The models that are most commonly built are those from Nicholson, although there are several other suppliers. For a 200 - residue protein you will probably pay about $2 per residue (please check this someone!). It will take you, on and off, about a month.
The normal way to build these is to build stretches of the backbone, add side chains, and adjust all the torsion angles using a special protractor. When you have done this, the subfragment (e.g. a helix) should be correct. But errors are cumulative, and here it may be helpful to compare your plastic creation with a picture on the screen. By rotating each you will be easily able to see any major errors.
Normally the model will be mounted on a base - ideally of perspex. Support rods should be measured (using the cartesian coordinates or a particular atom) so that x and y give the position on the base, whilst z gives the height. Some careful thought in working out where the supports go is well worth while. Corresponding holes are drilled in the atoms to take the support rods.
Hydrogen bonds can be extremely useful in stabilising the structure. Sometimes these are rods mounted into holes in atoms, whilst at other stages spring clips can be used. At this stage, also, any disulphide bonds will be made.
Now the elements can be mounted on their rods and fitted together. Also at this stage you may want to add a substrate, or other prosthetic group. The model requires final checking against the graphical display, and also careful inspection to see if there are bad non-bonded contacts.
Despite the wonders of virtual reality and computer graphics, you can still learn a great deal from physical models and they can be a great help to others, particular at communal meetings.
up to index
pmr for pps
Sept 18 1994