Structural Insight into Anaphase Promoting Complex 3 Structure and Docking with a Natural Inhibitory Compound.

BACKGROUND: Anaphase promoting complex (APC) is the biggest Cullin-RING E3 ligase and is very important in cell cycle control; many anti-cancer agents target this. APC controls the onset of chromosome separation and mitotic exit through securin and cyclin B degradation, respectively. Its APC3 subunit identifies the APC activators-Cdh1 and Cdc20.
MATERIALS AND METHODS: The structural model of the APC3 subunit of APC was developed by means of computational techniques; the binding of a natural inhibitory compound to APC3 was also investigated.
RESULTS: It was found that APC3 structure consists of numerous helices organized in anti-parallel and the overall model is superhelical of tetratrico-peptide repeat (TPR) domains. Furthermore, binding pocket of the natural inhibitory compound as APC3 inhibitor was shown.
CONCLUSION: The findings are beneficial to understand the mechanism of the APC activation and design inhibitory compounds.

Introduction

Many proteins regulate the cell cycle and progression through the cell cycle is dependent on their sequential degradation,[1] which is carried out by the anaphase promoting complex (APC) and the SKP1-CUL1-F-boxprotein (SCF).[2] APC contains two active shapes: APCCdc20, which is active over the M-step of the cell cycle and APCCdh1, which controls mitotic exit and the G1 step.[2] APC controls the metaphase-anaphase transition, mitotic exit, and G1 progression through mediating, respectively, securin and cyclin degradation.[3] So as to develop the metaphase-anaphase transition, APCCdc20 adds various ubiquitin molecules to securin in the M phase. Separase is activated after securin degradation, resulting in cohesion cleavage, chromosome segregation and, ultimately, cell entrances into the anaphase.[4] At the final phase of the mitotic exit, cytokinesis, APC acts as a regulating agent by means of targeting aurora kinase.[5] Moreover, the cell cycle is correlated with other pathways by pathways by APC, e.g., the chromosome segregation,[6] transcription,[7] Deoxyribose Nucleic Acid (DNA) replication,[8] transforming growth factor beta (TGF-β) signaling, and glycolysis pathways[9] to ensure cell growth and division at the right time. APC also controls the mitochondrial function through selective protein degradation to prepare the energy required over the cell cycle progression.[10]

An enzymatic cascade including the ubiquitin-activating enzyme (E1), ubiquitin conjugating enzymes (E2s) and E3 ligase enzymes happen in the ubiquitination reaction.[11] First, the ubiquitin is attached to the last residue G76, which is a thioester linkage with the cysteine residue in the active site of ubiquitin-activating enzyme (E1). Next, it is transferred from E1 to the cysteine residue in the active site of the E2, needing hydrolysis of ATP. Under physiological situations, UbcH10 and UbcH5 act as E2 enzymes for APC,[1213] whereas the E3 ligase acts as the scaffold for the ubiquitination reaction. The E3 ligase is bound to the E2 and the substrate simultaneously and the transfer of the ubiquitin from UbcH10 to a lysine residue on the substrate[14] is done easier. Finally, the poly-ubiquitinated substrates degradation happens by the 26S proteasome.[15]

The activation of APC happens through binding by two co-activators: Cdh1 and Cdc20, and phosphorylation on various subunits.[16] A vast spectrum of substrates via their destruction box and KEN (Lys-Glu-Asn) motif can be recognized by the co-activators.[17] APC is inhibited by the early mitotic inhibitor-1 (Emi1), mitotic checkpoint complexes (including Mad2, Bub3, BubR1/Mad3, and Cdc20) and RASSF1A in the S, G2, and prometaphase steps.[1819]

APC, which has 13 subunits, is the most complex representative of the RING/cullin family of multi-subunit E3 ligases.[20] Its subunits are located in three sub-complexes: I. the catalytic sub-complex (Doc1/APC10, APC11, and APC2), II. A structural or scaffold sub-complex (APC1, APC4, and APC5), and III. A tetratrico-peptide repeat (TPR) sub-complex (APC3/6/7/8/13).[21] The function of tetratrico-peptide repeat (TPR) subunits is binding to the co-activators (Cdh1 and Cdc20) and enhancing the self-assembly of the complex.[22] The scaffold sub-complex is comprised of the largest subunits like APC1 (1,944 residues), and prepares the platform for substrate recognition and reaction catalysis through the other sub-complexes.[23]

The first time TRP motif was discovered in APC subunits.[2024252627] This motif is also found in the N-terminal region of BubR1,[28] Hsp90,[29] the nicotinamide adenine dinucleotide phosphate (NADPH) oxidase subunit p67phox,[30] Hsp90-binding immunophilins,[31] and transcription factors.[32] Each TPR motif has 34 degenerative residues which make a helix-turn-helix structure.[33] TPRs are placed in a parallel arrangement. APC3 containing 822 residues is a critical subunit of APC for binding to Cdh1 or Cdc20 and play a role as linker between catalytic sub-complex and TPR subunits in TRP sub-complex of APC.[434] It was shown that APC3 is bound to Microcephalin (MCPH1), which is causative for main recessive autosomal microcephaly.[35]

Despite the fact that there is much information regarding APC function in different conditions and cell cycle phases, the structure of APC is still unknown. This is, in part, due to the fact that there are many technical problems in the expression of complicated protein complexes such as APC. In addition, such protein complexes still are assembled and crystallized difficultly. Thus, in this study, a 3D structural model of APC3 was developed and the binding of tosyl-L-arginine methyl ester (TAME), as an inhibitory compound with APC3 subunit, was investigated using full flexible docking algorithms to elucidate the TAME binding pocket on the APC3. TAME, known as APC inhibitor, inhibits APC function via competition with Inverted Repeat motif of Cdc20 and Cdh1, APC co-activators. The effective concentration of TAME for APC inhibition in vitro is 12 µM.[36] It is recommended that tosyl, arginine groups of TAME are involved in receptor binding. The results prepare a structural viewpoint of the model of APC3. The binding pocket, which is crucial to design new inhibitory compounds, was identified.

Materials and Methods

Multiple sequence alignment and homology modeling

In order to find homologous sequences, the UniProt database was applied.[37] After the fragmented and unrelated sequences were removed, repossessed sequences were introduced to ClustalW[38] for multiple sequence alignment using Clustal algorithm, BLOSUM62 scoring matrix and default parameter settings. The JalView software[39] was used to visualize the results of the alignment. Secondary structure patterns were projected through Jpred.[40]

So as to perform homology modeling, the MODELLER 9.10 package[41] was used. Sequence-structure alignment against Protein Data Bank (PDB) database was conducted by means of HHpred toolkit[42] to find a homologous structure as a template for homology modeling (HM). For APC3 modeling, 10,000 models were produced and the best model was chosen in accordance with Discrete Optimized Protein Energy (DOPE) score.

Molecular dynamics

Energy minimization was carried out for the models of the APC3 subunit by using molecular dynamic (MD) simulations in Gromacs 4.5 software[43] with Amber99SB-ILDN force field[44] and TIP3P explicit waters.[45] 0.15 M Na+ and Cl- were added to the water box to neutralize the system for simulation of the physiological environment. In order to provide the system, energy minimization was initiated by using a steepest-descent algorithm with a tolerance of 10 kJ/mol/nm and a 2 fs step size. Hydrogen atoms were allowed to be relaxed with fixed heavy atoms for 50 ps with a step size of 2 fs during the position-restraining step. All bonds were restrained through the linear constraint solver (LINCS) algorithm.[46] There were two steps in system equilibration: Firstly, the system was heated for 100 ps by means of V-rescale algorithm for temperature stabilization (T = 300 K with temperature coupling τP = 0.1 ps).[47] Next, the pressure coupling was run for 50 ps using the Parrinello-Rahman method in order to stabilize pressure (P = 1, τP = 0.1 ps).[48] For the calculation of electrostatics and van der Waals (VDW) interactions with grid spacing of 0.16 nm and 1.4 nm as cut off, the smooth Particle Mesh Ewald method was employed.[49] Finally, the MD simulation was done with a time step of 2.0 fs for 10 ns.

Analysis of the molecular dynamic simulations

To analyse and visualize the result of the APC3 simulation, Gromacs 4.5 and PyMol softwares were used. For each trajectory, energy analysis including temperature, pressure, potential energy, and kinetic energy was conducted. Root mean square deviation (RMSD) between the starting structure and the average structure was applied to study the structural convergence. Protein stability and flexibility was controlled by means of root mean square fluctuation (RMSF) and secondary structure analysis. The average structure was extracted from trajectory between 3,000 and 10,000 frames (steps) and then energy was minimized by steepest-descent and the conjugate-gradients method. The structure validation was performed by torsion angles analysis of the protein backbone (phi and psi) through PROCHECK software,[50] Z-scor, Verify3D, prosaII score, Molprobity.[51]

Molecular docking

Ligand and receptor preparation involve add Hydrogen, charge and miss-atom was carried out using ADT4.2 toolkit.[52] Interacting residue diagram was drawn using LigPlot + software.[53]

Results

Architecture of the anaphase promoting complex 3

In APC3 sequence, eight TPR domains were found in the 115-148 and C-terminal regions (499-770) using InterProScan and Prosite [Figure 1]. Conservation in C-terminal region (580-822) was shown in multiple sequence alignment of APC3 from different organisms [Figure 2].

Figure 1

APC3 domain distribution. TPR domains are located in N-terminal and C-terminal regions. Although just one repeat was seen in N-terminal region, other repeats are arranged in tandem in C-terminal region

Figure 2

Multiple sequence alignment of C-terminal region APC3. Sequence alignment shows conservation in C-terminal region. Secondary structure prediction was performed by JPred software. Most of regions were predicted as α-helix region. Alignment includes APC3 sequence from the following organisms: sp | P30260 | CDC27_HUMAN, sp | A7Z061 | CDC27_BOVIN, sp | Q4V8A2 | CDC27_RAT, sp | A2A6Q5 | CDC27_MOUSE, sp | P10505 | APC3_SCHPO, tr | B6JZC7 | B6JZC7_SCHJY, tr | C1E0V5 | C1E0V5_MICSR, sp | Q8LGU6 | CD27B_ARATH, tr | Q017V9 | Q017V9_OSTTA, tr | F8KQM7 | F8KQM7_HELBC, tr | C1DPD6 | C1DPD6_AZOVD and sp | P38042 | CDC27_YEAST

Anaphase-promoting complex subunit CUT9 (PDB ID: 2XPI) was selected as the template for APC3 structure prediction based on the structure-sequence alignment. This protein is related to Schizosaccharomyces pombe and contains TPR domains. Because of the high resolution (2.06 A°) and the presence of TPR domain, the CUT9 is a suitable template for HM. Homology was shown from the middle to end of APC3 (260-822) in the structure-sequence alignment. To select the best model based on DOPE score, 10,000 models were generated by means of MODELLER. The best structural model contains many α-helices localized in anti-parallel [Figure 3a]. Each TPR domain contains two α-helices (A and B) that are connected to each other via a turn region [Figure 3b]. Adjacent TPR domains were bound via a long loop. Each loop has internal curve that accelerates the interaction among close TPR domains. In the N-terminal, six TPR domains form a circle around the central axis. The outside of the circle is made up of β helix from each TPR domain and α helix is localized in the internal side [Figure 3b]. Other TPR domains are placed in the second circle. These circles are bound to each other via the seventh TPR domain tending toward the C-terminal region. The α-helix of the last TPR domain has an unstructured region. From the top view, the overall structure is superhelix, similar to the regular helices that were seen in a solenoid shape [Figure 3b]. MD simulation was performed for 10 nano-seconds to reveal the structural stability of structural model. RMSF plot was shown that the structural model is stable during the MD simulation [Figure S1] except for some residues in C-terminal. It was found that all of residues are in allowed region as shown in the Ramachandran plot [Figure 4b]. Moreover, superposition of the average and B-factor structures [Figure 4a] indicates high stability in the all regions, apart from the last TPR domain.

Figure 3

Three dimensional structure of human APC3. (a and b) The front view of APC3 structure that contain TPR domains. (c and d) The top view of APC3, which is similar to super-helix. Purple color represents helix and cyan color shows loop region. All panels were prepared by PyMOL software (www.PyMOL.org)

Figure 4

APC3 structure validation. (a) RMSF (root mean square fluctuation) showed that the APC3 structure model is stable during MD, only the last TPR domain has more flexibility. (b) Superposition of the average and B-factor structures, red color showed more flexible region and blue color represent the stable region. Here, only the last TPR domain has conformational changes. (c) In Ramachandran plot, 91% residues were in most favored regions and the other residues (8.9%) put in allowed additional region

RMSF (root mean square fluctuation) showed that the APC3 structure model is stable during MD, only the last TPR domain has more flexibility

Click here for additional data file.

In this study, molecular docking was used to investigate the binding of TAME to APC3. ADT4.2 toolkit and vina-autodock were applied to dock; the LigPlot software was used for visualization of the binding pocket. It was found that TAME was bound to APC3 with high affinity (−6.2 kcal/mol) [Figure 5a–c]. The docking environment was set up to cluster the different structural conformations into nine classes. As can be seen in Table 1, binding affinity of the different classes is very near each other (from −5.4 to −6.2). Furthermore, it was found that all classes were bound to a same binding pocket [Figure 5d], except for class 7, which had much conformational change (RMSD = 20.672) and was bound to another lactation on APC3 [Figure 5d]. Representation of the LigPlot result was shown that binding pocket was comprised of 12 residues; four of which (Val179, Tyr148, Glu291, and Lys247) involve in the hydrogen bond. Eight non-hydrogen (van der waals) bonds were also seen in the binding pocket [Figure 6].

Figure 5

TAME binding to APC3. (a) Two dimentional representation of TAME. (b and c) Binding pocket of TAME, which is located in TPR repeats. (d) Orentation of different conformational clusters of TAME showed that all of them are localized in same binding pocket expet for the seventh cluster

Table 1

Binding affanity of diffrent conformational structures of TAME to APC3

Figure 6

TAME binding pocket. This binding pocket is composed of 12 residues in which four of them involve in hydrogen bond (Val179, Tyr148, Glu291, and Lys247)

Discussion

Anaphase promoting complex 3 structure

APC, is a prominent member of the cullin-RING E3 ligase family, includes both cullin and RING domains in the catalytic subunits.[54] TPR domains are seen in proteins with different functions such as synaptosomal-associated protein (SNAP) secretary proteins, N-terminal region of BubR1,[28] Hsp90,[29] APC subunits Cdc16, Cdc23, and Cdc27,[2024252627] the NADPH oxidase subunit p67phox,[30] Hsp90-binding immunophilins,[31] and transcription factors.[32] The TPR domain plays a role in complex assembly,[55] cytoplasmic accumulation[56] and it is considered as ligand (peptide) in domain binding[57]; a number of TPR domains are important in protein-protein interaction.[58] In APC3, the TPR domains were bound to Cdh1 and accelerate APC activation.[34] In this work, the structural model of APC3 was developed and the binding of the natural inhibitory compound (TAME) to APC3 was studied. It was found that APC3 had eight TPR domains in the central and C-terminal region. The overall structure of APC3 is solenoid shaped [Figure 3]. Structural analysis of Protein phosphatase 5 (PP5) containing TPR domains revealed that the TPR motif is a pair of anti-parallel α-helices associated together with a packing angle of ~ 24° between helix axes.[59] APC3 has a structural conformation change in C-terminal [Figures 4a and b], which helps to Cdh1 recognition because this structure is very compact, and provides enough space to interact with Cdh1. Such conformational change is found in other Cdh1 recognition proteins such as APC10,[606162] which makes an activator recognition site in association with APC3.

Tosyl-L-arginine methyl ester molecular docking

TAME is a natural small molecule extracted from Xenopus extract. Previous studies showed that TAME is bound to APC3 and inhibits APC complex activity.[36] Many of researchers reported that C-terminal region of APC3 (499-824) has a role in binding to co-activators and APC inhibitors (TAME and pseudosubstrate).[3663] In the present study, it was found that TAME is bound to APC3 in high affinity; the binding pocket is located in TPR domains. This binding may result in arresting in domain movement which inhibits APC function. This hypothesis is confirmed by a study on binding of peptide to the TPR domai[63] that showed binding of peptide to TPR domain increased structure rigidity. Therefore, the TPR domain couldn’t associate with other proteins. Moreover, TAME was bound to the TPR domain in the C-terminal region, where play a role in binding to Cdh1. Based on docking pattern, similarity search methods were used in small molecules database to find molecules with high affinity and low toxicity.

Financial support and sponsorship

This work was funded by PhD student grant from Pasteur Institute of Iran.

Conflicts of interest

There are no conflicts of interest.