Molecular Cloning, Expression, Function, Structure and Immunoreactivities of a Sphingomyelinase D from Loxosceles adelaida, a Brazilian Brown Spider from Karstic Areas

Loxosceles is the most poisonous spider in Brazil and, at least, three different species of medical importance are known in Brazil (L. intermedia, L. gaucho, L. laeta), with more than 5000 cases of envenomation reported each year. In South Africa, L. parrami and L. spinulosa are responsible for cutaneous loxoscelism (Newlands et al., 1982). In Australia, a cosmopolitan species, L. rufescens, is capable of causing ulceration in humans. In the USA, at least five Loxosceles species, including L. reclusa (brown recluse), L. apachea, L. arizonica, L. unicolor and L. deserta are known to cause numerous incidents (Ginsburg W Gendron, 1990; Bey et al., 1997; Desai et al., 2000). Several studies have indicated that sphingomyelinase D (SMase D) present in the venoms of Loxosceles spiders is the main component responsible for the local and systemic effects observed in loxoscelism (Forrester et al., 1978; Kurpiewski et al., 1981; Tambourgi et al., 1998, 2000, 2002, 2004, 2005, 2007; van den Berg et al., 2002, 2007; Fernandes Pedrosa et al., 2002; Paixao Cavalcanti et al., 2006, Tambourgi et al., 2010). SMases D hydrolyze sphingomyelin resulting in the formation of ceramide-1-phosphate and choline (Forrester et al., 1978; Kurpiewski et al., 1981; Tambourgi et al., 1998) and, in the presence of Mg2+, are able to catalyze the release of choline from lysophosphatidylcholine (van Meeteren et al., 2004). All spider venom SMases D sequenced to date display a significant level of sequence similarity and thus likely possess the same (┙/┚)8 or TIM barrel fold (Murakami et al., 2005, 2006). Based on sequence identity, biochemical activity and molecular modelling, a scheme for classification of spider venom SMases D was proposed (Murakami et al., 2006). The class 1 enzymes include SMase I and H13, SMases D from L. laeta, which possess a single disulphide bridge and contain an extended hydrophobic loop or variable loop (Murakami et

. All other SMases D, such as SMases P1 and P2 from L. intermedia (Tambourgi et al., 1998, Lr1 and Lr2 from L. reclusa and Lb1, Lb2 and Lb3 from L. boneti (Ramos-Cerrillo et al., 2004) belong to class 2, which contains an additional intra-chain disulphide bridge that links the flexible loop with the catalytic loop (Murakami et al., 2006). The class 2 enzymes can be further subdivided into class 2a and class 2b depending on whether they are capable of hydrolysing sphingomyelin or not, respectively (Murakami et al., 2006). One representative of class 2b is the isoform 3 from L. boneti an inactive SMase D isoform (Ramos-Cerrillo et al., 2004) (Table 1). Studies on the effect of venoms from synanthropic species of Loxosceles spiders have been reported, however, analysis from those living in natural environment have been poorly performed. Loxosceles species are present in several different habitats, including the karstic environment, and in Brazil it is the most common troglophile arachnid. The spiders are commonly found on the walls at the entrance of caves, especially in the shady rocky areas. In order to characterize the venom of Loxosceles species living in natural environment, and compare their venom to those of synanthropic species, we have explored the caves of 'Parque Estadual do Alto do Ribeira' (PETAR -Ribeira Valley, SP, Brazil), which is an area of importance to both tourism and scientific research, due to the combination of tropical forest and extensive cave systems. Spiders captured in the caves of PETAR were identified by morphological analysis as Loxosceles adelaida Gertsch (1967), which belongs to 'gaucho group' (Gertsch, 1967). This group includes four species: L. adelaida, L. gaucho, L. similis and L. variegata. L. gaucho belongs to the synanthropic fauna of Brazilian arachnids and is considered an important cause of the loxoscelic accidents in the south-eastern region of the country. Thus, the aims of the present study were to clone and express SMases D from the spider gland of L. adelaida, captured in the caves of PETAR (Brazil), to compare the functional activities of the recombinant proteins with toxins from synanthropic species and to investigate the inter-and intra-species cross-reactivities of antibodies raised against the purified recombinant proteins.

Spiders and venoms
Loxosceles intermedia, L. laeta, L. gaucho and L. adelaida spiders were provided by Immunochemistry Laboratory, Butantan Institute, Brazil (permission Nº 1/35/95/1561-2 was provided by the Brazilian Institute of Environment and Renewable Natural Resources -IBAMA -a Brazilian Ministry of the Environment's enforcement agency). The venoms were obtained (permission Nº 01/2009/IBAMA) by electrostimulation by the method of Bucherl (1969) with slight modifications. Briefly, 15-20 V electrical stimuli were repeatedly applied to the spider sternum and the venom drops were collected with a micropipette in PBS, aliquoted and stored at -20°C. The protein content of the samples was evaluated using the BCA Protein Assay Kit (Pierce Biotechnology, MA, USA).

Animals
The adult New Zealand white rabbits weighing approximately 3 kg were supplied by Butantan Institute animal facilities, SP, Brazil. All the procedures involving animals were in accordance with the ethical principles in animal research adopted by the Brazilian College of Animal Experimentation.

cDNA generation
The venom glands were removed from 50 specimens of Loxosceles adelaida 5 days after venom extraction, when the maximum level of RNA synthesis was achieved. Then, the glands were quickly frozen in liquid nitrogen and stored at −80°C until used. The total RNA was extracted from the glands using the Trizol reagent, according to the manufacturer's instructions (Invitrogen, USA). The quality and yield of total RNA were verified by the integrity of 28S and 18S rRNA, through denaturing agarose gel electrophoresis, and spectrophotometrically using the ratio 260/280 nm. The cDNAs were synthesised from 5 µg of total RNA using the cDNA Cycle® Kit for RT-PCR (Invitrogen, USA).

PCR Amplification of Sphingomyelinase D homologue from L. adelaida
The SMase D cDNA was amplified by PCR using total reverse transcriptase-PCR products as template. Degenerate primers AdeF1 SMases D P1 and P2 from L. intermedia (accession numbers AY304471 and AY304472, respectively) and Smase D I from L. laeta (accession number AY093599) Fernandes Pedrosa et al, 2002). The PCR protocol included denaturation at 96°C for 3 min, followed by 35 cycles of denaturation (30 s at 95°C), annealing (30 s at 58°C), and extension (2 min at 72°C), and a final extension for 7 min at 72°C. The amplified fragments were purified by low melting using the UltraPure™ Low Melting Point Agarose (Invitrogen, USA) and cloned in a pGEM-T easy vector (Promega, USA). E. coli XL1-Blue competent cells (Strategene, USA) were transformed following the manufacturer's instructions. Positive clones were selected by growing the transformed cells in Luria broth (LB) medium containing 100 g/ml ampicilin and blue-white color screening. The nucleotide sequence was determined by the dideoxy chain-termination method using the BigDye™ Terminator Cycle Sequence Kit and the ABI 3100 automatic system (Applied Biosystems, USA).

Subcloning in expression vector
The cDNA corresponding to the mature L. adelaida SMase D was amplified from plasmids containing full length sequence using primer AdeF3 (5'-CTCGAGGCGGATAAACGTCGACCCATATG-3'), which contains a XhoI restriction site (underlined) and AdeR4 (5'-CCATGGTCAATTCTTGAAAGTCTCC-3'), which includes a restriction site for NcoI (underlined) and a stop codon (in italics and bold). PCR fragments (approximately 900 bp) were digested with XhoI and NcoI and cloned into the corresponding sites in pRSET-B bacterial expression vector (Invitrogen, USA). The use of the pRSET-B bacterial expression system results in the expression of a recombinant fusion protein, including a 6xHis-tag at the N-terminus and a 26 amino acid linker followed by the mature protein (N-terminal amino acid sequence before the coding sequence of the mature protein: 'MRGSHHHHHHGMASMTGGQQMGRDLYDDDDKDPSSR').

Recombinant protein expression and purification
Recombinant proteins were produced as described previously . In brief: pRSETB-L. adelaida SMase D cDNA transformed Escherichia coli BL21 (DE3) (Invitrogen, USA) cells were inoculated in 500 mL of 2YT/amp and grown overnight at 37°C and induced with IPTG. Recombinant proteins were harvested from the pellet by frenchpressure and purified on a Ni (II) Chelating Sepharose Fast Flow column (Pharmacia, Sweden, 1.0×6.4 cm). The fractions containing the recombinant proteins were pooled and concentrated using Centricon-30 (30,000-mw cutoff; Amicon, Inc., USA). Smase D P1 from L. intermedia and SMase D I from L. laeta were obtained as previously described Fernandes-Pedrosa et al., 2002).;The protein content of the samples was evaluated by the BCA protein kit assay, following the manufacturer's protocol (Pierce, USA).

Enzymatic activity
The SMase D enzymatic activity was estimated by determining the choline liberated from lipid substrates, using a fluorimetric assay (Tokumura et al., 2002). Briefly, sphingomyelin (SM -50 g) was diluted in 1 mL HEPES-buffered saline (HBS). Samples of the recombinant proteins or venom (2.5 to 20 g) was added and the reaction was developed for 20 min at 37°C. After incubation, a mixture composed by 1 unit/mL choline oxidase, 0.06 unit/mL of horseradish peroxidase and 50 M of 3-(4-hydroxy-phenyl) propionic acid in HBS was added and incubated for 10 min. The choline liberated was oxidized to betaine and H 2 O 2 and this product determined by fluorimetry at em=405 nm and ex=320 nm, using 96-well microtiter plates, in a spectrofluorimeter (Victor 3 TM , Perkin-Elmer, USA).

Normal human serum and erythrocytes
Human blood was obtained from healthy donors. Blood samples drawn to obtain sera were collected without anticoagulant, allowed to clot for 4 h at 4ºC and the normal human serum (NHS) was stored at -80 o C. Blood samples drawn to obtain erythrocytes for subsequent use as target cells were collected in anticoagulant (Alsever´s old solution: 114 mM citrate, 27 mM glucose, 72 mM NaCl, pH 6.1).

Treatment of erythrocytes with Loxosceles recombinant proteins/venom
Human erythrocytes were washed and resuspended at 1.5% in VBS 2+ and incubated with different concentrations of the recombinant proteins or venom for 1 h at 37 o C. Control samples were incubated with VBS 2+ . The cells were washed, resuspended to the original volume in VBS 2+ and analysed in a haemolysis assay as described (Tambourgi et al., 2000).

Dermonecrotic activity
Samples of L. adelaida venom (5 µg), SMase P1 from L. intermedia (5 µg) or SMase D from L. adelaida (5 µg and 10 µg), in PBS, were injected intradermally in the shaved back of adult rabbits. Control sites were injected with equal volume of PBS. The size of the lesions was measured over a period of 16, 24, 48 and 72 h. After 72 h, the animals were euthanized and skin specimens were obtained for histological examination.

Histological analysis
Skin samples were fixed in 4% buffered formalin solution, and then embedded in paraffin. Tissue sections were stained with haematoxylin and eosin and examined for the presence of epithelial necrosis, epithelial slough, dermal infiltrates, haemorrhage and level of collagen dissociation in the dermis and skin muscle fiber degeneration.

Homology molecular modelling and quality analysis
The atomic coordinates of SMase I, a SMase D from L. laeta (PDB accession code: 1XX1) was used as 3D-model for restraint-based modelling as implemented in the program MODELLER (Fiser and Sali, 2003). The overall model was improved enforcing the proper stereochemistry using spatial restraints and CHARMM energy terms, followed by conjugate gradient simulation based on the variable target function method (Fiser and Sali, 2003). Ten models were built for the L. adelaida SMase D sequence based on the (m)GenThreader alignment (Lobley et al., 2009) and the model with best global score was selected for explicit solvent MD simulations using the package Yasara (http://www.yasara.org) to check its stability and consistency. The overall and local quality analyses of the final model were assessed by VERIFY3D (Eisenberg et al., 1997), PROSA (Wiederstein and Sippl, 2007) and VADAR (Willard et al., 2003). Three-dimensional structures were displayed, analyzed and compared using the program COOT (Emsley and Cowtan, 2004).

Molecular dynamics simulation of the enzyme-substrate complex
The SMase D I structure (PDBid 1XX1) obtained from the Protein Data Base (www.rcsb.org) and a predicted structure of the sphingomyelin (SM) was manually docked using the sulfate ion from the experimental structure, as a reference of the position of the phosphate moiety of the SM. The structure was prepared for energy minimization (EM) molecular dynamics (MD) simulation using YASARA program for building missing atoms and hydrogens in the model. The parameters for the force field were obtained from YAMBER3 (Krieger et al., 2004). The pKa values for Asp, Glu, His and Lys residues were predicted. Based on the pH 7.0, the protonation states were assigned according to convention: Asp and Glu were protonated if the predicted pKa was higher than the pH; His was protonated if the predicted pKa was higher than the pH and it did not accept a hydrogen bond, otherwise it was deprotonated; Cys was protonated; Lys was deprotonated if the predicted pKa was lower than the pH and; Tyr and Arg were not modified (www.yasara.org). A simulation box was defined at 15 Å around all atoms of each macromolecular complex. Then, the simulation box was filled with water molecules and Na/Cl counter ions, that were placed at the locations of the lowest/highest electrostatic potential, until the cell neutralization, and the requested NaCl concentration reached 0.9%. A short MD simulation was performed for the solvent adjust, and water molecules were subsequently deleted until the water density reached 0.997 g/mL. A short steepest descent EM was carried until the maximum atom speed dropped below 2200 m/s. Then 500 steps of simulated annealing EM were performed with a target temperature at 0 K. Finally, a 20 ns MD production simulation was performed at 298 K and a non-bonded cutoff of 7.86 A. A snapshot was saved every 7.5 ps (picoseconds). The graphical analysis was carried out using Visual Molecular Dynamics (VMD) software (Humphrey et al., 1996) and the plots were generated using R (cran.r-project.org). Ligand binding analysis was carried out taking into account the potential energy obtained with the current force field for the complex and components: E binding = (E protein + E ligant ) -E complex In our equation, the complex energy is placed after the energy of the components; therefore, higher energies equated to higher affinity between the protein and the substrate.

Identification and characterization of L. adelaida recombinant protein
Analysis of L. adelaida SMase D clone revealed that the mature cDNA covers an open reading frame of 843 nucleotides encoding 280 amino acid residues (GenBank JN202927). The complete nucleotide sequence of this cDNA clone and the deduced amino acid sequence are shown in Figure 1A. Sequence analysis revealed that SMase D from L. adelaida exhibits similarity with previously characterized class 2 SMase D and shares 75% and 59% homology with the sequences of L. intermedia SMase P1 and L. laeta SMase I, respectively ( Figure 1B). The theoretical molecular mass for the putative recombinant mature protein of L. adelaida is 31.545 kDa, with a pI = 8.85. Expression of L. adelaida SMase D resulted in an approximately 36 kDa recombinant His6-tagged protein, clearly visible by SDS-PAGE in the bacterial cell extracts (Figure 2A). The recombinant protein was purified from the soluble fraction of cell lysates by Ni +2 -chelating chromatography and eluted from the resin in extraction buffer containing 0.8 M imidazole at greater than 95% purity ( Figure 2B). The supernatant was loaded onto Ni(II)-Chelating Sepharose column, the flowthrough was collected (FT), the column was washed (W1 -W4), and the recombinant protein was eluted with the elution buffer. Samples were separated by SDS-PAGE (12% gel) under reducing conditions and stained by silver.

Enzymatic activity
The ability of the L. adelaida recombinant SMase D protein to degrade sphingomyelin was investigated and compared with the activity of the previously characterized active recombinant isoform from L. intermedia gland, SMase D P1, and with L. adelaida crude venom. Figure 3 shows that the crude L. adelaida venom, as well as the recombinant L. adelaida SMase D A1 and Smase D P1 proteins present significant sphingomyelinase activity as shown by the breakdown of the substrate, being the activity of L. adelaida SMase D protein lower than that of the L. intermedia SMase P1. However, L. adelaida recombinant SMase D activity was approximately ten times higher than that of the crude L. adelaida venom. Since the recombinant L. adelaida protein was endowed with sphingomyelinase activity, we named it as SMase D A1.

Cross-reactivities of SMase D A1 of L. adelaida
The protein profile from the Loxosceles spp venoms and the recombinant SMases D were analyzed by SDS-PAGE followed by silver staining. Figure 4A shows that the venoms from L. intermedia, L. laeta, L. gaucho and L. adelaida differ in composition, number and intensity of bands. It can also be observed that the recombinant proteins were purified to homogeneity and that P1 and A1 exhibit similar molecular weight. SMase I presents Mr of approximately 33 kDa, and SMases P1 and A1, around 37 kDa and 36 kDa, respectively. In order to analyze the inter-and intra-species cross-reactivities, horse polyclonal antiserum raised against a combination of the SMases D P1 and P2 from L. intermedia and Smase D I from L. laeta, were used in western blotting reactions. Figure 4B shows that the antiserum was able to recognize the purified recombinant proteins, SMases D A1, I and P1, as well as bands of approximately 35 kDa in the Loxosceles spp spider venoms. The slightly higher Mr of the recombinant proteins is attributed to the extra N-terminal tag, which increased the size of L. adelaida SMase A1, L. intermedia SMase P1 and L. Laeta SMase I proteins by, approximately, 4 kDa, 3 kDa and 1 kDa, respectively.

L. adelaida recombinant SMase D induces haemolysis and dermonecrosis
Sphingomyelinases D isolated from Loxosceles spider venoms have been shown to be able to transform human erythrocytes in activators of the complement system (Tambourgi et al., 1995(Tambourgi et al., , 1998(Tambourgi et al., , 2000. In order to assess whether the recombinant SMase A1 could also induce Complement-dependent haemolysis, erythrocytes were incubated with increasing amounts of the SMases A1, P1 or crude venom. Although the activity of SMase A1 was relatively lower compared with the SMase P1 and L. adelaida venom, as shown in Figure 5, L. adelaida recombinant SMase A1, as well as SMase P1 and L. adelaida venom were able to render erythrocytes susceptible to lysis by autologous Complement.  The ability of L. adelaida recombinant SMase A1 to induce dermonecrotic lesions was tested by injecting rabbits with 5 µg of the toxin. The animals received buffer, as negative control, and 5 µg of SMase D P1 or venom as positive controls. A typical loxoscelic lesion, as revealed by the presence of oedema, oerythema and mild tenderness, developed in the skin area injected with the recombinant proteins or venom within a few hours of injection and approximately 24 h post injection, necrosis with gravitational spread and scar were observed at the inoculation site ( Figure 6A). Figure 6B shows that the dermonecrotic action of L. adelaida SMase D A1 is dose dependent, but that this activity was less intense than that exhibited by the crude venom and SMase D P1. Histological analysis of the skin samples obtained from PBS inoculated animals showed a thin epidermis and a normal pattern for the collagenous area and muscle fibers (Figure 7, panels A/A1/A2). Despite the L. adelaida recombinant SMase D A1 has exhibited lower dermonecrotic activity than the Smase D P1, skin samples obtained from the recombinant proteins or venom inoculation sites showed a thin epidermis, dissociation of the collagenous fibers due to the oedema, degeneration of muscle fibers, moderate haemorrhage in the superficial dermis and intense neutrophil infiltration in the deep dermis and musculature (Figure 7).  www.intechopen.com 3.5 Overall structure description MD analysis of L. adelaida SMase D A1 in silico model converged to a RMSD of 1.17 Ǻ and showed a stable behavior over the simulation. Global and local stereochemical assessment indicated a very good model for comparative structural analyses. SMase D A1 displays a typical TIM ( / ) 8 -barrel fold and its active-site cleft, formed by the metal-binding site (Asp, Glu and Asp) and the two catalytic histidines, is furthermore surrounded by the catalytic loop (residues 46-60), variable loop (residues 167-175), flexible loop (residues 196-203) and other short hydrophobic loops (Figure 8). Based on the current classification of SMases D (Murakami et al., 2006), SMase A1 belongs to class II containing an additional disulphide bridge (Cys53-Cys201), which connects the catalytic loop to flexible loop (Figure 8). This feature, not present in class I SMases D, diminishes significantly the active-site volume and also alters the inherent flexibility exhibited by the flexible loop. Beyond that, all the structural features observed in class I SMases D are fully conserved and details concerning the action mechanism are well described in Murakami et al., 2005 andMurakami et al., 2006. Fig. 8. Structure superposition of Smase D I (class I) and L. adelaida SMase D A1 (class II). The residues involved in metal-ion binding and catalysis are presented in atom colors (PDB code: 1XX1). Differences in the catalytic, flexible, and variable loops in L. laeta SMase D I (blue) and L. adelaida SMase D A1 (green).

Sphingomyelin-binding mode to SMases D
Despite the structure of a SMase D member has been solved, there is neither structural nor biophysical data relating the binding mode of the sphingomyelinase (SM) into the active-site cleft of SMases D. Thus, in order to shed light into the structural determinants for recognition and binding of SM by SMases D, a MD simulation was performed using the crystal structure of the SMase D I and the SM docked into enzyme taking into account the crystallographic position of the sulphate ion, which provides a good notion how the phosphate group from SM is oriented in the active-site cavity. As observed in Figure 9, the RMSD of the protein is low (~ 1.5 Å), whereas the SM shows a higher variation in the first 3000 ps (~ 5.15 Å) and after that become more stable (~ 3 Å). Although, the aliphatic tails of the substrate exhibit high flexibility as observed by RMSD analysis over the simulation, the polar head is stable. The binding analysis showed that the substrate-enzyme interaction increases during the simulation and it stabilizes around 291 Kcal/mol (Figure 9). In 10 ns of MD simulation, the model achieved a stable conformation with some fluctuations in the aliphatic tails as expected (Figure 9). The choline head is buried in the active-site cleft making van der Waals contacts with Trp230, which is conserved among SMases D ( Figures  10A-B). The phosphate moiety is coordinated by the magnesium ion as expected, forming a tetrahedral cage of the ion along with the three acidic residues ( Figure 10B). The residue Lys93, which is also highly conserved, is found in a distance range that permits to interact with the carbonyl oxygen of the sphingosine backbone. Val89, Ser132, Asp164, Ser166, Pro134, Pro168, Tyr169, Leu170, Leu198, Tyr228 and Met250 are also participating in the coordination of the substrate ( Figure 10B). These data corroborates with previous crystallographic studies, whose suggested the importance of Lys93, Trp230 and other aromatic residues in the recognition and interaction of the substrate.

Discussion
Previously, we have characterized the biochemical and biological properties of L. adelaida venom and evaluate the toxic potential of envenomation by this non-synanthropic Loxosceles species (Pretel et al., 2005). The biological activities of the L. adelaida venom was compared to that of Loxosceles gaucho, a synanthropic species of medical importance in Brazil. L. adelaida venom showed a similar potential to induce haemolysis, dermonecrosis and lethality as L. gaucho venom. Thus, showing that the troglophile Loxosceles species, L. adelaida, commonly