Classification scheme of H3-rules
H3-rules were first proposed from the analysis of 45 antibody structures .
We have classified antibody sequences into 4 patterns as bellow
and, therein we have designated them a each base type, Kink or Extend.
After this proposal, H3-rules were conformed from the analysis
of other newly determined 55 structures .
Recently, we reclassified the antibody structures more precisely
and rivesed H3-rules using updated 315 datset .
Four sequences patterns:
i-a. the (n-1)st residue not is Asp.
i-b. the (n-1)st residue is Asp and the 0th residue is not basic.
i-c. the (n-1)st residue is Asp and the 0th residue is basic.
i-d. the (n-1)st residue is Asp and the 0th and (-1)st residue are both basic.
rule i. Base type identification (Kinked, K, or Extended, E, base).
simplified flow chart here.
Pattern i-c "always" take a K form.
Pattern i-d "always" take an E form.
Pattern i-a and i-b are more complex... See ref .
rule ii. Further classification of each base type.
Either when position (n-3) is Trp, or
when position (n-2) is Gly and (n-3) is a large hydrophobic residue,
the second bulge is additionally inserted in the kinked base.
rule iii. Formation of the hydrogen bond ladder.
iii-a When Pro is located at positions 2, 4, m-3, or m-1,
the beta-hairpin should be broken.
iii-b When the base is kinked, the above rule iii-a is not satisfied,
and both the 2nd and the (m-1)st residues or both the 3rd and (m-2)nd residues
are occupied by aromatic residues,
a beta-hairpin with a hydrogen bond ladder is formed.
rule iv. Formation of the beta-turn.
schematic diagram of beta-hairpin here.
When the above rule iii-a is not satisfied,
and Gly, Asp, or Asn is located at the position of (m/2 or m/2+1 in class A) ,
at ((m+3)/2 in class B), or at (m/2+2 in class C),
a typical beta-turn in the corresponding class is formed.
Also, when m equal to 5 and the 3rd residue is Gly, a typical turn conformation is formed.
Note: rule iv works well when m is shoter than 9.
When m is from 9 to 12, only if CDR-H3 is also satisfied with rule iii-b,
rule iv works well.
Shirai, H., Kidera, A., Nakamura, H. (1996) FEBS Lett. 399, 1-8.
Shirai, H., Kidera, A., Nakamura, H. (1999) FEBS Lett. 455, 188-197.
>Kuroda, D., Shirai, H., Kobori, M., Nakamura, H. (2008) PROTEINS. 73(3):608-620.
Return to the Laboratory of Protein Informatics