Quantitative proteomics reveals olfactory input-dependent alterations in the mouse olfactory bulb proteome

J O U RN A L OF P ROT EO M IC S 1 0 9 ( 2 01 4 ) 1 2 5 – 14 2

Available online at www.sciencedirect.com

ScienceDirect www.elsevier.com/locate/jprot

Quantitative proteomics reveals olfactory input-dependent alterations in the mouse olfactory bulb proteome Hao-Long Zenga , Xiaoping Raob , Lei-Ke Zhanga , Xiaolu Zhaoa , Wei-Ping Zhanga , Jie Wangb , Fuqiang Xub,⁎, Lin Guoa,⁎⁎ a

State Key Laboratory of Virology, College of Life Sciences, Wuhan University, Wuhan, China State Key Laboratory of Magnetic Resonance and Atomic and Molecular Physics, Wuhan Institute of Physics and Mathematics, Chinese Academy of Sciences, Wuhan, China b

AR TIC LE I N FO

ABS TR ACT

Article history:

Olfactory sensory information is processed and integrated by circuits within the olfactory bulb

Received 1 March 2014

(OB) before being sent to the olfactory cortex. In the mammalian OB, neural activity driven by

Accepted 23 June 2014

external stimuli can lead to experience-dependent changes in structures and functions. In this

Available online 3 July 2014

study, quantitative proteomics techniques were employed to study proteome-wide changes in

Keywords:

transgenic model), wild-type control, and short-term and long-term odor exposures. Our results

Olfactory bulb

revealed that proteins related to various processes were altered in the OBs of odor-deprived and

Quantitative proteomics

odor-stimulated mice compared to the wild-type controls. These changes induced by odor

Odor deprivation

stimulation were quite different from those induced by a deficit in peripheral olfactory inputs.

the OB under four levels of neural activity (from low to high): devoid of peripheral input (using a

Odor exposure

Detailed analysis demonstrated that metabolic process and synaptic transmission were the most

Metabolic process

commonly altered pathways and that the effects of peripheral deprivation were more profound.

Synaptic transmission

Our comparative proteomics analysis indicated that olfactory deprivation and odor exposure lead to different alterations in the OB proteome, which provides new clues about the mechanisms underlying the olfactory deprivation- or odor stimulation-induced plasticity of OB function and organization. Biological significance By combining quantitative proteomics, bioinformatics and WB/IHC analysis, this study reports the results of the first comparative study on proteome-wide changes in the olfactory bulb under different levels of olfactory input. Odor deprivation and stimulation induced proteomic changes clearly demonstrate significant metabolic shifts and alterations on synaptic transmission. This quantitative system biology study leads to a new level of

Abbreviations: OB, olfactory bulb; GCLs, granule cell layers; EPLs, external plexiform layers; TH, tyrosine hydroxylase; OCNCX, olfactory cyclic nucleotide-gated channel subunit 1 knockout; LT, long-term odor stimulation; ST, short-term odor stimulation; GO, gene ontology; SVC, synaptic vesicle cycle; WB, Western blot; IHC, immunohistochemistry. ⁎ Correspondence to: F. Xu, Wuhan Institute of Physics and Mathematics, Chinese Academy of Sciences, Wuhan 430071, China. Tel.: +86 27 87197091; fax: + 86 27 87199543. ⁎⁎ Correspondence to: L. Guo, College of Life Sciences, Wuhan University, Wuhan 430072, China. Tel.: + 86 27 68753800; fax: + 86 27 68753797. E-mail addresses: [email protected] (F. Xu), [email protected] (L. Guo).

http://dx.doi.org/10.1016/j.jprot.2014.06.023 1874-3919/© 2014 Published by Elsevier B.V.

126

J O U RN A L OF P ROT EO M IC S 1 0 9 ( 2 01 4 ) 1 2 5 – 14 2

understanding in the development of olfactory bulb plasticity induced by odor deprivation or stimulation, and provides many new clues for the olfactory bulb related functional studies. © 2014 Published by Elsevier B.V.

1. Introduction In all living things, from bacteria to mammals, detecting chemicals in the environment is critical for survival. This is particularly true for higher eukaryotes, as nearly 4% of the genomes of these organisms are involved in olfactory function [1]. Olfactory system is linked to many important biological functions, including mating success, predator–prey balance, food preferences, orientation, social interactions, and mother care [2]. In the mammalian main olfactory system, olfactory sensory neurons (OSNs) in the epithelium detect odor molecules via the large family of G protein-coupled receptors (GPCRs) and send information to the olfactory bulb (OB) via the projection of axons that terminate at the dendrites of second-order interneurons and output neurons (mitral/tufted cells) in glomeruli [3,4]. Although the odorant receptors comprise the largest family of GPCRs, each OSN expresses only one GPCR, and the OSNs that express the same type of receptors converge into one or two glomeruli in the OB. These precise axon connections enable the about 1000 different types of receptors to be sorted into 1800 glomeruli in the mouse OB [5]. As the initial site involved in the processing of olfactory information in the brain, the OB exhibits a multi-layer cellular architecture. The glomerular modules interact with each other through neuronal circuits via local interneurons: periglomerular cells, short axon cells and granule cells. Periglomerular cells and short axon cells modulate intro- or interglomerular signal processing, contributing to lateral inhibitory/excitatory circuits that regulate functional specificity in small populations of glomeruli [5–7]. Granule cells, which do not have axons, extend a few short neurites toward deeper parts of the granule cell layers (GCLs) and one large dendrite toward the external plexiform layers (EPLs), where it branches extensively [8]. The interneurons are thought to modulate the activity of mitral and tufted cells, which are the principal neurons in mammalian olfactory bulb, to optimize discrimination of external olfactory stimuli by sharpening the odor representation by these projection neurons [8–10]. In addition to the complicated architecture of the OB, its neuronal circuitry is characterized by high adaptability, as it is one of the few structures in the mammalian nervous central system in which there is a continuous supply of newly generated neurons [11]. This neural activity is mainly driven by external stimuli, which can lead to experience-dependent morpho-functional changes in adult circuits, thus allowing ongoing integration of new and different smells [12]. Previous studies have suggested that sensory experiences such as olfactory enrichment or deprivation influence OB organization and function [13–15]. A particularly striking example of this phenomenon is provided by the rapid down-regulation of the catecholamine biosynthetic enzyme tyrosine hydroxylase (TH) in dopamine neurons intrinsic to the OB that has been observed in many odorant deprivation studies employing unilateral naris closure in neonatal rats [16,17]. In addition to the effects on OB neural transmission, energy demands in OB could also be

affected by external stimuli energy, as odor information processing elicits high energy demands on activated glomeruli [18,19]. Olfactory dysfunctions or anosmia are common symptoms associated with many neurological diseases, such as Alzheimer's disease, Parkinson's disease and major depression. The biochemical and molecular mechanisms linking the olfactory system, and especially the OB, to these neurological diseases are not clear [20,21]. Studies on OB function and organization are traditionally conducted following odor stimulation or olfactory deprivation by means of a hypothesis-driven approach focusing on biochemical changes in one or a few pathways [22]. However, because olfactory function/dysfunction involves the interactions of multiple cellular pathways (synaptic transmission, metabolic process, transcription regulation etc.), it is necessary to study the olfactory function/dysfunction from a global and integrative point of view. To identify proteins associated with olfactory function, in this study, we employed quantitative proteomic techniques to study proteome-wide alterations in the OB under four neural activity levels, from low to high: devoid of peripheral input (using a transgenic model), a wild-type control, and short-term and long-term odor exposures. Through stable isotope dimethyl labeling coupled with high-resolution nano-liquid chromatography–mass spectrometry (LC–MS), we were able to generate large-scale quantitative proteomics data from ex vivo OBs. Western blot (WB) and immunohistochemistry (IHC) analyses were performed on selected proteins, providing additional evidence validating the MS-based quantification results. Our data revealed that proteins related to various processes were altered in the OBs of odor-deprived and odor-stimulated mice compared with wild-type mice, and the changes induced by odor stimulation (both short and long terms) were quite different from those induced by deficits in peripheral olfactory inputs. Bioinformatics analysis demonstrated that the synaptic transmission and metabolic process are the most commonly altered pathways under all three conditions and that the effects of peripheral deprivation are more profound. Our functional study by magnetic resonance spectroscopy (MRS) using a 13C isotopic labeled glucose further validated our conjecture. Our quantitative proteomic analysis may provide new clues about the mechanism underlying the peripheral olfactory input-dependent plasticity of OB organization, leading to a new level of understanding of OB function.

2. Materials and methods 2.1. Animals Olfactory cyclic nucleotide-gated channel subunit 1 knockout (OCNCX) mice are a well-defined model of olfactory deprivation. The olfactory cyclic nucleotide-gated ion channel is necessary for the olfactory sensory nerve to generate odor-induced action potentials, rendering the OCNCX mice essentially anosmic [22].

J O U RN A L OF P ROT EO M IC S 1 0 9 ( 2 01 4 ) 1 2 5 – 14 2

The transgenic mice (male) and wild-type littermates (male) used in this study were generated by crossing heterozygous female OCNCX mice with male C57BL/6 background mice. The female heterozygous OCNCX mice were obtained from Dr. Minmin Luo's laboratory at the National Institute of Biological Sciences (NIBS, Beijing, China). The genotype of the mice was verified using a standard protocol through tail DNA PCR analysis. All animal experiments were carried out according to the protocols approved by the Wuhan Institute of Physics and Mathematics, Chinese Academy of Sciences (no. 00012092). The animals were maintained under a 12 h/12 h light–dark cycle at room temperature (22 ± 2 °C) and used for experiments at the age of 12–14 weeks. Food and water were provided ad libitum.

2.2. Odor deprivation and exposure For odor deprivation, OCNCX mice (n = 4) and wild-type mice (n = 4) were used to perform comparative quantitative proteomics experiment (OCNCX experiment). For odor exposure, two comparative quantitative proteomics experiments were performed: 1) between the long-term odor treated and no odor treated controls (LT experiment); 2) between the short-term odor treated and no odor treated controls (ST experiment). Male wild-type (WT) mice (n = 18) were separated equally into three groups (6 in each group). One group of mice was held separately for long-term (with isoamyl acetoacetate, 6 h), short-term (with isoamyl acetoacetate, 30 min) or no treatment. The mice were placed in a customized cage (size: 30 cm × 30 cm × 30 cm). For both long and short-term odor treatments, odor stimulation was carried out using the same regime as described previously [23]. Briefly, each repeating cycle of stimulation consists of a 50 s odor exposure followed by a 5 min break. Isoamyl acetoacetate odors were generated by directing 250 ml/min of humidified air through an isoamyl acetoacetate (liquid) containing chamber, and then mixed with a carrier stream of charcoal-filtered, humidified air (1 l/min) through a Teflon tube. Following odor exposure, mice were anesthetized with pentobarbital (1 mg/kg), and their OBs were quickly removed and immediately frozen in liquid nitrogen. The tissues were stored at −80 °C until experimental use.

2.3. Extraction and in-solution digestion of OB proteins The average OB tissue weight to total body weight was measured as 0.6 mg/g (OB tissue/total body). Tissues from each mice group were homogenized together (a pooled approach to minimize the biological variations) with a douncer in RIPA buffer containing various protease and phosphatase inhibitors (25 mM Tris–HCl pH 7.6, 150 mM NaCl, 1% NP-40, 1% sodium deoxycholate, 0.1% SDS, 1 mM NaF, 1 mM Na3VO4, with an added protease inhibitor cocktail mixture (COMPLETE, Roche Applied Science)). Buffer to tissue ratio was ~ 8 ml buffer per gram of tissue. The homogenates were held at 4 °C for 30 min and then centrifuged at 16,000 g for 30 min. The supernatants were collected and divided into 300 μl aliquots, with the estimated protein concentration from different mice groups in the range of 7–10 mg/ml. The in-solution digestion of OB proteins was performed as described in a previous study [24]. Proteins were precipitated by mixing with 50% acetone/50% methanol/0.1% acetic acid,

127

followed by centrifugation at 2000 g for 20 min. The protein pellets were re-suspended with 8 M urea/4 mM CaCl2/0.2 M Tris–HCl, pH 8.0, and the proteins were reduced with 10 mM dithiothreitol (DTT) at 50 °C for 30 min and alkylated with 40 mM iodoacetamide in the dark for 30 min. After measuring protein concentrations via Bradford assay, 500 μg of proteins were digested with trypsin at a ratio of 1:50 (trypsin/protein w/w). The digested peptides were desalted using a SepPak C18 cartridge (Waters) and dried with a SpeedVac.

2.4. Stable isotope dimethyl labeling and SCX fractionation Desalted peptides were re-suspended in 0.1 M sodium acetate, pH 6.0. Next, 4% formaldehyde (HCHO, which serves as “light labeled”) was added to the peptides from stimulated or knockout mice, and 4% deuterated formaldehyde (DCDO, which serves as “heavy labeled”) was added to the peptides extracted from control mice. After mixing, 0.6 M sodium cyanoborohydride (NaBH3CN) was added, and the mixtures were incubated at room temperature (20 ± 2 °C) for 1 h. The samples were then quenched by adding 1% ammonium hydroxide, followed by the addition of 5% formic acid. After labeling, the peptides were mixed at a ratio of 1:1 and desalted again prior to separation via strong cation exchange (SCX) chromatography. For SCX fractionation, Buffer A contained 5 mM KH2PO4, pH 2.7, in 20% acetonitrile/80% ddH2O. Buffer B contained 5 mM KH2PO4 and 0.5 M KCl, pH 2.7, in 20% acetonitrile/80% ddH2O. SCX was performed on a polysulfoethyl column (2.1 × 50 mm, 5 μm × 200 Å, The Nest Group, Inc.) using a KCl gradient from 0 to 0.5 M at a flow rate of 0.2 ml/min. During a 60 min gradient elution, nearly twelve fractions were collected and desalted with a C18 ZipTip (Millipore) prior to MS analysis.

2.5. LC–MS/MS and data processing All ESI-based LC–MS/MS experiments were performed on a QSTAR Elite LC–MS/MS system (Applied Biosystems) coupled to a nano-flow multi-dimensional HPLC system (Tempo™ nano MDLC system, Applied Biosystems) equipped with a micro-well plate autosampler. Dried peptides were re-suspended in 0.2% formic acid/2% ACN and loaded into a CapTrap (0.5 mm × 2 mm, MICHROM Bioresources, Inc.) at a flow rate of 5 μl/min, then eluted from the CapTrap over a C18 analytic column (100 μm × 150 mm, 3 μm particle size, 200 Å pore size, MICHROM Bioresources, Inc.) at a flow rate of 300 nl/min over a 130 min gradient. The mobile phase consisted of two components: component A was 2% acetonitrile with 0.1% formic acid, and component B was 98% acetonitrile with 0.1% formic acid. The gradient comprised of: 5 min in 5% B, 25 min of 5–15% B, 55 min of 15–40% B, 15 min of 40–80% B, the gradient was maintained in 80% B for 10 min, followed by 5 min of 80–5% B, and a final step in 5% B for 15 min. All of the parameters applied in the LC–MS method were set as described by Chen et al. [25]. For MS data analysis, tandem mass spectra were extracted using Analyst version 2.0. All MS/MS samples were processed with Mascot Distiller (version 2.3, Matrix Science, London, UK), and the generated files were analyzed using Mascot (version

128

J O U RN A L OF P ROT EO M IC S 1 0 9 ( 2 01 4 ) 1 2 5 – 14 2

2.4.0 Matrix Science, London, UK) against SwissProt_2013_01 (selected for Mus musculus (house mouse), 16,587 entries), assuming the digestion enzyme trypsin. The Mascot search parameters were set as follows: parent ion tolerance of 100 ppm and fragment ion mass tolerance of 0.40 Da. Carbamidomethyl cysteine was selected as a fixed modification and oxidation of methionine was chosen as a variable modification. Serine, threonine, and tyrosine phosphorylation were set as variable modifications. The peptide charge was set to 2+ or 3+. Up to two missed cleavages were allowed, and the significance threshold was set at p > 0.05. The false discovery rates (FDRs) of the peptide-spectra matches determined by a decoy database search were 1.22% for the odor deprived experiments, 1.23% for the long-term odor treatment experiments and 1.54% for short-term odor treatment experiments. Proteins were considered to be successfully identified when at least one correct assigned peptide was obtained (Mascot scores at or above the homology score). For single peptide-identified proteins, the spectrum of the peptide was also manually verified. Quantitation analysis was performed as described previously with minor modifications [24]. (1) For all of the LC–MS/MS analyses in each sample group, individual rov files that derived from the Mascot Distiller were processed. (2) All of the quantitation reports were combined (as Microsoft Excel files), and only peptides passing the defined stringent tests (threshold values for fraction, correlation, and standard error set to >0.5, >0.7, and <0.2, respectively) were used for protein-level quantitation. (3) Peptide ratios determined to be outliers were eliminated. The detection of peptide outliers was conducted via Dixon's Q test at 95% confidence as described previously [26]. (4) If several peptides for the same protein were present, a weighted average ratio (taking the ion intensity into consideration) was calculated for the protein. The standard deviation of the protein ratio was defined as the standard deviation of these peptide ratios.

2.6. Western blot analysis Equal amounts of proteins from each sample group were separated in 12% SDS-PAGE gels and transferred to polyvinylidene difluoride (PVDF) membranes (Millipore) via electroblotting. The membranes were blocked with Tris-buffered saline Tween-20 (TBST) buffer containing 5% BSA or 5% nonfat milk and then stained with a primary antibody. The rabbit polyclonal antibodies used in this study included anti-Cycs (10993-1-AP), anti-Ndufs3 (15066-1-AP), anti-Atp6d (18274-1-AP) and anti-Syt1 (14511-1-AP) (all from the Proteintech Group); anti-Gad1 (ab97739), anti-Calm1 (ab45689) and anti-Calb2 (ab101812) (all from Abcam); anti-TH (ab152, Millipore) and anti-Cplx1/2 (sc-33603) (all from Santa Cruz Biotechnology). A mouse antiActa2 polyclonal antibody (sc-1616, Santa Cruz Biotechnology) was also used. The membranes were then washed with TBST and incubated with horseradish peroxidase (HRP)-conjugated secondary antibodies at room temperature for 1 h. HRP was subsequently detected via enhanced chemiluminescence, and the obtained film was visualized, and band intensities were measured using the Quantity One software package (Bio-Rad Laboratories, version 4.6.2). The densitometry results of the WB were expressed as the mean ± SEM. Expression is reported

relative to β-actin in the same sample and normalized to control OBs. The data were analyzed using the nonparametric Kruskal–Wallis test, followed by the Mann–Whitney test. Differences were considered statistically significantly when the p values were less than 0.05.

2.7. Immunohistochemistry Mice were weighed, anesthetized through intraperitoneal injection of sodium pentobarbital (Nembutal 50 mg/kg body weight in 0.9% saline solution) and then transcardially perfused with 0.1 M PBS followed by 4% paraformaldehyde (PFA) in PBS. Their brains were post-fixed in PFA overnight and incubated in a 30% sucrose solution for 48 h. The tissues were embedded in optimal cutting temperature compound (OCT, Sakura Finetek, Torrance, CA) and frozen, after which they were sequentially sectioned at 20 μm intervals with a freezing microtome (LEICA CM1850, Germany). OB sections were stored at −20 °C until use. IHC was performed on free-floating tissue sections as follows: first, the tissue sections were incubated in 0.3% Triton-X in PBS for 1 h at room temperature and then rinsed in PBS or directly blocked in 10% normal goat serum in PBS for 1 h at room temperature; next, the sections were incubated overnight at 4 °C with a primary antibody against Cycs (10993-1-AP, rabbit, Proteintech, 1:100), Atp6d (18274-1-AP, rabbit, Proteintech, 1:100) or Calm1 (ab45689, rabbit, Abcam, 1:500), diluted in PBS. After rinsing in PBS, the sections were incubated with a fluorescein isothiocyanate (FITC) conjugated secondary antibody (KPL, 172-1506, 1:200) for 90 min at 37 °C, then rinsed again and counterstained with the nucleus stain DAPI (1 μg/ml) for at least 30 min. The sections were finally mounted on gelatin-coated slides and cover-slipped with a 1,4-Diazabicyclo[2.2.2]octane (DABCO) based anti-fade mounting medium.

2.8. Metabolic kinetics analysis The metabolic kinetics was measured with the infusion of [1-13C] glucose [27]. Seven OCNCX mice and eight WT mice were used. All animals were fasted overnight (15–19 h). On the experimental day, the mice were anesthetized with 1.5–2.5% isoflurane. Blood glucose concentration was measured with glucose meter using one drop of blood sample from one lateral tail vein. The other lateral tail vein was catheterized with PE10 tubes for [1-13C] glucose infusion. After catheterization, the animals were allowed to recover for 0.5 h. After the mice were totally recovered and free-moving, [1-13C] glucose was infused for 10 min, 30 min, 60 min and 120 min (~2 mice in every point). After the infusion, the mice were fast anesthetized with 3.5% isoflurane, and blood samples (~ 100 μl) were collected from orbit, and the animals were decapitated and fast fitted using a microwave method within 8 s. The OBs were removed and homogenized [28]. The supernatants were collected and lyophilized for MRS detection. The lyophilized samples were dissolved with 100 μl D2O (with 5 mM formate and phosphate buffer, pH = 7.2). The 13C enrichment of plasma glucose and metabolites were measured using 1 H–[13C] NMR at 600 M Bruker Avance vertical bore spectrometer. The detailed description of the method was described in our previous study [28]. The metabolic kinetics was calculated with CWave in Matlab [29].

J O U RN A L OF P ROT EO M IC S 1 0 9 ( 2 01 4 ) 1 2 5 – 14 2

2.9. Bioinformatics Statistical analysis of the MS results derived from the three experimental groups was performed using Microsoft Office Excel. Biological Networks Gene Ontology (BiNGO) 2.44 was used to calculate the gene ontology (GO) term enrichment of significantly up- or down-regulated proteins for each experimental group (defined as showing at least a 1.5-fold change between the treated and untreated samples) and determine significantly under- and over-represented functional GO categories. The Cytoscape network visualization platform (http://www.cytoscape.org/) implementing the latest release of the BiNGO plug-in was used to identify proteins that were annotated on the basis of biological process categories. The analyses were conducted using the default BiNGO M. musculus database and the GOSlim_Generic ontology. Statistical significance was determined by means of hypergeometric analysis, followed by Benjamini and Hochberg's false discovery rate correction (p < 0.001) [30]. For global analysis of metabolic process and synaptic transmission pathways, the KEGG PATHWAY database was accessed via the KEGG automatic annotation server [31]. Based on the quantified MS results, proteins matched in the KEGG PATHWAY database were extracted and submitted to STRING 9.0 (the Search Tool for the Retrieval of Interacting Genes/Proteins) to qualify the physical and functional interactions of these proteins. The proteins and their interactions were then uploaded to Cytoscape (version 2.8.3) for data visualization.

3. Results

129

subjected to trypsin digestion, and the resulting peptides were labeled with isotope dimethyl labels, then separated into ~12 fractions via SCX and analyzed via RPLC–MS/MS. Overall, after the data from all experimental replicates were combined, 1437 proteins were identified from the OCNCX experiment (3 replicates), 1121 proteins were identified from the LT experiment (4 replicates), and 965 proteins were identified from the ST experiment (4 replicates). In total, 1618 non redundant proteins were identified when the data from all experiments were combined (Table S1). Among the identified proteins from the three experiments, 1285, 946 and 859 proteins were successfully quantified in the OCNCX, LT and ST experiments, respectively (Table S2). Data reproducibility was analyzed by calculating the standard deviation (S.D.) of the quantified proteins across all three or four replicates from each comparative experiment. As shown in Fig. 1B, the standard deviations of more than 90% of the proteins quantified in at least two of the replicates were lower than 0.5, which indicates an acceptable reproducibility. To compare the proteome alterations in mice under odor deprivation and odor stimulation, we further performed a data correlation analysis among the three experiments and found that the Pearson correlation coefficient between the two odor-exposed experiments is 0.46, compared with 0.27 between the LT and OCNCX experiments and 0.31 between the ST and OCNCX experiments, which indicates that quantitative datasets of the two odor exposure experiments show a better correlation while each of them exhibits a poor correlation with the OCNCX dataset (Fig. 1C). The correlation analysis indicated that distinctive OB proteome landscapes may exist in mice under odor deprivation and odor stimulation.

3.2. Alternative evidence supporting the mass spectrometry findings

3.1. Experimental design and data overview To study the effects of external stimuli on the OB proteome, odor stimulation and deprivation were applied to achieve upand down-regulation of the peripheral olfactory inputs, respectively. For down-regulation, OCNCX mice, which are a common transgenic mouse model for anosmia, were used. Odor exposure was employed for up-regulation because it is known that ongoing neuronal activity driven by external stimuli can lead to morpho-functional changes in adult neurocircuits [12]. As different stimulatory paradigms that differ in timescales lead to distinct mechanisms within different regions of the olfactory system [32,33], two modes, involving short- and long-term odor stimulations (ST and LT), were selected. Therefore, a total of four different neuronal activity levels were generated: deficit in input (OCNCX), spontaneous input (as control), ST and LT. An outline for the quantitative proteomics experiments, based on duplex stable isotope dimethyl labeling, is shown in Fig. 1A. Comparative analyses were performed between OCNCX mice and wild-type controls, odor-stimulated mice (ST and LT) and non-odor-stimulated controls, respectively, forming a total of three quantitative proteomics experiments: OCNCX, LT and ST. The sample from each OB group was lysed and then divided into multiple identical aliquots, which allowed experimental replicates to be conducted for each group. Proteins were extracted and

To validate the MS quantification data, ten proteins were selected to be analyzed by WB (Fig. 2A). These proteins came from different functional groups and included 2 energy metabolic proteins (Cycs, Ndufs3), 3 neurotransmitter-related proteins (Gad1, Atp6d), 3 synaptic vesicle cycling-related proteins (Cplx1/2, Syt1, Syp), 2 calcium-binding proteins (Calm1, Calb2), and β-actin which did not show differences in the proteomics profiles for the OCNCX and odor-stimulated mice and was therefore used as a loading control. When the WB results were compared with the quantitative MS results, except in a few cases (e.g. Calm1 in the OCNCX experiment, Gad1 in the ST experiment), for the majority of the proteins compared, the results from the WB and MS analyses agreed with each other regarding general alteration trends (Fig. 2B). Because our MS analysis used OB samples pooled from multiple mice, we next tested a few proteins via IHC to validate the MS data at the level of individual mouse. We selected Cycs, Atp6d and Calm1 for localization and protein expression assessment through IHC analysis, which are three proteins that were regulated in the OCNCX dataset that belong to different functional groups. We found that compared to the control mice, the IHC signals for the Cycs protein were markedly decreased in the cell dendrites of the EPLs and GCLs of OCNCX mice OBs (Fig. 3A, B). These findings were consistent with the MS data on Cycs which showed an OCNCX to control ratio of ~0.6, a large

130

J O U RN A L OF P ROT EO M IC S 1 0 9 ( 2 01 4 ) 1 2 5 – 14 2

Fig. 1 – Quantitative proteomic analysis workflow and data overview. A, flowchart of the quantitative proteomics experiment. B, distribution of the standard deviation (S.D.) of proteins in three or more replicates from each quantitative proteomic experiment. A total of 90.1%, 92.2% and 91.7% of the proteins quantified in at least two experimental replicates in the OCNCX, LT and ST experiments, respectively, presented an S.D. value lower than 0.5 (the left of the dashed line). C, correlation analysis among the different proteomic experiments. The Pearson correlation coefficient (PCC) indicated that the quantitative datasets from the two odor-exposed experiments showed a better correlation (0.46), while each of these datasets exhibited a poor correlation with the OCNCX experiment (0.27 for LT and OCNCX, 0.31 for ST and OCNCX).

J O U RN A L OF P ROT EO M IC S 1 0 9 ( 2 01 4 ) 1 2 5 – 14 2

decrease (Table S3). For the protein Atp6d, slight decreases of the IHC signal could also be observed in the EPLs dendrites of the OBs of OCNCX mice (Fig. 3C, D). These findings were also consistent with the MS data on Atp6d which showed an OCNCX to control ratio of ~0.7, a slight decrease as well (Table S2). Calm1 (Calmodulin), a key calcium-binding protein has been reported to be intensively immunoreactive in mitral cells at postnatal day 1 but decreases during development, until totally disappearing at adulthood in wild-type rats [34]. The Calm1 IHC results for control mice showed an extremely low expression level in both the EPLs and GCLs, which were consistent with the theory regarding its disappearing in adult mitral cells. However, in OCNCX mice, strong signals could be observed in GCLs, which indicated that olfactory deficits may increase the expression level of Calm1 in granule cells, including both the cell body and dendrites (Fig. 3E, F). The Calm1 IHC data in OCNCX mice were consistent with the MS data which showed a ~1.7 fold increase compare to the control (Table S1).

3.3. Functional classification of differentially regulated proteins Among the 1285 proteins quantified in the OCNCX experiment, 112 and 269 proteins were determined to be up- and down-regulated, respectively, based on an OCNCX-to-control ratio of less than 0.67 or more than 1.5 (a cut-off threshold set

131

as 50%). From the 946 proteins quantified in the LT experiment, 51 and 89 proteins were observed to be up- and down-regulated, respectively. Of the 859 proteins quantified in the ST experiment, 46 and 57 proteins were found to be upand down-regulated, respectively. Overall, the up- and down-regulated proteins combined accounted for 29.6% of the quantified proteins from the OCNCX experiment, 14.8% from the LT experiment, and 12% from the ST experiment. All of the regulated proteins across all experimental groups are summarized in Table S3. To better understand the impact of olfactory deprivation or odor exposure on OB function, we subjected the regulated proteins (those whose expression level was changed by 50%) to a gene ontology-based enrichment analysis to identify the overrepresented functional categories for each of the experimental groups. The BiNGO program [30] was used to search for the GO terms which were statistically overrepresented among the proteins affected by odor deprivation or odor exposure for all three experimental groups. In the OCNCX mice, blockage of peripheral olfactory inputs had a profound effect on OB function. Based on the default BiNGO M. musculus database, the down-regulated proteins displayed overrepresentation among 11 GO terms in the biological process category, including metabolic process, transport and cell–cell signaling (Fig. 4A). The up-regulated

Fig. 2 – Western blot analysis of OB proteins in mice subjected to olfactory deprivation or following odor exposure. A, a total of ten proteins related to the energy metabolic process, synaptic vesicle cycling or calcium signaling pathways were selected to be analyzed by immunoblotting using β-actin as a loading control. B, comparison between WB ratios and MS ratios indicated that most of our WB results were consistent with the quantitative MS ratios (OCNCX: fold change of OCNCX to control; LT: fold change of LT to control; ST: fold change of ST to control). The nonparametric Kruskal–Wallis test, followed by Mann–Whitney test was applied. *p < 0.5, **p < 0.01, ***p < 0.001, #the ratio of Calb2 is 2.02. The group data shown are the average ± SEM.

132

J O U RN A L OF P ROT EO M IC S 1 0 9 ( 2 01 4 ) 1 2 5 – 14 2

proteins showed overrepresentation among 5 GO terms, with the most highly enriched terms being related to metabolic process (Fig. 4B). The results demonstrated first that the down-regulated proteins displayed much greater enrichment than the up-regulated proteins, which suggested a major role of down-regulation induced by olfactory deficits, and second that the terms related to energy metabolic process and cell communication appeared mainly to be downregulated, while the opposite was observed for those involved in translation. For the two odor-exposed experimental groups, as stated previously, a good correlation was observed. A total of 5 and 8 GO terms related to metabolic process and transport were enriched among the down-regulated proteins from the LT and ST groups, respectively (Fig. 4C, E). Among the up-regulated proteins in the two groups, metabolic process appeared to be the only relevant term (Fig. 4D, F). These results suggested a similar effect of the two odor stimulation experiments on OB function. For further comparison between the alterations of OB function induced by odor deprivation and stimulation, we categorized the proteins that were differentially expressed in both the odor-deprived and odor-stimulated (LT and ST) experimental groups into several typical functional groups. As shown in Table 1, the function of the regulated proteins

identified in the OCNCX mice and odor-stimulated mice covered many aspects of biological processes, among which signal transduction, transport, metabolic process (redox reactions & substance metabolism), and transcription/translation regulation were the four most-affected groups. In addition, a few of the regulated proteins were involved in proteolysis, calcium signaling, or other biological processes.

3.4. Alterations in metabolic process Our GO analysis revealed that proteins related to various processes were altered in the OBs of odor-deprived and odor-stimulated mice compared with wild-type mice. Nevertheless, in both types of experimental groups, among the upand down-regulated proteins, we found that metabolic process was the only highly overrepresented biological process term. To systematically investigate the effects on the metabolic system induced by olfactory deprivation or odor exposure, the three most important energy metabolic process pathways: oxidative phosphorylation, glycolysis/gluconeogenesis and the TCA cycle, were studied [35]. We extracted all of the proteins that matched the three pathways in the KEGG PATHWAY (http://www.genome.jp/kegg/) database from all of the quantified results (see Table S3). The resulting lists

Fig. 3 – Immunohistochemistry analysis of regulated proteins in OCNCX mice and control mice. A, Cytochrome c (Cycs) was widely expressed in the wild-type mice OB, especially in the external tufted cell layers (EPLs) and granule cell layers (GCLs). B, the Cycs signal was significantly decreased in the EPLs of OCNCX mice. C, the Atp6d signal was distributed similarly to that of Cycs in the wild-type mice OB. D, The Atp6d expression level in the EPLs of OB was slightly decreased in OCNCX mice. E, Calm1 (Calmodulin) was expressed in the glomerular layers but was nearly undetectable in the EPLs and GCLs of the wild-type mice OB. F, clear Calm1 signals were observed in GCLs in the OCNCX mice. Scale bar: A–D, E–F, 50 μm.

J O U RN A L OF P ROT EO M IC S 1 0 9 ( 2 01 4 ) 1 2 5 – 14 2

were then submitted to STRING for network analysis. Fig. 5 shows the metabolic networks of the quantified proteins in the three pathways derived from the three experiments. The interactions (edges) of the submitted proteins (nodes) are based on known or predicted protein–protein interactions, and the color represents the relative quantitation ratios of the proteins between the experimental and control groups. The node shape represents the involved pathways: oxidative phosphorylation (round), glycolysis/gluconeogenesis (square), TCA cycle (triangle) and multiple pathways (hexagon). The quantitative network analysis demonstrated that odor deprivation and odor stimulation induced distinct regulatory patterns (p < 0.001 (OCNCX vs LT), p < 0.001 (OCNCX vs ST) based on Student's t-test), and the two odor stimulation groups (LT vs ST) showed very similar impacts on OB metabolic process (p > 0.3 (LT vs ST) based on Student's t-test), but of different magnitudes. For OCNCX mice, there was significant downward co-regulation observed among the

133

three pathways, especially for oxidative phosphorylation, which is the main energy-generating process [36], indicating strikingly reduced energy metabolic activity (Fig. 5A). However, for the LT and ST mice, the detected co-regulation was not remarkable (Fig. 5B and C); slight down-regulation was found among most of the proteins, while in the protein networks related to oxidative phosphorylation and glycolysis/gluconeogenesis, several proteins, such as Gapdhs, Atp5j, Atp5k, Ndufv1, and Cox6b1, were significantly up-regulated after long-term or short-term odor stimulation. In contrast, another set of proteins, including Aldh7a1, Adh5, Pfkm, and Pgm2, showed a marked decrease in expression levels.

3.5. Alteration of synaptic transmission In the sensory processing of odors, the OB acts as an important relay station, exhibiting a multi-layered cellular architecture responsible for information transmission from

Fig. 4 – Gene ontology analysis of the distribution of biological processes in clusters of regulated proteins in the OCNCX group (A & B), LT group (C & D) and ST group (E & F). A, C, and E refer to the clustering of down-regulated proteins. B, D, and F refer to the clustering of up-regulated proteins. Cytoscape was used to visualize the BiNGO analysis of the biological process categories. Terms are depicted as nodes connected by arrows that represent hierarchies and relationships between terms. Node size is proportional to the number of proteins assigned to a given ontology term, whereas node color represents the corrected p-value (Benjamin Hochberg false discovery rate correction) corresponding to enrichment of the term.

134

J O U RN A L OF P ROT EO M IC S 1 0 9 ( 2 01 4 ) 1 2 5 – 14 2

Table 1 – Proteins that were differentially expressed in both the odor-deprived and odor-stimulated experimental groups. Accession

Gene name

Protein ratio a

Description OCNCX/WT

LT/Control

ST/Control

0.66 0.64 0.59 0.61 0.06

0.98 0.47 0.62 0.62 0.60

0.46 – 0.59 0.68 0.85

24.53 2.07

– 0.66

5.74 1.02

2.83 1.59 0.53 0.52 0.63 0.55 0.63

– – 0.59 0.41 0.16 2.62 1.53

0.42 1.76 – 0.70 0.33 – 1.23

0.62 0.66 0.64 0.64

0.52 – 0.62 1.27

1.02 0.45 1.10 0.65

Calcium signaling ANXA5_MOUSE CAPS2_MOUSE KPCB_MOUSE KPCE_MOUSE S10A5_MOUSE

Anxa5 Cadps2 Prkcb Prkce S100a5

Annexin A5 Calcium-dependent secretion activator 2 Protein kinase C beta type Protein kinase C epsilon type Protein S100-A5

Signal transduction ABI1_MOUSE GBG13_MOUSE

Abi1 Gng13

GNAI3_MOUSE STK39_MOUSE CADM4_MOUSE DCLK1_MOUSE DCLK2_MOUSE DLG4_MOUSE GBG4_MOUSE

Gnai3 Stk39 Cadm4 Dclk1 Dclk2 Dlg4 Gng4

GPSM1_MOUSE RHEB_MOUSE SAC1_MOUSE SEPT4_MOUSE

Gpsm1 Rheb Sacm1l Sept4

Abl interactor 1 Guanine nucleotide-binding protein G(I)/G(S)/G(O) subunit gamma-13 Guanine nucleotide-binding protein G(k) subunit alpha STE20/SPS1-related proline-alanine-rich protein kinase Cell adhesion molecule 4 Serine/threonine-protein kinase DCLK1 Serine/threonine-protein kinase DCLK2 Disks large homolog 4 Guanine nucleotide-binding protein G(I)/G(S)/G(O) subunit gamma-4 G-protein-signaling modulator 1 GTP-binding protein Rheb Phosphatidylinositide phosphatase SAC1 Septin-4

Proteolysis PSA1_MOUSE PSMD6_MOUSE

Psma1 Psmd6

Proteasome subunit alpha type-1 26S proteasome non-ATPase regulatory subunit 6

0.60 0.61

0.62 1.46

0.80 1.68

Redox reaction CYB5_MOUSE AL7A1_MOUSE CX6B1_MOUSE DHSA_MOUSE

Cyb5a Aldh7a1 Cox6b1 Sdha

4.86 0.59 0.33 0.60

3.06 0.39 1.90 1.01

– 0.70 1.27 0.65

NCPR_MOUSE NDUA2_MOUSE

Por Ndufa2

0.32 0.55

0.34 0.20

– –

NDUB9_MOUSE

Ndufb9

0.46

3.04

–

NDUBA_MOUSE

Ndufb10

0.54

0.60

1.13

NDUC2_MOUSE NDUS4_MOUSE

Ndufc2 Ndufs4

0.60 0.59

1.76 0.56

0.96 0.74

NDUV1_MOUSE

Ndufv1

0.53

2.00

1.43

USMG5_MOUSE

Usmg5

Cytochrome b5 Alpha-aminoadipic semialdehyde dehydrogenase Cytochrome c oxidase subunit 6B1 Succinate dehydrogenase [ubiquinone] flavoprotein subunit, mitochondrial NADPH–cytochrome P450 reductase NADH dehydrogenase [ubiquinone] 1 alpha subcomplex subunit 2 NADH dehydrogenase [ubiquinone] 1 beta subcomplex subunit 9 NADH dehydrogenase [ubiquinone] 1 beta subcomplex subunit 10 NADH dehydrogenase [ubiquinone] 1 subunit C2 NADH dehydrogenase [ubiquinone] iron–sulfur protein 4, mitochondrial NADH dehydrogenase [ubiquinone] flavoprotein 1, mitochondrial Up-regulated during skeletal muscle growth protein 5

0.63

0.42

–

Long-chain-fatty-acid–CoA ligase ACSBG1 Acyl carrier protein, mitochondrial Glyceraldehyde-3-phosphate dehydrogenase, testis-specific Glycogen phosphorylase, muscle form Dolichyl-diphosphooligosaccharide–protein glycosyltransferase subunit 1 40S ribosomal protein S13

0.61 0.58 1.59

0.80 0.59 1.98

0.29 0.85 1.18

0.65 0.64

0.76 3.17

0.45 –

0.62

0.60

0.77

Heterogeneous nuclear ribonucleoprotein F Heterochromatin protein 1-binding protein 3 Integrator complex subunit 5 60S ribosomal protein L32 40S ribosomal protein S30

0.65 0.55 0.05 1.59 0.33

0.26 0.83 0.05 0.72 –

0.80 0.64 0.02 0.61 0.06

Substance metabolism ACBG1_MOUSE Acsbg1 ACPM_MOUSE Ndufab1 G3PT_MOUSE Gapdhs PYGM_MOUSE RPN1_MOUSE

Pygm Rpn1

RS13_MOUSE

Rps13

Transcription or translation HNRPF_MOUSE Hnrnpf HP1B3_MOUSE Hp1bp3 INT5_MOUSE Ints5 RL32_MOUSE Rpl32 RS30_MOUSE Fau

135

J O U RN A L OF P ROT EO M IC S 1 0 9 ( 2 01 4 ) 1 2 5 – 14 2

Table 1 (continued) Accession

Gene name

Protein ratio a

Description OCNCX/WT

LT/Control

ST/Control

RT36_MOUSE SET_MOUSE TADBP_MOUSE

Mrps36 Set Tardbp

28S ribosomal protein S36, mitochondrial Protein SET TAR DNA-binding protein 43

0.64 0.57 0.32

0.47 0.56 –

0.86 0.99 0.66

Transport AT1B3_MOUSE

Atp1b3

0.51

–

2.14

CLCA_MOUSE SNP47_MOUSE VATG2_MOUSE

Clta Snap47 Atp6v1g2

Sodium/potassium-transporting ATPase subunit beta-3 Clathrin light chain A Synaptosomal-associated protein 47 V-type proton ATPase subunit G 2

0.63 0.10 0.62

– 0.12 0.62

0.55 – –

Others CRYAB_MOUSE MPP2_MOUSE NIPS2_MOUSE TAGL3_MOUSE THIO_MOUSE TRAP1_MOUSE

Cryab Mpp2 Gbas Tagln3 Txn Trap1

Alpha-crystallin B chain MAGUK p55 subfamily member 2 Protein NipSnap homolog 2 Transgelin-3 Thioredoxin Heat shock protein 75 kDa, mitochondrial

0.10 0.51 0.65 17.58 0.58 0.60

0.56 1.09 0.63 1.12 0.26 0.74

0.48 0.65 0.78 6.66 – 0.39

a

Proteins that were differentially expressed in both the odor-deprived experimental group (OCNCX/WT) and odor-stimulated experimental group (ST/Control or LT/Control) were categorized into several typical functional groups.

the nose to the brain [5]. Among the proteins found to be significantly regulated by olfactory deprivation or odor exposure, many were associated with synaptic transmission,

covering a wide range of synaptic-related functions, including neurotransmitter metabolic process and transport, vesicle release and recycling, and pre- or postsynaptic signal transduction.

Fig. 5 – Metabolic network of quantified OB proteins identified in response to olfactory deprivation (A), long-term odor stimulation (B) and short-term odor stimulation (C). Proteins quantified in our MS results that matched the three typical metabolic pathways (oxidative phosphorylation, glycolysis/gluconeogenesis and TCA cycle) in the KEGG PATHWAY database were extracted and submitted to STRING 9.0 for network analysis. The network includes 101 nodes and 1904 edges. The node color represents the quantitative ratio. The node shape represents the involved pathways: oxidative phosphorylation (round), glycolysis/gluconeogenesis (square), TCA cycle (triangle) and multiple pathways (hexagon).

136

J O U RN A L OF P ROT EO M IC S 1 0 9 ( 2 01 4 ) 1 2 5 – 14 2

To compare the changes in the synaptic transmission-related proteome associated with different experimental conditions, we constructed a network based on the quantified proteins related to synaptic transmission identified in our proteomics results (Fig. 6). According to the KEGG PATHWAY annotation, a total of 61 proteins in the neurotransmitter (GABA, dopamine and glutamate) metabolic process, synaptic vesicle cycle (SVC) and pre-/postsynaptic signal transduction pathways were mapped in all three of the experimental datasets (Table S4). The node shape represents the involved pathways: SVC (round), synaptic signal transduction (square) and neurotransmitter metabolic process (triangle). Our network analysis demonstrated that compared with the control mice, the OCNCX mice showed a significant alteration of synaptic transmission-related proteins, while the odor-stimulated (LT and ST) mice showed smaller changes, indicating rather different effects of the two conditions (Fig. 6). For the OCNCX mice, proteins involved in neurotransmitter metabolic process displayed a consistent down-regulation tendency, whereas those related to the SVC and signal transduction pathways changed discordantly (Fig. 6A). Compared to the OCNCX mice, the two odor-exposed groups (LT and ST) showed different changes. Most of those changes were moderate, but of different magnitudes between the two odor-exposed groups (Fig. 6B, C). Additionally, there were also a few proteins, such as Slc17a7 (~7.5 fold change in LT mice and

~3.7 fold change in ST mice), Gng4 (~1.5 fold change in LT mice and ~1.2 fold change in ST mice), and Prkcb (~0.6 fold change in LT mice and ~0.6 fold change in ST mice), that exhibited more significant changes, suggesting a more direct role in odor-induced regulation. To illustrate the changes in synaptic transmission induced by olfactory deficits in greater detail, we generated a cellular synaptic transmission map for the OCNCX mice group, covering proteins involved in the neurotransmitter metabolic process, SVC and pre-/postsynaptic signal transduction pathways (Fig. 7). Several observations were made based on the synaptic transmission map. First, regarding the neurotransmitter metabolic process, the proteins necessary for dopamine or GABA synthesis, such as Th, Gad1, and Gls were all decreased, and proteins involved in GABA degradation (Abat) and transport (Slc38a3, Slc32a1) also showed the trend of down-regulation (Fig. 7A). Additionally, our MS data showed that GABAA receptors (Gabrb1/2) were also decreased (~ 0.5-fold reduction). Second, among SVC proteins, we found that although many vesicle proteins, such as Syp (synaptophysin) were unchanged, Cplx1/2 (Complexin), one of the key proteins involved in calcium influx-induced vesicle release, was significantly down-regulated, and Syt1, the Ca2+ sensor in vesicle membranes, was slightly reduced (Fig. 7A, also see Fig. 2 for WB data). Finally, concerning pre- or

Fig. 6 – Network of synaptic transmission-related proteins in the OB of OCNCX mice (A), long-term odor stimulated mice (B) and short-term odor stimulated mice (C). The network includes 57 nodes and 198 edges, covering protein functional groups related to the synaptic vesicle cycle (SVC), synaptic signal transduction and neurotransmitter metabolic process. The node color represents the quantitative ratio. The node shape represents the involved pathways: SVC (round), synaptic signal transduction (square) and neurotransmitter metabolic process (triangle).

J O U RN A L OF P ROT EO M IC S 1 0 9 ( 2 01 4 ) 1 2 5 – 14 2

137

Fig. 7 – Summary of the changes in synaptic transmission pathways observed in response to olfactory deprivation. A, proteins involved in neurotransmitter metabolism and uptake, vesicle-mediated transmitter release. B, proteins involved in pre- or postsynaptic signal transduction pathways. The dotted, solid and unconnected solid arrows represent the pathways related to neurotransmitter metabolism and uptake, vesicle-mediated transmitter release and synaptic signal transduction, respectively. The dot color of proteins represents the quantitative ratio in OCNCX groups.

postsynaptic signal transduction, proteins in the G protein and calcium signaling pathways showed differential expression levels (defined as a > 1.5 fold change) (Fig. 7B). All of these findings implied an adaptation mechanism related to synaptic transmission in the olfactory-deprived mice.

3.6. Alterations of the metabolites and metabolic kinetics Our comparative proteomics has demonstrated that the metabolic activities of energy and neurotransmitters in OCNCX mice were decreased, as shown above. Since changes

138

J O U RN A L OF P ROT EO M IC S 1 0 9 ( 2 01 4 ) 1 2 5 – 14 2

of the metabolic proteins are closely related to the metabolites or metabolic kinetics of the tissue, the metabolite analysis can be used as an alternative evidence supporting the proteomics finding. The metabolic kinetics and metabolites in OB in both OCNCX and WT mice were analyzed by MRS. During the metabolic kinetic measurements, a balanced [1-13C] glucose enrichment is the most important parameter. With the infusion protocol, blood [1-13C]-glucose enrichment reached a stable value rapidly (Fig. S2). The concentrations of major metabolites were measured and were found that most were decreased, especially for glutamate (Glu) and GABA, in the OCNCX mice compared to the WT (Fig. 8). These two amino acids are the most important neurotransmitters for the brain functions. This result was consistent with our proteomics funding. Furthermore, in order to further validate the changes in the OCNCX OB, the metabolic kinetics was also investigated. The 13C enrichment in GABA-C2, Glu-C4, and glutamine (Gln-C4) were measured and the results revealed that during the entire infusion periods, the enrichments of these metabolites in OCNCX group were lower than the control (Fig. S3). Comparing these two groups, the estimated metabolic rates were all significantly reduced (p < 0.001) (Table S5). The decreased metabolic rates agree with the results of the proteomics.

4. Discussion In the mammalian OB, neural activity driven by external stimuli can lead to experience-dependent changes in structures and functions [37]. In this study, we employed a quantitative proteomics approach to study proteome-wide alterations in the OB under different levels of neural activity, with the aim of finding new clues related to olfactory input-dependent signaling alterations using wild type mice as control. The four tested OB conditions were as follows: devoid of peripheral input, wild-type control, and short-term and long-term odor exposures. Among the various experimental models for the blockade

of olfactory inputs [38,39], we choose OCNCX mice because they possess the characteristics of general anosmia and minimal invasiveness. To adjust the peripheral olfactory inputs, we designed short-term and long-term treatment groups. Odor stimulation paradigms that differ in timescales have been thought to induce different mechanisms mediated by distinct regions of the olfactory system [23]. In total, we obtained relative quantitation information on 1285, 946 and 859 proteins in the OCNCX, LT and ST experiments, respectively. To verify the MS-based data, WB and IHC were performed on selected proteins, validating the quantification results. For functional validation, the concentration of metabolites and metabolic kinetics in OB were analyzed by MRS using the 13C isotopic labeled glucose method. Based on the obtained data, the percentages of significantly regulated proteins in the OCNCX, LT and ST datasets were 29.6%, 14.8% and 12.0%, respectively. The different percentages of significantly regulated proteins across the three experiments indicate varying degrees of impact on the host OB. Olfactory deprivation as a result of permanent genetic knockout induced much greater alterations in the OB proteome than odor exposure, while the two odor exposure experiments had a similar effect, which was further confirmed by our correlation analysis (Fig. 1C). Our quantitative proteomics analysis suggested that odor exposure and olfactory deprivation lead to different alterations in the OB proteome, involving a wide range of biological processes, especially for metabolic and synaptic transmission pathways (Fig. 4, Table 1). The results from the subsequent functional studies by MRS demonstrated that the concentrations and metabolic kinetics of the neurotransmitters were decreased significantly in OCNCX mice compared to the wild type, which further confirmed our MS finding. Our study provides new clues related to peripheral olfactory input-specific OB signaling, which may be used to achieve a better understanding of the molecular and biochemical mechanisms of OB plasticity.

4.1. Energy metabolism

Fig. 8 – The total concentrations of metabolites in OB of OCNCX and WT mice. NAA: N-Acetylaspartic acid; Ala: Alanine; Lac: Lactate; Tau: Taurine; Asp: Aspartate; Gln: Glutamine; Glu: Glutamate; GABA: gamma-aminobutyric acid.

The glomerular layer of the OB, which exhibits the most intensive capillary network, has been reported to be an area presenting particularly high energy consumption [40]. Most energy is produced for presynaptic action potential generation and the activity of postsynaptic receptors [41–43]. Coincidentally, in the present study, among the up- and down-regulated proteins in both the odor deprivation and stimulation groups, metabolic process was found to be the only significantly over-represented biological process term (Fig. 4, Table 1), indicating the high sensitivity of energy metabolic processes to peripheral olfactory inputs. For OCNCX mice, although attenuation of electrophysiology and metabolic neural activity has been reported in the mice after naris closure [44], the quantitative proteomics data reported here provide new clues for understanding olfactory deprivation at the protein level. Our data showed that in the OCNCX mice, there was significant downward co-regulation among the proteins involved in metabolic pathways (Fig. 5A), especially for proteins related to oxidative phosphorylation, which suggested that energy metabolic activity in the OB of

J O U RN A L OF P ROT EO M IC S 1 0 9 ( 2 01 4 ) 1 2 5 – 14 2

OCNCX mice was reduced. Consistent with this, our MRS data showed that many metabolites involved in TCA cycles like Aspartate, Gln, Glu and GABA were decreased (Fig. 8) [23]. We speculate that this may be due to the long-term absence of olfactory inputs in OCNCX mice, which ultimately leads to reduced energy consumption. For the two odor-exposed groups, network analysis also showed a reduced status of energy metabolic activity, although this reduction was much smaller compared to the OCNCX group (Fig. 5B, C). This suggests that not only odor deprivation but also odor exposure attenuates energy metabolic activity to some extent, likely due to olfactory habituation [45]. Several unexpected protein alterations were also found in the two odor-exposed groups. For instance, from the control to the ST and then to the LT mice, the expression levels of several important metabolic proteins showed a gradual increasing trend (Fig. 5B, C). These proteins included glyceraldehyde-3-phosphate dehydrogenase (Gapdhs), Atp synthase-related protein (Atp5j and Atp5k), NADH dehydrogenase (Ndufv1 and Ndufc2), and Cytochrome c oxidase (Cox6b1). However, the question of why these proteins exhibit sensitivity to external stimulation of the OB has yet to be answered, and the biochemical mechanisms involved remain to be investigated.

4.2. Neurotransmitter metabolism Sensory experience influences neurotransmitter synthesis in the OB, the expression of the catecholamine biosynthetic enzyme tyrosine hydroxylase (TH), a key enzyme involved in dopamine biosynthesis, is rapidly down-regulated in the dopamine neurons intrinsic to the OB in naris-occluded or olfactory-deprived mice/rats [16,46,47]. In addition, Gad1, an isoform of glutamic acid decarboxylase (GAD) that catalyzes the production of GABA, has been reported to be reduced by naris occlusion-induced deafferentation [17]. In accord with these previous reports, down-regulation of both TH and Gad1 was also observed in OCNCX mice in the present study (see Fig. S1 for TH WB results, Fig. 2 for Gad1). This indicates that olfactory deprivation via the genetic knockout (OCNCX) results in down-regulation of GABA and dopamine synthesis. From the results of functional validation, two amino acids, glutamate and GABA, which are the most important neurotransmitters for the brain functions, were also reduced significantly (Fig. 8). Furthermore, our metabolic kinetics analysis of glutamine, glutamate and GABA revealed that metabolic rates of these amino acids were lower (Fig. S3, Table S5). The present study indicates that sensory experience, including blockage or intensification of peripheral olfactory inputs, may be able to directly regulate dopamine, glutamate and GABA synthesis. In addition to neurotransmitter synthesis, data from the OCNCX group showed that, several key proteins involved in transportation and degradation were also decreased (Fig. 7A). For example, Slc32a1, an inhibitory vesicular amino acid transporter involved in the uptake of GABA and glycine into synaptic vesicles, and Abat, the 4-aminobutyrate amino transferase that catalyzes the hydrolysis of GABA, were both reduced significantly by olfactory deficits, suggesting a possible reduction in the numbers of the inhibitory neurons in the OBs of olfactory-deprived mice. It has been reported

139

that nostril closure decreases the number of newborn granule cells in the mouse OB, and the frequency of miniature synaptic inhibitory events received by mitral cells is reduced, whereas action potential-dependent GABA release is dramatically enhanced [48,49]. This is considered to be a compensatory mechanism to increase remnant inhibitory neural cell excitability and maintain normal functioning of the mouse OB. It is also possible that this unique form of adaptive response is present in OCNCX mice.

4.3. Synaptic vesicle cycle Neurotransmitter release and uptake are mediated by synaptic vesicle exocytosis and endocytosis at nerve terminals. When an action potential invades a nerve terminal, calcium influx stimulates fast neurotransmitter release, thereby ensuring normal synaptic transmission [50,51]. In the present study, 25 proteins involved in the SVC in both the odordeprived and odor-stimulated groups were quantified (Fig. 6), but to our surprise, many SVC proteins were unchanged, even in the OCNCX group. However, several proteins that participate in synaptic vesicle exocytosis were found to be altered in OCNCX mice (Fig. 7A). Syt1, the Ca2+ sensor for fast exocytosis, was slightly reduced (~ 0.7-fold change), while Cplx1/2 (Complexin), the protein that stabilizes the SNARE complex (essential for the proper positioning of Syt1), showed marked down-regulation (~0.5-fold change). It has been reported that fast synaptic vesicle exocytosis is controlled by a Complexin/Syt1 switch. Prior to neurotransmitter release, Syt1 competes with Complexin for SNARE complex binding, thereby dislodging Complexin from the SNARE complex in a calcium-dependent manner, with Syt1 acting as the calcium sensor initiating this process [52,53]. Therefore, we speculate that the lower ratio of Complexin to Syt1 maybe a self-regulatory mechanism of the OB for increasing calcium influx-induced neurotransmitter release at presynaptic terminals in the absence of olfactory inputs. In accord with this notion, we found that two other membrane components of the SNARE complex, Snap25 (~1.2-fold change) and Stx1a (~ 1.8-fold change), were both increased in OCNCX mice. Snap25 and Stx1a are responsible for creating an unstable intermediate prior to fusion pore opening [54]. These results suggest that the capacity for calcium influx-induced fast neurotransmitter release was enhanced in the OB of OCNCX mice, which may constitute part of the neurochemical adaptation mechanism to compensate for deficits of peripheral olfactory inputs.

4.4. Synaptic signal transduction As a system involving a multi-layer structure, the OB exhibits a large number of local neural circuits, which critically determine the input–output functions of the OB [7]. In olfactory-deprived mice, a previous study found that early olfactory deprivation leads to up-regulation of dopamine D2 receptor sites in the rat OB [55], which was thought to represent an attempt to neurochemically adapt to reduced dopaminergic activity and maintain OB function. In the present study, approximately 22 proteins involved in the

140

J O U RN A L OF P ROT EO M IC S 1 0 9 ( 2 01 4 ) 1 2 5 – 14 2

modulation of pre- or post-synaptic signal transduction were identified in OCNCX mice (Fig. 6). Data from OCNCX mice showed a significant decrease in GABAA receptors (Gabrb1/2) (Fig. 7B), indicating similar sensory experience-dependent OB plasticity, with GABAA receptor-mediated inhibitory synapses being reduced in OCNCX mice. In addition, many synaptic signaling proteins were found to be regulated in OCNCX mice. A total of six out of eight Gi/o proteins were found to be increased. Two calcium signaling proteins, CaM and Ppp3cb, were up-regulated as well (Fig. 7B). There is evidence that local inhibitory regulation of neurotransmitter release in the nervous system is mediated via Gi/o-coupled receptors, and Gβα-mediated inhibition is thought to be involved in this mechanism downstream of the calcium signaling [56,57]. These phenomena may reflect an attempt of intracellular signals to increase local inhibitory regulation in OCNCX mice, and such G protein-mediated inhibitory regulation may play a role in gaining signaling control in the absence of olfactory inputs.

4.5. Transcription and calcium signaling Experience-driven synaptic activity causes membrane depolarization and calcium influx into select neurons within a neural circuit, which then triggers a wide variety of cellular changes that alter synaptic connectivity within the neural circuit. Calcium influx-induced synapse remodeling is linked to activation at the gene transcription level [58,59]. In this study, many proteins involved in transcriptional regulation were found to be regulated (Table 1, Fig. 7B). As for the up-regulated, given transcription intermediary factor 1-beta (Tif1b) as an example, Tif1b was found to be increased (~ 1.9 fold) in the OCNCX dataset. For down-regulated proteins, we identified Ndrg4, a protein related to the maintenance of intracerebral brain-derived neurotrophic factor (BDNF) levels [60], as being down-regulated in the OCNCX mice and long-term odor-stimulated mice (Table S3). It has been reported that BDNF is significantly decreased in olfactory-deprived mice [61]. Considering that the expression of BDNF is closely associated with the formation of GABAergic synapses [62,63],we speculate that the down-regulation of Ndrg4 may result in a decrease in BDNF in OCNCX mice, which is further involved in down-regulation of inhibitory neurons in the OB of OCNCX mice.

5. Conclusions In conclusion, the present study employed a quantitative proteomics approach, following the blockade or intensification of peripheral olfactory inputs, to study proteome-wide alterations in the OB under different neural activity levels. Mass spectrometry quantitation data on selected proteins were validated via WB and IHC analyses. Through systemic network analysis, alterations in synaptic transmission and metabolic pathways were identified. Additionally, a number of regulated proteins involved in transcription and other functions were also quantified. Further studies addressing how these proteins are involved in the development of OB plasticity induced by olfactory deprivation or odor exposure

may lead to a more comprehensive and in-depth understanding of OB function.

Conflict of interest The authors declare that they have no conflict of interest. Supplementary data to this article can be found online at http://dx.doi.org/10.1016/j.jprot.2014.06.023.

Acknowledgments We thank Dr. Minmin Luo (NIBS, Beijing) for supplying female heterozygous OCNCX mice, Yong Zhao for the assistance in mass spectrometry analysis, Zhixiang Xu and Yunling Gao for the assistance in IHC analysis, and Dr. Graeme F. Mason (Yale University) for the metabolic kinetics measurements. This work was supported by grants from the Ministry of Science and Technology of the People's Republic of China (2013CB911102, 2012BAI23B02), Natural Science Foundation of China (NSFC, 91132307/H09, 31171061/C090208, 21105116), the Chinese Academy of Sciences (XDB02050005), and the 111 Project of China (B06018) to Wuhan University.

REFERENCES

[1] Firestein S. How the olfactory system makes sense of scents. Nature 2001;413:211–8. [2] Su CY, Menuz K, Carlson JR. Olfactory perception: receptors, cells, and circuits. Cell 2009;139:45–59. [3] Buck L, Axel R. A novel multigene family may encode odorant receptors: a molecular basis for odor recognition. Cell 1991;65:175–87. [4] Kiyokage E, Pan YZ, Shao Z, Kobayashi K, Szabo G, Yanagawa Y, et al. Molecular identity of periglomerular and short axon cells. J Neurosci 2010;30:1185–96. [5] Mori K, Nagao H, Yoshihara Y. The olfactory bulb: coding and processing of odor molecule information. Science 1999;286:711–5. [6] Whitman MC, Greer CA. Adult neurogenesis and the olfactory system. Prog Neurobiol 2009;89:162–75. [7] Aungst JL, Heyward PM, Puche AC, Karnup SV, Hayar A, Szabo G, et al. Centre-surround inhibition among olfactory bulb glomeruli. Nature 2003;426:623–9. [8] Petreanu L, Alvarez-Buylla A. Maturation and death of adult-born olfactory bulb granule neurons: role of olfaction. J Neurosci 2002;22:6106–13. [9] Mori K, Shepherd GM. Emerging principles of molecular signal processing by mitral/tufted cells in the olfactory bulb. Semin Cell Biol 1994;5:65–74. [10] Yokoi M, Mori K, Nakanishi S. Refinement of odor molecule tuning by dendrodendritic synaptic inhibition in the olfactory bulb. Proc Natl Acad Sci U S A 1995;92:3371–5. [11] Carleton A, Petreanu LT, Lansford R, Alvarez-Buylla A, Lledo PM. Becoming a new neuron in the adult olfactory bulb. Nat Neurosci 2003;6:507–18. [12] Lledo PM, Saghatelyan A. Integrating new neurons into the adult olfactory bulb: joining the network, life–death decisions, and the effects of sensory experience. Trends Neurosci 2005;28:248–54.

J O U RN A L OF P ROT EO M IC S 1 0 9 ( 2 01 4 ) 1 2 5 – 14 2

[13] Jones SV, Choi DC, Davis M, Ressler KJ. Learning-dependent structural plasticity in the adult olfactory pathway. J Neurosci 2008;28:13106–11. [14] Rochefort C, Gheusi G, Vincent JD, Lledo PM. Enriched odor exposure increases the number of newborn neurons in the adult olfactory bulb and improves odor memory. J Neurosci 2002;22:2679–89. [15] Woo CC, Hingco EE, Taylor GE, Leon M. Exposure to a broad range of odorants decreases cell mortality in the olfactory bulb. Neuroreport 2006;17:817–21. [16] Cho JY, Min N, Franzen L, Baker H. Rapid down-regulation of tyrosine hydroxylase expression in the olfactory bulb of naris-occluded adult rats. J Comp Neurol 1996;369:264–76. [17] Parrish-Aungst S, Kiyokage E, Szabo G, Yanagawa Y, Shipley MT, Puche AC. Sensory experience selectively regulates transmitter synthesis enzymes in interglomerular circuits. Brain Res 2011;1382:70–6. [18] Sharp FR, Kauer JS, Shepherd GM. Local sites of activity-related glucose metabolism in rat olfactory bulb during olfactory stimulation. Brain Res 1975;98:596–600. [19] Chaigneau E, Oheim M, Audinat E, Charpak S. Two-photon imaging of capillary blood flow in olfactory bulb glomeruli. Proc Natl Acad Sci U S A 2003;100:13081–6. [20] Mesholam RI, Moberg PJ, Mahr RN, Doty RL. Olfaction in neurodegenerative disease: a meta-analysis of olfactory functioning in Alzheimer's and Parkinson's diseases. Arch Neurol 1998;55:84–90. [21] Serby M, Larson P, Kalkstein D. Olfactory sense in psychoses. Biol Psychiatry 1990;28:830. [22] Brunet GHG Lisa J, Ngai John. General anosmia caused by a targeted disruption of the mouse olfactory cyclic nucleotide gated cation channel. Neuron 1996;17:681–93. [23] Chaudhury D, Manella L, Arellanos A, Escanilla O, Cleland TA, Linster C. Olfactory bulb habituation to odor stimuli. Behav Neurosci 2010;124:490–9. [24] Zhang LK, Chai F, Li HY, Xiao G, Guo L. Identification of host proteins involved in Japanese encephalitis virus infection by quantitative proteomics analysis. J Proteome Res 2013;12:2666–78. [25] Chen X, Wu D, Zhao Y, Wong BH, Guo L. Increasing phosphoproteome coverage and identification of phosphorylation motifs through combination of different HPLC fractionation methods. J Chromatogr B Analyt Technol Biomed Life Sci 2011;879:25–34. [26] Rorabacher DB. Statistical treatment for rejection of deviant values: critical values of Dixon's “Q” parameter and related subrange ratios at the 95% confidence level. Anal Chem 1991;63:139–46. [27] Wang J, Jiang L, Jiang Y, Ma X, Chowdhury GM, Mason GF. Regional metabolite levels and turnover in the awake rat brain under the influence of nicotine. J Neurochem 2010;113:1447–58. [28] Wang J, Du H, Jiang L, Ma X, de Graaf RA, Behar KL, et al. Oxidation of ethanol in the rat brain and effects associated with chronic ethanol exposure. Proc Natl Acad Sci U S A 2013;110:14444–9. [29] Mason GF, Falk Petersen K, de Graaf RA, Kanamatsu T, Otsuki T, Shulman GI, et al. A comparison of (13)C NMR measurements of the rates of glutamine synthesis and the tricarboxylic acid cycle during oral and intravenous administration of [1-(13)C]glucose. Brain Res Brain Res Protoc 2003;10:181–90. [30] Maere S, Heymans K, Kuiper M. BiNGO: a Cytoscape plugin to assess overrepresentation of gene ontology categories in biological networks. Bioinformatics 2005;21:3448–9. [31] Moriya Y, Itoh M, Okuda S, Yoshizawa AC, Kanehisa M. KAAS: an automatic genome annotation and pathway reconstruction server. Nucleic Acids Res 2007;35:W182–5.

141

[32] McNamara AM, Magidson PD, Linster C, Wilson DA, Cleland TA. Distinct neural mechanisms mediate olfactory memory formation at different timescales. Learn Mem 2008;15:117–25. [33] Wilson DA, Linster C. Neurobiology of a simple memory. J Neurophysiol 2008;100:2–7. [34] Bastianelli E, Pochet R. Calmodulin, calbindin-D28k, calretinin and neurocalcin in rat olfactory bulb during postnatal development. Brain Res Dev Brain Res 1995;87:224–7. [35] Smeitink JA, Zeviani M, Turnbull DM, Jacobs HT. Mitochondrial medicine: a metabolic perspective on the pathology of oxidative phosphorylation disorders. Cell Metab 2006;3:9–13. [36] Mitchell P, Moyle J. Chemiosmotic hypothesis of oxidative phosphorylation. Nature 1967;213:137–9. [37] Sutton MA, Schuman EM. Dendritic protein synthesis, synaptic plasticity, and memory. Cell 2006;127:49–58. [38] Korol DL, Brunjes PC. Rapid changes in 2-deoxyglucose uptake and amino acid incorporation following unilateral odor deprivation: a laminar analysis. Brain Res Dev Brain Res 1990;52:75–84. [39] Xu Z, Wang L, Chen G, Rao X, Xu F. Roles of GSK3beta in odor habituation and spontaneous neural activity of the mouse olfactory bulb. PLoS One 2013;8:e63598. [40] Johnson BA, Leon M. Modular representations of odorants in the glomerular layer of the rat olfactory bulb and the effects of stimulus concentration. J Comp Neurol 2000;422:496–509. [41] Chaigneau E, Tiret P, Lecoq J, Ducros M, Knopfel T, Charpak S. The relationship between blood flow and neuronal activity in the rodent olfactory bulb. J Neurosci 2007;27:6452–60. [42] Lecoq J, Tiret P, Najac M, Shepherd GM, Greer CA, Charpak S. Odor-evoked oxygen consumption by action potential and synaptic transmission in the olfactory bulb. J Neurosci 2009;29:1424–33. [43] Nawroth JC, Greer CA, Chen WR, Laughlin SB, Shepherd GM. An energy budget for the olfactory glomerulus. J Neurosci 2007;27:9790–800. [44] Philpot BD, Foster TC, Brunjes PC. Mitral/tufted cell activity is attenuated and becomes uncoupled from respiration following naris closure. J Neurobiol 1997;33:374–86. [45] Buonviso N, Gervais R, Chalansonnet M, Chaput M. Short-lasting exposure to one odour decreases general reactivity in the olfactory bulb of adult rats. Eur J Neurosci 1998;10:2472–5. [46] Brinon JG, Crespo C, Weruaga E, Martinez-Guijarro FJ, Aijon J, Alonso JR. Bilateral olfactory deprivation reveals a selective noradrenergic regulatory input to the olfactory bulb. Neuroscience 2001;102:1–10. [47] Baker H, Cummings DM, Munger SD, Margolis JW, Franzen L, Reed RR, et al. Targeted deletion of a cyclic nucleotide-gated channel subunit (OCNC1): biochemical and morphological consequences in adult mice. J Neurosci 1999;19:9313–21. [48] Saghatelyan A, Roux P, Migliore M, Rochefort C, Desmaisons D, Charneau P, et al. Activity-dependent adjustments of the inhibitory network in the olfactory bulb following early postnatal deprivation. Neuron 2005;46:103–16. [49] Angely CJ, Coppola DM. How does long-term odor deprivation affect the olfactory capacity of adult mice? Behav Brain Funct 2010;6:26. [50] Katz B. The Release of Neural Transmitter Substances. Liver-pool. UK: Liverpool University Press; 1969. [51] Sudhof TC. The synaptic vesicle cycle. Annu Rev Neurosci 2004;27:509–47. [52] Fernandez-Chacon R, Konigstorfer A, Gerber SH, Garcia J, Matos MF, Stevens CF, et al. Synaptotagmin I functions as a calcium regulator of release probability. Nature 2001;410:41–9.

142

J O U RN A L OF P ROT EO M IC S 1 0 9 ( 2 01 4 ) 1 2 5 – 14 2

[53] Tang J, Maximov A, Shin O-H, Dai H, Rizo J, Südhof TC. A Complexin/synaptotagmin 1 switch controls fast synaptic vesicle exocytosis. Cell 2006;126:1175–87. [54] Fasshauer D, Margittai M. A transient N-terminal interaction of SNAP-25 and syntaxin nucleates SNARE assembly. J Biol Chem 2004;279:7613–21. [55] Guthrie KM, Pullara JM, Marshall JF, Leon M. Olfactory deprivation increases dopamine D2 receptor density in the rat olfactory bulb. Synapse 1991;8:61–70. [56] Dolphin AC. G protein modulation of voltage-gated calcium channels. Pharmacol Rev 2003;55:607–27. [57] Blackmer T, Larsen EC, Takahashi M, Martin TF, Alford S, Hamm HE. G protein betagamma subunit-mediated presynaptic inhibition: regulation of exocytotic fusion downstream of Ca2 + entry. Science 2001;292:293–7. [58] Flavell SW, Greenberg ME. Signaling mechanisms linking neuronal activity to gene expression and plasticity of the nervous system. Annu Rev Neurosci 2008;31:563–90.

[59] West AE, Greenberg ME. Neuronal activity-regulated gene transcription in synapse development and cognitive function. Cold Spring Harb Perspect Biol 2011;3. [60] Yamamoto H, Kokame K, Okuda T, Nakajo Y, Yanamoto H, Miyata T. NDRG4 protein-deficient mice exhibit spatial learning deficits and vulnerabilities to cerebral ischemia. J Biol Chem 2011;286:26158–65. [61] McLean JH, Darby-King A, Bonnell WS. Neonatal olfactory sensory deprivation decreases BDNF in the olfactory bulb of the rat. Brain Res Dev Brain Res 2001;128:17–24. [62] Kohara K, Yasuda H, Huang Y, Adachi N, Sohya K, Tsumoto T. A local reduction in cortical GABAergic synapses after a loss of endogenous brain-derived neurotrophic factor, as revealed by single-cell gene knock-out method. J Neurosci 2007;27:7234–44. [63] Sakata K, Woo NH, Martinowich K, Greene JS, Schloesser RJ, Shen L, et al. Critical role of promoter IV-driven BDNF transcription in GABAergic transmission and synaptic plasticity in the prefrontal cortex. Proc Natl Acad Sci U S A 2009;106:5942–7.