Phenyl acetaldehyde (PHA), 2-phenylethanol (2-PE), and its acetate ester are important scent compounds in numerous flowers, including petunias and roses. They also contribute to the aromas of tomato, grape, and tamarind, fruits and to the flavor of tea. PHA has been identified in some animals and fungi as well. Under aerobic conditions; phenylacetaldehyde synthase (PAAS) catalyze oxidative decarboxylation by a radical mechanism. PAAS is the first PLP enzyme to be described that, in its native state, catalyzes the stoichiometric oxidative decarboxylation of an L-amino acid substrate (Figure 1) [1-13]. PAAS by alignment is homologous with Aromatic L-amino acid decarboxylase (AADC) catalyzes the second enzymatic step in synthesis of the neurotransmitters dopamine and serotonin, which are found in neurons of all animals [4]. The Arabidopsis thaliana genome includes two genes, At2g20340 and At4g28680, encoding pyridoxal 5′-phosphate-dependent AADCs with high homology to PhPAAS [8].

Figure 1: 2-PE metabolic pathways. On the top, the plant metabolic pathway with two enzymatic steps is presented: the first step is the conversion of L-phe into phenylacetaldehyde by PAAS with PLP as the cofactor. The second step is the conversion of phenylacetaldehyde to 2-PE by adehydrogenase. On the bottom, the Ehrlich pathway abundant in yeast is shown including three well-described enzymatic steps. The scheme is based on the suggestions of Achmon et al. [1].

Since there is not much experimental data available about different structural properties of PAAS protein, in the present investigation, we studied different structural properties of PAAS protein by bioinformatics tools in Petunia and Rosa. The features of first, secondary and tertiary structure of this protein were compared between Petunia and Rosa.

Amino acid sequence of the PAAS protein

The nucleotide sequence of PAAS was determined using Gen Bank (GenBank no:ABB72475.1 and ABB04522.1; http://www.ncbi.nih.gov/gen bank/). According to Gen Bank, the PAAS protein is composed of 508 and 506 amino acid residues in RhPAAS and PhPAAS respectively (Table 1).

Table 1: Sequence information of PAAS in Rosa hybrid (RhPAAS) and Petunia hybrid (PhPAAS).

Analysis of the secondary structure of the PAAS protein

The protein sequence of the PAAS protein was input, and four conformational states, including helices, sheets, turns and coils, were analyzed. Protein features (attributes) such as the counts and frequencies of each element (carbon, nitrogen, sulfur, oxygen, and hydrogen) and each amino acid, weight, isoelectric point, aliphatic index, N-terminal amino acids, half-life and some more features for double strand elements were extracted using various bioinformatics tools and software including CLC Bio software (CLC bio, Finlandsgade 10-12, Katrinebjerg 8200 Aarhus N Denmark). Prosite features describe domains, families and functional sites as well as associated patterns and profiles of each protein. ScanProsite site (http://www.expasy.org/tools/scanProsite/) was used as a web-based tool to compute Prosite signature matches in protein sequences of favorite proteins [14-17].

Prediction of the 3D structure of the PAAS protein

Predictive analysis of the PAAS protein tertiary structure was conducted using 3D Ligand Site, the online ligand-binding site prediction server (http://www.sbg.bio.ic.ac.uk). This web server automates the manual processes used for the prediction of ligandbinding sites in the eighth round of the critical assessment of protein structure prediction (CASP8), and is a useful tool for the analysis of protein tertiary structure [18].

Prediction of the protein–protein interactions (PPI)

The understanding protein–protein interaction (PPI) in sequence of PAAS was determined by using of STRING (http://STRING.embl.de) [17].

Protein features (attributes) such as the counts and frequencies of each element (carbon, nitrogen, sulfur, oxygen, and hydrogen) and each amino acid, weight, isoelectric point, aliphatic index, N-terminal amino acids, half-life and some more features for double strand elements of PAAS were extracted using various bioinformatics tool (Table S1-S6). The results indicated that, the major aromatic amino acid for PAAS (Phe) in Petunia is more than Rosa. The frequency of negatively charged, Leucin, and frequency of Ser-Leu dipeptid, Pro- Glu dipeptid, Phe-Ser dipeptid, Thr-Thr dipeptid in Petunia are more than Rosa. PAAS as a protein be a member of group II PLP-dependent amino acid decarboxylases with histidine, glutamate, serine, and aromatic L-amino-acid decarboxylases [11]. The content of Arg, Pro, His, Try were the most important features used to prediction of protein characteristics. Also the frequency of Asn-Gln, Gly-Gly and Asp-Pro were the most important features used to build the rest of tree. This tree was the best model to demonstrate the importance of dipeptides in protein function like thermostability [3].

In contrast, in Petunia, the frequency of hydrophobic and hydrophilic residues, α-helix, β-sheet, β-strand is lower than Rosa. Neither PhPAAS nor RhPAAS contains signal peptides at its N-terminal suggesting cytosolic localization [11]. Predictive analysis of the PAAS protein tertiary structure was conducted using 3D Ligand Site, are presented in Figure 2. PAL2 (L-Phe ammonia-lyase 2), PAL4 and COMT (caffeic acid 3-O-methyltransferase) were as protein-protein interaction in this pathway (Figure S1). Phenylalanine ammonia-lyase (PAL) catalyzes the first step of the phenylpropanoid pathway by the deamination of L-phenylalanine to cinnamic acid, which synthesis the hundreds of most important phenolics group secondary metabolic products in plants comprising phenylpropanoids depravities, lignins, lignans, flavonoids and alkaloids [14]. In Arabidopsis thaliana have been annotated that PAL encoding by four genes including AtPAL1, 2, 3 and 4. In Arabidopsis thaliana have been shown that recombinant native AtPAL1, 2 and 4 were exhibited to be catalytically competent for L-phenylalanine deamination, whereas AtPAL3, achieved as a N-terminal His-tagged protein, was of very low activity and only measurable at high substrate concentrations. All four PALs demonstrated similar pH optima, but not temperature optima; AtPAL3 had a lower temperature optimum than the other three isoforms [2]. In plants, COMT an enzyme is belonging to lignin biosynthesis and converting caffeic acid to ferulic acid and 5-hydroxyferulic acid to sinapic acid. In Transgenic alfalfa plants have been shown that COMT cDNA sequences under control of the bean PAL2 promoter. This study demonstrated that strong downregulation of COMT resulted in decreased lignin content [7].

Figure 2: Predictive analysis of the PAAS protein tertiary structure by using 3D Ligand Site, the online prediction server (http://www.sbg.bio.ic.ac.uk).

The transmembrane (TM) secondary structures of membrane proteins were predicted by using the method of preference functions (http://split.pmfst.hr/split/4). The predicted TM helix position of PhPAAS was 108-131 amino acids, however this trends was 251- 265 in RhPAAS. Transmembrane domain usually denotes a single transmembrane α-helix of a transmembrane protein. Transmembrane domain is any three-dimensional protein structure which is thermodynamically stable in a membrane. The amino acid sequence or primary structure of a protein is the most important indication for its function. However, it is approved that prediction of protein characteristics from the primary amino acid sequence is not possible directly. Therefore, methods to predict protein characteristics have converged on tertiary and quaternary structures [3].

Understanding protein–protein interactions (PPI) is important for the analysis of intracellular signaling and metabolic pathways, and modeling of protein complex structures. When two or more than two proteins bind to each other protein-protein interaction occurs that frequently perform biological functions. Thus, knowledge of PPI can be significant insights into almost all biological processes within a cell. In addition to high-throughput experimental methods, computational approaches have been developed to predict functional associations between two proteins by extracting information from the genomic context and evolutionary relationship. These computational methods include gene cluster, gene neighborhood, gene fusion, phylogenetic profile, in silico two-hybrid, and mirror tree methods [15].

Recently the paas synthetic gene was cloned into the pETDuet-1 vector that contains two multiple cloning sites. However, protein over expression or biological activity was not observed for PAAS. Inspection of the PAAS amino acid sequence revealed the presence of 13 cysteine residues, signifying the probability of disulfide bonds in the PAAS tertiary structure [1]. In unfavorable situations for disulfide bonds, configuration of the tertiary structure can cause damage and misfolding of the protein in E. coli. Such problems were explained for bovine β-lactoglobulin having five cysteine residues [12] and also for Nogo-A-specific exon 3 that has eight cysteine residues [6]. Export from the plasma membrane into the periclinal cell wall engages transfer of the comparatively unpolar scent molecules from a lipophilic into an aqueous section. The unpolar character of scent constituents can be explained using their octanol-water partition coefficients, a parameter evaluating their comparative solubilities in lipophilic and aqueous situations. This demonstrates the very low solubility of scent molecules in any aqueous location, which will significantly hinder their transfer across the cell wall. Noticeably both physicochemical processes, disconnection from the plasma membrane, and transport across the cell wall have to be assisted energetically, possibly by direct or indirect action of proteins [10].

Although a novel system showed for the first time the possibility of using the IMPACT system intein tag as a PAAS protein solubility supporter [1], intensifying information about the structure of the PAAS will be needed for the applying protein engineering.

In this study, various supervised and unsupervised tools were applied to identify the most important features contribute to the prediction of PAAS protein features. PAASs have been known only in scent-producing P. hybrida and Rosa hybrida [11]. PAASs are a member of group II of amino acid decarboxylases, and their general amino acid character makes it complicated to predict the actual function of the protein based on amino acid sequences. A major dispute in the postgenomic epoch is to find out protein function at the proteomic level. Most organisms contain a large number of proteins whose functions are presently unidentified. Normally, computational techniques to refer protein function have relied principally on sequence homology. However, the recent appearance of high-throughput techniques for determining protein interactions has facilitated a new line of study where protein function is predicted by utilizing interaction data [16]. Additionally to interaction networks that have been determined experimentally, there are a number of computational methods for building functional interaction networks, where two proteins are linked if they are predicted to complete a common natural mission. The molecular function of a protein demonstrated its biochemical activity, while its biological process specifies the role it plays in the cell or the pathway in which it contributed.