Short-Term and Long-Term Sensitization Differentially Alters the Composition of an Anterograde Transport Complex in Aplysia

Abstract Long-term memory formation requires anterograde transport of proteins from the soma of a neuron to its distal synaptic terminals. This allows new synaptic connections to be grown and existing ones remodeled. However, we do not yet know which proteins are transported to synapses in response to activity and temporal regulation. Here, using quantitative mass spectrometry, we have profiled anterograde protein cargos of a learning-regulated molecular motor protein kinesin [Aplysia kinesin heavy chain 1 (ApKHC1)] following short-term sensitization (STS) and long-term sensitization (LTS) in Aplysia californica. Our results reveal enrichment of specific proteins associated with ApKHC1 following both STS and LTS, as well as temporal changes within 1 and 3 h of LTS training. A significant number of proteins enriched in the ApKHC1 complex participate in synaptic function, and, while some are ubiquitously enriched across training conditions, a few are enriched in response to specific training. For instance, factors aiding new synapse formation, such as synaptotagmin-1, dynamin-1, and calmodulin, are differentially enriched in anterograde complexes 1 h after LTS but are depleted 3 h after LTS. Proteins including gelsolin-like protein 2 and sec23A/sec24A, which function in actin filament stabilization and vesicle transport, respectively, are enriched in cargos 3 h after LTS. These results establish that the composition of anterograde transport complexes undergo experience-dependent specific changes and illuminate dynamic changes in the communication between soma and synapse during learning.


Introduction
Activity-dependent synaptic plasticity is key to the formation of long-term memories (LTMs) and involves changes in the expression of specific genes as well as the synthesis of new proteins. This then results in structural remodeling of the synapse (Martin et al., 2000;Kandel, 2001;Sutton and Schuman, 2006). Mechanistic investigations of LTM in vertebrate and invertebrate animal models have identified different phases of memory formation-an early short-term phase independent of gene expression changes and new protein synthesis, which lasts typically 30 min to 1 h, an intermediate-term phase prolonging for 1-3 h, which requires protein synthesis, but not gene transcription, and a later long-term phase, which lasts .8 h and is sensitive to gene expression disruptions and protein synthesis (Emptage and Carew, 1993;Ghirardi et al., 1995;Sutton and Carew, 2000;Kandel, 2001;Sutton and Schuman, 2006;Byrne and Hawkins, 2015). This alteration in gene expression resulting in synaptic structural changes poses a compelling question in cell biology: how are somatic-tosynaptic processes coordinated and comprised?
Such control of synaptic function by KIFs raises an important and unanswered question: do the contents of transported cargo change during learning and LTM? We speculated that identifying the transported proteome during the dynamic process of memory formation would provide insights into the molecules critical for synapse remodeling and/or new synapse formation. Furthermore, how the composition of anterogradely transported proteins is altered during learning could enlighten us to clinically significant mechanisms underlying transport-related neurologic diseases including Alzheimer's disease, amyotrophic lateral sclerosis, and Parkinson's disease, as well as Charcot-Marie-Tooth disease (Zhao et al., 2001;Lazarov et al., 2007;Bilsland et al., 2010;Chu et al., 2012;Gunawardena et al., 2013).
To understand the potentially evolving composition of anterograde transport complexes during learning and LTM, we took advantage of the robust learning paradigm of sensitization, a form of nonassociative learning in the sea slug Aplysia californica. Specifically, we used sensitization of siphon withdrawal reflex, whereby a single electric tail shock induces short-term sensitization (STS) in the siphon withdrawal response of Aplysia lasting 30 min to 1 h. Four repeated tail shocks result in long-term sensitization (LTS), with .24 h of siphon withdrawal (Pinsker et al., 1970(Pinsker et al., , 1973Carew et al., 1971Carew et al., , 1981Hawkins et al., 1989;Antonov et al., 1999Antonov et al., , 2003. Previously, we reported that the expression of a molecular motor protein kinesin called Aplysia kinesin heavy chain 1 (ApKHC1) in sensory and motor neurons of gill withdrawal reflex are upregulated by serotonin, a modulatory neurotransmitter involved in sensitization and learning (Puthanveettil et al., 2008). Given this finding, we here evaluated ApKHC1 cargos following both STS and LTS training in Aplysia. We isolated ApKHC1 complexes following training and studied their composition using quantitative mass spectrometry. Using isobaric tandem mass tags (TMTs), we found that 28% of identified anterogradely trafficked proteins participated in synaptic function. A quantitative comparison of cargo protein enrichment among different sensitization conditions demonstrated the presence of activity-dependent regulation of anterograde transport, suggesting that proteins on the transport complex undergo temporal alteration based on the different stages of memory formation.

Behavioral training and dissection
Eight-month-old Aplysia weighing 80-100 g were used for the study. Long-term sensitization training was conducted following the study by Pinsker et al. (1973). Since, sea slugs are hermaphroditic animals, sex-related differentiations are not applicable to this study. Briefly, each animal was housed individually 48 h before behavioral training and maintained in circulating seawater at 18°C. Responsiveness was pretested by light siphon touch using a brush. Animals with a response time of !10 s were selected for the experiment. These qualifying animals were then randomly selected into a control (untrained) group, a short-term sensitized (single shock) group, and two longterm sensitized (four shocks) groups, with no significant variance among the groups. Sensitization was induced by delivering four 6 V, 15 mA shocks for 1.5 s every 3 s. The short-term sensitized group received this once, and the long-term sensitized groups received it four times in 30 min intervals. Electrical shocks were delivered on the posterior body wall in the tail region.
The entire CNS (without buccal ganglia) was isolated 1 h after the delivery of single shocks (STS) or four shocks (1 h; LTS) or 3 h after four shocks (3 h; LTS). The percentage change in siphon withdrawal duration post-test versus pretest is provided in Figure 1, B and C (Extended Data Fig. 1-1, tables 1-5). To anesthetize the animals, an intracoelomic injection of 0.38 M MgCl 2 was administered. The cerebral, pedal, and pleural ganglia ring was collected with attached abdominal ganglia and, to block synaptic transmission, transferred to ice-cold 1:1 artificial seawater/isotonic MgCl 2 before lysis.

ApKHC1 complex isolation and Western blot analysis
For each N = 1, four CNSs (n = 4) were pooled together and lysed to extract one final sample in a buffer containing 50 mM Tris, pH 7.5, 1 mM EDTA, 150 mM NaCl, 1 mM DTT, 0.5% NP-40, EDTA-free protease inhibitor tablet (Roche), and phosphatase inhibitor cocktails 2 and 3 (Sigma-Aldrich) at 4°C in individual Dounce homogenizers. Lysates were spun down at 10,600 Â g for 10 min at 4°C. Supernatants were incubated in protein A/G agarose beads, pre-equilibrated, and blocked with lysis buffer for 2 h to control any nonspecific binding of unwanted proteins. After collecting the supernatant by brief centrifugation at 100 Â g, anti-ApKHC1 antibody (Extended Data Fig. 1-1, tables 1-5) was added and incubated/rotated for 12-14 h. Protein A/G beads were added to the supernatants and incubated for 1 h. The bead-antibody complex was collected and washed three times by brief centrifugation. For SDS-PAGE analysis, complexes were eluted in 2Â Laemmli buffer, and for mass spectrometric analysis, they were eluted in 5% SDS (v/v). To carry out Western blot (WB) analyses, the protein concentration was determined using a Qubit protein assay kit (Thermo Fisher Scientific). Approximately 100 mg of protein was loaded onto SDS-PAGE gels and used for WB analyses. The antibodies used are listed in Extended Data Figure 1-1 (tables 1-5). The target protein bands were detected using HRPconjugated anti-rabbit or anti-mouse secondary antibodies (Cell Signaling Technology) at a 1:5000 dilution followed by visualization using chemiluminescence (GE Health Care). The signal bands were analyzed by ImageJ.

TMT quantitative proteomics and mass spectrometry
Samples in 5% SDS (v/v) were brought to 24 ml with an additional 5% SDS and processed for digestion using micro-S-Traps (Protifi) according to manufacturer instructions. Briefly, Figure 1. Identification of kinesin complexes in Aplysia CNS. A, The approach used to capture ApKHC1-specific protein cargos in Aplysia ganglia. The scheme represents the workflow for pulling down the ApKHC1 cargo complex for proteomics. ApKHC1 complexes were immunoprecipitated using an anti-ApKHC1 antibody. B, C, Siphon withdrawal response on STS training (B) and LTS training (C). Protein complexes were then separated on SDS/PAGE. D, E, IP of ApKHC1 was validated by Western blot analysis (D) and silver staining (E). F, Image of ApKHC1 amino acid sequence showing ;74% coverage by 150 tryptic peptides identified in mass spectrometry analysis. G, ApKHC1 levels were studied after 1 h of STS and LTS by Western blot analysis. H, Fold increase in ApKHC1 protein levels after 1 h of STS and LTS. Relative protein levels are expressed as the mean fold change, with error bars showing SEM; statistical analyses were performed by one-way ANOVA followed by Dunnett's post hoc test (Extended Data Fig. 1-1, tables 1-5, Extended Data Fig. 1-2).
Research Article: New Research proteins in 5% SDS were reduced with 1 ml of 120 mM TCEP [Tris (2-carboxyethyl) phosphine hydrochloride] at 56°C for 20 min, followed by alkylation using 1 ml of 500 mM methyl methanethiosulfonate for 20 min at ambient temperature. Finally, 4 mg of sequencing-grade trypsin was added to the mixture and incubated for 1 h at 47°C. Following this incubation, 40 ml of 50 mM TEAB (tetraethylammonium bromide) was added to the S-Trap, and the peptides were eluted using centrifugation. Elution was repeated once. A third elution using 35 ml of 50% acetonitrile was also performed, and the eluted peptides were dried under a vacuum. The peptides were subsequently resolubilized in 30 ml of 50 mM triethylammonium bicarbonate, pH 8.5, labeled with TMT labels (10-plex) according to the manufacturer instructions (Thermo Fisher Scientific), and pooled. The multiplexing strategy is provided in Extended Data Figure 2-1 (tables 1-6). The pooled, plexed samples were then dried under vacuum, resolubilized in 1% trifluoroacetic acid, and finally desalted using 2 mg capacity ZipTips (Millipore) according to manufacturer instructions. Peptides were then online eluted into a Fusion Tribrid Mass Spectrometer (Thermo Fisher Scientific) from an EASY PepMap RSLC C18 column (2 mm, 100 Å, 75 mm Â 50 cm; Thermo Fisher Scientific), using a gradient of 5-25% solvent B (80:20 acetonitrile/water; 0.1% formic acid) for 180 min, followed by 25-44% solvent B in 60 min, 44-80% solvent B in 0.1 min, a 5 min hold of 80% solvent B, a return to 5% solvent B in 0.1 min, and finally a 10 min hold of solvent B. All flow rates were 250 nl/min delivered using a nEasy-LC1000 nano-liquid chromatography system (Thermo Fisher Scientific). Solvent A consisted of water and 0.1% formic acid. Ions were created at 1.7 kV using an EASY Spray source (Thermo Fisher Scientific) held at 50°C.
A synchronous precursor selection (SPS)-MS3 mass spectrometry method was selected based on the work of Ting et al. (2011). Scans were conducted between 380 and 2000 mass/charge ratio (m/z) at a resolution of 120,000 for MS1 in the Orbitrap mass analyzer at an AGC target of 4E5 and a maximum injection of 50 ms. We then performed collision induced dissociation (CID) in the linear ion trap of peptide monoisotopic ions with a charge 2-8 above an intensity threshold of 5E3, using quadrupole isolation of 0.7 m/z and a CID energy of 35%. The ion trap AGC target was set to 1.0E 4 with a maximum injection time of 50 ms. Dynamic exclusion duration was set at 60 s, and ions were excluded after one time within the 610 ppm mass tolerance window. The top 10 MS2 ions in the ion trap between 400 and 1200 m/z were then chosen for high-energy collisional dissociation at 65% energy. Detection occurred in the Orbitrap at a resolution of 60,000, an AGC target of 1E5, and an injection time of 120 ms (MS3). All scan events occurred within a 3 s specified cycle time.

Proteomic data processing and statistical analysis
Quantitative analysis of the TMT experiments was performed simultaneously with protein identification (ID) using Proteome Discoverer 2.5 software. The precursor and fragment ion mass tolerances were set to 10 ppm and 0.6 Da, respectively. Trypsin enzyme was used with a maximum of two missed cleavages. The following databases were used: NCBI identical protein group for A. californica (txid6500), a common contaminant FASTA, and P01867 (IgG2b) FASTA files. A SEQUEST search was performed, and the Percolator feature of Proteome Discoverer 2.5 was used to set a false discovery rate of 0.01. The impurity correction factors obtained from Thermo Fisher Scientific for each kit were included in the search and quantification. The following settings were used to search the phosphor-enriched data: dynamic modifications of oxidation/115.995 Da (M), and deamidated/10.984 Da (N, Q); and static modifications of TMT6plex/1229.163 Da (N terminus, K) and carbamidomethyl/157.021 Da (C). Only unique 1 Razor peptides were considered for quantification purposes. Coisolation threshold and SPS Mass Matches threshold were set to 50 and 65, respectively. Normalization was achieved using P01867 values in Proteome Discoverer 2.5. The normalized abundance values for master proteins were imported into JMP PRO 15.2.0 (SAS), and the resulting 824 proteins were analyzed by t test for unequal variances per comparison (control vs STS/LTS 1 h/LTS 3 h). To make sure that variations in the coimmunoprecipitations (co-IPs) are not because of different amounts of antibodies we used, we normalized to IgG levels in the co-IPs. We also considered the possibility that different amounts of cargos might be loaded to the same kinesin complex under different conditions. If this is true, we should observe different amounts of proteins (more or less of the same protein, and or different proteins) in the complex. We therefore also normalized to kinesin levels in the IPs.

Quantitative real-time PCR analysis
Quantitative real-time PCR (qRT-PCR) analyses were conducted following our previously stated protocols (Liu et al., 2014;Badal et al., 2019). RNA isolation was conducted using TRIzol after 1 h of STS and LTS training from the entire CNS of the trained animals (n = 2). Quantification of each transcript is normalized to the Aplysia 18S rRNA reference gene following the 2 ÀDDCt method (Livak and Schmittgen, 2001).

Statistical analyses
Statistical analyses were performed in R and Prism 9 (details provided in the Extended data Figs. 1-1, 2-1, 3-1, 4-1, 5-1). Statistical tests performed were unpaired two-tailed Student's t test and one-way ANOVA followed by Dunnett's post hoc test unless indicated otherwise. The results are graphically represented as the mean 6 Standard error of the mean (SEM) throughout the text, unless otherwise stated. N represents the number of independent samples for each experiment.

Isolation of kinesin transport complexes using antikinesin heavy chain antibody and TMT labeling
To determine whether compositions of anterograde transport complexes are altered during learning and used Research Article: New Research in activity-associated remodeling of synapses, we established co-IP of ApKHC1 complexes from Aplysia CNS (Extended Data Fig. 1-2). Following the study by Puthanveettil et al. (2008) and Liu et al. (2014) and using an anti-KHC1 monoclonal antibody, we established the ApKHC1 co-IP and assessed the efficiency of the IP by silver-staining analysis followed by mass spectrometry and WB (Fig. 1A,D,E). Immunoprecipitated beads alone served as the specificity control in the co-IP. Specific protein bands, corresponding to ApKHC1 and present only in the IP, were not detected in the "beads only" control in the WB. We therefore isolated the corresponding protein bands from the polyacrylamide gel for ApKHC1 and analyzed them by mass spectrometry to identify the band in detail. A NCBI database search of the liquid chromatography tandem mass spectrometry (LC-MS/MS) data against the IPG (Identical Protein Group) database for A. californica (txid6500) was conducted. More than 150 tryptic peptides with ;74% coverage were identified as corresponding to ApKHC1, suggesting the robustness of the IP (Fig. 1F, Extended Data Fig. 1-1, tables 1-5).
As mentioned above, and following previously described methods (Pinsker et al., 1973;Frost et al., 1985;Hawkins et al., 1989;Antonov et al., 1999Antonov et al., , 2003, we induced STS in sea slugs by single tail shock, and LTS by four spaced tail shocks. Single-shock training only gives rise to short-term memory and does not induce transcriptional changes in the same way that four tail shocks do (Kandel, 2001), so we decided to compare how anterograde transport is influenced by short-term versus long-term sensitization.
We first examined whether ApKHC1 expression levels are modulated by LTS training. Similar to previously reported changes in ApKHC1 levels induced by 5x5HT in culture (Puthanveettil et al., 2008), we found that LTS training induced a significant increase in ApKHC1 protein levels (fold increase: approximately twofold; n = 3; p , 0.05, one-way ANOVA followed by Dunnett's post hoc test; Fig.  1G,H, Extended Data Fig. 1-1, tables 1-5). Importantly, STS training did not alter ApKHC1 levels.
Short-term and long-term sensitization training results in the enrichment of distinct proteins in the ApKHC1 complex For the comparative quantitative analysis of the enriched proteins following STS and LTS training, we normalized the data with Ig g -2B chain C region (P01867) using an equal amount of anti-ApKHC1 antibody for all IP experiments. Data presented in volcano plots in Figure 2, A and B, indicate the enriched proteins compared with control on STS and LTS, respectively. Following the IPG database of txid6500, we learned that 3.5% of the total A. californica proteome is associated with the ApKHC1 complex.
These changes in ApKHC1 complex composition could have multiple sourcesthey could indicate the formation of more ApKHC1 complexes, they could be a product of alterations within the same complex, they could represent a combination of both or due to their regulated release to the synaptic end. Upon normalization by IgG, no significant differences were noted in KHC levels between LTS and STS. To address this uncertainty, and to confirm whether there were any differences in identified cargos to ApKHC1 ratio on induction of STS and LTS, we normalized the dataset with ApKHC1 abundance in the co-IP. The normalization also showed enrichment of different sets of proteins after STS and LTS, with cofilin protein enriched on both types of training ( Fig. 2D-F, Extended Data Fig. 2-1,  tables 1-6). Furthermore, analysis of global changes suggested that an uncharacterized protein LOC101861750 Figure 2. Graphical illustration of quantitative proteomics data showing differential enrichment of protein sets after STS and LTS (1 h). Proteins are ranked in a volcano plot according to their statistical -log2 p-value (y-axis) and their relative abundance ratio (log2 fold change) between higher enrichment and depletion (x-axis). Red dots indicate significantly regulated proteins (false discovery rate, ,0.01; s0 = 1). A, Volcano plot of control versus STS protein enrichment normalized to IgG2. B, Volcano plot of control versus LTS (1 h) protein enrichment normalized to IgG2. C, Venn diagram comparison of sets of proteins showed higher enrichment in response to STS and LTS (1 h) normalized to IgG2 (p , 0.05). D, Volcano plot of control versus STS protein enrichment normalized to ApKHC1. E, Volcano plot of control versus LTS (1 h) protein enrichment normalized to ApKHC1. F, Venn diagram showing a comparison of sets of proteins showed higher enrichment in response to STS and LTS (1 h) normalized to KHC1 (p , 0.05; Extended Data Fig. 2-1, tables 1-6, Extended Data Fig. 2-2). commonly depleted after behavioral training (Extended Data Fig. 2-2). These results suggest that STS and LTS induce changes both in the number of ApKHC1 complexes and in their cargo composition.
Based on our previous studies (Liu et al., 2014), we observed that the IP efficiency of samples, as well as the stability of specific proteins interacting with kinesin cargos, can influence the abundance of proteins in individual IP biological replicates. Thus, we further analyzed the dataset to identify protein cargos with synaptic importance-those that were 1.3-fold enriched in the ApKHC1 IPs compared with the control IPs-believing this would help us screen candidates for the WB validation, which is otherwise limited because of the reactivity of commercially available antibodies for Aplysia (Extended Data Fig.  3-1, tables 1-4). For further studies, we selected two candidate protein cargos from STS: FKBP2, an endoplasmic reticulum (ER) chaperon, and calumenin, a calcium-binding protein mainly involved in ER functions in protein folding and sorting. Three protein candidates selected from LTS are synaptotagmin-1, dynamin-1, and calmodulin (Extended Data Fig. 3-1, tables 1-4). These candidates were enriched on STS and LTS, respectively, and used for the validation of mass spectrometry data by WB as co-IP in ApKHC1 IPs.
As shown in Figure 3A, WB analyses revealed that FKBP2 and calumenin were significantly enriched as co-IP in ApKHC1 IPs after STS. FKBP2 and calumenin levels showed 1.2-fold ( Fig. 3B; N = 5, p , 0.05) and 1.8-fold ( Fig. 3C; N = 5, p , 0.05) increased enrichment compared with control, respectively. FKBP2 and calumenin levels in LTS IPs showed no significant difference (p 0.05) compared with control IPs in LTS samples. Examination of synaptotagmin-1, dynamin-1, and calmodulin levels in the WB (Fig. 3D), however, found that these synaptic proteins had been significantly enriched in ApKHC1 IPs in response to LTS. Synaptotagmin-1, dynamin-1, and calmodulin levels showed 1.2-fold ( Fig. 3E; N = 5, p , 0.05), 1.7-fold ( Fig. 3F; N = 5, p , 0.01), and 2.1-fold ( Fig. 3G; N = 5, p , 0.001) higher enrichment compared with control IPs (for further details, please see Extended Data Fig.  3-1, tables 1-4, and Extended Data Fig. 3-2). Moreover, no significant differences in enrichment were observed for these three cargo proteins in STS samples compared with  TMT-quantitative mass spectrometry of ApKHC1 complexes after 3 h of LTS establishes temporal changes in the composition of the anterograde cargo complex As described above, different sets of proteins are enriched in the ApKHC1 cargo complex as a response to STS versus LTS. Mechanistically, LTS strengthens existing synaptic connections and leads to the formation of new synapses. One hour after LTS training, we found enrichment of specific proteins functioning in synapse formation. This finding piqued our interest in whether the composition of the ApKHC1 complex could change with additional time elapsed. We wondered whether the kinesin complex transported an entirely different set of proteins to its final destination at the synapses, or whether the final set remained essentially the same. Since the 3-6 h window after LTS is crucial for the facilitation of synaptic connections between the sensory and motor neurons of the siphon-and gill-withdrawal reflex (Ghirardi et al., 1995;, we assessed the ApKHC1 cargos after 3 h of LTS training. To do this, we conducted TMT labeling and LC-MS/ MS analysis of the ApKHC1 complexes that we coimmunoprecipitated 3 h after LTS training, as described above. We normalized the dataset with IgG2 and conducted the quantitative comparison of the enriched proteins with the control. Following similar statistical parameters, differential analysis (Fig. 4A, Extended Data Fig. 4-1, tables 1-3) revealed that angiopoietinrelated protein 7 (ANGPTL7; XP_012938614.1), protein transport protein Sec24A (sec24A; XP_012934788.1), protein transport protein Sec23A (sec23A; XP_012936590.1), gelsolin-like protein 2 (XP_005100380.1), and the uncharacterized protein LOC101861914 (XP_005109882.2) are enriched in the kinesin complexes. We observed that, along with the presence of sec23A and gelsolin-like protein 2, uncharacterized protein LOC101861914 showed ongoing 14.32-fold enrichment after 3 h of LTS (Fig. 4B), maintaining the enrichment that it had initially established during STS. Protein transport protein sec24A and ANGPTL7, enriched after 1 h of LTS, also remained after 3 h-a notable finding given that a comparison of global changes suggested the depletion of a number of proteins in the kinesin complex after 3 h of LTS (Extended Data Fig. 4-2). Normalization of the quantitative dataset with ApKHC1 (Fig. 4C, Extended Data Fig. 4-1, tables 1-3) showed enrichment of different sets of proteins after 1 and 3 h of LTS (Fig. 4D), with only skin secretory protein xP2 isoform enriched in both STS and 3 h after LTS training. In sum, the analysis of global changes, in accordance with the IgG2-normalized dataset, indicated depletion of a number of proteins in the anterograde complex but displayed persistent of others (Extended Data Fig. 4-2).

qRT-PCR analysis of the candidates suggested translational regulation of anterograde cargo loading
Since the proteomics analyses showed differential cargo enrichment in anterograde complex on STS or LTS, we wanted to check whether these proteins were regulated at their transcript level. Thus, we studied the expression level of the candidates found highly enriched in the anterograde complex, namely FK506-binding protein 2 (ApFKBP2; XM_ 005099915.3), calumenin (ApCALU; XM_005110250.3), synaptotagmin-1 (ApSYT1; NM_001204624.1), dynamin-1 (ApDNM1; XM_005106526.3), calmodulin (ApCAM; XM_005095332.3); and ApSEC23A (XM_013081136.2), on induction of STS and LTS. A list of primers used in the study has been provided in Extended Data Figure 5-1  (tables 1, 2). We used Aplysia CCAAT enhancer-binding protein (ApC/EBP), an immediate-early gene, which is essential for the successful consolidation of long-term memory, and plasticity as the positive control (Alberini et al., 1994;. As reported previously by Puthanveettil et al. (2008), an expected rise in ApC/EBP (;93-fold increase over control; p , 0.001) and ApKHC1 (greater than twofold increase over control; p , 0.01) levels were found on LTS, which validates the induction of LTF in the LTS samples (Fig. 5A,B). Relative gene expression values (with p-values) are provided in Extended Data Figure  5-1 (tables 1, 2). Other than ApSEC23A, all analyzed candidates showed no significant alteration in gene expression on STS and LTS (1 h) training (Fig. 5C). This suggests that at that time point, the proteins are being regulated at their translational level or some factor that controls/selects the protein needs to be transported from the cell body based on the synaptic need. In our proteomics experiment, we observed ApSec23A enrichment after 3 h of LTS. This could suggest that the upregulation of ApSec23A gene expression after 1 h of LTS reflected protein enrichment after 3 h in ApKHC1 cargos. Given these observations, combining the proteomics study with the gene expression analysis, we can assume that most of the regulations of cargo loading onto the anterograde complex could take place at protein level, and the cargo is selected based on the designated activity.

Discussion
LTM formation is an intricately complex process. It is represented by the seamless orchestration of signal transduction, transcriptional regulation, and protein synthesis, resulting in a structural change to the neural continued spectrometry. E, F, Evaluation of Sec23A enrichment in STS or LTS (1 h) ApKHC1 co-IP by WB. G, H, Sec23A enrichment after 3 h of LTS was confirmed by WB. I-M, Calumenin, synaptotagmin-1, dynamin-1, and calmodulin enrichments were also evaluated in ApKHC1 complexes from LTS (3 h) animals. Relative protein levels are expressed as the mean fold change, with error bars showing SEM; statistical analyses were performed by one-way ANOVA followed by Dunnett's post hoc test and two-tailed paired t test (Extended Data Figs. 4-1, 4-2). circuitry (Kandel, 2001;Bliss et al., 2003). Physiologic regulation of soma-to-synapse communication mediated by KIFs serves as a critical component of this synaptic and structural plasticity (Puthanveettil et al., 2008;Swarnkar et al., 2021). Using the well characterized siphon withdrawal reflex of Aplysia, which facilitates robust behavioral plasticity and nonassociative learning-induced LTM in regard to both short-term and long-term sensitization (Pinsker et al., 1970(Pinsker et al., , 1973Carew et al., 1971Carew et al., , 1981Kandel, 2001), we assessed whether KIFs mediate the transport of specific proteins during learning. By quantitative proteomic analyses in this study, we explored the role of kinesins further by discovering that different sets of proteins are enriched in the anterograde cargo complex ApKHC1 in response to learning.
Our training protocol for inducing LTS involved four bouts of electric shocks separated by 20 min to the posterior of the body wall (Pinsker et al., 1970(Pinsker et al., , 1973Carew et al., 1971Carew et al., , 1981Kandel, 2001), taking a total of 2 h for stimulation. We then isolated the CNS 1 or 3 h after training for ApKHC1 complex immunoprecipitation. However, an intermediate form of sensitization [intermediate-term sensitization (ITS)], which lasts 1-3 h after training, could be produced by our stimulation as well. ITS could be induced by spaced tail shocks in 15 min intervals. ITS lasts ;90 min post-training and decays completely after 3 h (Ghirardi et al., 1995;Mauelshagen et al., 1996;Sutton and Carew, 2000;Sutton et al., 2001Sutton et al., , 2002Sutton et al., , 2004. In our sensitization experiments, gene expression analysis showed a significant increase in ApC/EBP (an immediate early gene) levels, in the LTS (1 h) group. Since intermediate forms of memory require only new protein synthesis, but not gene expression (Sutton et al., 2001), our LTS (1 h) data, which are from 3 h after the first tail shock, reflect an early phase of LTM (overlapping with ITS), whereas the LTS (3 h) data represent specific changes related to LTM.
We tried to carry out Gene Ontology analysis with the identified protein IDs associated with kinesin cargos, however, because of the lack of well defined available functional GO term annotations of Aplysia genome, we could not run the analysis for biological processes, molecular functions, and cellular components. Based on the STRINGdb 11.5 (string-db.org) and using the ShinyGO 0.76.2 online tool, we did network analysis of the protein domains (Pfam; Extended Data Fig. 1-2). RNA recognition motif, Ras family, and EF-hand domains are some of the nodes observed that are directly associated with neuronal functions as described in the previous paragraph.

Activity-dependent alterations in ApKHC1 cargos after STS and LTS training
The quantitative comparison of ApKHC1-transported cargos, analyzed from STS-trained and LTS-trained Aplysia showed activity-dependent differential enrichment and depletion of proteins in the anterograde cargo. For the co-IP, we used an equal amount of anti-ApKHC1 antibody, so we normalized the dataset to IgG2 to equilibrate any unwanted (nonbiological) variations among the samples and to achieve an accurate statistical assessment.
Among the proteins enriched on STS, we identified those with a fundamental role in performing a basal cellular function, such as Hsc70, DDX3Y in neural development (Vakilian et al., 2015); ERH in pyrimidine metabolism, DNA replication, and cell proliferation (Fujimura et al., 2012); and FKBP2, which processes peptidyl-prolyl cis/trans isomerase (PPIase) domains (Tong and Jiang, 2015), a subunit of the V1 complex of vacuolar(H1)-ATPase (V-ATPase)-ATP6V1C1. We also found the following RNA-binding proteins: hnRNP1, which participates in mRNA transport and protein translation (Clarke et al., 2021) as well as CELF1 functioning in RNA processing (Gallo and Spickett, 2010;Ladd, 2013). On the other hand, after 1 h of LTS, enrichment of synaptotagmin-1 along with dynamin-1 and other Ca 21 -sensing proteins, NECAB1 and calmodulin, suggests that, as a result of LTS, kinesin is transporting proteins participating in synapse strengthening and memory formation to the neuronal ends. Synaptotagmin-1, a presynaptic vesicle protein interacts with neurexin and SNAP-25, and is essential as a Ca 21 sensor for fast and synchronous synaptic vesicle fusion (Monje et al., 2012;Chang et al., 2018), whereas dynamin-1 is critically required for synaptic vesicle endocytosis (Raimondi et al., 2011;Fa et al., 2014). Calmodulin is required for the activation of CAMKII (Mayford et al., 1996;Murakoshi et al., 2017). In Aplysia, synaptotagmin-1 has been reported to have an inhibitory effect on synaptic vesicle release (Martin et al., 1995). Nonetheless, our findings are in agreement with those of the study by Monje et al. (2012), in which enhanced levels of synaptotagmin-1 are present in Aplysia cerebral ganglia after 24 and 48 h of serotonin treatment. Intriguingly, we could find only some of the components of synaptic vesicle proteins. While cargos of kinesins consists or multiprotein complexes, organelles, and vesicles, the precise mechanisms by which they transport these cargos remains to be understood in detail. One would reasonably expect to identify all the components of these cargos, probably because of the co-IP conditions, which include the use of mild detergents, we identified only some of their components.
Enrichment of F-actin barbed-end-capping and severing protein gelsolin-like protein 2 (Burtnick et al., 1997) after 3 h of LTS suggests that after establishing new synaptic connections, actin filaments at the end of the neurites of the newly formed synaptic ends might be stabilized for the maintenance of newly formed memory . Sensitization also showed depletion of the components of multiprotein complexes such as ribosomal complex among the protein cargos following both STS and LTS. STS led to 21.4%, LTS (1 h) led to 42.85%, and LTS (3 h) led to 40% depletion of the ribosomal proteins of the total depleted proteins (according to the p , 0.05, IgG2-normalized data). Reduced association of these proteins with ApKHC1 in response to STS/LTS training further confirms that sensitization training alters the composition of protein cargos transported by ApKHC1, including those that are the integral components of protein translation and degradation machineries. Furthermore, the proteins that show decreased association with ApKHC1 on training could be regulated in different ways, for example by local translation or transport by other motor proteins. Together, these results further confirm that kinesin-mediated transport and its cargos are targets of regulation during learning and LTM.
A conserved domain search of LOC101861914, which is noted to be highly enriched on STS and 3 h after LTS, has shown that the uncharacterized protein contains EF hand motifs, with helix-loop-helix structural domain predominantly observed in Ca 21 -binding proteins. Using an online tool at www.swissmodel.expasy.org, structural prediction of the amino acid sequence showed homology with other Ca 21 -interacting proteins calcium binding domain of CpCDPK3, calexcitin, CDPK1, and calcyphosin. Crystal structure and biochemical characterization of recombinant human calcyphosin delineates a novel EF hand-containing protein family that shows 21.43% sequence identity (Extended Data Fig. 1-2) with uncharacterized protein LOC101861914. Thus, the current study uncovers a putative calcium-binding protein that is significantly enriched in the anterograde complex on sensitization; however, detailed characterization is required to discover its role during synaptic plasticity.
We also observed enrichment of protein transport proteins sec23A/sec24A after 3 h of LTS. Sec23/24 heterodimer is an integral component of the COPII coat complex that forms on the endoplasmic reticulum and transports cargo from the endoplasmic reticulum to Golgi (Bi et al., Figure 6. In response to LTS, ApKHC1 gets upregulated and increases anterograde transport, but the ApKHC1 level remains unaltered in response to STS. Results from the current investigation showed that on STS training, protein enrichment into each ApKHC1 complex increased significantly compared with LTS. However, in response to LTS, selective enrichment of proteins occurs within the same complex along with the increase in the number of ApKHC1 complex formations. 2002; Lord et al., 2013). Furthermore, coordinated local protein synthesis and degradation are essential for the maintenance of homeostatic plasticity (Cajigas et al., 2010), and we assume that sec23/24 protein delivery plays a key role in this plasticity by transporting vesicles from ER to Golgi after 3 h of LTS in newly formed, stabilized neural ends. Since the function of gelsolin-like proteins and sec23/24 have not yet been studied extensively in learning and memory, these results serve as a platform for further exploration. We assume that after LTSbased on the changing temporal needs of neurons in the ApKHC1 complex, and after the establishment and remodeling of new and existing synaptic connectionsthe system might be moving toward homeostasis. Thus, these findings suggest that protein loading in kinesin cargos happens in an activity-dependent manner and that there might be a motor-cargo code determining the specificity of anterograde transport of a distinct cargo.
We also found that after LTS training, unlike STS, there was a significant increase in ApKHC1 protein expression. Based on our qRT-PCR analysis, we observed that cargos loaded onto ApKCH1 are regulated mostly at the protein/translational level. Therefore, the question is then raised as to whether the changes in ApKHC1 complex composition could be because of the formation of more ApKHC1 complexes, the loading of more proteins into the same complex, or simply the combined effects of both. Therefore, we further analyzed the dataset after normalizing it with ApKHC1 to check whether more or fewer cargos were being loaded onto the ApKHC1 complex. This analysis showed that following STS, a significantly higher number of proteins was enriched in the ApKHC1 complex compared with 1 h after LTS. Since STS did not alter ApKHC1 levels, these results suggest an increase in the number of enriched proteins in the ApKHC1 complex during STS, whereas LTS might involve both an increase in the number of ApKHC1 complexes as well as the specific enrichment of proteins within the same complex.
Our studies establish that both the composition and number of ApKHC1 complexes are targets of regulation during learning (Fig. 6). These dynamic changes in ApKHC1 complexes facilitate soma-to-synapse communication. They also reflect adjustments in the synaptic proteome that facilitate structural changes and synaptic plasticity during learning. Identification and characterization of circuit-specific KIF complexes will help bring deeper insights into the regulation of the communication between soma and synapse during LTM.