Plant Actin-Depolymerizing Factors Possess Opposing Biochemical Properties Arising from Key Amino Acid Changes throughout Evolution

Abstract

Functional divergence in paralogs is an important genetic source of evolutionary innovation. Actin-depolymerizing factors (ADFs) are among the most important actin binding proteins and are involved in generating and remodeling actin cytoskeletal architecture via their conserved F-actin severing or depolymerizing activity. In plants, ADFs coevolved with actin, but their biochemical properties are diverse. Unfortunately, the biochemical function of most plant ADFs and the potential mechanisms of their functional divergence remain unclear. Here, in vitro biochemical analyses demonstrated that all 11 ADF genes in Arabidopsis thaliana exhibit opposing biochemical properties. Subclass III ADFs evolved F-actin bundling (B-type) function from conserved F-actin depolymerizing (D-type) function, and subclass I ADFs have enhanced D-type function. By tracking historical mutation sites on ancestral proteins, several fundamental amino acid residues affecting the biochemical functions of these proteins were identified in Arabidopsis and various plants, suggesting that the biochemical divergence of ADFs has been conserved during the evolution of angiosperm plants. Importantly, N-terminal extensions on subclass III ADFs that arose from intron-sliding events are indispensable for the alteration of D-type to B-type function. We conclude that the evolution of these N-terminal extensions and several conserved mutations produced the diverse biochemical functions of plant ADFs from a putative ancestor.

Figure 1.

Relative F-Actin-Severing/Depolymerization and -Bundling Activities of 11 AtADFs at pH 6.6 and 7.4. The F-actin-severing/depolymerizing and -bundling activities of AtADFs were determined using high-/low-speed cosedimentation assays at pH 6.6 and pH 7.4. The severing/depolymerization activity of 6 μM ADF3 at pH 7.4 was used to normalize the activity of all other AtADFs at pH 6.6 and 7.4; the resulting activities ranged from 0 to 100%. The bundling activity of AtADF5 at pH 6.6 was used to normalize the activity of all other AtADFs at pH 6.6 and 7.4; the resulting activities ranged from 0 to 100%. Functional type D indicates a protein having only severing/depolymerizing activity, whereas functional type B indicates a protein having only bundling activity. The data are presented as the mean ± sd from at least three replicate assays.

Figure 2.

Phylogenetic Analysis of ADF Gene Subclasses in Plants under the Bayesian Information Criterion. (A) Phylogenetic analysis of ADF gene subclasses in plants under the Bayesian information criterion. The four subclasses of ADFs are highlighted in differently colored boxes. Species abbreviations are as follows: At, Arabidopsis thaliana; Atr/AMTR, Amborella trichopoda; Pt/POPTR, Populus trichocarpa; Zm, Zea mays; Os, Oryza sativa; Vv, Vitis vinifera; Cs, Cucumis sativus; Sm, Selaginella moellendorffii; and Pp, Physcomitrella patens. PpADF was used as an outgroup, and the numbers on the branch nodes of the phylogenetic tree refer to posterior probabilities. (B) Schematic diagram of ADF gene structures. In Arabidopsis, all ADF variants contain three exons. The first intron is marked by a light blue line, and the second intron is marked by a black line. The first exon, which only contains the ATG codon, is marked by a black box, the first exon displaying an N-terminal extension due to intron sliding is marked by a red box, and the second and third exons are marked by blue boxes. Lines and boxes are proportional in length and represent the full-length genomic sequences of ADFs. The numbers in the bracket represent the length of the first exon in the ADF gene; the red asterisks show that the conserved exon pattern of ADFs are lacking. The scale bar represents 500 nucleotides.

Figure 3.

Relative F-Actin-Severing/Depolymerization and -Bundling Activities of 10 Putative ancADFs Corresponding to AtADFs at pH 6.6 and 7.4. (A) Reconstruction of the most recent common ancestral protein of each subclass of ADFs in Arabidopsis. Bootstrap percentages >50 are shown next to the corresponding nodes. Nodes A to J refer to the recent common putative ancestral proteins. (B) The actin-severing/depolymerization and -bundling activities of ancADFs at pH 6.6 and pH 7.4 were determined using high-/low-speed cosedimentation assays. The severing/depolymerization activity of 6 μM ancADF-E at pH 7.4 was used to normalize the activity of all other ancADFs at pH 6.6 and 7.4; the resulting normalized activities ranged from 0 to 100%. The bundling activity of ancADF-J at pH 6.6 was used to normalize the activity of all other ancADFs at pH 6.6 and 7.4; the normalized activities ranged from 0 to 100%. PpADF and SmADFs were used as a positive control. Functional type B indicates a protein having only bundling activity, whereas functional type D indicates a protein having only severing/depolymerizing activity. Data are presented as the mean ± sd from at least three replicate assays.

Figure 4.

The Putative Fate of ancADF Activity after Functional Mutational Shifts. Relationships among Arabidopsis ADFs and putative ancADFs, simplified from the illustration shown in Figure 3A. (A) and (B) F-actin-severing/depolymerization or -bundling activities of forward mutants of ancADFs or reverse mutants of current ADFs were determined. (A) The ancADF at node I had no actin-bundling activity but subsequently evolved into ADF9, which possesses F-actin-bundling activity at pH 7.4. This change was concomitant with the replacement of Asn at position 10 and Thr at position 23 by Leu, Lys, and Thr at positions 2 to 4 and Lys at position 18, respectively. ancADF-I2 involved the replacement of Ala and Val at positions 16 and 17 by Trp and Met, respectively; ancADF-I3 involved the replacement of Asn at position 10 by Leu; ancADF-I4 and I5 arose from the replacement of Asn and Thr at positions 10 and 23 by Lys, respectively; ancADF-I1 m involved the replacement of Asn by Leu, Lys, and Thr at position 10 and Thr at position 23 by Lys, respectively; ADF9 m1 involved the replacement of Lys by Ala at position 4; ADF9 m2 involved the replacement of Lys by Ala at position 18; and ADF9 m3 involved the replacement of Lys by Ala at positions 4 and 18. (B) ancADF at node B appears to have low F-actin-severing/depolymerization activity but subsequently evolved to possess increased activity, as shown for node E. This change was partly concomitant with the replacement of Ser at position 18 by His. ancADF-B m involved the replacement of Ser at position 18 by His, ancADF-E m arose from the replacement of His at position 18 by Asn, and ancADF-1-4 m arose from the replacement of His at position 11 by Asn. Data are presented as the mean ± sd from at least three replicate assays.

Figure 5.

Mapping of Hot Spot Residues Responsible for the Functional Divergence of ADF Proteins onto the Crystal Structure of ADF1 (PDB Code 1F7S). (A) The two previously proposed F-actin binding surfaces, respectively labeled as surfaces F1 and F2, are circled, and the side chains of the key residues involved in F-actin binding are shown in ball-and-stick format. (B) The residues responsible for the enhanced actin-bundling activity of certain ADFs are shown in yellow spheres, and the residue responsible for the enhanced actin-severing activity of certain ADFs is shown in a blue sphere. All structural images were prepared using PyMOL (DeLano).

Figure 6.

Alignment of Key Basic Residues and Functional Verification of Subclass I and III ADFs from Different Plant Species. (A) Alignments of ADFs between different species were generated using DNAMAN. The asterisks indicate certain key conserved amino acid residues that are crucial sites for ADF function. At, Arabidopsis; Os, O. sativa; Pt, P. trichocarpa; and Zm, Z. mays. PtADF indicates POPTR_0009S13570g. (B) Functional verification of predicted key basic residues of subclass I and III ADFs from rice and P. trichocarpa based on high-/low-speed cosedimentation assays. Values plotted are means, and the error bars represent sd, n = 3. At least three independent experiments were performed.