| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
,2
* State Key Laboratory of Biomembrane and Membrane Biotechnology, Department of Biological Sciences and Biotechnology, Tsinghua University, Beijing, China;
Laboratory of Molecular Biology, NIDDK, National Institutes of Health, Bethesda, Maryland, USA;
Cold Spring Harbor Laboratory, New York, USA; and
Cancer and Developmental Biology Laboratory, National Cancer Institute, Frederick, Maryland, USA
2Correspondence: Z.X., State Key Laboratory of Biomembrane and Membrane Biotechnology, Department of Biological Sciences and Biotechnology, Tsinghua University, Beijing, China 100084; E-mail: zuopingx{at}mail.tsinghua.edu.cn or Y.Z., Cold Spring Harbor Laboratory, P.O. Box 100, Cold Spring Harbor, NY 11724, USA; E-mail: zhongyi{at}cshl.edu
 |
ABSTRACT |
|---|
 TOP ABSTRACT INTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
Key Words: L-type Ca2+ channel • heat shock • Alzheimer’s disease • synaptic function
|
INTRODUCTION |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
2
3
4
5)
. The structure and function of presenilin proteins appear to be highly conserved from C. elegans and Drosophila to mammals (6
, 7)
. A large number of identified missense mutations in the two human presenilin genes (PS1 and PS2) cause early-onset Alzheimer’s disease (AD) (8
, 9)
. One major effect of these PS mutations is the hyperproduction of Aß42 peptides generated from the cleavage of amyloid precursor protein (APP) by 
-secretase, a multiprotein protease in which presenilin is the crucial catalytic component (10
11
12
13)
. Increased Aß42 secretion leads to accelerated formation of extracellular amyloid plaques in specific brain regions, potentially providing a mechanistic explanation for the role of this class of presenilin mutations in early-onset AD. However, it is puzzling that so many different mutations that are distributed over the entire length of the presenilin protein all lead to an apparent gain-of-function phenotype, namely, hyperproduction of Aß42 peptides. In fact, in transgenic C. elegans, Drosophila and mouse models, several of these conserved AD-linked presenilin mutations instead reduce -secretase-like activity in the proteolysis of another membrane protein, the Notch receptor (14
15
16)
. To gain insight into how presenilin mutations are linked to the pathogenesis of AD, presenilin functions in addition to cleavage of APP and Notch have also been studied extensively.
Presenilin interacts directly with a number of proteins (17)
and affects a range of physiological functions, including intracellular Ca2+ homeostasis (18
19
20)
, capacitative Ca2+ entry (20
21
22)
, the endoplasmic reticulum (ER) stress response (23
24
25)
, and apoptosis (26
27
28)
. Some of these effects have been suggested to contribute to hyperproduction of Aß42 or otherwise promote neuronal toxicity that might be relevant to AD pathogenesis (22
, 29)
. However, reported data concerning the involvement of presenilin in responses to various stresses including oxidative stress, ER stress, and glutamate excitotoxicity have been inconsistent. While some studies have shown that AD-linked presenilin mutations or presenilin deficiency increase neuronal vulnerability to oxidative stress (30
31
32)
and glutamate toxicity (18
, 33
, 34)
, other work has shown that overexpression of wild-type presenilin or AD-linked mutant presenilins cause no change in response to oxidative stress (35
, 36)
or glutamate toxicity (37
, 38)
. The potential role of presenilin in the ER stress response is particularly controversial. Misfolded proteins that accumulate in the ER lumen stimulate expression of molecular chaperones, such as BiP, and transcription factors, such as CHOP (39
40
41)
. Some studies claim that this ER stress response is significantly diminished by AD-linked presenilin mutations (23
, 42)
, while others demonstrate that up-regulation of the ER stress-induced signaling pathway and chaperones are independent of presenilin expression (25)
and are not altered by knock-in presenilin-1 mutations (43)
. In addition to these discrepancies, the potential impact of alterations in the stress response caused by presenilin mutations has not been assessed at the synaptic level. Addressing this latter issue is important in understanding the causes of AD because subtle synaptic alterations precede overt neuronal degeneration in AD pathogenesis (44)
.
Since the above-mentioned studies about the role of presenilin in ER stress were conducted with either cell lines or primary neuronal cultures, we reasoned that it would be worth investigating the functional effects of presenilin mutations in response to stress in the context of an intact organism. We elected to perform an analysis of these mutants in Drosophila at two levels: 1) the synaptic level, where we examined voltage-gated K+ and Ca2+ currents at the neuromuscular junction (NMJ) in presenilin mutant larvae, and 2) the behavioral level, where we evaluated learning and memory in adult flies. The Drosophila NMJ has been widely used as a model system to study synaptic functions. It is a glutamatergic synapse with boutons, which in many ways resemble mammalian CNS synapses. The Drosophila presenilin homologue is highly conserved (53% amino acid identity), including strong conservation at amino acid residues associated with familial AD mutations (45
46
47)
. It is widely expressed in both neuronal and non-neuronal tissues (45)
. We subjected wild-type and presenilin mutant-expressing Drosophila larvae to heat shock (HS) or drug-induced ER stress and recorded voltage-gated currents from the NMJ, which consist of K+ and Ca2+ currents (48
, 49)
. Our analysis showed that voltage-gated K+ and Ca2+ currents were largely intact in larvae expressing loss-of-function and AD-linked presenilin mutations under both normal and stress conditions. However, Ca2+ tail currents were altered after stress conditions were halted. Similarly, long-term memory but not learning was reduced in the presenilin mutants only during the recovery period following cessation of stress treatments. These observations lead us to propose that presenilin plays a critical role in maintaining the normal function of Ca2+ channels after recovery from stress conditions.
|
MATERIALS AND METHODS |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
. Both homozygous PsnS3 and heterozygous combination of PsnS3 and PsnC4 (PsnS3/PsnC4) were used in our analyses. Since the two alleles were generated independently from separately mutagenized fly populations, the transheterozygous genotype (PsnS3/PsnC4) can be used to control for any possible background chromosome lesions. The following transgenic lines were also used: 1) w– (, or ;) P[w+, Psn-prom-PsnL235P]; PsnB3/TM6C, Tb
2) w– (, or ;) P[w+, Psn-prom-PsnE280A]; PsnB3/TM6C, Tb
3) w– (, or ;) P[w+, Psn-prom-Psnwt]; PsnB3/TM6C, Tb
To construct these transgenes, various full-length Psn cDNAs were subcloned as BamHI/XbaI fragments downstream of an 
1.5 kb Psn promoter XhoI/EcoRI fragment (Psn-prom) amplified from fly genomic DNA by PCR. This fragment contains 5' flanking and 5' UTR sequences and supports full rescue of Psn mutant flies when used to express wild-type Psn cDNAs (Y. Ye and M. E. Fortini, unpublished data). The two wild-type Psn (Psnwt) variants used as control transgenes, Psn-14a and Psn-14b, are splice variants encoding closely related isoforms that include or lack a 14 amino acid segment in the large cytoplasmic loop domain of the protein, respectively (45)
. PsnL235P and PsnE280A are transgenes encoding mutant presenilins with AD-associated mutations in residues conserved between Drosophila and human presenilin (16)
. Transgenes were inserted into the Drosophila transformation vector pCaSpeR for microinjection to generate the transgenic fly strains. For functional assays, all transgenes were introduced into a homozygous mutant PsnB3 background that is genetically null for endogenous Psn (50)
. Flies were grown on standard cornmeal medium at 23°C unless otherwise indicated (e.g., drug treatment regimes; see below). Homozygous loss-of-function mutant and transgenic larvae were selected according to body morphology using the dominant Tb marker and were reared under the appropriate conditions as specified for each particular experiment.
Drug treatments
Tunicamycin, thapsigargin, diltiazem hydrochloride, and amiloride hydrochloride were obtained from Sigma Chemical Company. All drugs were stored at –20°C in the dark, diluted to their respective desired concentrations immediately before each experiment, and used under minimal light conditions. Tunicamycin and thapsigargin were dissolved in ethanol. The maximum final ethanol concentration used in the experiments was 0.1%.
Recording from the larval NMJ
Larvae were affixed to a dissection dish with the dorsal side up and dissected in Ca2+-free saline. An incision was made along the dorsal midline, and the cuticle was pulled back and affixed to the dish. All internal organs were removed. Stimulation of the motor nerve in PPF experiments was achieved by 1-ms positive current stimulation of a loop of motor nerve in a suction electrode. The stimulation threshold was determined by varying stimulation intensity until the amplitude of junction currents did not increase. Stimulation intensity was set 20% above the threshold. Current recordings were obtained from ventral-lateral longitudinal muscle fiber 12 in abdominal segments 3 or 4 (48
, 51)
using the two-microelectrode voltage-clamp method as described (48
, 52)
. Both the voltage electrode and the current electrode were filled with 3 M KCl, with a resistance in the range of 5–10 M
. All electrophysiological recordings were performed at room temperature (1923°C) and the holding potential for all electrophysiological recordings was set at –80 mV.
I/V curves were generated by 250 ms voltage steps from a holding potential of –80 mV to potentials between –40 and +40 mV in 10 mV increments. To test the tail current kinetics, currents were elicited by 30 ms depolarization to –15 mV and then 160 ms repolarization to voltages between –110 and –70 mV in 20 mV increments. Pulses were given at 5-second intervals. All recordings were conducted at room temperature. K+ currents were recorded in Ca2+-free saline solution containing (in mM) 70 NaCl, 5 KCl, 10 NaHCO3, 115 sucrose, 5 trehalose, 5 HEPES, 10 MgCl2, at pH 7.1. For Ca2+ currents recordings, 10 mM Ba2+ was added to block K+ channels and serve as the charge carrier.
Data were recorded with a Gene Clamp 500B amplifier (Axon Instruments, Foster City, CA, USA), using Clampex software (Axon Instruments) and a Digidata 1320 A/D converter (Axon Instruments). Data were sampled at 50 KHz after Gauss filtering at 500 Hz. All data were leak-subtracted during analysis with the P/4 protocol in Clampex software except for PPF experiments. Cell capacitance was tested to evaluate the current densities. All current traces shown are representative data obtained from larvae as denoted. Current amplitude was measured at the peak of the current. Each tail current was fitted with a single exponential by Clampfit (Axon Instruments) to measure the time constant. Data were expressed as mean ± SEM.
Learning and memory tests
Olfactory associative learning was measured with a Pavlovian conditioning procedure (53)
. Briefly, a group of 100 flies was trained by exposing them sequentially to one odor (conditioned stimulus, CS+) paired with electric shock, then to a second odor (CS–) without electric shock. Learning was tested by exposing the flies simultaneously to the CS+ and CS– in a T-maze for 2 min. Flies trapped in CS+ and CS– T-maze arms were counted and a performance index (PI) was calculated as the number of flies avoiding the CS+ less the number avoiding the CS–, divided by the total number of flies and multiplied by 100. Hence, a 50:50 random distribution that yields a PI of zero indicates no learning. Twenty-four hour memory was evaluated after massed training, which consists of 10 cycles of one-session training used in the learning test, without any rest between successive training sessions. Flies were subsequently tested for memory as in the learning test in the T-maze after 1 day.
|
RESULTS |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
. Two wild-type variants of the presenilin gene, Psn-14a and Psn-14b, were also included in the study as control transgenes (see Materials and Methods). PsnS3 is an ethyl methanesulfonate (EMS) -induced hypomorphic mutant derived from w1118 (50
, 54)
. 2202U, an isogenic w1118 strain, was used as the wild-type control for PsnS3. Transgenic flies were constructed to express the AD-linked PsnL235P and PsnE280A mutant transgenes and the matching wild-type control transgenes Psn-14a and Psn-14b (see Materials and Methods). To avoid overexpression artifacts and examine transgene function in the absence of endogenous presenilin gene function, the transgenes were placed under the direct transcriptional control of the endogenous presenilin promoter in a presenilin genetic null background where the endogenous presenilin gene was deleted (see Materials and Methods). We confirmed that this approach enables us to express presenilin protein in the transgenic flies at levels similar to that of normal endogenous presenilin in the wild-type 2202U flies (Supplemental Fig. S1). The homozygous PsnS3 and transgenic Psn mutant flies were viable until the third instar larvae stage, but were lethal for PsnS3 and semi-lethal for PsnL235P and PsnE280A transgenic flies at later pupal stages due to diminished Notch cleavage activity reflecting complete (PsnS3) or nearly complete (PsnL235P and PsnE280A) elimination of presenilin activity (16
, 54)
. Since the PsnS3 mutant contains a second lethal mutation, we also examined the heterozygous combination of this PsnS3 allele with another EMS-induced strong hypomorphic mutant allele PsnC4 (50
and supplemental data). All larvae used in this study were assayed as homozygotes except for this PsnS3/PsnC4 genotype.
Voltage-gated currents were recorded from the late third instar larvae using the two-electrode voltage-clamp method (52
; see Materials and Methods). Different components of membrane currents can be isolated via physiological and pharmacological manipulations. Since Drosophila larval muscle fibers do not have Na+ channels, the outward and inward currents are carried by K+ and Ca2+, respectively (52)
. K+ currents are recorded in Ca2+-free saline whereas voltage-gated Ca2+ currents are recorded in the presence of a high concentration of Ba2+. Ba2+ serves as the charge carrier for Ca2+ currents and simultaneously blocks K+ currents (55)
. Since K+ currents were not affected significantly by presenilin mutations at all conditions tested (data not shown), only results on voltage-gated Ca2+ currents are shown.
We assessed the effects of HS stress on parameters of voltage-gated Ca2+ currents. It is known that high temperatures induce stress responses in flies, such as expression of HS proteins (56)
. Flies were reared under three conditions, including 1) room temperature (19–23°C; designated as RT), 2) HS treatment (reared at 29–30°C for 2–3 days prior to recording, designated as HS), and 3) post-HS treatment (reared at 29–30°C for 2 days, followed by RT for 1 day prior to recording; designated as post-HS). Voltage-gated Ca2+ currents were examined at various depolarizing membrane potentials using voltage clamp. In both the wild-type 2202U flies and flies expressing either wild-type or mutant presenilins, the peak value of the Ca2+ current was reached at membrane potentials of approximately –10 to –20 mV (Fig. 1
A; for mutants, only PsnL235P and PsnS3 data are shown). The current/voltage (I/V) relationship for the Ca2+ current was not altered by HS treatment in either wild-type or presenilin mutant larvae.
|
Voltage-gated Ca2+ channels deactivate as their voltage-dependent open gates close on repolarization of the membrane potential. However, the channels continue to carry current for a fraction of a millisecond after repolarization, giving rise to a tail current. The time course of the tail current reflects the kinetics of the channel closing process (deactivation). The tail current in wild-type and mutant presenilin larvae was examined by a 30 ms depolarization to –15 mV, followed by repolarization to different membrane potentials (Fig. 1B
). Its time constant (
) was determined by fitting the tail current with a single exponential function. The AD-linked presenilin mutations had no significant effects on Ca2+ tail currents at RT. In the loss-of-function mutant PsnS3, we observed a trend that the tail current was slower than in other genotypes, but the differences were not statistically significant (Fig. 1C
). The tail current was not changed by HS treatment in the wild-type 2202U strain larvae or in larvae expressing the wild-type or mutant presenilins, including PsnS3 and PsnE280A (Fig. 1B
). An exception was noted in larvae expressing the mutant PsnL235P, in which the tail current was accelerated under HS conditions when measured at the –70 and –110 mV repolarization potentials. To compare the change in tail current decay rate between RT and HS among different genotypes, the ratio of tail current time constant at RT vs. at HS was calculated. As shown in Fig. 1D
, no significant difference was observed between the mutant presenilin flies and controls. Overall, these data indicate that Ca2+ channel functions are largely intact in presenilin mutant animals under normal and HS conditions.
We then examined Ca2+ currents under the post-HS condition, in which larvae were heat-shocked for 2 days and allowed to recover at RT for 1 day. With this treatment, the I/V curve of the Ca2+ current was not changed by presenilin mutations. Compared with wild-type controls, there was no significant difference in the peak Ca2+ current or the depolarizing membrane potential at which the peak current was reached in larvae expressing either the loss-of-function or AD-linked mutant presenilins (Fig. 2
A). However, the Ca2+ tail current in these mutants was noticeably slower under the post-HS condition (Fig. 2B
and Supplemental Fig. S2A). The time constant of the Ca2+ tail current was increased at all three membrane potentials tested. When changes in the rate of tail current from RT to post-HS were compared, it was obvious that there was significant slowing of the tail current in the presenilin null mutants and the transgenic flies expressing mutant presenilins compared with corresponding controls (Fig. 2C, D
).
|
The HS experiments demonstrated that the Ca2+ current was largely unaffected by loss-of-function or AD-linked presenilin mutations under either normal or HS conditions. However, the Ca2+ tail current was altered after the HS was terminated. To further confirm that presenilin is involved in maintaining homeostasis of the Ca2+ tail current during poststress recovery, we treated larvae with another form of stress. We fed larvae with thapsigargin and tunicamycin, which are known inducers of the ER stress response. Tunicamycin is a nucleoside antibiotic that prevents protein glycosylation and thapsigargin is an inhibitor of the ER Ca2+-ATPase. Constant administration of tunicamycin by feeding had no discernable effects on Ca2+ tail currents in presenilin null mutant flies (Supplemental Fig. S2B) or in transgenic flies expressing the wild-type or AD-linked mutant presenilin PsnE280A (Fig. 3
A, C). In transgenic PsnL235P flies, there was an apparent weak but not statistically significant slowing in tail current kinetics at –90 and –70 mV after tunicamycin treatment. Consistent with the post-HS findings, 1 day after the larvae were transferred to food without tunicamycin significantly slower kinetics in the tail current were observed in the presenilin null mutant (Supplemental Fig. S2B) and both transgenic fly strains expressing mutant presenilins, even though the tail current in the wild-type control Psn-14a flies exhibited faster kinetics at the –70 mV repolarization potential (Fig. 3A, C
). Similarly, the Ca2+ tail current was not altered by feeding of thapsigargin to the wild-type Psn-14a and mutant PsnE280A transgenic lines, but was prolonged in the mutant PsnE280A 1 day after thapsigargin treatment ceased (Fig. 3B, D
). A prolonged tail current was also observed in the presenilin null mutant 1 day after cessation of thapsigargin treatment (Supplemental Fig. S2C). In the mutant PsnL235P, the tail current was slower upon administration of thapsigargin and the slower kinetics persisted after thapsigargin treatments ended (Fig. 3B
). Overall, these results further demonstrate that presenilin is required for maintaining normal Ca2+ tail currents after termination of the stress treatments, and extend our initial findings with heat shock to include other stress inducers.
|
The Ca2+ currents recorded include at least two components: a diltiazem-sensitive one that is similar to the vertebrate L-type Ca2+ channel (LTCC) current and an amiloride-sensitive one similar to the vertebrate T-type Ca2+ channel current (49)
. To determine which types of Ca2+ channel are affected by presenilin mutations under poststress conditions, we used diltiazem and amiloride to block the L-type and T-type Ca2+ channel currents, respectively (49)
. Under post-HS conditions, the tail current in the presence of amiloride was prolonged in mutant PsnS3 larvae (Fig. 4
). On the other hand, the tail current in the presence of diltiazem was not affected, suggesting that the L-type rather than the T-type Ca2+ channel tail current is specifically affected by presenilin mutations during the poststress recovery period (Fig. 4)
.
|
The Drosophila larval NMJ exhibits dynamic functional plasticity resembling that found in vertebrate central synapses (57)
. To investigate whether presenilin mutations affect synaptic plasticity, we examined paired pulse facilitation (PPF), a form of short-term plasticity at the NMJ. PPF is generally believed to reflect enhanced transmitter release during the twin pulse stimulation due to residual increases in presynaptic Ca2+ after the first pulse (58)
. Recently it was proposed that PPF may result from gradual saturation of an endogenous Ca2+ buffer (59)
. We examined PPF with multiple intervals (
100 ms) that separated the two stimulation pulses. The wild-type larvae showed an 2-fold facilitation at 20–30 ms interstimulus intervals, and this facilitation was not affected by either HS or post-HS treatment (Fig. 5
A). In PsnS3 mutant larvae, the PPF was comparable to the wild-type at RT (Fig. 5B
). The HS treatment resulted in a slight but not statistically significant enhancement of the facilitation. However, the post-HS treatment significantly reduced the PPF at 20–30 ms intervals. We then examined functional synaptic transmission. In a presenilin null mutant, the amplitude of the excitatory junctional current (EJC) was not affected by HS but was markedly increased in post-HS conditions (Supplemental Fig. S3). The amplitude and frequency of spontaneous miniature excitatory junctional current (mEJC) showed no significant difference between controls and the presenilin null mutant (data not shown), indicating an increased probability of quantal transmitter release.
|
To gain insight into the potential behavioral consequences of these electrophysiological perturbations, we examined the effects of HS treatments on learning and memory in the presenilin mutants. To overcome the lethality at the pupal stage in the presenilin mutants, we constructed presenilin null background flies carrying a wild-type Drosophila presenilin transgene and a second mutant presenilin transgene bearing an AD mutation, expressing both transgenes under the control of the endogenous presenilin promoter as direct transcriptional fusions. This constructed genotype closely mimics the genetics of AD patients carrying an AD-associated presenilin mutation, which is autosomal dominant and therefore involves one copy of a normal presenilin gene together with one copy of an AD mutant presenilin gene. Learning in these flies was not affected by either HS or post-HS treatment, nor was 24 h memory by HS alone (Fig. 6
A). However, under post-HS conditions, 24 h memory in these flies expressing the AD mutant presenilin was significantly reduced (Fig. 6B
). Neither learning nor memory was affected by either treatment in flies expressing a wild-type presenilin control transgene (Fig. 6A, B
). Taken together, our data indicate that presenilin AD mutations impair a specific aspect of poststress Ca2+ channel function that is implicated in long-term memory formation.
|
|
DISCUSSION |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
By performing our studies in an intact whole-animal model, the observed effects of presenilin mutations on the stress response could be correlated with their in vivo physiological consequences. At the synaptic level, the effects of presenilin were most obvious after stress treatments had terminated and animals were in the recovery phase. We found that both K+ and Ca2+ currents remained largely unaltered in loss-of-function or AD-linked presenilin mutants under both normal and stress conditions. This outcome is consistent with findings in mice where partial or complete loss of presenilin function does not alter basal neurotransmission at hippocampal area CA1 synapses (60
, 61)
. Under stress conditions, the Ca2+ tail current was in general not changed significantly, although some minor variability was detected among the different genotypes. For example, the Ca2+ tail current in flies expressing PsnL235P became faster under HS but was prolonged under ER stress. It is uncertain whether this might reflect a possible presenilin involvement in the stress response because of the variability among the different genotypes. However, a consistent finding observed in all the presenilin mutants examined is that the Ca2+ tail current was prolonged at 1 day after cessation of the stress treatment; further investigation revealed that this effect was specifically attributable to the LTCC tail current.
In vertebrates, prolonged LTCC tail currents have been observed during development (62)
or after strong depolarizations (63
64
65)
. The LTCC consists of multiple subunits, and each subunit has multiple isoforms as a result of differential gene expression, alternative splicing, and post-translational modification (66)
. Different channel compositions give rise to different gating properties of the LTCC and may influence the different rates of deactivation tail current (67
68
69
70
71)
. The class Cav1.2 
1 subunit has been shown to contribute to forming an LTCC with anomalous gating properties characterized by rare and short openings with very low open probability during depolarization but long openings with high open probability after repolarization (70
, 72)
. Isoforms of ß subunits suppress the prolonged tail current, probably by facilitated inactivation of the channel (73
, 74)
. Phosphorylation of the LTCC by the cAMP-dependent protein kinase A (PKA) or protein kinase C (PKC) can also slow down the deactivation tail current (75
76
77)
.
How presenilin might be involved in modification of the LTCC tail current after stress termination is not clear. Despite differences in some pharmacological properties compared with their vertebrate counterparts (49)
, the Drosophila larval muscle LTCCs are also modulated by PKA (78)
and PKC (79)
. Perturbations of the PKA or PKC pathways that enhance phosphorylation by pharmacological or genetic means increase the Ca2+ current (78
, 79)
. However, whether the tail current might be affected by these phosphorylation pathways has not been examined in these studies. Nevertheless, the effect of presenilin mutations on the LTCC tail current during recovery from stress is unlikely to be merely a result of channel phosphorylation. Instead, we observed a specific effect on the LTCC tail current only, and all other properties of the LTCC current remained unaffected. This effect is distinct from the other known effects of phosphorylation on both vertebrate and Drosophila LTCCs in which the current-voltage relationships and/or channel activation and inactivation are also affected (77
, 80)
. It was suggested recently that presenilin also forms ER Ca2+ leak channels, a function quite distinct from its recognized activity as an intramembranous aspartyl protease (81)
. In accordance with this model, altered intracellular Ca2+ levels due to the inability of mutant presenilins to conduct Ca2+ properly might affect the tail current via Ca2+-dependent inactivation of the LTCC. However, the proposed channel function for presenilin is associated with the full-length presenilin holoprotein, which is virtually undetectable at steady-state levels in neuronal and non-neuronal tissues in vivo (82)
(see also Supplemental Fig. S1). Thus, although an ER-localized Ca2+ leak channel function of presenilin cannot be ruled out as a potential explanation for our findings, we believe that this newly proposed function of presenilin is unlikely to account for the effects of loss-of-function and AD-associated presenilin mutants on Drosophila LTCC properties.
A more likely explanation for our findings is that the prolonged LTCC tail current is the final outcome of multiple events that occur in the poststress recovery period, possibly involving functions of presenilin in the regulation of gene transcription. This possibility is supported by the observation that the LTCC tail current in presenilin mutants was not affected immediately after cessation of stress (Y. Lu and Y. Zhong, unpublished data). Stresses, including heat, can induce modifications of the Ca2+ channels, such as changes in channel composition (83
84
85
86)
. Presenilin might be involved in reversing the changes that occur under stress conditions once these conditions are alleviated, perhaps by transcriptional regulation of specific isoforms of channel subunits, ancillary subunits, and modulatory components. Levels of protein kinases might also be affected, changing the dynamic equilibrium of Ca2+ channel phosphorylation. Such precedents for the involvement of presenilin in neuronal transcriptional control mechanisms exist. Indeed, presenilin is essential for release of the nuclearly translocating transcriptional regulatory factor NICD during Notch signal transduction (54
, 87)
, and its processing of APP and APLP is likewise involved in transcriptional activation through release of an analogous membrane-released fragment, termed AICD (88
89
90
91)
. Furthermore, loss of presenilin function has been shown to reduce CREB/CBP-dependent gene expression (61)
. It remains to be determined whether these or other presenilin-dependent transcriptional mechanisms might underlie the modulatory effects of presenilin on LTCC tail current during recovery from stress.
A reduction in PPF is generally regarded as the result of increased probability of transmitter release (92)
. The reduced PPF in presenilin mutants after stress termination could be related to the prolonged LTCC tail current seen in these animals. LTCCs have been shown to be involved in the regulation of neurotransmitter release (77
, 93
, 94)
as well as prepulse-induced facilitation (64
, 65)
. A putative neurosteroid, pregnenolone sulfate modulates presynaptic transmission when applied to hippocampal slices: acute EPSP is augmented whereas PPF is decreased (94)
. This effect has been partly attributed to elevated intracellular Ca2+ from activation of the LTCC. In hippocampal and cerebellar Purkinjie neurons and some sensory neurons, LTCCs may also contribute to normal functions of neurotransmitter release (95
96
97
98
99)
. A slower deactivation LTCC tail current in the presenilin mutants during recovery from stress-inducing treatments will result in elevated Ca2+ influx and increased release probability of synaptic vesicles, which is consistent with our observation of the increased EJC amplitude (Supplemental Fig. S3), and therefore a reduced PPF. An alternative explanation for the reduced PPF involves a change in endogenous Ca2+ buffer capacity. Alterations in the fast endogenous Ca2+ buffer calbindin-D28k affect the magnitude of PPF (59)
. As discussed above, transcription may be affected in presenilin mutants during stress recovery, which could affect the level of Ca2+ buffer proteins.
Our behavioral assay data raise the intriguing possibility that long-term plasticity (LTP) might also be affected after termination of the stress response. It has been shown that the LTCC is involved in NMDA receptor-independent LTP and spatial memory formation in mouse behavioral tests (100)
. As predicted by this established role for the LTCC in LTP and memory formation, we found that long-term memory in mutant presenilin flies was specifically affected during recovery from stress. It is interesting that A-kinase anchoring proteins, which are required for PKA phosphorylation-induced prolongation of LTCC tail current in mammals (75)
, are also involved in long-term memory formation in Drosophila (Y. Lv and Y. Zhong, unpublished data), hinting at a potentially conserved function for LTCC function in LTP in the fly and mammalian nervous systems.
Alzheimer’s disease-associated amyloid-ß (Aß) peptides selectively up-regulate the LTCC by direct binding to the channel and promoting insertion of the subunit into the plasma membrane (101)
. However, flies do not make toxic Aß peptides (102)
, so our observed impairment of LTCC function during poststress recovery is an interesting amyloid-independent effect of presenilin mutants on neuronal function, perhaps distinct from its amyloid-related effects on neuronal survival. The poststress Psn mutant effects could thus reflect more subtle perturbations in neuronal function than those caused by massive neurodegeneration. Impaired presenilin function thus might disrupt the synaptic activities of neuronal networks before any overt neurodegeneration occurs, and could also render neurons less able to recover from stress when exposed to toxins, during aging, and in the later neurodegenerative stages of Alzheimer’s disease. Furthermore, because the flies we used in the learning and memory studies carry a wild-type presenilin allele and an AD-linked presenilin allele, and thus closely recapitulate the autosomal dominant presenilin mutant genotype found in human AD patients, the specific defect uncovered in long-term memory during stress recovery might be clinically relevant to AD neuropathophysiology. In humans, even among people carrying the identical presenilin mutation, there is significant variability in the age of disease onset (103
104
105
106)
, indicating contributions from other genetic modifiers and/or environmental factors. Our observation of prolonged Ca2+ tail currents and impaired plasticity at synaptic and behavioral levels in AD-associated presenilin mutants during recovery from stress raises the possibility that neuronal stress response and recovery mechanisms might be intimately linked to the pathogenesis of Alzheimer’s disease and could serve as a potential target in considering therapeutic strategies for prevention and treatment of this dementia.
|
ACKNOWLEDGMENTS |
|---|
|
FOOTNOTES |
|---|
Received for publication August 12, 2006. Accepted for publication February 22, 2007.
|
REFERENCES |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
-Secretase: proteasome of the membrane?. Nat. Rev. Mol. Cell. Biol. 5,499-504[CrossRef][Medline]-Secretase, Notch, Aß and Alzheimer’s disease: where do the presenilins fit in?. Nat. Rev. Neurosci. 3,281-290[CrossRef][Medline]-secretase: a complex story of GxGD-type presenilin proteases. Trends Cell Biol. 12,556-562[CrossRef][Medline]-secretase through coordinated expression of presenilin, nicastrin, Aph-1, and Pen-2. J. Neurosci. Res. 77,309-322[CrossRef][Medline]-secretase complex: machinery for intramembrane proteolysis. Curr. Opin. Neurobiol. 14,379-383[CrossRef][Medline]-secretase complex. Neuron 38,9-12[CrossRef][Medline]-secretase. J. Neurosci. Res. 74,353-360[CrossRef][Medline]1 subunits. J. Physiol. (London) 554,609-6192/
-subunit on ionic and gating current in transiently transfected HEK 293 cells. Am. J. Physiol. 270,H1521-1528[Medline]1-subunit of smooth muscle Ca2+ channel preserves multiple open states induced by depolarization. J. Physiol. (London) 526,47-561C L-type calcium channel. EMBO J. 13,5032-5039[Medline]-Secretase cleavage and binding to Fe65 regulate the nuclear translocation of the intracellular C-terminal domain (ICD) of the APP family of proteins. Biochemistry 42,6664-6673[CrossRef][Medline]-secretase activity and a blistered-wing phenotype. Proc. Natl. Acad. Sci. U. S. A. 95,13703-13708This article has been cited by other articles:
|
|
 |
|
  S. D. Buckingham, A. K. Jones, L. A. Brown, and D. B. Sattelle Nicotinic Acetylcholine Receptor Signalling: Roles in Alzheimer's Disease and Amyloid Neuroprotection Pharmacol. Rev., March 1, 2009; 61(1): 39 - 61. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
), the same amount of the drug vehicle, ethanol, was added to the food. Tail currents were recorded immediately after drug treatments (
, drug) or 1 day after transferring treated larvae to fresh food (
, drug+rest). Representative current traces (n=5) and graphs plotting the tail current time constants vs. repolarizing potentials are shown. *Values significantly different from control treatment values (P<0.05, ANOVA). C, D) Comparison of tail current changes among different genotypes between drug-treated and control (top panel) or between drug + rest and control (bottom panel). The ratio of tail current time constant was calculated from data in panels A, B. *Significantly different from wild-type presenilin control value (ANOVA, P<0.05).
