Present address: School of Biological and Biomedical Sciences, University of Durham, Durham, UK.
Interactions between mature Der p 1 and its free prodomain indicate membership of a new family of C1 peptidases
Article first published online: 4 OCT 2007
Volume 62, Issue 11, pages 1302–1309, November 2007
How to Cite
Zhang, J., Hamilton, J. M., Garrod, D. R. and Robinson, C. (2007), Interactions between mature Der p 1 and its free prodomain indicate membership of a new family of C1 peptidases. Allergy, 62: 1302–1309. doi: 10.1111/j.1398-9995.2007.01492.x
- Issue published online: 4 OCT 2007
- Article first published online: 4 OCT 2007
- Accepted for publication 13 June 2007
- cysteine protease;
- Der p 1;
- dust mite allergy;
- peptidase allergen
Background: Studies in vivo have shown that the cysteine peptidase activity of group 1 house dust mite allergens contributes to their allergenicity. These allergens are synthesized initially as proenzymes and removal of the propiece is necessary to unmask their proteolytic activity. In related C1 family cysteine peptidases of enzyme clan CA, liberated propieces continue to inhibit the mature peptidase as tight binding inhibitors. As it is not known whether mite peptidase allergens behave similarly, our objective was to investigate the effect of the Der p 1 propiece on the catalytic activity of Der p 1 and Der f 1.
Methods: Enzymatic activity of natural Der p 1 and Der f 1 was assessed using a specific substrate and the effect of the recombinant propiece on its enzyme kinetics defined. The integrity of the propiece during these interactions was studied functionally and by analysis of the reaction mixtures.
Results: Der p 1 propiece was a potent competitive inhibitor of Der p 1 and Der f 1. In contrast to other cysteine peptidase prodomains, which are cognate tight binding inhibitors, the Der p 1 propiece behaves as a substrate and is fully degraded during this interaction.
Conclusion: Mature Der p 1-prodomain interactions differ from other C1 family cysteine peptidases, suggesting that group 1 mite allergens are a new subgroup among C1 family cysteine peptidases. The rapid inactivation of Der p 1 prodomain is a newly identified mechanism that may contribute to the potency of this allergen.
Accumulating evidence suggests that the enzymatic activity of some allergens underpins the development of allergy to nonenzymatic bystander allergens (1–3). This evidence implies that a small cadre of enzymes from diverse sources make decisive contributions to the pathogenesis of allergy. The best studied of these enzymatic contributions is proteolysis. In the case of fungi and dust mites, proteolysis is a vital driver of transepithelial allergen delivery and creates a signalling environment that favours the development of Th2 immunity (4–17).
Group 1 house dust mite allergens (e.g. Der p 1, Der f 1 and Eur m 1) are cysteine peptidases of enzyme clan CA and are involved in the digestion of resistant structural proteins that form the staple diet of these mites (18, 19). All are initially synthesized as latent zymogens. Latency is maintained by a prodomain that folds into the binding cleft of the enzyme, thereby preventing its cleavage while the substrate binding sites are blocked. This propiece must be removed for the peptidase to mature and become activated. This step is crucial because it reveals both the peptidase activity of the allergen and, as demonstrated by various approaches, unmasks important epitopes (20–22). Without removal of the prodomain, proDer p 1 is significantly less immunogenic than mature Der p 1 (20–22). This reduced immunogenicity is likely to be due in large measure to the absence of enzymatic activity because mature, but enzymatically inactive, Der p 1 has a greatly reduced capacity to elicit sensitization to itself or nonpeptidase bystander allergens (1, 23–25).
In general, the removal of prodomains from peptidases does not unleash uncontrolled enzymatic activity. Typical exemplifications of this behaviour in Clan CA are the propieces from cathepsins, enzymes closely related to group 1 mite allergens. It is well established that cathepsin prodomains act as potent inhibitors even when their covalent attachment to the enzyme has been severed (26–30). Interactions between some of these cysteine peptidases and their tightly, but noncovalently, bound (‘untethered’) propieces show remarkable potency and selectivity (26, 27, 29), while others inhibit potently but less specifically (28–30). This continuity of inhibition by untethered propieces is thought to prevent unwanted proteolysis by enzymes with fundamental roles in the regulation of cell function.
An evolutionary tree for clan CA cysteine peptidases traditionally divides the C1 family into two subfamilies, the larger of which contains secreted and lysosomal enzymes of various origins, including Der p 1. Among this sub-family it is possible to discern two general types of propiece structure (31). In one, exemplified by cathepsin B, the propiece is relatively short (62 residues), whereas in the other type, exemplified by cathepsins, L, S and K, the propieces are longer (ca 100 residues). These longer propieces contain a conserved sequence of EX3RX2(I/V)FX2NX3IX3 N (the ERFNIN motif), which is involved in maintaining the globular structure of the propiece and its tight interaction with the catalytic cleft (32, 33).
There are grounds to believe that the propiece of Der p 1 and related allergens differ from both groups of currently recognized propieces. Mite allergen propieces are of intermediate length (80 residues) and lack the classically described ERFNIN motif, suggesting that their mode of binding to the peptidase may differ. The possible uniqueness of the mite allergens is strongly reinforced by the recent crystal structure reported for a mutant of Der p 1 expressed in its zymogen form (22). Der p 1 is the archetype of group 1 house dust mite allergens and so our purpose was to express the prodomain of Der p 1 and to use this as a model to characterize the potential inhibitory behaviour of mite group 1 prodomains generally. We sought to discover if the untethered prodomain is a regulator of enzymatic activity as anticipated from its similarity to well-studied enzymes, which exemplify the C1 family.
Expression and purification of Der p 1 prodomain
Two forms of the Der p 1 prodomain (PP-01 and PP-02) were produced from a D. pteronyssinus cDNA library. The underlying rationale was an initial desire to use an expression construct and vector that minimized potential problems because of proteolysis and toxicity during expression and, by incorporating a cleavable signal peptide, to have a preferred route to a soluble expressed protein. Having established the feasibility of expressing soluble PP-01, we then expressed the prodomain with an authentic N-terminus by engineering a flush-ended construct without a signal sequence. The resulting protein (PP-02) was then used as a comparator.
In PP-01, DNA encoding amino acids 18–98 of the entire Der p 1 sequence was amplified using 5′-CGGGATCCGGCTCGTCCATCA TCGATCAA-3′ and 5′-CCCTCGAGTTCAGCATTCAAATCGAATTGAGTT-3′ containing BamHI and AvaI restriction sites (bold text) in the sense and antisense primers respectively. PCR products were ligated into T-vector pCR2.1 TOPO for subsequent insertion into the pET22b+ plasmid and expression in Escherichia coli (OrigamiTM B(DE3) pLysS, Merck Chemicals Ltd, Nottingham, UK).
The second form of the prodomain (PP-02) was made using a construct with flush ends. Residues 19–98 of Der p 1 were amplified using 5′-GAGTCTC CATATGCGTCCATCATCGATCAAAACTTTTG-3′ and 5′-GGAATCTCATATGTTATTCAGC ATTCAAATCGAATTGAGT-3′ containing an NdeI restriction site (bold text) in both. PCR products were then cloned into pET22b+ plasmid and expressed in E. coli (BL21DE3, Merck, Germany).
Escherichia coli propiece constructs were inoculated into Luria-Bertani (LB) medium containing 100 μg/ml ampicillin and incubated at 37°C overnight. Aliquots were then inoculated (1 : 100 v : v) into fresh LB/ampicillin medium for 2.5–3 h at 37°C. Isopropyl β-d-thiogalactoside (IPTG) (Sigma-Aldrich, Poole, Dorset, UK) was added (1 mM) to induce expression and incubation continued at 30°C for a further 3.5–4 h. Negative controls were performed without IPTG. After centrifugation at 13 000 g and 4°C to collect cells, the pellet was washed once in phosphate buffer saline (PBS), and then frozen overnight at −20°C. Phosphate buffer saline (1–2 ml) was added to the thawed cells prior to extraction on ice (5s/pulse with 1 s pauses for 2 min) using a sonicator (Jencons PLS Ltd, Leighton Buzzard, UK). The supernatant was collected by centrifugation at 13 000 g and 4°C for 10 min, as PBS extracts.
Ammonium sulphate was added to the PBS extracts at 95% saturation and the resulting preparation purified using an Äkta Explorer system (GE Healthcare, Little Chalfont, Buckinghamshire, UK). The ammonium sulphate cut was applied to a Superdex G75 10/30 column and eluted with PBS. A peak eluting in a volume of 12–16 ml, was collected and re-chromatographed before being loaded on a Resource Q column (1 ml) in 10 mM Tris–HCl pH 8.0, and eluted by a gradient of 0–0.5 M NaCl. Final purification of PP-01 and PP-02 involved chromatography on Superdex G200 and desalting on a PD 10 column.
Der p 1 purification
House dust mites were cultured under barrier conditions with controlled temperature and relative humidity (25°C, 75%). Der p 1 was purified chromatographically from D. pteronyssinus culture medium enriched in faecal pellets. Purified Der p 1 was analysed by electrophoresis and mass spectrometrically by matrix-assisted laser desorption ionisation-time of flight mass spectrometry (MALDI-TOF). Purity was typically >95% and the protein was devoid of serine peptidase contamination as judged by complete active site titration with the selective Der p 1 inhibitor ADZ 50,000.
Synthesis of Der p 1 substrate and specific inhibitor
(3S,6S,9S,12S,15S,18S)-1-(2-aminophenyl)-9-butyl-18-carbamoyl-15- (4-hydroxy-3-nitrobenzyl)-12-(hydroxymethyl)-3-isopropyl-6-methyl-1,4,7,10,13,16-hexaoxo-2,5,8,11,14,17-hexaazaicosan-20-oic acid (ADZ 50,059) is a preferred substrate for Der p 1. ADZ 50,059 was synthesized by in-house chemistry using a conventional solid phase approach and successive rounds of protection, activation and coupling.
(S)-3-((S)-2-((S)-2-benzamido-3-methylbutanamido)propanamido)-2-oxoheptyl-2,6-bis(trifluoromethyl)benzoate (ADZ 50,000) is a potent inhibitor of Der p 1 (34) and was synthesized by in-house chemistry. In brief, Bz-Val-Ala-Nle-OH was prepared on trityl resin and converted to the corresponding bromomethylketone by reaction with isobutylchloroformate and N-methylmorpholine under argon. This was followed by treatment with ethereal diazomethane and reaction with hydrogen bromide in acetic acid. The resulting bromomethylketone derivative was reacted with potassium fluoride and 2,6-bis(trifluoromethyl)benzoic acid to yield the desired product as confirmed by electrospray mass spectrometry and proton magnetic resonance analyses.
Synthetic peptides from the Der p 1 prodomain sequence were prepared by conventional solid phase synthesis. The compounds were: peptide 1 (LMSAE); peptide 2 (NRFLMSAEAFE); peptide 3 (Ac-NRFLMSAEAFE-NH2) and peptide 4 (Ac-GSNGGAIN-NH2). Compounds were named using chemoffice ultra v. 9.0 (CambridgeSoft, Cambridge, MA, USA).
Mass spectrometry and electrophoresis
Tricine sodium dodecylsulphate-polyacrylamide gel electrophoresis (SDS-Tricine PAGE) was performed according to the method of Schagger and von Jagow (35). Matrix-assisted laser desorption ionisation-time of flight mass spectrometry used a Kratos Axima instrument in positive reflectron mode at an accelerating voltage of 100 kV. For analysis, 1 μl of solution was applied to a MALDI target and mixed with 1 μl of matrix solution (composition: α-cyano-4-hydroxycinnamic acid in 700 μl of acetonitrile and 300 μl of 0.1% trifluoroacetic acid).
Der p 1 enzyme activity assay
Reaction mixtures were assembled in 96-well plate format using a Perkin Elmer Multiprobe II Plus HTS EX robot and comprised 10 μl of substrate (12.5 μM final concentration), 70 μl of reaction buffer (potassium phosphate buffer pH 8.25 containing 1 mM EDTA), and 10 μl of dithiothreitol (DTT, 1 mM final concentration). Reactions were initiated by adding 10 μl of Der p 1 dissolved in reaction buffer at 2.5 μg/ml. The plate was immediately transferred to a fluorescence plate reader (Perkin Elmer Fusion Alpha-FP, Seer Green, UK) equipped with a temperature-controlled carrier set at 30°C and the reaction followed by excitation/emission at 330/420 nm. For inhibition assays, reaction buffer volume was 60 μl and inhibitors were added in 10 μl aliquots. Assays for Der f 1 (Indoor Biotechnologies Ltd, Warminster, UK) were performed similarly.
Assays for cathepsin L and cathepsin S
Reaction mixtures comprised 10 μl of Z-Phe-Arg-7-Amino-4-methylcoumarin substrate (Sigma-Aldrich, 10 μM final concentration in the cathepsin L assay, 20 μM in the cathepsin S assay), 70 μl of reaction buffer (400 mM of sodium acetate buffer, pH5.5 containing 4 mM EDTA for cathepsin L and 0.1 M sodium phosphate buffer, pH7.5 containing 2 mM EDTA for cathepsin S), and 10 μl of DTT (8 and 2 mM final concentrations respectively). Reactions were initiated by adding 10 μl of human liver cathepsin L or recombinant human cathepsin S dissolved in reaction buffer to give final concentrations of 2.9 and 25 nM respectively. Reactions were performed at 30ºC and followed by excitation/emission at 360/460 nm. For reactions containing Der p 1 prodomain, the reaction buffer volume was 60 μl and prodomain added in 10 μl aliquots.
Expression and purification of Der p 1 propiece
The propiece of Der p 1 was PCR cloned and expressed in E. coli. As expected from the chosen expression strategy, PP-01 was expressed as a soluble protein on induction with IPTG (Fig. 1A). It was found that PP-02 was also obtainable as a soluble protein (data not shown). An ammonium sulphate cut of the crude PBS extract of induced cells was loaded onto a Superdex G75 column and eluted with PBS to yield major peak (I) with an elution volume of 12–16 ml (Fig. 1B). This was purified further by ion exchange chromatography (Fig. 1C) from which peak I was prepared by gel filtration and desalting. The propiece was produced at a yield of 0.15 mg/50 ml cells and analysed as >95% pure by SDS-Tricine PAGE (Fig 1D). A MALDI-TOF mass spectrum obtained after in-gel tryptic digestion of the expressed protein confirmed its identity (Fig. 1E,F).
Inhibition of Der p 1 peptidase activity
In characterizing the action of Der p 1 propiece on the catalytic activity of Der p 1 we first examined the effects of the crude PBS extract made from IPTG-induced cells. Whereas PBS extracts made from transformed but noninduced cells had no effect on Der p 1 peptidase activity, extracts made from IPTG-induced cells inhibited Der p 1 reaction rate by an average of 86% in four experiments.
We then tested the effects of purified Der p 1 propiece. Studies were undertaken with both PP-01 and PP-02 with indistinguishable results. Figure 2A displays representative progress curves for Der p 1 catalysed cleavage of ADZ 50,059. In the absence of inhibitors, cleavage of the substrate proceeded quickly. Addition of ADZ 50,000, an irreversible inhibitor of Der p 1, prevented hydrolysis of the substrate. In contrast, purified PP-01 or PP-02 significantly reduced the rate of reaction compared with the positive control, but did not fully prevent cleavage of the substrate under the conditions examined. The IC50 measured for PP-01 was 60.5 ± 4.3 nM (Fig. 2B). As expected from its sequence similarity, Der f 1 is also capable of cleaving ADZ 50,059 and the Der p 1 propiece showed potent inhibitory activity against mature Der f 1 protein (Fig. 2C). Next, we investigated four synthetic peptides as potential inhibitors, chosen because modelling experiments and the reported crystal structure of proDer p 1 suggested them to span, or be proximal to, the catalytic site of the enzyme. As shown representatively in Fig. 2D for the longer pair, none of the synthetic peptides showed inhibitory behaviour.
In further experiments, we studied the kinetics of ADZ 50,059 cleavage of by Der p 1 in the presence and absence of propiece. The apparent Km was significantly increased from 15.4 ± 2.1 μM to 56.1 ± 7.6 μM (n = 4) by 20 nM propiece (Fig. 3) whereas Vmax was not reduced. Although the mechanisms are not revealed by this experiment, the data apparently suggest that the propiece could be a competitive inhibitor of Der p 1.
Cleavage of Der p 1 propiece by mature Der p 1
To investigate the inhibitory interaction between the propiece and Der p 1, both proteins were incubated together at 20°C for 0–5 h and the mixtures then analysed in the activity assay and by SDS-Tricine PAGE. Co-incubation of the prodomain and Der p 1 resulted in the time-dependent loss of inhibitory activity and, after 5 h, the propiece was ineffective as an inhibitor (Fig. 4A). Control studies showed that the loss of activity was not because of an inherent instability of the propiece because proteins incubated separately for 5 h prior to the peptidase assay retained their full inhibitory capacity (Fig. 4B). SDS-Tricine PAGE revealed that co-incubation of the proteins resulted in the gradual disappearance of the propiece with the appearance of new bands at lower molecular masses (Fig. 4C), and it confirmed the stability of the propiece when incubated alone. The degradation of the propiece is likely to be the explanation why, under the conditions employed, PP-01 and PP-02 were unable to inhibit Der p 1 fully. Confirmation that Der p 1 degraded the propiece was provided by inclusion of ADZ 50,000 in the incubation mixture. This prevented proteolysis and the appearance of cleavage products (Fig. 4C). To examine the general susceptibility of the propiece to cleavage by other cysteine peptidases, we examined its effects on cathepsins L and S. This revealed that neither prodomain inhibited the activity of these enzymes (Fig. 5), suggesting that PP-01 and PP-02 are neither tight binding inhibitors nor competitive substrates of cathepsins L or S.
Here, we provide the first evidence that Der p 1 differs from other C1 cysteine peptidases in the regulation of its activity by its prodomain. Whereas rDer p 1 prodomain superficially behaves as a potent inhibitor of the enzyme, this is actually because of competitive proteolysis that results in extensive degradation of the propiece by low concentrations of Der p 1. This starkly contrasts with the well-established behaviour of other C1 family cysteine peptidases in which untethered propieces control unchecked proteolysis by the mature enzyme because they are persistent, nondegradable inhibitors (26–30, 32, 33).
For Der p 1, the significance of rapid inactivation of its prodomain means it is quickly able to achieve full catalytic activity and this may be significant for the potency of Der p 1 as an allergen. This proposal is supported by a growing body of evidence which suggests that, at least for some causes of allergy, peptidase activity makes an important contribution to allergic sensitization by increasing the transepithelial delivery of allergens and the creation of a signalling environment with a Th2 bias in which the immune response is maintained (2, 3, 5–8, 24). If the cleaved prodomain of Der p 1 were stable and retained noncovalent attachment to the mature protein, this would potentially limit the amount of proteolytic activity and its potency as an allergen. Given their sequence similarity, group 1 allergens from other house dust mite species probably behave in a similar manner to Der p 1. This is supported by our experimental observation of excellent substrate cross-reactivity between the Der p 1 propiece and mature Der f 1. In addition, if the cleaved Der p 1 prodomain remained noncovalently attached, the immunogenicity of the molecule would be reduced. The reasoning here is that unprocessed proDer p 1 is known to be less immunogenic than mature Der p 1 (20–22) and that not all of this reduction in immunogenicity is because of the absence of proteolytic activity. A combination of structural and functional investigations of proDer p 1 indicates that access to at least two major B-cell surface epitopes is sterically hindered by the prodomain (21, 22). Intriguingly, the IgE binding of rproDer p 1 with a loss of function mutation to the active site cysteine is the most different from native Der p 1 (22). One of the main contact zones between Der p 1 or Der f 1 and their respective prodomains appears to lie within the substrate binding grove proximal to the catalytic site, presumably preventing nucleophilic attack by C132/133 (Der p 1/Der f 1). The other is the prodomain binding loop where D228 is key (22). Although these interactions remain to be fully characterized, present evidence implies that the conformation of the bound prodomain is crucial to making the complex less immunogenic than expected. Thus, mechanisms capable of destroying any untethered prodomain should result in Der p 1 gaining maximum immunogenicity because they would remove molecules capable of blocking surface epitopes considered to be important in the triggering of disease.
Inspection of the prodomain sequences of group 1 mite allergens suggests that they differ from other cysteine peptidases, and the functional significance of these differences is now revealed by our studies. Some members of this C1 sub-family, such as cathepsin B, have short prodomains, whereas in others, exemplified by cathepsins K, L and S, the prodomains are longer (96–99 residues). The prodomains of these cathepsins all contain the ERFNIN motif which maintains the correct conformation of the propiece for inhibitory activity. Similarly, when other C1 cysteine peptidases are compared, the pattern of long prodomains and the ERFNIN sequence persists. Thus, caricain (116 residue prodomain) has a perfect ERFNIN motif, while actinidain (102 residues), vignain (106 residues) and Dictyostelium discoideum cysteine peptidase 1 (99 residues) have ERFNIN motifs of correct length and any substitutions are few, conservative and essentially structure-neutral.
In contrast, group 1 mite allergen prodomains are of intermediate length and lack clear ERFNIN motifs. The prodomains of Der p 1, Der f 1 and Eur m 1 all consist of 80 residues and by CLUSTALW sequence alignment the equivalents of the EX3RX2(I/V)FX2NX3IX3 N motifs are DX3AX2NFX2SX3VX2NG, EX3AX2NFX2SX3VX2NK and KX3AX3NFX2SX3VX2NK respectively. In the structure-based sequence alignment reported for recombinant proDer p 1, procathepsin L and procaricain (22), the true spatial equivalent of an ER(I/V)FNIN sequence in the Der p 1 prodomain is DEX2ARXNFX2SX3VX2NG, in which bold text indicates residues that would align with an expected ER(I/V)FNIN motif and underscores illustrate where any identical ERFNIN residues actually occur. The CLUSTALW and previously published structure-based alignment highlight the lack of identity and suggest a significant structural impact, which may contribute to the Der p 1 prodomain having a looser fit with the enzyme’s binding grove. These and possibly other features translate into the effect revealed by our data, which demarcate Der p 1 from other members of the C1 sub-family to which it has hitherto been assigned.
The extensive work concerning the prodomains of cathepsins shows that the potency for homologous inhibition is typically sub-nanomolar. For example, Carmona et al. (26) showed that a propiece construct consisting of 87 of the 96 residues of the complete cathepsin L prodomain sequence had a Ki of 0.088 nM against cathepsin L, similar to the inhibitory activity (Ki 0.12 nM) reported for a recombinant cathepsin L prodomain preparation largely comprising a mixture of truncated (71 and 84 residue) forms of the propiece (28). Maubach et al. (30) reported a Ki of 0.27 nM for recombinant cathepsin S prodomain on cathepsin S, compared 0.46 nM for a mixture comprising a full length and an 11 residue truncated version of the cathepsin S propiece (28). In Der p 1, the issue of potency and selectivity is entirely different because the prodomain is a competitive substrate and is therefore clearly able to inhibit substrate degradation by the related dust mite allergen Der f 1 because of this. It is noteworthy that the Der p 1 prodomain was not degraded by either cathepsin L or S, indicating that it behaves selectively as a substrate among related with C1 cysteine peptidases.
Although the full-length Der p 1 prodomain is not therapeutically useful as a Der p 1 inhibitor and its truncated forms are not useful templates for inhibitor design, our observations do not imply that selective small molecule inhibition of Der p 1 is unattainable. As shown here, the peptoid ADZ 50,000, which is unrelated to the propiece sequence, exemplifies the feasibility of achieving excellent inhibition, and offers scope for further tuning its selectivity of action. The findings from this study have revealed a novel and unexpected feature of Der p 1. They support a proposal from structure predictions (22) by providing the first functional demonstration that Der p 1, and by implication all related group 1 mite allergens, should be considered a new sub-family amongst C1 cysteine peptidases.
We thank Asthma UK and the Wellcome Trust for financial support.
- 18House dust mites. In: BarnesPJ, GrunsteinMM, LeffAR, WoolcockAJ, editors. Asthma. Philadelphia, PA: Lippincott-Raven, 1997:1089–1103., .
- 19Allergen structure and function. In: AdkinsonNF, YungingerJW, BusseWW, BochnerBS, HolgateST, SimonsFER, editors. Middleton’s Allergy. Principles & Practice. Philadelphia, PA: Mosby, 2003:585–609., .