Diversity in the architecture of ATLs, a family of plant ubiquitin-ligases, leads to recognition and targeting of substrates in different cellular environments

Abstract

Ubiquitin-ligases or E3s are components of the ubiquitin proteasome system (UPS) that coordinate the transfer of ubiquitin to the target protein. A major class of ubiquitin-ligases consists of RING-finger domain proteins that include the substrate recognition sequences in the same polypeptide; these are known as single-subunit RING finger E3s. We are studying a particular family of RING finger E3s, named ATL, that contain a transmembrane domain and the RING-H2 finger domain; none of the member of the family contains any other previously described domain. Although the study of a few members in A. thaliana and O. sativa has been reported, the role of this family in the life cycle of a plant is still vague. To provide tools to advance on the functional analysis of this family we have undertaken a phylogenetic analysis of ATLs in twenty-four plant genomes. ATLs were found in all the 24 plant species analyzed, in numbers ranging from 20-28 in two basal species to 162 in soybean. Analysis of ATLs arrayed in tandem indicates that sets of genes are expanding in a species-specific manner. To get insights into the domain architecture of ATLs we generated 75 pHMM LOGOs from 1815 ATLs, and unraveled potential protein-protein interaction regions by means of yeast two-hybrid assays. Several ATLs were found to interact with DSK2a/ubiquilin through a region at the amino-terminal end, suggesting that this is a widespread interaction that may assist in the mode of action of ATLs; the region was traced to a distinct sequence LOGO. Our analysis provides significant observations on the evolution and expansion of the ATL family in addition to information on the domain structure of this class of ubiquitin-ligases that may be involved in plant adaptation to environmental stress.

Figure 1. Number of retrieved ATLs in 24 plant species.

The phylogenetic relationship between the 24 plant species is displayed at the bottom; it was adapted from the National Center of Biotechnology Information (NCBI) taxonomy server (http://www.ncbi.nlm.nih.gov/Taxonomy). The three major groups of plants are: basal plants (blue), monocots (green) and eudicots (orange). The species abbreviation is: ppp, Physcomitrella patens; smo, Selaginella moellendorfii; osa, Oryza sativa; bdi, Brachypodium distachyon; sit, Setaria italica; zma, Zea mays; sbi, Sorghum bicolor; aco, Aquilegia coerulea; mgu, Mimulus guttatus; vvi,Vitis vinifera; egr, Eucalyptus grandis; ccl, Citrus clementina; csi, Citrus sinensis; cpa, Carica papaya; aly, Arabidopsis lyrata; ath, Arabidopsis thaliana; csa, Cucumis sativus; mtr, Medicago truncatula; gma, Glycine max; ppe, Prunus persica; mdo, Malus domestica; pop, Populus trichocarpa; rcu, Ricinus communis; mes, Manihot esculenta.

Figure 2. A broad view of domain architecture in ATLs.

(A) The sequence LOGO represents the consensus sequence of the ATL RING-H2 domain. Numbers indicate the residues involved in zinc ligation, asterisks indicate the residues conserved in RING fingers, and arrowheads residues conserved in ATLs. (B) Schematic representation of a canonical ATL with its seven regions (I–VII). The second column displays the size diversification in number of amino acids sorted independently according to the statistic percentile.

Figure 3. Phylogeny of 1815 ATL proteins from 24 plant species.

The 42 amino acid sequence of the RING-H2 was used to generate the tree with FastTree; the tree is displayed in Figure S1. The branches were classified in 9 groups, A to I, collapsing branches with local support below 80%. Basal species are colored in blue, monocots in green and eudicots in orange. An outside circle represents presence of one or more transmembrane regions. ATLs have one or more transmembrane helix region toward the amino-terminal end, according to TMHMM, these are shown on a scale of shading. The basal gray color corresponds to one transmembrane helix.

Figure 4. Distribution of ATLs from 24 plant species in 9 groups.

Heat map representation of the number of ATLs from the 24 species in each one of the nine groups by a gray scale. The total number of genes in each group is shown at the bottom (the distribution of the 1815 ATLs in 9 groups in displayed in Table S2). The species tree is as in Figure 1.

Figure 5. Sequence LOGOs mapped to ATL regions.

Below the canonical ATL diagram, geometric figures represent the seven regions and the sequence LOGOs mapped to each region are shown. The total number of logos per region is indicated at the bottom. The catalog of the 75 sequence LOGOs is shown in Table S3.

Figure 6. Distribution of 75 sequence LOGOs on the 9 ATL groups.

A heat map shows the frequency of the conserved motifs in each of the nine ATL groups. Asterisks on top point to sequence LOGOs that are specific to a group. The GLD motif and RING-H2 are widely spread among all groups except for the GLD motif that is absent from group B. Roman numbers I–VII represent the seven ATL regions; the frequency of conserved motifs is shown according to a blue scale. The species tree is as in Figure 1.

Figure 7. Inferred evolutionary history and organization of ATL10.

(A) The inferred evolutionary history of ATL10 is depicted. A segmental duplication event involving chromosomes 1 and 3 generated ATL78 and ATL75, from ATL81 and ATL77, respectively; ATL10 was then generated by tandem duplication events from ATL75. (B) The upper diagram depicts the modular organization of ATL10 based on sequence LOGOs mapped to it. Below is the schematic representation of seven clones encompassing different regions of ATL10 that were used in yeast two-hybrid assays.

Figure 8. Interaction of DSK2a/ubiquilin with ATL10.

(A) Schematic representation of DSK2a/ubiquilin showing UBL, UBA, and STI-like domains and derivative clones generated for the yeast two-hybrid assays; ΔUBL represents a clone obtained from the yeast two-hybrid screening. DSK2a/ubiquilin-containing fragments were ligated into the activation domain of pGADT7. (B) Representative plates showing yeast two-hybrid interactions between DSK2a/ubiquilin derivative clones and ATL10(I); the left panel shows the template of the plates. The yeast strain AH109 was cotransformed with pGBKT7 and pGADT7 derivatives, selecting for transformants on SC lacking Trp and Leu. Transformants were then streaked onto SC medium lacking Trp and Leu (SC) and onto SC medium lacking Trp, Leu, His and Ade (high-stringency selective conditions). The plates were incubated at 30°C for four days; growth is seen as dense streaks of yeast over background. The interaction of ATL10(IV–VII), a clone lacking region (I), with ΔUBL was also included as a negative control.

Figure 9. Interaction of DSK2a/ubiquilin with diverse ATLs.

(A) Representative plates showing yeast two-hybrid interactions between DSK2a/ubiquilin and region I from six ATL genes; the left panel shows the template of the plates. The 6 ATL-containing fragments were ligated into the DNA-binding domain of pGBKT7, the two DSK2a/ubiquilin clones, ΔUBL and ΔUBLΔUBA (depicted in gray in the template plate), are as in Figure 8. The yeast two-hybrid assay was performed as in Figure 8. (B) A sequence LOGO in region I of ATLs. The MEME suite was used to generate sequence LOGOs from the 5 yeast two-hybrid clones shown in (A) and from the ATL2 clone. A sequence LOGO resulted from the region highlighted in yellow in the five clones; this region was not found in ATL2. As comparison, this sequence LOGO placed over the sequence LOGO 71 is shown below.

Figure 10. The ATL10 RING-H2 domain interacts with the ubiquitin-conjugase UBC11.

Representative plates showing yeast two-hybrid interactions between UBC11 and ATL10 clones carrying the RING-H2 domain. A scheme of the seven regions in ATL10 is shown above as a guide. The left panel shows the template of the plates. ATL10-containing fragments were ligated into the DNA-binding domain vector pGBKT7 and UBC11 was recombined into the activation domain of pHOST pACT2 Lox vector; description ATL10 segments is shown in Figure 7, region IV corresponds to GLD, region VI to the RING-H2 domain and the asterisk to a mutated RING-H2 domain. The yeast two-hybrid assay was performed as in Figure 8, except that transformants were then streaked onto SC medium lacking Trp and Leu (SC), into SC medium lacking Trp, Leu, and His, and onto SC medium lacking Trp, Leu, His, and Ade. No growth was observed after 7 days of incubation on SC medium lacking Trp, Leu, His, and Ade; this plate is not shown.