Uses of phage display in agriculture: sequence analysis and comparative modeling of late embryogenesis abundant client proteins suggest protein-nucleic acid binding functionality

Abstract

A group of intrinsically disordered, hydrophilic proteins-Late Embryogenesis Abundant (LEA) proteins-has been linked to survival in plants and animals in periods of stress, putatively through safeguarding enzymatic function and prevention of aggregation in times of dehydration/heat. Yet despite decades of effort, the molecular-level mechanisms defining this protective function remain unknown. A recent effort to understand LEA functionality began with the unique application of phage display, wherein phage display and biopanning over recombinant Seed Maturation Protein homologs from Arabidopsis thaliana and Glycine max were used to retrieve client proteins at two different temperatures, with one intended to represent heat stress. From this previous study, we identified 21 client proteins for which clones were recovered, sometimes repeatedly. Here, we use sequence analysis and homology modeling of the client proteins to ascertain common sequence and structural properties that may contribute to binding affinity with the protective LEA protein. Our methods uncover what appears to be a predilection for protein-nucleic acid interactions among LEA client proteins, which is suggestive of subcellular residence. The results from this initial computational study will guide future efforts to uncover the protein protective mechanisms during heat stress, potentially leading to phage-display-directed evolution of synthetic LEA molecules.

Figure 1

A graphic depiction of the region of the client proteins to which the SMP1 or GmPM28 proteins bound. In each graph, the full-length protein is depicted as a black bar centered at zero on the Hopp/Woods hydrophilicity plot [57] for the protein (retrieved from Expasy Protscale [35]). Recognizable motifs present in the protein are represented as red bars on the black bar under which the Pfam [58] acronym defining the motif superfamily is displayed. The region of the full-length protein displayed on the phage and captured by the LEA is shaded grey. When this region overlaps with a recognizable motif, the protein is assigned to (b) (A motif bound). When it coincides with the most, or among the most, hydrophilic of the proteins regions, it is placed in (a) (or marked by an asterisk in (b)). If the LEA-bound fragment is neither the most hydrophilic nor encoding a recognizable motif, it is placed in (c) (no recognizable attribute bound). In each graph, the size, in amino acids, of the protein moiety bound by the LEA is provided as well as the number of independently acquired clones. If the clones were of different lengths, the number of clones of a specific length is provided. Whether the clone was bound by SMP1 or GmPM28 and the temperature at which the binding occurred are also provided.

Figure 2

Analysis of the inclusive bound regions of the LEA client proteins identified using phage display. (a) Amino acid composition of the LEA-bound client protein regions is given here by % of the entire individual bound region. The regions are identified by the TAIR locus identifier for the full-length protein, though only composition of the bound region is represented in the plot. (b) A comparison of the GRAVY hydropathicity of the bound regions of the LEA client proteins is given here, again identified by the full-length protein TAIR locus identifier.

Figure 3

Amino acid sequences of the bound regions of the LEA client proteins. Recurring patterns within the set of sequences have been identified by red and blue text. The red characters indicate the K-x(2,4)-V-x(4)-[ACDGNSTV] pattern. Blue characters indicate the R-x(1,2)-R-x(0,1)-S pattern.

Figure 4

Chemical structure of a nucleotide dimer, left, and glycerol 3-phosphate (glycerophosphate), right. The red lettering on the nucleotide dimer represents the chemical similarity to the glycerophosphate molecule.

Figure 5

Seven homology models of LEA client proteins, focused on the regions containing the protein moiety to which the LEA proteins bound, were developed. The sequences, numbered by the full-length protein TAIR locus identifier, are shown annotated by secondary structure elements. Secondary structure annotation was accomplished using the ESPript web utility [59]. β-Sheets are labeled with a solid black arrow, α-helices with medium curly script, β-turns with TT, and 3₁₀-helices (η) with small curly script. Sequence number is also indicated in frequency of ten and corresponds to that of the full-length sequence. Below the sequences, the seven homology models of the bound regions only are shown in cartoon representation. The homology models are labeled, as with the sequences, according to the TAIR locus identifier of the full-length protein to which they belong. The homology model PDB files have been included in Supplementary Materials available online at http://dx.doi.org/10.1155/2013/470390.