Phylogenetic analysis of Sec7-domain-containing Arf nucleotide exchangers

Abstract

The eukaryotic family of ADP-ribosylation factor (Arf) GTPases plays a key role in the regulation of protein trafficking, and guanine-nucleotide exchange is crucial for Arf function. Exchange is stimulated by members of another family of proteins characterized by a 200-amino acid Sec7 domain, which alone is sufficient to catalyze exchange on Arf. Here, we analyzed the phylogeny of Sec7-domain-containing proteins in seven model organisms, representing fungi, plants, and animals. The phylogenetic tree has seven main groups, of which two include members from all seven model systems. Three groups are specific for animals, whereas two are specific for fungi. Based on this grouping, we propose a phylogenetically consistent set of names for members of the Sec7-domain family. Each group, except for one, contains proteins with known Arf exchange activity, implying that all members of this family have this activity. Contrary to the current convention, the sensitivity of Arf exchange activity to the inhibitor brefeldin A probably cannot be predicted by group membership. Multiple alignment reveals group-specific domains outside the Sec7 domain and a set of highly conserved amino acids within it. Determination of the importance of these conserved elements in Arf exchange activity and other cellular functions is now possible.

Figure 1.

Unrooted phylogenetic tree relating the Sec7-domain ARF-GEF family members. (A) Tree was constructed using neighbor-joining with core Sec7-domain gene sequences. Pie sections demarcate seven phylogenetically related groups of the Sec7 family plus one in mammals and one in bacteria. The bar on the lower left-hand shows distance. (B) Confidence measurements of phylogentic tree branches. Branches as in A, except that lengths no longer show phylogenetic distance and branch color indicates data from bootstrap analysis (1000 iterations). Colors range from red, indicating 50% bootstrap support, to green, indicating 100% support.

Figure 4.

Multiple alignments of the Sec7 domain arranged by group. (A) Visual representation of the multiple alignment strength in the Sec7 catalytic domain. As in Figure 3, except all cross-group homologies, including those with limited quality, are shown as colored areas. In addition, the gray areas indicating regions with no intragroup sequence similarity are embedded in white spaces whose width indicates the maximum insert size. (B) Positions of the HC amino acids identified here (further described in Table 4). (C) Positions of previously characterized mutations that affect Arf GEF activity. The top row indicates the wild-type sequence and the bottom row shows the target mutations. Color from red to green indicates the proportion (from 0 to 100%) of wild-type activity expressed by the mutant. Mutations that abolish Arf binding are indicated by ^*. CYH1 mutant coordinates: E157K, V179A, Y187A, M195A, D207A, K208A, K208E, R219A, and E157A (Betz et al., 1998). ARNO/CYH2 mutant coordinates: E117K, R152E, E156K, M194A, and N201A (Cherfils et al., 1998). (D) Arf contacts. The positions of the 15 amino acids that make contact with Arf in the cocrystal (Goldberg, 1998). See coordinates in F legend. (E) BFA sensitivity consensus. The positions of the amino acids, which are proposed here to define a consensus for BFA sensitivity (Table 5). See structural positions in G. (F) Arf contact sites. The Arf-binding surface in the hydrophobic groove is shown on the three-dimensional structure of the Sec7 domain of S. cerevisiae GEA2 (by x-ray crystallography). Numbered green solid arrows indicate the 10 predicted alpha helices. Light blue indicates predicted loops. Bright blue sequences correspond to the Arf-contact sites positioned in the hydrophobic grove as defined in the Arf–Sec7 domain complex (Renault et al., 2002). Yellow indicates the catalytic glutamic acid. The N-side of the groove is composed of helices α1–5, the glutamate finger loop, and N-the terminal half of helix α7. The C-terminal side includes helices α6, the C-terminal half of helix α7 and helices α8–10. Positions for amino acids are, in S. cerevisiae Gea2 coordinates: R650, L651, G653, S655, Q656, I658, D696, F699, I700, Y703, I706, M707, D711, and V717. (G) BFA sensitivity consensus. As in F, except that yellow sequences correspond to amino acids that are predicted to define BFA sensitivity shown in E and Table 5. Positions for consensus sites are in S. cerevisiae Gea2 coordinates: Y703, S704, M707, D711, and M721. (H) HC amino acids. HC amino acids are positioned on both sides of the Sec7 domain. Two 180° opposite views of the Sec7 domain are shown: front and back. Red indicates HC sites shown in B and Table 4. Limited overlap, 4/16, exists between HC amino acids and Arf contact sites of the Sec7 domain.

Figure 3.

Multiple alignments of whole proteins arranged by group. Visual representation of the multiple alignment strength in the Sec7-domain–containing proteins. Each group is shown in a different color and row. For a given row, colored bands indicate the strength of the alignment between members of that group and each other group for which there is significant alignment. For example, the orange areas in the BIG/SEC7 row indicate intragroup homologies, whereas the yellow areas represent BIG/SEC7-GBF/GEA intergroup homologies. Color intensities indicate average PAM scores between all pairs of sequences in the comparison, with gray indicating no sequence similarity (0 or less) to bright colors indicating strong sequence similarity (3). Tall gray areas indicate regions with no sequence similarity even within a group. The width of these gray areas is proportional to the average insert size. Boxes mark areas of large cross-group sequence similarity, including the central Sec7 domain. Names are attached to regions of sequence similarity either within or across groups. Actual sequences corresponding to these regions can be viewed at http://www.uic.edu/~nava/papers/SEC7.

Figure 2.

Distribution of Sec7-domain Arf-GEFs in different organisms. Trees are as in Figure 1 except that the thick black lines show the members of the family in each organism. All fungi have Sec7 proteins in the same four groups. Animals each have members in five groups. Plants have members in just two groups. F, human and mouse genomes also contain a novel gene, FXB8, which is only weakly related to any other group. B, in addition to the eukaryotic sequences, two related bacterial sequences were also found, which are not closely related to any other group.

Figure 5.

Model for the evolution of the Sec7-domain protein family. Based on the phylogenetic analysis of the eukaryotic Sec7-domain proteins, we propose a model for the evolution of the seven groups in fungi, animals, and plants, from a single Sec7 domain progenitor (see text for DISCUSSION). The PH domain is shown as a gray circle.