Access the full text.
Sign up today, get DeepDyve free for 14 days.
Lactoferrin (Lf) - member of the transferrin family of proteins responsible for many different functions in the body of mammals participates in regulation of free iron level in the body fluids making the protein bacteriostatic. The main goal of studies was to test the suitability of molecular dynamic simulation to study structural changes in the tertiary structure of lactoferrin. According to ConSurf Server analysis one of the most conservative amino acids was found not only in iron- but also in carbohydrates- binding pockets which may suggest a significant impact of carbohydrates on the functions performed by lactoferrin. Pocket-Finder program applied to find iron-binding pockets revealed the potential Fe binding area. The stability of the ligand deprived protein was verified performing the 50 ns dynamic simulation using the Gromacs program. The tertiary structure changes during the simulation were observed in N-lob solely. No structural changes were observed in C-lob iron-binding pocket. KEYWORDS: ligand binding site in proteins; lactoferrin; molecular dynamic simulation; Introduction Lactoferrin (called also Lf in this paper) is a member of the transferrin family of proteins it is a glycoprotein composed of about 690 residues and weighs about 80-kDa. Two globular homologous lactoferrin lobes are connected flexibly by -helices. Each lobe consists of two domains (Figure 1) with iron binding pocket in central part of the protein body. One of the most important lactoferrin features is that it participates in regulation of the levels of free iron in the body fluids of mammals which makes the protein bacteriostatic. Lactoferrin binds two Fe3+ ions together with two synergistically bound CO32- ions (K~ 1022 M). This process is reversible and pH dependent. Typically, iron is released in pH 3-4, compared with pH 5-6 for transferrin . Iron in the lactoferrin pockets (in N- and C-lobe iron pocket) creates four coordination bonds with residues: Asp, His and two residues of Tyr. Distances between the nearest iron pocket residues and CO 32- ion are longer than 2.5 Å. The main goal of studies was to test the suitability of molecular dynamic simulation to study structural changes in the tertiary structure of large protein. Lactoferrin was selected because it is a well-know protein which gives the opportunity to compare theoretical and experimental data. Materials and Methods Bovine lactoferrin structure The crystallographic structure of lactoferrin PDB id = 1BLF was selected from PDB  for studies. This structure is composed of 689 residues. Four domains can be distinguished in the lactoferrin structure (Table 1). Three carbohydrate are located on the C-lobe surface (Figure 1). Chemical component of carbohydrate are D-mannose (MAN), -D-mannose (BMA) and N-acetyl-D-glucosamine (NAG) . Figure 1. Bovine lactoferrin (PDB id 1BLF). Colors selected domains on tertiary structure draft as well as on the residue sequence. Table 1. Lactoferrin residues classification. *- residue Domain N1 A N1 B N2 C1 A C1 B C2 N-terminus* 5 251 92 340 595 433 C-terminus* 91 339 250 432 689 594 Programs Analysis of phylogenetic relationship between homologous sequences The ConSurf Server  was applied to analyze the phylogenetic relationship between homologous sequences of Lf. Current sequence databases contain Lf sequences from nine species: human, mouse, cow, horse, pig, goat, sheep, buffalo and camel. Similarity of human lactoferrin to Lf of another origin is located in the range from 65 to 100% . The ConSurf Server was enabled to find a conservative position in the homologous amino acid sequences of lactoferrin. The Pairwise DaliLite Server  was applied to calculate RMSD for structure in 0 ns and X ns of MD simulation. Molecular Dynamics (MD) Simulation of lactoferrin The GROMACS program (version 4.5)  and GROMOS96 43A1 force file was applied to prepare the 50 ns molecular dynamic simulation Lf (1BLF) with ligands and carbohydrates removed from the file. The simulation was carried out in the cubic water box with the side length equal to 116 Å at 300 K temperature. Calculations were performed using the PL-Grid infrastructure. The Pocket-Finder Server  was applied in analyse the result of MD simulation. This program applies geometric criterion to search binding pockets. It helped to find iron binding pockets in N- and C-lobe. Results Analysis of phylogenetic relationship between homologous sequences The analysis of amino acid position in the Lf sequence using by the ConSurf Server showed that the iron binding pocket includes the most conserved amino acids as well as some of the most conservative amino acid was find near carbohydrates. Lactoferrin Molecular Dynamics (MD) Simulation The simulation goal was the observation of changes in the tertiary structure of Lf in pH close to 7, in the absence of carbohydrate, Fe3+ and CO32- ligands. The simulation can be divided into two parts. First part of the simulation the relaxation period - from 1 to about 20 ns (FSP) is a term of major changes in the tertiary structure while in second part of the simulation term from about 20 to 50 ns (SST) reveals the structural stability (Figure 2a). a. 1,4 1,2 1,0 RMSD average RMSD RMSD (nm) 0,8 0,6 0,4 0,2 0,0 0 10 20 30 40 50 Time (ns) 3,0 b. RMSD (nm) 2,5 RMSD average RMSD 2,0 1,5 1,0 0,5 0,0 0 10 20 30 40 50 Time (ns) Figure 2. The changes in Lf tertiary structure during the MD simulation: a. backbone after lsq (Least-Squares) fit to backbone, b. N-lobe after lsq fit to the C-lobe. Comparison of the graphs (Figure 2) suggests that movements of N-lobe to C-lobe determine the greatest changes in tertiary structure during the simulation due to high similarity of these two graphs. This simulation confirms the experimental observation that the -helical hinge between Nand C-lobe is very flexible . RMS- D analysis to compare the two lobes The different structural forms of the Lf lobes were observed in the simulation (Figure 3). a. b. c. d. Figure 3. The Lf lobes structure changes during the MD simulation: a. N-lobe before and b. after the MD simulation, c. C-lobe before and d. after the MD simulation. On the N-lobe surface the channel leading to iron binding pocket was formed in contrast to the C-lobe. The channel was formed mainly by changing the relative position of domain N1 and N2. To understand this process the detailed RMSD analysis of the selected part of each lobe was done. The RMSD analysis of the N1 B domain (Table 1) was started by finding the iron binding pocket with the Pocket-Finder Server . As a result, the three binding pockets were found near the site of iron binding (Figure 4a). a. b. c. Figure 4. The binding pockets found by Pocket-Finder Server near the: a. N-lobe iron binding site (localized in the N1 B domain) Pocket 1: residues: 253,254,255; Pocket 2: residues: 304,306,307; Pocket 3: residues: 288,289,290, 291, 298,300,301,302; b. C-lobe iron binding site (localized in the C1 B domain) Pocket 1: residues: 595 ,596, 597, 598, 599, 604, 607, 608, 611, 612, 648, 649, 650, 651, 655,656, 657, 660, 661, Pocket 2: residues: 631, 632, 633, 634, 639, 641, 642, 643; Pocket 3: residues: 632, 641, 645, 647, 648; c. N-lobe iron binding site (localized in the fragment of the N-lobe) Pocket 1: 211, 229, 239, 242, 245, 246; Pocket 2: 60, 61, 82, 89, 211, 250, 251, 252, 253; Pocket 3: 8, 9, 11, 15, 42, 45, 57, 58, 59, 60, 253, 297, 299, 300, 301. On the figure, the clamp indicates the amino acid sequence fragment which was analyzed. The Fe3+ and CO32- ions were marked in yellow. The relative position of the binding pockets appeared rather more stable in period from 12 to 50 ns. The conformational changes of the iron binding site surrounding in the N-lobe seem to correlate in the time with changes in the relative positions of N- and C-lobe as was confirmed experimentally [7,8]. The same procedure was applied to analyze the conformational changes of the C-lobe (Figure 4b).The relative position of the binding pockets near the site of iron binding in the C1 B domain was very stable during the MD simulation. In conclusion, the C-lobe can be "closed" even if it is not associated with the iron and the N-lobe is opening up. The similar process was observed experimentally [7,8]. In the next step, the more detailed analysis of the N-lobe was done (Figure 4c). As a result it was discovered that changes in relative position N1 and N2 domains are the cause of the "open" N-lobe structure. These changes are very fast and big, while the changes in relative position N1 A and N1 B are rather small and slow (Figure 5). These results are in a good agreement with experimentally observed . N-lob Hinge C-lob RMSD [nm] Time [ns] Figure 5. The changes in Lf tertiary structure during the MD simulation: N-lob - N1 A, N2 domains, Hinge (part of the protein connecting two lobes ) - N1 B, C1 A domains, C-lob - C1 B, C2 domains, The Pairwise DaliLite Server  was applied to calculate RMSD for structure in 0 ns and X ns of MD simulation. Summary Despite the absence of ligands in the implementation of the MD simulation the good agreement of simulation result with experimental data was obtained. This may be due to high conservativity of amino acid near carbohydrates. This is why they do not affect the iron binding process due to their different function. The MD simulation helped determine that the formation of the "open" N-lobe structure (30 ns of MD simulation) is preceded by a conformational change within the relative position of the N- and C-lobe (the hinge conformational change in 15 ns of MD simulation) (Figure 5). The results of performed MD simulations provide the consistent with experi-mental data information which demonstrates the effectiveness of MD simulation to structural studies of Lf. Studies applying this method will be continued in order to better understand the Lf functions. Acknowledgements Calculations were performed using the PL-Grid infrastructure (CYFRONETLCG2 grant id 7239).
Bio-Algorithms and Med-Systems – de Gruyter
Published: Dec 1, 2012
Access the full text.
Sign up today, get DeepDyve free for 14 days.