Mutation of the H-bond acceptor S119 in the ADAMTS13 metalloprotease domain reduces secretion and substrate turnover in a patient with congenital thrombotic thrombocytopenic purpura

Thrombotic thrombocytopenic purpura (TTP) is an acute thrombotic disorder characterized by consumptive thrombocytopenia and microangiopathic hemolytic anemia.¹  In most cases TTP is acquired, but a small number of cases bears mutations in the adisintegrin and metalloprotease with thrombospondin motifs (ADAMTS13) gene, resulting in the same phenotype.^2,3 

ADAMTS13 is a metalloprotease specifically hydrolyzing the Y1605-M1606 peptide bond in multimeric von Willebrand factor (VWF). VWF multimer size reduction is pivotal in preserving normal hemostasis because uncleaved multimers spontaneously bind platelets causing occlusion of microvasculature, the TTP hallmark.

More than 50 ADAMTS13 gene mutations have been reported since its identification in 2001.⁴  Most of these were not investigated by in vitro mutagenesis and expression analysis, but those that were cause mainly secretion defects.³  This is in line with the observation that ADAMTS13 antigen is absent in plasma samples from patients with hereditary TTP.⁵  Nonetheless, some ADAMTS13 mutants are secreted^3,6,7  but are less or not functional.

Recently, high-resolution structures of the metalloprotease domains of ADAMTS1, ADAMTS4, and ADAMTS5 were published, providing insight in active-site and small molecule inhibitor orientation.^8,9  The structures of the metalloprotease domains of ADAM/ADAMTS proteins generally superimpose with little variation except in the variable loops making up domain periphery.¹⁰ 

In this study, we characterize a recently identified ADAMTS13 metalloprotease domain missense mutation that causes a partial secretion defect and a 3-fold reduction in the turnover number (k_cat). We correlate kinetic data with rational assumptions from a homology model of the ADAMTS13 metalloprotease domain.

This study was approved by the Katholieke Universiteit Leuven Ethical Committee of the Katholieke Universiteit Leuven and its hospitals. Informed consent was obtained in accordance with the Declaration of Helsinki. The propositus is a 14-year-old white pediatric patient with typical congenital TTP successfully treated with prophylactic plasma infusion therapy. ADAMTS13 exons with exon-intron boundaries were amplified from genomic DNA using polymerase chain reaction with primers described elsewhere.¹¹  Sequencing reactions were outsourced (GATC Biotech AG).

Citrated plasma was used for analysis of ADAMTS13 activity and antigen. Normal human pooled plasma (NHP, n = 20) was used as a control. Antigen was measured by enzyme-linked immunosorbent assay (ELISA) as described⁵  and by immunoblotting with monoclonal antibody (mAb) 17B10 binding to the third ADAMTS13 thrombospondin-1 repeat. Bound mAb was detected by horseradish peroxidase labeled goat anti–mouse antiserum (GAM-HRP; Jackson Immunoresearch).

Mutagenesis was performed with the QuikChange XL Site-Directed Mutagenesis kit (Stratagene) using primers bearing the desired mutations. All mutants were constructed using the inducible expression plasmid pGene/V5-His (Invitrogen) containing the ADAMTS13 cDNA.⁵  Vectors were lipofected in Chinese hamster ovary GeneSwitch cells and human embryonic kidney 293 (HEK293) GeneSwitch cells (Invitrogen) for stable and transient expression, respectively. Serum-free conditioned medium was purified by ion exchange (Q-sepharose; GE Healthcare).¹² 

Activity toward multimeric VWF was measured using mild denaturants as described previously.¹³  Reactions contained 10 μg/mL purified VWF (CSL Behring) and 0.6nM recombinant ADAMTS13 (rADAMTS13) and were incubated for 16 hours. Proteolysis was analyzed by residual collagen binding.¹⁴  Proteolytic fragments were visualized by Western blotting with anti-VWF monoclonal antibodies RG46, 52K-2, and 52K-8 (Dr Zaverio M. Ruggeri, Scripps Research Institute, La Jolla, CA)¹⁵  and GAM-HRP.

Determination of kinetic parameters k_cat and Michaelis constant (K_M) using the FRETS-VWF73¹⁶  substrate and 0.1nM enzyme was as described previously.¹⁷ 

MODELLER 9v4¹⁸  was used to generate a homology model based on the structures of ADAMTS1, ADAMTS4, and ADAMTS5 (respective Protein Data Bank codes: 2V4B, 3B2Z, and 3B8Z).¹⁹  Sequence identity between ADAMTS13 and templates varied between 25% and 30%. Modeling preserved both the active-site zinc ion and the zinc-coordinating water molecule. The dead-end elimination method²⁰  was used to optimize orientation of all side chains and polar hydrogens. Idealization of bond geometry and removal of unfavorable nonbonded contacts was performed by a conjugate gradient energy minimization with the GROMACS Version 3.3.1 software package²¹  in the presence of explicit water and counter ions. We applied the GROMOS96 43a1 force field. Mutant structure S119F was generated using the BRUGEL package²²  and least squares fitting was done using PyMOL (DeLano Scientific).

ADAMTS13 antigen was severely reduced to 5% in remission phase and undetectable (< 5%) in acute phase plasma as determined by Western blot (Figure 1A left) and ELISA. Both samples contained no measurable VWF73 cleaving activity (not shown). A homozygous missense mutation 356C>T and a known single nucleotide polymorphism 1342C>G were identified, encoding a serine to phenylalanine substitution (S119F) and a glutamine to glutamic acid exchange (Q448E), respectively. Meyer et al recently reported S119F as the first congenital TTP case on the African continent.²³  Importantly, the serine residue is conserved among ADAMTSs except in ADAMTS16 and ADAMTS18²⁴  where a threonine resides at the corresponding position, alluding to a role of the side-chain hydroxyl group.

Transient expression of both the single (S119F) and the double (S119F/Q448E) mutant revealed normal expression but markedly impaired secretion (Figure 1A right). VWF multimers were completely cleaved by 0.6 nM of ADAMTS13 in NHP, whereas an identical concentration of either single or double mutant only partially cleaved the substrate (Figure 1B). Western blotting of the reduced VWF N-terminal reaction products revealed no obvious size difference between VWF fragments generated by S119F and wild type (WT), indicating site specificity was unaltered (Figure 1C).

Importantly, k_cat of S119F (0.88 ± 0.04 s⁻¹, mean ± SD) was significantly smaller than WT (2.78 ± 0.11 s⁻¹), whereas K_M was only slightly smaller (1.4 ± 0.2μM vs 2.3 ± 0.3μM) as determined by FRETS-VWF73 kinetic measurements (Figure 2A). A reduced k_cat indicates that substrate turnover and/or product release is affected by the mutation. Furthermore, K_M is only slightly different, indicating ground-state binding is comparable with WT or somewhat better. As a consequence, the specificity constant (k_cat/K_M) reduces by half from 1.2μM⁻¹s⁻¹ (WT) to 0.6μM⁻¹s⁻¹ (S119F). This parameter is inversely related to the free energy of transition state complex formation.²⁵  Therefore, more energy is required to attain transition state in S119F following the enzyme substrate binding step. Hence, higher concentrations of S119F are needed for optimal VWF processing and TTP may precipitate at enzyme levels that would be safe if no mutation were present.

In our metalloprotease domain model (Figure 2B), S119 resides on a short α-helix juxtaposed to the C-terminus of the S1′ loop (residues 247-261; Figure 2C).^9,26  This loop lacks secondary structure and is important for substrate recognition.¹⁰  The side-chain hydroxyl group from S119 accepts an H-bond from the backbone amine of conserved residue W262. Furthermore, the W262 indole side chain is oriented in a hydrophobic pocket that is conserved among ADAMTSs and compiled of aliphatic residues from nearby strands including the short α-helix harboring S119 (Figure 2C).^8,9  Both the H-bond and the optimal orientation of the short α-helix are hence affected in S119F. To distinguish between these mechanisms underlying the abnormal kinetics, a mutant S119A was constructed. This mutant prevents H-bond formation but exerts no steric strain on the helix bearing S119. Interestingly, kinetic parameters were not different between S119A and S119F (k_cat = 0.82 ± 0.03 s⁻¹ and K_M = 1.1 ± 0.1μM). This indicates that steric interference in S119F is not the leading cause of decreased catalytic efficiency. The data show that the S119-W262 H-bond is crucial for optimal turnover (reduced k_cat), although not required for substrate recognition or binding (equal K_M). Alternatively, disruption of the H-bond induces structural changes in the active-site shape that would concomitantly explain the poor secretion. In conclusion, as the C- and N-terminus of the S1′ loop meet, it adopts an Ω shape, which is in part sustained by the hydrophobic interactions including C-terminal W262 keeping the ends of the flexible strand together in the pocket. Loss of the W262 H-bond with S119 might render the backbone of the S1′ loop more flexible, influencing overall structure and/or optimal alignment for transition state complex formation.

We thank Dr Zaverio M. Ruggeri (Scripps Research Institute) for monoclonal antibodies against VWF.

H.B.F. and K.V. are fellows and C.V.G. is clinical investigator of the Research-Foundation Flanders (Fonds voor Wetenschappelijk Onderzoek-Vlaanderen [FWO]). R.V. is a graduate student of the FWO. This work is supported by grants from the FWO (G.0299.06) and the Katholieke Universiteit Leuven (GOA/2004/09).

Contribution: H.B.F. designed and performed research, analyzed and interpreted data, and wrote the paper; I.P. performed research and interpreted data; R.V. performed research and critically reviewed the manuscript; M.D.M., H.D., and C.V.G. provided essential reagents and/or patient material and critically reviewed the manuscript; and K.V. designed research, analyzed and interpreted data, and wrote the paper.

Conflict-of-interest disclosure: The authors declare no competing financial interests.

Correspondence: Karen Vanhoorelbeke, Laboratory for Thrombosis Research, Katholieke Universiteit Leuven Campus Kortrijk, E Sabbelaan 53, 8500 Kortrijk, Belgium; e-mail: karen.vanhoorelbeke@kuleuven-kortrijk.be.