Crystal structure of Schistosoma mansoni cathepsin D1 in complex with a nanobody reveals the conformation of the propeptide-bound state

Schistosoma mansoni cathepsin D1 (SmCD1) has been shown to be an essential enzyme for helminth metabolism due to its role in haemoglobin degradation: a key amino-acid source for the developing parasite. Therefore, the enzyme is a potential target for the development of antischistosomal inhibitors. SmCD1 has significant sequence identity to cathepsin D-like proteases found in other schistosome species and homology to mammalian aspartic proteases. Here, we report the first crystal structures of a helminth cathepsin D, SmCD1, and have identified a single-domain antibody (nanobody) that specifically binds to SmCD1 with nanomolar affinity but does not recognize human cathepsin D. We have mapped the epitope of the nanobody by determining the crystal structure of the enzyme–nanobody complex, revealing the conformation of SmCD1 in the propeptide-bound state.

1. Introduction

Schistosomiasis presents a devastating global health challenge, prevalent in sub-Saharan Africa, and is classified as a widespread neglected tropical disease (NTD) by the World Health Organization (WHO; Van Der Werf et al., 2003; Steinmann et al., 2006; World Health Organization, 2021). Of the three major species infecting humans, Schistosoma mansoni is causative of acute intestinal and hepatic schistosomiasis (Han et al., 2009). Despite its significant impacts, existing treatments for schistosomiasis are heavily reliant on mass administration of praziquental (PZQ), which shows negligible effects against juvenile parasitic stages and introduces challenges for drug resistance (Humphries et al., 2012; Xiao et al., 2018; Aboagye & Addison, 2022). Similarly, current antibody diagnostic tests exhibit variable accuracy for low-intensity infections, displaying inconsistent sensitivity between schistosome species, and lack of correlation to clinical status (Valli et al., 1997; Ochodo et al., 2015; Wilson, 2017; Ogongo et al., 2022). Thus, there remains a need for further study to aid understanding and contribute to the development of accessible, sensitive tools for rapid detection and neutralization (Ajibola et al., 2018; Molehin et al., 2022; Chala, 2023).

Helminth aspartyl proteases have been extensively reported as playing crucial roles in parasite survival, particularly in the developing trematode (Cesari et al., 1998; Liu et al., 2014; Brindley et al., 2001). S. mansoni cathepsin D1 (SmCD1) is considered a potential therapeutic target because of the pivotal role that the enzyme plays in host haemoglobin degradation and parasite maturation, as indicated by RNAi studies (Morales et al., 2008). SmCD1 is highly conserved across species, including >82% shared identity with orthologues from S. japonicum and S. bovis, and retains significant sequence similarity to human cathepsin D (hCD) and other mammalian and invertebrate homologues (Morales et al., 2004; Silva et al., 2011; Kang et al., 2019). SmCD1 is known to be expressed within the gut of the developing parasite, where the acidic environment is crucial for enzyme activation (Timms & Bueding, 1959; Koehler et al., 2007; Figueiredo et al., 2015). In common with other aspartyl proteases, the propeptide of SmCD1 is presumed to occupy the active site of the enzyme and is cleaved upon activation at acidic pH, permitting the access of substrates to the catalytic aspartyl residues. There is evidence that SmCD1 exists as a dimer, which shifts to a monomeric state upon activation at acidic pH (Araujo-Montoya et al., 2020). The functional implications of this reversible interaction have yet to be determined, and the dimerization interface remains to be definitively identified.

Despite the potential of SmCD1 as a therapeutic target, structural characterization of the enzyme is lacking and there are limited studies utilizing SmCD1 to formulate novel agents (Dougall et al., 2014; Gomes et al., 2023; Balogun et al., 2024). Here, we report crystal structures of SmCD1 alone and in complex with a high-affinity nanobody, which stabilizes the propeptide of the enzyme, providing insight into the conformation of the SmCD1 zymogen.

2. Materials and methods

2.1. Production of recombinant SmCD1 and human homologues

The genes encoding SmCD1_15–385 (UniProt G4VEV6), human cathepsin D_211–412 (UniProt P07339) and human cathepsin E_20–398 (UniProt P14091) were inserted into the vector pOPINTTGneo by In-Fusion cloning (Berrow et al., 2007), which adds an N-terminal mu-phosphatase signal sequence and a C-terminal 6×His tag (Nettleship et al., 2009). Recombinant cathepsins were expressed in Expi293 cells (Le Bas et al., 2022) and purified by tandem affinity size-exclusion chromatography with HisTrapFF (Cytiva) and HiLoad 16/600 Superdex 200 columns (Cytiva) on an ÄKTAxpress HPLC system (Nettleship et al., 2009). Eluted fractions were analysed by reducing SDS–PAGE, pooled and mass-validated by mass spectrometry. For crystallization, SmCD1 was expressed in the presence of kifunensine (1 µg per millilitre of culture) to inhibit complex glycosylation (Chang et al., 2007).

The activity of purified SmCD1 was assayed as described previously (Araujo-Montoya et al., 2020) using the FRET substrate [7-methoxycoumarin-4-acetyl-GKPILFFRLK(DNP)-D-R-amide]. Briefly, enzyme (1 µM–100 nM in 20 mM sodium acetate, 150 mM NaCl buffer at pH 3.0–7.5) was incubated with the substrate (1 µM) in a final reaction volume of 100 µl. Immediately after adding substrate, relative fluorescence units (RFU) at an excitation of 320 nm and emission of 420 nm were measured with a CLARIOStarPlus Plate Reader (BMG Labtech), recording every 30 s for 2 h. For inhibition studies, 100 nM SmCD1 was incubated with a twofold excess of Nb10C9. RFU data were analysed in GraphPad Prism (v.10.3.1) and linear regression was used to calculate the slope of fluorescence over time.

2.2. Generation of nanobodies to SmCD1

Nanobodies to SmCD1 were generated as described by Eyssen et al. (2024), adapted for solid-phase panning as outlined by Zimmermann et al. (2020). One llama was immunized in accordance with CEDAR policies and adhering to Animal Research Reporting of In Vivo Experiments (ARRIVE) principles. Blood sampling was conducted following the UK Government Animal (Scientific Procedures) Act 1986 under UK Home Office project licence PA1FB163A. For immunization, 200 µl of 1 mg ml⁻¹ purified SmCD1 was administered on day 0, followed by two boosts of 200 µg antigen per immunization on days 28 and 56. Ten days following the final boost, a 170 ml blood draw was taken. A seroconversion ELISA confirmed the camelid immune response to SmCD1 (Supplementary Section S1.1). Peripheral blood mononuclear cells (PBMCs) were isolated from blood samples by density-gradient centrifugation. RNA was extracted from PBMCs and used as a template for rtPCR to amplify sequences corresponding to the variable domain of heavy-chain-only antibodies (VHH or nanobodies). Following PCR, nanobodies were cloned into the phagemid-display vector pADL23c to prepare a phagemid-display library. Phage-displaying nanobodies specific for SmCD1 were enriched by two rounds of bio-panning with 50 and 10 nM SmCD1, respectively, confirmed by polyclonal ELISA (Supplementary Section S1.1). 93 individual phage clones were picked from each round of panning; nanobody-displaying phages were recovered by infection with M13K07 helper phage and tested for binding to SmCD1 by anti-M13 ELISA (Supplementary Section S1.1). Positive phage binders were sequenced and analysed for CDR3 sequence identity using IMGT/V-QUEST (Giudicelli et al., 2011). Nanobodies were subcloned into the pOPINVHH_his vector (Eyssen et al., 2024) by In-Fusion cloning (Berrow et al., 2007) for the production of proteins for biophysical and structural studies.

2.3. Production of nanobodies and complex with SmCD1

Nanobodies were first expressed on a small scale using WK6 cells in 4 ml Terrific Broth medium supplemented with 100 µg ml⁻¹ ampicillin. Cell pellets were resuspended in 300 µl 1 mg ml⁻¹ polymyxin B sulfate (Sigma–Aldrich) to isolate the periplasm. Nanobodies were purified using Ni–NTA spin columns (Qiagen) and desalted with Zeba desalting columns (Thermo Scientific). Purified nanobodies were analysed on SDS–PAGE and binding was evaluated by titration ELISA (Section 2.4).

To obtain a higher yield for further characterization, nanobodies were expressed on a large scale using WK6 cells in 1 l Terrific Broth medium supplemented with 100 µg ml⁻¹ ampicillin, 0.1%(v/v) glucose and 2 mM MgCl₂. The periplasm was isolated by osmotic shock with TES buffer, and nanobodies were purified by tandem affinity size-exclusion chromatography with HisTrapFF (Cytiva) and HiLoad 16/600 Superdex 75 columns (Cytiva) on an ÄKTAxpress HPLC system as per established protocols (Nettleship et al., 2009; Le Bas et al., 2022). Eluted fractions were analysed by reducing SDS–PAGE, pooled and validated for size by mass spectrometry.

For crystallization trials, SmCD1 was mixed with a 1.2 molar excess of Nb10C9 for 3 h at 4°C and agitated at 25 rev min⁻¹. Following this, Endo F1 was added at 0.05 molar equivalents relative to the total protein and incubated at 4°C overnight to remove high-mannose glycoforms. The sample containing SmCD1, Nb10C9 and Endo F1 was first manually passed over a GSTrap glutathione S-transferase affinity column (Cytiva) equilibrated with 20 mM reduced glutathione diluted in SEC buffer (20 mM Tris, 300 mM NaCl pH 7.5) to remove Endo F1 and cleaved oligosaccharides. The SmCD1–Nb10C9 complex was co-eluted using a Superdex 200 10/300 GL Increase column (Cytiva) equilibrated in 20 mM Tris, 300 mM NaCl pH 7.5 on an ÄKTApure HPLC (Le Bas et al., 2022). Fractions were assessed for the presence of the complex by reducing SDS–PAGE and those of interest were spin-concentrated with a Vivaspin 20 10 000 MWCO column (Sartorius). Co-purified complex was stored at 4°C and used for crystallization screening on the same day as purification.

2.4. ELISA and biolayer interferometry

Small-scale purified nanobodies were analysed for binding to SmCD1 by α-VHH titration ELISA as described in Eyssen et al. (2024), using wells coated overnight at 4°C with 50 nM SmCD1. Binding was assessed using 25 µg ml⁻¹ as the highest concentration of nanobody with a twofold dilution series. Following successful large-scale purification, a further α-VHH titration ELISA was performed with Nb10C9, starting with a concentration of 20 nM. To assess potential cross-reactivity to human cathepsin D (hCD) and human cathepsin E (hCE), ELISA wells were coated overnight at 4°C with 50 nM antigen diluted in PBS. Nb10C9 was titrated from 1 µM in a series of tenfold dilutions. For evaluation of nanobody pH-sensitivity, ELISAs were adapted as per Supplementary Section S1.2.

For all ELISAs, wells were washed thrice between incubations with 300 µl PBST. 0.1 µg ml⁻¹ Monorab Rabbit Anti-Camelid VHH-HRP monoclonal antibody [GenScript, catalogue No. A01861; diluted in 0.1%(w/v) BSA–PBS] was added to detect bound nanobody. ABTS peroxidase substrate (LGC SeraCare) was added for antibody capture and incubated in the dark for 15 min. Absorbance was measured as microplate endpoint at 405 nm using a CLARIOStarPlus plate reader (BMG Labtech). ELISA data were analysed using GraphPad Prism (v.10.3.1).

Biolayer interferometry (BLI) was used to estimate a dissociation constant (K_d) as a measure of nanobody affinity. All experiments were designed using Octet BLI Discovery Software (v.13.0) and were conducted on an Octet R8 system (Sartorius). Assays were performed at 30°C using 0.1%(w/v) BSA–PBST as the assay buffer. Biosensors were first washed in buffer-only wells for 600 s to generate a baseline. SmCD1 was biotinylated using a One-Step Antibody Biotinylation Kit (Miltenyi Biotec) and free biotin was removed using Zeba Biotin Removal Columns (Thermo Scientific). 50 nM biotinylated SmCD1 was immobilized on streptavidin biosensors during a 300 s loading step. After a 60 s wash and 30 s re-equilibration into buffer, sensors were incubated with wells containing nanobody titrated from 200 nM in a series of twofold dilutions. Association was measured for 600 s and dissociation was measured in buffer-only wells for a further 600 s. Data points were collected every 0.2 s. Kinetic characterization was completed in Octet Analysis Studio (v.13.0). Graphs were generated in GraphPad Prism, with data represented as nanometre shift plotted against time. Replicate BLI measurements were combined.

2.5. Crystallization and structure determination

For crystallization of the SmCD1 enzyme alone, droplets consisting of 100 nl protein solution at 20 mg ml⁻¹ and 100 nl reservoir solution were dispensed into MiTeGen In Situ (2-drop) plates using a Mosquito liquid-handling robot (TTP Labtech). Crystals were obtained using sitting-droplet vapour diffusion at 20°C (293 K) in the presence of 0.1 M bis-Tris propane pH 8.5, 0.02 M sodium/potassium phosphate, 20%(v/v) PEG 3350. For the SmCD1–Nb10C9 complex, crystal screening was also carried out by sitting-droplet vapour diffusion. Droplets consisting of 50 nl protein complex at 10 mg ml⁻¹ and 100 nl reservoir solution were dispensed as described previously. Crystals were obtained at 20°C (293 K) in the presence of 0.1 M PCTP pH 6.00, 0.2 M CaCl₂, 20%(v/v) PEG 3350. X-ray diffraction (XRD) data for all structures were collected in situ to 1.95 Å resolution at 293 K using a Dectris EIGER2 X 4M detector at the VMXi instrument, Diamond Light Source (Sanchez-Weatherby et al., 2019; Mikolajek et al., 2023).

The data set for SmCD1 alone was produced from a single crystal using xia2/DIALS (Winter, 2010; Winter et al., 2022; Gildea et al., 2022), as implemented in the Diamond Light Source SynchWeb interface ISPyB (2025-R.4). The data set was extended to 2.20 Å resolution, yielding a corrected R factor of 0.292 overall and 0.094 for the inner shell. All shells had 100% completeness, and CC_1/2 for the outer shell was 0.311 (Table 1).

Table 1
X-ray crystallographic data-collection and refinement statistics

Values in parentheses are for the outer shell.

	SmCD1 (PDB entry 9snt)	SmCD1–Nb10C9 (PDB entry 9rxi)
Data collection
Space group	P62	P1211
a, b, c (Å)	136.62, 136.62, 51.80	70.14, 181.80, 78.85
α, β, γ (°)	90.00, 90.00, 120.00	90.00, 92.41, 90.00
Temperature (K)	293	293
Resolution range (Å)	44.72–2.20 (2.28–2.20)	90.90–2.59 (2.63–2.59)
Rmerge	0.292 (2.768)	0.671 (7.642)
Rmeas	0.314 (2.977)	0.699 (7.966)
Rp.i.m.	0.114 (1.080)	0.194 (2.202)
〈I/σ(I)〉	8.5 (1.1)	6.7 (1.0)
CC1/2	0.990 (0.311)	0.984 (0.262)
Completeness (%)	99.94 (100.00)	99.73 (100.00)
Multiplicity	7.4	12.1
Refinement
Resolution range (Å)	44.72 (2.20)	90.90 (2.59)
No. of reflections	28306	86279
Rwork/Rfree	0.175/0.200	0.173/0.234
No. of atoms
Protein	A, 2795	A, 5604; B, 5640: C, 5489; D, 5485; E, 1909; H, 1916
Ions/buffer	0	Ca, 2; Na, 2
Water	119	53
Total	2914	26100
Residual B factors (Å^2)
Protein	A, 40.14	A, 45.3; B, 46.1; C, 52.3; D, 63.4; E, 43.1; H, 40.6
Ligand/ion	0	Ca, 71.1, 51.7; Na, 79.7, 60.7
Water	42.93	34.8
R.m.s. deviations
Bond lengths (Å)	0.008	0.0067
Bond angles (°)	0.95	1.68
Ramachandran plot
Most favoured (%)	98.03	92.35
Clashscore	4.29	6

The final data set for SmCD1–Nb10C9 was produced by multiplexing 18 data sets using xia2/DIALS (Winter, 2010; Winter et al., 2022; Gildea et al., 2022), again implemented in the Diamond Light Source SynchWeb interface ISPyB (2025-R.4). The multiplexed enzyme data set was extended to 2.59 Å resolution, yielding a corrected R factor of 0.671 overall and 0.165 for the inner shell. All shells had 100% completeness, and the CC_1/2 for the outer shell was 0.262 (Table 1).

All diffraction data were phased and the atomic model was built and refined using the Phenix suite (Liebschner et al., 2019) or the CCP4 suite via the CCP4i2 interface (CCP4 v.9.0, CCP4i2 v.1.1.0; Winn et al., 2011; Agirre et al., 2023) for the enzyme and the enzyme–nanobody complex, respectively.

Phasing was performed by molecular replacement using a search structure derived from the experimentally solved structure of human cathepsin D (PDB entry 1lya; Baldwin et al., 1993) using Phaser-MR v.2.7.0 (Read, 2001). Density modification and phase improvement produced interpretable electron-density maps, which were used for automated model building using Phenix AutoBuild. The resulting model was manually refined through iterative cycles of model adjustment in Coot (Emsley et al., 2010) and refinement in phenix.refine (Afonine et al., 2012). Model quality was assessed using MolProbity (Chen et al., 2010) and R-factor statistics (R, 0.17; R_free, 0.20), and the final structure was validated against the experimental data. Coordinates and structure factors were deposited in the Protein Data Bank with code 9snt. All data were included throughout refinement, and consequently the resolution of the refined structure was 2.20 Å (Table 1).

The refined SmCD1 structure and a predictive model of Nb10C9 generated with NanoBodyBuilder2 (Dunbar et al., 2016; Abanades et al., 2023) were used as initial search models with Phaser-MR v.2.7.0 (Read, 2001) for structural determination of the SmCD1–Nb10C9 complex. The positioned model from Phaser-MR was iteratively rebuilt and refined using WinCoot v.9.8.95 (Emsley et al., 2010) and REFMAC5 v.5.5 (Vagin et al., 2004; Murshudov et al., 2011). All data were included throughout refinement, and consequently the resolution of the refined structure was 2.59 Å (Table 1).

Following structural refinement, PDBePISA was used for the estimation of intermolecular interactions (Krissinel & Henrick, 2007), identifying two potential binding epitopes for Nb10C9 on SmCD1. Identical interactions between Nb10C9 and SmCD1 were predicted using the Protein–Ligand Interaction Profiler (PLIP; Adasme et al., 2021). Graphical representations were generated using PyMOL (v.3.0.4; Schrödinger) and LigPlot+ v.2.2.9 (Laskowski & Swindells, 2011).

2.6. Generation of Nb10C9 mutants

Knockout constructs were designed with the proposed interacting residues of the nanobody mutated to alanines. Nb10C9_KO1 was mutated at Gln1, Gln3, Tyr95, Leu122, Thr124 and Gln125. Nb10C9_KO2 was mutated at Gly26, Arg27, Thr28, Tyr32, Asn77, Tyr100 and Glu103. The gBlock templates were purchased from IDT (Integrated Technologies Inc.) and cloned into the pOPINVHH_his vector by In-Fusion cloning (Berrow et al., 2007).

Mutants were expressed on a large scale using WK6 cells in 1 l Terrific Broth medium as described previously (Section 2.3). To assess binding, ELISA wells were coated with 50 nM SmCD1 at 4°C. Mutant nanobodies were titrated, alongside wild-type Nb10C9, from 100 µM in a series of tenfold dilutions. For the detection of bound nanobody, 0.1 µg ml⁻¹ anti-camelid VHH-HRP [GenScript; diluted in 0.1%(w/v) BSA–PBS] was added. ABTS peroxidase substrate (LGC SeraCare) was added for antibody capture and incubated in the dark for 15 min. Absorbance was measured as microplate endpoint at 405 nm using a CLARIOStarPlus plate reader (BMG Labtech). ELISA data were analysed using GraphPad Prism (v.10.3.1).

3. Results

3.1. Structure of SmCD1

Recombinant SmCD1, lacking the intrinsically disordered carboxy-terminal extension that appears to be unique to the S. mansoni enzyme (Wong et al., 1997), was produced by transient expression in Expi293 cells in the presence of kifunensine. Purified SmCD1 was treated with EndoF1 to remove high-mannose sugars and the protein crystallized at pH 8.5. The structure of SmCD1 was determined by molecular replacement using a search structure derived from the experimentally solved structure of human cathepsin D (PDB entry 1lya) to a resolution of 2.20 Å (Table 1). A single copy of SmCD1 was placed within the asymmetric unit, consisting of 14 α-helices and 25 β-sheets which form two domains stabilized by three disulfide bonds and typical of aspartyl proteases, with each domain contributing an aspartate residue to the active site (Asp84 and Asp270). Connective density was observed for most of the propeptide sequence (Val15–Leu45) which forms the sixth β-strand in the inter-domain β-sheet. However, the electron-density map was insufficient to distinguish the remainder of the propeptide-including region, where it is cleaved on enzyme activation (Ser46–Pro53; Supplementary Fig. S1). The S2 subsite residues (Met336 and Gly337) of the protease (Silva et al., 2002) are also not resolved in the crystal structure. The catalytic aspartic acids and active-site flap of SmCD1 were positioned in equivalent locations as in human cathepsin D (PDB entry 1lyb; Baldwin et al., 1993), exhibiting an average r.m.s.d. of 0.72 Å. In addition, homology is observed across the cathepsin gene family, with human cathepsin E sharing 46.7% identity with SmCD1 (63.6% similarity; Supplementary Fig. S2). Notably, human cathepsin E (PDB entry 1tzs; Ostermann et al., 2004) also shows a markedly similar overall structure to SmCD1, with an average r.m.s.d. of 0.87 Å over all residues, with a short stretch of the propeptide also present in the structure.

Overall, SmCD1 was observed in an inactive conformation, as expected from the basic pH of the crystallization conditions, with the propeptide occluding the active-site cleft and the N-terminal strand (residues Gln54–Tyr67) inserted into the catalytic site, with Lys59 forming salt bridges with the two active-site aspartates (Fig. 1).

Figure 1 Structure of SmCD1. (a) Overall structure of the SmCD1 backbone with the propeptide sequence shown in maroon, disulfide bridges in green and catalytic aspartates shown in stick representation coloured by element. (b) Salt-bridge interactions between Asp84, Asp270 and Lys59.

3.2. Identification and characterization of a nanobody to SmCD1, Nb10C9

A total of 186 individual clones (93 from each round of panning) were randomly picked and assayed by ELISA, and the eight positive phage clones were sequenced. The results showed that all had unique CDR3 sequences and, of these, six clones were expressed on a small scale in Escherichia coli for preliminary binding studies. Titrations from 25 µg ml⁻¹ identified a lead candidate, Nb10C9, which retained binding to SmCD1 at the highest dilution (Fig. 2a). This nanobody was progressed to large-scale expression and successfully purified, yielding a solution at 22 mg l⁻¹. Further titration ELISA and BLI allowed an estimation of affinity (Fig. 2b). Nb10C9 exhibited a strong interaction with SmCD1, presenting a weighted average K_d of 2.67 nM (±3.31 × 10⁻³ nM), calculated from n = 3 technical replicates and fitted to a 1:1 binding model. ELISA showed that Nb10C9 did not bind to either human cathepsin D or E (Fig. 2c).

Figure 2 Characterization of nanobody binding. (a) Titration ELISAs using 50 nM SmCD1 immobilized in wells of a 96-well plate to assess the binding of purified nanobody, detected using anti-camelid VHH-HRP. Data shown represent single measurements from small-scale expressed material used for initial candidate prioritization. (b) Biolayer interferometry using streptavidin biosensors loaded with 50 nM biotinylated SmCD1 binding to Nb10C9. Association occurred during the first 600 s; samples were then transferred to a buffer-containing well to measure dissociation for a further 600 s. Values shown in the table are weighted averages calculated from n = 3 technical replicates. Error values represent standard error of the mean. (c) Titration ELISA of Nb10C9 bound to SmCD1, human cathepsin D (hCD) or human cathepsin E (hCE). (d) Amino-acid sequence of Nb10C9 with CDRs shown in bold (IMGT numbering; Giudicelli et al., 2011), with secondary structures of Nb10C9 depicted along the top of the sequence. Green digits indicate disulfide bridges.

SmCD1 is activated under acidic conditions, with optimal activity at pH 3–3.5, consistent with previous findings (Fig. 3a; Dunn, 2002; Araujo-Montoya et al., 2020). Therefore, the pH-sensitivity of nanobody binding was assayed to determine whether the nanobody could affect enzyme activity. Kinetic assays suggested there may be a small inhibitory effect of Nb10C9 binding on SmCD1 activity (Figs. 3b and 3c). However, this inhibition is limited by the pH-dependent binding of Nb10C9, which only shows titratable binding between pH 5.0 and 7.5 (Fig. 3c).

Figure 3 Activity of SmCD1. (a) SmCD1 activity measured over time under different pH conditions, recording the fluorescence of the fluorogenic substrate 7-methoxycoumarin-4-acetyl-GKPILFFRLK(DNP)-D-R-amide. (b) SmCD1 activity at pH 3.5, incubated with 200 nM Nb10C9 or pepstatin. Data were background-subtracted and are displayed as the mean of n = 3 repeats. (c) Estimation plot of unpaired Student's t-test for activity data of 50 nM SmCD1 or 50 nM SmCD1 + 200 nM Nb10C9 at the point of enzyme saturation, between 3000 and 4000 s. (d) Titration ELISA of Nb10C9 bound to 50 nM SmCD1 between pH 3.0 and pH 7.5. Data were background-subtracted and a line of best fit is displayed.

3.3. Structure of the SmCD1–Nb10C9 complex

Nanobody Nb10C9 and SmCD1 were mixed in a molar ratio of 1:1.2 (enzyme:nanobody), treated with EndoF1, co-purified (Supplementary Fig. S3) and crystallized at pH 6.0. The structure of the complex was solved using the experimentally determined SmCD1 structure and a model-built structure of Nb10C9. Matthews coefficient analysis placed four copies of the SmCD1–Nb10C9 ensemble within the asymmetric unit using Phaser-MR (Table 1). In the final model, four copies of SmCD1 and two copies of Nb10C9 were observed within the asymmetric unit (Supplementary Fig. S4). Two distinct assemblies are noted, an SmCD1 homodimer and an SmCD1–Nb10C9 complex, that replicate across symmetry operator 00−100−1. Regular solvent channels throughout the crystalline layering prohibit sufficient space for additional copies of Nb10C9, with no evidence of obvious missing electron density.

For all copies of SmCD1 (residues Val15–Pro385), connective density was observed for the entirety of the propeptide sequence and cleavage site (Ser49/Gly50) (Fig. 1a and Supplementary Fig. S1). Two of the four placed copies of SmCD1 were in complex with two Nb10C9 nanobodies and showed complete connective backbone density over residues 15–385, including residues 46–53 of the propeptide sequence, a tract that was not resolved in the SmCD1 structure (Fig. 1 and Supplementary Fig. S5). Otherwise, structural superimposition of SmCD1 alone and with bound nanobodies showed that the two structures were identical, with an average r.m.s.d. of 0.28 Å between nanobody-bound chains of the complex and SmCD1 alone (Supplementary Fig. S5). The unbound SmCD1 molecules did not show connective density in this region, indicating that nanobody binding has stabilized the conformation of the propeptide towards the carboxyl end of the sequence.

Intermolecular interactions between Nb10C9 and SmCD1 were identified using PDBePISA (Krissinel & Henrick, 2007). Nb10C9 recognizes residues positioned either side of the SmCD1 active-site cleft, namely Arg47, Val48, Ser49, Asp101, Ser133, Gln160 and Gly163, described here as the `CDR interface' (Figs. 4a and 4b). The critical interaction is a salt bridge between Glu103 of the Nb10C9 CDR3 and Arg47 of SmCD1. Glu103 is stabilized by a hydrogen bond to the Gly163 amide, whilst Tyr100, also belonging to the Nb10C9 CDR3, hydrogen-bonds to Asp101 via side-chain contacts. Further hydrogen bonds are provided by the Nb10C9 CDR1. Backbone hydrogen bonds exist between residues Gly26 and Arg27 with Ser49 of SmCD1 and between Thr28 and the Ser133 hydroxyl, and hydrogen bonds between the side chains of Tyr32 and Gln160. Finally, the carboxamide of framework-region Asn77 hydrogen-bonds to the carbonyl of Val48. A second binding site of Nb10C9 was also predicted by PDBePISA (Krissinel & Henrick, 2007) and by PLIP (Adasme et al., 2021), suggesting an interaction between the framework regions of Nb10C9 and SmCD1 (the `FR interface'). This secondary interaction primarily involves a salt bridge between Gln1 of Nb10C9 FR1 and Asp239 of SmCD1 (Fig. 4c). Gln1 may be stabilized by a hydrogen-bond network involving Lys374 and Gln3, although the intermolecular distances captured within the crystal structure are strained at 3.75 and 3.87 Å, respectively. Gln196 also participates in the FR interface, forming hydrogen bonds to the phenolic hydroxyl of Tyr95 of FR3, adjacent to CDR3, as well as to Leu122, Thr124 and Gln125 of FR4. The secondary interaction between SmCD1 and Nb10C9, via the FR interface, comprises only a handful of interactions which appear strained in the crystal structure and may be promoted by van der Waals forces.

Figure 4 Molecular interactions between SmCD1 and Nb10C9. (a) Molecular interactions between the backbone of SmCD1 (light yellow) and Nb10C9 (turquoise), with the backbone of the propeptide sequence coloured orange. PDBePISA-identified interacting groups are shown in stick notation. Salt bridges are indicated by red dashed lines and hydrogen bonds by blue dashed lines. (b) LigPlot+ schematic (Laskowski & Swindells, 2011) of Nb10C9 (upper subpanel) and SmCD1 (lower subpanel) residues participating in the CDR interface. Identified interactions are represented by dashed lines, whilst residues present at the interface are represented as surfaces. (c) Binding of SmCD1 to Nb10C9 via the FR interface. Molecular interactions between the backbone of SmCD1 (light yellow) and Nb10C9 (turquoise) are shown. Residues involved in the FR interface are coloured teal. PDBePISA-identified interacting groups are shown in stick notation. Salt bridges are indicated by red dashed lines and hydrogen bonds by blue dashed lines. Bond distances are represented in Å. (d) Titration ELISA of Nb10C9 mutants, Nb10C9_KO1 and Nb10C9_KO2, bound to 50 nM SmCD1.

Side-directed mutagenesis of Nb10C9 was carried out to test the importance of the interactions between Nb10C9 and SmCD1 observed in the crystal structure for nanobody binding. Two Nb10C9 mutants were designed: Nb10C9_KO1, with the identified FR interface residues (Gln1, Gln3, Tyr95, Leu122, Thr124, Gln125) substituted by alanines, and Nb10C9_KO2, with alanine substitutions made for CDR interface residues (Gly26, Arg27, Thr28, Tyr32, Asn77, Tyr100, Gln103). Mutated nanobodies were expressed and purified and their binding activity was tested by ELISA compared with the parent nanobody (Fig. 4d). The results showed that Nb10C9_KO2 had reduced Nb10C9 binding, which was rationalized by the mutation of residues interacting with the propeptide, and confirmed the biological relevance of the CDR interaction. Interestingly, a low-affinity interaction persists at concentrations above 1 µM. This could be attributed to the FR interaction, taking place with a low free energy of dissociation, since these residues are intact in the Nb10C9_KO2 mutant.

4. Discussion

We report the first crystal structure of schistosome cathepsin D1 as the inactive zymogen with the propeptide intact and covering the active site. A contiguous sequence for the propeptide was only observed in the structure of the enzyme in complex with a nanobody selected from a phage-display library following immunization of a llama with purified SmCD1. This suggests that binding to the nanobody stabilizes the conformation of the propeptide. By analogy to human cathepsin D (Lee et al., 1998), activation of the enzyme at acidic pH must involve a major conformational change such that following release of the propeptide the amino-terminal strand is rearranged to replace the β-strand of the propeptide in the six-stranded β-sheet between the two domains of the enzyme. In this way, the active site becomes accessible.

Comparison of the structures of SmCD1 alone and in complex with Nb10C9 showed that the nanobody binds to the sequence Arg47–Ser49 immediately adjacent to the propeptide cleavage site between Ser49 and Gly50 (Brindley et al., 2001). This sequence is identical in cathepsin D from S. bovis, and approximately 70% similar to the enzyme from S. japonicum, but differs from both human cathepsins D and E (Supplementary Fig. S2), consistent with the ELISA results (Fig. 2c). Therefore, Nb10C9 could be used as a schistosome-specific antibody for detecting SmCD1 in cellulo, enabling localization studies within a native context.

Helminth infection, and the resulting schistosomiasis, remains difficult to treat and diagnose. Here, we have identified an epitope in SmCD1 as a schistosome-specific biomarker and determined the structure of the SmCD1 zymogen. This may guide future structure-based drug design for the development of novel therapeutics against schistosomiasis, a World Health Organization-recognized neglected tropical disease.

5. Related literature

The following references are cited in the supporting information for this article: Esparza et al. (2023), Robert & Gouet (2014) and Sievers & Higgins (2014, 2021).