Water mediated contacts

The presence of water in the interface is pointed to by numerous NMR and X-ray experiments and previously studied by molecular dynamics simulations [1,2,10,11]. It has been argued that water allows for more movement in the interface, contributing thus to larger entropy and serving as a lubricant. Although the presented molecular dynamics simulations are not sufficiently long to draw any statistical conclusions on the life time of a water molecule in the interface, or frequency of forming a bridge between the macromolecules, one can still compare the four systems and obtain some information on behaviour and importance of water in different complexes.

 

Table 3.5: Average number of water bridges per snapshot in the four systems. Bridges to residues 50 and 51 were calculated separately in order to determine their relative importance.

	residue 50	residue 51	total
wild type	$1.2 \pm 1.1$	$0.5 \pm 0.7$	$2.0 \pm 1.5$
protein mutant	$0.0 \pm 0.2$	$1.3 \pm 1.4$	$1.4 \pm 1.4$
DNA mutant	$0.7 \pm 0.9$	$0.4 \pm 0.6$	$1.1 \pm 1.2$
double mutant (160 ps)	$0.6 \pm 0.7$	$0.6 \pm 0.8$	$1.2 \pm 1.1$

 

  

Figure 3.9: The number of water bridges to residues 50 (left) and 51 (right) in the four complexes.

The first test to provide some information is to determine the number of water bridges during the runs. A water bridge is defined as a contact between two heavy hydrogen bonding atoms (nitrogen or oxygen) of the macromolecules mediated by one water molecule through two hydrogen bonds to that molecule. A hydrogen bond is as in the case of direct contacts defined as a distance shorter than 3 Å. If a molecule mediates two bridges, both of them are counted separately.

Table 3.5 and Figure 3.9 show that as in the case of direct contacts, residue 50 is more important in the wild type complex, and residue 51 in the complex with the mutated protein. The total number of water bridges is, however, less informative since the averages are not significantly different in all complexes.

  

Figure 3.10: The interface of the wild type complex after 50 ps of the run. The green, red and blue balls are CPK representations of C₆ of Ade8, and C$\beta$ of Asn51 and Gln50, respectively. The dotted surface surrounds the interface defined by these atoms according to the text with 4 Å radius. The side chains shown are Ile47 (yellow), Gln50 (blue), Asn51 (red) and Met54 (white).

Another test to elucidate the role of water in the interface is to follow individual water molecules and the contacts they form to the macromolecules. In order to do this, one has to define the interface. Below the interface is understood as a sphere around a point between the macromolecules, whose coordinates are obtained by averaging the coordinates of selected atoms of the protein and the DNA. In particular, from visual inspection of the systems, the $\beta$ carbons of residues 50 and 51 and C₆ carbon of Ade8 in the $\alpha$-strand (XXATTA) were selected to define the center of the interface. The radius of the sphere was chosen to be 4 Å, so that it does not protrude beyond the cavity between the macromolecules. Figure 3.10 illustrates the interface in the wild type complex.

  

Figure 3.11: Contacts formed by two representative water molecules to the hydrogen-bonding atoms of residues 50 and 51 and the major groove side of base pairs 6 to 10 in the wild type complex (left) and protein mutant (right). The distances to the protein are in the upper graphs, and to the DNA in the lower.

  

Figure 3.12: Contacts formed by two representative water molecules to the hydrogen-bonding atoms of residues 50 and 51 and the major groove side of base pairs 6 to 10 in DNA mutant (left) and double mutant (right) complexes. The distances to the protein are in the upper graphs, and to the DNA in the lower.

Only those water molecules, which entered the interface were further considered. Furthermore, if a molecule stayed in the interface shorther than 10ps (from its first appearance to its last appearance) it was not considered as actually entering the interface and was discarded. In this way two molecules were discarded in the complex with the mutant DNA, one in the double mutant and in the protein mutant complexes, and no molecules were discarded in the wild type complex.

The general result is that in the wild type complex and in the double mutant there were fewer water molecules than in the other two (4 and 5 vs. 9 and 8). One of the five waters in the double mutant may furthermore be disregarded since it appeared after 160 ps and cannot be used in the present analysis, as explained in the section on the RMSD test. However, in order to be able to compare this complex with the others, the number of the remaining waters should be multiplied by 200/160, which gives 5 molecules in any case.

Another general result is that they form hydrogen bonds to the DNA bases more often than to the protein side chains. Only one water molecule in the ``wrong'' complexes keeps a hydrogen bond to Asn51 for a longer time (ca. 80 ps), all other waters bond to the protein occasionaly (up to ca. 20 ps). In the ``correct'' complexes on the other hand, bonding time to one side chain may last up to 180 ps in the wild type system and 120 ps in the double mutant.

There is, however, no significant difference in bonding to the DNA in the different complexes (see Figures 3.11 and 3.12). These figures also indicate larger mobility in the complexes with only one mutation. This result is also in agreement with the studies of free DNA, which indicate presence of long-lived ``water spine'' in the minor groove (residence times longer than 1 ns) and less ordered water molecules in the major groove (with residence time up to 500 ps) [12,13]. One can describe the water molecules at the DNA surface as belonging to the DNA, thus a protein recognizes a hydrated DNA.

  

Figure 3.13: Movement of water molecules in the wild type complex. Three molecules are followed in time. The colour of the oxygens gradually becomes whiter. The side chains of the protein shown are Ile47, Gln50, Asn51 and Met54. Colour code for the side chains is the same as in Figure 3.10. The cyan water molecule is the one described by Figure 3.11.

Finally, these simulations show that there are more water molecules in the interface of the ``wrong'' complexes, than of the wild type and double mutant systems. The waters in the two complexes with only one mutation form fewer contacts (per molecule) to the protein side chains, which indicates higher mobility and possibly larger distance between the macromolecules. The water molecules in the interface of the ``correct'' complexes form more stable contacts to the polypeptide. The contacts to the DNA are similar in all four complexes, which agrees with studies of the free DNA in solution. As in the case of direct contacts, residue 50 forms more water bridges in the wild type and double mutant complexs, whereas Asn51 forms most contacts in the complex with only the protein mutated.

Already during 200 ps simulation an individual water may enter the interface or leave it, however, at least one molecule stays during the whole run in the interface. An example of movement of water molecules in the wild type complex is given by Figure 3.13.