Amino acid sequences structure prediction from

The ultimate goal of protein engineering is to design an amino acid sequence that will fold into a protein with a predetermined structure and function. Paradoxically, this goal may be easier to achieve than its inverse, the solution of the folding problem. It seems to be simpler to start with a three-dimensional structure and find one of the numerous amino acid sequences that will fold into that structure than to start from an amino acid sequence and predict its three-dimensional structure. We will illustrate this by the design of a stable zinc finger domain that does not require stabilization by zinc. 

Invent computer methods to predict the three-dimensional folded structure of a protein—and the pathway by which folding occurs—from its amino acid sequence, so information from the human genome can be translated into the encoded protein structures. 

Fig. 1. Secondary structure of E. colt ribosomal proteins LI 1 and SI 1 as predicted from their amino acid sequences. The prediction was carried out using four different methods represented by four different lines (S74, F82, N77, and R76). The line PRE summarizes the secondary structure obtained when at least three out of the four predictions are in agreement. The symbols represent residues in helical (A), turn or bend (B), extended (C), and coil (D) conformational states, respeaively. For details see Wittmann-Liebold et al. (1977b) and Dzionara rl cd. (1977).

Urease (urea amidohydrolase) is an enzyme first identified over a hundred years ago in bacterial extracts [22], The presence of urease is a virulence factor for some pathogenic bacteria [23,24], It is now known to occur also in plants, fungi, and invertebrates (see [24,25] for reviews). Urease from jack bean was the first enzyme to be crystallized, in 1926. Almost 50 years later its metal content was reexamined and it was found to contain two atoms of nickel per subunit [26]. Finally in 1995 the crystal structure of the enzyme from the enteric bacterium Klebsiella aerogenes was determined [27], Amino-acid sequence comparisons predict that the structures of the plant and bacterial enzymes are similar, although with different subunit arrangements. 

One of the ultimate goals of protein modeling is the prediction of 3D structures of proteins from their amino acid sequences. The prediction of protein structures rely on two approaches that are complementary and can be used in conjunction with each other ... 

Marsden, R. L., L. J. McGuffin, and D. T. Jones. 2002. Rapid protein domain assignment from amino acid sequence using predicted secondary structure. Protein Sci 11 2814-24. 

Ransom et al. 193 cloned the gene for mandelate racemase from Pseudomonas putida in Pseudomonas aeruginosa on the basis of the inability of the latter strain to grow on D-mandelate as a sole carbon source. The amino acid sequence was deduced from the nucleotide sequence, and the predicted molecular mass of the enzyme was 38 750[193. The enzyme is composed of eight identical subunits. The crystal structure of mandelate racemase has been solved and refined at 2.5 A resolution [194. The secondary, tertiary and quaternary structures of mandelate racemase are quite similar to those of muconate lactonizing enzyme 195, 196. Mandelate racemase is composed of two major structural domains and a small C-terminal domain. The N-terminal domain has an a + P structure, and the central domain has an a/p-barrel topology. The C-terminal domain consists of an L-shaped loop. 

Do you know the amino acid sequence (e.g., from cDNA cloning) and do you have purified protein Then you can determine the MW of the protein via MALDI-TOF and compare to the MW calculated from the cDNA sequence. If the MW are identical, the primary structure of your protein is correct. Deviations are evidence of modifications (e.g., phosphorylations, glycosylations, point mutations). With smaller proteins, the acetylation of the N-terminal amino acid becomes apparent in the MALDI-TOF. You determine the position of a modification by cutting the protein into peptides, measuring the MW of the individual peptides, and comparing to the MW predicted from the sequence. 

The protein folding problem is the task of understanding and predicting how the information coded in the amino acid sequence of proteins at the time of their formation translates into the 3-dimensional structure of the biologically active protein. A thorough recent survey of the problems involved from a mathematical point of view is given by Neumaier [22]. 

Chou P Y and G D Fasman 1978. Prediction of the Secondary Structure of Proteins from Tlieir Amino Acid Sequence. Advances in Enzymology 47 45-148. 

To understand the biological function of proteins we would therefore like to be able to deduce or predict the three-dimensional structure from the amino acid sequence. This we cannot do. In spite of considerable efforts over the past 25 years, this folding problem is still unsolved and remains one of the most basic intellectual challenges in molecular biology. 

Different side chains have been found to have weak but definite preferences either for or against being in a helices. Thus Ala (A), Glu (E), Leu (L), and Met (M) are good a-helix formers, while Pro (P), Gly (G), Tyr (Y), and Ser (S) are very poor. Such preferences were central to all early attempts to predict secondary structure from amino acid sequence, but they are not strong enough to give accurate predictions. 

The most common location for an a helix in a protein structure is along the outside of the protein, with one side of the helix facing the solution and the other side facing the hydrophobic interior of the protein. Therefore, with 3.6 residues per turn, there is a tendency for side chains to change from hydrophobic to hydrophilic with a periodicity of three to four residues. Although this trend can sometimes be seen in the amino acid sequence, it is not strong enough for reliable stmctural prediction by itself, because residues that face the solution can be hydrophobic and, furthermore, a helices can be either completely buried within the protein or completely exposed. Table 2.1 shows examples of the amino acid sequences of a totally buried, a partially buried, and a completely exposed a helix. 

Loop regions exposed to solvent are rich in charged and polar hydrophilic residues. This has been used in several prediction schemes, and it has proved possible to predict loop regions from an amino acid sequence with a higher degree of confidence than a helices or p strands, which is ironic since the loops have irregular structures. 

We have described a general relationship between structure and function for the a/p-barrel structures. They all have the active site at the same position with respect to their common structure in spite of having different functions as well as different amino acid sequences. We can now ask if similar relationships also occur for the open a/p-sheet structures in spite of their much greater variation in structure. Can the position of the active sites be predicted from the structures of many open-sheet a/p proteins ... 

What can be done by predictive methods if the sequence search fails to reveal any homology with a protein of known tertiary structure Is it possible to model a tertiary structure from the amino acid sequence alone There are no methods available today to do this and obtain a model detailed enough to be of any use, for example, in drug design and protein engineering. This is, however, a very active area of research and quite promising results are being obtained in some cases it is possible to predict correctly the type of protein, a, p, or a/p, and even to derive approximations to the correct fold. 

Figure 17.2 An example of prediction of the conformations of three CDR regions of a monoclonal antibody (top row) compared with the unrefined x-ray structure (bottom row). LI and L2 are CDR regions of the light chain, and HI is from the heavy chain. The amino acid sequences of the loop regions were modeled by comparison with the sequences of loop regions selected from a database of known antibody structures. The three-dimensional structure of two of the loop regions, LI and L2, were in good agreement with the preliminary x-ray structure, whereas HI was not. However, during later refinement of the x-ray structure errors were found in the conformations of HI, and in the refined x-ray structure this loop was found to agree with the predicted conformations. In fact, all six loop conformations were correctly predicted in this case. (From C. Chothia et al.. Science 233 755-758, 1986.)...

Homologous proteins have similar three-dimensional structures. They contain a core region, a scaffold of secondary structure elements, where the folds of the polypeptide chains are very similar. Loop regions that connect the building blocks of the scaffolds can vary considerably both in length and in structure. From a database of known immunoglobulin structures it has, nevertheless, been possible to predict successfully the conformation of hyper-variable loop regions of antibodies of known amino acid sequence. 

A particular goal of chemical theory is to predict protein structure from the amino acid sequence—to calculate how polypeptides fold into the compact geometries of proteins. One strategy is to develop methods (often based on bioinformatics) for predicting structures approximately and then refining the structures... 

---