Sandbox Reserved 1440

This page describes validity of the proposed binding structure at the catalytic triad of the DPP-4 protein. There are a variety of ligands that were tested using electron density clashes and real space R-values to flag potential clashes in the current model.

What is a Validation Report?

A validation report (see linked image) [[1]] is an analysis of a molecular dataset model and the real space electron density of the proposed model. There are several factors that affect the validity of the model. Perhaps the most important is the resolution in Angstroms. A low resolution of less than 2 will likely have few electron density clashes, while a resolution of above 3 will have several clashes due to overlapping atomic radii.

Validaton report guidelines.

Rfree - measures the fit of the model using a subset of the data available.

RSRZ outliers - Real Space R values, measures fit of atomic model vs. the collected data in real space.

Clashscore - Derived from number of atoms in the model that are unusually close together.

Ramachandran outliers - Measures φ (phi) and ψ (psi) angle irregularities.

Sidechain outliers - Measures protein sidechain outliers and deviations from expected backbone angles.

JMol Images

This page will allow the user to view Jmol renderings of each protein along with the ligand in the binding pocket. The validation report will also identify the clashes in the model, which will be indicated as follows:

Red atoms indicate that there are significant clashes in the area of the displayed atom

Blue atoms indicate that there are less significant but still present clashes in the area of the displayed atom.

Overall Blue indicates a minor issue, while Red indicates a major discrepancy between the measured electron density and the location in the model.

The magnitude of any clashes are also indicated with the size of the atom where BIG atoms indicate a larger clash than small atoms.

3Q8W

Outside view of 3Q8W binding pocket (orange) with clashes shown

3Q8W is the structure with the worst resolution (3.64 Å), but the majority of the structure’s validation issues are due to its large clashscore. It has little to no RSRZ issues. The structure itself is valid, although its resolution is poor.

Click to see catalytic triad of 3Q8W: [[2]]

1NU6

1NU6 protein along with the ligand and catalytic triad displayed. The data was captured with a 2.10 Å resolution. The ligand NDG is not found in the binding site, rather it is bound to the outside of the binding pocket. This indicates that the ligand may be binding to an allosteric site and modulating the activity at the binding site. The ligands present in this model are two sugars (NAD and NDG), a Mercury (II) ion, and water. Of these, the mercury ion is closest to the triad, but is not close enough to represent a binding interaction. NDG however is found near the opening of the binding pocket, and thus may be an allosteric site for the protein.

The catalytic triad of the protein is shown in Figure 1b. Important to note that there are no clashes surrounding the site, and thus the model may provide an accurate representation of the binding capabilities of this binding pocket.

1PFQ

Image of the 1PFQ binding pocket and surrounding clashes

1PFQ has the highest number of RSRZ issues and a resolution of 1.9 Å. However, the largest gaps in electron density are located on the edges of the protein (red), and the majority of the gaps are relatively small (blue).

1PFQ’s binding pocket is surrounded by gaps in the electron density data. The residues of the catalytic triad do not have any missing data. Therefore, while the overall structure of 1PFQ isn’t particularly accurate, studies focusing on the catalytic triad should have accurate data.

Click to see the catalytic triad: [[3]]

5T4B

5T4B has a resolution of 1.76 Å and can be seen here with the 75N ligand within the binding pocket. There are more RSRZ outliers near the entrance to the binding cavity, but few clashes near the actual binding site. This indicates that the binding pocket itself has validity, but the mechanism for the ligand entering the cavity may be in question.

Shown in Figure 1c is a view of a the 75N ligand inside the binding pocket of DPP-4.

The 5T4B catalytic triad shown with electron density around the SER630, HIS740 and ASP706 residuals. There are no clashes at the triad.

To see the binding pocket of the 75N ligand at the catalytic triad, click on Figure 1d. There is electron density from the catalytic triad that extends into the binding pocket, and surrounding the functional binding groups on the ligand. No RSRZ outliers are found within the binding pocket, and this conformation of the molecule is supported by the electron density surrounding the catalytic triad.

6B1E

In the 6B1E protein with resolution 1.77 Å, the big density sizes are far from the ligand, which doesn’t have a significant impact on the binding pocket interaction. All of the missing electron density is not impacting the actual validity of the binding site.

Figure 1g:

In the 6B1E crystal structure, the catalytic triad seem to have some missing density but still not significant enough to be an invalid structure and interaction with the LF7 ligand.

Figure 1h:

4N8D

4N8D with resolution 1.65 Å: an overview of the binding site, missing densities, and clashes. Overall, the missing density sizes are not significant, which gives this crystal a valid structure. And the clashscore is not high, which means it’s not conflicting with the binding site. This crystal seems to be less valid than 6B1E in terms of the RSRZ score, but not enough to be claimed an invalid structure.

4N8D: the catalytic triads missing densities are shown. The densities of the amino acids are not significant enough to be viewed as an invalid structure, but does have more missing than the 6B1E crystal.

4A5S

Broad view of 4A5S molecule with binding pocket and clashes

4A5S has the highest resolution of all of the DPP4 structures at 1.62 Å. It has a relatively low amount of RSRZ issues, although the gaps are larger. Also, its RSRZ issues are generally only around the edges of the protein chain.

There are no clashes around the binding pocket [[4]], meaning this should be an accurate model. Also the electron density doesn’t show any gaps.

Structural highlights

This is a sample scene created with SAT to by Group, and another to make of the protein. You can make your own scenes on SAT starting from scratch or loading and editing one of these sample scenes.

Authors

Lana Jevtic, Kaeli Jacobson, Tayler Aarness, Eric Ruterbories.

St. Olaf College.

Medicinal Chemistry, Interim 2018.

Professor Robert Hanson.

Shown in Figure 1a is a close-up view of a morphinan antagonist (PDBid BF0) bound to the µ-opioid receptor. The green circles indicate hydrogen bonds.