Introduction

Increased knowledge about the geometries of interacting groups within proteins, such as sidechain sidechain pairs, sidechain mainchain pairs, atom sidechain pairs and atom atom pairs can help further our understanding of the factors involved in the folding, stability and functioning of proteins. This requires detailed analysis of high resolution proteins to extract three-dimensional information. The first detailed study of interacting groups within proteins was carried out by Warme & Morgan (1978a,b) who analysed the interactions between sidechains and different atom types, o f which 19 types were identified (see Table 2.2). The authors were able to calculate the propensities of interaction between each of these 19 atom types and each o f the 2 0 different types of sidechain. This study did not, however, analyse the geometrical distribution of the atomic groups around the sidechains and thus could not provide information on their preferred packing arrangements. Since then a number of further studies have been carried out. These have included the distributions of oxygen and sulphur atoms around aromatic sidechains (Thomas et al, 1982; Reid et al, 1985), ion-pairs within proteins (Barlow & Thornton, 1983), carboxylate groups around arginine (Singh & Thornton, 1987), aromatic-aromatic sidechains (Burley & Petsko, 1985,1986a; Singh & Thornton, 1985; Blundell et al, 1986), amino-aromatic groups (Burley & Petsko, 1986b; Singh & Thornton, 1990) and hydrogen bonding atoms around amino-acid sidechains containing atoms capable o f hydrogen bonding (Ippolito et al, 1990). In addition to this, similar studies have been conducted on the distributions of metal ions (Chakraborti, 1989,1990; Yamashita et al, 1990), and water molecules (Thanki et a l,

1988,1990,1991) around sidechains. The main conclusions emerging from these studies are listed in Table 2.1 together with brief descriptions of the datasets employed.

Atomic Environments o f Arginine Sidechains Within Proteins

However, descriptions of the studies involving aromatic rings o f sidechains have been reserved for Chapter 3, which seems more appropriate as Chapter 3 deals with the atomic environments around the aromatic ring of phenylalanine. There have also been a number o f related studies regarding the rationalization of the spatial orientation of such interacting groups from an energetic viewpoint (Hunter et a/., 1991; Mitchell et a l, 1992).

Studies of interacting groups have also been based on small molecule data. These include studies conducted by Murray-Rust et a l (1983) and Vedani & Dunitz (1985). The data are available from the Cambridge Crystallographic Database (Allen et a l,

1979). One particular study to note is that of Murray-Rust & Glusker (1984) in that the methodology employed there is similar to the methodology used for the results

I analysis

presented in this and the following chapter. In theirj^Murray-Rust and Glusker (1984) selected compounds containing certain types of ester, ketones, epoxides and ether groups and examined the distributions of hydrogen bonding H-X groups (where X = N or O) within them. Contouring routines were used to identify and highlight regions o f high population density of the X atoms, which congregate in specific regions. The hope was that this approach could be extended to ligand-macromolecule interactions to see if similar patterns of ‘congregating atoms’ emerged. These patterns might then help predict the binding sites of chemical groups o f ligands around the macromolecules. At that time, however, there were too few known macromolecular structures to allow such an analysis, and it is only in the past few years that it has been possible to look at the distribution of chemical groups o f ligands around other chemical compounds in the proteins to which they are bound (Singh et a l, 1991; Moodie & Thornton, 1993).

The aim of this chapter, and the next, is to present a detailed analysis o f atomic distributions around two sidechains which have dissimilar natures. This chapter deals with arginine which has a hydrophobic chain capped by a polar positively charged guanidinium group. The following chapter deals with phenylalanine which is an apolar aromatic amino-acid. The atom types considered in both cases (See Table 2.2) are the

19 types defined by Warme & M organ (1978a) and the results show how their geometrical preferences differ for the two sidechains.

As can be deduced from Table 2.1 previous studies regarding the sidechain portion o f arginine have onljl focussec^on its hydrogen bonding properties (Ippolito et a l, 1990; Singh et a l 1987). Other works published on the hydrogen bonding properties of arginine, but not mentioned in Table 2.1 as they were approached from a different angle, include that by Baker & Hubbard (1984) and Mitchell et a l (1992).

Atomic Environments o f Arginine Sidechains Within Proteins

Table 2.1: Description of Previous Studies Regarding Sidechains

Year Conunents

1983 Barlow & Thornton

A total of 229 ion-pairs were identified in 38 proteins solved to resolutions of 2.5Â or less. Data from these proteins revealed that a third of charged residues are involved in ion-pairs and 76% of these stabilize tertiary structure^ rather than secondary structure.

1987 Singh

\el al­

Evaluated geometries of interacting arginine-carboxyl oxygens from 37 proteins resolved to 2.0Â or better. Carboxyl oxygens were found to aggregate in certain regions around

arginine where they could hydrogen bond

m s

Thanki et al.

Examined distribution of water molecules in 16 proteins of high resolution around main­ chain and sidechains of 20 amino-acids. Distinct non-random distributions of water molecules were found. Proportion of waters in contact with the sidechains showed an inverse correlation with the hydrophobicity whereas it was fairly constant for mainchains.

1989 Chakrabo

-rti.P

Examined geometry of metal ions around the sidechains of methionine and cysteine. They found that the geometry follows the same pattern in proteins and is independent of the nature of the metal centre. The metal ions usually bind to these residues involved in turns

or coils.

1990