Table 2.

Specificity-determining positions in water and glycerol transport proteins from the MIP family

Position		Amino acid residues in proteins of the training set			Description of the position
GlpF (E. coli)	AQP1 (bovine)	AQP group	GLP group	Z-scores	Water contacts, Å	AQP1 channel side	Glycerol contacts, Å	GlpF channel side	Interhelical interactions
207	198	SSSSSSS	DDDDDDDDDD	7.65E–12	5–7	Channel ?	5–7	Channel
236	214	FFFFFFF	PPPPPPPPPP	1.25E–11	5–7	Channel	7–10	Channel
48	58	FFFFFFF	WWWWWWWWWW	1.33E–11	<5*	Channel	<5*	Channel
135	127	GGGGGNG	FFFFFFFFFF	4.95E–11	7–10	Not helix	5–7	Not helix
159	151	FFFFFFF	LLLLLLLLLL	7.08E–11	<5*	Channel	<5*	Channel ?
187	178	LLLLLLL	IIIIIIIIII	8.33E–11	<5*	Channel	<5*	Channel
22	25	VVVVVVV	IIIIIIIIII	2.84E–09	5–7	Channel ?	7–10	Channel ?
195	186	IIIIILI	GGGGGGGGGG	3.88E–09	7–10	Channel ?	<5*	Channel ?
191	182	HHHHHIH	GGGGGGGGGG	1.60E–08	<5*	Channel	7–10	Channel
201	192	WSSSSSS	AAAAAAAAAA	2.48E–07	<5*	Not helix	<5*	Not helix
108	118	AAAAALA	YYYYYFYYYY	1.80E–06		Channel ?		Channel	+
137	131	LHHHHLH	TTTTTTTTTT	1.12E–05	5–7	Not helix	<5*	Not helix
211	202	AAAAAAA	KKKKRRRRRR	3.74E–05		Not channel	7–10	Channel ?
43	53	GGGGGGG	EEEEVVVVVV	7.92E–05		Not channel	7–10	Not helix	+
136	135	EEDEEEE	SSSSSSCSAS	9.77E–05		Not helix	5–7	Not helix
195	190	NNNNNGN	GGGGGGGGGG	1.02E–04	5–7*	Not channel	<5*	Not channel
194	185	SSSSSGS	MMMMLLLLLL	1.78E–04	7–10	Channel	7–10	Channel
24	27	GGGGGVG	FFFFFLLFLL	2.74E–04		Not channel	7–10	Not channel	+
20	23	WWWWWMW	LLLLLILIII	3.14E–04		Not channel	7–10	Not channel	+
200	191	TTTTTLT	FFFFYYYYYY	3.17E–04	<5*	Not helix	<5*	Not helix
193	184	VIIVIVI	SAASSSSSSS	8.53E–04		Not channel		Not channel	+

Asterisk (*): residues contacting the substrate (Fu et al. 2000; Sui et al. 2001).

Contacts: the minimal distance between the atoms of the current amino acid residue and the atoms of the substrate.

Channel side: orientation of the residues with respect to the channel, identified as follows: a vector perpendicular to the helix axis and pointing to the most exposed surface of the helix is calculated based on the residues solvent accessibility data published in the DSSP database (www.sander.ebi.ac.uk/dssp); the channel vector of a transmembrane helix is defined as the one opposite to the above defined vector; then for each residue the radius vector is computed as the vector perpendicular to the helix axis and pointing to Cα; finally, if the angle between the radius vector and the channel vector is smaller than 45°, the residue is labeled “channel”; if the angle is larger than 45° but smaller than 90°, “channel?”; in all other cases, “not channel.”

Not helix: residues not belonging to TM helices according to the secondary structure description in the corresponding PBD file.