Evolutionary trajectory of transcription factors and selection of targets for metabolic engineering

Abstract

Transcription factors (TFs) provide potentially powerful tools for plant metabolic engineering as they often control multiple genes in a metabolic pathway. However, selecting the best TF for a particular pathway has been challenging, and the selection often relies significantly on phylogenetic relationships. Here, we offer examples where evolutionary relationships have facilitated the selection of the suitable TFs, alongside situations where such relationships are misleading from the perspective of metabolic engineering. We argue that the evolutionary trajectory of a particular TF might be a better indicator than protein sequence homology alone in helping decide the best targets for plant metabolic engineering efforts. This article is part of the theme issue 'The evolution of plant metabolism'.

Keywords: anthocyanins; biosynthesis pathway; phylogenetic relationship; specialized metabolism.

Figure 1.

Identification of maize C1 homologous genes and their use for metabolic engineering in various plant species. R2R3-MYB TFs corresponding to maize C1 were identified by sequence similarity-based gene predictions followed by functional characterization (dark blue arrows) or by mutant/population-based gene identification followed by phylogenetic relationship analysis (light blue arrows). The references for the TFs mentioned are as follows: C1 from maize [46] and rice [47]; AN2 from tobacco [48], petunia [49] and potato [50]; ANT1 from tomato [51]; MYB1 and MYB10 from apple [–54]; PAP1/PAP2 [55] and TT2 [56] from Arabidopsis; MYBAs [57,58] and MYBPAs [59,60] from grape; Rubys from citrus [61]; MYB10 from gerbera [62]; MYB1 from eggplant [63]; MYBs from strawberry [64,65]; MYB1 from radish [66] and Ros1/Ros2/Ve from snapdragon [67].

Figure 2.

Phylogenetic relationship of flavonoid R2R3-MYBs TFs from different species. The amino acid sequences of the R2R3-MYB domains were aligned using MUSCLE, and the phylogenetic tree was reconstructed using the neighbour-joining method and 1000 bootstraps implemented in MEGA (v. 7.0). Bootstrap values greater than 50 are displayed on the corresponding phylogenetic branch. The species information and accession numbers of the proteins shown are as follows: Arabidopsis: AtTT2 (AT5G35550), AtPAP1 (AT1G56650), AtPAP2 (AT1G66390), AtGL1 (AT3G27920), AtWER (AT5G14750); maize: ZmC1 (GRMZM2G005066), ZmP1 (GRMZM2G084799); beet: BvMYB1 (M1ETK3.1); rice: OsC1 (XP_015642631.1); gerbera: GhMYB10 (CAD87010); snapdragon: AmROS1 (ABB83826.1), AmROS2 (ABB83827.1), AmVe (ABB83828.1); petunia: PhAN2 (AAF66727.1); grape: VvMYBPA1 (CAJ90831.1), VvMYBPA2 (NP_001267953.1), VvMYBA1 (BAE96751.1), VvMYBA2 (BAD18978.1); apple: MdMYB10 (ABB84753.1) and tomato: SlANT1 (WDP81135.1).

Figure 3.

Phylogenetic tree illustrating the evolutionary trajectory of a TF. The focal TF is circled, and it has at least three potential interaction partners and at least three potential target genes. Most TFs would have more partners and targets (the ellipses in the partner and target vectors indicate that additional partners and targets may exist). The focal TF exists in two different states (see arrows) in the four species. This illustrates a hypothetical scenario where partner 2 represses the ability of the focal TF to activate target 3. We also assume that there are no gene duplications, a single gene loss (the loss of partner 2 in species B and D) and that ancestral populations do not remain polymorphic for the presence or absence of partner 2 for long periods of time. Experimental data from the four taxa would allow one to infer that the partner 2 interaction represses target 3 activation; observing similar data with additional taxa would lead to increased confidence in that hypothesis. Other scenarios that might influence target 3 activation are discussed in the text.

Figure 4.

Cartoon illustrating the pros and cons of relying on TF homology and phylogeny to select targets for plant metabolic engineering.