Information Technology Reference
In-Depth Information
and Ile, Val, Phe (best three favored residues) at position 8. The nonameric and
octameric alleles show both similarities and differences in amino acids preferred at
various binding positions. Preferences for primary anchors show certain similarities:
all models exhibit some preference for small amino acids (H2-D
b
(Asn), H2-K
b
(Val). and H2-K
k
(Pro, Ala)), while C-terminal amino acids are strongly hydropho-
bic: H2-D
b
(Leu), H2-K
b
(Val), and H2-K
k
(Ile, Val). The most noticeable difference
between the nonameric and octameric alleles is at position 5, where H2-D
b
exhibits a
preference for polar Asn, while H2-K
b
shows a preference for Phe (aromatic hydro-
phobic residue) and H2-K
k
for Pro (small amino acid residue).
As well as refining and confirming our understanding of sequence dependence at
anchor positions, our results throw new light on all other positions within the
peptide. Although this study supports the importance of both primary and secondary
anchor residues, it is clear that other positions also play a key role in peptide-binding
(Hudrisier, Mazarguil, Laval, Oldstone, and Gairin 1996). Table 3 shows a summary
of residues associated with both favored and disfavored binding to all three alleles.
Looking at Table 3, for weak binding peptides, hydrophobic residues are present at
position 1 (Phe) and position 3 (Leu, Ile, Tyr, Phe) in abundance, and there is a prob-
able electrostatic repulsion of both negatively charged polar side chains (Asp and
Glu) and positively charged polar side chains (Lys, Arg, and His).
Each class I mouse MHC allele binds a mixture of structurally diverse peptides,
typically 8-10 amino acids in length, with each allele exhibiting defined peptide
specificity. From our work (Doytchinova and Flower 2002a; Doytchinova and
Flower
2002b; Doytchinova et al. 2002c; Doytchinova and Flower
2003;
Guan et al.
2003a; Guan et al. 2003b; Hattotuwagama et al. 2004), previous peptide binding
experiments, and X-ray crystallographic studies of human class I MHC molecules, it
is clear that the molecule binds short peptides, most of which are nonamers
(Bjorkman, Saper, Samraoui, Bennett, Strominger, and Wiley 1987). Topologically
position 1 corresponds to pocket A of the cleft of the peptide-binding site on
HLA-A*0201 (Saper, Bjorkman, and Wiley 1991). Anchor residues at position 2 and
at the C-terminus (position 9) are seen to be of primary importance for binding,
where pocket B interacts with the side chain of the residue at position 2. The struc-
ture of pocket A is mainly polar residues and consists of a network of hydrogen
bonding residues. A hydrophobic ridge cuts through the binding cleft forcing the
peptide to arch between position 5 and the carboxyl-terminal residue (position 9)
which are anchored into the D and F pockets in the floor of the cleft (Fremont,
Matsumura, Stura, Peterson, and Wilson 1992).
Equivalent data for mice show clear
differences and significant similarities. The crystal structure of several mouse class I
molecules has revealed that the peptide binding cleft is also closed at both ends, that
the length of the cleft is similar for all class I molecules (Fremont, Stura, Matsumara,
Peterson, and Wilson 1995; Zhang, Young, Imarai, Nathenson, and Sacchettini 1992;
Young, Zhang, Sacchettini, and Nathenson 1994; Smith, Reid, Harlos, McMichael,
Stuart, Bell, and Jones 1996a; Smith, Reid, Stuart, McMichael, Jones, and Bell
1996b), and that the carboxyl-terminal peptide position is an anchor residue deeply
buried in the F pocket. Analysis of the structure and binding results of the H2-K
b
and
H2-K
k
octameric complex reveals that there is a strong preference for an aromatic
and hydrophobic residues Tyr and Phe (H2-K
b
) and Leu (H2-K
k
) at positions 3 and 5
