Multiplace subcellular localization of dehydrin-like proteins in embryonic axes of common beech (Fagus sylvatica L.) seeds during maturation and dry storage

SUPPLEMENTAL INFORMATION FOR ONLINE PUBLICATION ONLY

Ewa Marzena Kalemba1, Agnieszka Bagniewska-Zadworna2, Ewelina Ratajczak1

1) Laboratory of Seed Biochemistry, Institute of Dendrology, Polish Academy of Sciences, Kórnik, Poland

2) Department of General Botany, Institute of Experimental Biology, Faculty of Biology, AdamMickiewiczUniversity, Umultowska 89, Poznań, Poland

Tab. 1Results of DP-Bind ( Hwang and others 2007) Analyses

DP-Bind is a web server for sequence-based prediction of DNA-binding residues in DNA-binding proteins. Dehydrin/response ABA protein from (Fagus sylvatica) with accession number CAE54590.1 was used as a query and PSSM-based encoding parameter. [POS] and [RES] refer to position and residue, respectively, in the analyzed protein sequence. Labels in columns S_LBL, K_LBL, and P_LBL correspond to binding labels predicted by the three prediction methods: Support Vector Machine (SVM), Kernel Logistic Regression (KLR), and Penalized Logistic Regression (PLR), respectively. Binding labels in columns MAJ_CON and STR_CON show the consensus predictions.Columns S_PRB, K_PRB, and P_PRB show the probabilities of the predicted labels. These probabilities are in range [0.5 to 1.0]. The higher the probability, the greater the confidence of the predicted label. Table contains residues which were given “1”labels for DNA-binding activity in all examined predictions. Moreover POS and RES are indicated in color when predicted residues are arranged into a row amino acids sequence.

POS / RES / S_LBL / S_PRB / K_LBL / K_PRB / P_LBL / P_PRB / MAJ_CON / STR_CON
5 / R / 1 / 0.8613 / 1 / 0.9178 / 1 / 0.6653 / 1 / 1
20 / A / 1 / 0.7101 / 1 / 0.6522 / 1 / 0.5760 / 1 / 1
26 / N / 1 / 0.5441 / 1 / 0.6455 / 1 / 0.7365 / 1 / 1
32 / P / 1 / 0.8151 / 1 / 0.8304 / 1 / 0.8089 / 1 / 1
33 / I / 1 / 0.6884 / 1 / 0.6188 / 1 / 0.6642 / 1 / 1
34 / R / 1 / 0.5945 / 1 / 0.5077 / 1 / 0.6737 / 1 / 1
35 / S / 1 / 0.6998 / 1 / 0.7249 / 1 / 0.7443 / 1 / 1
36 / D / 1 / 0.7257 / 1 / 0.5786 / 1 / 0.7595 / 1 / 1
37 / E / 1 / 0.6665 / 1 / 0.6142 / 1 / 0.5989 / 1 / 1
38 / Y / 1 / 0.5840 / 1 / 0.6429 / 1 / 0.5760 / 1 / 1
39 / G / 1 / 0.7716 / 1 / 0.7827 / 1 / 0.7799 / 1 / 1
50 / P / 1 / 0.6454 / 1 / 0.5880 / 1 / 0.6844 / 1 / 1
54 / T / 1 / 0.5360 / 1 / 0.5738 / 1 / 0.7280 / 1 / 1
63 / N / 1 / 0.6225 / 1 / 0.5667 / 1 / 0.7701 / 1 / 1
64 / W / 1 / 0.7003 / 1 / 0.6739 / 1 / 0.7893 / 1 / 1
65 / W / 1 / 0.5755 / 1 / 0.6501 / 1 / 0.6486 / 1 / 1
108 / G / 1 / 0.7207 / 1 / 0.6132 / 1 / 0.5623 / 1 / 1
111 / S / 1 / 0.5785 / 1 / 0.5312 / 1 / 0.5050 / 1 / 1
112 / S / 1 / 0.6476 / 1 / 0.5856 / 1 / 0.6842 / 1 / 1
113 / S / 1 / 0.7497 / 1 / 0.5430 / 1 / 0.5447 / 1 / 1
124 / R / 1 / 0.7550 / 1 / 0.7537 / 1 / 0.5896 / 1 / 1
125 / K / 1 / 0.7330 / 1 / 0.5442 / 1 / 0.6419 / 1 / 1
126 / K / 1 / 0.8158 / 1 / 0.6815 / 1 / 0.5371 / 1 / 1
127 / K / 1 / 0.7656 / 1 / 0.6036 / 1 / 0.5601 / 1 / 1
139 / G / 1 / 0.5195 / 1 / 0.5346 / 1 / 0.5433 / 1 / 1
169 / K / 1 / 0.6536 / 1 / 0.6636 / 1 / 0.5852 / 1 / 1
170 / K / 1 / 0.7342 / 1 / 0.6376 / 1 / 0.5634 / 1 / 1
177 / K / 1 / 0.6973 / 1 / 0.7380 / 1 / 0.7008 / 1 / 1
179 / K / 1 / 0.7765 / 1 / 0.8357 / 1 / 0.5588 / 1 / 1
182 / G / 1 / 0.8241 / 1 / 0.8545 / 1 / 0.6656 / 1 / 1
183 / H / 1 / 0.5619 / 1 / 0.6309 / 1 / 0.5229 / 1 / 1

Fig. 1 Profisis (ISIS), a machine learning-based method that identifies interacting residues from sequence alone using transient protein-protein interfaces from complexes of experimentally known 3D structures, was used to predict 12 potential protein-protein interaction sitespointed with red diamonds and described below as a single residues or tracks of residues in the amino acid sequence

Fig. 2 Results of putative dehydrin-lipid interactions predicted with MCPep server ( that performs Monte Carlo (MC) simulations of helical peptides in association with lipid membranes. Three 50 amino acid peptides of dehydrin (CAE54590.1) were analyzed, two of them contained consensus K segment that potentially form an amphipathic helix and one N’-terminal peptide containing none of the consensus dehydrin segment as a reference probe. The localization of analyzed peptides in dehydrin/response ABA [Fagus sylvatica] sequence (CAE54590.1) is given below. Peptides are indicated with bolded and colored letters put in frames. Overlaping sequence of two analyzed peptides is pointed with black bolded letters and put in double frame.

Simulations of peptides in water (A, C, E) and in membrane (B, D, F) are compared. Comparision of two peptides that were predicted to interact with membranes (G) based on the average distance of each residue from the membrane midplane.

A) Simulation of KEKIPGVGHKDDQYHQGQTTSTITPDYYDQGQTHEKKGIMEKIKEKLPGH peptide in water, along the x-z and y-z planes. Residues' types are designated in circles. The hydrophobic residues (A, G, F, I, L, V) are in orange, charged residues (K, R, E, D) are blue, and the polar residues are green:

B)Simulation of the KEKIPGVGHKDDQYHQGQTTSTITPDYYDQGQTHEKKGIMEKIKEKLPGH peptide in the membrane along the x-z and y-z planes. Residue type is designated in circles. Hydrophobic residues (A, F, G, I, L, V) are orange, charged residues (K, R, E, D) are blue, and polar residues are green. The horizontal dotted line marks the location of the phosphate groups of the lipid polar heads in each frame.

C) Simulation of GLSGMLHRSGSGSSSSSEDDGFGGRRKKKGVMEKVKEKIPGVGHKDDQYH peptide in water, along the x-z and y-z planes. Residues' types are designated in circles. The hydrophobic residues (A, G, F, I, L, V) are in orange, charged residues (K, R, E, D) are blue, and the polar residues are green:

D) Simulation of GLSGMLHRSGSGSSSSSEDDGFGGRRKKKGVMEKVKEKIPGVGHKDDQYH peptide in the membrane along the x-z and y-z planes. Residue type is designated in circles. Hydrophobic residues (A, F, G, I, L, V) are orange, charged residues (K, R, E, D) are blue, and polar residues are green. The horizontal dotted line marks the location of the phosphate groups of the lipid polar heads in each frame.

E) Simulation of the N’-terminal 50 amino acid peptide (MKAVREKPNLFELFLQNQYAGAAPQNDEFGNPIRSDEYGNRVRTHEFGNP) in water, along the x-z and y-z planes. Residues' types are designated in circles. The hydrophobic residues (A, G, F, I, L, V) are in orange, charged residues (K, R, E, D) are blue, and the polar residues are green:

F) Simulation of the N’-terminal 50 amino acid peptide (MKAVREKPNLFELFLQNQYAGAAPQNDEFGNPIRSDEYGNRVRTHEFGNP) in membrane

G) Comparision of two peptides that were predicted to interact with membranes. The orientation graph shows the average distance of each residue from the membrane midplane. Residue type is designated in circles. Hydrophobic residues (A, F, G, I, L, V) are orange, charged residues (K, R, E, D) are blue, and polar residues are green. The horizontal dotted line marks the location of the phosphate groups of the lipid polar heads in closer distatnce for GLSGMLHRSGSGSSSSSEDDGFGGRRKKKGVMEKVKEKIPGVGHKDD-QYH the peptide.