The discontinuous nature of the GST domain interfaces was analyzed by grouping amino acids into segments. Amino acids separated by less than 10 residues were allocated to a segment. Figure 12 (A) and (B) shows the segments that make up the N- and C-interfaces of a typical GST-family protein. It was found that the N-C interface residues could be grouped into five segments. The N1 and N2 segments are found in the N-interface. The C-interface is divided into the C1, C2 and C3 segments. Visual inspection of the interface segments shows that over 90 % of the N-C interface is α- helical with only one β-strand found in the N1 segment (see Figure 12 A and B). This classifies the domain interfaces of GST-family proteins as α-class interfaces (Jones and Thornton, 1995). Interestingly, in terms of CLIC1 and CLIC4 the bulk of the inter-domain contacts are formed through the N1 segment (h1 in case of CLIC1 and CLIC4). On the other hand the N2 segment (h3 in CLIC1 and CLIC4) forms only one inter-domain contact with segment C2 (h6 in CLIC1 and CLIC4).
Hydrogen bonds and/or salt bridges connecting an N-interface residue and a C- interface residue were referred to as N-C hydrogen bonds and/or salt bridges (see Appendix Tables C and D). Only 20 % of the inter-domain hydrogen bonds were formed between non-polar residues. Of the 80 % of polar residues involved in N-C hydrogen bonds, 60 % were charged (basic or acidic). Thirty four of the 40 pairs of GST-interfaces, including that of CLIC1, were found to form hydrogen bond networks. This is a common phenomenon in proteins where each donor/acceptor is bonded to multiple acceptors/donors (Stickle et al., 1992). The 40 analysed GST family proteins had a total of 2289 domain-interface residues that formed 316 inter- domain hydrogen bonds (0.14 bonds/residue), 194 inter-domain salt-bridges (0.08 bonds/residue), and in total, 510 inter-domain hydrogen bonds/salt bridges (0.22 bonds/residue). Figure 13 illustrates the break down of the above-mentioned contacts in terms of the various GST classes of proteins. The classes are arranged according to ascending number of total inter-domain hydrogen bonds/salt bridges. The monomeric
C3
C2
N2
N1
C1
C2
N1
N2
B
C-terminal
domain
N-terminal
domain
A
C3
C1
Figure 12: GST-family domain interface segmentation
Ribbon diagram showing a (A) front and (B) side view of the segments that make up the domain interface of a typical GST-family protein. The interface of the N-terminal domain is made up of two segments N1 (green) and N2 (red). The domain interface of the C-terminal domain consists of three segments C1 (cyan), C2 (blue) and C3 (pink). Interface residues separated by more than 10 residues were allocated to different segments. The structure used to represent the interface segmentation within the GST-family was picked randomly. It is that of Sj. GST (pdb file: 1gta). The model was generated using the Swiss-PDB viewer (Guex and Peitsch, 1997).
Elong ation facto r Grx2 PfZetaCLIC DeltaUre2PTau Sj Omeg a ThetaSigma Unkn ownBeta Phi Mu Alpha Pi 0 0.05 0.1 0.15 0.2 0.25 0.3 0.35 0.4 0.45 N o . bonds pe r r es idue GST class
Figure 13: Inter-domain hydrogen bonds and salt-bridges per amino acid of GST class proteins
Bar chart depicting number of inter-domain salt-bridges (blue), hydrogen bonds (red) and sum of hydrogen bonds and salt bridges (green) per domain-interface residue. The x-axis shows the different GST classes in ascending order according to the sum of hydrogen bonds and salt bridges per residue. In cases where GST classes were represented by more than one protein member the average of the number of domain-interface residues, hydrogen bonds and salt bridges was used to plot the bar chart. The average number of inter-domain salt-bridges, hydrogen bonds and sum of salt bridges and hydrogen bonds for the GST family are shown as blue, red and green dashed lines respectively. Inter-domain contacts were identified using iMOL server (http://i.moltalk.org) and confirmed by PPI server (http://www.biochem.ucl.ac.uk/bsm/PP/server/index.html) as well as visual inspection of the GSTs crystal structures. The cut-off distance for H-bonds was 3.9Å (<90°), while that for salt bridges was 4.0 Å. The chart was generated using Microsoft Excel.
Grx2 and CLICs (CLIC1 and CLIC4 used for analysis) form far fewer inter-domain contacts than the average for the GST family (see Figure 13). It seems that the number of hydrogen bonds between the N- and C-interfaces is proportional to the interface size although, the correlation coefficient was very small (R2 = 0.27, result not shown). The statistical significance of this correlation is supported by the observation by Jones and Thornton that the number of hydrogen bonds in dimeric interfaces is roughly proportional to the interface size (Jones and Thornton, 1995). In addition to the fact that, the average number of GST domain interface H-bonds (0.8 ± 0.38 per 100 Å2 of iASA) is comparable to the average number of dimer interface hydrogen bonds (0.88 ± 0.4 per 100 Å2 of iASA) as reported by Jones and Thornton (1995). No relationship was found between the number of N-/C-interface salt bridges and interface size (R2 = 0.15, result not shown). Interface ionic interactions were less common than hydrogen bonds with an average of 5 ± 3.4 salt-bridges compared to an average of 8 ± 4.0 hydrogen bonds per domain interface. This can be attributed to the relatively few number of charged residues found in the N-/C-interfaces of GST proteins.
The N-C hydrogen bonds and salt bridges were assigned to an interface segment (Figure 14). The highest numbers of contacts exist between N1 and C3, followed by N2-C1 and N1-C2. No contacts were found between N2-C3. Hot-spot residues were not favored to form N-C hydrogen bonds and salt-bridges and no apparent conservation of these contacts was established. Visual inspection of the segments revealed that most inter-domain hydrogen bonds and salt bridges are found on the opposite side of the domain linker. This suggests that these non-local contacts act as pegs clamping the N- and C-domain segments together and possibly forming during the later stages of folding.
The percentages of polar and non-polar residues in the N- and C-interfaces of 40 GST proteins were recorded (see Appendix Table B). Figure 15 illustrates the mean percentage hydrophobic and hydrophilic residues found in the N- and C-interfaces. The percentage non-polar residue means of the N- and C-interfaces were found to be 65.4 % ± 7.1 and 67.5 % ± 4.5, respectively. The percentage polar residue means of the N-and C-interfaces were 34.6 % ± 7.2 and 32.4 % ± 4.2, respectively. Thus on average, the number of non-polar atoms are approximately double that of the polar
N1-C1 N1-C2 N1-C3 N2-C1 N2-C2 N2-C3 0 40 80 120 160 200 240 Number of Bonds Interacting Segments
Figure 14: Total number of N-C interface hydrogen bonds and salt bridges in the GST-family
Bar chart depicting the total number of inter-domain hydrogen bonds (red), salt bridges (blue), and sum of hydrogen bonds and salt bridges (green) (see Appendix Tables C and D for full data). The x-axis, labelled interacting segments, shows the hydrogen bonds and salt bridges between the N-terminal (N1, N2) and C-terminal interface segments (C1, C2, C3). Inter-domain contacts were identified using iMOL server (http://i.moltalk.org) and confirmed by PPI server (http://www.biochem.ucl.ac.uk/bsm/PP/server/index.html) as well as visual inspection of the GSTs crystal structures. The cut-off distance for H-bonds was 3.9 Å (<90 °), while that for salt bridges was 4.0 Å. The chart was generated using Microsoft Excel.
N polar
N non-polar
C polar
C non-polar
% Residues in interface
0 10 20 30 40 50 60 70 80Figure 15: Hydrophobicity of the N- and C- interfaces belonging to GST-family members
Bar chart depicting the percentage of polar and non-polar residues found at the N- (blue) and the C- (red) interfaces belonging to GST-family members. The graph was generated by plotting the means of the N- and C- percentage polar and non-polar interface residues (see Appendix Table E for full data). The error bars represent the samples standard deviation. The dashed lines indicate the averaged CLIC1 and CLIC4 values of percentage polar and non-polar residues. For each member belonging to the GST- family the percentage polar and non-polar N and C-terminal interface residues were calculated using PPI server (http://www.biochem.ucl.ac.uk/bsm/PP/server/index.html). The chart was plotted using SigmaPlot v9.0
atoms. The percentage of non-polar amino acids present at the domain interfaces of CLIC1 and CLIC4 is higher than the average percentage for the GST family (see Figure 15). Hence, the domain interfaces of CLIC1 and CLIC4 are somewhat more hydrophobic than the rest of the GSTs. This is also true for Grx2, the other monomeric GST structural homologue (74.9 % and 71.9 % non-polar residues for the N- and C-domains of Grx2, respectively).