protein does not generate functional NES activity in the context of the full-length protein under these experimental conditions. Consequently, ChAT may contain an NES that was not detected in this assay, or it may not contain an NES and instead bind to another NES-containing protein to be co-transported across the nuclear membrane into the cytoplasm.
Previous studies from this laboratory (9) showed that 82-kDa ChAT contains a functional NLS within the first nine amino acids at the amino terminus of the protein; when these residues were deleted from the full-length protein, 82-kDa ChAT was found predominantly in the cytoplasm. Further confirmation of the functional nature of this amino-terminal region of 82-kDa ChAT was obtained in this study. Expression of these nine amino acids (MGLRTAKKR) tagged to eGFP yielded a fusion protein that had a predominantly nuclear distribution (Fig.  5A), similar to that observed for full-length 82-kDa ChAT (Fig.  5B). As the functional NLS domain that we identified in 69-kDa ChAT is also present in 82-kDa ChAT (amino acid residues 481-486), we investigated whether this common NLS also influences the subcellular compartmentalization of full-length 82-kDa ChAT. To address this, we mutated the same critical amino acid residues as in the wild-type 69-kDa ChAT yielding the mutant construct 82-kDa ChAT R481A,R482A,R484A,K486A -eGFP. Comparison of the subcellular distribution of the mutant and wild-type 82-kDa ChAT in HEK 293 cells by confocal microscopy revealed a predominantly cytoplasmic distribution for the mutant fusion protein (Fig. 5C) in contrast to a predominantly nuclear distribution of the wild-type protein (Fig. 5B). This indicates that this novel NLS, common to both forms of the ChAT protein, is also involved in the nuclear transport of 82-kDa ChAT. Therefore, deleting or mutating either of the two functional NLSs identified in 82-kDa ChAT results in cytosolic distribution similar to 69-kDa ChAT. Thus both NLSs are necessary to give a predominantly nuclear distribution and retention of 82-kDa ChAT protein.
Fluorescence Recovery after Photobleaching (FRAP) Assay-A variety of proteins shuttle between the nucleus and the cytoplasm in the cell, including hnRNP (37), histone deacetylase-7 (38), QKI-5 protein (39), FKBP12-rapamycin-associated protein (40), and Smad1 (41). The FRAP assay has been successfully adapted and characterized previously as an experimental method for evaluating whether proteins that have a predominantly nuclear localization can move out of the nucleus to the cytoplasm and then shuttle back into the nucleus (27). We used the FRAP assay in the present studies to test if 82-kDa ChAT has the ability to shuttle from the nucleus to the cytoplasm and between nuclei in bikaryon cells. Briefly, HEK 293 cells were co-transfected with peGFP-82-kDa ChAT and DsRed1-hnRNP; hnRNP has been characterized previously to be a nuclear shuttling protein and thus serves as a positive control in the experiment (28). Transfected cells were fused together with 50% polyethylene glycol treatment to form bikaryon cells. By using confocal microscopy, the nuclei of bikaryon cells that contain both 82-kDa ChAT-eGFP and DsRed1-hnRNP exhibit a yellow fluorescence (overlay of red and green). Therefore, in the bikaryon cell there are two distinct nuclear pools of our protein of interest. One of the two nuclei in the fused cells was targeted and subsequently photobleached, after which the cells were allowed a recovery period of 1 h and examined for evidence of fluorescence recovery in the photobleached nucleus. If the protein of interest is a shuttling protein, it should translocate from the unbleached nucleus through the cytoplasm to the bleached nucleus resulting in recovery of fluorescence in the bleached nucleus. In our experiments, recovery of yellow fluorescence in the bleached nuclei would indicate that both the eGFP-tagged ChAT and the DsRed1-tagged hnRNP proteins shuttled out of the unbleached nucleus and were transported into the bleached nucleus. As shown in Fig. 6, at 60 min after photobleaching, we observed the recovery of red fluorescence in the bleached nucleus. This confirms that shuttling pathways in the bikaryon cell were intact allowing the positive control DsRed1-hnRNP to translocate into the bleached nucleus. However, we did not observe recovery of green fluorescence (seen as yellow in the presence of the positive control) indicating that a measurable amount of 82-kDa ChAT does not shuttle between the nuclei. Based on this observation, it appears that 82-kDa ChAT present in the nucleus remains there, and unlike 69-kDa ChAT, it does not move back to the cytoplasm along nuclear export pathways.
ChAT Is Catalytically Active in the Nucleus-Intact nuclei were isolated from wild-type HEK 293 cells or cells stably expressing 69-or 82-kDa ChAT and incubated with choline and [ 3 H]acetyl-CoA to determine whether the enzymes are catalytically active and can catalyze O-acetylation of choline within the nucleus. As illustrated in Fig. 7A, [ 3 H]ACh was synthesized in nuclei isolated from cells expressing either form of ChAT but not from wild-type HEK 293 cells. Previous studies (9) demonstrated that 69-and 82-kDa ChAT tagged with GFP had similar specific activities and that activity of the enzymes in the cytoplasm and the nucleus reflected the relative abundance of these proteins in the subcellular compartments.
We also analyzed the nuclear fractions for ChAT protein by immunoblot to determine whether either form of the enzyme undergoes proteolytic processing in the nucleus. In particular, we wanted to establish whether 82-kDa ChAT was degraded to yield a smaller form of the enzyme by cleavage of the unique amino-terminal domain from the protein. As shown in Fig. 7B, both 69-and 82-kDa ChAT were found in nuclei with the apparent molecular mass expected for the full-length forms of FIG. 5. The novel NLS identified in 69-kDa ChAT is also present and functional in 82-kDa ChAT. Previously, we determined by deletion mutagenesis that the first nine amino acids at the amino terminus of 82-kDa ChAT contained a functional NLS (9); this NLS is unique to 82-kDa ChAT. A, we show that when these nine amino acids are fused to eGFP the resulting protein product 1 MGLRTAKKR 9 -eGFP is localized predominantly to nucleus. This is compared with the subcellular localization of the wild-type 82-kDa enzyme shown in B. To test if the novel NLS also influenced the localization of 82-kDa ChAT, site-directed mutagenesis was used to produce the mutant construct peGFP-82-kDa ChAT R481A,R482A,R484A,K486A , with the subcellular distribution of its protein product illustrated in C. Whereas the wild-type enzyme was localized largely in the nucleus (B), the mutant protein was found predominantly in the cytoplasm (C). These data indicate that function of both NLSs is required to lead to nuclear accumulation of 82-kDa ChAT. These images are representative of data obtained in 3 independent experiments. the proteins. Based on this immunoblot analysis, it appears that the levels of 82-kDa ChAT are lower than that of 69-kDa ChAT in the two cell lines used for these experiments. However, immunocytochemical analysis of the cell lines using the anti-ChAT antibody CTab revealed that many cells in the 82-kDa ChAT-expressing cell line did not express the enzyme in detectable amounts, whereas the enzyme was expressed in most cells in the 69-kDa ChAT-expressing cell line. This likely also accounts for the lower level of catalytic activity observed in the nuclei from cells expressing the 82-kDa ChAT compared with nuclei from cells expressing the 69-kDa ChAT.
To date, a functional role for ChAT in the nucleus has not been determined. Some possible functions can be considered. In the present study, we found that the nuclear 69-and 82-kDa ChAT pools are catalytically active within the context of the nucleus and can produce ACh when provided with choline and acetyl coenzyme A as substrates in this experimental situation. There is evidence that some G-protein-coupled receptors including muscarinic receptors can be located on the nuclear envelope (42,43), thus offering the potential for ACh produced in the nucleus to serve as a ligand if the receptor-binding site was oriented to the inner surface of the nucleus. Whereas the two substrates for ACh biosynthesis are likely present in the nucleus, further studies are needed to determine whether this reaction normally occurs in cholinergic neurons. To date, choline has been identified as the endogenous substrate for ChAT with this enzyme carrying out O-acetylation. It is not known whether ChAT could have alternative substrates or other biological functions in the nucleus, or whether this enzyme is capable of catalyzing N-acetylation as well as O-acetylation.
Finally, whereas 82-kDa ChAT is localized predominantly in the nucleus, 69-kDa ChAT shuttles into and out of the nucleus with the bulk of the protein located in the cytoplasm. This differential subcellular distribution could be a way of regulating availability of both proteins for participation in ACh biosynthesis or alternative cellular functions.
In the present study, we determined that the 69-kDa ChAT protein contains a single NLS motif responsible for its entry into the nucleus. However, based on its subcellular distribution, 69-kDa ChAT also leaves the nucleus along nuclear export pathways. We identified one of these to be the LMB-sensitive Crm-1 transporter, but it is possible that this enzyme also shuttles out of the nucleus by other mechanisms. Conversely, movement of 69-kDa ChAT into and out of the nucleus could be regulated and occur in response to changes in cellular function. For example, stimulus-mediated activation of signal transduction pathways could lead to post-translational modification and conformational rearrangement of the 69-kDa enzyme. This could result in favorable exposure of the NLS causing an increase in the translocational driving force for the protein into the nucleus under defined cellular conditions. Alternatively, post-translational modification of 69-kDa ChAT could lead to altered interaction of the enzyme with another protein(s) that undergoes nuclear import. These types of regulatory mechanisms can be seen for a wide range of other nuclear shuttling proteins. For example, NFB is retained in the cytoplasm by binding with IB until IB is phosphorylated by mitogen-activated protein kinase leading to the release of NFB (44); the Smad1 protein is cytoplasmic until it undergoes phosphorylation (41); histone deacetylase-4 nuclear translocation is facilitated by binding to 14-3-3 which shuttles it into the ChAT present in the nuclei retained the ability to produce ACh, with no choline acetylating activity measured in the nuclei from wild-type cells. These data represent the mean Ϯ S.E. of 4 independent experiments. Statistically significant differences were found between cells expressing either form of ChAT and wild-type cells (analysis of variance, p ϭ 0.0056, n ϭ 4). B, to determine whether either 69-and 82-kDa ChAT proteins undergo proteolytic processing in the nucleus, we analyzed the nuclear fractions studied for enzymatic activity by immunoblot. In this representative blot, it is clear that both 69-and 82-kDa ChAT were found in the nuclei at the apparent molecular mass expected for the full-length forms of the proteins. nucleus (45), and both 4E-T (46) and mitogen-activated protein kinase-activated protein (MAPKAP) (47) redistribute between the nuclear and cytoplasmic compartments following phosphorylation. ChAT is a substrate for a number of protein kinases, with phosphorylation of the enzyme regulating its catalytic activity and membrane association (48 -50); it is not know if phosphorylation of ChAT regulates its partitioning between the cytoplasm and the nucleus. It is also not know whether nuclear accumulation of 69-kDa ChAT can be regulated by an endogenous factor or whether physiological or pathological perturbations of the neuron alter steady-state distribution of the enzyme. For example, if a required co-factor that regulates nuclear transport of 69-kDa ChAT is expressed in low abundance, it is likely then that only a small proportion of the total cellular content of the enzyme may be translocated into the nucleus.
Furthermore, we have now identified a second functional NLS motif in 82-kDa ChAT. The distribution of the 82-kDa enzyme differs substantially from that of 69-kDa ChAT. This suggests either that the NLS that is unique to the 82-kDa protein in association with the newly identified NLS is involved in a process leading to a predominantly nuclear accumulation for the protein or that the presence of two NLSs in the 82-kDa protein are sufficient to overcome the driving force(s) resulting in nuclear export. We have also established that differential subcellular localization of 69-and 82-kDa human ChAT occurs in neural cells and is not a function of expression and mislocalization of the proteins in HEK 293 cells; subcellular distribution of the two forms of ChAT was tested in the current study in IMR32 human neuroblastoma cells (data not shown) and previously in PC12 cells (9) and found to be indistinguishable from the patterns observed in HEK 293 cells. The M-ChAT transcript isoform encoding both 69-and 82-kDa ChAT has been found in the human brain and spinal cord by reverse transcriptase-PCR (6). It has been shown previously that this transcript is translated to both forms of the enzyme in cultured cells (6,9) as well as in vitro (6). However, the presence of 82-kDa ChAT in cholinergic neurons and tissues has not yet been demonstrated. The M-ChAT transcript appears to be the most abundantly expressed human ChAT mRNA in brain and spinal cord, but the presence of the two translation initiation sites lowers the translational efficiency of this transcript. Consequently, it was determined by in vitro translation studies that the expression of 69-and 82-kDa ChAT from the M-ChAT transcript proceeds at about one-fifth of the rate of production of the 69-kDa ChAT from the N1-ChAT transcript (6). This suggests that expression levels of 82-kDa ChAT would be considerably lower than those of the 69-kDa enzyme.
In summary, we investigated the ability of 69-kDa ChAT to shuttle between the nuclear and cytoplasmic compartments of the cell, thereby redefining 69-kDa ChAT as a nucleocytoplasmic protein. Additionally, we investigated the functional role of the novel NLS identified in 69-and 82-kDa ChAT in the subcellular compartmentalization of 82-kDa ChAT. We illustrated that the nuclear localization of the 82-kDa ChAT protein appears to require the combined effect of both NLS motifs that translocate 82-kDa ChAT into the nucleus and may lead to its retention there.