Information Technology Reference
In-Depth Information
five residues that directly interacted with the peptide in the F pocket are identical in
both the A3 family and HLA-B27 (Leu81, Asp116, Tyr123, Thr143, and Trp147).
Arg and Lys bound to pocket F and interacted with negatively charged residues
Asp116 or Asp77 in both the A3 family and HLA-B27. B27 had been shown to
accept hydrophobic residues such as Leu, Ala, and Tyr because of their interaction
with Leu81, Tyr123, Thr143, and Trp147 in the binding pocket (Jardetzky, Lane,
Robinson, Madden, and Wiley 1991). In the present study, the specificity at position
9 was restricted to Arg and Lys only; both Ala and Tyr had deleterious effects on
peptide binding. This suggests a possible difference in the conformation of the bind-
ing pocket in spite of sequence similarity. Also, this may be the result of a change in
conformation after the binding of other amino acids in the peptide. A peptide-binding
motif for the HLA-A3 superfamily has been defined previously (Sidney, Grey,
Southwood, Celis, Wentworth, del Guercol, Kubo, Chestnut, and Sette 1996;
Rammensee, Friede, and Stevanovic 1995). Some useful similarities can be found on
comparing the present motif with those defined by the above two groups. The amino
acid preferences for the primary anchor residues are similar. All the motifs show
preference for Arg and Lys at position 9 and have a preference for various hydro-
phobic residues at position 2, such as Ile and Thr. The preferences for secondary
anchor residue positions 3 and 7 in the three motifs are hydrophobic amino acids
such as Phe.
The amino acid contributions to the affinity of peptides binding to the A2 family:
A*0201, A*0202, A*0203, A*0206, and A*6802 alleles using the Additive-PLS
Method have also been analysed quantitatively. Certain discrepancies between
A*6802 and A*02 molecules concerning the amino acid preferences at P1-P9 were
seen in the present study. These discrepancies throw doubt on whether the A*6802
allele belongs to the A2 supertype. The sequence comparison showed that there are
only one or two differences in the residues forming the six pockets of A*0201,
A*0202, A*0203, and A*0206 molecules. The number of these differences between
A*6802 and A*02 molecules is seven residues. Five of them concern pockets A, B,
and C and are so substantial that they alter the amino acid preferences at the primary
anchor P2 and the secondary anchors P1 and P6. The preferred Val and Thr for P2
brings the A*6802 allele closer to the A3 supertype (Sidney et al. 1996) rather than
to the A2 one. But the A3 supermotif requires positively charged residues, such as
Arg and Lys, at the C-terminus (Sidney et al. 1996), which is not true in the case of
A*6802. Obviously, A*6802 is an intermediate allele standing between A2 and A3
supertypes: in anchor position 2 it is closer to A3 and in anchor position 9 it is nearer
to A2. Residues identified as preferred for two or more A*02 molecules, without
being deleterious for any molecule, are considered as preferred. Residues identified
as deleterious for two or more molecules are considered as deleterious in the com-
mon motif. The expansion concerns all positions and especially the anchor P2.
The Additive-PLS results for the mouse alleles are in good agreement with pre-
vious studies of the preferred primary anchor positions: 5 and 9 (nonamers); 2, 5, and
8 (octamers - H2-K
k
and H2-K
b
, respectively). All three models also agree with
previous analyses of the preferred residue type at the anchor positions. For H2-D
b
:
Asn at position 5 and Leu at position 9; for H2-K
b
: Phe at position 5 and Val at
position 8; and for H2-K
k
: Glu, Pro, Gly (best three favored residues) at position 2
