Crystal structure of native Anopheles gambiae Serpin-2, a negative regulator of melanization in mosquitoes

Abstract

Serpins are the dominant group of protease inhibitors in metazoans that control a wide variety of biological processes including major innate immune reactions. One of these inhibitors, SRPN2, controls melanization in mosquitoes – a powerful, arthropod-specific innate immune response. SRPN2 depletion from the hemolymph of adult female mosquitoes significantly reduces longevity and therefore this serpin is a potential target for novel insecticides. We report here the crystal structure of SRPN2 in its native conformation from the African malaria mosquito, Anopheles gambiae to 1.75 Å resolution. SRPN2 adopts a similar fold as observed for other serpins with a core of three β-sheets surrounded by nine α-helices with an exposed reactive center loop (RCL) that extends from the protein body. Similar to other native serpin structures, several residues within the reactive center loop were disordered and could not be modeled. Intriguingly, the N-terminal hinge of the RCL in SRPN2 was found to be inserted into β-sheet A, suggesting a potential activation mechanism analogous to heparin-mediated activation of Antithrombin III.

INTRODUCTION

Extracellular proteolytic cascades regulate a wide variety of biological processes in vertebrates including blood clotting and complement activation. Similarly, in invertebrates they regulate major innate immune reactions, including hemolymph coagulation, the Toll pathway, and prophenoloxidase (proPO) activation that leads to melanization. Serpins, members of a large, wide-spread superfamily of serine proteinase inhibitors, play a critical role in these immune reactions by controlling accidental triggering of the cascades and regulating the spread of the signal once the cascades are activated (recently reviewed in ¹). Serpins typically consist of 350–400 amino acid residues and adopt a canonical fold of three β-sheets (A–C) in the center of the protein surrounded by up to nine α helices. The reactive center loop (RCL), located at the carboxy terminus, is positioned away from the body of the serpin, highly flexible, and contains the scissile bond (designated P1–P1'), which is cleaved by the target protease during inhibition (recently reviewed in ²).

Most serpins inhibit S1 family serine proteinases while others function as cross-class inhibitors that can also target cysteine proteinases ³. Other serpins do not function as proteinase inhibitors but play a role in e.g. hormone transport ⁴. Inhibitory serpins function as single-use suicide substrates. Upon cleavage of the scissile bond in the RCL by the target protease, the serpin undergoes a large conformational change. The RCL is inserted into β-sheet A in the core of the serpin, distorting the active center of the protease. This traps the acyl-intermediate of the reaction, resulting in the formation of SDS-stable, covalent complexes between serpin and target protease ⁵.

We recently described Serpin 2 (SRPN2) from the African malaria mosquito, Anopheles gambiae, to be a key regulator of the melanization response. Depletion of SRPN2 from the hemolymph of adult female mosquitoes causes the formation of melanotic tumors, and killing of allochthonous malaria parasite species ^6–8. SRPN2 is a functional inhibitor of serine proteases, and one of its endogenous targets is the clip-serine protease CLIPB9, which was shown to cleave and activate proPO ⁹. This result indicated that SRPN2 regulates melanization by inhibiting proPO activation. In addition, SRPN2 depletion significantly shortens mosquito longevity, making it an attractive insecticide target for malaria control. Here we report the crystal structure of An. gambiae SRPN2, the first serpin fold from a dipteran insect, in its native conformation to 1.75Å resolution.

MATERIALS AND METHODS

Protein production and purification

The pET28a_His-SRPN2 plasmid ⁹ was used to transform Escherichia coli BL-21 DE3 pRare; two colonies were picked and inoculated in 15 ml of LB medium containing 50 μg/ml kanamycin and 25 mg/ml chloramphenicol. A starter culture was transferred to 1.5 l of LB medium containing antibiotics and grown to an OD₆₀₀ ~0.8 at 37°C. Induction was carried out with 0.1mM IPTG at 15°C overnight with shaking. Cells were harvested via centrifugation at 8,500 rpm for 10 minutes. Cell pellets were resuspended in 50 mL of buffer A (50mM NaCl, 20mM Tris-HCl, pH of 8.0) supplemented with protease inhibitor cocktail (Roche) and lysed by sonication. The insoluble material was pelleted by centrifugation at 19,000 rpm for 30 minutes and the clarified lysate was loaded onto a 5mL HisTrap HP column. All purification steps were conducted using an AKTA Xpress purification system (GE Healthcare) at 4°C. Non-specifically bound proteins were washed using 10% buffer B (500mM Imidazole, 50mM NaCl, 50mM Tris-HCl, pH 8.0). Elution was carried out with a linear gradient from 10% to 100% buffer B and all elution peaks of interest were collected and analyzed by SDS-PAGE. Fractions containing SRPN2 were pooled and loaded onto a HiTrap Q HP anion exchange column equilibrated with buffer A. Elution was carried out with a linear gradient from 0% to 100% Buffer C (1M NaCl, 20mM Tris, pH 8.0) and the purity was analyzed by SDS-PAGE. SRPN2 fractions, which eluted at approximately 50% buffer C, were pooled and concentrated to 6.2 mg/mL for crystallization screening.

Crystallization and data collection

Crystallization screening was conducted in Compact Jr. (Emerald biosystems) sitting drop vapor diffusion plates at 20°C using 0.5μl of protein and 0.5μl of crystallization solution equilibrated against 100μl of the latter. Prismatic, hexagonal crystals were obtained in two days from the Index Screen (Hampton Research) condition #24 (2.8 M Sodium Acetate, pH 7.0). Samples were transferred to a solution containing 70% crystallization solution and 30% glycerol for approximately 30 seconds before freezing in liquid nitrogen for data collection. X-ray diffraction data were collected at 100K at the Advanced Photon Source (APS) IMCA-CAT beamline 17ID using an ADSC Quantum 210r CCD detector. Intensities were integrated and scaled using the HKL2000 package ¹⁰.

Structure Solution and Refinement

Structure solution was carried out by molecular replacement with BALBES ¹¹ in the space group P6₃ searching for one molecule in the asymmetric unit, which produced a homology model for the rotation and translation searches from PDB: 2GD4 (Antithrombin-S195A factor Xa heparin structure). The R/R_free converged at 44%/49% following initial refinement. At this stage, main chain atoms of the model fit the electron density maps reasonably well. However, the side chains were difficult to discern. The side chains were then removed, the model refined and the resulting model was improved using the automated model building package in PHENIX ¹². The refined model obtained from the autobuilding step converged at R=22%/R_free=27%. This model was then used for molecular replacement search with PHASER ¹³ against higher resolution diffraction data that was subsequently collected. The final model was obtained by iterative rounds of refinement and manual model building with PHENIX and COOT ¹⁴, respectively. Anisotropic displacement parameters were modeled using TLS ¹⁵ refinement, in which the protein chain was partitioned into 3 groups determined from the TLSMD server ¹⁵. Disordered side chain atoms were truncated to the point where electron density could be observed. Figures were prepared using the CCP4MG package ¹⁶. Relevant crystallographic data are provided in Table 1. Structure comparison was conducted using secondary structure matching with the program Superpose ¹⁷ via the CCP4 interface ¹⁸. Coordinates and structure factors have been deposited to the Protein Databank with the accession code 3PZF.

Table 1.

Crystallographic data for SRPN2 refined to 1.75Å resolution.

	SRPN2
Data Collection
Unit-cell parameters (Å, o)	a = 96.54, c = 78.29
Space group	P63
Resolution (Å)1	30.0–1.75 (1.81–1.75)
Wavelength (Å)	1.0000
Temperature (K)	100
Observed reflections	448,554
Unique reflections	40,410
<I/σ(I)>1	35.9 (2.3)
Completeness (%)1	96.5 (78.7)
Redundancy1	11.1 (8.6)
Rsym (%)1, 2	5.9 (41.5)
Refinement
Resolution (Å)	26.25–1.75
Reflections (working/test)	38,251 / 2,042
Rfactor / Rfree (%)3	17.3 / 20.7
No. of atoms (protein/water)	2,953 / 165
Model Quality
R.m.s deviations
Bond lengths (Å)	0.011
Bond angles (o)	1.233
Average B factor (Å2)
All Atoms	35.7
Protein	35.6
Water	37.2
Coordinate error, maximum likelihood (Å)	0.19
Ramachandran Plot
Most favored (%)	97.0
Additionally allowed (%)	3.0

¹Values in parenthesis are for the highest resolution shell.

²R_sym = Σ _hklΣ_i |I_i(hkl) - <I(hkl)>| / Σ_hklΣ_i I_i(hkl), where I_i(hkl) is the intensity measured for the ith reflection and <I(hkl)> is the average intensity of all reflections with indices hkl.

³R_factor = Σ_hkl ||F_obs (hkl) | - |F_calc (hkl) || / Σ_hkl |F_obs (hkl)|; Rfree is calculated in an identical manner using 5% of randomly selected reflections that were not included in the refinement.

RESULTS AND DISCUSSION

Crystallization screening of SRPN2 yielded two crystal forms at 20°C. The first is a primitive hexagonal form (P6₃) yielding hexagonal plates obtained in 2.8 M Sodium Acetate, pH 7.0 from a sitting drop vapor diffusion experiment. The same crystallization conditions in a microbatch experiment under paraffin oil produced isomorphous crystals with a modified habit in which the crystals were elongated along the crystallographic c-axis. This particular crystal form consistently yielded diffraction to at least 2.0 Å resolution or better. A second crystal form of SRPN2 was obtained in 2M Li₂SO₄, 100mM Hepes pH 7.6 that belonged to the space group I4₁32 with a = 282.9 Å. These crystals diffracted to significantly lower resolution to a maximum of 3.6 Å.

The structure of the hexagonal form (P6₃) of SRPN2 was solved by molecular replacement using diffraction data to 1.75Å resolution and contained a single molecule in the asymmetric unit. Data collection and refinement statistics are provided in Table 1. The structure of the I4₁32 form was solved by molecular replacement searching for 4 molecules in the asymmetric unit. Two sets of dimers were located using Molrep in the space group I4₁32. However, no solution was found in I432. Following refinement, the average B-factors were approximately 2-fold higher for one of the subunits and the resulting electron density maps were ambiguous for certain regions making it difficult to build an accurate model. Therefore, the structure was not refined further.

SRPN2 adopts a similar fold as observed for other serpins (Fig. 1), however significant differences are present which is indicated by the RMS deviations calculated between Cα atoms for homologs (Table S1). Similar to several other native serpin structures ¹⁹, the RCL residues between I366 to D375 in SRPN2 were disordered and could not be modeled. The most notable difference between SRPN2 and most native serpin folds is the conformation of the N-terminal hinge region of the RCL, which is partially inserted between strands 2 and 3 of β-sheet A (Fig. 2A). This conformation has so far only been seen in a small number of serpins, non-heparin bound antithrombin III (ATIII) ²⁰, heparin cofactor II ²¹, murine anti-chymotrypsin (mACT) ²², and Spn48 from Tenebrio molitor ²³. The N-terminal hinge of the RCL that is inserted into and interacts with the β-sheet is mainly localized to the residues from Leu 356 to Ala 360. This region could be accurately modeled to the electron density maps (Fig. S1). The insertion is stabilized by several hydrogen bonds, with the side chains of Arg 174, Tyr 251, Asn 354, Ser 358 and Glu 359 forming the majority of the main contacts (Fig. 2B). Similarly, the electron density maps of the Cubic I form obtained following refinement clearly showed that the residues Leu 356 to Ala 360 are inserted into β-sheet A (Fig. S2). As was observed for the initial crystal form, no crystal contacts were present near these residues which indicated that this observed conformation is not an artifact of crystal packing.

Figure 1. Stereoview of the overall structure of SRPN2.

The A-sheet is colored red, B-sheet blue, C-sheet in tan, α-helices are shown in green and the RCL is colored cyan. PDB: 3PZF

Figure 2. The hinge region of SRPN2 is partially inserted into β-sheet A.

(A) Superposition of serpin structures that contain loop insertions at the N-terminal hinge region of the RCL. SRPN2 (magenta, PDB: 3PZF), non-heparin bound ATIII (coral, PDB: 1T1F), heparin cofactor II (blue, PDB: 1JMJ), mACT (green, PDB: 1YXA), and SPN48 (cyan, PDB: 3OZQ). (B) Hydrogen bond interactions between the RCL loop insertion and strands 2 and 3 of β-sheet A in SRPN2.

Partial insertion of the hinge region can reduce serpin activity as it moves the RCL closer to the main body of the protein, making the scissile bond less accessible to the protease target as compared to a fully exposed loop. In some cases, this can be overcome by binding of heparin to the serpin molecule. In ATIII, and most likely also HCII, heparin binding on or near to homologous regions of helix D favors expulsion of the hinge and increased inhibitory activity of the serpin ^21,24,25. Similarly, heparin binding to Spn48 increases its activity ²³. However, the lysine and arginine residues responsible for binding of heparin in ATIII and HCII are not conserved in Spn48, and heparin binding might be conferred by basic residues on helix I ²³ (Fig. S3). In SRPN2, the basic residues in helices D and I are not conserved (Fig. S3) and calculations of surface charges on the protein do not indicate clearly an alternative heparin binding site. In accordance with these observations, heparin pentasaccharide does not increase the inhibitory activity of SRPN2 on its cognate protease, AgCLIPB9 (An and Michel, unpublished). Future work is needed to elucidate whether the partial hinge insertion in the native SRPN2 molecule is indicative of a novel regulatory mechanism of serpin function.