Intracellular Trafficking of Histone Deacetylase 4 Regulates Long-Term Memory Formation.код для вставкиСкачать
THE ANATOMICAL RECORD 294:1025–1034 (2011) Intracellular Trafficking of Histone Deacetylase 4 Regulates Long-Term Memory Formation WEN-HAN WANG,1 LI-CHENG CHENG,1 FEI-YAN PAN,1,2,3 BIN XUE,1 DA-YONG WANG,4 ZHONG CHEN,5* AND CHAO-JUN LI1,2,3* 1 The Jiangsu Key Laboratory for Molecular and Medical Biotechnology, College of Life Sciences, Nanjing Normal University, Nanjing, China 2 Model Animal Research Center (MARC), Nanjing University, Nanjing, China 3 The School of Medicine, Nanjing University, Nanjing, China 4 The School of Medicine, South-East University, Nanjing, China 5 College of Life Sciences, Hainan Normal University, Haikou, China ABSTRACT Histone acetylation is important for gene transcription, which is controlled by the balance between two kinds of opposing enzymes: histone acetyltransferases and histone deacetylases (HDACs). HDACs repress gene transcription by decreasing histone acetylation levels. Our hypothesis was that shuttling of Class II HDACs, such as HDAC4, between the nucleus and cytoplasm is critical for its function. We constructed mutants of mammalian HDAC4 that had different cellular locations and checked their function during memory formation using Caenorhabditis elegans as a model. The deletion of hda4, a homolog of HDAC4, was able to enhance learning and long-term memory (LTM) in a thermotaxis model. Transgenic experiments showed that mammalian wild-type HDAC4 rescued the phenotype of hda4-deleted worms but impaired LTM formation in wild-type worms. The cytosol-localized HDAC4 mutant was not able to alter the phenotype of knock-out worms but led to enhanced LTM formation in wild-type worms similar to hda4-deletion mutants. Constitutive nuclear localization of HDAC4 rescued the phenotype of deletion worms similar to wild-type HDAC4 but had no effect on wild-type worms. These results support our hypothesis that HDAC4’s biological function is regulated by its intracelluC 2011 Wiley-Liss, Inc. lar distribution. Anat Rec, 294:1025–1034, 2011. V Key words: HDAC4; long-term memory; Caenorhabditis elegans Chromatin remodeling, especially through histone-tail acetylation, which alters the compact chromatin structure and changes the accessibility of DNA to regulatory proteins, is emerging as a fundamental mechanism for regulating gene expression (Kurdistani and Grunstein, 2003). Increased acetylation of individual histones at speciﬁc residues is often associated with transcriptional activation, whereas decreased histone acetylation is often associated with transcriptional silencing (Turner, 2002; Kurdistani and Grunstein, 2003). It has been shown that the balance between acetylation and deacetylation of eukaryotic chromatin is very important for many biological processes, including synaptic plasticity and learning behavior (Alarcón et al., 2004; Levenson et al., 2004; Guan et al., 2009). Thus, sodium butyrate, a C 2011 WILEY-LISS, INC. V nonselective histone deacetylase (HDAC) inhibitor, which has been used for cancer therapy, may provide a therapeutic avenue for memory impairment caused by neuron degeneration and other diseases (Stefanko et al., *Correspondence to: Zhong Chen, Ph.D., College of Life Sciences, Hainan Normal University, Haikou 571158, China. Fax: 0898-68083831. E-mail: firstname.lastname@example.org or Chao-Jun Li, Ph.D., Professor, The School of Medical, Nanjing University, Nanjing 210093, China. Fax: 86-25-83596289. E-mail: licj@nju. edu.cn Received 24 August 2010; Accepted 31 January 2011 DOI 10.1002/ar.21389 Published online 3 May 2011 in Wiley Online Library (wileyonlinelibrary.com). 1026 WANG ET AL. 2009). However, this broadly acting nonselective HDAC has shown a range of side effects, which may limit their utility in memory impairment therapeutics. Chromatin acetylation is controlled by histone acetyltransferases (HATs), which transfer acetyl groups to the lysine residues of histone, thereby disrupting the interaction between histone and DNA chromatin. Deacetylation is controlled by HDACs, which are able to catalyze the removal of acetyl groups from lysine residues (Varga-Weisz and Becker, 1998). It has been shown that HDACs can regulate memory formation (Oliveira et al., 2007). Memory can be classiﬁed into short-term memory (STM) and long-term memory (LTM). LTM requires new protein synthesis, which involves transcription and gene expression (Levenson and Sweatt, 2005). Thus, disruptions in the balance of acetylation and deacetylation in eukaryotic chromatin would result in the changes in the gene expression pattern that relates to learning and memory formation. To deacetylate histones and repress transcription, HDACs must be present in the nucleus. Thus, HDAC nuclear localization will promote an increase in HDAC activity (Chawla et al., 2003). HDACs have been classiﬁed into four different subfamilies based on their sequence homology and cofactor requirements. Of these, Class I and Class II HDACs have been widely examined. Class I HDACs are ubiquitously expressed, whereas Class II HDAC expression may be more restricted (de Ruijter et al., 2003). Class I HDACs are found almost exclusively in the nucleus, whereas a signiﬁcant characteristic of Class II HDACs is that they are able to shuttle in and out of the nucleus in response to certain cellular signals (Zhao et al., 2001). In some studies, the administration of an HDAC inhibitor has been shown to restore normal LTM formation, which indicates that nuclear HDAC activity can cause memory deﬁcits in mice. Nonetheless, the function of HDAC4 localization has not been clariﬁed. Based on the above research reports, we hypothesize that the function of Class II HDAC is related to its subcellular localization during LTM formation. The little worm’s system has only 302 neurons, and a complete diagram of its neuronal chemical and electrical connections is available. As such, Caenorhabditis elegans (C. elegans) is an excellent model for studying the mechanism of memory formation. In this study, thermotaxis of C. elegans provides an ideal index to analyze the molecular mechanisms that drive learning and memory (Gomez et al., 2001; Kodama et al., 2006; Kuhara and Mori, 2006; Ye et al., 2008). We used C. elegans to examine whether constitutive targeting of HDAC4 to various subcellular locations affects LTM formation. MATERIALS AND METHODS Plasmids Construction and Conﬁrmation Human HDAC4, HDAC4 (3S) (resides in cytoplasm), and HDAC4 (D118) (resides in nucleus) mutants were provided by Dr. Yang (Duke University, Durham, NC; Bolger and Yao, 2005) and were inserted into the pEGFP-N1 vector (GeneBank accession number #U55762). The Neuro-2A cells were maintained in minimum essential medium containing 10% fetal calf serum, 100 U/mL penicillin G, and 100 lg/mL streptomycin at 37 C in 5% CO2. The cells were transfected using Lipofectamine 2000 (Invitrogen, Carlsbad, CA) according to the manufacturer’s instructions. Immunoﬂuorescent Staining The cells were stained for immunoﬂuorescence essentially as described previously (Zhao et al., 2001). When samples were harvested, the cells were rinsed with phosphate buffer solution (PBS) and ﬁxed with 4% paraformaldehyde for 1 hr at room temperature. Then the samples were rinsed three times with PBS and were blocked with 5% bovine serum albumin (BSA) for 1 hr. After blocking, the samples were incubated with HDAC4 (L-19) antibody (dilution, 1:200; sc-5246, Santa Cruz Biotechnology, Santa Cruz, CA) overnight at 4 C followed by incubation with antigoat immunoglobin G (IgG)-ﬂuoresceine isothiocyanate antibody (dilution, 1:100; sc-2048, Santa Cruz Biotechnology, Santa Cruz, CA) for 1 hr at 37 C after thorough washing. After rinsing three times with PBS, the samples were exposed to 4,6-diamino-2-phenyl indole (5 lg/mL; #8226, Sigma–Aldrich) for 10 min at room temperature for nuclear staining. The ﬂuorescence images of the cells were photographed using a cooled charge-coupled device camera (Diagnostic Instruments, MI) and processed using SPOT software (Diagnostic Instruments, MI). Sodium Dodecyl Sulfate-Polyacrylamide Gel Electrophoresis and Western Blot Analysis The Neuro-2A was obtained from Cell Resources Research Centre of Shanghai Institute for Biological Sciences, Chinese Academy of Sciences. The cells were washed with ice-cold PBS. The cells were lysed with modiﬁed radio-immunoprecipitation assay buffer. Protein lysates were centrifuged at 12,000g for 15 min, and then the supernatant was collected for protein quantiﬁcation by Bradford protein assay (Bio-Rad). An aliquot of 10 lg of total protein was boiled and separated electrophoretically on an sodium dodecyl sulfate-polyacrylamide gel electrophoresis gel and transferred onto a Hybond-N nitrocellulose membrane (GE Healthcare) by electroblotting. After blocking in Tris-buffered saline buffer containing 5% BSA for 1 hr, the nitrocellulose membrane sheets were incubated overnight at 4 C with green ﬂuorescent protein (GFP) (1A5) antibody (dilution, sc101536; Santa Cruz, CA). After washing with Tris-buffered saline and Tween 20 for three times, the membrane sheets were incubated with horseradish peroxidase (HRP)-conjugated goat antirat IgG-HRP secondary antibody (dilution, 1:5,000, sc-2006, Santa Cruz, CA) at 37 C for 1 hr. The immunoreactive bands were visualized using GE Healthcare’s ECL detection kit according to the manufacturer’s instructions (GE Healthcare). C. elegans Strain Maintenance Worm breeding and handling were conducted as described by Bredy et al. (2007). Wild-type Bristol strain N2 and mutant strains were obtained from the Caenorhabditis Genetics Center at the University of Minnesota. Mutant strains used are as follows: RB2357 [hda1 (ok1595)]; VC983 [hda2 (ok1479)]; RB1618 [hda3 (ok1991)]; RB758 [hda4 (ok518)]; VC220 [cmk-1 (ok287)]; and VC1052 [unc-43 (gk452)]. Behavioral Analysis The method of head thrash frequency, body bend frequency, and basic movements were performed as 1027 HDAC4 REGULATES LTM FORMATION previously described (Tsalik and Hobert, 2003; Murakami et al., 2005). To assay the head thrashes, young adult nematodes were washed with double-distilled water, followed by washing with M9 medium. Every nematode was transferred into a microtiter well containing 60 lL of M9 medium on top of agar. After a 1-min recovery period, the head thrashes were counted for 1 min. A thrash was deﬁned as a change in the direction of bending at the midbody. Fifteen nematodes were examined per treatment. To assay the frequency of body bends, young adult nematodes were ﬁrst fed on OP50 food in covered plates and then placed on a fresh and foodless nematode growth medium (NGM) plate. The worms were scored for the number of body bends in an interval of 20 sec. A body bend was counted as a change in the direction of the part of the nematode corresponding to the posterior bulb of the pharynx along the Y-axis, assuming that the nematode was traveling along the X-axis. Fifteen nematodes were examined per treatment. To assay the three basic movements, young adult nematodes were ﬁrst fed on OP50 food in covered plates and then placed on a fresh and foodless NGM plate. Three basic body movements, including forward sinusoidal movement (forward turns), reversal movement (backward turns), and turns in which nematodes change direction (Omega/U turns), were measured over a 30-sec interval. In Omega turns, a young adult nematode’s head touches the tail and its shape looks similar to the shape of Greek letter Omega, whereas the angle of the body bend is typically >90 degree in U turns. Fifteen nematodes were examined per treatment. applied, and so, solution ﬂows freely and smoothly in both directions throughout the gonad until the gonad noticeably swells up. For DNA transformation, the best results seem to occur, when the gonad was ﬁlled seemingly to the point of damage. The ﬂow was stopped at this point and pulled the worm off the needle. The coverslip was returned to the dissecting scope, and a drop (20 lL) of recovery buffer was added on the worms. It was made sure that buffer surrounds the worms, so that they were released off of the pad. Thermotaxis Tracking Behavior Assay for Learning The learning assay was performed basically as previously described by Mohri et al. (2005). The synchronized young adult animals were incubated on NGM plates with fresh food at 20 C overnight. The young adult worms were then incubated with fresh food at 25 C for different incubation intervals (1, 2, 4, 8, 12, and 18 hr). The young adult animals (25 worms) were picked onto a 9-cm dish that contained agar. A frozen acetic acid containing vial was placed on the center of the plate, and then, the plate was incubated at 25 C for 30 min in the presence of a constant humidity of 60% to create a radial gradient of temperature ranging from 17 C to 25 C. The index of thermotaxis was determined as the percentage of worms left on the region of the plate at 25 C. All of the assays were replicated more than three times. The index of thermotaxis Microinjection Constructs (unc-64::HDAC4, unc-64::HDAC4(3s), and unc-64::HDAC4(D118)) were generated by amplifying the promoter of unc-64 sequences and HDAC4 or mutant genes and cloning them into the pBluescript SKII(þ) expression vector (Stratagene, La Jolla). The constructs were injected into N2 wild-type or into hda4 (ok518) mutant worms (Fire, 1986; Mello et al., 1991). For injection into these unmarked strains (N2 and hda4 (ok518)), the unc-122::dsRed or dop-1::gfp (gifts from Dr. Piali Sengupta of the University of California, CA) was used as coinjection markers. A needle-loading pipette was ﬁlled with 1 lL DNA injection mix by capillary action. The pipette tip was inserted into the back of the needle, and injection mix was expelled onto the needle’s internal ﬁlament. When the injection solution was drawn into the needle tip, the step was ﬁnished. This prepared needle was placed into the needle holder and mounted on the manipulator (Micromanipulator 5171, Eppendorf). The position of the needle tip was in the center of the microscope’s ﬁeld of view using the 5 objective. Once positioned, the needle was moved up using the Z-axis control, so that it was slightly out of focus. A drop of oil was placed on an injection pad and under a dissecting microscope (Olympus, MVX10) on the top of a small Petri plate cover. Several worms were moved from a bacteria-free region of an NGM plate with a naked pick and transferred to the oil drop. The worms were oriented in rows with their ventral sides facing the same direction (opposite the needle direction). Using the ﬁne X-axis control, the tip was placed into center of the cytoplasmic core. Pressure was ¼ The number of wormsleft on the 25 C region The number of total worms Thermotaxis Tracking Behavior Assay for Memory The memory assay was performed basically as previously described by Mohri et al. (2005). The synchronized young adult animals were incubated on NGM plates with fresh food at 20 C overnight. The young adult worms were washed twice with M9 buffer and then incubated in fresh food at 25 C for additional 12 hr. These young adult worms were moved on to NGM plates without food and incubated at 25 C for different incubation intervals (6, 12, 18, 24, 36, and 48 hr). The animals (30 worms) were picked onto a 9-cm dish that contained agar with a radial gradient of temperature ranging from 17 C to 25 C. The index of thermotaxis was determined as described above. hda4 RNAi in C. elegans Ninety-nine base pair fragments of hda4 genes were ampliﬁed by polymerase chain reaction (PCR) from C. elegans cDNA and cloned into the L4440 vector that was a gift from Dr. Wang (South-East University, Nanjing, China) (A. Fire laboratory database, available at ftp://ftp.wormbase.org/pub/elegans vector). The following primers were used: hda4 (forward, GCTCTAGAACAGGCAATCCAACAAC; reverse, CTAGCTAGCTGTGACTTCATATCCGCCAA). RNAi feeding was performed essentially as described in Conte and Mello (2003). The L4440 feeding vector was transformed into HT115 bacteria, a gift from Dr. Wang 1028 WANG ET AL. Fig. 1. Effects of hda4 deletion on the learning and the memory formation for thermosensation. All four hda deletions were used to check their learning and memory ability. These deletions had no difference from wild-type worms on their head thrash frequency (A), body bend frequency (B), and forward, backward, and Omega turn (C), which mean that any learning and memory changes of worms are not because of their activity differences. (D) Learning ability evaluation showed that only hda4 deletion was able to signiﬁcantly increase the learning ability (P < 0.05 at 1 and 18 hr). (E) Long-term (from 18 to 48 hr) memory evaluation showed that hda4 deletion was able to signiﬁcantly enhance LTM (P < 0.05). Data are expressed as mean SD. *P < 0.05; **P < 0.01. (South-East University, Nanjing, China) and then cultured on luria bertani (LB) plates with 50 lg/mL ampicillin and 15 lg/mL tetracycline (Santa Cruz Biotechnology, Santa Cruz, CA). One single colony was inoculated into 20 mL bacterial cultures in LB containing ampicillin overnight; these cultures were used for seeding standard solid NGM including adenosine monophosphate and isopropyl b-D-1-thiogalactopyranoside (Santa Cruz Biotechnology, Santa Cruz, CA) plates. The wild-type or hda4 worms were put on individual plates and allowed to eat HT115 bacteria containing the hda4 RNAi plasmid. RNAi efﬁciency was detected with semiquantitative RT-PCR. Primers speciﬁc for hda4 (C10E2.3) sequences are as follows: forward, ACAGGAGTAAAGGGGAG; reverse, AAGTGTGGCGAGGAGAC. Pri- mers speciﬁc for act1 (T04C12.6) sequences were used as internal standards and are as follows: forward, CCAAAGGCTAACCGTGAAAA; reverse, GGAAGC GTAGA GGGAGAGGA. Statistical Analysis Results were calculated from three or more independent experiments. Data are presented as Mean SD. An overall analysis of variance between groups was used to compare the control group and the treated or mutant group, followed by post hoc Tamhane’s or least signiﬁcant difference multiple comparisons test, as appropriate (*P < 0.05; **P < 0.01). HDAC4 REGULATES LTM FORMATION 1029 Fig. 2. Effects of cmk1 or unc43 mutation and inhibition of hda4 mRNA expression on the long-term memory formation for thermosensation. (A) The IT phenotypes of cmk1 or unc43 mutant worms in 36 or 48 hr. (B) RT-PCR of 99 bp hda4 fragment in wild-type worms silenced by hda4 RNAi. RNAi worm is signiﬁcantly lower than wildtype worm. (C) The IT phenotypes of unc43 mutant worms on a radial thermal gradient ranging from 17 to 252 after feeding hda4 RNAi plasmid. Behavioral analysis of ﬁve mutant and N2 worms, including head thrash frequency (D), body bend frequency (E), and forward, backward, and Omega turn (F), suggested that there were no differences among these worms. Data are expressed as mean SD. *P < 0.05; **P < 0.01. RESULTS The Genetic Deletion of hda4 Enhances Learning and LTM Formation in Worms assess the formation of learning and memory for thermosensation (Gomez et al., 2001). The basic principle is that on a radical temperature gradient, worms will avoid migrating toward their starvation temperature. We found that the ability of hda1, hda2, and hda3 mutants to learn was not signiﬁcantly different from wild type (P > 0.05), whereas hda4-deletion mutants exhibited significantly increased learning potential at 1 and 18 hr (P < 0.05) (Fig. 1D). For the memory detection, we found that hda deletions had no signiﬁcant effect on STM formation for 6 and 12 hr. In contrast, hda4 deletion could enhance LTM formation at 18, 24, 36, and 48 hr, although other hdas did not exhibit this effect (Fig. 1E). From this result, we concluded that the hda4 deletion could promote both learning ability and LTM formation in worms. The HDAs of C. elegans are divided into two classes: I and II. HDA1, HDA2, and HDA3 belong to HDA Class I, whereas the HDA4 is included in HDA Class II. To ensure that the effect of hda deletion on altered memory function was not due to defects in locomotion, we ﬁrst examined head thrash, body bends, and movement in four hda mutants (hda1, hda2, hda3, and hda4) and N2 worms. The results showed that there were no signiﬁcant changes (P > 0.05) between mutant and wild-type worms (Fig. 1A–C). The learning or memory detection model was described by Mohri et al. and was used to 1030 WANG ET AL. Fig. 3. Mammalian HDAC4 is able to rescue the phenotype of hda4 deletion in worm. (A) The green ﬂuorescent dots in C. elegans’ nerve ring and dorsal cord nervous system indicate the F1 worms that expressed exogenous GFP-HDAC4 in C. elegans’ neurons. (B) Western blotting showed that exogenous GFP-HDAC4 was really expression in Neuro-2A. Behavioral analysis of three mutant and N2 worms, including head thrash frequency (C), body bend frequency (D) and forward, back- ward, and Omega turn (E), suggested that there were no differences between HDAC4 and control worms. (F) When mammalian HDAC4 was expressed in wild-type worms, the long-term memory formation was largely impaired (P < 0.05 in 36 hr and P < 0.01 at 48 hr). (G) When mammalian HDAC4 was expressed in hda4 deleted worms, long-term memory formation was recovered to the normal level of wild-type worms. Data are expressed as mean SD. *P < 0.05; **P < 0.01. HDA4 Mediates CaMKII-Related Memory Function (3s) and HDAC4 (D118) plasmids expressed under the control of the unc64 promoter that drives unc-64 expression in C. elegans neurons of the nerve ring and dorsal cord nervous system. We microinjected these constructs into the ovary of N2 and hda4 mutant worms to obtain F1 transgenic C. elegans (Fig. 3A). As shown in Fig. 3B, the expression of exogenous HDAC4 was detected by western blot analysis with GFP antibody. F1 transgenic C. elegans were used to detect LTM formation using the thermotaxis tracking model. The effects of exogenous HDAC4 expression on locomotion behavior of N2 or hda4 mutant worms were measured, and the results indicated HDAC4 expression had no effect on the locomotion behavior of worms (Fig. 3C–E). As shown in Fig. 3F, exogenous HDAC4 expression in wild-type worms signiﬁcantly impaired LTM formation as compared with control worms. When HDAC4 was expressed in hda4-deletion worms, the amelioration of LTM formation was recovered to the normal level (Fig. 3G). Therefore, expression of exogenous mammalian HDAC4 can effectively abrogate the amelioration of LTM formation induced by hda4 deletion. The nuclear-cytosol shufﬂing of HDAC4 is controlled by Ca2þ/calmodulin-dependent protein kinase (CaMK) promoted phosphorylation at its C-terminal (Backs et al., 2006). The cmk-1 and unc-43 encode worm Ca2þ/calmodulin-dependent protein kinase I (CaMKI) and II (CaMKII), respectively. CaMKII deletion (unc-43) signiﬁcantly impairs LTM formation (Fig. 2A), possibly by impacting HDA4 nuclear localization. When HDA4 was knocked down by siRNA, the impaired LTM formation resulting from CaMKII deletion was largely recovered (Fig. 2B,C). The results of behavior analysis indicate that cmk-1 and unc-43 deletion have no effect on locomotion similar to hda4 RNAi (Fig. 2D–F). These results suggest that HDA4 regulates CaMKII-dependent LTM formation. Mammalian HDAC4 is Able to Rescue the Phenotype of hda4 Deletion in Worms To study the function of mammalian HDAC, we constructed GFP-tagged HDAC4 and two mutant (HDAC4 1031 HDAC4 REGULATES LTM FORMATION Fig. 4. Nuclear-localized HDAC4 (3S) is able to rescue the phenotype of hda4 deletion in worms, similar to wild-type HDAC4. (A) The expression of HDAC4 (3S)-GFP (left) or HDAC4-GFP (right) fusion proteins in Neuro-2A. HDAC4 (3S) mutant protein was located in the nucleus of neurons, whereas most HDAC4 protein was located in the cytosol of neurons. Behavioral analysis of three mutant and N2 worms, including head thrash frequency (B), body bend frequency (C), and forward, backward, and Omega turns (D), suggested that there were no differences between HDAC4 (3S) and control worms. (E) When HDAC4 (3S) was expressed in wild-type worms, long-term memory formation was not signiﬁcantly impaired, although there was slight decrease at 36 and 48 hr. (F) When mammalian HDAC4 (3S) was expressed in hda4 deleted worms, long-term memory formation was recovered to the normal level of wild-type worms. Data are expressed as mean SD. *P < 0.05; **P < 0.01. Nuclear Localization of HDAC4(3S) is Able to Rescue the Phenotype of hda4-Deleted Worms The Cytoplasmic HDAC4 is Favor of LTM Formation The three serines (246/467/632) in the HDAC4 are important for HDAC4 nuclear export (Zhao et al., 2001). When the three serines were mutated to alanine (HDAC4 (3S)), abolishing serine-speciﬁc phosphorylation sites, HDAC4 nuclear export was inhibited by preventing 14-3-3 binding (Zhao et al., 2001). Thus, the serinemutated HDAC4 is only found in the nucleus, as shown in Neuro-2A cells (Fig. 4A). The results of behavior analysis showed that HDAC4 (3S) expression in N2 or hda4 mutant worms had no effect on locomotion behavior, similar to HDAC4 expression (Fig. 4B–D). Figure 4E shows that exogenous expression of HDAC4 (3S) in wild-type worms did not cause any defect in LTM formation (P > 0.05). Exogenous expression of nuclear HDAC4 (3S) in hda4-deletion worms was able to signiﬁcantly rescue the hda4-deletion phenotype at 36 hr (P < 0.01) and 48 hr (P < 0.05), similar to HDAC4 (Fig. 4F). These results indicate that the nuclear localization of HDAC4 is important for LTM formation. HDAC4 (D118) is mutated such that the resulting HDAC4 protein can only locate in cytoplasm, as shown in Neuro-2A cells (Fig. 5A). The results of behavior analysis showed that HDAC4 (D118) expression in N2 or hda4 mutant worms had no effect on locomotion behavior, similar to wild-type HDAC4 expression (Fig. 5B–D). Exogenous expression of HDAC4 (D118) in wild-type worms signiﬁcantly accelerated LTM formation at 48 hr (P < 0.05), similar to hda4 deletion (Fig. 5E). When expressed in hda4-deletion worms, HDAC4 (D118) had no effect on the hda4-deletion phenotype (Fig. 5F). These results further suggest that nuclear-cytosol shufﬂing is a key factor affecting LTM formation. DISCUSSION Learning and memory have been successfully investigated in some animal models including Aplysia, Drosophila, and mice using various learning and memory 1032 WANG ET AL. Fig. 5. Cytosol-localized HDAC4 (D118) is able to ameliorate longterm memory formation, similar to the hda4-deletion mutant. (A) The expression of HDAC4 (D118)-GFP or HDAC4-GFP (right) fusion proteins in Neuro-2A. The HDAC4 (D118) mutant protein was located in the cytosol of neurons, similar to the subcellular localization of wildtype HDAC4 protein. Behavioral analysis of three mutant and N2 worms, including head thrash frequency (B), body bend frequency (C), and forward, backward, and Omega turns (D), suggested that there were no differences between HDAC4 (D118) and control worms. (E) When HDAC4 (D118) was expressed in wild-type worms, long-term memory formation was signiﬁcantly accelerated, as in hda4-deletion mutants. (F) When mammalian HDAC4 (D118) was expressed in hda4deleted worms, the phenotype was not recovered. Data are expressed as mean SD. *P < 0.05; **P < 0.01. paradigms (Mayford and Kandel, 1999; Mohri et al., 2005). C. elegans contains Class I and Class II HDACs that include HDA4 (Grozinger et al., 1999; Knoepﬂer and Eisenman, 1999). HDA4 protein is highly abundant in the neurons of C. elegans, suggesting that it has a function in the nerve system (van der Linden et al., 2007). A result of hda4 research shows that this gene shares a high homology with human Class II HDACs, and that the C. elegans HDA4 protein contains an myocyte enhancer factor-2 (MEF-2)-binding domain and a nuclear localization signal domain in the N-terminus, as well as a single catalytic domain in the C-terminus similar to mammalian HDAC4 (Choi et al., 2002). In addition to the high homology with human Class II HDACs, the expression pattern of HDA4 is similar to that of mammalian Class II HDACs. For instance, phosphorylation of S198 in HDA4, which is similar to S246, S467, and S632 in HDAC4, is catalyzed by KIN-29 or UNC-43 and can regulate gene expression through the MEF-2/ HDA4 pathway similar to the mammalian system (van der Linden et al., 2007). Thus, HDAC4 would likely behave in C. elegans the same as in the mammalian sys- tem. Using C. elegans as a genetic model, we suggest that HDA4 has an inhibitory effect in regulating memory formation. Moreover, the intracellular distribution of HDA4 is important for its function. Exogenously introduced mammalian HDAC4 is able to rescue the phenotype of a hda4 deletion in C. elegans. Further, nuclear HDAC4 could function to inhibit gene transcription and thereby impair memory formation. Recently, histone acetylation has been implicated in synaptic plasticity and learning behavior (Levenson et al., 2004; Vecsey et al., 2007). HDAC inhibitors are able to recover learning potential and promote the retrieval of LTM in mice even after massive neuronal loss (Fischle et al., 1999). Cohen et al. (2009) found that nuclear accumulation of HDAC4 and transcriptional reduction of MEF-2-regulated gene expression can lead to a loss of neural input. In Aplysia, Guan et al. (2002) reported that HDAC5 plays an important role in the blocking of long-term facilitation and synapse-speciﬁc long-term depression induced by 5-hydroxytryptamine and phe-met-arg-phe-amide. This LTM-related synaptic plasticity was blocked, because cAMP response element- HDAC4 REGULATES LTM FORMATION binding (CREB)2 and HDAC5 can displace CREB1CREB-binding protein leading to histone deacetylation (Guan et al., 2002). Although chromatin remodeling seems to be a general modiﬁcation to gene transcription, it is likely the equilibrium of histone acetylation and deacetylation regulated by HDACs and HATs that is important in the expression of memory-related genes (Alarcón et al., 2004; Vecsey et al., 2007). HDAC2 preferentially binds to the promoters of genes implicated in synaptic remodeling/plasticity or those regulated by neuronal activity (Guan et al., 2009). Thus, the protein levels of these genes are markedly increased in the brain of HDAC2KO mice and decreased in HDAC2OE mice, providing a mechanism by which HDAC2 is able to negatively regulate memory formation. Here, we propose additional layers of HDAC regulation. HDAC4, a Class II HDAC, has the ability to shuttle between the nucleus and cytoplasm. These cellular locations regulate its deacetylation function inside and outside the nucleus. We found that the nuclear restriction of HDAC4 is very important for its ability to regulate memory formation. Nuclear-restricted, but not cytoplasm-restricted, HDAC4 could rescue the enhancement in memory formation observed in hda4-deletion mutants. HDAC4 export from the nucleus is dependent on CaMK (Zhao et al., 2001). CaMK is able to directly phosphorylate HDAC4 at three serine residues (246/467/ 632), whereupon 14-3-3 binds to HDAC4 and facilitates its export out of the nucleus. CaMK is a critical modulator of neuronal development and plasticity (Wayman et al., 2008). Thus, CaMK-dependent HDAC4 shufﬂing between the nucleus and the cytoplasm provides more precise regulation of memory formation. ACKNOWLEDGEMENTS Human HDAC4, HDAC4 (3S) (resides in the cytoplasm), and HDAC4 (D118) (resides in the nucleus) were kindly provided by Yang Xiangjiao (McGill University Health Centre, Montreal, Canada). LITERATURE CITED Alarcón JM, Malleret G, Touzani K, Vronskaya S, Ishii S, Kandel ER, Barco A. 2004. Chromatin acetylation, memory, and LTP are impaired in CBPþ/ mice a model for the cognitive deﬁcit in Rubinstein-Taybi syndrome and its amelioration. Neuron 42:947– 959. Backs J, Song K, Bezprozvannaya S, Chang S, Olson EN. 2006. CaM kinase II selectively signals to histone deacetylase 4 during cardiomyocyte hypertrophy. J Clin Invest 116:1853–1864. Bolger TA, Yao TP. 2005. Intracellular trafﬁcking of histone deacetylase 4 regulates neuronal cell death. J Neurosci 25:9544–9553. Bredy TW, Wu H, Crego C, Zellhoefer J, Sun YE, Barad M. 2007. Histone modiﬁcations around individual BDNF gene promoters in prefrontal cortex are associated with extinction of conditioned fear. Learn Mem 14:268–276. Chawla S, Vanhoutte P, Arnold FJ, Huang CL, Bading H. 2003. Neuronal activity-dependent nucleocytoplasmic shuttling of HDAC4 and HDAC5. J Neurochem 85:151–159. Choi KY, Ji YJ, Jee C, Kim DH, Ahnn J. 2002. Characterization of CeHDA-7, a class II histone deacetylase interacting with MEF-2 in Caenorhabditis elegans. Biochem Biophys Res Commun 293:1295–1300. Cohen TJ, Barrientos T, Hartman ZC, Garvey SM, Cox GA, Yao TP. 2009. The deacetylase HDAC4 controls myocyte enhancing factor- 1033 2-dependent structural gene expression in response to neural activity. FASEB J 23:99–106. Conte DJ, Mello CC. 2003. RNA interference in Caenorhabditis elegans. Curr Protoc Mol Biol Chapter 26:Unit 26.23. de Ruijter AJM, van Gennip AH, Caron HN, Kemp S, van Kuilenburg AB. 2003. Histone deacetylases (HDACs): characterization of the classical HDAC family. Biochem J 370:737–749. Fire A. 1986. Integrative transformation of Caenorhabditis elegans. EMBO J 5:2673–2680. Fischle W, Emiliani S, Hendzel MJ, Nagase T, Nomura N, Voelter W, Verdin E. 1999. A new family of human histone deacetylases related to Saccharomyces cerevisiae HDA1p. J Biol Chem 274:11713–11720. Gomez M, De Castro E, Guarin E, Sasakura H, Kuhara A, Mori I, Bartfai T, Bargmann CI, Nef P. 2001. Ca2þ signaling via the neuronal calcium sensor-1 regulates associative learning and memory in C. elegans. Neuron 30:241–248. Grozinger CM, Hassig CA, Schreiber SL. 1999. Three proteins deﬁne a class of human histone deacetylases related to yeast Hda1p. Proc Natl Acad Sci USA 96:4868–4873. Guan JS, Haggarty SJ, Giacometti E, Dannenberg JH, Joseph N, Gao J, Nieland TJ, Zhou Y, Wang X, Mazitschek R, Bradner JE, DePinho RA, Jaenisch R, Tsai LH. 2009. HDAC2 negatively regulates memory formation and synaptic plasticity. Nature 459:55–63. Guan Z, Giustetto M, Lomvardas S, Kim JH, Miniaci MC, Schwartz JH, Thanos D, Kandel ER. 2002. Integration of long-term-memory-related synaptic plasticity involves bidirectional regulation of gene expression and chromatin structure. Cell 111:483–493. Knoepﬂer PS, Eisenman RN. 1999. Sin meets NuRD and other minireview tails of repression. Cell 99:447–450. Kodama E, Kuhara A, Mohri-Shiomi A, Kimura KD, Okumura M, Tomioka M, Iino Y, Mori I. 2006. Insulin-like signaling and the neural circuit for integrative behavior in C. elegans. Genes Dev 20:2955–2960. Kuhara A, Mori I. 2006. Molecular physiology of the neural circuit for calcineurin-dependent associative learning in Caenorhabditis elegans. J Neurosci 26:9355–9364. Kurdistani SK, Grunstein M. 2003. Histone acetylation and deacetylation in yeast. Nat Rev Mol Cell Biol 4:276–284. Levenson JM, O’Riordan KJ, Brown KD, Trinh MA, Molfese DL, Sweatt JD. 2004. Regulation of histone acetylation during memory formation in the hippocampus. J Biol Chem 279:40545–40559. Levenson JM, Sweatt JD. 2005. Epigenetic mechanisms in memory formation. Nat Rev Neurosci 6:108–118. Mayford M, Kandel ER. 1999. Genetic approaches to memory storage. CNS Spectr 15:463–470. Mello CC, Kramer JM, Stinchcomb D, Ambros V. 1991. Efﬁcient gene transfer in C. elegans: extrachromosomal maintenance and integration of transforming sequences. EMBO J 10:3959–3970. Mohri A, Kodama E, Kimura KD, Koike M, Mizuno T, Mori I. 2005. Genetic control of temperature preference in the nematode Caenorhabditis elegans. Genetics 169:1437–1450. Murakami H, Bessinger K, Hellmann J, Murakami S. 2005. Agingdependent and -independent modulation of associative learning behavior by insulin/insulin-like growth factor-1 signal in Caenorhabditis elegans. J Neurosci 25:10894–10904. Oliveira AM, Wood MA, McDonough CB, Abel T. 2007. Transgenic mice expressing an inhibitory truncated form of p300 exhibit long-term memory deﬁcits. Learn Mem 14:564–572. Stefanko DP, Barrett RM, Ly AR, Reolon GK, Wood MA. 2009. Modulation of long-term memory for object recognition via HDAC inhibition. Proc Natl Acad Sci USA 106:9447–9452. Tsalik EL, Hobert O. 2003. Functional mapping of neurons that control locomotory behavior in Caenorhabditis elegans. J Neurobiol 56:178–197. Turner BM. 2002. Cellular memory and the histone code. Cell 111:285–291. van der Linden AM, Nolan KM, Sengupta P. 2007. KIN-29 SIK regulates chemoreceptor gene expression via an MEF2 transcription factor and a class II HDAC. EMBO J 26:358–370. Varga-Weisz PD, Becker PB. 1998. Chromatin-remodeling factors: machines that regulate? Curr Opin Cell Biol 10:346–353. 1034 WANG ET AL. Vecsey CG, Hawk JD, Lattal KM, Stein JM, Fabian SA, Attner MA, Cabrera SM, McDonough CB, Brindle PK, Abel T, Wood MA. 2007. Histone deacetylase inhibitors enhance memory and synaptic plasticity via CREB: CBP-dependent transcriptional activation. J Neurosci 27:6128–6140. Wayman GA, Lee YS, Tokumitsu H, Silva AJ, Soderling TR. 2008. Calmodulin-kinases: modulators of neuronal development and plasticity. Neuron 59:914–931. Ye H, Ye B, Wang D. 2008. Trace administration of vitamin E can retrieve and prevent UV-irradiation- and metal exposure-induced memory deﬁcits in nematode Caenorhabditis elegans. Neurobiol Learn Mem 90:10–18. Zhao X, Ito A, Kane CD, Liao TS, Bolger TA, Lemrow SM, Means AR, Yao TP. 2001. The modular nature of histone deacetylase HDAC4 confers phosphorylation-dependent intracellular trafﬁcking. J Biol Chem 276:35042–35048.