Protein stability

Reading material:   The DeepView manual (also available as a PDF file here) and the SPDV FAQ.

Goals: to gain a better understanding of how proteins are held together, and why some proteins are more stable than others.

Enzymes are used in a variety of industrial processes, many of them under extreme conditions such as high temperature or high pH. Common household items may also contain stabilized enzymes or ones engineered for special purposes. Examples of these are proteases, lipases and cellulases in washing powder, and amylase in flour.

Structural biology and protein-engineering experiments have provided an understanding of the basic protein stabilizing forces, which include hydrophobic and electrostatic interactions and entropic effects. Furthermore, studies of proteins from organisms living in extreme surroundings, such as high temperature, high salt etc., have given insights into why some proteins are more stable than others.

In this exercise, you will analyze mesophilic (PDB: 1st3) and thermophilic (PDB: 1thm) subtilisin. You will analyze the enzyme using different methods and also suggest and perform (in silico) mutations that you propose would stabilize the mesophilic form. (Remember, we wouldn’t want to change anything that affects the overall structure or activity of the protein, so knowing something about that is important, too! A brief overview can be found at: http://en.wikipedia.org/wiki/Subtilisin.

You need to be familiar with some of the basic commands, and know how to do mutations, in Swiss-PdbViewer (DeepView).

1. Conformational entropy

The entropy of the system is very important for the stability of a protein. By decreasing the conformational entropy of the main chain in the unfolded state, increased stability can be achieved.

- Glycine and proline

Two amino acids have significantly different conformational properties than the others: Gly and Pro. By changing a Gly residue into an Ala (or any other amino acid) increased stability might be achieved. The same effect could be obtained by replacing a non-proline residue with a Pro. However, the mutations should not destroy the ability to form secondary structure elements, which, for example, means that Pro should preferably be introduced in turns or loops, i.e. places where the structure is more likely to accommodate this sort of change.

Don’t forget to investigate the surroundings of the suggested mutations – will the introduced amino acid residues be happy here?

- Disulfide bridges

Disulfide linkages are also an effective way to reduce the conformational entropy of the unfolded protein. When designing such linkages you need to consider the geometry of the residues involved. Ideally, the C-alpha/C-alpha distance should be 6-8 Å, the side chains should point towards each other (or be able to do so, in a favorable side-chain conformation), and the distance between the side-chain gamma atoms (if they exist) should be less than 6 Å. You might find this site helpful here: http://cptweb.cpt.wayne.edu/DbD2/

Q2: Find a suitable place to create a disulfide linkage in mesophilic subtilisin (1st3).

- B-factors

Clues about which are the most flexible regions of the protein (and any other uncertainties) can be obtained from the temperature factors (the B-factors) for the atoms involved. The B-factors are given in the coordinate file (i.e. the pdb file). You can take a look at the actual coordinate file by clicking on the small document symbol on the left-hand side of the toolbar. The B-factors are in the very right-hand column (the last number). A B-factor of 30 Ų, for example, means a root-mean-square error in the atomic coordinate of 0.6 Å, while one of 80 Ų means a coordinate error of 1 Å (more than half the length of a C-C bond). In the program you can color the protein according to the B-factors. Lower B-factors are blue and higher are green, and higher still are red. Click on Color and select “B-factor”. 

Q3: Investigate the B-factors of subtilisin (1st3) and see if these can lead you to residues that could be good targets for stabilization. Suggest mutations that could improve the stability.

2. Electrostatic forces

- Stabilizing helices

Alpha helices are very common structure elements of proteins, and helix ”caps” are important determinants of protein stability. In addition, the introduction of a negatively-charged side chain at the N-terminal end of a helix will create favorable interactions with the partial positive charge that exists there because of the aligned carbonyl groups of the helix’s main chain (see lecture notes on protein structure). Similar effects have been reported for the removal of negatively-charged amino acid residues near the C-terminal end of alpha helices.

- Improving electrostatic contacts

An optimized network of electrostatic contacts can contribute substantially to the stability of a protein.

Q5: Investigate the subtilisin (1st3) structure and suggest two mutations that could improve the stability by introducing new electrostatic interactions. This could be in terms of improved hydrogen bonding networks or new salt bridges.

3. Holes and hydrophobic contacts

The core of globular proteins is mostly hydrophobic in nature. If there are any buried charges, they have to be counteracted by other factors such as opposite charges (salt bridges). In general, thermostable enzymes have fewer internal cavities than their mesophilic counterparts. Removal of such cavities can stabilize protein structures.

We will now take a look at the hydrophobic core of subtilisin (1st3) and analyze the packing of the core. Then we will try to find ways of improving that packing by filling in cavities and pockets.

We will use the web server CASTp found at: http://sts.bioe.uic.edu/castp/calculation.php.  This program identifies protein pockets and cavities based on a specified pdb file. The calculation uses a probe with a radius of 1.4 Å, which is the size of a water molecule, to search for holes in the structure. If this one isn't working, remember, backups are http://voronoi.hanyang.ac.kr/betacavityweb/ and DoGSiteScorer.

Enter 1st3 in the window marked “Type the 4 letter pdb ID:”, and click the “Submit” button. Jmol is preselected for visualization. The calculation might take a few moments. Be patient! The pockets are shown in different colours. On the left-hand side, you will find a list of the pockets. Here you get information on their volume and area. Select a pocket in the list to identify its position and to see which residues are lining it.

Q6: Select one of the pockets, and suggest a suitable mutation that could stabilize the protein by filling the pocket.

4. Comparisons with a thermophilic protein

Comparison of sequences as well as structures of thermostable proteins can give us clues about changes that could increase the stability. Therefore, load the 1thm file and superimpose it on the mesophilic 1st3 structure.

In addition to comparing the structures, you should also investigate the sequence alignments. Use the sequence alignment tools in the Fit menu of Swiss-PdbViewer and open the Alignment window to see the result. Put the cursor on the letters in the sequence align window; this will allow you to see the position of the amino acid residues in the structure.

  Q7: Compare the two proteins (amino acid sequences, 3D structures, B-factors) and see if you can find any differences that might be related to the higher stability of the thermophilic variant.

Extra exercise A. Conformational preferences

- Analysis of the protein structure: Ramachandran plot

The Ramachandran plot presents the backbone torsion (phi and psi) angles for residues in a protein structure (see lecture notes), and can help identify regions that have strained conformations. These places might be targets for genetic engineering aimed at relaxing those regions and so stabilizing the protein.

It is very easy to make use of the Swiss-PdbViewer program for the analysis. First, select "all atoms”, then choose "Ramachandran" under the "Wind" menu to calculate the plot. This plot shows how well the phi and psi angles in the structure conform to the most common (favorable) pairs of values. Right-handed alpha helices and beta sheets, for example, have very different phi/psi combinations, and so appear in two separate regions of a Ramachandran plot. Put the pointer on the symbols in the diagram and see where residues are located. Residues falling outside of the "allowed" regions are called outliers. These could actually be wrong in the structure (we’ll talk about that later). However, in most structures, there are residues that are “true” outliers, in the sense that their peptide bond angles are strained in the structure. So, one reasonable idea for a stabilizing mutation could be to relax strained parts of the structure. This often means that the residue is mutated to glycine, which can have almost any values for phi and psi. (Be careful, though, as some Ramachandran outliers are needed for reasons of structure or function.)

Or, you can make plots in PyMol with the Dynoplot plugin (https://pymolwiki.org/index.php/DynoPlot). First, go to the plugin manager and click on the "installed plugins" to check whether you already have it installed. If you don't , click on "install new plugin" and paste https://raw.githubusercontent.com/Pymol-Scripts/Pymol-script-repo/master/plugins/dynoplot.py in the URL box and click fetch. Once dynoplot is installed, you can write 'rama all' in the command line to visualize the plot, or select residues to view. If you encounter problems, and have Python installed in your system, PyRAMA is another (although less interactive) option.

Otherrwise, the sever at https://zlab.umassmed.edu/bu/rama/ coiuld be useful. Turn on Gly and Pro, and labels for outliers.

Investigate the 1st3 structure to find out if the secondary structure propensities might be improved. Secondary structure propensities are found in this file.

Extra exercise B. Literature search

Because subtilisin is a protease frequently used in detergents, there are many structures of this enzyme in the PDB. How many entries in the database can you find? What kind of mutants are there, and with which purposes were they created? Do you find any examples of mutations made to stabilize the protein?

Last modified: S. Mowbray, 6 September, 2022.