Model of β-Sheet of Muscle Fatty Acid Binding Protein of Locusta migratoria Displays Characteristic Topology.

The β-sheet of muscle fatty acid binding protein of Locusta migratoria (Lm-FABP) was modeled by employing 2-D NMR data and the Rigid Body Assembly method. The model shows the β-sheet to comprise ten β-strands arranged anti-parallel to each other. There is a β-bulge between Ser 13 and Gln 14 which is a difference from the published structure of β-sheet of bovine heart Fatty Acid Binding Protein. Also, a hydrophobic patch consisting of Ile 45, Phe 51, Phe 64 and Phe 66 is present on the surface which is characteristic of most Fatty Acid Binding Proteins. A "gap" is present between βD and βE that provides evidence for the presence of a portal or opening between the polypeptide chains which allows ligand fatty acids to enter the protein cavity and bind to the protein.

Background

In insects, body fat plays a major role [1]. Insects store lipids as glycogen and triglycerides in the adipocytes. The metabolism of lipids is important in insects for growth and reproduction and provides energy during non-feeding periods. In the resting locust, the major source of energy is Trehalose (a hemolymph) [2]. One advantage of using lipids as a source of energy is that they weigh less than isocaloric amounts of carbohydrates [3, 4]. The main storage form of lipids is as triglycerides which are released as diglycerides during flight [4, 5].

Based on the tissue of origin, the intracellular fatty acid-binding proteins (FABPs) are divided into three categories: (i) hepatic-type FABPS, (ii) intestinal-type FABPS and (iii) muscle/cardiac-type FABPs. These proteins are 14-15 kDa in size and are important for the uptake, metabolism and transport of long-chain fatty acids. They are also responsible for the modulation of cell growth and proliferation. Human muscle and bovine heart fatty acid binding proteins bind long-chain fatty acids in the cytosol of muscle tissues.

According to the Portal hypothesis proposed for the uptake of fatty acids by FABPS, the fatty acid first adsorbs to the protein surface and then searches for an opening or portal by which it can enter the protein cavity. After finding the opening, the fatty acid enters the protein cavity. The next step is the protonation of the ligand, before de-solvation, which demands a large amount of energy (30 kJ mol-1). Since the insertion of the negative charged ligand into the low dielectric matrix of the protein is energetically unfavorable, ligand-binding to the protein occurs before the protonation of the ligand. The insertion of the charged head-group of the ligand fatty acid inside the protein cavity requires energy in the amount of 300 kJ mol-1[6]. Molecular dynamics simulations of I-FABP have shown that the ligand first adsorbs to the protein surface. The rate limiting step is the de-solvation of the carboxylate head group of the fatty acid anion [7].

The β-sheet of human muscle fatty acid binding protein displays ten β-strands arranged anti-parallel to each other in which each strand is hydrogen bonded to its neighbor. This arrangement is completed by the formation of hydrogen bonds between the first and the last β-strand (Figure 1). The β-sheet is folded to form a β-barrel and contains an interior and an exterior surface. The interior surface serves as a cavity or a pit where the ligand binds [8]. The volume of this cavity varies between 300 and 700 Å3 [9]. In this project, 2-D NMR data, specifically information provided by the Nuclear Overhauser Effect (NOE), was used to build a model of the β-sheet of Lm-FABP using the Rigid Body Assembly method and bovine heart FABP as a template [10].

Figure 1

Backbone of human muscle FABP showing the arrangement of two α-helices (αI and αII) and ten b-strands (bA, bB, bC, bD, bE, bF, bG, bH, bI and bJ) [13].

Methodology

The methodology used in the study is summarized in schematic 1 (see supplementary material).

Protein expression and purification:

Lm-FABP was expressed in E.coli cells using previously published protocols [11]. The cDNA of the protein was isolated after digestion with restriction endonucleases, NcoI and BamHI and ligated with the pET3d vector. The plasmid pET3d/Lm-FABP was used to transform the E. coli D12 strain BL21 (DE3) cells. Lm-FABP was purified by the use of previously published protocols [11]. The protocol included cell lysis, centrifugation and purification by a Sepharose HR 26/10 FPLC column. Lm-FABP eluted with 50 mM Tris/HCl (pH 8.0). The fractions were analyzed by SDS/PAGE and Western blotting. Fractions containing Lm-FABP were further purified by gel filtration chromatography using a Sephacryl S-100 column.

NMR experiments:

2-D homonuclear NMR spectra were acquired for protein sample bound to native fatty acids in 20 mM phosphate buffer (10% D2O, 0.05% NaN3) at pH 5.5 and 35 °C. These spectra were acquired in the “phase sensitive” mode with pre-saturation of the water signal to obtain maximum signal-to-noise ratio. All the chemical shifts were referenced with respect to sodium 3- (trimethylsilyl) [2,2,3,3-2H4] propionate (TSP). Once the spectra had been acquired, they were analyzed to assign chemical shifts and to collect NOE constraints that consist of sequential, medium range and long range constraints.

Amino Acid Sequence Alignment:

Before analysis of the NMR data, the amino acid sequence of Lm-FABP was analyzed. The sequence is characterized by the presence of thirty one amino acids with aliphatic side chains. It also contains three Proline residues (Pro39, Pro91 & Pro103). The sequence also contains one Cysteine residue (Cys115). There are ten aromatic amino acid residues (seven Phenylalanines and three Tyrosines) and two Histidines (His102 and His96), also present in the sequence. In addition, the sequence is characterized by two Methionines, (f-Met 0 and Met 21).

A comparison of the sequence of Lm-FABP with the bovine heart FABP was performed using the FSSP data-base. The alignment revealed 42% sequence homology between two proteins. There are three insertions (Ile 7, Leu 46 and Asn 92) and two deletions (Gly 99 and Ala 132) in the sequence of Lm-FABP as compared to bovine heart FABP. The amino acid insertions are exactly the same if the sequence of Lm-FABP is compared to human muscle FABP. The only amino acid deleted is Gly 99.

Spin System Identification:

2-D homonuclear COSY and TOCSY spectra were used to identify amino acid spin systems comprising J-coupled 1H. The amide 1H chemical shift dispersion in the 2-D TOCSY ranged from 10.18-6.78 ppm. The scalar coupled NH and CαH 1H were recognized in the “finger print” region of the 2-D COSY spectrum, as shown in (Figure 2). To identify spin systems representing individual amino acids (comprising NH and CαH along with the side-chains 1H), the program AURELIA [12] was used to peak-pick the 2-D data. Initially, about 165 spin systems were identified. These included some doublets due to the presence of spin system heterogeneity. This list later was narrowed down to about 145 spin systems for 130 non-Proline amino acids in the protein. The three Proline residues present in the sequence were identified separately since they don't possess an amide 1H.

Figure 2

Comparison of secondary structure elements of Lm-FABP and bovine heart FABP showing the three insertions (Ile 7, Leu 46 and Asn 92) and the two deletions (Gly 99 and Ala 132).

NMR Chemical Shift Assignment:

The strategy of sequential 1H chemical shift assignment was followed to assign the backbone and side-chain NMR chemical shifts of Lm-FABP (data not shown).

Identification of Secondary Structure Elements:

Secondary structure elements (Figure 2) were identified by characteristic Nuclear Overhauser Effect (NOE) cross-peaks as shown in schematic 2 (see supplementary material). The sequential NOEs observed for α-helices consist of a strong dNN and a weak dαN NOE cross-peak for each “i-1” and “i” amino acid residue. In addition to these medium range NOEs are also observed as shown in schematic. For β-strands, the characteristic NOE information is a strong dαN and a weak dNN cross-peak for “i-1” and “i” amino acid residues. For the β- turns, a strong-weak dNN and strong to medium dαN NOE crosspeaks are expected for “i-1” and “i” amino acid residues.

Homology Modeling of β-sheet:

The β-sheet of Lm-FABP (Figure 3) was modeled after the β- sheet of Bovine Heart FABP using the Rigid Body Assembly method [10]. This method starts with the identification of the conserved β-sheet of bovine H-FABP and then assembly of the model of β-sheet of Lm-FABP from the H-FABP template using the inter-strand NOE cross-peak intensities in schematic 3,Table 1 (see supplementary material).

Figure 3

β-sheet of Lm-FABP modeled after bovine heart FABP by the use of 2-D NMR data and the Rigid Body Assembly method [10].

Discussion

Many structures of human, bovine and insect FABPs are known via X-ray crystallography or NMR [13-17]. These structures have provided details about ligand-binding property of FABPs. The FABPs display a ten-stranded β-sheet structure and a large, hydrophilic, water-filled inner cavity that serves as a ligand-binding site for hydrophobic ligands such as fatty acids and retinoids [6]. The β-sheet of Lm-FABP shown in (Figure 3) closely resembles that of the other FABPs. The fatty acid is bound to FABPs by electrostatic and hydrogen bond interactions of its carboxylate head group with charged or polar residues inside the protein and by interactions of its tail with hydrophobic residues. A “gap” is observed between βD and βE in the β-sheet which can serve as a portal for entry of long chain fatty acids in the inner cavity of the protein.

A description of the ligand-binding process is given by the Portal hypothesis which describes the entry of long chain fatty acids inside the protein cavity through an opening or portal. The Portal hypothesis is supported by many experimental studies, most of which were carried out using either the I-FABP or ALBP. Crystallographic studies of both these proteins have indicated that the lipid tail is located near the suggested entry site [18, 19]. NMR measurements of I-FABP have suggested that the protein exhibits a pronounced backbone disorder at the portal region and is more mobile than the rest of the protein suggesting that this region may be involved in ligand insertion [17, 20, 21, 22].

Conclusion

The information provided by the Nuclear Overhauled Effect was used to build a model of the β-sheet of Lm-FABP using the Rigid Body Assembly method and bovine heart FABP as a template. The β-sheet structure of Lm-FABP displays a typical structure of 10 anti-parallel β-strands arranged to form a sheet. A “gap” is observed between βD and βE. A β -bulge is also present between Ser 13 and Gln 14 which is a difference from the structure of bovine heart fatty acid binding protein. The observed “gap” between βD and βE provides evidence for the presence of a portal between the polypeptide chains which allows long chain fatty acids to enter the protein cavity and bind to the protein.