RICERCA ITALIANA     

Structural studies on hydrophobic molecule-binding proteins

Università degli Studi di Verona

Abstract

X-ray crystallography, NMR and other spectroscopic techniques, mass spectrometry, non-conventional electrophoretic techniques, and amino acid sequencing will be used to characterize at the molecular level and with the highest possible degree of detail the members of a structural protein family that share the function of lipophilic ligand binding. The molecules included in this group are known to belong to a very limited number of protein folds and for two of these we have more than one protein present in the list we intend to study. Our research will also focus on the relevant role of some members of this protein family in the important function of vitamin transport in vertebrates. Our program includes two vitamin transporters: riboflavin-binding protein and human intrisic factor, transporter of vitamin B12 (cobalamin). <<<

Principal Investigator

Ugo Luigi MONACO Università degli Studi di VERONA

Research Objectives

This program is a continuation of a research proposal that was already funded three times by this agency which allowed the participating operating units to establish sound collaborations that have produced significant results in the field. The goal of this research program was and is the application of the different and complementary techniques that are in use in the participating research units to several related aspects of the important problem of the interaction of a hydrophobic ligand with a protein evolved to bind it specifically or together with a family of structurally related ligands. The techniques available are those most prominent in modern structural analysis of biological macromolecules: X-ray crystallography, NMR, UV and visible spectroscopy, mass spectrometry, epitope mapping and other immunological methods advanced electrophoretic and conventional chemical sequencing techniques. The proteins we have selected, which were in some cases substituted as the program evolved, represent a very wide spectrum of hydrophobic binders and are actively being worked on by many important groups all over the world. We intend to study the properties of human serum albumin, which is probably the most thoroughly studied example of transport proteins. Our list includes also four lipocalins, members of a very well known structural family, alpha 1 microglobulin, beta-lactoglobulin, apolipoprotein D and prostaglandin D synthase (beta-trace). This last protein is the first lipocalin in which the presence of enzymatic activity has been proved. To a structural family closely related to that of the lipocalins belong the bile acid-binding proteins (BABPs). During this funding period we will continue our work on chicken liver (basic) fatty acid-binding protein, a small protein belonging to the family of intracellular (the four lipocalins are extracellular) lipophilic ligand binders and will extend our research field to include the structurally and functionally related human ileal lipid-binding protein. Riboflavin-binding protein is a member of another structural family that includes the very important folate receptors, extrinsic membrane proteins that participate in the folate cellular uptake, that are found altered in some neoplastic processes. In this protein family, the best characterized member is hen egg-white riboflavin-binding protein, the only protein of this family whose three-dimensional structure was determined by X-ray diffraction. Although cobalamin can not be defined as a hydrophobic ligand, we decided to spend part of our efforts in trying to structurally characterize one of its specific transporters, intrinsic factor, since there is currently no structural information on any transporter of this important vitamin. Finally, a new addition to our list is the velella astaxanthin-binding protein, an intensely blue molecule of unknown sequence that might be structurally related to the proteins we have been studying for several years.
To each of the proteins listed above we will attempt to apply all the techniques available to us with the final aim of describing in as much detail as possible the structure of the protein and its mechanism of ligand recognition <<<

First Results

a) Albumin.
The characterization of the amniotic fluid protein post-translational modifications will be initially based on the hypothesis that the chromophore is a derivative of tryptophan catabolism. At the end of the 1st year the hypothesis will be checked and, if found to be correct, the structural properties of the adduct will be determined. If not, a detailed analysis performed using UV-vis spectrophotometry, spectrofluorimetry and CD will allow the definition of the optical properties of the adduct. In addition, the mass spectrometry analyses of the isolated tryptic fragments will have revealed the identity of the modified peptides and some properties of the bound molecule.
The molecular defects causing analbuminemia and hypoalbuminemia as well as the amino acid substitutions of the variant albumins isolated from the sera sent to us will be defined.
b) Beta-Lactoglobulin.
During the first phase we expect to produce doubly enriched (13C and 15N) porcine and bovine lactoglobulins (PLG and BLG) and to determine the 3D structures, via NMR and dynamic simulations, of PLG. At the same time we expect to have all the listed mutants and the new ones designed on the basis of experimental folding and interaction data and to determine the thermodynamic parameters relative to the unfolding of PLG
The molecular docking calculations will provide deeper knowledge on the properties of the calyx as well as on the features required for a chemical to become a ligand. New complexes of BLG with drugs selected on the basis of the results of docking experiments will be prepared.
The anomalous behavior of apo BLG in DGGE at pH 5 will be clarified.
c) Alpha-1-Microglobulin.
During the 1st year the purification of the dimeric form and of the higher molecular mass oligomers will be completed. The proper conditions for the chemical deglycosilation of the amniotic protein, already under investigation, will be established. We expect that a suitable amount of the lipocalin will be available for crystal screening experiments.
d) Apolipoprotein D.
The problems to be solved in order to be able to obtain sufficient quantities of recombinant protein for structural work will have become evident.
e) Prostaglandin-D-synthase.
The protein, isolated from amniotic fluid in parallel with alpha-1-microglobulin, will be available.
The affinity of hydrophobic molecules for human PGDS will have been screened in comparison with the known properties of rat PGDS. These results will help to better understand the role of this protein in biological fluids and to compare its characteristics with those of other lipocalins.
f) Bile Acid-binding Proteins.
New information on the structure of co-crystals of chicken Lb-FABP will be available.
During the first phase we also expect to produce doubly enriched (13C and 15N) liver Lb-FABP and to determine its 3D structures, via NMR and dynamic simulations. We also expect to determine the dynamics of the protein in the apo and holo forms, through relaxation NMR measurements at different magnetic fields and to characterise the dynamics of binding.
The bioinformatic approach on the fatty acid binding protein family will contribute to define the properties of the calyx. Specifically, the amino acid side chains involved in direct interactions with the ligands will be identified and their role in binding and selectivity will be described. For Lb-FABP, the affinity towards various bile acids in comparison with already established ligands will have been assessed. A comparison between theoretical and experimental results will be useful to define the nature of the natural ligands of the protein and the determinants of the affinity (e.g. the balance between hydrophobic and hydrophilic substituents).
g) Riboflavin Binding Protein.
Hopefully the recombinat Xenopus laevis protein, at least partially purified, will be available. Possibly specific antibodies against the Xenopus laevis protein will be available.
h) Human Intrinsic Factor.
The first steps towars the production of good quantities of the recombinant protein will have been taken.
i) Astaxanthin Binding Protein.
During the 1st year we will prepare the isolated subunits which, when complexed with the carotenoid, forms dimers. The main problem is that the very high affinity requires very harsh conditions for ligand removal and, therefore, the apoprotein yield is low.a) Albumin.
The results expected are dependent upon the identity of the molecule affecting the optical properties of the amniotic fluid protein. If the adduct turns out to be a triptophan derivative, its structural characterization will be completed, otherwise the identity of a different compound and the properties of the linkage will be probably at least partially known. The properties of the characterized genetic variants will establish whether and to what extent the residues are involved in ligand binding.
b) Beta-Lactoglobulin.
During this second phase we expect to complete the analysis of residues playing a key role in folding of BLG and PLG, on the basis of both theoretical computations and site-directed mutagenensis experiments. In addition the role of molecular crowding in folding will be analysed.
The detailed study of the behavior of apo-BLG and of its complexes at high temperatures will clarify the conditions for complex formation after thermal perturbation of BLG in the presence of selected ligands. Drug-BLG complexes will be tested for physico-chemical (pH, ionic strength) and biological (proteolysis) stability: these results will serve to confirm the feasibility of using BLG as a vehicle for water insoluble and/or acid-sensitive drugs for intestinal absorption
c) Alpha-1-Microglobulin.
The factors causing the occurrence of complexes of the lipocalin with albumin and IgA, and also with other plasma proteins, will be known. The availability of tyrosinase, that we have isolated from Agaricus bisporus, and that catalyzes the conversion of 3-OH-kynurenine into xanthommatin, will also allow us to assess whether the adducts are substrates for monophenoloxidases. The crystallization trials using the deglycosilated amniotic and recombinant lipocalin will be completed.
e) Prostaglandin D Synthase.
The best conditions for chemical deglycosylation with trifluoromethansulphonic acid will have been applied to natural as well as recombinant prostaglandin-D-synthase. The conditional stability of the most relevant complexes characterized during the first year will be compared with those of the apoprotein and the structural modifications due to ligand binding will be clarified.
f) Bile Acid-binding Proteins.
More structural information on co-crystals with different complexes of the chicken Lb-FABP will be available. For the most stable complexes, the modifications in accessibility and flexibility of protein regions occurring upon binding will have been assessed through spectroscopic and biological tests. Computation of pKhalf and evaluation of pKa vs pH curves for residues associated to anomalous pKhalf values will have been carried out. The hypothesis of co-binding of different ligands will have been checked. The possibility that the binding of vastatins to lipid binding proteins might mediate some adverse effects of hypolipidemic treatment with these molecules will be evaluated. The folding and dynamics studies will also be completed.
g) Riboflavin Binding Protein.
Important data concerning the Xenopus laevis egg protein will have been derived from 2D electrophoresis analysis resolved with the antibodies in Western blot: pI, molecular mass, relative abundance. These data will allow the development of a purification protocol of the protein from this source.
i) Astaxanthin Binding Protein.
The protein subunits, most likely two and with very few differences in their primary structure, will be available in a pure form. Their amino acid sequences will be known and the disulfide bonds will have been assigned. On the basis of published data on alfa-crustacyanins, we presume that it will be possible to obtain the recostituted complex and we hope that well diffracting crystals can be grown. <<<

Timescale

24 months

National and international background

Participation of the water insoluble metabolites in the biochemical reactions that take place in an aqueous environment requires a mechanism to render them water soluble and amenable to transport among different aqueous compartments. The mechanism developed by evolution is binding of these ligands to a water soluble protein. The molecules we call hydrophobic molecule-binding proteins have thus as a first and most obvious function that of rendering water soluble a hydrophobic ligand whose presence is necessary in an aqueous phase. Although solubilization and transport is the most thoroughly studied and the best understood of the functions of the lipophilic molecule-binding proteins, it is by no means the only function proposed for these macromolecules that are in many cases known to recognize other macromolecules and are believed in some cases to act as metabolic regulators.
The prototype of hydrophobic transporters and one of the proteins most intensively studied with every physicochemical technique is serum albumin (a1), the best known member of a protein family that is known to include other molecules that have not been so extensively characterized. The three-dimensional structure of human serum albumin has been determined by X-ray diffraction studies (a2, a3) and the coordinates of the model of the apoprotein as well as those of its complexes with fatty acids (a4) and warfarin (a5) are available . The study of natural mutants with substitutions in the ligand-binding site can be of great help in the definition of the role of these specific amino acids.
Another well known protein family is that of the lipocalins which includes several members of known three-dimensional structure that share the same fold (1) although their sequence similarity is not as high as could be expected for molecules which are so similar in their three-dimensional structure. Characteristic of this fold, identified for the first time in retinol-binding protein, is the presence of an anti-parallel beta-barrel that in all the molecules of this family studied so far is found to encapsulate the bound ligand. Our research group will concentrate on four lipocalins, three of unknown function: beta-lactoglobulin, alpha 1 microglobulin and apolipoprotein D and a fourth which is one of the few lipocalins known to possess enzymatic activity, prostaglandin D synthase, also known as beta-trace.
Although beta-lactoglobulin is not present in the milk of all mammals and even though its function remains to be determined, abundance and ease of preparation have made this protein a model to which every physicochemical technique has been applied (b1). There are several X-ray structures of different crystalline forms of the bovine apo protein and of its complexes with a few hydrophobic ligands (b2, b3, b4). In every case the ligands are found to bind in the internal beta-barrel cavity, as expected from the analogy with the other lipocalins and the Milano unit has shown that the binding of fatty acids is pH and ionic strength dependent. An NMR structure determined by the Verona 2 operating unit is also available and studies of ligand binding (b5) and of the unfolding process (b6) as well as theoretical calculations that attempt to explain the protein behavior have also begun and promise to give us a very detailed picture of the behavior of this model protein.
Human alpha 1 microglobulin is believed to belong to the lipocalin family because of its sequence similarity to other members of the group, in particular its sequence is most similar to that of the epididymal retinoic acid-binding protein (c1, c2). The protein is glycosylated and shows the peculiarity of possessing a covalently bound chromophore whose structure was determined by the Pavia unit during the previous funding period. Its function is unknown and it is also not known whether it has a high affinity lipophilic binding site.
Apolipoprotein D (ApoD) is a 29 KDa glycoprotein that is predicted to be a lipocalin on the basis of its amino acid sequence (d1, d2). There is currently no experimental information on its three-dimensional structure. Although it has been shown that the protein can bind cholesterol and progesterone (d2), bilirubin (d3), arachidonic acid (d4) and a lipophilic human axillary odorant believed to act as pheromone (d5), the exact physiological ligand has not been identified. Its gene is expressed in many tissues and significant variations of the protein level have been observed in plasma and/or cerebro spinal fluid in several diseases (d6), including diseases of the nervous system of great importance like Alzheimer's disease.
Prostaglandin D synthase (beta-trace) is the only protein of our list which is also known to be an enzyme. It is very abundant in the cerebro-spinal fluid and it is also present in significant concentrations in the seminal and amniotic fluids (e1, e2, e3). The reaction it catalyzes is the conversion of prostaglandin H2 (PGH2) into prostaglandin D2 (PGD2), the latter is a major PG in the central nervous system of mammals where it shows several important functions such as the induction of hypothermia and sleep and is thought to be a new type of neuromodulator or perhaps neurotransmitter (e2, e3). The recombinant enzyme has also been shown to bind several hydrophobic ligands such as retinoids, bile pigments and steroid hormones (e3) but the existence of a non-substrate physiological ligand has not been demonstrated. Although the presence of a highly significant sequence homology with other lipocalins makes it certain that this protein belongs to that family, its three-dimensional structure has not been determined.
Structurally related to the lipocalins, is the family of the intracellular lipid-binding proteins in which the beta-barrel presents 10 instead of 8 strands of beta-structure (f1). The three-dimensional structure of several members of this structural family is known and in particular that of rat liver fatty acid-binding protein (FABP) has been reported (f2). The protein is unusual in that unlike all the other members of the family, it binds two instead of one fatty acid molecule. Some years ago we discovered, purified, crystallized (f3) and determined the X-ray structure of a member of this family that has the interesting property of having an unusually high isoelectric point (f4). We called this protein chicken liver (basic) fatty acid-binding protein. The presence of an analogous protein has been shown in the liver of several other vertebrates (f2) but so far it has not been unequivocally demonstrated that it is also present in the liver of mammals. During the previous funding period, the Verona 1 unit was able to show that the protein binds two molecules of bile acids, a property that renders it similar to the mammalian ileal lipid-binding proteins (or gastrotropins), which are also the subfamily with highest sequence similarity (f5). Although there are currently two NMR structures of the ileal lipid-binding proteins (f6, f7) there is no X-ray structure available.
The specific riboflavin-binding proteins (RfBPs) belong to another structural family. The term specific should be emphasized for there are other, non specific transporters of the vitamin in plasma like for example albumin. The specific RfBPs are believed to be crucial for riboflavin transport during pregnancy when adequate amounts of riboflavin are essential for the survival of the fetus (g1, g3). The only RfBP whose three-dimensional structure is known is hen egg-white RfBP which has been shown to belong to a new fold in which the structure is conditioned by the high degree of cross-linking due to the presence of nine disulphide bridges (g2). Although convincing evidence for the existence of non avian RfBPs has been presented, no RfBP other than those of the hen and quail, has been characterized at the molecular level.
Cobalamin (Cbl, vitamin B12) absorption is mediated in mammals by a series of specific binding proteins. The vitamin is first bound by haptocorrin in the stomach and then transferred to intrinsic factor (IF) in the duodenum following the proteolytic degradation of the first binding protein. The intrinsic factor-Cbl complex is taken up by the apical border of the cell and after some time the vitamin is released to the blood where it travels bound to another specific transporter, transcobalamin II (Tc II). Both IF and Tc II bind the vitamin with high affinity and share some structural features although their levels of expression in different tissues are completely different (h1, h2, h3). The receptor for the IF-Cbl complex, cubilin, has also received attention in the literature and models have been proposed for its structrure (h4). Although the cDNA coding for human IF has been available for some time (h5) and in spite of the fact that the protein has been expressed in Pichia pastoris there is no information on the three-dimensional structure of the protein. Tc II has been crystallyzed but its structure has not yet been solved and therefore there is currently no structural information on any of these important vitamin transporters.
Carotenoproteins are responsible for the varied colourations of many invertebrates and show a wide variety of physical and chemical properties. With the exception of alpha-crustacyanin, the blue astaxanthin-binding protein of lobster carapace, there is currently no information on the structure of the members of this protein family. The Pavia unit has isolated in homogeneous form, the astaxanthin binding protein of Velella velella, a surface-living oceanic organism with a characterisic blue mantle (i1). The protein is an oligomer of unknown primary structure but, on the basis of its ligand binding properties, it is believed to belong to the lipocalin family. The unit has undertaken the structural characterization of the Velella velella carotenoprotein and found that partial aminoacid sequencing indicates that its primary structure has homology neither with the lobster alpha-crustacyanin nor with any other protein of known sequence. Another goal of this project will thus be the structural characterization of this protein and the clarification of its relationship with the other proteins in the list. <<<