Structural exploration of acid sphingomyelinase at different physiological pH through molecular dynamics and docking studies

Abstract

Acid sphingomyelinase (ASM) is an enzyme involved in the hydrolysis of sphingomyelin leading to the production of ceramide. It is responsible for the stimulation of apoptosis and has a pH dependent efficacy for its substrate binding. Deficiency of this enzyme causes an autosomal recessive lysosomal storage disorder called Niemann–Pick disease. To understand the full length structure of ASM at different physiological pH, the protein was modeled based on its structure homologue purple acid phosphatase (PAP) and topological folding matching through ab initio calculations. Results from molecular dynamics simulations at different physiological pH, PCA and FEL analysis clearly depict the folding mechanism of the loop region (D206–L248) towards the catalytic domain. The ASM structure at neutral pH and pH 3.0 adopts a conformational fold that subsequently either closes or partially opens the catalytic domain making it inaccessible for the binding of substrate, whereas the ASM structure at pH 5.0 maintains a completely open conformation of the loop that creates a tunnel in the catalytic domain for the accessibility of substrate. This observation was also confirmed through molecular docking studies, showing that sphingomyelin binds at a different site which is away from the catalytic domain of neutral ASM. Though the substrate binds near the catalytic domain in the pH 3.0 ASM structure, it is unable to form a proper mode of interaction which is essential for its hydrolysis process. Only in the pH 5.0 ASM structure, a proper mode of interaction formulated by the phosphate ion that is posing towards the two catalytic zinc ions confirmed their ability to effectively engage the substrate for its hydrolysis. The structural insights of ASM at different physiological pH and subsequent substrate interactions provide an indispensable understanding of their catalytic mechanism which can lead to the enhancement of apoptosis and control of Niemann–Pick disease.

---

Introduction

Acid sphingomyelinase (ASM) is an enzyme (E.C. 3.1.4.12) that has a strategic task in sphingolipid metabolism through hydrolysis of sphingomyelin (SM) into ceramide at an optimum pH of 4.5 to 5.5 in the primary localization of the endo-lysosomal compartment.^1,2 This bioactive ceramide influences a variety of cellular processes such as apoptosis, senescence, differentiation, migration, proliferation and inflammation.³ The enzyme is abundantly present in lysosomes, particularly in monocytic, hepatic and endothelial cells, scarcely in mitochondria and it can also be secreted into the extracellular space.⁴ L-ASM regulates the lysosomal deprivation of SM, a main lipid component of the outer leaflet of eukaryotic membranes. S-ASM produces LDL-ceramide from LDL-sphingomyelin and this LDL-ceramide takes advantage of sub endothelial LDL aggregation and retention, and consequently forms cell formation.⁵ Upon stimulation effect of pathogens, ASM is believed to be relocalized to cellular membranes.⁶ The gene SMPD1 generates two different ASM enzymes, lysosomal sphingomyelinase (L-ASM) and secretory sphingomyelinase (S-ASM), via differential trafficking of a common protein precursor.^4,7 Though both are zinc metalloenzymes, their association with zinc differs. Accordingly, the L-ASM obtain zinc ions during the process of trafficking to lysosomes as a result it attains the cation independent nature whereas S-ASM is shown to behave either fully or partially dependent of Zn²⁺ for enzymatic activity. The environmental concentration of zinc ion and the zinc chelator EDTA does not affect the activity of L-ASM enzyme which defines a long-standing body in the literature that it is a cationic-independent nature.^4,7–9 The recent study shows that the purified L-ASM from human placenta has vital requirement of acid pH for its proper functionality of substrate binding (K_m) and not for its catalytic velocity.^1,6 Enzymatic dysfunction of L-ASM leads to an inherited autosomal recessive lysosomal storage disorder called Niemann–Pick disease.¹⁰ Though many enzymes can generate ceramide, ASM gets appreciation for the unique feature as it hydrolyzes sphingomyelin in three different cellular location viz., the endo-lysosomal compartment, the outer leaflet of the plasma membrane and lipoproteins,⁶ thereby increasing the availability of ceramide.

In the past three decades, a group of enzymes sharing similar activities have been reported viz., acidic, neutral, and alkaline sphingomyelinases which differ mostly in their tissue distribution and pH optimum.¹¹ The L-ASM is one of the spliced variants of ASM which initially produced as pre-pro form (75 kDa) and enters into ER through co-translational process where it becomes pro form (72 kDa). The conversion of pro into mature form of ASM (70 kDa) takes place only in the acidic environment which indirectly illustrates their localization at the lysosomal compartment. Experiments with brefeldin A, which particularly blocks the vesicular transport from ER to Golgi revealed that acidic environment is necessary for the maturation of L-ASM (70 kDa) from pre-pro form of 75 kDa, but had no effect on S-ASM (57 kDa). Studies based on site directed mutagenesis and protein expression reported that the size of 75 kDa and 72 kDa L-ASM reduced as expected when the two glycosylation sites are eliminated, whereas the size of S-ASM 57 kDa remains unaffected.¹²

The L-ASM protein includes a signal peptide, a predicted saposin domain, a phosphoesterase domain with predicted di-metal center and five metal ion binding sequence motifs,¹³ and a carboxy-terminal region translated from ASM gene.^13–17 The enzymes exonucleases, Ser/Thr phosphatases, 5′ nucleotidase come under phosphoesterase family.¹⁸ Kidney bean purple acid phosphatase (PAP) is one of the best-described members of this family.¹⁷ There are many variations between the catalytic mechanism of ASM and other di-metal phosphoesterases (DMP). For example, a large number of DMPs hydrolyze phosphomonoester substrates whereas ASM hydrolyzes the phosphodiester substrate named SM. Another difference is the presence of dissimilar metal ions in the di-metal center of DMPs, namely Fe³⁺, Zn²⁺, or Mn²⁺ ions, whereas the di-metal center of ASM contains only two Zn²⁺ and so named as a zinc-dependent enzyme.⁷ Across the species, the sequence motif “NX3CX3N” of ASM is highly conserved and this motif stretch present in a loop is believed to be involved in substrate recognition.¹⁹

Since there are different types of sphingomyelinase at different physiological pH based on their cellular locations, understanding the structural details responsible for this difference is indispensable for unraveling substrate metabolism. Accordingly, the present study focuses on the structural details of acid sphingomyelinase. We have explored in detail the full length structure of L-ASM, and its structural transition upon influence of different physiological pH using molecular modeling and dynamics simulation studies.

Computational methods and materials

Tertiary structure prediction of acid sphingomyelinase (ASM)

In order to understand the structural stability and transition of full length L-ASM protein (625aa, UniProt Accession Number: Q0VD19) (http://www.uniprot.org), the three dimensional structure was predicted using Robetta server.^20,21 The unavailability of homologous templates for the N and C-terminal region made us to model the protein using Robetta server. Here, the N-terminal region (1–184) which includes signal peptide (1–41), saposin-B-type domain (81–165) and proline rich region (175–193) is found to be homologous with oxidoreductase (3PQA) for the region of 1–80 and lipid binding protein (2DOB, region 81–184).²² The phosphodiesterase domain and C-terminal region are found to be homologous with hydrolase (2XMO, region 185–524)²³ and transcription protein (2IS9, region 525–625).²⁴ The model generated from Robetta server was further remodeled using Modeler 9v8 program²⁵ (PDB ID: 1X9O) in order to fix the two zinc metals which are proven to essential for the coordination interaction with active catalytic residues such as D202, H204, D274, N314, H421, H453 and H455. Finally, the best model was selected based on the least Modeler objective function (MOF) and discrete optimized protein energy (DOPE)²⁶ score. Further, the selected model was verified by the Structural Analysis and Verification Server (SAVES) which includes PROCHECK²⁷ and ERRAT.²⁸

Molecular dynamics simulation (MDS) of ASM with different pH

The modeled ASM was prepared for MD simulation studies at three different pH conditions to check the stability nature using GROMACS 5.0.6 molecular dynamics package.^29,30 The functionality of L-ASM has been proven to be at an optimum pH of 5.0 and hence ±2 unit (neutral and pH 3.0) was used for understanding its structural stability and consequent functionality through comparative studies. First, the topology for pH 3.0 and pH 5.0 along with neutral ASM were generated using pdb2gmx command by setting the protonation and de-protonation state of K, R, D, E and H residues which was identified using H⁺⁺ server.^31–34 Here, the proper bond order and geometrical parameters of the proteins were assigned according to GROMOS96 43a1 force field. Further, proteins were solvated by SPC216 water model in the cubic box size of 1.0 nm distance which is calculated from the center of the protein to edge of the box. The direction of prepared systems was maintained by applying periodic boundary conditions. Most importantly, net charge of the system was maintained for the pH 3.0 and pH 5.0 ASM structure after protonation and de-protonation step, whereas the native structure was neutralized by replacing the water molecules with CL⁻ ad NA⁺ counter ions based on their net charge. All three proteins were subjected to a maximum of 50 000 energy minimization steps using steepest descent algorithm with a tolerance of 1000 kJ mol⁻¹ nm⁻¹. The van der Waals and electrostatic long range interactions were applied by using fast Particle-Mesh Ewald electrostatics (PME)³⁵ with 1 nm cut off. The bond angles and geometry of the water molecules were constrained with LINCS³⁶ and SETTLE³⁷ algorithm respectively. In addition, Parrinello–Rahman³⁸ method was used to regulate the pressure, while the modified weak coupling berendsen thermostat and V-rescale algorithm were used to regulate the temperature of the system. In order to retain the coordination bonds formed between Zn²⁺ metals and metal binding residues D202, H204, D274, N314, H421, H453 and H455, the bond distance was set in range between 2.3 and 3.3 Å using distance restraints method in the topology. Both NVT (constant number of particles, volume and temperature) and NPT (constant number of particles, pressure and temperature) were accomplished for 100 ps and checked for their equilibration status. At last, the equilibrated different pH systems were subjected to 50 ns of production MD run with a time frame of 2 fs. Further, the analyses were carried out using the tool which is embedded in the GROMACS package to define the protein dynamics and its structural transition for the entire MD simulation.

T-pad and mode vector analysis

Protein plasticity is essential for molecular recognition and in numerous cellular processes like protein aggregation, metabolism, gene expression and molecular signaling.^39–42 T-pad analysis calculates residue-by-residue plasticity nature of the given protein and deciphers the transition for its biological functionality. It also identifies backbone transitions pertaining to two conformations of Ramachandran plot which helps to explain the structural adaptation of hinge point in protein structure responsible for molecular mechanism. Here, we used T-pad analysis to understand the structural integrity of full length ASM structure at different physiological pH condition. In order to create the protein plasticity nature, each 50 ns trajectory of neutral, pH 3.0 and pH 5.0 was prepared by removing the water molecules, ions and un-equilibrated trajectories by applying the trjconv tool and was subjected to T-pad analysis. Earlier, the plasticity of protein was obtained based on the circular spread Ramachandran angles Φ or ψ (CSΦ or CSψ) which was not able to represent the backbone conformation. Hence, the introduction of Protein Angular Dispersion (PADω) overcomes this complication and the assessment of both CSω and PADω enumerates the protein backbone plasticity clearly. Here, the PADω has two advantageous features over the CSω: (i) the function ω (ω = Φ + ψ) is dependent on both the Ramachandran angles; (ii) it is formulated in the range between 0° and 180°.⁴³ The above mentioned protocol was adopted from our previous study.^44,45

The PCA and FEL analysis

The protein high amplitude motions of simulated trajectories via eigenvectors of the covariance matrix were defined by using PCA and dPCA analysis.⁴⁶ Accordingly, in our study the PCA for different pH ASM enzymes are calculated to define the atomic motions of the protein throughout the entire MD simulation. Further, the cosine content (c_i) of each principal component (p_i) of covariance matrix were calculated to identify whether the MD time interval that is used to extract the concerned trajectory is sufficient or not and to represent the free energy landscape defined by PCA analysis.^47,48 The above mentioned protocol describes an absolute and very sensitive measure of trajectory which subsequently deduces numerous free energy minima that corresponds to conformational mapping of respective energy basins that are constructed based on the free energy landscape of chosen principal components (PCs). The values of cosine content normally ranges between 0 (no cosine) and 1 (perfect cosine) in the total time of simulation period (T):

In general, the first eigenvector cosine contribution defines the nature of protein structural dynamics but in most of the cases the first eigenvectors cosine values shows a value closer to 1. This high cosine content describes the large scale motions in the protein dynamics which is not used for analyzing the behavior of protein in terms of FEL. However, recent studies shows that the cosine contents which is equal to or below 0.2 or sometime up to 0.5 of first two PCs can yield qualitatively related and appealing results with observation of single basin.⁴⁸ Therefore, the first 20 PCs of each different pH ASM protein were extracted and analyzed based on their cosine values. FEL was constructed using cosine contents lesser than 0.2 of the first two projection eigenvectors and were defined as PC1 and PC2 respectively. These ultimately provide the energy minima of the clustered protein which significantly helps to identify best energy conforma in order to understand the structural dynamics and transition of ASM at different physiological pH.

Molecular interaction of ASM with sphingomyelin

The best free energy representative structure of neutral, pH 3.0 and pH 5.0 ASM were retrieved and subjected to protein preparation wizard of Maestro 9.2 software (Schrödinger Software Suite 2012). The hydrogen and bond order were assigned properly to coordinate the protein molecules along with hydrogen bond assignment. In addition, the sample water orientation was fixed and optimized using exhaustive sampling methods and subsequently allowed for energy minimization using OPLS 2005 force field. On the other side, the substrate molecule (Sphingomyelin) was retrieved from PubChem (CID: 45266629) (http://pubchem.ncbi.nlm.nih.gov/search/search.cgi) database and prepared using LigPrep module embedded in the Maestro 9.1 software by applying the same OPLS 2005 force filed. The parameter was set for possible ionization states at pH 7.0 ± 2.0 and was allowed for possible 32 combination generation of tautomers. Finally, receptor grid was generated for all the three free energy representative structures by selecting the H315 and NX₃CX₃N motif denoting residues N377, C381 and N385 as centroid with 22 Å radius to dock the substrate molecule. All possible tautomers molecules generated were allowed for the docking protocol using standard precision method of Maestro 9.1 with default parameter settings.⁴⁹ Further, the best docked conformation was subjected to PDBePISA server for the interaction energy calculation and its stability analysis.^50,51

Results and discussion

Structure of ASM

The full length L-ASM structure contains three distinctive domains, N-terminal (NTD), a phosphoesterase domain (PED) and C-terminal (CTD) domain (Fig. 1). Both NTD and CTD contain only helical segments [NTD: α1–α8 and CTD: α17–α22] placed at an angle and distance of 68.5° and 55.2 Å respectively with the basis of domain axis. The PED is positioned between the NTD and CTD domains with an axis distance of 29.3 and 5.7 Å between PED and NTD, CTD respectively.

Fig. 1 The three dimensional structure of full length ASM contains N-terminal domain (NTD: 1–184), phosphoesterase domain (PED: 185–524) and C-terminal domain (CTD: 525–625). NTD holds two sub domain called signal peptide (1–40) and saposin B type domain (81–165). The zoomed image shows the co-ordinate interaction of two zinc ions with the catalytic cavity of PED domain. The co-ordination interactions are shown in dashed line, color codes: ASM-PED domain interacting residues or catalytic cavity-tane stick hetero atom type and Zn²⁺-magenta.

NTD. The NT domain contains 8 helices and has two sub domains separated by a loop which connects α3 and α4. The first sub domain is enriched with signal peptide residues in the region M1–A40 containing the α1 (R23–G26) and α2 (C29–A42) helices. The α3 (K66–Q73) helix of first sub domain is placed parallel to the α2 helix with a distance and angle of 10.1 Å and 37.5° respectively. Though the helix α2 and α3 are stabilized by means of helicity through maintaining the n, n + 4 hydrogen bond, the series of hydrophobic interactions formed between α2 (L30, L33, W34 and L37) and α3 (I70 and F67) also contribute for their inter helical stability. The second sub domain contains 5 helices among which the first four amphipathic (α4–α7) helices fold like oblate ellipsoid and are identified as saposin fold B type domain as like other human saposin A (PDB ID: 2DOB) and C (PDB ID: 2GTG) domains (Fig. 2). The superimposition and comparison of NTD saposin domain with saposin A and C reveals that it has an RMS deviation of 2.03 and 1.55 Å respectively (Fig. 2A). It was identified that the saposin fold contains two stem (α4: C85–R100 and α7: P149–L157) and one hairpin formed by helices α5 (V105–L119) and α6 (P124–R144).^15,52 The saposin fold is composed of non-conserved charged residues on the surface while the conserved hydrophobic residues are pointed towards the core region. The superimposition of NTD saposin B-type domain with homo sapiens saposin B (PDB ID: 1N69) clearly depicts the difference between the open (saposin B) and closed (NTD saposin B type domain)^53–55 conformation of hairpin fold with respect to (Cα terminal helical residues) α5 and α6 which are displaced by distance of about 34.19 and 32.13 Å respectively (Fig. 2B). Unlike saposin A and C, the crystallographic structure of saposin B showed an open conformation upon binding with membrane leading to their functional activation. The open conformation of saposin B is attained by interacting with membrane and it subsequently helps in the substrate catalysis process by making an enriched cavity.¹⁹ However, the NTD saposin B type domain resembles with the closed conformation of saposin A and C wherein the stem α4 is oriented towards the hairpin α5–α6 and stem α7. Such an orientation is maintained by a series of hydrophobic interaction formed between α4: C88, F92, I95, L99, T87, P124 and A94, α5: L119, L115, C116, A112 and V105, α6: C127, V131, F134, M138, V139, W142 and V141 and α7: L157, L156, A152, P149 and C153 (Fig. 2C). Similar to saposin A and C, the α5 helix of NTD saposin B type domain is also found to have a kink in the helices which are remarkably similar but deviates from their main-chain torsion angles (φ = −83, ψ = −13) and is unable to form canonical α-helices at the base of F134 residue, which is not observed in saposin B. Accordingly, the P134–D137 intra helical n, n + 4 hydrogen bond was lost leading to the formation of a kink at α7 helix. The sequence comparison of saposin A, B and C with NTD saposin B type domain describes that six conserved cysteine residues plays major role for disulfide bond formation and its structural stability (Fig. 2D). The small helix α8 of NTD connects the PED and NTD via lengthy coil region composed of 26 residues (S169–S195) which is believed to be involved in accelerating the hydrolysis process of sphingolipids by PED.

Fig. 2 The comparison of saposin B-type domain. (A) Describes the superimposition of saposin B-type (green) domain with saposin A (purple) and C (cornflower blue). (B) Explains the open (dim gray) and closed (green) conformation of saposin B-type domain based on the comparison with the crystallographic structure of 1N69 upon binding with membrane. (C) Defines the saposin B-type domain containing the hydrophobic core region (residues are represented as ball and stick model) for its functional folding. (D) The sequence alignment clearly depicts the saposin fold conserved in comparison with other saposin type domain and show two stem composed of α4: C85–R100 and α7: P149–L157 which are linked by one hairpin formed by helices α5 (V105–L119) and α6 (P124–R144). There are six cysteine residues which are proven to form disulfide bond for its functional stability are marked in purple color. The hydrophobic residue which is involved for the kink formation in the entire saposin type domain is marked in red color.

PED. The phosphoesterase domain (PED) of ASM comes under the family of di-metal phosphoesterase (DMP) comprising of five metal binding motifs referred as MM1 to MM5. The PED was modeled based on the coordinates of hydrolase enzyme (PDB ID: 2XMO) and was observed to have two metallophosphoesterase signature motifs A and B described by Koonin et al. as like other phosphoesterase structure such as PAP (PDB ID: 1UTE). The predicted domain model contains an αβ sandwich topological protein fold as like other DMP structures and a di-metal center located on the two fold symmetry axis with each half comprised by a βαβαβ structural motif which is correspond to the DMP in PAP structure. In total, the PED domain contains eight helices spanning from α9–α16 and ten sheets named as β1–β10. The aforementioned five metal binding motifs are responsible for their functional activity by forming coordination interaction with two Zn²⁺ ions (Fig. 1). MM1, the first metal binding motif has D202 and forms coordination interaction with Zn²⁺(1) with a contact distance of 2.07 Å. Likewise, MM2 to MM5 metal binding residues D274, N314, H421 and H453 form coordination interaction with two zinc ions by a contact distance of 2.22, 2.26, 2.14 and 2.29 Å respectively. Though the metal coordination interactions are well conserved as observed in other di-metal phosphoesterase domain, MM1 has an extra residue H204 pointed towards the metal ion. This residue resides in the lengthy coil region formed by the residues D206–L248 and helps in maintaining the coordination interaction. Another residue H455 which is present beside the MM5 motif also contributes for metal coordination interaction with zinc with a contact distance of 3.30 Å. As like other di-metal phosphoesterase domain, ASM PED also contains conserved C246–D247 residues referred as CD peptide which resides at the end of the long coil region and before the α9 helix. The substrate recognition loop motif NX₃CX₃N (N377-MNF-C381-SRE-N385) lies in between β5–α13 and is accountable for the functional activity. In addition, the Niemann–Pick mutation residues (M378, N379 and W387) were also identified and are found to present near the substrate recognition loop motif and the di-metal center playing a vital role in substrate binding. Also, the hydrophilic/aromatic residues (D206–Y209 and H278) lying near the di-metal center also contribute in accelerating substrate binding and its subsequent functionality.

CTD. The C-terminal domain (CTD) contains six helices and is classified as the helix-only domain with helices α17 (Y533–Y539), α18 (L545–K558), α19 (T560–G573), α20 (G581–S594), α21 (S599–L606) and α22 (L612–V615). The α18 and α19 constitute the EF hand like motif and are placed equatorial to each other with distance and angle of 5.3 Å and 72.0° respectively (ESI Fig. 1†). The only helix–helix interaction was observed between α19–α20 helices which are placed bisectionally (angle: 48.3°) with a helical axis distance of 9.7 Å. The loop region (D206–L248) is placed between the α18 and α19 helices and is assumed to play a role in controlling PED domain structural transition.

Stability of the simulated ASM structure at three different pHs: neutral pH, pH 3.0 and pH 5.0

The RMSD profile of neutral, pH 3.0 and pH 5.0 ASM structure was obtained based on their backbone information and shown in the Fig. 3A. The structure of neutral, pH 3.0 and pH 5.0 of ASM were observed to be equilibrated after the 30^th, 20^th and 15^th ns respectively with the RMS deviation of ∼0.8 to 0.9 nm, which subsequently explains the stability nature of ASM. Accordingly, the pH 5.0 ASM structure shown, it is well equilibrated around a time period of 35 ns whereas pH 3.0 was equilibrated for about 30 ns only. The RMS fluctuation graphs (Fig. 3B) show that the NTD-saposin B type domain of neutral and pH 5.0 has similar RMS fluctuations to pH 3.0 ASM structure with RMS fluctuation ranging from ∼0.4 to 0.6 nm. The PED domain of ASM also shows well stabilized and similar kind of fluctuation except at the loop region (D206–L248). This loop shows higher fluctuation in neutral ASM structure of about 1.0 nm whereas pH 3.0 and pH 5.0 of ASM shows less fluctuation with RMS fluctuation up to 0.75 and 0.70 nm respectively. The CTD domain of neutral ASM structure shows minimal fluctuation in comparison with pH 3.0 and pH 5.0 with RMS fluctuation ranges from 0.5 to 0.75 nm and this elucidates the structural transition of CTD domain specifically in the pH 5.0 of ASM structure.

Fig. 3 The stability of ASM structure via molecular dynamics simulation studies. (A) The RMSD graph of ASM structure at different physiological pH (neutral, pH 3.0 and pH 5.0). (B) The RMS fluctuation of ASM structure with the clear description of each domain NTD, PED and CTD at different physiological pH.

The metal co-ordination interactions were also checked throughout the simulation period for the neutral, pH 3.0 and pH 5.0 ASM structure with Zn²⁺(1) and Zn²⁺(2) (ESI Fig. 2†). The neutral ASM structure shows that, upon simulation, the residues D202, D274, N314 maintain the co-ordination interaction with metal whereas the residues H204, H455, H421 and H453 lose their co-ordination interaction (ESI Fig. 2A†). As like neutral ASM, the pH 3.0 ASM structure also maintained D202, D274 interaction along with H204 metal interaction and lost their coordination interaction with H421, H455, N314 and H453 (ESI Fig. 2B†). In case of pH 5.0 ASM structure, D202 and D274 interactions were maintained with metal, and H421 and N314 maintain the metal interaction for half of the simulation period, whereas H204, H455 and H453 lost their co-ordination interaction (ESI Fig. 2C†). To summarize, in all structures, the dual interaction formed by D274 with the metal ion is intact but fluctuations in the histidine-metal co-ordination were present due to the flexible transitions gained in the loop region D206–L248 of ASM structure.

T-pad and mode vector analysis

The angular dispersion plot was generated for the 50 ns MD simulation of neutral, pH 3.0 and pH 5.0 ASM structures using T-pad tool (Fig. 4) which describes the fluctuation, transition and short transition derived from the trajectory. In neutral ASM structure, the residues of signal peptide show more number of transitions (PAD degree range: 42–101°) as well as short transition (50–146°) whereas fluctuation (19–96°) was observed by the contribution of minimal residues. The loop which connects the signal peptide and saposin B type domain shows fluctuation with less short transition observed in the residues 75–79 which indirectly characterizes their secondary structural changes. But, the saposin B-type domain is classified as a well stabilized domain showing 60% of residues involved only for fluctuation based on their PAD degree ranging from 20–60°. The proline rich region which connects NTD with PED domain is exposed to mainly transitions and short transitions with PAD degree 27–100° and is believed to be involved in the domain movement of NTD away from the PED domain which is confirmed from the mode vector analysis via the influence of this proline rich region. In PED domain, the region D206–L248 was supposed to control the hydrolysis of substrate with CTD which act as a controlling factor via its molecular packing towards catalytic domain. Accordingly, this region adopts 45% short transition with PAD degree ranging from 41–97° and is correlating with the expected structural transition observed in the mode vector analysis by moving towards the catalytic domain which subsequently regulates their functionality. Other region of PED shows 50% of the residues attain only fluctuations whereas transition and short transition was observed for only 16 and 34% of residues respectively, indicating the well stable nature of PED domain as like saposin B type domain which is also confirmed from the mode vector analysis by showing the restricted movement of their structural transition. The CTD domain illustrates that the residues attain mainly fluctuation than transition and short transition and moves in anti-parallel fashion against PED domain.

Fig. 4 The residual plasticity such as fluctuations, transitions and short transitions of ASM structure at different physiological condition (A) neutral, (B) pH 3.0 and (C) pH 5.0 were calculated using T-pad tool and plotted based on their PAD degree. The mode vector graph for corresponding physiological pH were also calculated from the 50 ns production MD simulation and represented for the understanding of domain motion. The fluctuations of residues were represented as starred connected lines, full transition are shown in red plus symbol whereas the short transition were shown in magenta rectangle shape.

The signal peptide of pH 3.0 ASM structure, in contrast to neutral ASM structure shows equal residual plasticity (fluctuation and transition) by showing PAD degree ranging from 19 to 98° and is exposed to move towards the saposin B type domain which is observed from their mode vector analysis. The same mode vector analysis describes the saposin B type domain movement towards the PED domain that is believed to be formulated by the aforementioned force created by signal peptide and it illustrates similar kind of conserved stability based on their PAD degree which is maintained well enough by more than 70% of the residues involved only in fluctuation (ranging from 19–80°). As like neutral ASM structure, the proline rich region shows only transition and short transition and helps in domain movement. In contrast to neutral ASM, the region which is believed to control the catalytic cavity was shown to have 51% of residues involved in fluctuation which indirectly explain the less plasticity nature of residues and its subsequent structural transition mimics somehow the opened conformation. This is also confirmed from the mode vector analysis showing restricted movement towards catalytic domain whereas the other region of PED domain shows 51% of residues involved only in fluctuation residual plasticity thereby confirming their structural stability as like neutral ASM structure. The residues of CTD domain shows 60, 15, 25% as fluctuation, transition and short transition respectively with PAD degree ranging from 19–147° and defines the vertical nature of movement towards the PED domain which significantly formulates the force to maintain the region (D206–L248) responsible for the regulation of catalytic domain unlike neutral ASM structure.

The mode vector analysis of pH 5.0 shows that the signal peptide moves towards the PED domain with residual plasticity of 57, 14 and 29% of fluctuation, transition and short transition respectively by influencing the PAD degree ranging from 20–65, 44–79 and 45–124° respectively. As like neutral and pH 3.0 ASM structure, pH 5.0 saposin B-type domain has conserved stability by having residual plasticity of 59, 19, 22% of the residues attaining fluctuation, transition and short transition with the PAD degree ranging from 19–60, 41–128 and 32–116° respectively. The region D206–L248 is responsible for its regulatory catalytic mechanism, which unlike pH 3.0 has higher residual plasticity towards short transition attained by 48% residues and less fluctuation showing the plasticity of about 25% ranges from 23–88° of PAD degree. Though, it has higher short transition residual plasticity, the movement observed in CTD domain which influences this loop region and formulates an open conformation and believed to form a tunnel for its catalytic mechanism. Accordingly, CTD domain moves vertically towards PED domain as like NTD domain mimicking the flying nature of bird wing and subsequently showing 59% of the residues involved only in the fluctuation along with transition and shorts transition of about 16 and 25% respectively. As like neutral and pH 3.0, pH 5.0 also maintains well the residual plasticity of about 50% of fluctuating residues whereas 24 and 25% of residues are involved for transition and short transition respectively as confirmed through their restricted movement seen in the mode vector analysis (Fig. 4). To summarize, the T-pad and mode vector analysis clearly explains that the loop region at pH 5.0 which regulates the catalytic domain developed an open conformation in comparison to the partially opened conformation at pH 3.0 and a closed conformation at neutral ASM structure. This visibly makes a note on the possibility of substrate binding at catalytic site under different physiological pH conditions. The detailed mechanism will be further exploited in the forthcoming section.

Structural transition of ASM in different physiological pH

In order to understand the structural transition of full length ASM structure the different physiological pH was applied using molecular dynamics simulation. Further, the free energy landscape (FEL) was constructed based on their Cα principal component analysis (PCA) to characterize the structural transition. The first two principal components which have the cosine content lesser than 0.2 were used for constructing FEL. The contour map generated from the FEL graph describes two, five and four free energy minimum cluster for neutral, pH 3.0 and pH 5.0 simulated ASM structure respectively. The free energy minimum cluster and its corresponding representative structure of neutral, pH 3.0 and pH 5.0 is clearly described in the ESI Fig. 3, 4† and 5 respectively along with the domain distance and volume (Tables 1 and 2). The superimposition of free energy representative structure and its subsequent structural changes are shown in Fig. 6.

Table 1 The domain volume measurements calculated for modeled and free energy minimum cluster generated from the FEL graph of neutral, pH 3.0 and pH 5.0 ASM were listed

Structure conformation		Domain volume measurements (Å^3)
		NTD	PED	CTD
Modeled ASM		11 170.342	39 201.935	11 402.067
Neutral	C1	12 277.476	42 814.842	12 181.407
	C2	12 394.853	43 223.428	12 049.152
pH 3.0	C1	11 896.071	43 053.605	12 253.649
	C2	11 866.154	43 077.705	12 157.727
	C3	11 898.198	42 642.728	12 150.488
	C4	11 860.277	42 967.019	12 089.513
	C5	11 883.340	42 555.454	12 195.041
pH 5.0	C1	12 115.566	42 676.324	12 067.844
	C2	11 975.899	42 466.175	12 001.785
	C3	11 822.492	42 791.628	11 899.003
	C4	11 833.770	42 641.645	11 786.069

---

Table 2 The domain–domain distance of free energy representative structure derived from each free energy minimum clusters of neutral, pH 3.0 and pH 5.0 ASM were listed

FEL generated structure of ASM		Domain–domain distance (Å)
		NTD–PED	PED–CTD
Neutral	C1	25.4	8.4
	C2	19.1	11.9
pH 3.0	C1	27.4	4.5
	C2	28.2	4.7
	C3	27.5	4.5
	C4	28.2	6.1
	C5	29.1	4.5
pH 5.0	C1	18.8	0.6
	C2	11.4	1.2
	C3	12.9	2.1
	C4	10.4	0.2

---

Structural transition of neutral ASM structure. Two free energy representative structure (C1: 24 840 ps and C2: 33 270 ps) were retrieved and analyzed for the calculation of domain distance and volume changes (Tables 1 and 2). Accordingly, the first free energy representative structure shows that the NTD–PED domain is placed of about 25.4 Å and PED–CTD is placed with a distance of 8.4 Å whereas in second free energy representative structure NTD moves closer towards PED domain with a distance of 19.1 Å (ESI Fig. 3†). Based on their free energy minimum cluster, the best structure (C2) was selected among the free energy representative structure. Further, the volume measurement of each domain in the best representative structure shows that the NTD, PED and CTD has increased volume of 1224, 4022 and 647 ų than the native ASM structure (Table 1) upon simulation and represents the relaxed folding nature of each domain and its subsequent structural transition. In connection with that, the superimposition of each free energy representative structure clearly explores the secondary structural changes observed in the region of P47–W50 and Q75–G79 by forming two parallel β-sheets via backbone hydrogen bond between V48-N with E76-O and F78-N with V48-O by a contact distance of 2.120 and 2.079 Å respectively. Here, the saposin B type domain and PED has shown very less fluctuation which is already observed in the T-pad analysis (Fig. 4). But, the loop connecting the β1 and α9 missing in the crystallographic structure of PAP shows much structural displacement by 18.8 Å with respect to CA atom of A327 in each representative structure (Fig. 6A). This displacement was also observed in the T-pad and mode vector analysis which clearly depicts the role of this loop in maintaining the closed conformation of catalytic cavity and hindering the hydrolysis process.

Structural transition of pH 3.0 ASM structure. The FEL graph shows five free energy minimum cluster (C1: 14 050 ps, C2: 17 240 ps, C3: 29 450 ps, C4: 39 280 ps and C5: 47 450 ps) among which the most populated free energy representative structure (C3) was selected for further analysis (ESI Fig. 4†). In all five representative structures, the NTD domain displaced by an average distance of 28.44 Å away from the PED domain whereas the CTD domain moved closer to the PED with an average distance of 4.86 Å which is lesser than that of the neutral ASM representative structures. The domain volume information from C3 shows that NTD (11 898.198 ų) and PED (42 642.728 ų) domain restricted their volume changes in comparison with neutral ASM representative structure whereas CTD (12 150.488 ų) has higher volume with an increase of 101 ų than the neutral ASM (12 049.152 ų) and illustrates the structural changes in the CTD domain (Table 2). The superimposition of C3 with other representative structures of ASM describes that the saposin B type and PED domains maintained their structural plasticity as like neutral ASM whereas the loop region 206–248 expected to regulate the catalytic cavity attained their displacement with an average distance of 8.05 Å (Fig. 6B). Here, the loop region depicts the partially opened conformation which is also confirmed based on the T-pad and mode vector analysis showing the partially restricted residual plasticity and loop movement. This may lead for the formulation of free space near the catalytic cavity but owing to the partial open conformation, the substrate binding efficiency would not be optimal.

Structural transition of pH 5.0 ASM structure. In case of pH 5.0 ASM structure, four free energy minimum cluster was generated based on the first two principal components and the corresponding free energy representative structure (C1: 21 730 ps, C2: 19 940 ps, C3: 31 030 ps and C4: 42 640 ps) were pulled out, which is displayed in the Fig. 5 with the distance of domain motions (Table 2). In all four representative structures, the NTD domain moves closer to the PED domain with a distance of 18.8, 11.4, 12.9 and 10.4 Å and at the same time CTD domain also moves closer to the PED domain with a domain axis distance of 0.6, 1.2, 2.1 and 0.2 Å respectively. Of these, the C4 is considered the best based on cluster size and here, the NTD and CTD domains act as wings for PED domain and move together in a compact manner to create well stabilized PED. This movement is also confirmed based on their residual plasticity and mode vector analysis which is discussed in earlier section. The volume information of best representative structure (C4) clearly shows that the NTD (11 833.770 ų) and CTD (11 786.069 ų) shows restricted volume in comparison to the volume occupied by neutral and pH 3.0 ASMs. The superpose image of all representative structure explains the structural restrained transition in the region of saposin B type and PED domain and the loop region which is expected to regulate the catalytic cavity shows an average displacement of 6.4 Å with respect to CA atom of A327 and pose for the fully opened conformation due to their tightly controlled minimal displacement (Fig. 6C). The structural transitions are very much limited in pH 5.0 ASM which holds the loop wide open allowing free space for substrate entry thereby leading to better binding efficiency and subsequent hydrolysis.

Fig. 5 The free energy landscape of pH 5.0 ASM structure derived based on the function of first two principal components whose cosine content were less than 2.0 using projection eigenvector of PCA. There are four representative structures identified and extracted from the most populated free energy minimum clusters and displayed with the details of domain–domain distance acquired based on the domain axis.

Fig. 6 The superimposition of each free energy representative structure obtained from the FEL graph of different physiological pH molecular dynamics simulation. (A) Superimpose of free energy representative structure of ASM at neutral pH showing the open and closed conformation (green color) of the loop region which is predicted to involve in the catalytic activity of PED domain. The secondary structural changes among the representative structure were marked using circular ellipsoid shape. (B) The five representative structures of pH 3.0 ASM structure superimposed and showing the partially opened conformation of the loop portraying towards their functional activation. (C) Shows superimposition of four free energy representative structure and its open conformation of loop region responsible for its catalytic mechanism.

Mechanism of ASM structure transition at different physiological pH

The opening and closure of the loop under different physiological pH regulates the substrate movement into the catalytic cavity. To elucidate the above, structural transition of full length ASM under different pH was studied to foresee any favorable hydrophobic interactions that may regulate the catalytic domain for its substrate hydrolysis (Fig. 7). Accordingly, the free energy representative structure for neutral ASM shows a closed conformation of the loop region (206–248) towards the catalytic domain thereby blocking the tunnel and making it inaccessible for the substrate hydrolysis. This closed conformation was maintained via series of hydrophobic interactions and folding mechanism, in which α18, α19 and α20 of CTD domain form parallel triple helical orientation along the loop region (217-CENPLCCR-224) axis (aL). Specifically, the α19 (560-TQLFQTFWFLY-570) helix is placed parallel to the aL with a distance of 5.4 Å and forms interaction through non-polar and polar uncharged amino acids. Here, the aL contains both polar and non-polar amino acids which are not stimulated by their protonation state; hence the aliphatic side chains of aL form a series of hydrophobic interaction (C584, F568, A587, A591, F563, F566, L569, P220, P583, V553, Y554, L221, C228 and L562) along with α18 and α20 and helps in the folding nature of loop. Near the upper part of loop region, the parallel helical orientation of α18 and α20 makes a second folding pattern towards the catalytic domain and forms series of polar and non-polar interaction between P231, A232, P235, Y239, V456, G236, A237, F459, G452, G425 and F451. The tail region of CTD (V615, L612, L618, P613), loop connecting α11–α12 (I328, P326 and G330) and loop connecting β5–α13 (N391, S392, L389, L388 and T393) of PED domain also contribute for this folding mechanism of loop region in a closed conformation (Fig. 7A). All these interacting mechanism are observed to be due to the triple helical orientation and parallel hydrophobic interaction with aL region. In case of pH 3.0 and pH 5.0 free energy representative structure of ASM, the protonation and de-protonation state of D551, R555, R557 and K558 of α18 (546-DLVYRMRK-558) and the residues E218 and R224 of aL region (217-CENPLCCR-224) repel the aL orientation and displaced away from the α19 with an axis distance of 8.6 and 8.4 Å in pH 3.0 ASM and pH 5.0 ASM respectively (ESI Fig. 5†). In both the case, the hydrophobic interaction were maintained among the triple helix (α18, α19 and α20) by the residues of pH 3.0 ASM (L552, L562, D551, C217, C222, Y554, V548, L562, C584, T588, Q564, F568, L569 and P546) and pH 5.0 ASM (A548, L552, V553, C217, E218, F566, L562, N589, Y570, P583, F568, L589, C590, A591 and F563). The different kind of residues and protonation states of triple helix involved in the interaction towards aL region makes the difference in the folding and transition mechanism of loop that leads to partially and completely opened conformation (Fig. 7B and C). On the other side, the tail region of CTD, the loop connecting α11–α12 and the loop connecting β5–α13 form triangle polar and non-polar interaction and the difference in the interacting pattern also prompts for the structural changes of partially and completely opened conformation (Fig. 7C). By studying the interacting mechanism we observed that compare to neutral and pH 3.0, at pH 5.0 the ASM structure is able to maintain the fully opened confirmation which is in turn stabilized by the hydrophobic interactions and protonation states of triple helix and aL loop region. This clearly depicts that physiological pH 5.0 is the most optimal pH for the structural and conformational stability of the ASM which opens up the substrate cavity for its effective functionality. This observation correlates with the experimental data showing that the L-ASM required acidic pH (5.0) for its binding with substrate by providing the opened conformation.^1,6

Fig. 7 The structural mechanism illustrates the open and closed conformation of the loop region which is believed to regulate the catalytic mechanism of PED domain. (A) The neutral ASM structure adopts some structural mechanism through series of interacting pattern (hydrophobic interaction) observed between the helical orientations of CTD with loop region 206–248. Other hydrophobic and polar interaction formed between tail region of CTD and loop of PED domain also contributes significantly for the closed conformation of the loop. (B) pH 3.0 ASM representative structure explains their interacting mechanism for the partially opened conformation of the loop. (C) Describes the interacting mechanism of completely opened conformation of the loop region illustrates a way for the substrate binding for its catalytic mechanism. Color code: hydrophobic interaction of CTD helical orientation-yellow sphere hetero atom types, loop region-green sphere hetero atom type, other polar and non-polar interaction between PED and CTD-cornflower blue sphere hetero atom type.

Molecular interaction of sphingomyelin with ASM

In order to check the molecular interaction of ASM structure at different physiological pH with substrate sphingomyelin, the best free energy representative structures were subjected to docking studies using SP Glide docking protocol under the Maestro 9.2 software. The docked complexes were also subjected to PDBePISA server for their mode of binding and interaction energy as displayed in the Fig. 8 and Table 3.

Fig. 8 The bio-molecular complex of different physiological pH ASM structure with SM (sphingomyelin). Interaction of SM with best free energy representative structure of (A) neutral ASM, (B) pH 3.0 ASM and (C) pH 5.0 ASM structures were displayed. The loop region which regulates catalytic domain is shown in green color. The zoomed image shows the interacting residues of PED domain with substrate and metal ions which clearly illustrate the best binding mode of substrate (SM) in pH 5.0 condition rather than pH 3.0 and neutral. Color code: interacting residues-cornflower blue stick mode, substrate-magenta, loop region-green and zinc ions-magenta sphere representation. All the interacting residues and substrate were represented as stick model with hetero atom types.

Table 3 The interacting residues with substrate and zinc ions, interface area and solvation free energy change with Glide docking energy are shown for each complexes of substrate with neutral, pH 3.0 and pH 5.0 free energy representative structure of ASM^a

Docked complexes	Structure 1			Structure 2			Interface area, Å^2	Δ^iG kcal mol^−1	Glide docking energy
	Mol	iNat	iNres	Mol	iNat	iNres
a The Δ^iG corresponds to hydrophobic interface or positive protein affinity towards the substrate binding.
Neutral-sphingomyelin	Residues	50	17	Lig	26	1	355.4	−1.5	−36.929
	Residues	13	6	Zn^2+(1)	1	1	56.9	−36.1
	Residue	5	4	Zn^2+(2)	1	1	46.7	−36.4
	Zn^2+(1)	1	1	Zn^2+(2)	1	1	7.0	−5.6
pH 3.0-sphingomyelin	Residues	54	19	Lig	30	1	412.7	0.6	−44.532
	Residues	6	5	Zn^2+(1)	1	1	48.8	−30.5
	Residues	9	5	Zn^2+(2)	1	1	46.4	−23.4
	Zn^2+(2)	1	1	Lig	3	1	30.1	−13.8
	Zn^2+(1)	1	1	Zn^2+(2)	1	1	0.4	−0.3
pH 5.0-sphingomyelin	Residues	59	24	Lig	28	1	407.5	−2.2	−37.946
	Residues	10	7	Zn^2+(1)	1	1	42.4	−22.7
	Residues	10	6	Zn^2+(2)	1	1	40.5	−25.6
	Zn^2+(1)	1	1	Lig	4	1	35.1	−16.4
	Zn^2+(2)	1	1	Lig	6	1	30.4	−13.3

---

Interaction of sphingomyelin with neutral ASM representative structure. The interacting complex of Sphingomyelin (SM) with neutral ASM shows that the substrate is bound at a different site which is away from the catalytic cavity of about 15.8 Å with respect to Zn²⁺(2). This different kind of binding originated due to the closed conformation of catalytic domain by the loop region (206–248) observed in the neutral ASM structure and displayed in Fig. 8A. Accordingly, the substrate SM interacts with ASM by residues of S382, F380, E384, N385, S392, T393, D394, L428, K429, W433, N434, S430, D394 and A396 with a solvation free energy of −1.5 kcal mol⁻¹ and an interface area of 355.4 Ų upon complex formation. Though SM binds at different binding cavity, the N2 and O5 atom of SM forms two hydrogen bonds with OG-S430 and N-K429 with a contact distance of 3.05 and 2.86 Å respectively. Other passive residues were involved for the polar and non-polar interaction with SM. Besides, the two zinc ions Zn²⁺(1) and Zn²⁺(2) maintain their coordination interaction with H315, D274, D202 and D226 with the solvation free energy of −36.1 and −36.4 kcal mol⁻¹ respectively. All these results indicate that the neutral ASM structure is not contributing well for its structural adaptability for substrate binding leading to the hydrolysis of SM.

Interaction of SM with pH 3.0 representative ASM. The interacting conformation of SM with best free energy representative of pH 3.0 ASM structure (Fig. 8B) has the solvation free energy of 0.6 kcal mol⁻¹ with interface area covered by about 412.7 Ų (Table 3). Though the SM is bound near the catalytic domain by the distance of 3.6 Å with respect to O3 atom with Zn²⁺(2), the mode of interaction expected was not observed as like in the crystallographic structure of PAP due to their partially opened conformation of the loop which restricts the SM binding towards the catalytic domain. This is an important finding which reasons out why pH 3.0 ASM might not be optimal for substrate binding. Here, the residues N314 and L388 form hydrogen bonds with O5 and O4 atom of SM with a contact distance of 2.96 and 2.81 Å respectively. In addition, the N atom of SM forms anion-π interaction with H421 with a distance of 5.0 Å which stabilizes the SM in the catalytic cavity. The residues I328, F327, P326, P324, T318, F386, L388, W387, I390, H315, N314 and I390 of ASM structure also form series of hydrophobic interactions with aliphatic chain of SM of makes their strong binding towards catalytic cavity. In the metal binding domain, the Zn²⁺(1) and Zn²⁺(2) form co-ordination interaction with the residues H204, D274 and D202 by solvation free energy of −30.5 and −23.4 kcal mol⁻¹ respectively. The substrate SM also forms interaction with Zn²⁺(2) with interaction free energy of −13.8 kcal mol⁻¹. However, since the SM bound near to catalytic domain has lesser solvation free energy and the difference in the mode of binding, there is very less effectual interaction with ASM structure in the physiological pH of 3.0.

Interaction of SM with pH 5.0 representative ASM. The docked conformation of SM with ASM structure (Fig. 8C) of best free energy representative structure was stabilized well in the cavity by having the solvation free energy of −2.2 kcal mol⁻¹ and interface area of 407.5 Ų. Most importantly, the phosphate ion containing two oxygen atoms O4 and O6 of SM faces the Zn²⁺(2) and Zn²⁺(1) with a distance of 3.16 and 3.00 Å by the solvation free energy of −13.4 and −16.4 kcal mol⁻¹ respectively (Table 3). The interaction pattern observed here is very similar to the crystallographic structure of PAP for its catalytic mechanism of ASM. Here, the residues H453 and H421 form hydrogen bond with O3 and N2 atom of SM with a distance of 3.11 and 2.87 Å respectively. A C–H⋯π interaction is also formed between C12 atom of SM with aromatic ring of H204 with a distance of 3.6 Å. In addition, the residues N391, S392, T393, S430, C427, I422, H453, H421, H315, and H204 also form polar and non-polar interaction with SM by forming perfect tunneling orientation towards the catalytic domain. Hence, least solvation free energy and the interacting mode of SM clearly depicts the proper catalytic mechanism by the ASM at physiological pH 5.0. Further, only in physiological pH 5.0, the zinc coordination with both substrate and ASM was observed whereas no such coordination interactions were observed in other physiological pH (Fig. 8). Here, the Zn²⁺(1) and Zn²⁺(2) have strong co-ordination interaction with residues H204, D274, H315 and D202 with a solvation free energy of −22.7 and −25.6 kcal mol⁻¹ respectively which is considerably lesser than the physiological pH 3.0.

On comparing the ASM structures at all three physiological pHs, the pH 5.0 ASM structure shows better solvation free energy and the perfect mode of interaction with SM as observed in the crystallographic structure of PAP which is not observed in neutral and pH 3.0 ASM structure. Hence, these observations give us an insight to the structural features that are crucial for the selection of optimal pH for ASM to exhibit its most efficient catalytic activity. This also confirms the experimental data based on their structural transition describing that the acidic pH (5.0) is very essential for the substrate binding.

Conclusion

In order to understand the full length ASM structure, the ASM structure was modeled based on their homologue PED domain and the remaining NTD and CTD domain were predicted based on their topological folding matching through ab initio calculation and thereby validated. In order to understand the influence of different physiological pH and its subsequent effect on the structural stability of ASM, a molecular dynamics simulation for a time period of 50 ns was performed. RMSD and RMSF profile of different physiological pH ASM structure reveal that pH 5.0 ASM structure has larger equilibrated time period of 35 ns whereas the neutral and pH 3.0 ASM has less in comparison. In addition, the loop region (D206–L248) of the pH 5.0 ASM structure which is proven to be involved in the regulation of catalytic cavity has restricted movement with RMS deviation of about 0.7 nm. But, the neutral and pH 3.0 ASM structures fluctuate around 1.0 and 0.75 nm respectively showing the huge displacement in the region. Additionally, the PCA and FEL graphs explored the series of changes which stimulate the folding mechanism of loop region (D206–L248) ranging from closed and, partially open to fully opened conformation towards the catalytic cavity. Accordingly, neutral ASM best representative structure depicts that series of hydrophobic interaction formulated between the α18, α19 and α20 (triplet helical orientation) of CTD domain with loop region axis (aL) and holds the loop in closed conformation. There is no protonation stimulation in the neutral ASM leading polar and non-polar interaction which subsequently helps for the folding nature of aL. Also, the tail region of CTD, loop of α11–α12 and β5–α13 of PED domain also contribute significantly for this folding of aL in a closed conformation and thus leading for the inaccessible nature of catalytic domain to substrate. In case of pH 3.0 and pH 5.0 ASM structure, the protonation and de-protonation residues reside in the triplet helical orientation and aL region repel the interactions which were observed in the neutral ASM and displaces the helix α19 by an axis distance of 8.6 and 8.4 respectively. Moreover, the difference in the interacting pattern observed in the other regions of PED and tail region of CTD makes aL region to be partially restrained (partially opened) and fully restrained (fully opened) conformation and providing the way for the substrate entry. Further, the molecular docking studies of SM with ASM structure at different physiological pH also confirms that the neutral and pH 3.0 ASM structure has a totally no and different binding mode of substrate in the catalytic cavity which is unsuitable thereby establishing that at these pH the substrate does not possess a proper binding mode for its subsequent catalysis. But, pH 5.0 ASM structure formulates proper binding mode of substrate by forming zinc coordination interaction along with oxygen atom of phosphate ions (SM) leads for the effective functional catalysis through an effective binding mode that adopts the same interaction profile as like in the crystallographic structure of PAP. Our study is in consensus with the experimental data explaining that L-ASM functionally active and mediates substrate binding at the physiological pH of 5.0 and has elucidated the structural transitions that bring about the difference.

Acknowledgements

Authors would thank Centre for Bioinformatics, Pondicherry University for the computational facilities.

References

1. J. W. Callahan, C. S. Jones, D. J. Davidson and P. Shankaran, J. Neurosci. Res., 1983, 10, 151–163.
2. R. W. Jenkins, D. Canals, J. Idkowiak-Baldys, F. Simbari, P. Roddy, D. M. Perry, K. Kitatani, C. Luberto and Y. A. Hannun, J. Biol. Chem., 2010, 285, 35706–35718.
3. Y. A. Hannun and L. M. Obeid, Nat. Rev. Mol. Cell Biol., 2008, 9, 139–150.
4. S. L. Schissel, E. H. Schuchman, K. J. Williams and I. Tabas, J. Biol. Chem., 1996, 271, 18431–18436.
5. S. L. Schissel, X. Jiang, J. Tweedie-Hardman, T. Jeong, E. H. Camejo, J. Najib, J. H. Rapp, K. J. Williams and I. Tabas, J. Biol. Chem., 1998, 273, 2738–2746.
6. R. W. Jenkins, D. Canals and Y. A. Hannun, Cell. Signalling, 2009, 21, 836–846.
7. S. L. Schissel, G. A. Keesler, E. H. Schuchman, K. J. Williams and I. Tabas, J. Biol. Chem., 1998, 273, 18250–18259.
8. T. Levade, R. Salvayre and L. Douste-Blazy, J. Clin. Chem. Clin. Biochem., 1986, 24, 205–220.
9. C. Muhle, H. B. Huttner, S. Walter, M. Reichel, F. Canneva, P. Lewczuk, E. Gulbins and J. Kornhuber, PLoS One, 2013, 8, e62912.
10. S. L. Sturley, M. C. Patterson, W. Balch and L. Liscum, Biochim. Biophys. Acta, 2004, 1685, 83–87.
11. F. M. Goni and A. Alonso, FEBS Lett., 2002, 531, 38–46.
12. R. Hurwitz, K. Ferlinz and K. Sandhoff, Biol. Chem. Hoppe-Seyler, 1994, 375, 447–450.
13. E. V. Koonin, Protein Sci., 1994, 3, 356–358.
14. E. H. Schuchman, M. Suchi, T. Takahashi, K. Sandhoff and R. J. Desnick, J. Biol. Chem., 1991, 266, 8531–8539.
15. C. P. Ponting, Protein Sci., 1994, 3, 359–361.
16. S. Zhuo, J. C. Clemens, R. L. Stone and J. E. Dixon, J. Biol. Chem., 1994, 269, 26234–26238.
17. T. Klabunde, N. Strater, R. Frohlich, H. Witzel and B. Krebs, J. Mol. Biol., 1996, 259, 737–748.
18. J. B. Vincent and B. A. Averill, FASEB J., 1990, 4, 3009–3014.
19. M. Seto, M. Whitlow, M. A. McCarrick, S. Srinivasan, Y. Zhu, R. Pagila, R. Mintzer, D. Light, A. Johns and J. A. Meurer-Ogden, Protein Sci., 2004, 13, 3172–3186.
20. D. Chivian and D. Baker, Nucleic Acids Res., 2006, 34, e112.
21. D. E. Kim, D. Chivian and D. Baker, Nucleic Acids Res., 2004, 32, W526–W531.
22. V. E. Ahn, P. Leyko, J. R. Alattia, L. Chen and G. G. Prive, Protein Sci., 2006, 15, 1849–1857.
23. Y. G. Kim, J. H. Jeong, N. C. Ha and K. J. Kim, Proteins, 2011, 79, 1205–1214.
24. X. Yang, J. Zhou, L. Sun, Z. Wei, J. Gao, W. Gong, R. M. Xu, Z. Rao and Y. Liu, J. Biol. Chem., 2007, 282, 24490–24494.
25. M. A. Marti-Renom, A. C. Stuart, A. Fiser, R. Sanchez, F. Melo and A. Sali, Annu. Rev. Biophys. Biomol. Struct., 2000, 29, 291–325.
26. M. Y. Shen and A. Sali, Protein Sci., 2006, 15, 2507–2524.
27. C. Colovos and T. O. Yeates, Protein Sci., 1993, 2, 1511–1519.
28. R. Luthy, J. U. Bowie and D. Eisenberg, Nature, 1992, 356, 83–85.
29. B. Hess, C. Kutzner, D. van der Spoel and E. Lindahl, J. Chem. Theory Comput., 2008, 4, 435–447.
30. D. Van Der Spoel, E. Lindahl, B. Hess, G. Groenhof, A. E. Mark and H. J. Berendsen, J. Comput. Chem., 2005, 26, 1701–1718.
31. M. S. Chimenti, V. S. Khangulov, A. C. Robinson, A. Heroux, A. Majumdar, J. L. Schlessman and B. Garcia-Moreno, Structure, 2012, 20, 1071–1085.
32. G. R. Grimsley, J. M. Scholtz and C. N. Pace, Protein Sci., 2009, 18, 247–251.
33. A. Onufriev, D. A. Case and G. M. Ullmann, Biochemistry, 2001, 40, 3413–3419.
34. A. Onufriev, A. Smondyrev and D. Bashford, J. Mol. Biol., 2003, 332, 1183–1193.
35. G. Collier, N. A. Vellore, R. A. Latour and S. J. Stuart, Biointerphases, 2009, 4, 57–64.
36. B. Hess, H. Bekker, H. J. C. Berendsen and J. G. E. M. Fraaije, J. Comput. Chem., 1997, 18, 1463–1472.
37. S. Miyamoto and P. A. Kollman, J. Comput. Chem., 1992, 13, 952–962.
38. R. Martonak, A. Laio and M. Parrinello, Phys. Rev. Lett., 2003, 90, 075503.
39. R. Huber, Nature, 1979, 280, 538–539.
40. J. H. Lin, Curr. Top. Med. Chem., 2011, 11, 171–178.
41. M. Nocker and P. Cozzini, Curr. Top. Med. Chem., 2011, 11, 133–147.
42. K. Teilum, J. G. Olsen and B. B. Kragelund, Biochim. Biophys. Acta, 2011, 1814, 969–976.
43. R. Caliandro, G. Rossetti and P. Carloni, J. Chem. Theory Comput., 2012, 8, 4775–4785.
44. K. Muthu, M. Panneerselvam, N. S. Topno and K. Ramadas, RSC Adv., 2016, 6, 15960–15975.
45. K. Muthu, M. Panneerselvam, N. S. Topno, M. Jayaraman and K. Ramadas, Cell. Signalling, 2015, 27, 739–755.
46. A. Amadei, A. B. M. Linssen and H. J. C. Berendsen, Proteins: Struct., Funct., Bioinf., 1993, 17, 412–425.
47. B. Hess, Statistical physics, plasmas, fluids, and related interdisciplinary topics, Phys. Rev. E, 2000, 62, 8438–8448.
48. G. G. Maisuradze and D. M. Leitner, Proteins, 2007, 67, 569–578.
49. R. A. Friesner, R. B. Murphy, M. P. Repasky, L. L. Frye, J. R. Greenwood, T. A. Halgren, P. C. Sanschagrin and D. T. Mainz, J. Med. Chem., 2006, 49, 6177–6196.
50. E. Krissinel, J. Comput. Chem., 2010, 31, 133–143.
51. E. Krissinel and K. Henrick, J. Mol. Biol., 2007, 372, 774–797.
52. M. Rossmann, R. Schultz-Heienbrok, J. Behlke, N. Remmel, C. Alings, K. Sandhoff, W. Saenger and T. Maier, Structure, 2008, 16, 809–817.
53. B. C. Bryksa, P. Bhaumik, E. Magracheva, D. C. De Moura, M. Kurylowicz, A. Zdanov, J. R. Dutcher, A. Wlodawer and R. Y. Yada, J. Biol. Chem., 2011, 286, 28265–28275.
54. D. C. De Moura, B. C. Bryksa and R. Y. Yada, PLoS One, 2014, 9, e104315.
55. K. Popovic, J. Holyoake, R. Pomes and G. G. Prive, Proc. Natl. Acad. Sci. U. S. A., 2012, 109, 2908–2912.

---

Footnote

† Electronic supplementary information (ESI) available: Fig. S1 to S5 and PDB file of modeled ASM. See DOI: 10.1039/c6ra16584b

---