From a PDB structure to an annotated figure

p53 is the protein that stops a damaged cell from turning into cancer. Many tumours survive by switching it off, and one of the ways they do that is a second protein, MDM2, that clamps onto p53 and keeps it quiet. Drugs that pry those two apart are a real and active area of cancer research.

In this first RBAi Academy lesson you will open the structure of that exact grip — a piece of MDM2 holding a piece of p53 — and turn it into a clean, labelled figure you could put in a talk. You will not write a line of code, install PyMOL (the usual desktop program for drawing molecules), or download anything from a separate database. You ask RBAi for what you want, in plain English, in the RBAi main chat: the same chat where you do the rest of your research, and where every prompt in this lesson goes.

We will use PDB 1YCR (entry 1YCR in the Protein Data Bank, the public archive of solved 3D structures): a fragment of human MDM2 bound to a short p53 peptide. It is small, famous, and a perfect first structure.

What you will make

The grip. MDM2 is the light grey-blue surface (a drawing of the protein's outer shape), with the residues that line its binding groove shaded darker. The orange ribbon (a line that traces the protein chain) lying in that groove is the p53 peptide, seen from the side it binds on. PDB 1YCR.

The annotated pocket. MDM2 is the pale ribbon, and the slate sticks are the MDM2 side chains (the R groups that make each amino acid different) that line the pocket. The three p53 side chains that do most of the gripping, Phe19, Trp23 and Leu26 (orange), plug into it. The dashed line is a hydrogen bond (a weak attraction between a hydrogen on an N or O atom and another N or O nearby) from the ring nitrogen of Trp23 to the backbone oxygen of MDM2's Leu54: 2.83 Å in the deposited coordinates. Phe19 means the phenylalanine at position 19 of the chain (three-letter amino-acid code plus position); Trp is tryptophan and Leu is leucine.

Both figures were drawn by RBAi's PyMOL figure renderer from the unmodified 1YCR coordinates; nothing was drawn or edited by hand. The prompts below ask RBAi for the same two scenes. Your first render may frame them a little differently, and asking for a specific change is part of the lesson.

1. Open the structure

Start a new chat in RBAi and paste this:

Open PDB 1YCR and show it as a 3D structure. In plain language, tell me what I am looking at: which chain is MDM2 and which is p53, the residue numbers that are actually modelled, and the experimental method and resolution. Do not change the structure.

RBAi fetches the entry, opens it in an interactive viewer you can rotate (the 3D structure card in the chat), and explains it. You do not need to find the file anywhere else first.

What you should see: MDM2 is chain A (residues 25–109) and p53 is chain B (residues 17–29). A chain is one separate protein strand in the file, named by a letter, and a residue is one amino acid in it, numbered along the protein. The structure was solved by X-ray diffraction at 2.6 Å: the atom positions were worked out from how a crystal of the complex scatters X-rays, and 2.6 Å is the resolution, the finest detail the data can show (smaller is sharper; one angstrom, Å, is 0.1 nanometres, about the width of an atom). Those residue ranges are the parts present in this model — not the full-length proteins.

2. Ask for the overview figure

Stay in the same chat. You describe the picture you want; RBAi does the selections (which atoms to show, and how) and the rendering (drawing the final image).

Make a publication-quality figure of this complex: MDM2 (chain A) as an opaque light grey-blue molecular surface, with the residues that line its binding groove shaded a darker blue, and the p53 peptide (chain B) as an orange cartoon lying in the groove. Look straight at the groove from the side p53 binds on, on a white background. Keep the original coordinates.

That gives you the first picture above: the peptide lying in a groove on MDM2's surface. Coordinates are the 3D positions of every atom in the file, so keeping them means the picture changes but the structure does not. Rotate the structure in the viewer to see the groove from a few angles before you settle on one.

3. Zoom in and highlight the key residues

The overview shows where p53 binds. The next figure shows how. Ask for a close-up:

Now make a close-up of the binding pocket with no surface: MDM2 as a pale grey-blue cartoon with the side chains that line the pocket as thin slate sticks, and p53 as an orange cartoon with Phe19, Trp23 and Leu26 as orange sticks. Draw the hydrogen bond between the Trp23 ring nitrogen and the backbone oxygen of MDM2 Leu54 as a dashed line and tell me its length. Label Phe19, Trp23, Leu26 and Leu54. Keep the deposited numbering.

In our run the Trp23 to Leu54 hydrogen bond measured 2.83 Å.

Sticks draw every bond as a short rod, so you see each atom of a side chain, and the deposited numbering is the residue numbering in the original PDB entry. Open the result at full size and check it like an editor: are all three side chains visible, are the labels on the right residues, is anything cut off at the edge? If something is wrong, say exactly what — "make the labels larger" or "turn it so Trp23 isn't hidden behind the helix." Asking for a specific fix and looking again is the normal way to work; it is not a failure.

4. Ask why those residues matter, then download

Explain why Phe19, Trp23 and Leu26 are the important residues here. Keep separate what the coordinates show and what comes from the original paper. Then give me the two figures and the original PDB file to download.

Then take your files straight from RBAi:

• The structure: on the 3D structure card, use Download → Structure (.pdb).
• A figure: use Expand → Download on the image.

A render may take a little while; if it runs as a background job, wait for RBAi to mark it done and show the file before asking again.

What the picture shows — and what it doesn't

• What you can point to: three p53 side chains fitting into a groove on MDM2's surface. That much is in the coordinates.
• What comes from the paper: that this grip is how MDM2 holds p53, and why blocking this groove is a drug-design strategy — from Kussie and colleagues (1996).
• What it cannot tell you: how tightly they bind, how the two move in solution, or any clinical effect. A single static structure does not measure those, and colour marks a component, not an energy.

Your turn

Before you read RBAi's explanation, write three sentences yourself: one thing you can see in the figure, one thing that comes from the paper, and one thing the picture cannot establish. Then compare. Learning to keep those three apart is the real skill — the figure is just how you practise it.

Try it

Open RBAi, paste the first prompt, and build your own version of these figures. The structure used here is the unmodified deposited entry — you can download the exact 1YCR file and check it against RBAi's.

RBAi is in beta, so the exact wording and the first render can vary with the model and your account. That is normal — you ask for a surface view, then a labelled close-up, and refine until each one reads well. The two figures in this lesson were made in RBAi exactly that way.