In Silico Investigation of Molecular Mechanisms Underlying the Function of NLuc and its Variants

Bioluminescence, the emission of light by living organisms, results from chemiluminescent reactions facilitated by enzymes like luciferases. Among these, NanoLuc (NLuc) stands out due to its exceptional brightness, stability, and compact structure, making it a valuable tool in bioassays and imaging applications. NLuc is a 19.1 kDa monomeric enzyme derived from the deep-sea shrimp Oplophorus gracilirostris . Its structure comprises eleven antiparallel β-strands forming a β-barrel, capped by four α-helices. To enhance its versatility, NLuc has been engineered into split forms. The two-component split system divides NLuc into a large fragment and a small peptide, which can be either low-affinity (SmBiT) or high-affinity (HiBiT). Building upon this, a three-component system incorporates an additional peptide, offering improved control and potential applications in chemical biology. Despite the advancements in split NLuc applications, several impediments exist that can be addressed to improve these systems. Recent studies have illuminated the allosteric mechanisms of NLuc. The enzyme exhibits homotropic negative allostery, where product binding to an allosteric site inhibits substrate binding at the catalytic site. Ongoing research into its structural dynamics and allosteric behaviors continues to expand its potential applications, while efforts to enhance the efficiency of its split forms aim to broaden its utility in complex biological assays. In this in silico assay, we clarify the sources of malfunctions in split NLuc and also explore aspects of split NLuc technologies. We examine some hypotheses of NLuc mechanisms that display the complex behavior of this luciferase. In Silico Investigation of Molecular Mechanisms Underlying the Function of NLuc and its Variants Mina Oliayi a, Rahman Emamzadeh a *, Mohamad Reza Ganjalikhany a a Department of Cell and Molecular Biology and Microbiology, Faculty of Biological Science and Technology, University of Isfahan, Isfahan, Iran. * correspondence: [email protected]

Bioluminescence, the emission of light by living organisms, results from chemiluminescent reactions facilitated by enzymes like luciferases. Among these, NanoLuc (NLuc) stands out due to its exceptional brightness, stability, and compact structure, making it a valuable tool in bioassays and imaging applications. NLuc is a 19.1 kDa monomeric enzyme derived from the deep-sea shrimp Oplophorus gracilirostris . Its structure comprises eleven antiparallel β-strands forming a β-barrel, capped by four α-helices. To enhance its versatility, NLuc has been engineered into split forms. The two-component split system divides NLuc into a large fragment and a small peptide, which can be either low-affinity (SmBiT) or high-affinity (HiBiT). Building upon this, a three-component system incorporates an additional peptide, offering improved control and potential applications in chemical biology. Despite the advancements in split NLuc applications, several impediments exist that can be addressed to improve these systems. Recent studies have illuminated the allosteric mechanisms of NLuc. The enzyme exhibits homotropic negative allostery, where product binding to an allosteric site inhibits substrate binding at the catalytic site. Ongoing research into its structural dynamics and allosteric behaviors continues to expand its potential applications, while efforts to enhance the efficiency of its split forms aim to broaden its utility in complex biological assays. In this in silico assay, we clarify the sources of malfunctions in split NLuc and also explore aspects of split NLuc technologies. We examine some hypotheses of NLuc mechanisms that display the complex behavior of this luciferase. Key words: Nanoluciferase (NLuc), Split Nanoluciferase, NanoBiT, tri-part NLuc, Molecular Dynamics (MD), Docking

Bioluminescence is chemiluminescence that occurs in living organisms, where light is emitted through a catalytic reaction [1]. There is considerable interest in harnessing bioluminescence for designing more efficient, sensitive bioassays and developing sustainable, environmentally friendly lighting technologies [2,3]. Most bioluminescence based applications use photoproteins and luciferases [4]. Although numerous luciferase-luciferin systems have been developed, NanoLuc (NLuc) is considered the most efficient luciferase due to its brightness, long-lasting emission (half-life > 2 hours), and stable structure. This system, along with its substrate furimazine (FMZ), has been engineered by Hall et al. [5] NLuc is a monomeric enzyme composed of eleven antiparallel β-strands (S1-11) aligned into β-barrel which is capped by 4 α-helices (H1-4) (Fig. 1). It is predicted that the active site is located in a central cavity of β-barrel structure (PDB ID 5IBO) [6–8]. In 2016, through mutations in the structural amino acids of this enzyme, a two-component split version of NLuc was introduced, consisting of a large fragment and a terminal peptide of β-barrel structure [9]. A notable feature of this system is its design in two formats: with a low-affinity peptide (SmBiT) and with a high-affinity peptide (HiBiT), both of which bind to the large fragment but possess distinct characteristics and functionalities tailored for different applications. One year after the introduction of this system, a three-component technology [10], derived from the two-component system contains the HiBiT peptide, was developed, which has gained significant attention in recent years [11–14] (Fig. 1). Several structures of crystallized NLuc have been reported, including structures that have been solved with substrate analogs, product, or bound inhibitors; however, the crystal structure of the catalytically favored luciferin-bound enzyme complexes of NLuc with its substrates furimazine (FMZ) and coelenterazine (CTZ) has not been reported yet [5,6,15]. The reported structures are found in two forms, which, according to the study by Nemergut et al., correspond to the open and closed forms of the enzyme [15]. The open form is characterized by a spacious internal tunnel, a collapsed surface pocket, and the presence of Y94 inside the inner channel. In contrast, in the closed form, the internal tunnel becomes significantly narrower or partially collapsed, Y94 is oriented outside the inner tunnel, and the surface pocket, consisting of the amino acids R43, H57, K89, and Y94, is rearranged to accommodate the binding of the product furimamide (FMA). Nemergut et al. describe the behavior of NLuc as exhibiting homotropic negative allostery, indicating that binding of the product to the surface allosteric site prevents simultaneous substrate binding to the catalytic site. However, the precise processes leading to the conversion between the open and closed forms remain unclear. Interestingly, Y94 participates in a flip-flop process, transitioning between the inside and outside of the internal tunnel in the open and closed conformations, respectively, while H93 moves in the opposite direction. This suggests that when the internal tunnel is closed, the surface pocket is optimally positioned to bind FMA [15]. Despite efforts to elucidate the substrate oxidation mechanism, the exact reaction and molecular conformational changes remain unclear [15,16]. Our results show that allosteric structural changes induced by the substrate and product facilitate the conversion between the open and closed conformations. In addition, apparent changes in efficiency in previous studies suggest that the split systems shows some level of decreased function compared to the original form of NLuc [17]. Although a direct comparison between the intact and split systems has not been conducted so far, Hall’s study, which involved modifications to the B9 peptide and the large fragment of the tri-part system, reported that NanoBiT exhibits less than half the activity of intact NLuc. As well as, the tri-part NLuc system retains 60% of NanoBiT’s activity and 30% of intact NLuc’s activity. Moreover, due to challenges in attaching the peptides of the tri-part system to nanobodies recognizing AFP—such as the lack of expected emission from NLuc and insufficient information about the adequate length of linkers, as well as the proper orientation and positioning of tags [13]—a structural investigation of these split systems could elucidate the reasons for their incomplete efficiency. Here, we investigate, on one hand, intact NLuc to gain more insight into its various conformations and to understand the mechanisms involved in transitioning between its open and closed forms. On the other hand, we aim to address challenges in the split forms to improve split technologies and facilitate the design of split systems using molecular dynamics methods.

Structural Preparation Structures were obtained from Protein Data Bank for the open form (PDB ID 5B0U), closed form (PDB ID 5IBO), the binary split NLuc with the SmBiT peptide (NanoBiT, PDB ID 7SNX), the large fragment of the binary split NLuc (11S, PDB ID 7SNY). Other structures were obtained from modeling using RoseTTAfold method by Robetta server[18] for the intact NanoBiT—NLuc containing the mutations of NanoBiT system (NBit-o)—the binary split NLuc with the HiBiT peptide (NanoHiBiT), the ternary split NLuc (tri-part NLuc), the large fragment of the tri-part NLuc (Δ11S), and Δ11S with peptide β9 (Δ11S-β9). These structures were then used as initial models for MD simulations. Structural analyses were conducted using PyMOL 2.5.7 [19] and Visual Molecular Dynamics (VMD) [20]. Molecular Dynamics Simulations Molecular dynamics simulations were carried out using the AMBER package (version 22) [21] with the ff19SB force field. The appropriate protonation state of histidine residues was adjusted using the H++ server [22]. Structural neutralization was achieved by adding sodium ions to the system. All molecules were solvated in a 10 Å layer of explicit water using the OPC model within a truncated octahedral box, prepared by xLEaP (from ambertools 23). The coordination and topology files were used for subsequent minimization and MD simulations. The simulation process began with energy minimization. Initially, the protein complex was kept fixed while the ions and water molecules were minimized over 8,000 steps. This was followed by minimizing the entire system, utilizing 4,000 steps of the steepest descent method and 4,000 steps of the conjugate gradient algorithm. Hydrogen bond constraints were applied using the SHAKE algorithm [23]. Non-covalent interactions were computed with the Particle Mesh Ewald (PME) method [24], employing periodic boundary conditions and a 10 Å cutoff. The system was then heated from 0 to 300 K over 400 ps in a constant-volume (NVT) ensemble, controlled by a Langevin thermostat [25]. Next, equilibration was conducted for 5 ns under constant pressure with isotropic position scaling in the NPT ensemble. Finally, production molecular dynamics (MD) simulations were carried out for 100 ns and 200 ns (with FMA and FMZ) using the pmemd.cuda engine, with the NPT ensemble. The system’s coordinates were saved every 4 ps, and a time step of 2 fs was applied throughout the simulations. Analysis of Simulations Trajectories were analyzed using Cpptraj (V6.18.1) [26] from AmberTools23 [27]. The root mean square deviations and fluctuations were calculated referring to the initial structure. The distances between specific atoms excluding hydrogen, the number of contacts and also secondary structures were calculated using the DSSP method of Kabsch and Sander. Graphs were plotted using Xmgrace [28] and Gnuplut [29]. Graphs were smoothed using python scripts. Tunnel exploration The tunnel structures were analyzed using the Caver server [30]. The main tunnel is reported by Nemergut et al., which has an opening between H2 (Helix 2) and H3 (Helix 3). Docking of the substrate through the tunnel was performed using the transport analysis panel, which employs Autodock Vina [31] and fetches the substrate structure from PubChem [32] on this server.

In the current in silico study, MD simulation has been undertaken to investigate the structure-function relationship and other functional features of the open (5B0U) and closed (5IBO) forms of NLuc, the large fragment of the binary system (11S), the binary system with the low-affinity peptide SmBiT (NanoBiT), the binary system with the high-affinity peptide HiBiT, the large fragment of the ternary system (Δ11S), the large fragment of the ternary system with the β9 peptide (Δ11S-β9), the ternary system (Tri-NLuc), and the non-split NanoBiT (NBiT). Analyzing the Structural Characteristics The trajectories form the MD simulations were analyzed and RMSD, RMSF and structural clustering have been studied to measure the structural flexibility, stability and obtain predominant conformations of each structure during the simulation. stability and flexibility of different parts. Based on the RMSD results (Fig. 2), the large fragments in the binary and ternary systems (11S and Δ11S, respectively) showed higher RMSD values compared to the complete form of NLuc, while subsequent peptide attachment increases the RMSD. RMSF analysis showed the same structural flexibility in both open and closed conformations of NLuc, which can be regarded as relatively rigid constructions in either forms (Fig. 3). Additionely, the RMSF analysis reveals that the overall flexibility in Tri-part NLuc structure decreased following the binding of peptides. This observation indicates enhanced structural coherence of the enzyme upon formation of the complete construct. In contrast, both the 11S fragment and NanoBiT exhibit almost the same flexibility, indicating the higher stability of large fragment (11S) in the binary system compared to the ternary systems (Δ11S). Noteworthy that three uncompleted structures such as 11S, Δ11S, and Δ11S-β9 demonstrate more flexibility around residues 25-35 which are located around H3 (Helix 3) (Fig. 4). Summary data from hierarchical clustering methods shows that all structures display acceptable homogeneity in clustering (Supplementary Table 1). As such, the first cluster of each structure has been chosen in order to more analyzeing. Analysis of the Open and Closed Forms of NLuc The position of surface pocket amino acids of the first cluster of open and closed structures of NLuc are shown in Fig. 5. The transformation from open to closed confirmation is displayed in Fig. 6, showing the C-terminal of S6 and N-terminal of S7 strands turn into loops as the structure is altered from closed to open conformation which may create the adequate space to flip-flop Y94 and H93. However, this flip-flop wasn’t captured in 100 ns MD. To investigate inter-conformational changes, we monitored the distances between R162, a key residue in the active site, and Y94 and H93. The graphs indicate that the distance between R162 and Y94 in the open form structure is shorter than in the closed form structure throughout the simulation. Conversely, the shortest distance between R162 and H93 is observed in the closed form structure (Fig. 7). As illustrated in the figure, the open and closed conformations do not interconvert during the simulation. This suggests that the transition between the open and closed forms of the enzyme is not a spontaneous process and is likely hindered by a significant energy barrier between the two conformations. Binary split NLuc structures analysis Structural investigations determined that, on one hand, in the NanoBiT molecule, Y94 is located inside the structure, similar to open form (Fig. 6), which is possibly related to the H93P mutation. While the narrowed tunnel structure of this protein closely resembles that of the closed form structure. (Fig. 8C). Further analysis revealed that H3 is extended in NanoBiT, a feature commonly observed in closed form compared to open form (Supplementary Fig. 1). This extension may contribute to the narrowing of the tunnel (Fig. 8C). The distance between H2 and H3 was measured during the simulation, as the tunnel entrance is located between these two helices. The results indicate that the distance between H2 and H3 decreases in the closed structure compared to the open structure, with the shortest distance observed in the NanoBiT structure compared to both the open and closed conformations of NLuc (Fig. 9B). To investigate whether the disruption of the main tunnel is a cause of splitting or a result of mutations within these structures, the NanoBiT system without any cleavage has been investigated during 100 ns of MD simulation. The results indicated the disappearance of the tunnel in this structure as well, suggesting that the observed changes are mutation-induced. One notable mutation is G35A. Given that alanine has a higher propensity to participate in alpha-helical structures rather than glycine [33], the extension of H3 in the split systems appears to be a predictable outcome. Based on the results of the native contacts in the binary system, the interactions between 11S and SmBiT appear to be stronger and longer-lasting for residues in the middle and C-terminal region of SmBiT, which can be considered the most critical linkage for strengthening the overall structure of NLuc (Supplementary Fig. 2). Additionally, as key residues in the active site, R164 and R162 are part of this highly connected fragment, providing the necessary stability to position these residues appropriately within the active site. Therefore, it is advisable to avoid disrupting these crucial interactions when fusing or tagging the fragments to other proteins - one of the significant challenges in split-system development. According to previous studies, fusing tags to the N-terminal of both SmBiT and 11S is more efficient than using the C-terminal. For example, in the study by Dixon et al. [9], the best pairs with the highest signal-to-background (S/B) ratios were those in which both fusion proteins were attached to the N-terminal of SmBiT and 11S, which correlates with the data presented here. Similarly, in the study by Dixon et al. [10], 86% of N-terminal fusions to β10 pairs exhibited brighter luminescence compared to C-terminal fusions. Furthermore, as noted by Nemurgut et al., the C-terminal region of NLuc is a critical area, making it sensitive to truncation or extension, which can significantly affect the functionality and stability of the protein [15]. All clusters of 11S and Δ11S exhibit a relatively stable structure with the open scares remaining of the peptide segments (Supplementary Fig. 3). This suggests that during the reconstruction of the complete NLuc, the larger fragment does not require significant conformational changes, thereby facilitating the attachment process. However, as shown in the figure and supported by the RMSF analysis of the Δ11S fragment, this segment appears to undergo more structural alterations compared to 11S. Structural analysis of tri-part NLuc Regarding the tri-part NLuc, the interactions of the B9 and B10 peptides in the tri-part complex were separately calculated with a cutoff of 5 Å. According to percentage interaction during the simulation, the overall interaction stability for B9 was approximately 77%, while for B10, it was 63%. This suggests that the number of interactions between Δ11S-β10 and peptide β9 was greater and more persistent during the simulation compared to peptide β10. Therefore, the hypothesis arose as to whether the binding of these two peptides to the large fragment is a sequential interaction. To investigate this hypothesis further, the molecualr docking analysis of Δ11S with β9 and β10 was performed, using the Haddock server. The docking results revealed that the binding affinity of β10 to Δ11S was lower compared to β9, based on the Haddock score. The first cluster, with a score of -77, shows the most favorable binding and the best score for this complex, while the binding of β9 to Δ11S has a Haddock score of -89 (Fig. 10). This observation can be somewhat predicted by considering the structure of Δ11S at the binding sites of each peptide. The second structure in the binding region to β9 remains unchanged throughout the simulation, whereas the binding region to β10 undergoes significant alterations over time. Additionally, a comparison of the binding energies obtained from the Haddock server for β10 binding to Δ11S and to the Δ11S-β9 complex shows a dramatic decrease from -77 to -162, indicating an increased affinity of β10 after the binding of β9. Meanwhile, the score for β9 binding to Δ11S and to the Δ11S-β10 complex increased from -89 to -150, demonstrating that the change in structure and enhanced affinity for β10 occurs more significantly upon the binding of β9 compared to the binding of β9 to the Δ11S-β10 complex. These score differences suggest that the sequential binding of the peptides is likely necessary for the formation of the complete structure in the tri-part system. Analysis of open and closed forms in the presence of Furimazine and Furimamide Since previous studies have proposed that the presence of a product in the surface pocket may induce a conformational change from open to closed, to investigate the structural changes between the open and closed conformations due to the presence of the product or substrate in the central tunnel and the product in the surface pocket, first the molecular docking has been performed on the open and closed conformations of the NLuc structures and then MD simulations of three systems were performed for 200 ns . These include the open form structure with the FMA product in both the surface pocket and central tunnel separately, as well as the closed form structure with the FMZ substrate in the opening of the central tunnel. In our simulation time The flip-flop conformational swapping has not been observed for Y94 and H93 which is in line with Nemergut’s study. The orientation of Y94 in the closed conformation in the presence of substrate/product did not shift from the surface pocket to the tunnel interior during 100 μs of MD simulation in their work. [15]. The comparison of the RMSD of the open structure (5B0U) in the presence of the FMA product in the surface pocket, as well as in the central tunnel, with the closed structure (5IBO) in the presence of the FMZ substrate in the central tunnel opening, revealed that the most significant structural changes are associated with the binding of FMA to the enzyme surface. These changes occur particularly at simulation times when the product is closer to the entrance of the central tunnel. Furthermore, the data indicates that surface interactions play a more significant role in the structural changes of this enzyme compared to interactions within the central cavity (Fig. 11). As it has been depicted in Fig. 12, FMA moved on the surface of the open form towards the space between H2 and H3, to the vicinity of the entrance of the internal tunnel. The RMSF analysis of this system indicates increased mobility in the residues 100–115, corresponding to the β-strand 6 (S6), caused by the substrate’s presence in the tunnel of the closed form of the enzyme (5IBO) (Fig. 13A and C). This increased instability is also observed in the open form without the product (5B0U) (Fig. 13A and B). Based on earlier discussion, this suggests that instability in this region may facilitate the flip-flop motion between Y94 and H93. Additionally, the presence of FMA on the enzyme surface enhances flexibility in regions H2-H3 and particularly in H4, which collectively form the enzyme’s cap structure. The data obtained comparing the flexibility of the open structure without the product and in the presence of the product bound to the central tunnel and the external surface revealed that FMA binding increased flexibility in the H2-H3 and H4-S4 regions compared to open form without the product (Fig. 13A). Additionally, a reduction in flexibility was observed in the residues 110–120 region (ending half of S6 and beginning half of S7) due to the presence of FMA in both the central tunnel and external pocket. This reduction in flexibility mirrors the flexibility pattern of closed form in the same region (Fig. 13C). According to the study by Nemergut et al., the surface pocket in the open conformation open form exhibits cramped structure [15]. The condition of the spacious and cramped surface pocket can be calculated using the distance between residues H57 and H93 (in open form) and H57 and Y94 (in 5IBO). By measuring these distances in the presence of FMA and FMZ respectively, and by comparing the pocket structure in the first cluster of open and closed conformations with and without the presence of the product and substrate, the status of the surface pocket in these two structures can be analyzed. These analyses revealed that the distance between residues H57 and H93 increases in the open conformation during the simulation time in the presence of the product within the central tunnel, with observations from the first cluster structures indicating an opening of this pocket (Fig. 14). Tunnel investigation The structure of the first cluster of closed and open conformation have been studied by Caver server. Based on the data obtained, although the entrance of closed form (Fig. 15B) has become wider compared to open form (Fig. 15A), its depth has significantly decreased, which could be attributed to the closure of the channel (Supplementary Table 2). Additionally, Furimazine as the substrate is used to dock through the tunnels in both open form and closed form and the energy of Furimazine passing through the tunnel has been calculated using AutoDock vina by the caver server. As depicted in Fig. 16, the furimazine docking energy through open form is far less than that through closed form and the substrate can go deeper inside the tunnel in open form so as the energy of passing through furimazine is negative to depth of 8 Å in open form whereas it is so to depth of 2 Å in the case of closed form. Furthermore, the results display cramped tunnels with narrow opening compared to the closed form of NLuc (Fig. 15C), despite the internal position of Y94 which is observed in open conformation. In order to investigate the passage of substrate through the tunnel of the NanoBiT, the data indicates that the passage of this substrate is energy-consuming, suggesting that the tunnel of this structure becomes narrow and unsuitable for substrate passing (Fig. 16). The tunnels of the first cluster of HiBiT (Fig. 15D) and Tri-NLuc (Fig. 15E) were examined, and the results showed that the main tunnel in the HiBiT structure has significantly narrowed, while in the Tri-NLuc structure, it is completely closed. Docking results indicated that the passage of furimazine through HiBiT requires energy, whereas the secondary tunnel in Tri-NLuc could be a potential candidate for substrate entry (Fig. 16). The results obtained from substrate docking along the internal tunnel for the first cluster of 5IBO-FMA (Fig. 16) simulation showed that the binding energy for the substrate’s deeper penetration became more negative and favorable compared to closed form without the substrate. This data may suggest an allosteric mechanism in the catalytic process, so that the binding of the substrate to the tunnel gate of the closed form leads to expanding the tunnel like the open form of the enzyme, which has an accessible active site for the substrate. However, Y94 did not relocate into the central cavity during the 200 ns of MD simulation. The consistency of these results with the data obtained from native contact (Supplementary Fig. 4) calculations indicate that the interaction between the substrate and the enzyme is increasing during the simulation, and the substrate is stably bound to the enzyme. Finally, exploring the tunnel for the reverse mutant of A35G (Fig. 15G) showed that this mutation prevents the H3 extension followed by inducing suitable tunnel for passing FMZ through it (Fig. 16).

Nanoluciferase (NLuc) has attracted remarkable attention from researchers worldwide. Despite its usefulness, the molecular feature of this enzyme is still unclear. Computational data in this article are gathered to explore the molecular behavior of NLuc in complete and split forms to complement the available information and further explore the molecular mechanism and also to help more efficient case-designed approaches. The presnt in silico study was conducted in two phases. In the first phase, open, closed, and all structural configurations of the split systems were simulated in the absence of the product (FMA) or substrate (FMZ) over 100 ns of MD simulation. In the second phase, structural changes were examined in response to the presence of the FMA in the central tunnel and the surface pocket of the open form (5B0U) separately, as well as the presence of the substrate in the central tunnel of the closed form (5IBO). The results of the analysis of the first study indicated that the open and closed structures are inherently stable and likely require an external factor to transition between each other. Engineered NanoBiT has malfunctions affected by mutations in addition to limitations imparted by the split feature, as evidenced by Hall et al. report that luminescence of NanoBiT is less than half of the NLuc [17] Accordingly, analysis of the split structures revealed that these systems, due to mutations introduced to enhance efficiency under split conditions, possess a central tunnel that is too narrow and unsuitable for substrate passage. Among these mutations, G35A appears to have the most significant impact. As indicated by tunnel analysis data, a reverse mutation at this position improves substrate passage through the tunnel. Detailed examination of the split structures showed that, based on critical interactions between the SmBiT peptide and the larger fragment, as well as reports by Nemergut and colleagues [15], the N-terminal of the peptide is the most favorable site for tag attachment. Additionally, we observed that in the three-component system, the B9 and B10 peptides do not bind randomly to the large fragment but rather in a sequential manner, with B9 binding first, followed by B10. Structural analysis of the large fragment in the binary and ternary split systems indicated greater stability for the 11S, suggesting that the binary system is likely more efficient than the ternary system [17]. In the second phase of simulation study, and inline with simulation data obtained by Nemergut and colleagues [15], we did not observe flip-flop movement of the two residues Y94 and H93 during the simulation; however, significant structural changes were detected, indicating a transition conformation between two open and closed forms. Notably, structural stabilization in the S6 region and the open surface pocket, when FMA is positioned in the central tunnel of the open form (5B0U), suggests that the presence of the product in the central tunnel can induce the closed conformation, providing a suitable binding site in the surface pocket for FMA or FMZ. We suggest that the simultaneous binding of two products or a product and a substrate may reduce the energy barrier between the open and closed conformations. In line with that, simulation of the docked substrate at the channel entrance of the first cluster of the closed structure (5IBO) revealed that over 200 ns this substrate partially enters the inner channel, causing allosteric changes in the structure. As a result, the central tunnel becomes wider and deeper than the tunnel in the closed structure. This process could indicate the beginning of an induced-fit mechanisms caused by the binding of the substrate to the tunnel opening. Additionally, taken from results gained from simulation of open form and FMA in the surface pocket, there is no evidence to show altering the surface pocket to more spacious one and no conversion to closed form but FMA moves on the surface of the protein towards the entrance of the central pocket. Since the dimeric form of NLuc has not been observed in solution when bound to either the product or substrate, and the binding of the product to the surface pocket has so far been limited to crystallographic structures and computational studies, one hypothesis arising from the results of this research is that the surface pocket may play a role in increasing the local substrate concentration near the entrance of the internal tunnel. Considering that the conversion between closed and open structures is a time-consuming process, it can be interpreted that the prolonged emission duration of NLuc [5] is likely due to the time required for the formation of the appropriate structure during the enzymatic catalysis process. Despite significant efforts to uncover the mechanism of NanoLuc (NLuc) enzyme activity, the factor responsible for the opposing reorientation of the residues Y94 and H93 remains unresolved. This question highlights the need for further structural and functional studies to clarify the interplay between these residues during enzymatic catalysis and strategies to improve activity of the split-systems. Conflicts of interest: The authors declare that they have no known conflicts of interest or personal relationships that could have appeared to influence the work reported in this paper. Acknowledgments We would like to express our appreciation to Dr. Shawn Owen of the University of Utah for his valuable feedback. Research reported here was supported by Iran National Science Foundation (INSF).

[1] S.H.D. Haddock, M.A. Moline, J.F. Case, Bioluminescence in the sea, Ann Rev Mar Sci 2 (2010) 443–493. https://doi.org/10.1146/annurev-marine-120308-081028.[2] A.J. Syed, J.C. Anderson, Applications of bioluminescence in biotechnology and beyond, Chem. Soc. Rev. 50 (2021) 5668–5705. https://doi.org/10.1039/D0CS01492C.[3] Y. Su, J.R. Walker, Y. Park, T.P. Smith, L.X. Liu, M.P. Hall, L. Labanieh, R. Hurst, D.C. Wang, L.P. Encell, N. Kim, F. Zhang, M.A. Kay, K.M. Casey, R.G. Majzner, J.R. Cochran, C.L. Mackall, T.A. Kirkland, M.Z. Lin, Novel NanoLuc substrates enable bright two-population bioluminescence imaging in animals, Nat Methods 17 (2020) 852–860. https://doi.org/10.1038/s41592-020-0889-6.[4] S. Schramm, D. Weiß, Bioluminescence – The Vibrant Glow of Nature and its Chemical Mechanisms, ChemBioChem 25 (2024) e202400106. https://doi.org/10.1002/cbic.202400106.[5] M.P. Hall, J. Unch, B.F. Binkowski, M.P. Valley, B.L. Butler, M.G. Wood, P. Otto, K. Zimmerman, G. Vidugiris, T. Machleidt, M.B. Robers, H.A. Benink, C.T. Eggers, M.R. Slater, P.L. Meisenheimer, D.H. Klaubert, F. Fan, L.P. Encell, K.V. Wood, Engineered Luciferase Reporter from a Deep Sea Shrimp Utilizing a Novel Imidazopyrazinone Substrate, ACS Chem. Biol. 7 (2012) 1848–1857. https://doi.org/10.1021/cb3002478.[6] Y. Tomabechi, T. Hosoya, H. Ehara, S.-I. Sekine, M. Shirouzu, S. Inouye, Crystal structure of nanoKAZ: The mutated 19 kDa component of Oplophorus luciferase catalyzing the bioluminescent reaction with coelenterazine, Biochem Biophys Res Commun 470 (2016) 88–93. https://doi.org/10.1016/j.bbrc.2015.12.123.[7] S. Inouye, J.-I. Sato, Y. Sahara-Miura, Y. Tomabechi, Y. Sumida, S.-I. Sekine, M. Shirouzu, T. Hosoya, Reverse mutants of the catalytic 19 kDa mutant protein (nanoKAZ/nanoLuc) from Oplophorus luciferase with coelenterazine as preferred substrate, PLoS One 17 (2022) e0272992. https://doi.org/10.1371/journal.pone.0272992.[8] T. Altamash, W. Ahmed, S. Rasool, K.H. Biswas, Intracellular Ionic Strength Sensing Using NanoLuc, Int J Mol Sci 22 (2021) 677. https://doi.org/10.3390/ijms22020677.[9] A.S. Dixon, M.K. Schwinn, M.P. Hall, K. Zimmerman, P. Otto, T.H. Lubben, B.L. Butler, B.F. Binkowski, T. Machleidt, T.A. Kirkland, M.G. Wood, C.T. Eggers, L.P. Encell, K.V. Wood, NanoLuc Complementation Reporter Optimized for Accurate Measurement of Protein Interactions in Cells, ACS Chem. Biol. 11 (2016) 400–408. https://doi.org/10.1021/acschembio.5b00753.[10] A.S. Dixon, S.J. Kim, B.K. Baumgartner, S. Krippner, S.C. Owen, A Tri-part Protein Complementation System Using Antibody-Small Peptide Fusions Enables Homogeneous Immunoassays, Sci Rep 7 (2017) 8186. https://doi.org/10.1038/s41598-017-07569-y.[11] S.J. Kim, A.S. Dixon, S.C. Owen, Split-enzyme immunoassay to monitor EGFR-HER2 heterodimerization on cell surfaces, Acta Biomater 135 (2021) 225–233. https://doi.org/10.1016/j.actbio.2021.08.055.[12] S.J. Kim, A.S. Dixon, P.C. Adamovich, P.D. Robinson, S.C. Owen, Homogeneous Immunoassay Using a Tri-Part Split-Luciferase for Rapid Quantification of Anti-TNF Therapeutic Antibodies, ACS Sens. 6 (2021) 1807–1814. https://doi.org/10.1021/acssensors.0c02642.[13] M. Oliayi, R. Emamzadeh, M. Nazari, Nanobody-enhanced split-luciferase technology for innovative detection of the liver cancer biomarker alpha-fetoprotein, Biochemical Engineering Journal 209 (2024) 109402. https://doi.org/10.1016/j.bej.2024.109402.[14] M.E. Baghdadi, R. Emamzadeh, M. Nazari, E. Michelini, Development of a bioluminescent homogenous nanobody-based immunoassay for the detection of prostate-specific antigen (PSA), Enzyme and Microbial Technology 180 (2024) 110474. https://doi.org/10.1016/j.enzmictec.2024.110474.[15] M. Nemergut, D. Pluskal, J. Horackova, T. Sustrova, J. Tulis, T. Barta, R. Baatallah, G. Gagnot, V. Novakova, M. Majerova, S.M. Marques, M. Toul, J. Damborsky, D. Bednar, Z. Prokop, Y.L. Janin, Illuminating the mechanism and allosteric behavior of NanoLuc luciferase, (n.d.).[16] V. Orel, Kinetický mechanismus NanoLuc luciferázy, (n.d.).[17] M.P. Hall, V.A. Kincaid, E.A. Jost, T.P. Smith, R. Hurst, S.K. Forsyth, C. Fitzgerald, V.T. Ressler, K. Zimmermann, D. Lazar, M.G. Wood, K.V. Wood, T.A. Kirkland, L.P. Encell, T. Machleidt, M.L. Dart, Toward a Point-of-Need Bioluminescence-Based Immunoassay Utilizing a Complete Shelf-Stable Reagent, Anal. Chem. 93 (2021) 5177–5184. https://doi.org/10.1021/acs.analchem.0c05074.[18] M. Baek, F. DiMaio, I. Anishchenko, J. Dauparas, S. Ovchinnikov, G.R. Lee, J. Wang, Q. Cong, L.N. Kinch, R.D. Schaeffer, C. Millán, H. Park, C. Adams, C.R. Glassman, A. DeGiovanni, J.H. Pereira, A.V. Rodrigues, A.A. van Dijk, A.C. Ebrecht, D.J. Opperman, T. Sagmeister, C. Buhlheller, T. Pavkov-Keller, M.K. Rathinaswamy, U. Dalwadi, C.K. Yip, J.E. Burke, K.C. Garcia, N.V. Grishin, P.D. Adams, R.J. Read, D. Baker, Accurate prediction of protein structures and interactions using a three-track neural network, Science 373 (2021) 871–876. https://doi.org/10.1126/science.abj8754.[19] The PyMOL Molecular Graphics System, Version 3.0 Schrödinger, LLC., (n.d.).[20] Humphrey, W., Dalke, A. and Schulten, K., “VMD - Visual Molecular Dynamics”, J. Molec. Graphics, 1996, vol. 14, pp. 33-38., (n.d.).[21] D.A. Case, H.M. Aktulga, K. Belfon, Ben-Shalom, IY, Berryman, JT, Brozell, SR, Cerutti, DS, Cheatham, TE, III, Cisneros, GA, Cruzeiro, VWD, Darden, TA, Duke, RE, Giambasu, G., Gilson, MK, Gohlke, H., Goetz, AW, Harris, R., Izadi, S., Izmailov, SA,… Kollman, PA (2022).[22] J.C. Gordon, J.B. Myers, T. Folta, V. Shoja, L.S. Heath, A. Onufriev, H++: a server for estimating p Ka s and adding missing hydrogens to macromolecules, Nucleic Acids Research 33 (2005) W368–W371. https://doi.org/10.1093/nar/gki464.[23] J.-P. Ryckaert, G. Ciccotti, H.J.C. Berendsen, Numerical integration of the cartesian equations of motion of a system with constraints: molecular dynamics of n -alkanes, Journal of Computational Physics 23 (1977) 327–341. https://doi.org/10.1016/0021-9991(77)90098-5.[24] U. Essmann, L. Perera, M.L. Berkowitz, T. Darden, H. Lee, L.G. Pedersen, A smooth particle mesh Ewald method, The Journal of Chemical Physics 103 (1995) 8577–8593. https://doi.org/10.1063/1.470117.[25] R.J. Loncharich, B.R. Brooks, R.W. Pastor, Langevin dynamics of peptides: The frictional dependence of isomerization rates of N-acetylalanyl-N′-methylamide, Biopolymers 32 (1992) 523–535. https://doi.org/10.1002/bip.360320508.[26] D.R. Roe, T.E.I. Cheatham, PTRAJ and CPPTRAJ: Software for Processing and Analysis of Molecular Dynamics Trajectory Data, J. Chem. Theory Comput. 9 (2013) 3084–3095. https://doi.org/10.1021/ct400341p.[27] D.A. Case, H.M. Aktulga, K. Belfon, D.S. Cerutti, G.A. Cisneros, V.W.D. Cruzeiro, N. Forouzesh, T.J. Giese, A.W. Götz, H. Gohlke, S. Izadi, K. Kasavajhala, M.C. Kaymak, E. King, T. Kurtzman, T.-S. Lee, P. Li, J. Liu, T. Luchko, R. Luo, M. Manathunga, M.R. Machado, H.M. Nguyen, K.A. O’Hearn, A.V. Onufriev, F. Pan, S. Pantano, R. Qi, A. Rahnamoun, A. Risheh, S. Schott-Verdugo, A. Shajan, J. Swails, J. Wang, H. Wei, X. Wu, Y. Wu, S. Zhang, S. Zhao, Q. Zhu, T.E.I. Cheatham, D.R. Roe, A. Roitberg, C. Simmerling, D.M. York, M.C. Nagan, K.M.Jr. Merz, AmberTools, J. Chem. Inf. Model. 63 (2023) 6183–6191. https://doi.org/10.1021/acs.jcim.3c01153.[28] P.J. Turner, XMGRACE, Version 5.1. 19, Center for Coastal and Land-Margin Research, Oregon Graduate Institute of Science and Technology, Beaverton, OR 2 (2005) 19.[29] T. Williams, C. Kelley, H.B. Bröker, J. Campbell, R. Cunningham, D. Denholm, E. Elber, R. Fearick, C. Grammes, L. Hart, Gnuplot 4.5: an interactive plotting program. 2011, URL Http://Www. Gnuplot. Info 56 (2017).[30] J. Stourac, O. Vavra, P. Kokkonen, J. Filipovic, G. Pinto, J. Brezovsky, J. Damborsky, D. Bednar, Caver Web 1.0: identification of tunnels and channels in proteins and analysis of ligand transport, Nucleic Acids Research 47 (2019) W414–W422. https://doi.org/10.1093/nar/gkz378.[31] J. Filipovic, O. Vavra, J. Plhak, D. Bednar, S.M. Marques, J. Brezovsky, L. Matyska, J. Damborsky, CaverDock: A Novel Method for the Fast Analysis of Ligand Transport, IEEE/ACM Trans Comput Biol Bioinform 17 (2020) 1625–1638. https://doi.org/10.1109/TCBB.2019.2907492.[32] S. Kim, J. Chen, T. Cheng, A. Gindulyte, J. He, S. He, Q. Li, B.A. Shoemaker, P.A. Thiessen, B. Yu, L. Zaslavsky, J. Zhang, E.E. Bolton, PubChem 2025 update, Nucleic Acids Research 53 (2025) D1516–D1525. https://doi.org/10.1093/nar/gkae1059.[33] H.J. Bohórquez, C.F. Suárez, M.E. Patarroyo, Mass & secondary structure propensity of amino acids explain their mutability and evolutionary replacements, Sci Rep 7 (2017) 7717. https://doi.org/10.1038/s41598-017-08041-7. Legends: Fig. 1: Structure of NLuc. Eleven antiparallel β-strands (yellow) aligned into β-barrel which is capped by 4 α-helices (red). SmBiT is a peptide with lower affinity, while HiBiT is a peptide with higher affinity for the large fragment LgBiT. The large fragment of tri-part NLuc is called Δ11S. Fig. 2: The image depicts the RMSD of the open form, closed form, 11S, Δ11S, Δ11S-β9, NanoBiT, NanoHiBiT and, Tri-part-NLuc over the course of a 100-nanosecond simulation. Fig. 3: The image depicts the RMSF of the molecules open form, closed form, 11S, Δ11S, Δ11S-β9, NanoBiT, NanoHiBiT and, Tri-part-NLuc during 100 ns of MD simulation. Fig. 4: The secondary structure content of the A) open form, B) 11S, C) D11S, and D) D11S-β9 for the residues 26-41 during the simulation. The α-helix content in this region replaced by bend in the incomplete NLuc structures. Fig. 5: The superposition of the the first cluster of open and closed conformation. The surface pocket residues in the closed conformation are shown in red, while those in open conformation are in blue. Fig. 6: Secondary structure diagrams of enzymes A) open form and B) closed form in the left panel, reveal the loss of β-strand structures in regions S6 and S7. Schematic illustrations of enzymes A) open form and B) closed form, highlighting amino acids Y94 (blue) and H93 (yellow) in the righr panel. Fig. 7: The distance plot between amino acids Y94, H93 or P93 (in NanoBiT) and the key catalytic amino acid R162 was analyzed to examine the orientation of these residues in various NLuc enzyme structures. As observed, the distance between Y94 and R162 in the open form and NanoBiT structures is the shortest compared to the distance between H93/P93 and R162. This indicates that Y94 is positioned inside the tunnel, while H93/P93 is located in the external pocket. Conversely, for the closed NLuc structure, the results are reversed, showing that Y94 is oriented in the external pocket and H93/P93 resides in the central tunnel during the simulation time of 100 ns. Fig. 8: The tunnel configurations of the first cluster of the various forms of NLuc. A) open form, B) closed form, C) NanoBiT, D) NBiT, E) NanoHiBiT, and F) Tri-part-NLuc. As shown, the main tunnel located between helices H2 and H3 is observed in the structures of open form and closed form, while this tunnel is closed in the structures of NanoBiT, NBiT, NanoHiBiT, and Tri-part-NLuc. Additionally, the secondary tunnel near helix H4 is clearly observed in the structures of NanoBiT and Tri-part-NLuc. Fig. 9: the distance analysis between helices H2 and H3. A) the distances between H2 and H3 in open,closed and NanoBiT structures are highlighted in red. B) the distance graph for H2 and H3 for open,closed and NanoBiT during 100 ns of MD simulation. Fig. 10: Docking results of peptides β9 and β10 with the large fragment Δ11S, as well as with the respective complexes Δ11S-β10 and Δ11S-β9. The docking structures for each component are depicted at the top of their respective columns. The red color represents the docked peptides. Fig. 11: The RMSD graph of the 5B0U-FMA complex at the active site (ac) and surface pocket (sp), as well as the 5IBO complex with FMZ during 200 ns MD similation. The graph shows that the most significant structural changes in the enzyme occurred during the latter stages of the simulation when FMA was bound to the surface of the enzyme. Fig. 12: Time-lapse snapshots (1000 frames) of the surface pucket docked 5B0U-FMA complex during simulation. The red molecule represents the initial binding position of FMA to the surface pocket, while the remaining colored molecules illustrate 10 snapshots of the complex’s interaction with the enzyme. As shown, FMA accumulates near the tunnel entrance (brown) over the course of the 200-nanosecond simulation. Fig. 13: The RMSF graphs compare the structural flexibility of A) 5B0U with FMA bound to the surface pocket (sp) and active side (ac), as well as the structure of 5IBO with FMZ bound to the opening of the central cavity. Additionally, B) it illustrates the structure of 5B0U alone, with FMA bound to the surface pocket (sp) and active side (ac), and C) the structure of 5IBO both alone and with FMA bound to the central cavity opening. Fig. 14: The figure shows A) the first two clusters of the 5BOU structure in the presence and absence of FMA in the central tunnel, as well as B) the first cluster from the 5IBO simulation alone for comparison of the surface pocket (cramped pocket in blue and spacious pocket in red). These results demonstrate an increase in available space (similar to 5IBO) in the surface pocket of the 5BOU structure due to the presence of FMA. Additionally, the graph obtained from cpptraj confirms this increase by showing an increased distance between residues H57 and H93. Fig. 15: The tunnels of the first cluster for the various structures of NLuc. A) 5B0U, B) 5IBO, C) NanoBiT, D) NanoHiBiT, E) Tri-part-NLuc, F) 5IBO-FMZ, and G) NanoBiT A35G mutant. Fig. 16. The binding enery of FMZ while passign through the active site of various NLuc structures. The second tunnel of tri-part NLuc is evaluated in order to lack of main tunnel. Information & Authors Information Version history Peer review timeline Published Journal of Molecular Graphics and Modelling Version of Record1 Jan 2026Published Copyright This work is licensed under a Non Exclusive No Reuse License.

Authors Metrics & Citations Metrics Article Usage 215views 192downloads Citations Download citation Mina Oliayi, Rahman Emamzadeh, Mohamad Reza Ganjalikhany. In Silico Investigation of Molecular Mechanisms Underlying the Function of NLuc and its Variants. Authorea. 17 February 2025. DOI: https://doi.org/10.22541/au.173980499.98679318/v1 DOI: https://doi.org/10.22541/au.173980499.98679318/v1 If you have the appropriate software installed, you can download article citation data to the citation manager of your choice. Simply select your manager software from the list below and click Download. For more information or tips please see 'Downloading to a citation manager' in the Help menu.