PIc m Spec PLC

y

Figure 3.8 Aligned backbone traces com paring the sim ilarity in solved PH dom ain structures

Ribbon plots com paring the structures of dynamin (Dyn), N-terminal pleckstrin (N-Plc), mouse p-spectrin (M-Spec) and PLC Ô1 PH domains. Amino and carboxy termini are indicated by N and C, respectively. The Btk PH domain could not be compared with the aforementioned struc­ tures because of the unavailability of its structural coordinates.

N-Plk Dyn m -Spec d-S pec PLC Ô1 Btk N-Plk Dyn m -Spec d-Spec PLC Ô1 Btk p i P2 p3 p4 M E P K R I R E G Y L V K K G S V F N T W ...K P N W V V L L E D . . . G L E F Y K K K ... E D N S P K P G M I P L K G S T S G N Q D E I L V I R K G W L T I N N I G I M K G G S K . E Y W F V L T A E N . . . L S W Y K D D ... E E K E K K Y M L S V D N L M E G F L N R K H E W E A H N K K A S S R S W H N V Y C V I N N Q . . E M G F Y K D A K S A A S G I P Y H S E V P V S L K E G S G T G A G E G H E G Y V T R K H E W D S T T K K A S N R S W D K V Y M A A K A G R I S F Y K D Q K G Y K S N P E L T F R G E P S Y D L Y N A D P D L Q A L L K G S Q L L K V K S S S W ... R R E R F Y K L Q E D C K T I W Q E S R ... K V M R T P E S Q L F S I E D I L E S I F L K R S Q Q . . . K K K T S P L N F K K R L F L L T V H . . K L S Y Y E Y D F ... E R G R R G S K K G S I D V E K P5 P6 P7 a l T L T S P C Q D F G K ... R M F V F K I T T T K Q ... Q D H F F Q A A F L E E R D A W V R D I N K A I K C I E G K L R D V E K G F M S ... S K H I F A L F N T E Q R N V Y K D Y R Q L E L A C E T . . Q E E V D S W K A S F L R A G V Y P E R V A I C E V A L D Y K K ... K K H V F K L R L S D G N ... E Y L F Q A K D D . S E M N T W I Q A I S S A A I E I A S D Y T K ... K K H V L R V K L A N G A ...L F L L Q A H D D T S M S Q W V T S L K A Q S D S T A I Q E V R M G H R T ...E G L E K F A R D V P E D R C F S I V F K D Q R N T . . . L D L I A P S . . P A D A Q H W V L G L H K I I H H S S I T C V E T V V P E K N P P P E R Q I P R R G E E S S E M S Q I S I I E R F P Y P F Q V V Y D E G P L Y V F S P T . . . E E L R K R W I H Q L K N V I

Figure 3.9 Sequence alignment of known PH domain structures

The sequences were aligned using the three-dimensional structures of dynamin (Dyn), N-terminal pleckstrin (N-Plk), PLC 61, drosophila and mouse p-spectrin (d-Spec and m-Spec, respectively) PH domains, p-strand and a-helix regions that are conserved between structures are indicated by p and a , respectively, above the aligned sequences. The sequence numbers are indicated for the dynamin PH domain, and secondary structural elements are highlighted red for p-strands and blue for a-helices.

strands p5 and p6 as well as an additional a-helix between strands p5 and p6.

Furthermore PLC 61 PH domain contains an a-helix at the amino terminus. p-Spectrin PH domains have an additional two-tum alpha helix in the loop between strands p3 and p4 which has no counterpart in the other solved structures. Amino acid sequence identity between mouse and drosophila p-spectrin PH domains is only 42%, but the two structures do closely resemble each other, the most obvious difference being the two residue gap at the end of the first a-helix in the mouse p-spectrin PH domain. The amino acids in the vicinity of this gap also show considerable sequence variability; consequently the a l-p 4 loop has a more extended conformation in the drosophila structure compared to mouse.

3.3.3 Comparisons of PH domains with other known structures

A powerful technique for providing insights into the possible structure-function relationships of modular domains is through the homology screening of structural databases. Comparative analysis of protein structural databases has identified several candidates with structures or topologies similar to PH domains. Their overall topology has been remotely likened to the a-p fold of the monomer subunit of E. coli verotoxin 1 (Downing et al., 1994) and the lipocalin family of proteins (specifically the retinol binding proteins) (Timm et al., 1994; Yoon et al., 1994), whereas the shape of the spectrin PH domain has been compared to the rapamycin binding protein FKBP (Macias et al., 1994). However, the most striking structural similarity is with the PTB domains of She (Zhou et al., 1995) and IRSl (Eck et al., 1996; Zhou et al., 1996).

The two PTB domains do not share significant sequence similarity or contain the conserved tryptophan found in the C-terminal a-helix of PH sequences. However, both domains have highly homologous structures (Figure 3.10). The backbone structures of core regions of the PTB and PH domains can be superimposed with a root mean square deviation of 1-1.9 Â, no greater than the deviation observed between PH domains themselves. However the divergence between PH and PTB structures is most obvious in the length and conformation of loops connecting elements of secondary structure. Specifically the PTB domain of She contains an extra N-terminal a-helix and a large insertion between the first and second p-strands. The remarkable identity of the PTB core with that of the PH domain leads to the supposition that the PTB sequence is a member of the PH family.

PTB dom ains, analogous to SH2 domains, can specifically bind to tyrosine phosphorylated proteins. However these two domains exhibit markedly different