Skip to main content
NCBI home page
Protein Science : A Publication of the Protein Society logoLink to Protein Science : A Publication of the Protein Society
. 2026 May 23;35(6):e70634. doi: 10.1002/pro.70634

Investigating the structural basis of surfactant protein D oligomerization: The crucial role of Cys15‐Cys15/Cys20‐Cys20 disulfide bond linkage

Deng Li 1,2, Mona S Minkara 1,✉
PMCID: PMC13239783  PMID: 42175711

Abstract

Surfactant protein D (SP‐D) plays a vital role in innate immunity by recognizing and binding to glycans on the surfaces of various pathogens. The oligomerization of SP‐D into dodecameric structures enhances its pathogen‐binding and antiviral activities. Central to this oligomerization are two conserved cysteine residues (Cys15 and Cys20) in the N‐terminal domain, which form disulfide bonds essential for assembly. However, the precise disulfide bond linkages within the N‐terminal domain remain unclear, limiting our understanding of SP‐D's assembly mechanism at the molecular level. In this study, we investigated the four possible disulfide linkage types of the SP‐D N‐terminal domain to elucidate their effects on structure. Using replica exchange molecular dynamics (REMD) simulations, we extensively sampled the conformational space of each linkage type to identify predominant conformations. These representative structures were used to model SP‐D dodecamer assembly. Extensive unbiased molecular dynamics (MD) simulations with three independent replicates per linkage type assessed the structural stability of each dodecamer model. Our results indicate that the disulfide linkage, where disulfide bonds form between Cys15/Cys20 residues of adjacent monomers, leads to a stable dodecamer structure most consistent with experimental observations from atomic force microscopy (AFM). In contrast, the other three linkage types showed greater structural instability and did not align with experimental data. These findings provide molecular‐level insights into SP‐D assembly, suggesting that the disulfide linkage between Cys15/Cys20 residues is strongly supported as the physiologically relevant configuration for dodecamer formation, enhancing our understanding of SP‐D's role in innate immunity and informing future therapeutic strategies targeting SP‐D function.

Keywords: disulfide linkage, oligomerization, replica exchange molecular dynamics, surfactant protein D

1. INTRODUCTION

Surfactant protein D (SP‐D) is a key component of the innate immune system, central to the recognition and clearance of diverse pathogens, including viruses, bacteria, fungi, and pollen within the pulmonary environment (Allen et al., 1999; Hartshorn et al., 1998; Madan et al., 1997). By binding to specific carbohydrate patterns on these pathogens in a calcium‐dependent manner, SP‐D facilitates their agglutination and removal from the respiratory tract, thereby helping to maintain immune homeostasis (Colmorten et al., 2019; Griese, 1999; Kishore et al., 2006). Through engagement with pathogen‐associated molecular patterns (PAMPs), SP‐D not only enhances the efficiency of pathogen clearance but also modulates immune responses, underscoring its multifaceted role in host defense (Alonso‐Ojembarrena et al., 2025; Bhatt et al., 2024; Crouch, Kuan, & Persson, 1994; Hartl & Griese, 2006; Holmskov et al., 2003).

A critical factor influencing SP‐D's immune functions is its oligomeric state. SP‐D exhibits multiple oligomeric forms, with the dodecameric structure being the most prevalent and functionally significant (Arroyo et al., 2018; White et al., 2008). This dodecamer, often referred to as a cruciform due to its cross‐like appearance, comprises four SP‐D trimers assembled through interactions at their N‐terminal domains as shown in Figure 1a (Brown‐Augsburger, Chang, et al., 1996; Crouch, Persson, et al., 1994; Persson et al., 1989). Each SP‐D trimer comprises three monomers whose collagen‐like domains coil together to form a triple‐helical stalk (Figure 1b). Along each monomer, four distinct regions are arranged linearly from the N‐terminus to the C‐terminus: an N‐terminal domain harboring two conserved cysteine residues that mediate inter‐chain disulfide bonding, a collagen‐like domain responsible for trimerization, a short neck region, and a C‐terminal carbohydrate recognition domain (CRD) that binds pathogen‐associated glycans in a calcium‐dependent manner (Figure 1b). The oligomerization of SP‐D enhances its pathogen‐binding capacity by increasing the valency of its CRDs, thereby improving its ability to cross‐link and neutralize pathogens (Crouch, Persson, et al., 1994). Notably, the SP‐D dodecamer demonstrates a more efficient viral inhibition capability compared to its trimeric counterpart, underscoring the importance of its higher‐order structure in immune defense mechanisms (Arroyo et al., 2021; Crouch et al., 2011).

FIGURE 1.

FIGURE 1

Open in a new tab

Schematic illustration of the molecular architecture of Surfactant protein D (SP‐D). (a) Cruciform SP‐D is assembled from four trimeric SP‐D subunits. (b) Each trimeric SP‐D subunit consists of three monomers, and each monomer contains four major structural domains: the N‐terminal domain, collagen‐like domain, neck region, and carbohydrate recognition domain (CRD).

Central to the oligomerization process of SP‐D are two conserved cysteine residues located in the N‐terminal domain, Cys15 and Cys20 (Brown‐Augsburger, Chang, et al., 1996; Crouch, Persson, et al., 1994; Persson et al., 1989). These cysteines are essential for forming disulfide bonds that stabilize the assembly of SP‐D monomers into trimers and further into dodecamers. The precise arrangement of these disulfide bonds within the N‐terminal domain is crucial for the proper folding and function of SP‐D. However, despite extensive research, the detailed disulfide bond linkages in both trimeric and dodecameric forms of SP‐D remain unclear.

Disulfide bonds formed between cysteine residues play a fundamental role in stabilizing protein quaternary structures across a wide range of biological systems. Beyond structural stabilization, Cys‐Cys linkages are also key determinants of protein folding, self‐assembly, and redox‐dependent regulation. Reviews of disulfide‐rich proteins have emphasized that disulfide connectivity can strongly influence folding pathways and final structural outcomes, especially in secreted and multimeric proteins (Fass & Thorpe, 2018; Mainathambika & Bardwell, 2008). Experimental and computational studies of disulfide‐containing proteins have further shown that the specific cysteine‐pairing pattern can substantially affect protein stability, oligomeric organization, and biological activity (Manteca et al., 2017; Morand et al., 2021; Mossuto et al., 2011; Welker, Wedemeyer, Narayan, & Scheraga, 2001). Computational work has likewise shown that even a single disulfide bridge can reshape the accessible folding landscape and conformational stability of a protein assembly (Morand et al., 2021). In the context of pathological aggregation, aberrant or altered disulfide bonding has been implicated in amyloid and prion‐related structural transitions (Mossuto et al., 2011; Welker, Wedemeyer, & Scheraga, 2001). Together, these observations highlight the importance of defining the correct disulfide connectivity when interpreting the structure and function of multimeric proteins such as SP‐D.

Previous studies have highlighted the importance of these cysteines in SP‐D assembly. Guo et al. indicated that the two cysteines are critical to the SP‐D dodecamer and that there are inter‐chain disulfide bonds in SP‐D trimers (Guo et al., 2008). Their data also suggested that within SP‐D, these two cysteines exist at least partially in a reduced state (not forming a disulfide bond). Experimental mutagenesis studies have further demonstrated that substitution of Cys15 and/or Cys20 with serine residues disrupts dodecamer formation, confirming the essential role of these specific disulfide bonds in SP‐D oligomerization (Brown‐Augsburger, Hartshorn, et al., 1996; Guo et al., 2008). Similar observations have been reported for other members of the collectin family, such as surfactant protein A (SP‐A) and mannose‐binding lectin (MBL), where conserved N‐terminal cysteine residues serve analogous roles in higher‐order oligomeric assembly, suggesting a conserved mechanism across the collectin superfamily (Håkansson & Reid, 2000; Jensen et al., 2005; Palaniyar et al., 2001). Providing molecular‐level information on the N‐terminal domain of SP‐D would therefore be helpful to better understand the mechanism of SP‐D assembly. Such knowledge could guide future efforts to modulate SP‐D function, for example, by informing therapeutic strategies that harness or enhance its innate immune properties.

In this study, we aim to elucidate the structural characteristics of the SP‐D N‐terminal domain by systematically investigating the four possible disulfide linkages based on biochemical evidence from previous studies (Arroyo et al., 2018; Guo et al., 2008). The studies demonstrated that Cys15 and Cys20 form inter‐chain disulfide bonds in SP‐D trimers and that these cysteines exist at least partially in a reduced state. Given that each SP‐D trimer consists of three monomers (each contributing one Cys15 and one Cys20), and considering the trimeric symmetry, the four configurations we investigated represent the most plausible arrangements where each monomer participates equivalently in disulfide bonding. We performed enhanced sampling replica exchange molecular dynamics (REMD) simulations to extensively sample the conformational space of each linkage type. To assess the stability and dynamic behavior of the assembled dodecamers, we further carried out extensive unbiased molecular dynamics (MD) simulations for each disulfide linkage in three replicates. By analyzing the structural stabilities like root‐mean‐square deviation (RMSD) and radius of gyration (Rg), and constructing free energy landscapes (FELs), we evaluated the conformational preferences and stability associated with each linkage type. Furthermore, we compared our computational models with experimental observations from atomic force microscopy (AFM) studies to validate our findings and identify which disulfide bond linkage results in a dodecamer structure consistent with known SP‐D geometries (Arroyo et al., 2018).

This study leverages enhanced sampling MD simulation to explore the structural dynamics of SP‐D's N‐terminal domain. By identifying the disulfide bond linkage most consistent with experimental data for producing a stable dodecamer, our research addresses a critical gap in the understanding of SP‐D's structural biology, providing molecular‐level insights into the N‐terminal domain's disulfide linkages and their impact on oligomerization.

2. RESULTS AND DISCUSSION

2.1. Structural analysis of SP‐D trimer N‐terminal domain using replica exchange molecular dynamics

To investigate the impact of different disulfide bond linkages on the structure of the SP‐D trimer N‐terminal domain, we employed REMD simulations. This enhanced sampling method enabled extensive exploration of the conformational space, capturing the unique structural characteristics associated with each of the four possible disulfide bond linkages. Our REMD simulations revealed that each disulfide bond linkage induces distinct structural conformations in the SP‐D trimer N‐terminal domain.

To systematically characterize the conformational ensembles obtained from the REMD simulations, we performed clustering analysis using CPPTRAJ on the SP‐D trimer N‐terminal domain for each linkage type. The population distributions of the top three clusters for each linkage are illustrated in Figure 2. For the first linkage, we observed a predominant conformation, with Cluster 1 accounting for 3.8% of the total population, as shown in Figure 2a. The remaining clusters had significantly lower populations of 2.1% and 1.9%. Notably, the most populated cluster for the first linkage exhibited rich β‐strand secondary structures in the region between Cys15 and Cys20, as shown in Figure 3a. The corresponding secondary structure analysis for the dominant cluster is depicted in Figure 3, showing that the first linkage predominantly maintains well‐defined β‐strand secondary structures between residues Cys15 and Cys20, while other linkages exhibit a mix of random coil and helical structures with less pronounced β‐strands.

FIGURE 2.

FIGURE 2

Open in a new tab

Representative conformations of trimeric Surfactant protein D (SP‐D) N‐terminal domains with different intramolecular Cys15‐Cys20 disulfide linkages. Each panel (a‐d) shows the disulfide pattern (left) and the top three structural clusters (right) from replica exchange molecular dynamics (REMD) simulations. The percentages indicate the relative populations of each cluster.

FIGURE 3.

FIGURE 3

Open in a new tab

Secondary structure analysis of the Surfactant protein D (SP‐D) N‐terminal domain with different disulfide linkage patterns: (a) Cys15‐Cys15 and Cys20‐Cys20, (b) two Cys15‐Cys20, (c) Cys15‐Cys15 and Cys15‐Cys20, and (d) Cys20‐Cys20 and Cys15‐Cys20. The residue‐wise secondary structures for the top three clusters from replica exchange molecular dynamics (REMD) simulations are shown, classified as β‐strand (yellow), helix (blue), random‐coil (gray), or turn (green). The x‐axis indicates cluster numbers, and the y‐axis corresponds to residue numbers, highlighting structural variations driven by disulfide linkages.

In contrast, the other three linkages displayed more evenly distributed populations among the top clusters (Figure 2b–d). The second, third, and fourth linkages all exhibited evenly distributed populations across the top clusters, with each cluster accounting for approximately 1.6%–2.4% of the total population. This even distribution indicates that the conformations sampled were diverse, without any single dominant structural state. The lack of a predominant cluster suggests that the trimer structures for these linkages may be more flexible, lacking the specific structural features seen in the first linkage. Unlike the first linkage, no specific secondary structures were prominently observed in the corresponding conformations. For the second, third, and fourth linkages, the secondary structure analysis (Figure 3b–d) revealed a mix of random coil and helical structures with no dominant secondary structure elements, further highlighting the structural flexibility of these linkages compared to the well‐defined β‐strand structures seen in the first linkage.

Quantitative analysis of the secondary‐structure composition of the dominant clusters further supports these trends, as shown in Figures S1–S4. Among the four linkage types, the first linkage shows the strongest β‐strand propensity, with β‐strand populations reaching 15.0% in clusters 1 and 3. Importantly, although clusters 1 and 3 have the same overall β‐strand percentage, cluster 1 is particularly notable because Cys15 and Cys20 are located within the β‐strand region, indicating a more structurally organized arrangement around the disulfide‐linked cystines. In comparison, the second and third linkages exhibit more modest β‐strand content, ranging from 3.3% to 10.0%, whereas the fourth linkage shows minimal β‐strand character at only 1.7%–3.3%. Random coil remains the dominant structural feature in all linkage types, but is especially pronounced in the fourth linkage, where it accounts for 63.3%–72.5% of the population across clusters. Turn content is also substantially enriched in the second, third, and fourth linkages, reaching as high as 29.2%, 36.7%, and 33.3%, respectively, consistent with a greater degree of conformational disorder. Helical content is uniformly low across all systems, remaining between 0% and 5.8%.

2.2. Structural characterization of dodecamer SP‐D N‐terminal domain

To investigate the impact of different disulfide bond linkages on SP‐D assembly from trimers to dodecamers, we used the most representative conformation from cluster 1 of each linkage to construct the SP‐D dodecamer in N‐terminal domain by protein‐protein docking. This process was based on the biochemical basis that SP‐D dodecamers form through non‐bonded interactions at the N‐terminal domains of four SP‐D trimers. Figure S5 presents the docking results for all four linkages. Notably, the geometry of the dodecamer N‐terminal domain for the first linkage resembled a cruciform shape, closely aligning with experimental observations (Arroyo et al., 2018). In contrast, the geometries of the other linkage types deviated considerably from the cruciform shape, suggesting that these structures may be less favorable for proper dodecamer assembly.

To further investigate the effects of the different linkages on dodecamer assembly, extensive unbiased MD simulations were performed in three independent replicates per linkage type to assess their structural stability. We assessed the overall structural stability of each dodecamer by calculating the RMSD and Rg relative to the initial structure. As shown in Figure S6, the dodecamer structures corresponding to the first and third linkages remained relatively stable, with average RMSD values of 13.75 ± 2.90 Å and 8.33 ± 1.39 Å, and average Rg values of 20.98 ± 0.64 Å and 24.21 ± 0.61 Å, respectively. In contrast, the structures for the second and fourth linkages exhibited greater fluctuations, indicating instability. The fourth linkage showed average RMSD and Rg values of 10.60 ± 2.48 Å and 26.14 ± 1.75 Å, suggesting it was the most unstable and adopted a looser conformation. The second linkage displayed the largest RMSD and Rg values of 21.91 ± 7.48 Å and 29.54 ± 5.79 Å, indicating increased instability.

It is important to note that a low RMSD value, as observed for the third linkage, indicates structural stability relative to its initial docked conformation but does not necessarily imply that the adopted geometry corresponds to the experimentally observed cruciform shape. The third linkage maintains a compact structure (low Rg) that is dynamically stable, yet this compact conformation adopts a non‐cruciform arrangement that is geometrically inconsistent with AFM observations. Thus, structural stability as measured by RMSD is a necessary but not sufficient condition for identifying the physiologically relevant linkage; the geometry of the stable conformation must also match experimental constraints. This distinction between dynamical stability and geometric correctness is further supported by the inter‐arm angle analysis presented in Section 2.4.

2.3. Free energy landscape (FEL) of dodecamer SP‐D N‐terminal domain

To further explore the conformational states of the SP‐D dodecamer N‐terminal domain with different disulfide bond linkages, we constructed relative FELs based on RMSD and Rg. The FELs provide insights into the stability and conformational preferences of the dodecamer for each linkage type.

The FEL for the first disulfide linkage revealed three major local minima (Figure 4a), with the corresponding conformations depicted in Figure 4b. These minima were located at RMSD values ranging from 11 to 16 Å and Rg values between 20.4 and 21.4 Å. State A represents a more compact conformation, while states B and C exhibit slightly more extended structures. The narrow Rg range across all three minima suggests that the dodecamer generally maintained a consistent and stable structure throughout the simulation. The conformations corresponding to each of these local minima displayed geometries consistent with experimental observations, particularly the cruciform shape seen in AFM studies (Arroyo et al., 2018).

FIGURE 4.

FIGURE 4

Open in a new tab

Free energy landscape and representative conformations of the Surfactant protein D (SP‐D) N‐terminal domain with the first disulfide linkage type. (a) Free energy landscape as a function of the radius of gyration (Rg) and root‐mean‐square deviation (RMSD). Regions A, B, and C correspond to distinct energy basins. (b) Representative structures from regions A, B, and C, with their respective Rg values indicated. The free energy landscape was constructed from three independent replicates of unbiased molecular dynamics (MD) simulations.

In contrast, the FEL for the second linkage revealed numerous local minima, indicating multiple stable conformational states with varying degrees of stability and flexibility (Figure S7). These local minima, with RMSD values ranging from 15 to 25 Å and Rg values between 28 and 31 Å, highlight the structural diversity of the system. The presence of numerous local minima suggests a higher degree of structural flexibility compared to the first linkage, transitioning between compact and extended conformations. This is further supported by the observed gradual increase in Rg curve in Figure S6, indicating a shift from a compact to a more expanded conformation.

The FELs for the third and fourth linkages also revealed multiple local minima, but with fewer distinct states than the second linkage (Figures S8 and S9). The third linkage exhibited minima at RMSD values between 6.0 and 9.0 Å and Rg values from 24.1 to 24.5 Å. For the fourth linkage, the RMSD values ranged from 7.0 to 13.0 Å, and Rg values ranged from 25.0 to 27.5 Å. Although these landscapes were less rugged than that of the second linkage, they still indicated several conformational states whose geometries diverge substantially from the cruciform shape as observed in AFM.

These findings suggest that the first linkage promotes a more stable and consistent dodecamer structure with a geometry resembling the experimentally observed cruciform shape. The increased conformational variability in the other linkages may hinder proper assembly and function of the SP‐D dodecamer, as the N‐terminal domains do not adopt the necessary geometry for effective multimerization.

2.4. Comparison of dodecamer SP‐D N‐terminal domain structures with AFM observations

To further investigate the conformational characteristics of the dodecamer SP‐D N‐terminal domain in the first type of disulfide bond linkage, we analyzed representative conformations from the 1000 ns unbiased MD simulations via clustering analysis and compared them with previously reported AFM images as shown in Figure 5 (Arroyo et al., 2018).

FIGURE 5.

FIGURE 5

Open in a new tab

Comparison of dodecameric Surfactant protein D (SP‐D) N‐terminal domain structures with atomic force microscopy (AFM) observations for the first disulfide linkage type. (a) AFM images (left) and the corresponding representative conformations (right) illustrate the typical cruciform geometry. (b) An additional AFM image (left) and the corresponding representative conformations (right) illustrate an alternative geometry, in which three arms are positioned on one side and the remaining arm is positioned on the opposite side. The AFM images were reproduced from a study by Arroyo et al. with permission from Elsevier (Arroyo et al., 2018).

Predominant conformations 1 to 6 were consistent with AFM images shown in Figure 5a. These clusters, representing 61.5% of the conformations, exhibited the typical cruciform geometry observed experimentally. Conformation 6 and 7, representing 8% of conformations, showed a compact structure with three arms on one side and the remaining arm on the other side, corresponding to the AFM image in Figure 5b.

To quantify the geometry of the dodecamer SP‐D N‐terminal domain in the first linkage, we measured two specific angles defined by the positions of the trimeric arms, as described previously (Arroyo et al., 2018). The angle distributions in our simulations are shown in Figure 6. Our simulations revealed a populated angle distribution within 25–40°, consistent with the experimentally observed range. Additionally, we observed an angle distribution between 125130°, which was not commonly observed in AFM images.

FIGURE 6.

FIGURE 6

Open in a new tab

Arm angles distribution of the dodecameric Surfactant protein D (SP‐D) N‐terminal domain in the first disulfide linkage type. The two measured angles between trimeric arms show a predominant range of 25–40°, consistent with atomic force microscopy (AFM) observations (30–50°), and an additional range of 120–130°, which is less commonly observed in AFM data. The histogram represents the frequency of observed angles from simulations (n ≈ 50,000 × 3).

Several factors may contribute to this additional angle population at 125‐130°. First, this range may represent transient intermediate conformational states that the dodecamer adopts during dynamic fluctuations in solution; such short‐lived states would be difficult to capture in AFM imaging, where molecules are adsorbed onto surfaces and may be kinetically trapped in preferred orientations. Second, the inherent conformational flexibility of the SP‐D N‐terminal domain likely allows for a broader distribution of inter‐arm angles in solution than observed under the surface‐immobilized conditions of AFM experiments. Third, AFM tip‐sample interactions and substrate effects may bias the observed angle distributions toward certain preferred orientations, potentially underrepresenting geometries that are more common in solution. Finally, the substantially larger sampling in our simulations (n ≈ 50,000 × 3 frames) compared to AFM experiments (n ≈ 1000 particles) provides greater statistical power to detect less populated conformational states. For the second, third, and fourth linkages, the angle distributions (Figure S10a–c) did not correlate with experimental observations, further supporting the conclusion that these linkages are less likely to represent the native SP‐D state.

Based on our simulations and analyses, the first type of disulfide bond linkage is most consistent with current experimental observations of SP‐D dodecamers. It yields dodecamer geometries that closely match AFM data, while the other three linkages exhibit greater structural instability and do not match the observed geometries, indicating that they are less likely to represent the native state of SP‐D.

3. CONCLUSION

This study investigated the structural implications of the four possible disulfide linkages in the N‐terminal domain of SP‐D and their impact on the assembly and stability of SP‐D dodecamers. Utilizing REMD simulations and clustering analysis, we identified predominant conformations for each disulfide bond linkage type. These representative structures were employed to model the assembly of SP‐D dodecamers, followed by extensive 1000 ns unbiased MD simulations in three independent replicates to assess their structural stability and dynamic behavior.

Our findings indicate that the first disulfide bond linkage, involving disulfide bonds between Cys15/Cys20 residues of adjacent monomers, leads to a stable dodecamer structure most consistent with experimental observations from AFM. The FEL for this linkage displayed a well‐defined global minimum, suggesting a preferred and energetically favorable conformation. Geometric analyses revealed that the inter‐arm angles of the dodecamer in the first linkage closely align with those observed experimentally, exhibiting the characteristic cruciform geometry of SP‐D.

In contrast, the second, third, and fourth disulfide bond linkages demonstrated greater structural instability. Their FELs exhibited multiple shallow minima, indicating a lack of a dominant stable conformation. The inter‐arm angle distributions for these linkages did not match experimental observations, suggesting that these configurations are less likely to represent the physiological state of SP‐D.

The disulfide bond connectivity identified in this study has important implications beyond structural stability. The Cys15‐Cys15/Cys20‐Cys20 linkage pattern appears to confer the appropriate mechanical rigidity to the N‐terminal domain, enabling the dodecamer to maintain the cruciform geometry necessary for effective multivalent pathogen binding. Aberrant disulfide configurations, as represented by the other three linkage types, lead to geometries that would likely compromise the spatial arrangement of the CRDs, reducing the protein's ability to cross‐link and agglutinate pathogens. Furthermore, the correct disulfide linkage may be important for SP‐D's interactions with immune cell receptors, which require specific spatial presentations of the SP‐D arms for productive signaling. Disruption of the native disulfide pattern could therefore have downstream effects on immune modulation, potentially contributing to the pathological consequences observed in pulmonary diseases associated with SP‐D dysfunction.

Our computational investigations indicate that the disulfide bonds, formed by Cys15‐Cys15 and Cys20‐Cys20 with trimer SP‐D, are strongly supported as the configuration most consistent with the dodecamers SP‐D's structural characteristics observed by AFM. By elucidating the most plausible disulfide bond linkage in the SP‐D N‐terminal domain, this study provides valuable molecular‐level insights into the assembly mechanism of SP‐D. Understanding the structural basis of SP‐D oligomerization enhances our knowledge of its role in innate immunity. These insights may inform future therapeutic strategies to modulate SP‐D function, such as designing SP‐D oligomers that enhance its antiviral activities. Definitive experimental confirmation of the disulfide connectivity of the N‐terminal domain would further validate these computational predictions.

4. MATERIALS AND METHODS

4.1. Replica exchange molecular dynamics simulation for trimer SP‐D N‐terminal domain

4.1.1. Disulfide linkages in trimer SP‐D N‐terminal domain

The initial three‐dimensional structure of the SP‐D trimer N‐terminal domain was constructed using AlphaFold2 (Akdel et al., 2022; Jumper et al., 2021). To explore the influence of disulfide linkages on the trimeric SP‐D N‐terminal domain, we considered four possible disulfide bond configurations corresponding to different pairings of the conserved cysteine residues Cys15 and Cys20 within and between SP‐D monomers (Figure 7). These configurations are referred to as the first, second, third, and fourth linkages throughout this study.

FIGURE 7.

FIGURE 7

Open in a new tab

Schematic representation of the four possible disulfide linkages between Surfactant protein D (SP‐D) monomers. (a) First linkage: Disulfide bonds are formed between Cys15 and Cys15, and between Cys20 and Cys20 of adjacent monomers. (b) Second linkage: Disulfide bonds are formed between Cys15 and Cys20 of adjacent monomers. (c) Third linkage: Disulfide bonds are formed between Cys15 and Cys15 of adjacent monomers and between Cys15 and Cys20 of another adjacent pair. (d) Fourth linkage: Disulfide bonds are formed between Cys20 and Cys20, and between Cys15 and Cys20 of adjacent monomers.

We note that the N‐terminal domain of SP‐D lacks a resolved experimental structure, which motivated the use of AlphaFold2 for generating the initial model. The per‐residue predicted local distance difference test confidence scores for the N‐terminal region are shown in Figure S11, consistent with AlphaFold2's typical performance on intrinsically flexible or disordered regions (Akdel et al., 2022; Ruff & Pappu, 2021). Given these moderate confidence scores, the AlphaFold2 prediction was used solely as a starting point for extensive REMD simulations, which provided deep conformational sampling (~12.8 μs per linkage type across 64 temperature replicas). This approach substantially reduces dependence on the accuracy of the initial structure, as the enhanced sampling enables the system to explore a wide range of alternative conformations and identify thermodynamically favorable states independent of the starting model.

The four configurations are as follows:

  • First linkage: Disulfide bonds are formed between adjacent monomers at Cys15‐Cys15 and Cys20‐Cys20.

  • Second linkage: Disulfide bonds are formed between adjacent monomers at Cys15‐Cys20.

  • Third linkage: Disulfide bonds include Cys15‐Cys15 and Cys15‐Cys20 between adjacent monomers.

  • Fourth linkage: Disulfide bonds are formed at Cys20‐Cys20 and Cys15‐Cys20 between adjacent monomers.

4.1.2. Replica exchange molecular dynamics simulations

To extensively sample the conformational space of each disulfide bond linkage, REMD simulations were performed using the graphic processing unit (GPU)‐accelerated implementation of Amber22 (Case et al., 2022). For each linkage type, 64 replicas were generated, spanning a temperature range from 290 to 410 K (Patriksson & van der Spoel, 2008). The acceptance rates are in the reasonable range at around 30% to 35% for all the REMD simulations.

Prior to the REMD simulations, each system underwent energy minimization. Initially, harmonic restraints of 500 kcal mol−1 Å−2 were applied to all atoms to fix the protein structure, followed by an unrestrained minimization to remove any residual structural strains. After minimization, each replica was equilibrated at its corresponding temperature for 1 ns. In this process, the temperature of each replica was increased from 0 to the corresponding target temperature. The production REMD simulations were then conducted for approximately 200 ns per replica, resulting in an aggregate sampling time of ~12.8 μs for each linkage type.

We use AMBER ff19SB force field for protein and OPC water model with a solvation box extending 10 Å from any box edge to the nearest atom of the protein (Kirschner et al., 2008; Tian et al., 2020). Sodium and chloride ions were added to neutralize the system and achieve a physiological ionic strength of 0.15 M. In our simulations, disulfide bonds were modeled as static covalent bonds, consistent with the standard treatment in classical MD force fields, which do not allow for bond formation or breakage during the simulation. This approach models the structural consequences of pre‐formed disulfide bonds under physiological conditions, reflecting the oxidizing environment of the extracellular pulmonary space where SP‐D functions. We acknowledge that this approach does not capture the kinetics or thermodynamics of disulfide bond formation and exchange, and future studies employing reactive force fields or quantum mechanics/molecular mechanics methods could provide complementary insights into the redox‐dependent aspects of SP‐D assembly. Temperature control was maintained using the Langevin dynamics algorithm with a collision frequency of 1.0 ps−1, and pressure was controlled using the Berendsen barostat with a pressure relaxation time of 2 ps (Berendsen et al., 1984; Sindhikara et al., 2009; Uberuaga et al., 2004). The SHAKE algorithm was applied to constrain bonds involving hydrogen atoms, and long‐range electrostatic interactions were treated using the Particle Mesh Ewald (PME) method with a cutoff of 10 Å.

4.1.3. Clustering analysis

Following the REMD simulations, clustering analysis was performed using the CPPTRAJ module in AmberTools23 to identify predominant conformations for each linkage type (Case et al., 2023; Roe & Cheatham, 2013). The density‐based spatial clustering of applications with noise algorithm was employed with a cutoff of 3 Å, using the RMSD of the heavy atoms after alignment as the distance metric (Ester et al., 1996). The top three most populated clusters for each linkage type were identified, and the representative structure from the most populated cluster was extracted for subsequent modeling.

4.2. Molecular dynamics simulation for dodecamer SP‐D N‐terminal domain

The representative conformations from the most populated clusters of each linkage type were used as initial structures to model the assembly of the SP‐D dodecamer N‐terminal domain. The dodecamer structures were built using the HDOCK server, a leading platform for multimeric protein structure prediction (Yan et al., 2017; Yan et al., 2020). Active residues in the N‐terminal domain involved in oligomerization were specified during the protein‐protein docking process. HDOCK generated the top 100 ranked docking poses for each linkage type, which were evaluated based on both the HDOCK scoring function and visual inspection for structural plausibility. The top‐ranked structure for each linkage type was selected based on the HDOCK scoring function. While explicit symmetry constraints were not imposed, the definition of binding regions effectively directed the docking toward biologically plausible configurations. The subsequent extensive MD simulations served as a rigorous test of whether the selected docking poses could maintain stable conformations over microsecond timescales.

4.2.1. Molecular dynamics simulations

To evaluate the stability and dynamic behavior of the dodecamer structures over time, extensive unbiased MD simulations were performed for each linkage type. Each simulation system was first optimized through two stages of minimization: initially, strong harmonic restraints of 500 kcal mol−1 Å−2 were applied to all atoms in the protein structure to fix it, followed by an unrestrained minimization of the entire system. The systems were then equilibrated in two stages. In the first stage, the temperature was gradually raised from 0 to 310 K over 1 × 106 steps of MD with a timestep of 2 fs in the canonical ensemble, applying a weak harmonic restraint of 10 kcal·mol−1·Å−2 to the entire protein structure. In the second stage, the systems were equilibrated at a constant temperature of 310 K and 1 bar pressure for 100 ns. Temperature control was maintained using the Langevin dynamics algorithm with a collision frequency of 1.0 ps−1, and pressure was controlled using the Berendsen barostat with a pressure relaxation time of 2 ps (Berendsen et al., 1984; Sindhikara et al., 2009; Uberuaga et al., 2004). The SHAKE algorithm was used to constrain all hydrogen‐containing bonds. The production MD simulations were then conducted in the isothermal‐isobaric ensemble for 1000 ns in three independent replicates per linkage type. All simulations were performed using the GPU‐accelerated implementation of Amber22 (Case et al., 2022).

4.2.2. Structural analysis methods

MD trajectories were analyzed using CPPTRAJ to assess structural stability and conformational changes (D. A. Case et al., 2023; Roe & Cheatham, 2013). The RMSD of non‐hydrogen atoms relative to the initial structure was calculated to monitor structural stability over time. The Rg was computed to evaluate the compactness of the dodecamer structures. Secondary structure analysis was performed using the define secondary structure of proteins algorithm implemented in visual molecular dynamics (Humphrey et al., 1996; Kabsch & Sander, 1983). Additional clustering was conducted on the MD trajectories to identify dominant conformations during the simulations. To quantitatively assess the geometry of the dodecamer structures, four inter‐arm angles were measured following the description by Arroyo et al., and compared to observations from AFM (Arroyo et al., 2018). We note that the AFM images used for structural comparison were obtained from the published study by Arroyo et al., and no AFM experiments were conducted as part of the present study (Arroyo et al., 2018).

4.2.3. Free energy landscape

To visualize the conformational space explored by each linkage type, FELs were constructed using RMSD and Rg as collective variables. Two‐dimensional histograms were generated from the trajectory data, and relative free energies (ΔG) were calculated using the Boltzmann inversion of the probability densities:

∆Grelative=−RTlnPi

where R is the ideal gas constant (8.314 J mol−1 K−1), T is the temperature (310 K), and P i is the probability of RMSD and Rg coinciding in a randomly selected trajectory frame. The FELs are calculated using a grid of 200 × 200.

AUTHOR CONTRIBUTIONS

Deng Li: Conceptualization; investigation; writing – original draft; methodology; validation; visualization; writing – review and editing; data curation; formal analysis. Mona S. Minkara: Funding acquisition; writing – review and editing; project administration; resources; supervision; data curation.

CONFLICT OF INTEREST STATEMENT

The authors declare no conflict of interest.

Supporting information

Figure S1. Quantitative secondary structure composition of the top three clusters from replica exchange molecular dynamics (REMD) simulations for the first disulfide linkage type. The bar plot shows the percentage of β‐strand (yellow), helix (blue), random‐coil (gray), and turn (green) content for each cluster.

Figure S2. Quantitative secondary structure composition of the top three clusters from replica exchange molecular dynamics (REMD) simulations for the second disulfide linkage type. The bar plot shows the percentage of β‐strand (yellow), helix (blue), random‐coil (gray), and turn (green) content for each cluster.

Figure S3. Quantitative secondary structure composition of the top three clusters from replica exchange molecular dynamics (REMD) simulations for the third disulfide linkage type. The bar plot shows the percentage of β‐strand (yellow), helix (blue), random‐coil (gray), and turn (green) content for each cluster.

Figure S4. Quantitative secondary structure composition of the top three clusters from replica exchange molecular dynamics (REMD) simulations for the fourth disulfide linkage type. The bar plot shows the percentage of β‐strand (yellow), helix (blue), random‐coil (gray), and turn (green) content for each cluster.

Figure S5. Initial structures from protein–protein docking for different disulfide linkages.

Figure S6. Structural dynamics of the Surfactant protein D (SP‐D) N‐terminal domain with different disulfide linkages. (a) Root‐mean‐square deviation (RMSD) and (b) Radius of gyration (Rg) over 1 μs for the first disulfide linkage in black, second disulfide linkage in green, third disulfide linkage in red, and fourth disulfide linkage in blue. Curves represent the mean across three independent replicates of unbiased MD simulations.

Figure S7. Free energy landscape and representative conformations of the Surfactant protein D (SP‐D) N‐terminal domain with the second disulfide linkage type. Free energy landscape as a function of the radius of gyration (Rg) and root‐mean‐square deviation (RMSD).

Figure S8. Free energy landscape and representative conformations of the Surfactant protein D (SP‐D) N‐terminal domain with the third disulfide linkage type. Free energy landscape as a function of the radius of gyration (Rg) and root‐mean‐square deviation (RMSD).

Figure S9. Free energy landscape and representative conformations of the Surfactant protein D (SP‐D) N‐terminal domain with the fourth disulfide linkage type. Free energy landscape as a function of the radius of gyration (Rg) and root‐mean‐square deviation (RMSD).

Figure S10. Arm angles distribution of the dodecameric Surfactant protein D (SP‐D) N‐terminal domain in the (a) second disulfide linkage type, (b) third disulfide linkage type, and (c) fourth disulfide linkage type. The histogram represents the frequency of observed angles from simulations (n ≈ 50,000*3).

Figure S11. AlphaFold2 prediction confidence metrics for the Surfactant protein D (SP‐D) N‐terminal domain (residues 1–25).

PRO-35-e70634-s001.docx (2.3MB, docx)

ACKNOWLEDGMENTS

The authors appreciate the financial support from Northeastern University start‐up funds and National Science Foundation (NSF) CAREER Award(2338401). We are grateful to the Discovery Cluster of Northeastern University for computer time and facilities. This work also used Delta GPU at the National Center for Supercomputing Applications (NCSA) through allocation(CHE220072) from the Advanced Cyberinfrastructure Coordination Ecosystem: Services & Support (ACCESS) program, which is supported by National Science Foundation grants(2138259, 2138286, 2138307, 2137603, and 2138296) (Boerner et al., 2023). We acknowledge Dr. Elizabeth Andrews (E.A.) for making the materials accessible to M.S.M. Grammarly and ChatGPT (OpenAI, version 5.4) were used to improve the grammar, language, and wording of individual sentences in the main text during manuscript preparation. Neither tool was used to generate content for this article. The final manuscript was reviewed and verified by the authors.

Li D, Minkara MS. Investigating the structural basis of surfactant protein D oligomerization: The crucial role of Cys15‐Cys15/Cys20‐Cys20 disulfide bond linkage. Protein Science. 2026;35(6):e70634. 10.1002/pro.70634

Review Editor: Lynn Kamerlin

DATA AVAILABILITY STATEMENT

The data that support the findings of this study are available from the corresponding author upon reasonable request.

REFERENCES

  1. Akdel M, Pires DE, Pardo EP, Jänes J, Zalevsky AO, Mészáros B, et al. A structural biology community assessment of AlphaFold2 applications. Nat Struct Mol Biol. 2022;29(11):1056–1067. 10.1038/s41594-022-00849-w [DOI] [PMC free article] [PubMed] [Google Scholar]
  2. Allen MJ, Harbeck R, Smith B, Voelker DR, Mason RJ. Binding of rat and human surfactant proteins A and D to aspergillus fumigatus conidia. Infect Immun. 1999;67(9):4563–4569. 10.1128/IAI.67.9.4563-4569.1999 [DOI] [PMC free article] [PubMed] [Google Scholar]
  3. Alonso‐Ojembarrena A, Poindexter B, Aleem S, Healy H, Aguar‐Carrascosa M, Moliner‐Calderón E, et al. A phase 1b randomized, multicenter, dose determination trial of zelpultide alfa (recombinant human surfactant protein D) in preterm neonates at high risk of developing bronchopulmonary dysplasia. Front Pediatr. 2025;13:1639573. 10.3389/fped.2025.1639573 [DOI] [PMC free article] [PubMed] [Google Scholar]
  4. Arroyo R, Grant SN, Colombo M, Salvioni L, Corsi F, Truffi M, et al. Full‐length recombinant hSP‐D binds and inhibits SARS‐CoV‐2. Biomolecules. 2021;11(8):1114. 10.3390/biom11081114 [DOI] [PMC free article] [PubMed] [Google Scholar]
  5. Arroyo R, Martin‐Gonzalez A, Echaide M, Jain A, Brondyk WH, Rosenbaum J, et al. Supramolecular assembly of human pulmonary surfactant protein SP‐D. J Mol Biol. 2018;430(10):1495–1509. 10.1016/j.jmb.2018.03.027 [DOI] [PubMed] [Google Scholar]
  6. Berendsen HJC, Postma JPM, Vangunsteren WF, Dinola A, Haak JR. Molecular‐dynamics with coupling to an external bath. J Chem Phys. 1984;81(8):3684–3690. 10.1063/1.448118 [DOI] [Google Scholar]
  7. Bhatt R, Madsen J, Castillo‐Hernandez T, Chant K, Dehbi HM, Marlow N, et al. Recombinant fragment of human surfactant protein D to prevent neonatal chronic lung disease (RESPONSE): a protocol for a phase I safety trial in a tertiary neonatal unit. BMJ Open. 2024;14(8):e086394. 10.1136/bmjopen-2024-086394 [DOI] [PMC free article] [PubMed] [Google Scholar]
  8. Boerner TJ, Deems S, Furlani TR, Knuth SL, Towns J. ACCESS: advancing innovation: NSF's advanced cyberinfrastructure coordination ecosystem: services & support. Paper presented at the practice and experience in advanced research computing. Portland, OR, USA: Association for Computing Machinery; 2023. 10.1145/3569951.3597559 [DOI] [Google Scholar]
  9. Brown‐Augsburger P, Chang D, Rust K, Crouch EC. Biosynthesis of surfactant protein D ‐ Contributions of conserved NH2‐terminal cysteine residues and collagen helix formation to assembly and secretion. J Biol Chem. 1996;271(31):18912–18919. 10.1074/jbc.271.31.18912 [DOI] [PubMed] [Google Scholar]
  10. Brown‐Augsburger P, Hartshorn K, Chang D, Rust K, Fliszar C, Welgus HG, et al. Site‐directed mutagenesis of Cys‐15 and Cys‐20 of pulmonary surfactant protein D ‐ expression of a trimeric protein with altered anti‐viral properties. J Biol Chem. 1996;271(23):13724–13730. 10.1074/jbc.271.23.13724 [DOI] [PubMed] [Google Scholar]
  11. Case DA, Aktulga HM, Belfon K, Ben‐Shalom IY, Berryman JT, Brozell SR, et al. AMBER 2022. San Francisco: University of California; 2022. [Google Scholar]
  12. Case DA, Aktulga HM, Belfon K, Cerutti DS, Cisneros GA, Cruzeiro VWD, et al. The AmberTools. J Chem Inf Model. 2023;63(20):6183–6191. 10.1021/acs.jcim.3c01153 [DOI] [PMC free article] [PubMed] [Google Scholar]
  13. Colmorten KB, Nexoe AB, Sorensen GL. The dual role of surfactant protein‐D in vascular inflammation and development of cardiovascular disease. Front Immunol. 2019;10:2264. 10.3389/fimmu.2019.02264 [DOI] [PMC free article] [PubMed] [Google Scholar]
  14. Crouch E, Kuan SF, Persson A. Surfactant protein‐D ‐ possible roles in pulmonary host‐defense. Lung Surfactant: Basic Research in the Pathogenesis of Lung Disorders. 1994;27:132–140. [Google Scholar]
  15. Crouch E, Nikolaidis N, McCormack FX, McDonald B, Allen K, Rynkiewicz MJ, et al. Mutagenesis of surfactant protein D informed by evolution and X‐ray crystallography enhances defenses against influenza A virus In vivo. J Biol Chem. 2011;286(47):40681–40692. 10.1074/jbc.M111.300673 [DOI] [PMC free article] [PubMed] [Google Scholar]
  16. Crouch E, Persson A, Chang D, Heuser J. Molecular‐structure of pulmonary surfactant protein‐D (Sp‐D). J Biol Chem. 1994;269(25):17311–17319. [PubMed] [Google Scholar]
  17. Ester M, Kriegel H‐P, Sander J, Xu X. A density‐based algorithm for discovering clusters in large spatial databases with noise. Paper presented at the proceedings of the second international conference on knowledge discovery and data mining. Portland, Oregon: Association for Computing Machinery; 1996. [Google Scholar]
  18. Fass D, Thorpe C. Chemistry and enzymology of disulfide cross‐linking in proteins. Chem Rev. 2018;118(3):293–322. 10.1021/acs.chemrev.7b00123 [DOI] [PMC free article] [PubMed] [Google Scholar]
  19. Griese M. Pulmonary surfactant in health and human lung diseases: state of the art. Eur Respir J. 1999;13(6):1455–1476. 10.1183/09031936.99.13614779 [DOI] [PubMed] [Google Scholar]
  20. Guo CJ, Atochina‐Vasserman EN, Abramova E, Foley JP, Zaman A, Crouch E, et al. S‐Nitrosylation of surfactant protein‐D controls inflammatory function. PLoS Biol. 2008;6(11):2414–2423. 10.1371/journal.pbio.0060266 [DOI] [PMC free article] [PubMed] [Google Scholar]
  21. Håkansson K, Reid KBM. Collectin structure: A review. Protein Sci. 2000;9(9):1607–1617. 10.1110/ps.9.9.1607 [DOI] [PMC free article] [PubMed] [Google Scholar]
  22. Hartl D, Griese M. Surfactant protein D in human lung diseases. Eur J Clin Investig. 2006;36(6):423–435. 10.1111/j.1365-2362.2006.01648.x [DOI] [PubMed] [Google Scholar]
  23. Hartshorn KL, Crouch E, White MR, Colamussi ML, Kakkanatt A, Tauber B, et al. Pulmonary surfactant proteins A and D enhance neutrophil uptake of bacteria. Am J Phys. 1998;274(6):L958–L969. 10.1152/ajplung.1998.274.6.L958 [DOI] [PubMed] [Google Scholar]
  24. Holmskov U, Thiel S, Jensenius JC. Collectins and ficolins: humoral lectins of the innate immune defense. Annu Rev Immunol. 2003;21:547–578. 10.1146/annurev.immunol.21.120601.140954 [DOI] [PubMed] [Google Scholar]
  25. Humphrey W, Dalke A, Schulten K. VMD: Visual molecular dynamics. J Mol Graph Model. 1996;14(1):33–38. 10.1016/0263-7855(96)00018-5 [DOI] [PubMed] [Google Scholar]
  26. Jensen PH, Weilguny D, Matthiesen F, McGuire KA, Shi L, Hojrup P. Characterization of the oligomer structure of recombinant human mannan‐binding lectin. J Biol Chem. 2005;280(12):11043–11051. 10.1074/jbc.M412472200 [DOI] [PubMed] [Google Scholar]
  27. Jumper J, Evans R, Pritzel A, Green T, Figurnov M, Ronneberger O, et al. Highly accurate protein structure prediction with AlphaFold. Nature. 2021;596(7873):583–589. 10.1038/s41586-021-03819-2 [DOI] [PMC free article] [PubMed] [Google Scholar]
  28. Kabsch W, Sander C. Dictionary of protein secondary structure ‐ pattern‐recognition of hydrogen‐bonded and geometrical features. Biopolymers. 1983;22(12):2577–2637. 10.1002/bip.360221211 [DOI] [PubMed] [Google Scholar]
  29. Kirschner KN, Yongye AB, Tschampel SM, Gonzalez‐Outeirino J, Daniels CR, Foley BL, et al. GLYCAM06: A generalizable biomolecular force field. Carbohydrates. J Comput Chem. 2008;29(4):622–655. 10.1002/jcc.20820 [DOI] [PMC free article] [PubMed] [Google Scholar]
  30. Kishore U, Greenhough TJ, Waters P, Shrive AK, Ghai R, Kamran MF, et al. Surfactant proteins SP‐A and SP‐D: structure, function and receptors. Mol Immunol. 2006;43(9):1293–1315. 10.1016/j.molimm.2005.08.004 [DOI] [PubMed] [Google Scholar]
  31. Madan T, Eggleton P, Kishore U, Strong P, Aggrawal SS, Sarma PU, et al. Binding of pulmonary surfactant proteins A and D to aspergillus fumigatus conidia enhances phagocytosis and killing by human neutrophils and alveolar macrophages. Infect Immun. 1997;65(8):3171–3179. 10.1128/iai.65.8.3171-3179.1997 [DOI] [PMC free article] [PubMed] [Google Scholar]
  32. Mainathambika BS, Bardwell JC. Disulfide‐linked protein folding pathways. Annu Rev Cell Dev Biol. 2008;24:211–235. 10.1146/annurev.cellbio.24.110707.175333 [DOI] [PubMed] [Google Scholar]
  33. Manteca A, Alonso‐Caballero A, Fertin M, Poly S, De Sancho D, Perez‐Jimenez R. The influence of disulfide bonds on the mechanical stability of proteins is context dependent. J Biol Chem. 2017;292(32):13374–13380. 10.1074/jbc.M117.784934 [DOI] [PMC free article] [PubMed] [Google Scholar]
  34. Morand J, Nunes A, Faísca PFN. The folding space of protein β2‐microglobulin is modulated by a single disulfide bridge. Phys Biol. 2021;18(5):56001. 10.1088/1478-3975/ac08ec [DOI] [PubMed] [Google Scholar]
  35. Mossuto ME, Bolognesi B, Guixer B, Dhulesia A, Agostini F, Kumita JR, et al. Disulfide bonds reduce the toxicity of the amyloid fibrils formed by an extracellular protein. Angew Chem Int Ed Engl. 2011;50(31):7048–7051. 10.1002/anie.201100986 [DOI] [PMC free article] [PubMed] [Google Scholar]
  36. Palaniyar N, Ikegami M, Korfhagen T, Whitsett J, McCormack FX. Domains of surfactant protein A that affect protein oligomerization, lipid structure and surface tension. Comparative Biochemistry and Physiology a‐Molecular & Integrative Physiology. 2001;129(1):109–127. 10.1016/S1095-6433(01)00309-9 [DOI] [PubMed] [Google Scholar]
  37. Patriksson A, van der Spoel D. A temperature predictor for parallel tempering simulations. Phys Chem Chem Phys. 2008;10(15):2073–2077. 10.1039/b716554d [DOI] [PubMed] [Google Scholar]
  38. Persson A, Chang D, Rust K, Moxley M, Longmore W, Crouch E. Purification and biochemical‐characterization of Cp4 (Sp‐D), a collagenous surfactant‐associated protein. Biochemistry. 1989;28(15):6361–6367. 10.1021/bi00441a031 [DOI] [PubMed] [Google Scholar]
  39. Roe DR, Cheatham TE 3rd. PTRAJ and CPPTRAJ: software for processing and analysis of molecular dynamics trajectory data. J Chem Theory Comput. 2013;9(7):3084–3095. 10.1021/ct400341p [DOI] [PubMed] [Google Scholar]
  40. Ruff KM, Pappu R. AlphaFold and implications for intrinsically disordered proteins. J Mol Biol. 2021;433(20):167208. 10.1016/j.jmb.2021.167208 [DOI] [PubMed] [Google Scholar]
  41. Sindhikara DJ, Kim S, Voter AF, Roitberg AE. Bad seeds sprout perilous dynamics: stochastic thermostat induced trajectory synchronization in biomolecules. J Chem Theory Comput. 2009;5(6):1624–1631. 10.1021/ct800573m [DOI] [PubMed] [Google Scholar]
  42. Tian C, Kasavajhala K, Belfon KAA, Raguette L, Huang H, Migues AN, et al. ff19SB: amino‐acid‐specific protein backbone parameters trained against quantum mechanics energy surfaces in solution. J Chem Theory Comput. 2020;16(1):528–552. 10.1021/acs.jctc.9b00591 [DOI] [PMC free article] [PubMed] [Google Scholar]
  43. Uberuaga BP, Anghel M, Voter AF. Synchronization of trajectories in canonical molecular‐dynamics simulations: observation, explanation, and exploitation. J Chem Phys. 2004;120(14):6363–6374. 10.1063/1.1667473 [DOI] [PubMed] [Google Scholar]
  44. Welker E, Wedemeyer WJ, Narayan M, Scheraga HA. Coupling of conformational folding and disulfide‐bond reactions in oxidative folding of proteins. Biochemistry. 2001;40(31):9059–9064. 10.1021/bi010409g [DOI] [PubMed] [Google Scholar]
  45. Welker E, Wedemeyer WJ, Scheraga HA. A role for intermolecular disulfide bonds in prion diseases? Proc Natl Acad Sci USA. 2001;98(8):4334–4336. 10.1073/pnas.071066598 [DOI] [PMC free article] [PubMed] [Google Scholar]
  46. White M, Kingma P, Tecle T, Kacak N, Linders B, Heuser J, et al. Multimerization of surfactant protein D, but not its collagen domain, is required for antiviral and opsonic activities related to influenza virus. J Immunol. 2008;181(11):7936–7943. 10.4049/jimmunol.181.11.7936 [DOI] [PubMed] [Google Scholar]
  47. Yan YM, Tao H, He J, Huang SY. The HDOCK server for integrated protein‐protein docking. Nat Protoc. 2020;15(5):1829–1852. 10.1038/s41596-020-0312-x [DOI] [PubMed] [Google Scholar]
  48. Yan YM, Zhang D, Zhou P, Li BT, Huang SY. HDOCK: a web server for protein‐protein and protein‐DNA/RNA docking based on a hybrid strategy. Nucleic Acids Res. 2017;45(W1):W365–W373. 10.1093/nar/gkx407 [DOI] [PMC free article] [PubMed] [Google Scholar]

Associated Data

This section collects any data citations, data availability statements, or supplementary materials included in this article.

Supplementary Materials

Figure S1. Quantitative secondary structure composition of the top three clusters from replica exchange molecular dynamics (REMD) simulations for the first disulfide linkage type. The bar plot shows the percentage of β‐strand (yellow), helix (blue), random‐coil (gray), and turn (green) content for each cluster.

Figure S2. Quantitative secondary structure composition of the top three clusters from replica exchange molecular dynamics (REMD) simulations for the second disulfide linkage type. The bar plot shows the percentage of β‐strand (yellow), helix (blue), random‐coil (gray), and turn (green) content for each cluster.

Figure S3. Quantitative secondary structure composition of the top three clusters from replica exchange molecular dynamics (REMD) simulations for the third disulfide linkage type. The bar plot shows the percentage of β‐strand (yellow), helix (blue), random‐coil (gray), and turn (green) content for each cluster.

Figure S4. Quantitative secondary structure composition of the top three clusters from replica exchange molecular dynamics (REMD) simulations for the fourth disulfide linkage type. The bar plot shows the percentage of β‐strand (yellow), helix (blue), random‐coil (gray), and turn (green) content for each cluster.

Figure S5. Initial structures from protein–protein docking for different disulfide linkages.

Figure S6. Structural dynamics of the Surfactant protein D (SP‐D) N‐terminal domain with different disulfide linkages. (a) Root‐mean‐square deviation (RMSD) and (b) Radius of gyration (Rg) over 1 μs for the first disulfide linkage in black, second disulfide linkage in green, third disulfide linkage in red, and fourth disulfide linkage in blue. Curves represent the mean across three independent replicates of unbiased MD simulations.

Figure S7. Free energy landscape and representative conformations of the Surfactant protein D (SP‐D) N‐terminal domain with the second disulfide linkage type. Free energy landscape as a function of the radius of gyration (Rg) and root‐mean‐square deviation (RMSD).

Figure S8. Free energy landscape and representative conformations of the Surfactant protein D (SP‐D) N‐terminal domain with the third disulfide linkage type. Free energy landscape as a function of the radius of gyration (Rg) and root‐mean‐square deviation (RMSD).

Figure S9. Free energy landscape and representative conformations of the Surfactant protein D (SP‐D) N‐terminal domain with the fourth disulfide linkage type. Free energy landscape as a function of the radius of gyration (Rg) and root‐mean‐square deviation (RMSD).

Figure S10. Arm angles distribution of the dodecameric Surfactant protein D (SP‐D) N‐terminal domain in the (a) second disulfide linkage type, (b) third disulfide linkage type, and (c) fourth disulfide linkage type. The histogram represents the frequency of observed angles from simulations (n ≈ 50,000*3).

Figure S11. AlphaFold2 prediction confidence metrics for the Surfactant protein D (SP‐D) N‐terminal domain (residues 1–25).

PRO-35-e70634-s001.docx (2.3MB, docx)

Data Availability Statement

The data that support the findings of this study are available from the corresponding author upon reasonable request.

Articles from Protein Science : A Publication of the Protein Society are provided here courtesy of The Protein Society

RESOURCES