Cell Biochem Biophys DOI 10.1007/s12013-014-9999-z
ORIGINAL PAPER
Membrane Topology of Human Presenilin-1 in SK-N-SH Cells Determined by Fluorescence Correlation Spectroscopy and Fluorescent Energy Transfer Krishna Midde • Ryan Rich • Ashwini Saxena Ignacy Gryczynski • Julian Borejdo • Hriday K. Das
•
Ó Springer Science+Business Media New York 2014
Abstract Presenilin-1 (PS1) protein acts as passive ER Ca2? leak channels that facilitate passive Ca2? leak across ER membrane. Mutations in the gene encoding PS1 protein cause neurodegeneration in the brains of patients with familial Alzheimer’s disease (FAD). FADPS1 mutations abrogate the function of ER Ca2? leak channel activity in human neuroblastoma SK-N-SH cells in vitro (Das et al., J Neurochem 122(3):487–500, 2012) and in mouse embryonic fibroblasts. Consequently, genetic deletion or mutations of the PS1 gene cause calcium (Ca2?) signaling abnormalities leading to neurodegeneration in FAD patients. By analogy with other known ion channels it has been proposed that the functional PS1 channels in ER may be multimers of several PS1 subunits. To test this hypothesis, we conjugated the human PS1 protein with an NH2-terminal YFP-tag and a COOH-terminal CFP-tag. As expected YFP–PS1, and PS1–CFP were found to be expressed on the plasma membranes by TIRF microscopy, and both these fusion proteins increased ER Ca2? leak
channel activity similar to PS1 (WT) in SK-N-SH cells, as determined by functional calcium imaging. PS1–CFP was either expressed alone or together with YFP–PS1 into SKN-SH cell line and the interaction between YFP–PS1 and PS1–CFP was determined by Fo¨rster resonance energy transfer analysis. Our results suggest interaction between YFP–PS1 and PS1–CFP confirming the presence of a dimeric or multimeric form of PS1 in SK-N-SH cells. Lateral diffusion of PS1–CFP and YFP–PS1 in the plasma membrane of SK-N-SH cells was measured in the absence or in the presence of glycerol by fluorescence correlation spectroscopy to show that both COOH-terminal and NH2terminal of human PS1 are located on the cytoplasmic side of the plasma membrane. Therefore, we conclude that both COOH-terminal and NH2-terminal of human PS1 may also be oriented on the cytosolic side of ER membrane.
K. Midde R. Rich I. Gryczynski J. Borejdo Department of Cell Biology & Immunology and Center for Commercialization of Fluorescence Technologies, UNT Health Science Center, 3500 Camp Bowie Boulevard, Fort Worth, TX 76107, USA
H. K. Das (&) Department of Pharmacology & Neuroscience, UNT Health Science Center, 3500 Camp Bowie Boulevard, Fort Worth, TX 76107, USA e-mail: [email protected]
Present Address: K. Midde Department of Medicine, George E. Palade Laboratories for Cell and Molecular Medicine, Room 331-333, 9500 Gilman Drive, La Jolla, CA 92093-7636, USA
H. K. Das Institute of Cancer Research, UNT Health Science Center, 3500 Camp Bowie Boulevard, Fort Worth, TX 76107, USA
A. Saxena Department of Physiology & Anatomy, UNT Health Science Center, 3500 Camp Bowie Boulevard, Fort Worth, TX 76107, USA
Keywords Presenilin-1 Membrane topology TIRF microscopy FRET FCS Lateral diffusion
H. K. Das Institute of Aging and Alzheimer’s Disease Research, UNT Health Science Center, 3500 Camp Bowie Boulevard, Fort Worth, TX 76107, USA
123
Cell Biochem Biophys
Abbreviations AD Alzheimer’s disease APP Amyloid precursor protein CFP Cyan fluorescent protein ER Endoplasmic reticulum FAD Familial Alzheimer’s disease FCS Fluorescence correlation spectroscopy FRET Fo¨rster resonance energy transfer PBS Phosphate buffer saline MEF Mouse embryonic fibroblast NT NH2-terminal PCR Polymerase chain reaction PS1 Presenilin-1 PS1CTF COOH-terminal fragment of PS1 PS1FL Full length PS1 PS1NTF NH2-terminal fragment of PS1 TIRF Total Internal Reflection Fluorescence WT Wild type YFP Yellow fluorescent protein
Introduction Alzheimer’s disease (AD) afflicts nearly 5 million Americans. Gamma (c)-secretase enzyme-mediated cleavage of amyloid precursor protein (APP) has been associated with the neurodegeneration observed in AD [1]. The catalytic subunit of c-secretase is presenilin-1 (PS1) protein [2]. Sequential cleavage of APP by b- and c-secretase enzymes generates b-amyloid (Ab) peptide fragments. Ab is the primary component of amyloid plaques and its accumulation leads to neuron degeneration in the brain during AD [3, 4]. Most of the AD cases are sporadic and characterized by late onset, but a small fraction of AD cases are familial (FAD). FAD cases are characterized by an early onset and genetic inheritance. Majority of the FAD cases are due to mutations in the coding sequences of the PS1 gene [5]. In addition to enhanced APP processing, many FADPS1 mutations lead to decreased ERCa2? leak, increased ERCa2? pool, decreased cytosolic Ca2?, and deranged neuronal Ca2? signaling [6–8], which facilitate amyloid accumulation in the brain, neuronal atrophy, and cell death [7]. Therefore, changes in neuronal Ca2? may play an early and important role in the pathogenesis of AD and FAD [7]. FAD-PS1 mutations also cause pathogenesis of AD by augmenting PS1-mediated c-secretase activity in AD brains [9]. We and others have recently reported that PS1 can function as passive ER Ca2? leak channels [6, 8]. In wild type cells steady state levels of ER intraluminal Ca2? are determined by a balance of between sarco/endoplasmic reticulum Ca2?-ATPase (SERCA) to pump Ca2? into ER lumen from cytoplasm and PS1-facilitated passive Ca2?
123
leak from ER lumen to cytoplasm [7]. FAD-PS1 mutations abrogate the function of passive ER Ca2? leak from ER lumen to cytoplasm which results in higher steady state intraluminal Ca2? levels, lower cytosolic Ca2?, deranged Ca2? signaling, leading to amyloid accumulation, neuronal atrophy, and neuronal cell death [7]. PS1 is a *50-kDa transmembrane protein and resides in the endoplasmic reticulum (ER) membrane [10]. PS1 undergoes endoproteolytic cleavage to NH2-terminal (*30-kDa-NTF) and COOH-terminal (*20-kDa-CTF) fragments. The cleaved PS1 subsequently forms complex with nicastrin (NCT), Aph-1, and Pen2. This PS1 complex moves from ER to plasma membrane where it functions as membrane-embedded c-secretase enzyme [1, 11]. On the other hand, PS1 which is expressed only on the ER membrane acts as passive ER Ca2? channels. Structure function analysis based on NMR study [12] and biochemical data suggest that PS1 has ten hydrophobic domains from which nine domains cross the membrane as transmembrane domains (TMD) [13–17]. Endoproteolysis of PS1 takes place in the large loop between TMD6 and TMD7 [18]. TMD6 and TMD7 form a water-containing cavity inside the membrane with two catalytic aspartates responsible for c-secretase activity face each other [18]. According to this nine TMD model of mouse PS1, the NH2terminal of PS1FL, or the PS1NTF fragment after endoproteolysis of PS1FL has been localized on the cytosolic side of the plasma membrane whereas the COOH-terminal of PS1FL or PS1CTF fragment after endoproteolysis of PS1FL has extracellular localization [13, 18]. Although the topology of the PS1 protein has been determined by biochemical analysis, we do not know much about how PS1 forms ER Ca2? channels on the ER membrane. By analogy with other known ion channels it has been proposed that the functional PS1 channels may be multimers of several PS1 subunits. Herl et al. [19] used co-immunoprecipitation and fluorescent lifetime imaging microscopy to show that PS1 forms a dimer. These authors used CHO cells transfected with either (His-PS1NT ? Flag-PS1NT) or (His-PS1 loop ? Flag-PS1FL) followed by anti-His and anti-Flag staining and Fo¨rster resonance energy transfer (FRET) analysis to determine the dimer formation of PS1 [19]. In contrast, we used SK-N-SH cells transfected with (YFP–PS1FL ? PS1FL–CFP) followed by live cell imaging with confocal microscopy, TIRF microscopy, and FRET analysis to show that YFP–PS1FL and PS1FL–CFP fusion proteins were expressed on the plasma membrane and formed dimer or multimers in plasma membrane. Calcium imaging of SK-N-SH cells transiently transfected with either YFP–PS1FL or PS1FL–CFP showed increased ERCa2? leak channel activity similar to PS1 (WT) suggesting that attachment of YFP or CFP to PS1FL retained the ERCa2? channel function of PS1 (WT) protein. This result also
Cell Biochem Biophys
Materials and Methods
transiently transfected with 10 lg of pCDNA3.YFP–PS1 or 10 lg of pCDNA3.PS1–CFP or (10 lg of pCDNA3.PS1–CFP ? 10 lg of pCDA3.YFP–PS1) constructs by lipofectamine 2000 (Invitrogen, CA) according to manufacturer’s guide. Transfection of these constructs generate two distinct PS1–CFP and YFP–PS1 fusion proteins in SK-N-SH cells. After 48 h of transfection, cells in the coverslips were washed with PBS (137 mM NaCl, 2.7 mM KCl, 10 mM Na2HPO4, 2 mM KH2PO4) and used for confocal microscopy, TIRF microscopy, calcium imaging, FRET analysis, and diffusion measurements.
Construction of YFP–PS1 and PS1–CFP Expression Vectors
Confocal Microscopic Analysis to Determine the Cellular Localization of YFP–PS1 and PS1–CFP
pCDNA3-PS1 expression vector containing full length (FL) human PS1 cDNA was received from Dr. Michael Wolfe (Harvard Medical School). cDNA encoding FL PS1 protein (amino acid 1–467) containing the stop codon were amplified by PCR using a forward primer 50 -GAT CCCGCGG.atg.aca.gag.tta.cct.gca.ccg-30 and a reverse primer 50 -GATCCTCGAG.aat.atg.cta.gat.ata.aaa.ttg.atg.g. The PCR product was digested with SacII and XhoI and purified on an agarose gel. Gel purified DNA was ligated with SacII and XhoI digested pcDNA3/NT.YFP vector (Gift from Dr. Robert Luedtke, UNTHSC). pcDNA3-YFP vector contains YFP sequence 50 to the Sac II site. The YFP sequence is also fused to a signal peptide at the 50 end in order to express (peptide signal-YFP–PS1) fusion protein in the membrane. For the preparation of PS1–CFP construct, cDNA encoding FL PS1 protein (amino acid 1–467) without the stop codon were amplified by PCR using a forward primer 50 GATCGAATC.atg.aca.gag.tta.cct.gca.ccg-30 and a reverse primer 50 -GATCTCTAGA.gat.ata.aaa.ttg.atg.gaa.tgc-30 . The PCR product was digested with EcoR1 and XbaI and purified on an agarose gel. Gel purified DNA was ligated with EcoR1 and XbaI digested pcDNA3/CT.CFP vector (Invitrogen, CA). Orientations and the sequence of PS1 were determined by DNA sequencing of entire construct.
SK-N-SH cells co-transfected with YFP–PS1 and PS1– CFP and non-transfected cells (control) were grown to 70 % confluence on 25 mm glass circular cover slips (#1). Growth media was aspirated and cells were washed with PBS and imaged under confocal laser scanning microscope—Carl Zeiss LSM 510. PS1–CFP protein was visualized by excitation with 430 nm laser diode and CFP fluorescence was observed through a 480 nm band-pass filter. YFP–PS1 protein was visualized by excitation at 488 nm with an Ar laser and fluorescence was detected through a 520 nm long pass filter. Laser beam was focused to diffraction-limited spot via a 409/1.2 W confocal objective (C-Apochromat). Out-of-plane fluorescence was rejected by using 100 lm confocal aperture. Axial scans with 1 lm thickness (Z-stacks) were collected and optical slices were reconstructed to determine the expression and localization of YFP–PS1 and PS1–CFP proteins.
suggests that YFP–PS1 and PS1–CFP fusion proteins are expressed on ER membrane. Lateral diffusion of PS1FL– CFP and YFP–PS1FL-in the plasma membrane of SK-N-SH cells was measured using fluorescence correlation spectroscopy (FCS). Results from FCS appear to suggest that both the COOH-terminal and the NH2-terminal of human PS1FL are located on the cytoplasmic side of the plasma membrane.
Cell Culture Human neuroblastoma SK-N-SH cell line was maintained in Dulbecco’s modified Eagle medium (DMEM) containing 10 % fetal bovine serum, 1 % penicillin/streptomycin (Gibco, CA). Transient Transfection of YFP–PS1 and PS1–CFP Constructs SK-N-SH cells were seeded in culture dishes on to glass coverslips the day before transfection. The cells were
TIRF Microscopic Analysis to Determine the Localization of YFP–PS1 and PS1–CFP on the Plasma Membrane of SK-N-SH Cells SK-N-SH cells were grown to 70 % confluence on 22 mm square glass cover slips. Cells were transiently transfected with YFP–PS1 and PS1–CFP. After 48 h of transfection, cells in the coverslips were washed with PBS (137 mM NaCl, 2.7 mM KCl, 10 mM Na2HPO4, 2 mM KH2PO4) and used for TIRF microscopy. Olympus TIRF illuminator coupled to Olympus IX71 microscope was used. To make sure TIRF condition was established, 100 nm FITC-polystyrene microspheres (Molecular Probes) were examined. Spheres were diluted 100 times with water. When TIRF angle was too small, microspheres together with significant amount of background fluorescence were visible. TIRF angle was progressively increased until only a few blinking microspheres on top of a weak background were visible. Blinking indicated that TIRF volume was established: when entering it the spheres were illuminated with
123
Cell Biochem Biophys
evanescent wave and were clearly visible. When leaving the volume, spheres were no longer visible. Ratiometric Calcium Imaging SK-N-SH cells were grown on 35 mm diameter glass bottom plates (MatTek Corporation, Ashland, MA) to 70–80 % confluency. The cells were transiently transfected with PS1 (WT), YFP–PS1, PS1–CFP for 24 h. Cells were serum deprived overnight before the day of experiment. On the day of experiment, coverslips with adherent SK-N-SH cells were washed with HBSS (Life Technologies, Grand Island, NY) for 30 min at 37 °C. The HBSS composition is (in mM): 1.26 CaCl2, 0.49 MgCl26H2O, 0.4 MgSO4 7H2O, 5.3 KCl, 0.44 KH2PO4, 137.9 NaCl, 0.33 Na2HPO47H2O, 5.5 D-glucose, and pH 7.4. Then cells were incubated with calcium-sensitive dye Fura-2AM (3 lM; Life technologies) for 30 min at 37 °C. The cells were then washed twice with HBSS and incubated at 37 °C for an additional 20 min in HBSS, for complete hydrolysis of Fura-2AM. The plates were then mounted on a motorized inverted microscope (Olympus IX81). The cells were viewed at 209 magnification and alternatively illuminated with 340 and 380 nm wavelengths using a xenon light source (Prior Lumen200PRO). The emitted light was captured at 520 nm wavelength using CCD camera (Hamamatsu camera controller C10600). After achieving at least 2 min of stabilized baseline data, cells were stimulated with 1 lM thapsigargin (Sigma Aldrich). The data were collected every 2 s using commercially available software (Slidebook 5.0, Intelligent Imaging Innovations Inc., Denver, CO). The ratio values (340/380 nm) were then normalized to the baseline values averaged for 2 min before the response to the drug was observed, as previously described [20]. Measurement of Fluorescence Resonance Energy Transfer The NH2- and COOH-termini of PS1FL protein were labeled with YFP and CFP, respectively. Energy transfer (ET) was monitored by the change of fluorescent lifetime of the donor in a single live cell. Lifetime was measured in Alba-ISS (Urbana, IL) confocal microscope. The sample was imaged using 409, Numerical Aperture 1.15 water immersion objective (Olympus UApo/340). Efficiency of ET was calculated as E = 1 - sDA/sD where sDA and sD are the fluorescence lifetimes of the donor in the presence and absence of acceptor, respectively. The distance r between donor and acceptor was calculated r = Ro(1/ E - 1)1/6 where Ro is the characteristic distance between ˚) . CFP and YFP (49 A
123
Measurement of Diffusion Coefficients of YFP–PS1 and PS1–CFP Proteins in the Plasma Membrane SK-N-SH cells were grown to 70 % confluence on sterile glass 20 mm square cover slips. Dynamics of YFP–PS1 and PS1–CFP protein in the cell membrane were assessed with and without 80 % glycerol in PBS. Alba-ISS (Champaign, IL) Fluorescence Correlation Spectrophotometer attached to Olympus IX71 microscope was used to measure the lateral diffusion of YFP–PS1 and PS1–CFP proteins in cell membrane. The optical alignment of Alba FCS spectrophotometer was similar to the setup previously used [21]. Fluorescence intensity fluctuations were collected at the rate of 1 photon/10 ls for a total of 20 s in two avalanche photo-detectors (APD-Perkin Elmer). Autocorrelation function of the time traces were plotted in VistaVision version 4.0.119 ISS Alba inbuilt software to compute diffusion coefficients. Statistical Analysis Statistical analyses were calculated using GraphPad software. Students unpaired t test was used to determine if the difference between groups is significant or not. p value of less than 0.05 at 95 % confidence interval was considered statistically significant.
Results YFP–PS1 and PS1–CFP are Expressed and Colocalized in ER and Plasma Membrane of SK-N-SH Cells Transiently Transfected with pCDNA3.YFP–PS1 and pCDNA3.PS1–CFP Laser scanning confocal imaging of SK-N-SH cells was performed to study the expression and co-localization of PS1–CFP and YFP–PS1 proteins. Data were collected at single molecule level from a single cell to avoid complications from averaging and inhomogeneities in ensemble studies. Using this approach, interaction of NH2-terminal and COOH-terminal of PS1 protein and its sub cellular location can be precisely determined. Figure 1a and b shows the membrane expression of PS1–CFP and YFP– PS1 fusion proteins, respectively. The merge of the CFP and YFP fluorescence in panel C (yellow) of Fig. 1 indicates co-expression and co-localization of PS1 protein. Colocalization of both proteins supports the idea of transient dimer or multimer formation of PS1 in the plasma and ER membranes. TIRF microscopy also confirms that YFP–PS1 and PS1– CFP are expressed and localized on the plasma membrane of SK-N-SH cells (Fig. 2).
Cell Biochem Biophys
Fig. 1 Expression and co-localization of YFP–PS1 and PS1–CFP in ER and plasma membrane of SK-N-SH cells. Laser scanning confocal imaging of SK-N-SH cells was performed to study the expression and colocalization of PS1–CFP and YFP–PS1 proteins. a represents a single SK-N-SH cell imaged when excited at 430 nm laser to show CFP
fluorescence (green) alone. b represents the same cell, but excited by 510 nm laser to detect the YFP–PS1 protein expression (red). c represents the merge of A and B depicting the co-localization (yellow) of both proteins to support the idea of transient dimer or multimer formation of PS1 in the plasma and ER membranes (Color figure online)
Fig. 2 TIRF microscopy confirms that YFP–PS1 and PS1–CFP are expressed on plasma membrane of SK-N-SH cells transiently transfected with pCDNA3YFP–PS1 and pCDNA3PS1–CFP. For TIRF microscopy, cells were grown on 22 mm glass coverslips (VWR International), fixed in 2 % ice cold paraformaldehyde in PBS for 10 min at 4 °C. The cells were observed under the Olympus IX71 microscope equipped with a commercial TIRF attachment, using
Olympus 60 9 NA = 1.45 PlanApo oil objective and Hamamatsu C4742-95 high-resolution digital camera utilizing a progressive scan interline transfer CCD chip with no mechanical shutter, and Peltier cooling. Images were acquired with identical image acquisition parameters to monitor differences in the fluorescence intensity between treated and untreated cells. The contrast was enhanced using GIMP2 program
123
Cell Biochem Biophys
Fig. 3 Thapsigargin-induced cytosolic Ca2? mobilization of PS1 (WT), YFP–PS1, and PS1–CFP in SK-N-SH cells. Transient transfection with PS1 (WT) or YFP–PS1 or PS1–CFP expression constructs shows thapsigargin (1 lM) induced increase in cytosolic calcium in SK-N-SH cells. After achieving stable baseline, all ratiometric data was normalized to the baseline ratios averaged over 2 min. After recording 2 min of stable baseline data, thapsigargin (1 lM) was applied to the cells. a All transiently transfected groups demonstrated exaggerated increase in cytosolic calcium as compared to pCDNA3-vector transfected control cells (red circles; n = 82). The cells transfected with PS1 (WT) (orange circle; n = 21), PS1– CFP (dark green upright triangles; n = 40), and YFP–PS1 (light green inverted triangles; n = 23) showed increased calcium
mobilization in the presence of thapsigargin. b The difference between mean baseline ratio and maximal ratio (D340/380 nm) was calculated and differences between groups were measured using oneway analysis of variance (ANOVA) and Dunn’s pairwise multiple comparison test. Statistical significance was set at p \ 0.05. D340/ 380 nm was noted to be higher in all transfected groups in comparison with control group (0.39 ± 0.016; p \ 0.001). There was no statistically significant difference between PS1–CFP (0.99 ± 0.061; light green bar) and YFP–PS1 (0.79 ± 0.025; dark green bar) groups. Data is presented as mean ± SEM. Asterisk significant difference from control. Dagger significant difference from PS1 (Color figure online)
Transient Transfection of YFP–PS1 and PS1–CFP Increases Thapsigargin-Induced Cytosolic Ca2? Mobilization Similar to PS1 (WT) in SK-N-SH Cells
efficiencies of YFP–PS1, PS1–CFP, and PS1 (WT) expression constructs. These calcium imaging experiments confirm that attachment of YFP or CFP to PS1FL does not alter the ERCa2? leak channel activity function of PS1 (WT) protein.
We and others have reported that presenilins (PS1/PS2) act as passive ERCa2? leak channels to promote passive leak of Ca2? from ER lumen to cytoplasm [6, 8]. To confirm that the ERCa2? leak channel activity of PS1 is not altered due to the attachment of a NH2-terminal YFP-tag (YFP– PS1) and a COOH-terminal CFP-tag (PS1–CFP), we performed Ca2? imaging experiments with SK-N-SH cell line transiently transfected with pCDNA3 vector or pCDNA3.PS1 or pCDNA3.YFP–PS1 or pCDNA3.PS1– CFP expression construct by lipofectamine 2000. Ca2? imaging was performed 48 h post-transfection. Cells were loaded with fura-2-AM Ca2? imaging dye and intracellular calcium mobilization was measured by measuring normalized 340/380 nm ratio before and after the application of thapsigargin. Over-expression of YFP–PS1 or PS1–CFP increased thapsigargin mediated Ca2? mobilization peak very similar to PS1 (WT) transfected cells relative to vector-transfected cells (Fig. 3a, b). In YFP–PS1 and PS1– CFP groups the final intracellular calcium was less than PS1 (WT), but it was higher than the control group. These differences could be due to variation in transfection
123
FRET Analysis to Show Protein–Protein Interaction Between PS1–CFP and YFP–PS1 Molecules Expression of FAD mutants of PS1 alone or co-expression of both wild type PS1 and FAD mutants of PS1 abrogate the ability of PS1 to conduct Ca2? ions supporting for Ca2? hypothesis of AD [6, 8]. By analogy with other known ion channels it has been proposed that the functional PS1 channels may be multimers of several PS1 subunits. We have used ET in live cells to determine if PS1 forms dimer or multimers on cell membranes (ER or plasma membranes). We have conjugated the human PS1 protein with an NH2-terminal YFP-tag and a COOH-terminal CFP-tag. PS1–CFP were expressed alone or together with YFP–PS1 into human neuroblastoma SK-N-SH cell line and interaction between YFP–PS1 and PS1–CFP was determined by measuring lifetime of donor (PS1–CFP) in absence or presence of acceptor (YFP–PS1) in a single live cell (as in Fig. 1). ET was monitored by the change of fluorescent
Cell Biochem Biophys
A
B
C
Donoor
E
D
Donor + Acceptor
D Donor
Donnor
F
Donor + Acceptor
Donor + Acceptor
Fig. 4 Fo¨rster resonance energy transfer was measured in SK-N-SH cells expressing NH2-terminal YFP–PS1 and COOH-terminal PS1– CFP fusion proteins. Fluorescence life-times of the donor CFP was measured in live cells with and without cells expressing YFPacceptor. Data was collected on ISS Alba confocal fluorescence correlation spectrometer with micrometer spatial resolution and nanosecond temporal resolution to measure the dynamics of Presenelin-1 protein subdomains at single molecule regime. A 470 nm frequency modulated polarized excitation was used to measure
lifetimes in frequency domain (FD)—fluorescence lifetime imaging microscopy (FLIM) mode. Panels a, b, c and d, e, f in figure are representative plots of phase delay and modulation ratio of donor fluorescence alone and donor plus acceptor, respectively. The quench in the fluorescence lifetime of donor in the presence of acceptor suggests energy transfer and interaction of the NH2-terminal and ˚ ) units, presumably due to formation COOH-terminal (*r = 63.9 A of a membrane channel (Color figure online)
lifetime of the donor. Lifetime was measured in Alba-ISS (Urbana, IL) confocal microscope in frequency domain. First, SK-N-SH cells transfected with PS1–CFP alone were imaged (FLIM) and lifetimes were obtained sD = *2.014 ± 0.11 ns (average of three independent experiments). Panels a, b, and c of Fig. 4 are the phase delay and modulation ratio plots obtained in three independent experiments. Next, the lifetimes of the donor (CFP) was calculated from SK-N-SH cells co-transfected with PS1–CFP and YFP–PS1YFP–PS1 (panels d, e, and f of Fig. 4). The average lifetime was calculated to be sDA = 1.649 ± 0.058 ns. Table 1 summarizes the lifetimes obtained and the FRET measurements and results. We must note that the lifetime varied across the image and we choose ET areas because in their vicinity the life time was approximately constant. The efficiency of ET E = 1 - sDA/sD was *20 %. The distance between donor and acceptor was ˚ . It has been demonstrated previously that the *63.9 A NH2-terminal of PS1FL is localized on the cytoplasmic side of the plasma membrane. Since in our study, one PS1 molecule is PS1–CFP and another PS1 molecule is YFP– PS1, our FRET data appear to suggest that both the NH2-
terminal and COOH-terminal of PS1 molecule are present on the cytoplasmic side of the trans-membrane as reported by Doan et al. [22] and PS1 forms either dimer or multimers on the cellular membrane. Our FRET data is in agreement with the eight TMD topology of PS1 [23, 24]. It is to be noted that there exist several other models including 6TMD, 7TMD, 8TMD, and 9TMD topology of PS1 [13, 14, 17, 18, 23, 24]. Diffusion Studies in the Absence and in the Presence of Glycerol Show that Both the NH2-Terminal and the COOH-Terminal of PS1 are Located on the Cytoplasmic Side of Plasma Membrane In order to determine on which side of the cell membrane the PS1 protein termini are located, we measured lateral diffusion of PS1–CFP in the plasma membrane of SK-NSH cells by FCS (Fig. 5). Alba-ISS confocal spectrophotometer with avalanche photodiodes capable of single molecule detection SMD) was used to measure the lateral motion of PS1 protein in the plasma membrane. The number of molecules in the elliptical confocal volume (ECV) was determined from the inverse of the
123
Cell Biochem Biophys Table 1 Lifetimes of PS1–CFP donor in SK-N-SH cells with and without acceptor YFP–PS1 were used to calculate the transfer efficiency (ET) and the distance (r) between the CFP and YFP proteins Experiment
s (CFP-D)
s (CFP ? YFP) (DA)
ET = 1 - (sDA/sD)
r = Ro [(1/ET) - 1]1/6 (nm)
1
2.145 (v2 .787) (a)
1.582 (v2 .197) (d)
2
0.263
5.837
2
2
1.934 (v .564) (b)
1.686 (v .216) (e)
0.129
6.75
3
1.964 (v2 .77) (c)
1.679 (v2 .95) (f)
0.146
6.596
s fluorescence life-time, Ro Forster distance at which energy transfer efficiency is 50 %, D represents donor, DA represents donor and acceptor
Fig. 5 Identification of the NH2-terminal and the COOHterminal of PS1 on the cytoplasmic side of plasma membrane. In order to determine on which side of the cell membrane the protein termini are located, we measured lateral diffusion of PS1–CFP and YFP–PS1 in the plasma membrane of SK-N-SH cells by fluorescence correlation spectroscopy (FCS). a Confocal image of SK-N-SH cells expressing PS1–CFP. Addition of 80 % glycerol to the cells that express b PS1–CFP and c YFP– PS1 did not change the diffusion of the protein (*0.15 lm2/s), suggesting that both the NH2and COOH-terminus are facing the cytoplasmic side of the plasma membrane. The number of molecules under observation is kept low (to avoid averaging) and was determined by the inverse of autocorrelation function at time 0 (Color figure online)
autocorrelation function at time 0 [21]. In our experiments the number of molecules under observation was limited to *10 (average of 15 experiments). Diffusion studies of PS1–CFP protein in SK-N-SH cells were performed in PBS and 80 % glycerol in PBS. Both sets of experiments could be well fitted with standard 2D diffusion equation. Time traces obtained from single point profile from the cells were used to compute autocorrelation function. From the autocorrelation function of the CFP and YFP proteins in PBS and 80 % glycerol, the diffusion coefficients were calculated through ISS Vista Vision Version 4.0.00209 custom software. The diffusion coefficients of both the proteins, when measured in PBS and immediately after addition of 80 % glycerol were similar. The average diffusion coefficient obtained for YFP–PS1 protein was 0.15 lm2/s out of 15 experiments in both the control and
123
80 % glycerol group (p = 0.8404, t = 0.2033 and there was no statistical significance between the groups determined by students t test). Similarly, the average diffusion coefficient for PS1–CFP protein was determined to be 0.15 lm2/s out of 15 experiments in both the control and 80 % glycerol group (p = 0.8894, t = 0.1404 and there was no statistical significance between the groups determined by students t test). This suggests that the viscous glycerol does not inhibit the motion of the NH2- or COOH-terminus of the PS1 protein in the plasma membrane, indicating that they are internalized and face the cytoplasm. This result contradicts the previously published biochemical data showing the cytoplasmic localization of the NH2-terminal only. Although, glycerol will not enter through the plasma membrane to alter the diffusion of YFP–PS1 and PS1–CFP
Cell Biochem Biophys
proteins expressed on the ER membrane, we presume that both the NH2- and the COOH-termini of PS1 protein may be oriented toward the cytosolic side of the ER membrane.
Discussion PS1 protein was shown to be expressed on the membranes of the ER and the Golgi apparatus [25]. To date several studies have been undertaken to determine the intracellular membrane topology of PS1 protein expressed on the ER membrane [13, 14, 17, 18, 23, 24]. All published models agree that the NH2-terminus and the large hydrophobic loop of PS1 are localized in the cytoplasmic side of the ER membrane. In contrast, diverse opinions exist with regard to the number of TMD and the location of the COO2terminus of PS1 protein [24]. Doan et al. [22] have transfected chimeric NH2-terminal PS1–APP fusion construct in cultured COS-1 cells to demonstrate that the NH2terminal epitope specific antibody interacts with PS1 on the cytosolic site of the ER membrane. This data suggests the cytoplasmic localization of NH2-terminal of PS1. Similar experiments by these authors also lead to conclude that PS1 may contain either six or eight TMD with NH2terminal, COOH-terminal, and long hydrophobic loop are oriented toward the cytoplasmic side of the ER membrane [22]. Strooper et al. [25] have demonstrated by immunocytochemical analysis that PS1 protein is mainly expressed in the ER and the Golgi apparatus. These authors have transfected COS1 and CHO cells myc-PS1FL or PS1FLmyc constructs followed by treatment of cells with low concentrations of digitonin to selectively permeabilize the plasma membrane and antibodies against NH2-terminal domain of PS1 or the myc-tag introduced at the NH2-terminus or the COOH-terminus of PS1 [25]. Their results showed that antibody against the NH2-terminal inserted myc-tag reacted with PS1 in digitonin-permeabilized cells suggesting that the NH2-terminal of PS1 was localized on the cytoplasmic side of the ER membrane. Similar experiments by these authors have shown that the long hydrophobic loop is also located on the cytoplasmic side of the ER membrane. In contrast, this model of PS1 by De Strooper et al. [25] predicts a 7-TMD topology of PS1 with the COOH-terminal of PS1 protein being localized on the lumen side of the ER membrane. A nine TMD topology for PS1 protein has been determined using more stable N-linked glycosylation scanning approach [13]. Glycosylation acceptor sequences were introduced into FLPS1, and the results show that NH2terminus and large hydrophobic loop are oriented are present in the cytosolic side of the ER membrane but the COOH-terminal of the PS1 is localized to the luminal side
of the ER membrane. Studies on glycosylation pattern after TMD deletions followed by transfection into cells derived from PS1-/- and PS2-/-mouse blastocytes, combined with computer-based TMD protein topology predictions and biotinylation assays of different PS1 mutants convinced the authors that PS1 has nine TMD and COO2terminus locates to the lumen/extracellular space [13]. A 9-TMD topology with NH2-terminal and the long hydrophobic loop being on the cytoplasmic side and the COOHterminal end on the lumen side of the ER membrane has been confirmed by Spasic et al. [14]. Although, the 9-TMD topology of PS1 with NH2-terminus in the cytoplasmic and COOH-terminal on the lumen side of the ER is now widely accepted, it is to be noted that several other models including 6TMD, 7TMD, 8TMD, and 9TMD topology of PS1 have also been proposed [13, 14, 17, 18, 23, 24]. Some of these models also predict that both the NH2-terminus and COOH-terminus of PS1 are oriented to the cytosolic side of ER [22, 24, 26]. To date we know very little about the topology of PS1 on the plasma membrane. We have used biophysical approach to identify the membrane topology of PS1 on the plasma membrane. Our results indicate that both YFP–PS1 and PS1–CFP chimeric proteins are expressed on the plasma membrane and intracellular membranes (Figs. 1, 2). We also present the lateral diffusion studies of PS1–CFP in the presence and the absence of 80 % glycerol. Since the lateral diffusion of YFP–PS1 and PS1–CFP chimeric proteins was not altered after addition of glycerol (Fig. 5), these data suggest that the CFP-tagged COOH-terminal and YFP-tagged NH2-terminal of PS1 are oriented on the cytoplasmic side of the plasma membrane. We also observed that attachment of YFP or CFP to PS1FL also increased the ERCa2? channel activity very similar to PS1 (WT) protein (Fig. 3). Taken these results together, we conclude that passive ERCa2? channels are formed by either dimerization or multimerization of PS1 proteins with both the NH2-terminus and COOH-terminus of PS1 oriented toward the cytosolic side of the ER. Acknowledgments We wish to thank Dr. Michael Wolfe (Harvard Medical School) for providing us with pCDNA3.1-PS1 expression vector. This research was supported by NIH Grant R01HL090786 to Dr. Julian Borejdo and research support from Graduate School of Biomedical Sciences of UNTHSC to Dr. Hriday K. Das. Mr. Krishna Midde is supported by Predoctoral Fellowship 12PRE8730003 from American Heart Association.
References 1. De Strooper, B., Saftig, P., Craessaerts, K., Vanderstichele, H., Guhde, G., Annaert, W., et al. (1998). Deficiency of presenilin-1 inhibits the normal cleavage of amyloid precursor protein. Nature, 391(6665), 387–390.
123
Cell Biochem Biophys 2. Wolfe, M. S., Xia, W., Ostaszewski, B. L., Diehl, T. S., Kimberly, W. T., & Selkoe, D. J. (1999). Two transmembrane aspartates in presenilin-1 required for presenilin endoproteolysis and gamma-secretase activity. Nature, 398(6727), 513–517. 3. Selkoe, D. J. (1997). Alzheimer’s disease: Genotypes, phenotypes, and treatments. Science, 275(5300), 630–631. 4. Selkoe, D. J., & Schenk, D. (2003). Alzheimer’s disease: Molecular understanding predicts amyloid-based therapeutics. Annual Review of Pharmacology and Toxicology, 43, 545–584. 5. Sherrington, R., Rogaev, E. I., Liang, Y., Rogaeva, E. A., Levesque, G., Ikeda, M., et al. (1995). Cloning of a gene bearing missense mutations in early-onset familial Alzheimer’s disease. Nature, 375(6534), 754–760. 6. Tu, H., Nelson, O., Bezprozvanny, A., Wang, Z., Lee, S. F., Hao, Y. H., et al. (2006). Presenilins form ER Ca2? leak channels, a function disrupted by familial Alzheimer’s disease-linked mutations. Cell, 126(5), 981–993. 7. Supnet, C., & Bezprozvanny, I. (2010). The dysregulation of intracellular calcium in Alzheimer disease. Cell Calcium, 47(2), 183–189. 8. Das, H. K., Tchedre, K., & Mueller, B. (2012). Repression of transcription of presenilin-1 inhibits gamma-secretase independent ER Ca(2)(?) leak that is impaired by FAD mutations. Journal of Neurochemistry, 122(3), 487–500. 9. Duff, K., Eckman, C., Zehr, C., Yu, X., Prada, C. M., Perez-tur, J., et al. (1996). Increased amyloid-beta42(43) in brains of mice expressing mutant presenilin 1. Nature, 383(6602), 710–713. 10. Annaert, W. G., Levesque, L., Craessaerts, K., Dierinck, I., Snellings, G., Westaway, D., et al. (1999). Presenilin 1 controls gamma-secretase processing of amyloid precursor protein in pregolgi compartments of hippocampal neurons. Journal of Cell Biology, 147(2), 277–294. 11. De Strooper, B. (2003). Aph-1, Pen-2, and Nicastrin with Presenilin generate an active gamma-Secretase complex. Neuron, 38(1), 9–12. 12. Sobhanifar, S., Schneider, B., Lohr, F., Gottstein, D., Ikeya, T., Mlynarczyk, K., et al. (2010). Structural investigation of the C-terminal catalytic fragment of presenilin 1. Proceedings of the National Academy of Sciences USA, 107(21), 9644–9649. 13. Laudon, H., Hansson, E. M., Melen, K., Bergman, A., Farmery, M. R., Winblad, B., et al. (2005). A nine-transmembrane domain topology for presenilin 1. Journal of Biological Chemistry, 280(42), 35352–35360. 14. Spasic, D., Tolia, A., Dillen, K., Baert, V., De Strooper, B., Vrijens, S., et al. (2006). Presenilin-1 maintains a nine-transmembrane topology throughout the secretory pathway. Journal of Biological Chemistry, 281(36), 26569–26577.
123
15. Tolia, A., Chavez-Gutierrez, L., & De Strooper, B. (2006). Contribution of presenilin transmembrane domains 6 and 7 to a water-containing cavity in the gamma-secretase complex. Journal of Biological Chemistry, 281(37), 27633–27642. 16. Tolia, A., & De Strooper, B. (2009). Structure and function of gamma-secretase. Seminars in Cell & Developmental Biology, 20(2), 211–218. 17. Bammens, L., Chavez-Gutierrez, L., Tolia, A., Zwijsen, A., & De Strooper, B. (2011). Functional and topological analysis of Pen-2, the fourth subunit of the gamma-secretase complex. Journal of Biological Chemistry, 286(14), 12271–12282. 18. Nelson, O., Supnet, C., Tolia, A., Horre, K., De Strooper, B., & Bezprozvanny, I. (2011). Mutagenesis mapping of the presenilin 1 calcium leak conductance pore. Journal of Biological Chemistry, 286(25), 22339–22347. 19. Herl, L., Lleo, A., Thomas, A. V., Nyborg, A. C., Jansen, K., Golde, T. E., et al. (2006). Detection of presenilin-1 homodimer formation in intact cells using fluorescent lifetime imaging microscopy. Biochemical and Biophysical Research Communications, 340(2), 668–674. 20. Story, G. M., Peier, A. M., Reeve, A. J., Eid, S. R., Mosbacher, J., Hricik, T. R., et al. (2003). ANKTM1, a TRP-like channel expressed in nociceptive neurons, is activated by cold temperatures. Cell, 112(6), 819–829. 21. Midde, K., Luchowski, R., Das, H. K., Fedorick, J., Dumka, V., Gryczynski, I., et al. (2011). Evidence for pre- and post-power stroke of cross-bridges of contracting skeletal myofibrils. Biophysical Journal, 100(4), 1024–1033. 22. Doan, A., Thinakaran, G., Borchelt, D. R., Slunt, H. H., Ratovitsky, T., Podlisny, M., et al. (1996). Protein topology of presenilin 1. Neuron, 17(5), 1023–1030. 23. Li, X., & Greenwald, I. (1996). Membrane topology of the C. elegans SEL-12 presenilin. Neuron, 17(5), 1015–1021. 24. Kim, J., & Schekman, R. (2004). The ins and outs of presenilin 1 membrane topology. Proceedings of the National Academy of Sciences USA, 101(4), 905–906. 25. De Strooper, B., Beullens, M., Contreras, B., Levesque, L., Craessaerts, K., Cordell, B., et al. (1997). Phosphorylation, subcellular localization, and membrane orientation of the Alzheimer’s disease-associated presenilins. Journal of Biological Chemistry, 272(6), 3590–3598. 26. Li, L., Cheung, T., Chen, J., & Herrup, K. (2011). A comparative study of five mouse models of Alzheimer’s disease: Cell cycle events reveal new insights into neurons at risk for death. International Journal of Alzheimer’s Disease, 2011, 171464.
ORIGINAL PAPER
Membrane Topology of Human Presenilin-1 in SK-N-SH Cells Determined by Fluorescence Correlation Spectroscopy and Fluorescent Energy Transfer Krishna Midde • Ryan Rich • Ashwini Saxena Ignacy Gryczynski • Julian Borejdo • Hriday K. Das
•
Ó Springer Science+Business Media New York 2014
Abstract Presenilin-1 (PS1) protein acts as passive ER Ca2? leak channels that facilitate passive Ca2? leak across ER membrane. Mutations in the gene encoding PS1 protein cause neurodegeneration in the brains of patients with familial Alzheimer’s disease (FAD). FADPS1 mutations abrogate the function of ER Ca2? leak channel activity in human neuroblastoma SK-N-SH cells in vitro (Das et al., J Neurochem 122(3):487–500, 2012) and in mouse embryonic fibroblasts. Consequently, genetic deletion or mutations of the PS1 gene cause calcium (Ca2?) signaling abnormalities leading to neurodegeneration in FAD patients. By analogy with other known ion channels it has been proposed that the functional PS1 channels in ER may be multimers of several PS1 subunits. To test this hypothesis, we conjugated the human PS1 protein with an NH2-terminal YFP-tag and a COOH-terminal CFP-tag. As expected YFP–PS1, and PS1–CFP were found to be expressed on the plasma membranes by TIRF microscopy, and both these fusion proteins increased ER Ca2? leak
channel activity similar to PS1 (WT) in SK-N-SH cells, as determined by functional calcium imaging. PS1–CFP was either expressed alone or together with YFP–PS1 into SKN-SH cell line and the interaction between YFP–PS1 and PS1–CFP was determined by Fo¨rster resonance energy transfer analysis. Our results suggest interaction between YFP–PS1 and PS1–CFP confirming the presence of a dimeric or multimeric form of PS1 in SK-N-SH cells. Lateral diffusion of PS1–CFP and YFP–PS1 in the plasma membrane of SK-N-SH cells was measured in the absence or in the presence of glycerol by fluorescence correlation spectroscopy to show that both COOH-terminal and NH2terminal of human PS1 are located on the cytoplasmic side of the plasma membrane. Therefore, we conclude that both COOH-terminal and NH2-terminal of human PS1 may also be oriented on the cytosolic side of ER membrane.
K. Midde R. Rich I. Gryczynski J. Borejdo Department of Cell Biology & Immunology and Center for Commercialization of Fluorescence Technologies, UNT Health Science Center, 3500 Camp Bowie Boulevard, Fort Worth, TX 76107, USA
H. K. Das (&) Department of Pharmacology & Neuroscience, UNT Health Science Center, 3500 Camp Bowie Boulevard, Fort Worth, TX 76107, USA e-mail: [email protected]
Present Address: K. Midde Department of Medicine, George E. Palade Laboratories for Cell and Molecular Medicine, Room 331-333, 9500 Gilman Drive, La Jolla, CA 92093-7636, USA
H. K. Das Institute of Cancer Research, UNT Health Science Center, 3500 Camp Bowie Boulevard, Fort Worth, TX 76107, USA
A. Saxena Department of Physiology & Anatomy, UNT Health Science Center, 3500 Camp Bowie Boulevard, Fort Worth, TX 76107, USA
Keywords Presenilin-1 Membrane topology TIRF microscopy FRET FCS Lateral diffusion
H. K. Das Institute of Aging and Alzheimer’s Disease Research, UNT Health Science Center, 3500 Camp Bowie Boulevard, Fort Worth, TX 76107, USA
123
Cell Biochem Biophys
Abbreviations AD Alzheimer’s disease APP Amyloid precursor protein CFP Cyan fluorescent protein ER Endoplasmic reticulum FAD Familial Alzheimer’s disease FCS Fluorescence correlation spectroscopy FRET Fo¨rster resonance energy transfer PBS Phosphate buffer saline MEF Mouse embryonic fibroblast NT NH2-terminal PCR Polymerase chain reaction PS1 Presenilin-1 PS1CTF COOH-terminal fragment of PS1 PS1FL Full length PS1 PS1NTF NH2-terminal fragment of PS1 TIRF Total Internal Reflection Fluorescence WT Wild type YFP Yellow fluorescent protein
Introduction Alzheimer’s disease (AD) afflicts nearly 5 million Americans. Gamma (c)-secretase enzyme-mediated cleavage of amyloid precursor protein (APP) has been associated with the neurodegeneration observed in AD [1]. The catalytic subunit of c-secretase is presenilin-1 (PS1) protein [2]. Sequential cleavage of APP by b- and c-secretase enzymes generates b-amyloid (Ab) peptide fragments. Ab is the primary component of amyloid plaques and its accumulation leads to neuron degeneration in the brain during AD [3, 4]. Most of the AD cases are sporadic and characterized by late onset, but a small fraction of AD cases are familial (FAD). FAD cases are characterized by an early onset and genetic inheritance. Majority of the FAD cases are due to mutations in the coding sequences of the PS1 gene [5]. In addition to enhanced APP processing, many FADPS1 mutations lead to decreased ERCa2? leak, increased ERCa2? pool, decreased cytosolic Ca2?, and deranged neuronal Ca2? signaling [6–8], which facilitate amyloid accumulation in the brain, neuronal atrophy, and cell death [7]. Therefore, changes in neuronal Ca2? may play an early and important role in the pathogenesis of AD and FAD [7]. FAD-PS1 mutations also cause pathogenesis of AD by augmenting PS1-mediated c-secretase activity in AD brains [9]. We and others have recently reported that PS1 can function as passive ER Ca2? leak channels [6, 8]. In wild type cells steady state levels of ER intraluminal Ca2? are determined by a balance of between sarco/endoplasmic reticulum Ca2?-ATPase (SERCA) to pump Ca2? into ER lumen from cytoplasm and PS1-facilitated passive Ca2?
123
leak from ER lumen to cytoplasm [7]. FAD-PS1 mutations abrogate the function of passive ER Ca2? leak from ER lumen to cytoplasm which results in higher steady state intraluminal Ca2? levels, lower cytosolic Ca2?, deranged Ca2? signaling, leading to amyloid accumulation, neuronal atrophy, and neuronal cell death [7]. PS1 is a *50-kDa transmembrane protein and resides in the endoplasmic reticulum (ER) membrane [10]. PS1 undergoes endoproteolytic cleavage to NH2-terminal (*30-kDa-NTF) and COOH-terminal (*20-kDa-CTF) fragments. The cleaved PS1 subsequently forms complex with nicastrin (NCT), Aph-1, and Pen2. This PS1 complex moves from ER to plasma membrane where it functions as membrane-embedded c-secretase enzyme [1, 11]. On the other hand, PS1 which is expressed only on the ER membrane acts as passive ER Ca2? channels. Structure function analysis based on NMR study [12] and biochemical data suggest that PS1 has ten hydrophobic domains from which nine domains cross the membrane as transmembrane domains (TMD) [13–17]. Endoproteolysis of PS1 takes place in the large loop between TMD6 and TMD7 [18]. TMD6 and TMD7 form a water-containing cavity inside the membrane with two catalytic aspartates responsible for c-secretase activity face each other [18]. According to this nine TMD model of mouse PS1, the NH2terminal of PS1FL, or the PS1NTF fragment after endoproteolysis of PS1FL has been localized on the cytosolic side of the plasma membrane whereas the COOH-terminal of PS1FL or PS1CTF fragment after endoproteolysis of PS1FL has extracellular localization [13, 18]. Although the topology of the PS1 protein has been determined by biochemical analysis, we do not know much about how PS1 forms ER Ca2? channels on the ER membrane. By analogy with other known ion channels it has been proposed that the functional PS1 channels may be multimers of several PS1 subunits. Herl et al. [19] used co-immunoprecipitation and fluorescent lifetime imaging microscopy to show that PS1 forms a dimer. These authors used CHO cells transfected with either (His-PS1NT ? Flag-PS1NT) or (His-PS1 loop ? Flag-PS1FL) followed by anti-His and anti-Flag staining and Fo¨rster resonance energy transfer (FRET) analysis to determine the dimer formation of PS1 [19]. In contrast, we used SK-N-SH cells transfected with (YFP–PS1FL ? PS1FL–CFP) followed by live cell imaging with confocal microscopy, TIRF microscopy, and FRET analysis to show that YFP–PS1FL and PS1FL–CFP fusion proteins were expressed on the plasma membrane and formed dimer or multimers in plasma membrane. Calcium imaging of SK-N-SH cells transiently transfected with either YFP–PS1FL or PS1FL–CFP showed increased ERCa2? leak channel activity similar to PS1 (WT) suggesting that attachment of YFP or CFP to PS1FL retained the ERCa2? channel function of PS1 (WT) protein. This result also
Cell Biochem Biophys
Materials and Methods
transiently transfected with 10 lg of pCDNA3.YFP–PS1 or 10 lg of pCDNA3.PS1–CFP or (10 lg of pCDNA3.PS1–CFP ? 10 lg of pCDA3.YFP–PS1) constructs by lipofectamine 2000 (Invitrogen, CA) according to manufacturer’s guide. Transfection of these constructs generate two distinct PS1–CFP and YFP–PS1 fusion proteins in SK-N-SH cells. After 48 h of transfection, cells in the coverslips were washed with PBS (137 mM NaCl, 2.7 mM KCl, 10 mM Na2HPO4, 2 mM KH2PO4) and used for confocal microscopy, TIRF microscopy, calcium imaging, FRET analysis, and diffusion measurements.
Construction of YFP–PS1 and PS1–CFP Expression Vectors
Confocal Microscopic Analysis to Determine the Cellular Localization of YFP–PS1 and PS1–CFP
pCDNA3-PS1 expression vector containing full length (FL) human PS1 cDNA was received from Dr. Michael Wolfe (Harvard Medical School). cDNA encoding FL PS1 protein (amino acid 1–467) containing the stop codon were amplified by PCR using a forward primer 50 -GAT CCCGCGG.atg.aca.gag.tta.cct.gca.ccg-30 and a reverse primer 50 -GATCCTCGAG.aat.atg.cta.gat.ata.aaa.ttg.atg.g. The PCR product was digested with SacII and XhoI and purified on an agarose gel. Gel purified DNA was ligated with SacII and XhoI digested pcDNA3/NT.YFP vector (Gift from Dr. Robert Luedtke, UNTHSC). pcDNA3-YFP vector contains YFP sequence 50 to the Sac II site. The YFP sequence is also fused to a signal peptide at the 50 end in order to express (peptide signal-YFP–PS1) fusion protein in the membrane. For the preparation of PS1–CFP construct, cDNA encoding FL PS1 protein (amino acid 1–467) without the stop codon were amplified by PCR using a forward primer 50 GATCGAATC.atg.aca.gag.tta.cct.gca.ccg-30 and a reverse primer 50 -GATCTCTAGA.gat.ata.aaa.ttg.atg.gaa.tgc-30 . The PCR product was digested with EcoR1 and XbaI and purified on an agarose gel. Gel purified DNA was ligated with EcoR1 and XbaI digested pcDNA3/CT.CFP vector (Invitrogen, CA). Orientations and the sequence of PS1 were determined by DNA sequencing of entire construct.
SK-N-SH cells co-transfected with YFP–PS1 and PS1– CFP and non-transfected cells (control) were grown to 70 % confluence on 25 mm glass circular cover slips (#1). Growth media was aspirated and cells were washed with PBS and imaged under confocal laser scanning microscope—Carl Zeiss LSM 510. PS1–CFP protein was visualized by excitation with 430 nm laser diode and CFP fluorescence was observed through a 480 nm band-pass filter. YFP–PS1 protein was visualized by excitation at 488 nm with an Ar laser and fluorescence was detected through a 520 nm long pass filter. Laser beam was focused to diffraction-limited spot via a 409/1.2 W confocal objective (C-Apochromat). Out-of-plane fluorescence was rejected by using 100 lm confocal aperture. Axial scans with 1 lm thickness (Z-stacks) were collected and optical slices were reconstructed to determine the expression and localization of YFP–PS1 and PS1–CFP proteins.
suggests that YFP–PS1 and PS1–CFP fusion proteins are expressed on ER membrane. Lateral diffusion of PS1FL– CFP and YFP–PS1FL-in the plasma membrane of SK-N-SH cells was measured using fluorescence correlation spectroscopy (FCS). Results from FCS appear to suggest that both the COOH-terminal and the NH2-terminal of human PS1FL are located on the cytoplasmic side of the plasma membrane.
Cell Culture Human neuroblastoma SK-N-SH cell line was maintained in Dulbecco’s modified Eagle medium (DMEM) containing 10 % fetal bovine serum, 1 % penicillin/streptomycin (Gibco, CA). Transient Transfection of YFP–PS1 and PS1–CFP Constructs SK-N-SH cells were seeded in culture dishes on to glass coverslips the day before transfection. The cells were
TIRF Microscopic Analysis to Determine the Localization of YFP–PS1 and PS1–CFP on the Plasma Membrane of SK-N-SH Cells SK-N-SH cells were grown to 70 % confluence on 22 mm square glass cover slips. Cells were transiently transfected with YFP–PS1 and PS1–CFP. After 48 h of transfection, cells in the coverslips were washed with PBS (137 mM NaCl, 2.7 mM KCl, 10 mM Na2HPO4, 2 mM KH2PO4) and used for TIRF microscopy. Olympus TIRF illuminator coupled to Olympus IX71 microscope was used. To make sure TIRF condition was established, 100 nm FITC-polystyrene microspheres (Molecular Probes) were examined. Spheres were diluted 100 times with water. When TIRF angle was too small, microspheres together with significant amount of background fluorescence were visible. TIRF angle was progressively increased until only a few blinking microspheres on top of a weak background were visible. Blinking indicated that TIRF volume was established: when entering it the spheres were illuminated with
123
Cell Biochem Biophys
evanescent wave and were clearly visible. When leaving the volume, spheres were no longer visible. Ratiometric Calcium Imaging SK-N-SH cells were grown on 35 mm diameter glass bottom plates (MatTek Corporation, Ashland, MA) to 70–80 % confluency. The cells were transiently transfected with PS1 (WT), YFP–PS1, PS1–CFP for 24 h. Cells were serum deprived overnight before the day of experiment. On the day of experiment, coverslips with adherent SK-N-SH cells were washed with HBSS (Life Technologies, Grand Island, NY) for 30 min at 37 °C. The HBSS composition is (in mM): 1.26 CaCl2, 0.49 MgCl26H2O, 0.4 MgSO4 7H2O, 5.3 KCl, 0.44 KH2PO4, 137.9 NaCl, 0.33 Na2HPO47H2O, 5.5 D-glucose, and pH 7.4. Then cells were incubated with calcium-sensitive dye Fura-2AM (3 lM; Life technologies) for 30 min at 37 °C. The cells were then washed twice with HBSS and incubated at 37 °C for an additional 20 min in HBSS, for complete hydrolysis of Fura-2AM. The plates were then mounted on a motorized inverted microscope (Olympus IX81). The cells were viewed at 209 magnification and alternatively illuminated with 340 and 380 nm wavelengths using a xenon light source (Prior Lumen200PRO). The emitted light was captured at 520 nm wavelength using CCD camera (Hamamatsu camera controller C10600). After achieving at least 2 min of stabilized baseline data, cells were stimulated with 1 lM thapsigargin (Sigma Aldrich). The data were collected every 2 s using commercially available software (Slidebook 5.0, Intelligent Imaging Innovations Inc., Denver, CO). The ratio values (340/380 nm) were then normalized to the baseline values averaged for 2 min before the response to the drug was observed, as previously described [20]. Measurement of Fluorescence Resonance Energy Transfer The NH2- and COOH-termini of PS1FL protein were labeled with YFP and CFP, respectively. Energy transfer (ET) was monitored by the change of fluorescent lifetime of the donor in a single live cell. Lifetime was measured in Alba-ISS (Urbana, IL) confocal microscope. The sample was imaged using 409, Numerical Aperture 1.15 water immersion objective (Olympus UApo/340). Efficiency of ET was calculated as E = 1 - sDA/sD where sDA and sD are the fluorescence lifetimes of the donor in the presence and absence of acceptor, respectively. The distance r between donor and acceptor was calculated r = Ro(1/ E - 1)1/6 where Ro is the characteristic distance between ˚) . CFP and YFP (49 A
123
Measurement of Diffusion Coefficients of YFP–PS1 and PS1–CFP Proteins in the Plasma Membrane SK-N-SH cells were grown to 70 % confluence on sterile glass 20 mm square cover slips. Dynamics of YFP–PS1 and PS1–CFP protein in the cell membrane were assessed with and without 80 % glycerol in PBS. Alba-ISS (Champaign, IL) Fluorescence Correlation Spectrophotometer attached to Olympus IX71 microscope was used to measure the lateral diffusion of YFP–PS1 and PS1–CFP proteins in cell membrane. The optical alignment of Alba FCS spectrophotometer was similar to the setup previously used [21]. Fluorescence intensity fluctuations were collected at the rate of 1 photon/10 ls for a total of 20 s in two avalanche photo-detectors (APD-Perkin Elmer). Autocorrelation function of the time traces were plotted in VistaVision version 4.0.119 ISS Alba inbuilt software to compute diffusion coefficients. Statistical Analysis Statistical analyses were calculated using GraphPad software. Students unpaired t test was used to determine if the difference between groups is significant or not. p value of less than 0.05 at 95 % confidence interval was considered statistically significant.
Results YFP–PS1 and PS1–CFP are Expressed and Colocalized in ER and Plasma Membrane of SK-N-SH Cells Transiently Transfected with pCDNA3.YFP–PS1 and pCDNA3.PS1–CFP Laser scanning confocal imaging of SK-N-SH cells was performed to study the expression and co-localization of PS1–CFP and YFP–PS1 proteins. Data were collected at single molecule level from a single cell to avoid complications from averaging and inhomogeneities in ensemble studies. Using this approach, interaction of NH2-terminal and COOH-terminal of PS1 protein and its sub cellular location can be precisely determined. Figure 1a and b shows the membrane expression of PS1–CFP and YFP– PS1 fusion proteins, respectively. The merge of the CFP and YFP fluorescence in panel C (yellow) of Fig. 1 indicates co-expression and co-localization of PS1 protein. Colocalization of both proteins supports the idea of transient dimer or multimer formation of PS1 in the plasma and ER membranes. TIRF microscopy also confirms that YFP–PS1 and PS1– CFP are expressed and localized on the plasma membrane of SK-N-SH cells (Fig. 2).
Cell Biochem Biophys
Fig. 1 Expression and co-localization of YFP–PS1 and PS1–CFP in ER and plasma membrane of SK-N-SH cells. Laser scanning confocal imaging of SK-N-SH cells was performed to study the expression and colocalization of PS1–CFP and YFP–PS1 proteins. a represents a single SK-N-SH cell imaged when excited at 430 nm laser to show CFP
fluorescence (green) alone. b represents the same cell, but excited by 510 nm laser to detect the YFP–PS1 protein expression (red). c represents the merge of A and B depicting the co-localization (yellow) of both proteins to support the idea of transient dimer or multimer formation of PS1 in the plasma and ER membranes (Color figure online)
Fig. 2 TIRF microscopy confirms that YFP–PS1 and PS1–CFP are expressed on plasma membrane of SK-N-SH cells transiently transfected with pCDNA3YFP–PS1 and pCDNA3PS1–CFP. For TIRF microscopy, cells were grown on 22 mm glass coverslips (VWR International), fixed in 2 % ice cold paraformaldehyde in PBS for 10 min at 4 °C. The cells were observed under the Olympus IX71 microscope equipped with a commercial TIRF attachment, using
Olympus 60 9 NA = 1.45 PlanApo oil objective and Hamamatsu C4742-95 high-resolution digital camera utilizing a progressive scan interline transfer CCD chip with no mechanical shutter, and Peltier cooling. Images were acquired with identical image acquisition parameters to monitor differences in the fluorescence intensity between treated and untreated cells. The contrast was enhanced using GIMP2 program
123
Cell Biochem Biophys
Fig. 3 Thapsigargin-induced cytosolic Ca2? mobilization of PS1 (WT), YFP–PS1, and PS1–CFP in SK-N-SH cells. Transient transfection with PS1 (WT) or YFP–PS1 or PS1–CFP expression constructs shows thapsigargin (1 lM) induced increase in cytosolic calcium in SK-N-SH cells. After achieving stable baseline, all ratiometric data was normalized to the baseline ratios averaged over 2 min. After recording 2 min of stable baseline data, thapsigargin (1 lM) was applied to the cells. a All transiently transfected groups demonstrated exaggerated increase in cytosolic calcium as compared to pCDNA3-vector transfected control cells (red circles; n = 82). The cells transfected with PS1 (WT) (orange circle; n = 21), PS1– CFP (dark green upright triangles; n = 40), and YFP–PS1 (light green inverted triangles; n = 23) showed increased calcium
mobilization in the presence of thapsigargin. b The difference between mean baseline ratio and maximal ratio (D340/380 nm) was calculated and differences between groups were measured using oneway analysis of variance (ANOVA) and Dunn’s pairwise multiple comparison test. Statistical significance was set at p \ 0.05. D340/ 380 nm was noted to be higher in all transfected groups in comparison with control group (0.39 ± 0.016; p \ 0.001). There was no statistically significant difference between PS1–CFP (0.99 ± 0.061; light green bar) and YFP–PS1 (0.79 ± 0.025; dark green bar) groups. Data is presented as mean ± SEM. Asterisk significant difference from control. Dagger significant difference from PS1 (Color figure online)
Transient Transfection of YFP–PS1 and PS1–CFP Increases Thapsigargin-Induced Cytosolic Ca2? Mobilization Similar to PS1 (WT) in SK-N-SH Cells
efficiencies of YFP–PS1, PS1–CFP, and PS1 (WT) expression constructs. These calcium imaging experiments confirm that attachment of YFP or CFP to PS1FL does not alter the ERCa2? leak channel activity function of PS1 (WT) protein.
We and others have reported that presenilins (PS1/PS2) act as passive ERCa2? leak channels to promote passive leak of Ca2? from ER lumen to cytoplasm [6, 8]. To confirm that the ERCa2? leak channel activity of PS1 is not altered due to the attachment of a NH2-terminal YFP-tag (YFP– PS1) and a COOH-terminal CFP-tag (PS1–CFP), we performed Ca2? imaging experiments with SK-N-SH cell line transiently transfected with pCDNA3 vector or pCDNA3.PS1 or pCDNA3.YFP–PS1 or pCDNA3.PS1– CFP expression construct by lipofectamine 2000. Ca2? imaging was performed 48 h post-transfection. Cells were loaded with fura-2-AM Ca2? imaging dye and intracellular calcium mobilization was measured by measuring normalized 340/380 nm ratio before and after the application of thapsigargin. Over-expression of YFP–PS1 or PS1–CFP increased thapsigargin mediated Ca2? mobilization peak very similar to PS1 (WT) transfected cells relative to vector-transfected cells (Fig. 3a, b). In YFP–PS1 and PS1– CFP groups the final intracellular calcium was less than PS1 (WT), but it was higher than the control group. These differences could be due to variation in transfection
123
FRET Analysis to Show Protein–Protein Interaction Between PS1–CFP and YFP–PS1 Molecules Expression of FAD mutants of PS1 alone or co-expression of both wild type PS1 and FAD mutants of PS1 abrogate the ability of PS1 to conduct Ca2? ions supporting for Ca2? hypothesis of AD [6, 8]. By analogy with other known ion channels it has been proposed that the functional PS1 channels may be multimers of several PS1 subunits. We have used ET in live cells to determine if PS1 forms dimer or multimers on cell membranes (ER or plasma membranes). We have conjugated the human PS1 protein with an NH2-terminal YFP-tag and a COOH-terminal CFP-tag. PS1–CFP were expressed alone or together with YFP–PS1 into human neuroblastoma SK-N-SH cell line and interaction between YFP–PS1 and PS1–CFP was determined by measuring lifetime of donor (PS1–CFP) in absence or presence of acceptor (YFP–PS1) in a single live cell (as in Fig. 1). ET was monitored by the change of fluorescent
Cell Biochem Biophys
A
B
C
Donoor
E
D
Donor + Acceptor
D Donor
Donnor
F
Donor + Acceptor
Donor + Acceptor
Fig. 4 Fo¨rster resonance energy transfer was measured in SK-N-SH cells expressing NH2-terminal YFP–PS1 and COOH-terminal PS1– CFP fusion proteins. Fluorescence life-times of the donor CFP was measured in live cells with and without cells expressing YFPacceptor. Data was collected on ISS Alba confocal fluorescence correlation spectrometer with micrometer spatial resolution and nanosecond temporal resolution to measure the dynamics of Presenelin-1 protein subdomains at single molecule regime. A 470 nm frequency modulated polarized excitation was used to measure
lifetimes in frequency domain (FD)—fluorescence lifetime imaging microscopy (FLIM) mode. Panels a, b, c and d, e, f in figure are representative plots of phase delay and modulation ratio of donor fluorescence alone and donor plus acceptor, respectively. The quench in the fluorescence lifetime of donor in the presence of acceptor suggests energy transfer and interaction of the NH2-terminal and ˚ ) units, presumably due to formation COOH-terminal (*r = 63.9 A of a membrane channel (Color figure online)
lifetime of the donor. Lifetime was measured in Alba-ISS (Urbana, IL) confocal microscope in frequency domain. First, SK-N-SH cells transfected with PS1–CFP alone were imaged (FLIM) and lifetimes were obtained sD = *2.014 ± 0.11 ns (average of three independent experiments). Panels a, b, and c of Fig. 4 are the phase delay and modulation ratio plots obtained in three independent experiments. Next, the lifetimes of the donor (CFP) was calculated from SK-N-SH cells co-transfected with PS1–CFP and YFP–PS1YFP–PS1 (panels d, e, and f of Fig. 4). The average lifetime was calculated to be sDA = 1.649 ± 0.058 ns. Table 1 summarizes the lifetimes obtained and the FRET measurements and results. We must note that the lifetime varied across the image and we choose ET areas because in their vicinity the life time was approximately constant. The efficiency of ET E = 1 - sDA/sD was *20 %. The distance between donor and acceptor was ˚ . It has been demonstrated previously that the *63.9 A NH2-terminal of PS1FL is localized on the cytoplasmic side of the plasma membrane. Since in our study, one PS1 molecule is PS1–CFP and another PS1 molecule is YFP– PS1, our FRET data appear to suggest that both the NH2-
terminal and COOH-terminal of PS1 molecule are present on the cytoplasmic side of the trans-membrane as reported by Doan et al. [22] and PS1 forms either dimer or multimers on the cellular membrane. Our FRET data is in agreement with the eight TMD topology of PS1 [23, 24]. It is to be noted that there exist several other models including 6TMD, 7TMD, 8TMD, and 9TMD topology of PS1 [13, 14, 17, 18, 23, 24]. Diffusion Studies in the Absence and in the Presence of Glycerol Show that Both the NH2-Terminal and the COOH-Terminal of PS1 are Located on the Cytoplasmic Side of Plasma Membrane In order to determine on which side of the cell membrane the PS1 protein termini are located, we measured lateral diffusion of PS1–CFP in the plasma membrane of SK-NSH cells by FCS (Fig. 5). Alba-ISS confocal spectrophotometer with avalanche photodiodes capable of single molecule detection SMD) was used to measure the lateral motion of PS1 protein in the plasma membrane. The number of molecules in the elliptical confocal volume (ECV) was determined from the inverse of the
123
Cell Biochem Biophys Table 1 Lifetimes of PS1–CFP donor in SK-N-SH cells with and without acceptor YFP–PS1 were used to calculate the transfer efficiency (ET) and the distance (r) between the CFP and YFP proteins Experiment
s (CFP-D)
s (CFP ? YFP) (DA)
ET = 1 - (sDA/sD)
r = Ro [(1/ET) - 1]1/6 (nm)
1
2.145 (v2 .787) (a)
1.582 (v2 .197) (d)
2
0.263
5.837
2
2
1.934 (v .564) (b)
1.686 (v .216) (e)
0.129
6.75
3
1.964 (v2 .77) (c)
1.679 (v2 .95) (f)
0.146
6.596
s fluorescence life-time, Ro Forster distance at which energy transfer efficiency is 50 %, D represents donor, DA represents donor and acceptor
Fig. 5 Identification of the NH2-terminal and the COOHterminal of PS1 on the cytoplasmic side of plasma membrane. In order to determine on which side of the cell membrane the protein termini are located, we measured lateral diffusion of PS1–CFP and YFP–PS1 in the plasma membrane of SK-N-SH cells by fluorescence correlation spectroscopy (FCS). a Confocal image of SK-N-SH cells expressing PS1–CFP. Addition of 80 % glycerol to the cells that express b PS1–CFP and c YFP– PS1 did not change the diffusion of the protein (*0.15 lm2/s), suggesting that both the NH2and COOH-terminus are facing the cytoplasmic side of the plasma membrane. The number of molecules under observation is kept low (to avoid averaging) and was determined by the inverse of autocorrelation function at time 0 (Color figure online)
autocorrelation function at time 0 [21]. In our experiments the number of molecules under observation was limited to *10 (average of 15 experiments). Diffusion studies of PS1–CFP protein in SK-N-SH cells were performed in PBS and 80 % glycerol in PBS. Both sets of experiments could be well fitted with standard 2D diffusion equation. Time traces obtained from single point profile from the cells were used to compute autocorrelation function. From the autocorrelation function of the CFP and YFP proteins in PBS and 80 % glycerol, the diffusion coefficients were calculated through ISS Vista Vision Version 4.0.00209 custom software. The diffusion coefficients of both the proteins, when measured in PBS and immediately after addition of 80 % glycerol were similar. The average diffusion coefficient obtained for YFP–PS1 protein was 0.15 lm2/s out of 15 experiments in both the control and
123
80 % glycerol group (p = 0.8404, t = 0.2033 and there was no statistical significance between the groups determined by students t test). Similarly, the average diffusion coefficient for PS1–CFP protein was determined to be 0.15 lm2/s out of 15 experiments in both the control and 80 % glycerol group (p = 0.8894, t = 0.1404 and there was no statistical significance between the groups determined by students t test). This suggests that the viscous glycerol does not inhibit the motion of the NH2- or COOH-terminus of the PS1 protein in the plasma membrane, indicating that they are internalized and face the cytoplasm. This result contradicts the previously published biochemical data showing the cytoplasmic localization of the NH2-terminal only. Although, glycerol will not enter through the plasma membrane to alter the diffusion of YFP–PS1 and PS1–CFP
Cell Biochem Biophys
proteins expressed on the ER membrane, we presume that both the NH2- and the COOH-termini of PS1 protein may be oriented toward the cytosolic side of the ER membrane.
Discussion PS1 protein was shown to be expressed on the membranes of the ER and the Golgi apparatus [25]. To date several studies have been undertaken to determine the intracellular membrane topology of PS1 protein expressed on the ER membrane [13, 14, 17, 18, 23, 24]. All published models agree that the NH2-terminus and the large hydrophobic loop of PS1 are localized in the cytoplasmic side of the ER membrane. In contrast, diverse opinions exist with regard to the number of TMD and the location of the COO2terminus of PS1 protein [24]. Doan et al. [22] have transfected chimeric NH2-terminal PS1–APP fusion construct in cultured COS-1 cells to demonstrate that the NH2terminal epitope specific antibody interacts with PS1 on the cytosolic site of the ER membrane. This data suggests the cytoplasmic localization of NH2-terminal of PS1. Similar experiments by these authors also lead to conclude that PS1 may contain either six or eight TMD with NH2terminal, COOH-terminal, and long hydrophobic loop are oriented toward the cytoplasmic side of the ER membrane [22]. Strooper et al. [25] have demonstrated by immunocytochemical analysis that PS1 protein is mainly expressed in the ER and the Golgi apparatus. These authors have transfected COS1 and CHO cells myc-PS1FL or PS1FLmyc constructs followed by treatment of cells with low concentrations of digitonin to selectively permeabilize the plasma membrane and antibodies against NH2-terminal domain of PS1 or the myc-tag introduced at the NH2-terminus or the COOH-terminus of PS1 [25]. Their results showed that antibody against the NH2-terminal inserted myc-tag reacted with PS1 in digitonin-permeabilized cells suggesting that the NH2-terminal of PS1 was localized on the cytoplasmic side of the ER membrane. Similar experiments by these authors have shown that the long hydrophobic loop is also located on the cytoplasmic side of the ER membrane. In contrast, this model of PS1 by De Strooper et al. [25] predicts a 7-TMD topology of PS1 with the COOH-terminal of PS1 protein being localized on the lumen side of the ER membrane. A nine TMD topology for PS1 protein has been determined using more stable N-linked glycosylation scanning approach [13]. Glycosylation acceptor sequences were introduced into FLPS1, and the results show that NH2terminus and large hydrophobic loop are oriented are present in the cytosolic side of the ER membrane but the COOH-terminal of the PS1 is localized to the luminal side
of the ER membrane. Studies on glycosylation pattern after TMD deletions followed by transfection into cells derived from PS1-/- and PS2-/-mouse blastocytes, combined with computer-based TMD protein topology predictions and biotinylation assays of different PS1 mutants convinced the authors that PS1 has nine TMD and COO2terminus locates to the lumen/extracellular space [13]. A 9-TMD topology with NH2-terminal and the long hydrophobic loop being on the cytoplasmic side and the COOHterminal end on the lumen side of the ER membrane has been confirmed by Spasic et al. [14]. Although, the 9-TMD topology of PS1 with NH2-terminus in the cytoplasmic and COOH-terminal on the lumen side of the ER is now widely accepted, it is to be noted that several other models including 6TMD, 7TMD, 8TMD, and 9TMD topology of PS1 have also been proposed [13, 14, 17, 18, 23, 24]. Some of these models also predict that both the NH2-terminus and COOH-terminus of PS1 are oriented to the cytosolic side of ER [22, 24, 26]. To date we know very little about the topology of PS1 on the plasma membrane. We have used biophysical approach to identify the membrane topology of PS1 on the plasma membrane. Our results indicate that both YFP–PS1 and PS1–CFP chimeric proteins are expressed on the plasma membrane and intracellular membranes (Figs. 1, 2). We also present the lateral diffusion studies of PS1–CFP in the presence and the absence of 80 % glycerol. Since the lateral diffusion of YFP–PS1 and PS1–CFP chimeric proteins was not altered after addition of glycerol (Fig. 5), these data suggest that the CFP-tagged COOH-terminal and YFP-tagged NH2-terminal of PS1 are oriented on the cytoplasmic side of the plasma membrane. We also observed that attachment of YFP or CFP to PS1FL also increased the ERCa2? channel activity very similar to PS1 (WT) protein (Fig. 3). Taken these results together, we conclude that passive ERCa2? channels are formed by either dimerization or multimerization of PS1 proteins with both the NH2-terminus and COOH-terminus of PS1 oriented toward the cytosolic side of the ER. Acknowledgments We wish to thank Dr. Michael Wolfe (Harvard Medical School) for providing us with pCDNA3.1-PS1 expression vector. This research was supported by NIH Grant R01HL090786 to Dr. Julian Borejdo and research support from Graduate School of Biomedical Sciences of UNTHSC to Dr. Hriday K. Das. Mr. Krishna Midde is supported by Predoctoral Fellowship 12PRE8730003 from American Heart Association.
References 1. De Strooper, B., Saftig, P., Craessaerts, K., Vanderstichele, H., Guhde, G., Annaert, W., et al. (1998). Deficiency of presenilin-1 inhibits the normal cleavage of amyloid precursor protein. Nature, 391(6665), 387–390.
123
Cell Biochem Biophys 2. Wolfe, M. S., Xia, W., Ostaszewski, B. L., Diehl, T. S., Kimberly, W. T., & Selkoe, D. J. (1999). Two transmembrane aspartates in presenilin-1 required for presenilin endoproteolysis and gamma-secretase activity. Nature, 398(6727), 513–517. 3. Selkoe, D. J. (1997). Alzheimer’s disease: Genotypes, phenotypes, and treatments. Science, 275(5300), 630–631. 4. Selkoe, D. J., & Schenk, D. (2003). Alzheimer’s disease: Molecular understanding predicts amyloid-based therapeutics. Annual Review of Pharmacology and Toxicology, 43, 545–584. 5. Sherrington, R., Rogaev, E. I., Liang, Y., Rogaeva, E. A., Levesque, G., Ikeda, M., et al. (1995). Cloning of a gene bearing missense mutations in early-onset familial Alzheimer’s disease. Nature, 375(6534), 754–760. 6. Tu, H., Nelson, O., Bezprozvanny, A., Wang, Z., Lee, S. F., Hao, Y. H., et al. (2006). Presenilins form ER Ca2? leak channels, a function disrupted by familial Alzheimer’s disease-linked mutations. Cell, 126(5), 981–993. 7. Supnet, C., & Bezprozvanny, I. (2010). The dysregulation of intracellular calcium in Alzheimer disease. Cell Calcium, 47(2), 183–189. 8. Das, H. K., Tchedre, K., & Mueller, B. (2012). Repression of transcription of presenilin-1 inhibits gamma-secretase independent ER Ca(2)(?) leak that is impaired by FAD mutations. Journal of Neurochemistry, 122(3), 487–500. 9. Duff, K., Eckman, C., Zehr, C., Yu, X., Prada, C. M., Perez-tur, J., et al. (1996). Increased amyloid-beta42(43) in brains of mice expressing mutant presenilin 1. Nature, 383(6602), 710–713. 10. Annaert, W. G., Levesque, L., Craessaerts, K., Dierinck, I., Snellings, G., Westaway, D., et al. (1999). Presenilin 1 controls gamma-secretase processing of amyloid precursor protein in pregolgi compartments of hippocampal neurons. Journal of Cell Biology, 147(2), 277–294. 11. De Strooper, B. (2003). Aph-1, Pen-2, and Nicastrin with Presenilin generate an active gamma-Secretase complex. Neuron, 38(1), 9–12. 12. Sobhanifar, S., Schneider, B., Lohr, F., Gottstein, D., Ikeya, T., Mlynarczyk, K., et al. (2010). Structural investigation of the C-terminal catalytic fragment of presenilin 1. Proceedings of the National Academy of Sciences USA, 107(21), 9644–9649. 13. Laudon, H., Hansson, E. M., Melen, K., Bergman, A., Farmery, M. R., Winblad, B., et al. (2005). A nine-transmembrane domain topology for presenilin 1. Journal of Biological Chemistry, 280(42), 35352–35360. 14. Spasic, D., Tolia, A., Dillen, K., Baert, V., De Strooper, B., Vrijens, S., et al. (2006). Presenilin-1 maintains a nine-transmembrane topology throughout the secretory pathway. Journal of Biological Chemistry, 281(36), 26569–26577.
123
15. Tolia, A., Chavez-Gutierrez, L., & De Strooper, B. (2006). Contribution of presenilin transmembrane domains 6 and 7 to a water-containing cavity in the gamma-secretase complex. Journal of Biological Chemistry, 281(37), 27633–27642. 16. Tolia, A., & De Strooper, B. (2009). Structure and function of gamma-secretase. Seminars in Cell & Developmental Biology, 20(2), 211–218. 17. Bammens, L., Chavez-Gutierrez, L., Tolia, A., Zwijsen, A., & De Strooper, B. (2011). Functional and topological analysis of Pen-2, the fourth subunit of the gamma-secretase complex. Journal of Biological Chemistry, 286(14), 12271–12282. 18. Nelson, O., Supnet, C., Tolia, A., Horre, K., De Strooper, B., & Bezprozvanny, I. (2011). Mutagenesis mapping of the presenilin 1 calcium leak conductance pore. Journal of Biological Chemistry, 286(25), 22339–22347. 19. Herl, L., Lleo, A., Thomas, A. V., Nyborg, A. C., Jansen, K., Golde, T. E., et al. (2006). Detection of presenilin-1 homodimer formation in intact cells using fluorescent lifetime imaging microscopy. Biochemical and Biophysical Research Communications, 340(2), 668–674. 20. Story, G. M., Peier, A. M., Reeve, A. J., Eid, S. R., Mosbacher, J., Hricik, T. R., et al. (2003). ANKTM1, a TRP-like channel expressed in nociceptive neurons, is activated by cold temperatures. Cell, 112(6), 819–829. 21. Midde, K., Luchowski, R., Das, H. K., Fedorick, J., Dumka, V., Gryczynski, I., et al. (2011). Evidence for pre- and post-power stroke of cross-bridges of contracting skeletal myofibrils. Biophysical Journal, 100(4), 1024–1033. 22. Doan, A., Thinakaran, G., Borchelt, D. R., Slunt, H. H., Ratovitsky, T., Podlisny, M., et al. (1996). Protein topology of presenilin 1. Neuron, 17(5), 1023–1030. 23. Li, X., & Greenwald, I. (1996). Membrane topology of the C. elegans SEL-12 presenilin. Neuron, 17(5), 1015–1021. 24. Kim, J., & Schekman, R. (2004). The ins and outs of presenilin 1 membrane topology. Proceedings of the National Academy of Sciences USA, 101(4), 905–906. 25. De Strooper, B., Beullens, M., Contreras, B., Levesque, L., Craessaerts, K., Cordell, B., et al. (1997). Phosphorylation, subcellular localization, and membrane orientation of the Alzheimer’s disease-associated presenilins. Journal of Biological Chemistry, 272(6), 3590–3598. 26. Li, L., Cheung, T., Chen, J., & Herrup, K. (2011). A comparative study of five mouse models of Alzheimer’s disease: Cell cycle events reveal new insights into neurons at risk for death. International Journal of Alzheimer’s Disease, 2011, 171464.
