An entomopathogenic fungus is a one kind of fungus, which can act as a parasite of insects and now extensively used as biocontrol agents against a wide spectrum of insects and other arthropod pests that are harmful to various plants, because of their high efficacy, safe to non-target organisms and ease of multiplication (Sahayaraj and Karthick Raja Namasivayam, 2008). It infects susceptible hosts through the integument, and utilizes different proteases to carry out the digestion of cuticle proteins for colonization and to inactivate the host immune systems (Clarkson and Charnley, 1996), (Frobius et al. 2000).

Insects lack the immune system of vertebrates involving antigen–antibody reactions, although they can protect themselves efficiently from entomopathogenic infections. The humoral part of the insect immune system is characterized by the rapid and transient synthesis of proteins with potent antibacterial and antifungal activity and inactivates the effect of the insect’s pathogenic activity (Gillespie et al. 1997) and expressing a wide spectrum of protease inhibitors to inhibit the proteases and other lytic enzymes produced by entomopathogenic fungi. Such type of protease inhibitors was reported in many insects viz Bombyx mori was effective against entomopathogenic fungi Beauveria bassiana. Larvae of the greater wax moth, Galleria mellonella is very resistant to various microbes and their infections such as Aspergillus melleus, Beauveria bassiana, Metarhizium anisopliae etc (Eguchi et al. 1986). Because it produces various inducible serine protease inhibitors and it was identified recently such as ISPI-1, ISPI-2 and ISPI-3 in hemolymph that are active against various toxic proteases produced from entomopathogenic fungi (Frobius et al. 2000). Based on the determined amino acid sequences, Galleria mellonella ISPI-2 represents a novel member of the Kunitztype inhibitor family, whereas ISPI-1 and ISPI-3 share no similarity with other known proteins (Clermont et al. 2004).

Serine protease inhibitors are mainly classified into three basic types such as canonical, non-canonical and serpins. Canonical inhibitors are the largest family and its size ranging from 14 to 200 amino-acid residues. These are rigid, stable and regularly have beta sheet or mixed alpha-beta topologies. A few can be only alpha helical or irregular proteins that are rich in disulfide bridges (Otlewski et al. 2005). Non-canonical protease inhibitors (for eg. hirudin and haemedin) are interacting with the active site of serine proteases through their N-terminal tails (Grutter et al. 1990). Serpins are 45–55 kDa proteins that are composed of three beta sheets and eight or nine alpha helices forming a single domain and inhibit through the reactive-site loop present at their C-terminus (Gettins, 2002).

The understanding of the three-dimensional structure of a protein would be a precious aid to understand the details of a particular protein. Active site analysis is a key step for identify or design the potential molecules for the purpose of molecular docking studies followed by ligand optimization. The main objective of this study is an attempt to predict the structural as well as active site information of inducible serine protease inhibitor-2 from Galleria mellonella, which is helpful to identify the potential lead molecules for prevent the function of inhibitors and activate the function of fungal serine proteases.

The amino acid sequence of inducible serine protease inhibitor- 2 was retrieved from commonly used primary protein sequence database i.e. Swissprot (http://www.expasy.org) and its accession numbers is P81906 and the source was of Galleria mellonella in origin. The molecular weight of this inhibitor is 6.3kDa obtained from mass spectroscopy (Frobious et al. 2000). The results of Protein BLAST (http://www.ncbi.nlm.nih.gov) search for suitable template structure related to the target sequence (ISPI-2) showed the crystal structure of kunitz domain 1 of tissue factor pathway inhibitor -2 of Bos taurus with highest sequence similarity (53%), as the most suitable template for modeling. The alignment of inducible serine protease inhibitor- 2 and its corresponding template was carefully examined and conserved regions was identified shown in Figure 1 using align-2D module in MODELLER9V2 and it was concluded that this alignment can be assisted to generate a three dimensional reliable model. Jalview (Clamp et al. 2004) was used to display the conserved regions in graphical representation shown in Figure 2.

Once target-template alignment was completed, a three dimensional structure of ISPI-2, was predicted using the program MODELLER9V2 produced three different conformations and its modeller objective functions were 262.7881, 260.3169, 275.1234. The second conformation had lowest value compared to other. Generally, the best model could be obtained from choosing the model with the lowest value of the MODELLER objective function, which is reported in the second line of the model PDB file (Eswar et al. 2008). Based on the lowest value of the modeller objective function, one predicted model was taken and processed in to Swiss-PDB Viewer for energy minimization.

An assessment of the refined model involved two independent tests. The first test was to compare the residue backbone conformations in our refined model with the preferred values obtained from Protein Data Bank of known structures. The results of RAMPAGE web server (http://mordred.bioc.cam.ac.uk/~rapper/rampage.php) explored that 98% residues are found to be most favoured region of the Ramachandran Plot of refined model of induced serine protease inhibitor-2 which is more then cut-off of 90% in most reliable models (Lovell et al. 2003) shown in Figure 3. The stereo chemical quality of the predicted model was found to be satisfactory and low percentage of residues having phi/psi angles in the outlier region.

The second test was carried out using Errat web server (http://nihserver.mbi.ucla.edu/ERRATv2/) for check the quality of the models. Generally, the quality factor of high resolution structures generally produce values around 90% or higher (Colovos and Yeates, 1993). Here, the overall quality factor of this refined model was 92.568 predicted from Errat. The evaluated final reliable model has been deposited in to Protein Model Database (http://mi.caspur.it/PMDB/) and is now publicly accessible [PMDB id: PM0075527]. A three dimensional structure of inducible serine protease inhibitor-2 is shown in Figure. 4.

CastP web server (http://sts-fw.bioengr.uic.edu/castp/calculation.php) was used to predict the active site of inducible serine protease inhibitor – 2 from Galleria mellonella. Six potential binding pockets were identified and displayed in Figure 5. The active site residues (Figure 5) obtained from CastP web server is essential for inhibition of toxic serine proteases produced from various entomopathogenic fungi (Frobius et al. 2000) and they are cysteine, leucine, glutamic acid, histidine, arginine, phenylalanine, glycine, tyrosine, aspartic acid, threonine, aspargine and lysine.

The understanding of the three-dimensional structure and active site of ISPI-2 is important step for identify the probable ligand candidates. However, a three dimensional structure of ISPI-2 is only a predictive, and needed to be confirmed experimentally. Further molecular docking and virtual screening approach is necessary to recognize specific ligands for ISPI-2, which will be validated by many in silico analysis such as molecular docking, QSAR (Quantitative Structure-Activity Relationship) and rule of five.

Future studies helpful to design ligand against ISPI-2 – a major immune defense factor of G. mellonella against M. anisopliae and it completely neutralizes the inhibitory activity of ISPI-2.