Electron density maps and their interpretation
Author: Dr. Karolina Michalska
Purpose of this exercise. In this exercise
you will learn how to interpret electron density maps and how to
build an atomic model of a protein molecule on their basis. For this purpose you will
be using the moleculr graphics modeling program Coot written by Paul Emsley. You will also learn in this exercise about an interesting enzyme called L-asparaginase.
Before you start this exercise, you should be familiar with the following review article:
and with section 3.5 of the following article, describing a very unusual cluster of aspartates and calcium ions in the crystal structure of the enzyme EcAIII (PDB code 2ZAL), which is a type-III (or plant-type) L-asparaginase from the bacterium Escherichia coli:
- Michalska K, Jaskolski M. 2006. Structural aspects of L-asparaginases, their friends and relations. Acta Biochim. Polon. 53: 627-640.
- Michalska K, Borek D, Hernandez-Santoyo A, Jaskolski M. 2008. Crystal packing of plant-type L-asparaginase from Escherichia coli. Acta Cryst. D64: 309-320.
Your protein in nutshell
- You will be working with an atomic model and electron density maps of EcAIII, a plant-type
L-asparaginase from E. coli (PDB code 2ZAL). It is called "plant-type" because similar enzymes were first discovered in plants. Plant-type
L-asparaginases catalyze the hydrolysis of asparagine with the
release of aspartate and ammonia. They belong to the family of
Ntn-hydrolases which utilize an N-terminal nucleophile during the
catalytic reaction. As many other hydrolytic enzymes, Ntn-hydrolases
are expressed as inactive precursor molecules, with the putative
nucleophilic residue embedded in the polypeptide
chain. A unique mode of maturation required to liberate a nucleophile
at the N-terminus, involves an autocalytic splitting of the protein
precursor. Self-processing utilizes the same nucleophile that carries
out the final catalytic reaction. As a consequence of the maturation
process, two subunits are created: the N-terminal alpha
subunit and the C-terminal beta subunit, which carries the catalytic residue (in the case of EcAIII, Thr179) at its N-terminus.
- The crystals of EcAIII were grown at 292 K using the hanging-drop vapor-diffusion method, from a protein solution at 15 mg/ml concentration containing a large molar excess of the reaction product, L-aspartate (added as sodium L-aspartate at 0.1 M concentration). The precipitant solution contained 0.1 M Tris-HCl, pH 8.5, 0.08 M CaCl2, 13% PEG400, 17% PEG4000.
- Knowing that the molecular mass of EcAIII is 33400 Daltons, calculate the protein:L-aspartate molar ratio used during crystallization.
A note about contouring of electron density maps
Electron density maps are calculated using the familiar Fourier-Transform formula. As a result, one gets a huge three-dimensional array (that is stored in the computer and can be displayed on a graphics screen) containing the value of electron density, expressed as rho(xyz), at each point described by the fractional coordinates xyz that corresponds to a node of a grid of points into which the crystal unit cell has been divided. Contouring of an electron density map means connecting the neighboring grid points (perhaps with suitable interpolation) that have the same value of rho. In theory, the units of rho should be, according to its name, electrons per Å3. However, because of various complications (for example, we do not know the magnitude of the structure factor F(000), which obviously is equal to the total electron count in the unit cell, because we do not know the precise contents of the unit cell - for instance the composition of the solvent regions), rho is usually calculated on an arbitrary scale. To provide some common reference for contouring of different electron density maps, scientists have come up with the idea of statistical properties of the rho distribution. Briefly, before drawing the contours, one calculates the mean of all rho values, <rho>, and the standard deviation, known as sigma of all the individual rho values from the mean. In the end, the map is contoured in the units of sigma. For 2Fo-Fc maps the most suitable sigma level is 1. (It could be higher when contouring an electron-rich area, such as a heavy atom site, but not lower.) Fo-Fc maps are typically contoured at + or - 2.5sigma. The positive contours are expected to show those fragments of the structure that are absent from the current model, the negative ones will appear in areas that have incorrectly placed atoms.
Some useful documents about Coot
Coot stands for Crystallographic Object-Oriented Toolkit
Get to know our molecule
- Using the Internet resources (e.g. Wikipedia, the suggested readings, other publications in PubMed), assemble the essential information about enzymes with L-asparaginase activity. In your notepad, describe briefly this activity and write the chemical reaction catalyzed by L-asparaginase. What is the E.C. number of L-asparaginase?
Create your Coot environment
- Before you start using Coot, you will need to download the necessary
files (map files, PDB models). To do so, download into your local disk the
file practical-4.zip (right
mouse-button click) and unzip it in a directory SERP-4. A successful download/unzip operation should create 11 files in your SERP-4 directory. All the files required by Coot during this practical will be found in this directory.
Starting a Coot session
- Open a Coot session
- Go to File - Open map, filter files in Select map window, and load 2zal-2fofc_mod_4A.map, which is a 2Fo-Fc electron density map calculated at 4 Å resolution
- Go to Edit - Map parameters and change map radius to 20 Å
Getting comfortable with Coot
Left-mouse Drag Rotate view
- If you have too many maps/models you can for convenience switch them off via Display manager.
- Map/model/background colors can be changed in Edit.
- If you are not satisfied with what you have done, try Undo in Model/Fit/Refine.
- You may want to save current display as an image. This can be done in Draw - Screenshot - Simple.
- After any changes have been done to the model, it is a good practice to save the new coordinates via File - save coordinates
- For further instructions go to the Coot
manual or ask your tutor.
- Familiarize yourself with mouse navigation. Try contouring the map at different sigma levels
Ctrl Left-Mouse Drag Translate view
Shift Left-Mouse or double-click Label Atom
Right-Mouse Drag Zoom in and out
Ctrl Shift Right-Mouse Drag Rotate View around Screen Z axis
Middle-mouse Center on atom
Scroll-wheel Forward Increase map contour level
Scroll-wheel Backward Decrease map contour level
Use + or - on the keyboard if you don't have a scroll-wheel.
Electron density maps at different resolution
- Once you have learned how to change the display, try to identify in the electron density map those regions that correspond to protein or to solvent. Are you able to identify the secondary structure elements?
- When you managed to find a map region corresponding to an alpha-helix - center it and go to Calculate - Other modeling tools - Place helix here. The program will generate a model oligopeptide chain consisting of alanine residues (poly-Ala) that corresponds to the map in the center of your display. Does the model fit the electron density? Is the helix orientation (direction) correct? Now, look carefully at the side chians of the model, which are, of course, all methyl groups (poly-Ala). Looking at the map, are you able to determine what side chains should be in reality placed in the electron density? Perhaps by careful analysis you will be able to figure out a fragment of the protein's sequence?
- Load 2zal-2fofc_mod_3A.map, a 2Fo-Fc map calculated at 3 Å resolution. Note the difference between 4 Å and 3 Å resolution maps.
- Switch off the 4 Å map via Display manager. Try now to assign the side chains in the 3 Å map.
- Load 2zal-2fofc_mod_2_5A.map, a 2Fo-Fc map calculated at 2.5 Å resolution. Analyze the results as before.
- Now load 2zal-2fofc_mod_1_9A.map, a 2Fo-Fc map calculated at 1.9 Å resolution, corresponding to the actual limit of diffraction for this structure. You can switch all the previous maps off. Are the side chains recognizable now? If you still have problems, have a look at this figure, which illustrates how individual side chains might look in electron density maps.
- If you are still unable to assign the residue types, you can try to get help by looking at the amino-acid sequence of EcAIII in the 2ZAL_fasta.txt file.
Step-by-step instructions for modeling in Coot
- Load the coordinates from the PDB file 2zal_prot_wat-mod.pdb with a model of your protein. Go to Display manager and choose C-alphas representation of the molecule. Was your prediction about alpha-helices and beta-stands correct?
- Now the individual polypeptide chains have different colors. How many polypeptide chains are included in the model? Can you detect any non-crystallographic symmetry (NCS)? Try to describe the observed protein fold.
- If you want to walk along the polypeptide chain, use the space bar.
- Go back to Display manager and choose Bonds (Color by atom).
- Decrease map radius to 10 Å.
- Load 2zal-fofc_mod_1_9A.map, an Fo-Fc difference electron density map with 1.9 Å resolution. Remember to tick Difference map in Select map window.
- Now the recently loaded map is the active one. If you want to change it, go do Display manager and tick Scroll at the desired map. Note the sigma level for the difference map appearing at the top of display while scrolling. The green color corresponds to positive contours, the red color corresponds to negative contours.
- Go to Draw - Go to Atom and select residue 219B.
- Is the residue correct, i.e. does it correspond to the electron density? If not, go to Calculate - Model/Fit/Refine. Choose Simple mutate and select the residue to mutate and then the type of residue that you think would fit the map better. What residue should it be?
- Using Edit chi angles from the Model/Fit/Refine window try to model the side chain in the correct orientation. Watch the values of the torsion angles appearing at the top of the Coot display window. What are their final values? What values would you expect for this residue?
- Repeat the above steps for residue 42A. Hint: check the signal of the positive Fo-Fc peak. At what contour level does it disappear? What does it tell you about the number of electrons?
- Check the conformation of residue 58A. Correct it, if necessary, guiding yourself by the electron density map and the chemical environment of this residue. Did you notice any interesting interactions in this region?
- Go to Validate - Ramachandran plot, select 2zal_prot_wat-mod.pdb. Click on the residue marked as an outlier (red). Check the electron density to decide if the protein backbone conformation in this area is correct, i.e. if it agrees with the electron density map. If it is not, try to correct it. Use the Model/Fit/Refine panel and the Rotate/Translate zone tool. Hints: You define a zone by clicking on two residues. If you want to translate (move) only one atom, drag it using the left mouse button & Ctrl key.
- Check the final phi and psi angles via the Measures - distances & angles - Torsion, tool by clicking on the atoms defining the desired angle.
- Go to residue 70C. There you will find a positive Fo-Fc peak near the carbonyl oxygen atom. What can it be? Center your view on this peak. From the Model/Fit/Refine window select Place atom at pointer and select Water to be added to the 2zal_prot_wat-mod.pdb model.
- In the Model/Fit/Refine window select Real space refine zone (make sure you are using the correct map!) and double click on the newly added water molecule.
- Go to Measures - Environment distances and tick Show residue environment.
- Analyze the environment of this water molecule. Write down the interatomic distances. Is the current map interpretation correct? If not, remove the water molecule via Model/Fit/Refine - Delete - Water and select an appropriate atom from Place atom at pointer.
- Repeat the steps 15-18 for residue 311B.
- Go to residue 179B. Analyze its environment. Is there something missing in the electron density map? What could it be? Hint: have a look at the crystallization conditions.
- Go to residue 137A. Display symmetry-related molecules via Draw - Cell & Symmetry, tick yes in Master switch: show symmetry atoms. Try to interpret the difference electron density map. Use different contour levels to get an idea on the atomic number. Remember about the composition of the crystallization buffer and the chemical environment. For distance measurements you need to add a new atom as you have done in step 15.
- If time permits, you can load the asp_clean.pdb PDB file and model the ligand in the electron density map.
- Load the 2ZAL.pdb PDB file and check if your predictions were correct. Analyze the interactions of the species identified in step 21 above. If there are any inorganic ions found, select one of them and describe its coordination sphere.
- In this exercise you have seen electron density maps calculated at different resolution. How does the resolution level influence the map appearance? What structural features can be gleaned from an
electron density map at different resolution levels?
- While interpreting an electron density maps, what other information can be helpful in building the model?
- How would you discriminate between simple inorganic ions and water molecules?
- Your Report from this practical, as from all other practicals, should be completed and sent to me by email (email@example.com) before the date of the following practical class. Please include "SERP practical no. - your name" in the Subject field.
Mariusz Jaskolski, 28.03.2011