|Home | About | Journals | Submit | Contact Us | Français|
Homology-based three-dimensional model for Pisum sativum sieve element occlusion 1 (Ps.SEO1) (forisomes) protein was constructed. A stretch of amino acids (residues 320 to 456) which is well conserved in all known members of forisomes proteins was used to model the 3D structure of Ps.SEO1. The structural prediction was done using Protein Homology/analogY Recognition Engine (PHYRE) web server. Based on studies of local sequence alignment, the thioredoxin-fold containing protein [Structural Classification of Proteins (SCOP) code d1o73a_], a member of the glutathione peroxidase family was selected as a template for modeling the spatial structure of Ps.SEO1. Selection was based on comparison of primary sequence, higher match quality and alignment accuracy. Motif 1 (EVF) is conserved in Ps.SEO1, Vicia faba (Vf.For1) and Medicago truncatula (MT.SEO3); motif 2 (KKED) is well conserved across all forisomes proteins and motif 3 (IGYIGNP) is conserved in Ps.SEO1 and Vf.For1.
The spindle-like protein bodies which are exclusively present in sieve tubes of legumes were originally described as crystalline P-proteins.1–3 These protein bodies were also found in the form of globular conformations and thus were supposed to undergo structural transformations as well.4–6 The dispersion7 and reversible dispersion/contraction of these bodies8,9 was visualized that gave rise to the novel term “forisomes” (gate bodies) coined by Knoblauch et al.9 In order to stop leakage of photoassimilates or invasion of phytopathogens, the forisomes disperse and occlude sieve tubes as a response to mechanical injury. As soon as the damage is repaired, forisomes can recontract.8,9 The elevation or decrease of free calcium (Ca2+) in sieve tubes results in dispersion/contraction of forisomes.8 Under in vitro conditions, an increase in Ca2+ concentrations10,11 or pH-alterations9,11 can cause anisotropic contractility in forisomes.
The molecular structure of forisomes proteins may be hard to crack. Within forisomettes (the independent subunits of forisomes),3 proteins may be spatially highly ordered and intimately associated to form larger complexes. A 3D model of a protein (target) can be generated based on its amino acid sequence using comparative modeling approach. At least one experimentally solved 3D structure (template) is required for successful modeling of the target protein that has significant amino acid sequence similarity to the target sequence. The procedure for homology modeling involves 4 steps—template selection, target template alignment, model building and evaluation. Based on the 3D protein structures, valuable insights can be gained into molecular basis of protein structure and experiments such as site directed mutagenesis and structure-based specific inhibitors can be designed. The information based on amino acid content of a protein might predict the possible secondary structure of a polypeptide. Here we describe the secondary structure for a part of Ps.SEO1 protein sequence and model its 3D structure. A 3D model of Ps.SEO1 might provide detailed information on the spatial localization of conserved domains and their possible interactions.
The Basic Local Alignment Search Tool (BLAST) search identified motifs on all known members of forisomes proteins with similarity to the TyrX-NRX domain of the tryparedoxin (TryX) and nucleoredoxin (NRX) subfamily of disulphide oxidoreductase (Table 1). Forisomes proteins (For1) from C. gladiata, M. truncatula and Vicia faba were aligned with the sieve element occlusion (SEO) proteins (SEO1, 2 and 3) of M. truncatula and Pisum sativum (SEO1). A well conserved stretch of amino acids (residues 320 to 460) was aligned for all forisomes proteins (Fig. 1A). This particular stretch of amino acids contained TryX-NRX domain which is conserved in all members of the forisomes proteins. We also obtained several non-specific hits that relate remotely to the TryX-like family (like TlpA-like family, TryX-like RdCVF, redoxin, thioredoxin (TRX) family, AhpC-TSA, TrxA and PRK03147) classified under the category of thioredoxin-like superfamily (Fig. 1B). The TRX-like domain is widely distributed among major taxonomic kingdoms and organisms within each kingdom (Fig. 2).
Protein alignment studies. Homology modeling of all known members of the forisomes proteins showed d1o73a_ [Structural Classification of Proteins (SCOP) code, Protein: Tryparedoxin I from Trypanosoma brucei (TaxId: 5702)] as a template hit (Table 1). The percent identity of forisomes proteins with d1o73a_ protein sequence (for the length of 144 amino acids) varied from 10 to 17% (Table 1). The estimated precision on which the model was constructed is in the range of 95 to 100% (Table 1). The fold/protein data bank (PDB) definition, the type of superfamily, and family level classification were thioredoxin fold, thioredoxin-like, glutathione peroxidase-like respectively, which always remained identical irrespective of the forisomes proteins (Table 1). Amino acid alignment of the modeled stretch of forisomes (Cg.For1, Mt.For1 and Vf.For1) and SEO (Mt.SEO1, Mt.SEO2, Mt.SEO3 and Ps.SEO1) proteins with the thioredoxin fold protein (SCOP code: d1o73a_) which was selected as a template hit was conducted (Fig. 3). As expected, the alignment studies showed conserved amino acids (indicated by asterisks), and conservative amino acid changes (indicated by dots) along the modeled amino acid stretch for all known forisomes proteins (Fig. 3).
Secondary structure prediction and homology modeling. The secondary structure content prediction for amino acid residues 320 to 456 of Ps.SEO1 revealed the presence of α, β and coiled contents (Fig. 4). We have successfully modeled this particular stretch of amino acids for its secondary structure content which included seven α helices, seven β strands and 10 coiled regions (Fig. 4).
Based on alignment studies (Figs. 1A and and33), a conserved stretch of amino acids from P. sativum SEO1 protein (residues 320 to 456) was used to model its 3D structure. This model based on sequence homology is produced by finding a sequence alignment to a known structure, copying the coordinates and labeling the residues accordingly. The entire Ps.SEO1 protein alignment was performed against the fold library however only the sequence section (amino acid 320 to 456) showing significant homology to the domain of TRX fold, has been modeled (Fig. 5). Attempts were also made to model the remaining sequence portions by deletion of the amino acids 320 to 456, and separate analysis was conducted. However, confident matches were not found with the known deposited structures and hence model prediction for the remaining amino acid stretches was not possible. The analysis revealed the protein with SCOP code d1o8xa_ as the best template hit. But when structural modeling was performed we were not able to superimpose the structures of Ps.For1 and d1o8xa_ (data not shown). Therefore, from the same analysis, another template with SCOP code d1o73a_ was selected.
Molecular 3D protein imaging is represented as ribbon diagram for the template d1o73a_ as shown in Figure 5A and the predicted structure of amino acids 320 to 456 of Ps.SEO1 is shown in Figure 5B. The superimposition shows that the structures partially overlap (Fig. 5C). Motif 1 (EVF), which forms the α helix is conserved in Ps.SEO1, Vf.For1 and Mt.SEO3, motif 2 (KKED) forms a coil and is well conserved across all forisomes proteins and motif 3, which also makes a coil (IGYIGNP) is conserved in Ps.SEO1 and Vf.For1.
The forisomettes3 may be structurally and physiologically independent units, but must act in a concerted fashion when required. The complex 3D structure of Ps.SEO1 probably confers a multitude of electro- and hydrostatic interactions within the protein and its hydration shell. The Ca2+ induced changes in surface charge and electrical density may allow massive and sudden binding of water molecules and induce conformation shifts. The structural coherence within forisomes is also mysterious. The passage of a weak EPW elicits dispersion of forisome ends only.12
The template (SCOP code d1o73a_) on which Ps.SEO1 was modeled is TryX protein, a member of the TRX-fold family from Trypanosoma brucei.13 The functional significance of this TryX-NRX subfamily of disulphide oxidoreductases is mostly unclear, particularly in plants.14 The conserved domain search located protein domains reminiscent of thioredoxin fold in SEO2 (amino acids 307 to 450) and SEO3 (amino acids 323 to 459) of Medicago truncatula, and For1 (353 to 440) of Cannavalia gladiata.3 TryX and NRX are thioredoxin-like protein disulfide oxidoreductases that alter the redox state of target proteins via the reversible oxidation of an active center CXXC motif. Vertebrate NRX is a 400-amino acid nuclear protein with one redox active TRX domain containing a CPPC active site motif followed by one redox inactive TRX-like domain. Plant NRX is longer than the vertebrate NRX by about 100–200 amino acids, is a nuclear protein containing a redox inactive TRX-like domain between two redox active TRX domains. Both vertebrate and plant NRXs show thiol oxidoreductase activity in vitro. Their localization in the nucleus suggests a role in the redox regulation of nuclear proteins such as transcription factors. The TRX-like domain is widely distributed among major taxonomic kingdoms and organisms within each kingdom.
Our earlier studies have predicted the secondary structural content from the entire amino acid composition for For1 and SEO1, 2 and 3 proteins based on analytical vector decomposition method using comparative sequence analysis—Secondary Structural Content Prediction (SSCP) tool (coot.embl.de/SSCP).3 The prediction of secondary structural content derived from amino acid composition of FOR1 and SEO1, 2 and 3 proteins indicated that α, β and coiled-coil configurations are present at approximately equal levels.3 Noll (2005)15 hypothesized that coiled-coil structures presumably present in Mt.SEO2 might play a role in forisome function; whereas Fontanellaz (2006)16 discussed the potential of Vf.SEO1 to form coiled-coil structures (Noll 2005).15
The amino acid sequences of forisomes proteins (For1) from Cannavalia gladiata (Cg), Medicago truncatula (Mt), Vicia faba (Vf), and the sieve element occlusion (SEO) proteins (Mt.SEO1, 2 and 3) of Medicago truncatula (Mt), Pisum sativum (Ps.SEO1) were downloaded from NCBI web server (www.ncbi.nlm.nih.gov). Conserved portion of amino acids (residues 320 to 460) was aligned with ClustalX software.17 The studies on TryX-NRX domain of the tryparedoxin (TryX) and nucleoredoxin (NRX) subfamily domains was performed using NCBI server (http://www.ncbi.nlm.nih.gov). For the construction of Figure 1B, the Ps.SEO1 was used as query sequence. The distribution of TryX-like domains across the major taxonomic kingdoms was studied using the Superfamily 1.73 (supfam.mrc-lmb.cam.ac.uk/SUPERFAMILY). The content of secondary structure of Ps.SEO1 protein was studied for the modeled stretch (amino acids 320 to 456) using the PHYRE web server (www.sbg.bio.ic.ac.uk/phyre).18 The studies on molecular graphics for 3D modeling of Ps.SEO1 and the template (d1o73a_) were conducted using the UCSF Chimera package from the resource for bio-computing, visualization and informatics at the university of California, San Francisco (supported by NIH P41 RR-01081).19
Work on calcium signaling and plant stress tolerance in N.T.'s laboratory is supported partially by the Department of Science and Technology (DST), Government of India and Department of Biotechnology (DBT), Government of India. We thank Dr. Ajay Kumar Saxena (Jawaharlal Nehru University, New Delhi, India) for preliminary help with homology modeling.
Previously published online: www.landesbioscience.com/journals/psb/article/11422