Subscriber access provided by UNIV OF DURHAM
Article
A family of papain-like fungal chimerolectins with distinct Ca2+-dependent activation mechanism Gabriele Cordara, Dipankar Manna, and Ute Krengel Biochemistry, Just Accepted Manuscript • DOI: 10.1021/acs.biochem.7b00317 • Publication Date (Web): 30 Jun 2017 Downloaded from http://pubs.acs.org on July 14, 2017
Just Accepted “Just Accepted” manuscripts have been peer-reviewed and accepted for publication. They are posted online prior to technical editing, formatting for publication and author proofing. The American Chemical Society provides “Just Accepted” as a free service to the research community to expedite the dissemination of scientific material as soon as possible after acceptance. “Just Accepted” manuscripts appear in full in PDF format accompanied by an HTML abstract. “Just Accepted” manuscripts have been fully peer reviewed, but should not be considered the official version of record. They are accessible to all readers and citable by the Digital Object Identifier (DOI®). “Just Accepted” is an optional service offered to authors. Therefore, the “Just Accepted” Web site may not include all articles that will be published in the journal. After a manuscript is technically edited and formatted, it will be removed from the “Just Accepted” Web site and published as an ASAP article. Note that technical editing may introduce minor changes to the manuscript text and/or graphics which could affect content, and all legal disclaimers and ethical guidelines that apply to the journal pertain. ACS cannot be held responsible for errors or consequences arising from the use of information contained in these “Just Accepted” manuscripts.
Biochemistry is published by the American Chemical Society. 1155 Sixteenth Street N.W., Washington, DC 20036 Published by American Chemical Society. Copyright © American Chemical Society. However, no copyright claim is made to original U.S. Government works, or works produced by employees of any Commonwealth realm Crown government in the course of their duties.
Page 1 of 48
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60
Biochemistry
A family of papain-like fungal chimerolectins with distinct Ca2+-dependent activation mechanism Gabriele Cordara1,†,‡,*, Dipankar Manna1, †,‡ and Ute Krengel1,* 1
Department of Chemistry, University of Oslo, PO Box 1033 Blindern, 0315 Oslo, Norway
Keywords: calcium-binding protein, crystal structure, cysteine protease, gating tyrosine, protease inhibitor
ACS Paragon Plus Environment
1
Biochemistry
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60
Page 2 of 48
ABSTRACT
An important function of fungal lectins is to protect their host. Marasmius oreades agglutinin (MOA) is toxic to nematodes and exerts its protective effect through protease activity. Its proteolytic function is associated with a papain-like dimerization domain. The closest homolog of MOA is Polyporus squamosus lectin 1a (PSL1a). Here we probed PSL1a for catalytic activity and confirmed that it is a calcium-dependent cysteine protease, like MOA. The X-ray crystal structures of PSL1a (1.5 Å) and MOA (1.3 Å) in complex with calcium and the irreversible cysteine protease inhibitor E-64 revealed the structural basis for their mechanism of action. The comparison with other calcium-dependent proteases (calpains, LapG) reveals a unique metaldependent activation mechanism relying on a calcium-induced backbone shift and intra-dimer cooperation. Intriguingly, the enzymes appear to use a tyrosine-gating mechanism instead of propeptide processing. A search for potential MOA orthologs suggests the existence of a whole new family of fungal-specific chimerolectins with these unique features.
ACS Paragon Plus Environment
2
Page 3 of 48
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60
Biochemistry
INTRODUCTION Lectins are defined as proteins that “possess at least one non-catalytic domain that binds reversibly to a specific mono- or oligosaccharide”.1 They are classified into merolectins (single carbohydrate-binding domain), hololectins (two or more carbohydrate-binding domains) and chimerolectins (carbohydrate-binding module fused to one or more differently purposed domains).1 Fungi are a well-known and mostly unexploited source of chimerolectins.2 Their bioactivity is often associated with the presence of a catalytic domain.3, 4 The Marasmius oreades agglutinin (MOA) is a histo-blood group B-specific chimerolectin extracted from the fruiting bodies of the fairy ring mushroom Marasmius oreades.5 The crystal structure of MOA shows a ricin B chainlike / papain-like two-domain partition.6 Its papain-like domain exhibits calcium- or manganese (II)-dependent proteolytic activity.7 MOA exerts a cytotoxic activity, similarly to other proteins carrying the ricin B chain-like fold.8,
9
The catalytic domain is a key element in mediating
toxicity against nematodes or NIH/3T3 cells.10, 11 MOA has been crystallized in either a calciumfree (PDB ID: 2IHO)6 or calcium-bound form (PDB ID: 3EF2)12. Calcium binding triggers a conformational change 12 that is essential for the lectin’s proteolytic activity.7, 10, 13 The closest homolog of MOA is the Polyporus squamosus lectin 1a (PSL1a; 38% sequence identity)14, 15, which like MOA exhibits cytotoxic activity.16 Compared to MOA, this 31.2 kDa homodimeric lectin has different ligand specificity, to terminal sialic acid-containing glycotopes.15, 17 The only available crystal structure of PSL1a (PDB ID: 3PHZ)18 closely mimics the tertiary structure of MOA in its calcium-free state.18 In particular, the dimerization domain of PSL1a retains the papain-like fold and its hallmark L(eft)/R(ight)-domain partition (Figure S1).19 The good conservation of the papain-like domain (r.m.s.d. = 1.0 Å) extends to the components of
ACS Paragon Plus Environment
3
Biochemistry
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60
Page 4 of 48
the putative catalytic machinery. The catalytic triad of MOA (Cys215, His257, Glu274) perfectly aligns with three equivalent residues of PSL1a (Cys208, His248, Glu266). Likewise, each MOA residue involved in calcium coordination has a counterpart in PSL1a. The good conservation of structural features, and their importance for cytotoxic activity,10,
11, 16
suggests a conserved
enzymatic function. However, the catalytic activity of PSL1a has not been probed to date. Here we present a thorough characterization of PSL1a catalytic activity, its structural underpinnings, and compare PSL1a to MOA. Both lectins were crystallized in complex with the E-64 cysteine protease inhibitor and calcium. The structures suggest an unusual metal-dependent activation mechanism, involving the interplay with the symmetry-related protomer, which appears to be common to a new family of fungal chimerolectins.
ACS Paragon Plus Environment
4
Page 5 of 48
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60
Biochemistry
MATERIALS AND METHODS
Expression and purification For PSL1a production (UniProt reference: Q75WT9), ArcticExpress cells (DE3) cells (Agilent Technologies) were transformed with the pET43.1a-PSL1a vector, and cultivated according to the protocol suggested by the supplier. Gene expression was induced at 11°C for 24h with 0.1 mM IPTG. Subsequently, E. coli pellets were collected by centrifugation and stored at -80°C overnight before lysis. After thawing, the bacterial pellets were resuspended in a lysis buffer containing 50 mM Tris-HCl pH 8.0, 0.15 M NaCl, 2 mM EDTA, 1x concentrated cOmplete protease inhibitor cocktail EDTA free (Roche Diagnostics Ltd), 1 µl/ml Benzonase nuclease (Thermo Scientific) and 4 mg/ml hen egg white lysozyme. After incubation on a shaker for two hours at RT, the insoluble fraction was removed by two rounds of centrifugation (20000 rcf, 45 min). The clarified cell lysate was passed through a D-Gal-sepharose affinity column (Thermo Scientific), followed by extensive washing with 20 mM imidazole pH 8.0 and elution of the protein using a single step 1.0 M D-Gal gradient. The fraction containing the eluted PSL1a was concentrated using a 10000 MWCO PES membrane (Vivaspin, Sartorius AG) to a volume of ≈500 µl. Polishing of the protein preparation was carried out by size-exclusion chromatography (SEC) using a Superdex 75 10/300 GL gel-filtration column (Tricorn, GE Healthcare Life Sciences) with 20 mM imidazole pH 8.0, 2 mM EDTA, 0.2M D-Gal, 0.15M NaCl and DTT 2 mM. The fractions containing the purified protein were pooled, concentrated to a final protein concentration of ≈10-15 mg/ml using concentrator tubes with a 10000 MWCO PES membrane (Vivaspin, Sartorius AG) and underwent three rounds of buffer exchange against 20 mM imidazole/HCl pH 8.0, 2 mM EDTA, 2 mM DTT.
ACS Paragon Plus Environment
5
Biochemistry
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60
Page 6 of 48
MOA (UniProt reference: Q8X123) was expressed in E. coli BL21 (DE3) cells transformed with the pT7-LO-MOA expression vector, and purified following an established protocol.7,
11
The protein was stored at -80°C at ≈10-15 mg/ml in 10 mM Tris-HCl pH 8.0, 5 mM DTT. The structure described here was obtained from the Cys63Ala variant of MOA, which was generated by site-directed mutagenesis through the QuikChange II kit (Stratagene) following the protocol provided by the manufacturer. This variant was produced as a preventive measure against a possible cross-reactivity with thiol-modifying compounds, since Cys63 was found to be chemically modified by N-ethylmaleimide in an earlier experiment. Parallel experiments with wild-type MOA yielded equivalent results, although with marginally lower resolution.
Activity assay The activity of PSL1a or MOA was tested using native α1-antitrypsin (α1-AT) from human plasma (Sigma-Aldrich) as the target substrate, following an established protocol.7 A mixture containing 0.4 mg/ml α1-AT, 50 mM Na-HEPES pH 7.5, 10 mM DTT, 10 mM CaCl2 and 0.2 mg/ml PSL1a or MOA was incubated at 37°C for 24 h. Variations of the protocol to profile the enzymatic activity of PSL1a included the use of a 50 mM concentration of a different buffer than Na-HEPES pH 7.5 or a 10 mM concentration of a different chloride salt than CaCl2. The reaction was stopped by adding 5 µl of SDS-PAGE sample buffer 4x to 15 µl of the reaction mixture and boiling at 100°C for 10 min. Samples were loaded alongside with SeeBlue Plus2 Pre-Stained Standard (Invitrogen) molecular weight markers on a NuPAGE SDS-PAGE 4-12% gel for the electrophoretic run; the gels were stained with Coomassie Brilliant Blue dye (Sigma-Aldrich).. Double digestion with a peptide:N-glycanase was carried out by pre-incubating α1-AT 24 h at 37°C in presence of 2 UN of PNGase F from Elizabethkingia miricola in 50 mM Na-HEPES pH
ACS Paragon Plus Environment
6
Page 7 of 48
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60
Biochemistry
7.5. The reaction mixture containing the deglycosylated α1-AT was complemented with 10 mM DTT, 10 mM CaCl2, 0.2 mg/ml PSL1a or MOA and further incubated as described for the untreated substrate. Gel images were analyzed using the Fiji distribution of the ImageJ manipulation software.20
Crystallization MOA crystals were obtained for the MOA Cys63Ala variant (concentrated to 5 mg/ml). The protein solution contained the Galα1,3(Fucα1,2)Gal trisaccharide (Dextra Laboratories) in a 1:20 MOA:sugar molar ratio and the proteolytic activity inhibitor E-64 in a MOA:E-64 1:3 molar ratio. Crystals grew after a two-week incubation period at 20°C from a crystallization mixture containing 0.1 M imidazole pH 8.0, 15% PEG 8000, 7.5% DMSO and 0.2 M calcium acetate, pre-mixed with the protein solution in a 1:1 ratio (drop volumes). PSL1a crystals were obtained from a solution containing 5 mg/ml PSL1a, 20 mM imidazole pH 8.0, 5 mM calcium chloride and the DMSO-dissolved E-64 inhibitor in a PSL1a:E-64 1:3 molar ratio. The protein-inhibitor mixture was pre-incubated for 24h before mixing in a 1:1 ratio (drop volumes) with the reservoir solution. Hanging drops were transferred after 1h incubation from initial conditions containing 0.1 M citrate pH 5.6, 0.2 NaSCN and 50% PEG 3350 to wells containing the same components, but a lower PEG 3350 concentration (40% w/v). All crystals were cryoprotected using the crystallization solution supplemented with 15% ethylene glycol, successively frozen in liquid nitrogen for data collection.
Data collection, processing, scaling and structure determination
ACS Paragon Plus Environment
7
Biochemistry
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60
Page 8 of 48
Diffraction data were collected at the European Synchrotron Radiation Facility (Grenoble, France) at beamlines ID29 and ID23-2. The images were processed and scaled using the XDS software package.21 Scaling statistics are given in Table 1. MOA and PSL1a crystals belong to different space groups, containing either a single protomer (MOA, P6322) or the biological dimer (PSL1a, P212121) in the asymmetric unit. The structures were solved by molecular replacement with PHASER,22 using the calcium-bound structure of MOA (PDB ID: 3EF2)12 or the calciumfree structure of PSL1a (PDB ID: 3PHZ)18 as search models, respectively. The MOA search model was modified by stripping off atoms with HETATM record (including the calcium ions), deleting the Pro54-Val56 loop, and by mutating residues showing flexible or generally poorly defined side chains to Ala. The mFo-DFc difference electron density maps showed well-defined, positive density peaks at the metal binding sites, the three sugar binding sites and in the active site cleft. In PSL1a, additional positive and negative density peaks suggest a conformational change involving residues Ala165-Gly177.
ACS Paragon Plus Environment
8
Page 9 of 48
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60
Biochemistry
Table 1. Data collection and refinement statistics
A. Data collection Beamline Wavelength (Å) Space group Cell parameters: a, b, c (Å) Resolution (Å)a Rmerge (%)ab Rmeas ac Rp.i.m. ad CC1/2ae Mean I/σ(I)a Completeness (%)a Multiplicitya Unique reflectionsa B. Refinement Resolution (Å)a Rwork/Rfree (%)f Macromolecules / a.s.u. No. atoms Protein Water Ligands B-factor (Å2) Protein Water Ligands r.m.s.d. from ideal values Bond lengths (Å) Bond angles (deg.) Ramachandran plot Core region (%) Outliers (%) PDB ID
a
PSL1a-E-64
MOA-E-64
ESRF ID23-2 0.8726 P212121 75.7 94.4 102.8 47.2-1.5 (1.57-1.54) 17.8 (>100.0) 20.3 (>100.0) 9.8 (>100) 99.1 (33.3) 6.2 (0.6) 99.0 (93.8) 4.1 (3.8) 107830 (4995)
ESRF ID23-2 0.8726 P6322 120.9 120.9 99.9 46.4-1.3 (1.32-1.30) 7.0 (56.4) 7.8 (68.7) 3.3 (38.2) 99.9 (61.6) 14.0 (1.6) 94.2 (61.2) 5.2 (2.5) 99395 (3104)
47.2-1.5 22.1 / 24.8 2
46.4-1.3 16.0/19.2 1
4372 451 102
2386 331 147
14.9 23.3 19.2
11.4 24.2 15.1
0.02 2.0
0.02 2.3
97.3 0.0 5MUA
97.4 0.0 5MU9
Values in parentheses refer to highest resolution shell
b
Rmerge = ΣhΣj |Ihj - 〈Ih〉| / ΣhΣj Ihj , where 〈Ih〉 is the mean intensity of symmetry-related reflections Ih
ACS Paragon Plus Environment
9
Biochemistry
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60
c
Page 10 of 48
Rmeas = Σh [Nh/(Nh-1)]1/2 Σi |Ihj - 〈Ih〉| / ΣhΣi Ihj , where N is the redundancy of reflection h 23
d
Rp.i.m.= Σh [1/(Nh-1)]1/2 Σj |Ihj - 〈Ih〉| / ΣhΣj Ihj 24
e
The high resolution cut-off was chosen despite the low CC1/2 ensuring the presence of a low signal-to-noise ratio by visual inspection of the electron density map f
Rfree was calculated from 5% of randomly selected data for each data set
Model building and refinement Model building and refinement were carried out in cycles, using Coot25 and REFMAC526, respectively. In an initial step, ill-defined side chains and loops were removed from the original PHASER output. The models were subsequently rebuilt by adding the missing structural elements in a step-wise fashion as the quality of the electron density map improved, including (in this order) the metal ions, the sugar ligands, water molecules and additional buffer components (e.g. acetate, ethylene glycol). The metal ions were modeled as calcium based on previous structural data12, the presence of 0.2 M Ca2+ in the crystallization solution and a compatible ligand coordination27, 28. The E-64 inhibitor was modeled at the end of the refinement process, when the difference electron density at the active site allowed unambiguous tracing of the inhibitor. Ligand occupancy was determined by minimizing the residual difference electron density for the inhibitor molecule and taking the B-factors of nearby interacting atoms into account. A final refinement step included anisotropic refinement of the MOA-E-64 model. PSL1a diffraction data beyond an Rmeas >0.6 and Ī/σ(Ī)
