| Introduction |
| The model must be realistic |
| Identifying charged sites |
| Identifying ionized residues |
| Ramachandran angles |
MOZYME is a technique that allows very large systems, in particular proteins and sections of DNA, to be modeled in a reasonable time. Although in principle, it is similar to conventional methods in that the underlying computational model is PM7 or PM6, etc., in practice there are several differences. Most of these differences are a result of the large size of the systems. In this page, some of the considerations will be discussed.
Most important of all, is the continuing requirement that in quantum chemical calculations each data set must represent a realistic system. In Molecular Mechanics, hydrogen atoms are often ignored or combined into the heavy atom that they are bonded to, to give the "combined atom" model. In quantum mechanical calculations, every hydrogen atom must be explicitly defined.
Some residues are usually described as being ionized, thus Arg is usually represented as [Arg]+, and Asp as [Asp]-. But if there is a large difference in numbers of anionic and cationic residues, the net charge on the system can be unrealistically large. Neutralizing individual residues requires making decisions that are hard to justify. To avoid having to make subjective decisions, a well-defined starting structure would be to use the completely neutral system. This is somewhat artificial in that there are almost certainly ions present in the system, but usually, during geometry optimization, protons will migrate from carboxylic acid groups to form carboxylate anions, and some Arg and Lys will become cations.
Systems involving most open-shell simple transition metal complexes, such as Cr(III) surrounded by six oxygen atoms, should not be run. If an attempt were made to run [CrIII(H2O)6]3+ it would fail because the number of electrons (51) is odd. Changing the charge from +3 to +4 would likely change the character of the chromium too much for the results to be useful.
Another example of a faulty starting geometry would be DNA in which every phosphate group was singly ionized. If the geometry of such a system was calculated, the results would be nonsense - the very large net negative charge would cause the DNA to spontaneously untwist, ruining the structure. Only when the phosphate groups are neutralized, or counterions are present to neutralize charges, can the characteristic twist be reproduced.
All ionized sites can be identified by keyword CHARGES. If this keyword is used, the charged sites and net charge will be printed, and the calculation stopped. This is a quick and reliable way of finding ionized sites. If an attempt is made to run a system with the wrong charge, either because a mistake was made in the structure, or because the assumed net charge was incorrect, the error will be detected and the job stopped. However, if keyword CHARGE is not present or GEO-OK is present, the supplied charge will be replaced with the calculated charge, and the job will be run using the calculated charge.
When any keyword that implies that the residue sequence should be used, e.g., PDBOUT or RESIDUES, or if the residue sequence can be deduced from the data set, then the net charge on each residue is printed. If the charge on a residue is larger than 0.5, the residue will be labeled "CATION", if it is less than -0.5, it will be labeled "ANION".
When RAMA is present, the three Ramachandran angles, ψ, φ, and α are printed for each residue.