SEARCH

SEARCH BY CITATION

Keywords:

  • Computational biology;
  • biotechnology education;
  • computers in research and teaching;
  • molecular modeling;
  • estrogen receptor;
  • estradiol;
  • molecular operating environment

Abstract

  1. Top of page
  2. Abstract
  3. METHODS
  4. RESULTS AND DISCUSSION
  5. FUTURE TASKS
  6. REFERENCES

Molecular modeling is pervasive in the pharmaceutical industry that employs many of our students from Biology, Chemistry and the interdisciplinary majors. To expose our students to this important aspect of their education we have incorporated a set of tutorials in our Biochemistry class. The present article describes one of our tutorials where undergraduates use modeling experiments to explore the structure of an estrogen receptor. We have employed the Molecular Operating Environment, a powerful molecular visualization software, which can be implemented on a variety of operating platforms. This tutorial reinforces the concepts of ligand binding, hydrophobicity, hydrogen bonding, and the properties of side chains and secondary structure taught in a general biochemistry class utilizing a protein that has importance in human biology.

The rapid developments in our understanding of biomolecular structures and their direct relationship to function have transformed biochemistry to a new level. Integration of this knowledge in a readily available framework is critical and this can be achieved through visualization of three-dimensional structures of biomolecules using interactive computer graphics. In addition, molecular modeling plays an important role in structure-bas drug design in the pharmaceutical companies, biotechnology companies, and university research since it is useful in predicting structures of macromolecules, as well as for analyzing chemical and biological functions of these molecules. For these reasons, interactive molecular modeling has now become an integrated part of standard Biochemistry and/or Biomolecular Science courses taught at many universities. In this work, we present a relatively simple tutorial for a specific software package for such modeling exercises in a typical undergraduate classroom environment.

Currently, a number of molecular modeling software packages are used for teaching three-dimensional structures of biological macromolecules in undergraduate courses [1–15]. Many of these packages are available for free, and several tutorials illustrating simulation details as well as pedagogical aspects also are available. The present report is a result of our own efforts for several years in this area, and focuses on the software package, Molecular Operating Environment (MOE), developed by Chemical Computing Group Inc [16]. It should be noted at this time that neither our university nor we received any compensation for this work by this company. In earlier years, we have used Insight II as a visualization platform and developed tutorials for introductory biochemistry classes. Examples of macromolecules that were used in these tutorials include: hemoglobin, myoglobin, porins, membrane proteins, ferrodoxin-NADP+ reductase, and citrate-synthase. When our IBM RISC stations using the UNIX operating system came to the end of their lifetime we converted these tutorials to be used on PCs with MOE.

MOE is quite versatile and can be used on a wide variety of platforms ranging from Intel computers running Microsoft Windows™, Linux to Mac OS X, IBM eServer, Sun Microsystems, Hewlett-Packard, and SGI systems. This powerful program is generally available for classroom use at an academically discounted cost. We found the PC (Windows) platform most useful for our purposes, because essentially all of our computer labs and classrooms are equipped with PCs and students are more familiar with this mode of operation. In order to run MOE, at least 1 GB of free hard-disk space and at least 64 MB memory are required. Use of MOE for molecular modeling in these classes allows students to view, display, analyze, interpret and manipulate biomolecular data in a straightforward way by utilizing interactive three-dimensional computer graphics tools. The system integrates visualization, molecular modeling, protein modeling, cheminformatics and bioinformatics all in one program. Although there are other UNIX programs available we found this superior because the students spent less time learning the operating system and more time focused on the biochemistry aspects of the tutorials. MOE is versatile and flexible due to certain additional capabilities including the pharmacophore-based combinatorial library design, potential energy evaluations, docking, crystallographic system and electron density map visualization. By introducing the student to the initial suite of visualization programs they can easily move to more complex utilities of the system in later research projects or in their future careers.

This manuscript focuses on a specific application of MOE, where we describe the molecular structure of the ligand-binding domain of the human alpha-estrogen receptor (LBD-ER). A biochemical modeling project on the estrogen receptor (ER) is suitable for an introductory biochemistry class or an environmental class about endocrine disruptors in the environment. Students planning to use this tutorial should have some familiarity with basic protein chemistry as described in any standard Biochemistry textbook [17–20]. We have designed the tutorial with two objectives in mind: 1) to show the correlation between protein structure and function and to 2) introduce the students to the ligand binding properties of the ER. This tutorial will reinforce the concepts of hydrogen bonding and hydrophobic interactions that occur in ligand binding in steroid receptors. Most importantly students are introduced to computational biochemistry research. A brief outline of the general course-structure is presented in Table I.

Table I. Sample outline of the general course-structure
TopicHours
Course introduction1
Search and gather information from the Internet, Protein Data Bank, Pub Med etc2
Nuclear Receptor Superfamily2
Why study ER2
Molecular Operating Environment2
Overview of the three dimensional structure of ER and EST2
Observing the Hydrophobic pocket in the LBD1
Conserved amino acids and their role2
Observing hydrogen bond, ionic bond and cation-pi interactions in the protein3
Surface properties of the protein2
Structural and comparative studies of ER-LBD in complex with different ligands4
In class oral presentations4

METHODS

  1. Top of page
  2. Abstract
  3. METHODS
  4. RESULTS AND DISCUSSION
  5. FUTURE TASKS
  6. REFERENCES

Working with the MOE Window

A typical MOE application window is shown in Fig. 1. The top tool bar contains all the basic commands that will be used in this tutorial and this is accessible by mouse click. The blank line below the top tool bar is used for command line computing that we will not use in this present tutorial. SEQ is used to open Sequence Editor Window and Cancel is used to cancel any job. A detailed description of the MOE Window can be accessed from the MOE Help menu. Here, we briefly summarize a number of main commands. The vertical tool bar on the far right is used for additional commands and is accessible again by mouse click. For example, under System one can open the Atom Manager Window. The Builder button will open the Molecule Builder that allows one to construct small molecules or macromolecules or nucleotides. The Minimize button will perform Energy Minimization on the molecule in the window. The Close button can be used to clear molecular data. Using View it is possible to view the protein in different display Modes including line, stick, ball and line, ball and stick, and space filling. Among other standard MOE commands, Label is used to identify atom type, charge, or residue. Color is used to color specialized items such as specific atoms residues, chains, or secondary elements. Hide and Show are used to highlight or hide selected elements of structure. Measure is used to determine bond lengths, bond angles and dihedrals. The bottom part of the MOE window contains “dials” that are used for rotating, moving and zooming of molecule.

thumbnail image

Figure 1. MOE Application Window.

Download figure to PowerPoint

Data Input

To view the three-dimensional structure of LBD-ER complexed to 17β-estradiol (EST), a natural estrogen, the PDB file 1ERE.PDB [21] from the protein data base [22] was first edited to include data for only one dimer of the three found in the file. This edited file was then imported in the MOE window. Fig. 2A illustrates the structure of LBD-ER in ribbon form with EST displayed in space filling mode. The 2D structure of EST alone is shown in Fig. 2B.

thumbnail image

Figure 2. (A) Ribbon diagram of hERα LBD complexed with 17β-estradiol (white space filling mode). The figure shows the dimeric form of 1ERE.PDB. (B) Structure of 17β-estradiol (C18H24O2).

Download figure to PowerPoint

RESULTS AND DISCUSSION

  1. Top of page
  2. Abstract
  3. METHODS
  4. RESULTS AND DISCUSSION
  5. FUTURE TASKS
  6. REFERENCES

Why Explore the Structure of the ER?

Natural estrogens play a critical role in the normal growth, development and maintenance of a diverse range of tissues. They exert their effects via the ER which functions as a ligand activated transcription regulator. The ligand-binding domain of this protein is autonomous and has multiple functions including ligand binding, dimerization and ligand dependent transactivation [23]. Study of this part of the LBD-ER provides crucial information about the receptor function and steroid receptors in general.

X-ray crystal structures of the LBD-ER have been reported with a variety of ligands including EST, raloxifene, diethylstilbestrol, tamoxifen, genistein, and a number of synthetic estrogens [21, 24–27]. LBD-ER binds a large repertoire of compounds with remarkable structural and chemical diversity. Most ligands have two hydroxyl groups separated by a rigid hydrophobic linker. They are classified as agonist or antagonist ligands by their effect on the structure and function of the protein. The three dimensional structure of LBD-ER as shown in Fig. 2A displays a canonical alpha-helical sandwich topology composed of 12 α helices that are arranged into three anti-parallel layers. This arrangement allows a sizeable buried binding pocket where the ligands are sequestered. In general, when an agonist binds in the pockets, helix 12 forms a “lid” over the binding pocket and exposes the transactivation site for coactivators. This binding is necessary for transcriptional activity. In contrast, when an antagonist binds to the pocket, it changes the conformation of the LBD such a way that helix-12 blocks the coactivator-binding site. Thus the orientation of helix 12 is an important factor for distinguishing agonist from antagonist. The underlying determination of the function of the LBD-ER with a particular ligand is revealed in its binding mode in the cavity. Exploration of the hydrogen bonding between ligand and protein and complementarity of the hydrophobic residues and the ligand give the students insight into steroid-ligand recognition. Further exploration of the protein architecture around EST reveals how the A ring pocket imposes a requirement for the planar ring group. The promiscuity of the LBD-ER is exhibited by the large unoccupied space above and below the steroid cavity.

Visualizing the Structure

Upon opening the MOE window and loading the data, use X, Y, and Z rotate wheels on the bottom of MOE window to rotate the molecule in the three planes. Zoom in by using the zoom tool located on the bottom of the screen. It is also possible to change the appearance of the protein (line/stick/ball and line/ball and stick or space filling) by using the Render menu from the top toolbar panel or Mode menu from the far right toolbar. Click on Render|Backbone|Color|Secondary Structure to view the Secondary structure of the protein or click on Render|Backbone|Color|Chain Color to color the Protein by Chain. To make viewing of the secondary structure easier, color the protein by clicking: Render|Backbone|Flat Ribbon. Viewing may be made even easier by hiding the backbone. To do this, click: Render|Hide|Backbone. Hide the sidechains by clicking on Render|Hide|Sidechain. Now we only can see the ribbons.

To view the EST in its binding site click: Render|Show| Ligand. To highlight it, open Sequence Editor. Highlight EST and, right click Atoms|Select and then Atoms|Show. EST will be displayed in the MOE window in pink. Change the appearance of the EST to space-filling mode by clicking Render| Space filling. In the MOE window, we can now see that the ligand-binding domain has the anti-parallel topology, with bound EST (Fig. 2B) in the lower portion of the domain (Fig. 2A).

View the dimer of 1ERE.PDB in stereo mode by clicking on Render|Stereo|left-Right. Adjust the stereoscopic view by zoom in or zoom out the molecule in the MOE window. View the protein by clicking on MOE|GizMOE| Rock and Roll. The protein in the MOE widow will then start to move slowly, and the three-dimensional structure of the molecule will become more apparent.

Save the molecular data by clicking: File|Save. You may want to save your molecular structural output either in moe format or in pdb format (e.g. *.moe). * is a wild card that represents any series of characters. You may also want to print the figure by clicking: File|Print (set the printer as postscript and select print to file option) and then save it as a postscript file (e.g., “ere.ps”). Or, you may save it on your computer using the Printer pull down menu. These choices are available depending on the printing devices installed on the computer. Details of saving and printing of a MOE file can be found on the help menu of MOE.

Observing the Hydrogen Bonding

To observe details of hydrogen bonding, click: Edit| Hydrogens|Add Hydrogen, then click: Render|Draw| Hydrogen Bonds. Before observing the hydrogen bonds, display the backbone by clicking Render|Show|Backbone. Using MOE it is possible to display all or a selected set of hydrogen bonds in a given macromolecule. The contact analysis option of MOE (MOE|Compute|Biopolymer|Protein Contacts) reports hydrogen bond contacts within chains and/or between different chains. Both side-chain-to-side-chain and side-chain-to-main-chain hydrogen bond contacts can be calculated (see web-based supplementary material [28]). In Fig. 3, hydrogen bonding between 17β-OH of EST and δ nitrogen of His 524 is displayed. Here, Glu 353 in the ER accepts the H-bond donated by the 3-OH group of EST. The Arg 394 helps to keep the glutamate side chain in its right position via a structurally conserved H2O mediated H bond. The side chain of Arg 394 is further supported by a H-bond to the carbonyl of a nearby Phe 404. A particular task for the students may be to trace the entire hydrogen-bonding network in the MOE window. They might also be asked to measure the hydrogen bond distances shown in Fig. 3.

thumbnail image

Figure 3. Hydrogen bonding between the 17β-estradiol (white) and the side chains of the ER protein in the ligand-binding pocket. The side chains, His 524 (orange), Glu 353 (red), Phe 404 (green), and Arg 394 (yellow) are illustrated.

Download figure to PowerPoint

Hydrophilic and Hydrophobic Regions

To display the hydrophilic residues of 1ERE.PDB in the MOE window, open Sequence Editor by clicking SEQ OR click on Window|Sequence Editor. In the sequence editor, click: Selection|Conserved Residues|Hydrophilic. This will highlight all the hydrophilic residues. To display these residues on the protein, click: Selection|Atoms|Of Selected Residues. The hydrophilic residues will now be displayed in pink on the protein. Repeat the same procedure to observe the hydrophobic residues of the protein. Be sure to select chains before selecting conserved residues. The contact analysis option of MOE can also be used to identify hydrophobic and ionic contacts (salt bridges), in this way, students can browse and isolate contacts in both the Sequence Editor and main MOE window [28].

Other Noncovalent Interactions

Cation-pi interactions play a crucial role in protein folding and stability [29, 30]. There are several papers that describe and demonstrate the importance of such interactions in the framework of undergraduate Biochemistry course [31, 32]. The CaPTURE program by Gallivan and Dougherty is an authoritative source for identifying energetically significant cation-pi interactions within proteins in the Protein Data Bank [33]. Although such interactions are often involved in ligand binding, the EST in 1ERE is not involved in such interactions. So it would be useful to leave this as an assignment for the students to determine whether there would be such an interaction between the protein and EST.

Another example of non-covalent interactions is the salt bridge, which also is important in protein stability and folding. The contact analysis option of MOE can identify ionic contacts (salt bridges) in the protein. The unique salt bridge that stabilizes the agonist conformation in 1ERE is between Arg-548 (end of helix 12) and Glu-523 (helix 11). Another relevant intra-molecular protein salt bridge is formed between Glu353 and Arg394 that are adjacent to the ligand binding pocket. A particular task for the students would be to identify each of them in the MOE window.

Amino Acids in the Binding Site

Particular amino acid sequences in the binding site are necessary for the receptor to bind ligands. MOE has a very versatile tool to display and identify particular amino acids. To display a specific amino acid of 1ERE.PDB in the MOE window, open Window|Sequence Editor. In the Sequence Editor window, click: Selection|Residue Selector. In the Residue Selector window, click on the amino acid of your choice; for example, Histidine—HIS or H. Then Click on Select Atoms and finally close the Residue selector window. This will highlight all the Histidine residues of the pdb file in the MOE window. From the previous section, we know that the important amino acids in the binding pocket include Glu 353, Arg 394, and His 524. Students can easily locate these residues in the MOE application window. Koffman et al. reported that tyrosine has an important role in the binding of EST in ER [34]. According to Anstead et al. [35], Tyr 537 may have a controlling role in ligand binding because Tyr 537 lies at the start of helix 12. Tyrosine 537 is highlighted and shown in Fig. 4.

thumbnail image

Figure 4. Alpha carbon backbone structure of 1ERE.PDB in dimeric form. Tyr 537 (white) and 17β-estradiol (yellow) are shown in space filling mode.

Download figure to PowerPoint

Students can be guided to observe some of the other critical residues in and near the active site. These are highlighted in Fig. 5. This include: Ala 350 (white) Arg 394 (violet), Gly 521 (dark yellow), Leu 346, Leu 349, Leu 387, Leu 384, Leu 391, Leu 525 (Red), Phe 404 (yellow), Met 388, Met 421(green). Another student task would be to locate Ile 424, Leu 354, Leu 428, Leu540, Met 343, Met 522, Met 528, Phe 425, Val 534 in order to determine each of the residue's orientation to the ligand and to each other.

thumbnail image

Figure 5. Critical residues in the binding pocket of the ER illustrated in stick form are depicted in this figure. Illustrated are Ala 350 (white), Arg 394 (violet), Gly 521 (dark yellow), Leu 346, 349, 384, 387, 391, 525 (Red), Phe 404 (yellow), Met 388, 421 (green), and 17β-estradiol in white.

Download figure to PowerPoint

Surface Representation

Using MOE, the students can create, display, and manipulate the molecular surface of a biomolecule. With this method, AnalyticConnolly, GaussAccessible, GaussConnolly, and Interaction surface of a biomolecule can be displayed [16]. To operate this mode in the MOE window one needs to click Compute|Molecular Surfaces. A dialogue box will appear. In the dialogue box for surface type, click: Interaction. Set the spacing at 0.75, and set the render as solid with transparency set to 0. This will generate the Interaction surface for the protein. To remove the radii by clicking Window|Graphic Objects, choose the surface and click Hide. Similar protocols are used to display the GaussConnolly surface of the biomolecule as shown in Fig. 6. The two later surface renderings reinforce the importance of these surface hydrophilic and hydrophobic interactions in the formation of the LBD dimer. The contrast in color allows the students to identify the areas of the biomolecule that will be solvent accessible and available for hydrogen bonding. They will also be able to understand where the two lobes of the dimer interact and why this happens. Students can also visualize the hydrophilic areas of the surface away from the dimer interaction.

thumbnail image

Figure 6. GaussConnolly surface of 1ERE.PDB. This depiction represents pocket areas on the protein. The red areas represent nonpocket regions or peninsula regions, while the white areas represent neutral regions. The blue and green areas are hydrophilic and hydrophobic pockets, respectively.

Download figure to PowerPoint

Contact Statistics

The structure of the LBD-ER shows a dimer that is stabilized by hydrophobic interactions. All nuclear receptors have a hydrophobic core in which a specific ligand binds.

To find out the receptor's contact preference in this process, open: Compute|Contact Statistics Grid. In the MOE window select EST atoms. Now in the contact Statistics window panel set the Atoms to be Unselected in any chain. Press the Apply button. We will see the contact statistics plots of the ER's preferences along with the ligand (Fig. 7).

thumbnail image

Figure 7. Contact Statistics plots of 1ERE.PDB. EST is shown in the white ball and line mode. The green solid area is hydrophobic when the settings are at 90% and red is hydrophilic at 90%.

Download figure to PowerPoint

FUTURE TASKS

  1. Top of page
  2. Abstract
  3. METHODS
  4. RESULTS AND DISCUSSION
  5. FUTURE TASKS
  6. REFERENCES

By using the tutorial presented here, it is possible to engage students in exploring the structure of ligand-binding domain (LBD) of the ER when complexed with diethylstilbestrol (3ERD.PDB), 4-hydroxytamoxifen (3ERT.PDB), and raloxifene (1ERR.PDB). It is also possible to assign the task of identifying, superposing and comparing the structures of beta ligand-binding domain complexed with full antagonist raloxifene (1QKN.PDB), and partial agonist genistein (1QKM.PDB).

REFERENCES

  1. Top of page
  2. Abstract
  3. METHODS
  4. RESULTS AND DISCUSSION
  5. FUTURE TASKS
  6. REFERENCES