Introduction

Pathogenesis-related class 10 (PR-10) proteins comprise a unique class of highly conserved phytoproteins found in both monocotyledonous and dicotyledonous plants. PR-10 are primarily cytosolic proteins, constitutively expressed in several plant tissues including roots, stems, flowering compartments, fruits, and pollen grains from certain plant species. Their expression is upregulated upon abiotic and biotic stress conditions, such as invading pathogenic viruses, bacteria and fungi, cold, salinity, drought, oxidative stress, ultraviolet radiation, and physical wounding [1]. Hence, it has been proposed that PR-10 proteins have no unique function but play a more general role in plant development and defense mechanisms [2, 3].

PR-10 proteins are 154 to 163 amino acids long and have a molecular weight of approximately 17 kDa. Their 3-dimensional structure consists of an anti-parallel, seven-stranded β-sheets wrapping around an amphipathic C-terminal α helix (α3) embraced by two short α-helices (α1, α2) forming a V-shape. The main structural feature of the PR-10 fold is a large solvent-accessible, hydrophobic internal cavity spanning the entire protein [4, 5]. This hydrophobic core of PR-10 molecules serves as a binding site for a wide variety of ligands, thus explaining its promiscuous binding behavior. Here we review recent findings on ligand binding of PR-10 proteins with a particular focus on the Bet v 1 family and the impact of bound ligands on the immunological properties of these allergenic proteins. In addition, published data on PR-10 ligand binding in the broad context of plant biology are briefly discussed (Fig. 1).

Fig. 1
figure 1

Overview on described functions associated with PR-10 ligand binding; PR-10, pathogenesis-related proteins class 10; IPR, intracellular pathogenesis-related proteins/classic PR-10 proteins; CSBP, cytokinin-specific binding proteins; MLP, major latex proteins; other PR-10-like, not yet classified proteins possessing a PR-10-like fold

Full size image

Ligand Binding of PR-10 Proteins

Ligand binding has been described for many PR-10 proteins, mostly for allergens derived from plants of the Fagales order, such as birch (Betula verrucosa), hazel (Corylus avellana), and beech (Fagus sylvatica). PR-10 proteins from more distantly related allergenic food sources like peanut (Arachis hypogaea), strawberry (Fragaria ananassa), cherry (Prunus avium), and peach (Prunus persica) were also shown to bind ligands [6••, 7••, 8,9,10]. Interestingly, not all reported ligand binding PR-10 proteins have been described as allergens. In fact, the taxon of PR-10 proteins can be subdivided into intracellular pathogenesis-related (IPR) proteins/classic PR-10, cytokinin-specific binding proteins (CSBP), and major latex proteins (MLP), although CSBP and MLP only share a sequence identity of about 17 to 25% with the IPR group [1]. In this review, we have additionally included “allergens” as another subgroup of PR-10 because not all proteins displaying the typical features and the canonical fold of classic PR-10 are allergenic. Nevertheless, their common ability to bind low molecular weight compounds within their hydrophobic cavity can be considered a general feature of PR-10 proteins. The promiscuous ligand binding behavior of PR-10 proteins has been extensively documented using a large panel of phytohormones and plant metabolites. Based on these studies, three major chemical classes of PR-10-binding ligands have been defined: (i) cytokinins, (ii) flavonoids, and (iii) sterols [10,11,12,13]. A detailed overview of PR-10 ligands described in the literature is given in Table 1.

Table 1 Protein-ligand interaction of PR-10 proteins
Full size table

Cytokinins are phytohormones involved in the regulation of plant development, growth and defense mechanisms, cell division, and the deceleration of senescence [9]. Binding to the naturally occurring cytokinins zeatin and kinetin has been mostly described for CSBPs [8, 9, 12,13,14,15,16,17,18, 20,21,22,23, 24••]. According to the current hypothesis, CSBPs might have evolved within PR-10 proteins in order to maintain cytokinin homeostasis by specifically binding to these molecules. Thus, the specific PR-10 fold is highly conserved within the CSBP group supporting the notion that their fold enables efficient shuttling and storage of bioactive molecules [18, 19]. Interestingly, the PR-10 from Medicago truncatula (MtN13) was observed to play a role as regulator of free cytokinins during the early phases of nodulation, i.e., the recruitment process of symbiotic nitrogen-fixing bacteria [13]. Based on the fact that zeatin is a nucleoside analog, PR-10 proteins have been proposed to possess nuclease activity [15]. In this respect, several studies focusing on the putative RNase or DNase activity of PR-10 proteins derived from several plant species including birch, cotton, ginseng, peach, and pepper have been reported. However, the biological relevance of these studies has not been further explored and the reported PR-10/DNA/RNA interactions are still a matter of debate [9, 12, 15, 25,26,27,28,29,30]. Apart from the cytokinins zeatin and kinetin, other phytohormones, such as phytomelatonin, brassinosteroids, and gibberellic acid, were frequently observed to bind PR-10 proteins [14, 31].

Another important class of PR-10 ligands are flavonoids, which are polyphenolic compounds and secondary plant metabolites involved in color and flavor production, UV protection, antioxidation, and pathogen defense [32]. PR-10 allergens, mostly of the Fagales order, were shown to bind flavonoids with high binding affinity within their hydrophobic pocket, but the specificity appears to differ between PR-10 from different species. For instance, the natural ligand of hazelnut Cor a 1, Q3O-(Glc)-Gal is highly similar to the birch Bet v 1 co-purified natural ligand, quercetin-3-O-sophoroside (Q3OS), differing just in the orientation of the hydroxyl group. Nevertheless, the Cor a 1 ligand does not interact with Bet v 1 and vice versa, despite the high structural and sequence similarity between both allergens [7••]. In addition, Q3OS bound only to the major isoform, Bet v 1.0101 and not to other isoforms [33]. These observations are in line with the notion that the presence of many PR-10 isoforms in a plant could be the basis for their functional diversification [34]. In contrast, a study by Mc Bride et al. reported a more general, less discriminative binding of Bet v 1 and hazelnut Cor a 1 to other flavonoids (e.g., quercetin, genistein, apigenin, daidzein, resveratrol), highlighting the relative binding promiscuity of PR-10 allergens for this class of ligand [11]. Accordingly, other PR-10 allergens, such as strawberry Fra a 1, white oak Que. a 1, hazel pollen Cor a 1, and peanut Ara h 8, also bound numerous flavonoids [10, 11, 32]. The exact role of flavonoid binding by PR-10s remains to be elucidated; however, it is speculated that this protein class possesses major regulatory functions in flavonoid biosynthesis via binding and storage of functionally inert glycosylated flavonoids, which are in turn prevented from early activation by enzymatic deglycosylation by glycosyltransferases [7••, 10, 32]. This hypothesis is supported by the observation that the major allergen of beech pollen, Fag s 1, binds with high affinity to naringenin, a metabolic intermediate of flavonoid synthesis, whereas it does not interact with unglycosylated quercetin [8]. Similarly, the binding affinity of Bet v 1 to Q3OS was 60-fold stronger compared with quercetin [4].

Ligand Binding of Bet v 1

IgE sensitization to Bet v 1, the major allergen of birch (Betula verrucosa) pollen, ranges from 53 to 95% among birch pollen allergic patients. Similar sensitization rates were observed for Bet v 1 homologues like Mal d 1 (apple) and Cor a 1 (hazelnut), mostly due to cross-reactivity occurring between structurally similar PR-10 proteins [35, 36]. This results in clinical manifestations described as the oral allergy syndrome (OAS), a very common allergic disorder affecting more than 50% of birch pollen allergic patients [37]. Bet v 1.0101 (formerly designated Bet v 1a) represents the most abundant isoform of Bet v 1 comprising 50 to 70% of the allergen in pollen [4, 5]. The natural function of Bet v 1 is still not fully understood; however, based on the structural similarity with the START domain of the human MLN64 protein, it has been suggested that Bet v 1 play a role in steroid binding [38]. The ability of Bet v 1 to bind a broad spectrum of plant intrinsic ligands, such as fatty acids, cytokinins, or flavonoids, has led to the suggestion of an involvement in different stages of plant reproduction (e.g., protection of pollen DNA from UV-damage, transportation of lipids or flavonoids to the stigmatic surface to support pollen hydration and germination). Bet v 1 may also act as an storage scaffold for such ligands, enabling their rapid release upon seed germination [39]. In addition, the combination of different ligands and Bet v 1 isoforms with differences in their ligand preferences could possibly serve as molecular fingerprints to prevent self-pollination [23, 33, 40].

Bet v 1 has been shown to bind a broad range of hydrophobic to amphipathic ligands, differing in size and shape, to distinct binding sites within its hydrophobic cavity [23]. In general, pollen-derived molecules that interact with Bet v 1 can be grouped into fatty acids, flavonoids, and phytohormones. A detailed description of Bet v 1 ligands reported in the literature is given in Table 2.

Table 2 Ligand binding of Bet v 1
Full size table

The affinity of Bet v 1 towards fatty acids depends on their chain length and reaches a maximum at 14 to 18 carbon atoms including stearate, palmitate, and myristate. Among flavonoids, flavone, naringenin, apigenin, genistein, quercetin, and daidzein were shown to interact with Bet v 1. The cytokinins N6-(2-isopentenyl)adenine (IPA), kinetin, and zeatin bind to Bet v 1 with lower affinities, whereas the phytohormones indole-3-acetic acid, gibberellic acid, and abscisic acid showed no interaction [11, 23].

Other phytohormones reported to bind Bet v 1 in vitro are the brassinosteroids, brassinolide, and 24-epicastasteron [41]. However, the interaction was analyzed using Bet v 1.0107 (formerly designated Bet v 1l), a hypoallergenic isoform constituting only up to 7% of the total Bet v 1 content in birch pollen [45, 46]. Dehydroergosterol, a frequently used cholesterol model compound, as well as progesterone and the plant sterol stigmasterol were demonstrated to bind Bet v 1 [11, 23]. Recently, phytoprostane E1 (PPE1) and its derivatives B1 (PPB1) and F1 (PPF1) were shown to bind to Bet v 1.0101 with dissociation constants (Kds) of 0.5, 1.0, and 2.4 μM, indicating high binding affinity [6••]. Phytprostanes were detected in relevant quantities in birch pollen and shown to be associated with the induction of immune responses relevant for allergic sensitization [47, 48].

In 2014, Seutter von Loetzen et al. isolated the glycosylated flavonoid Q3OS in complex with Bet v 1 directly from birch pollen extracts. Its binding to this PR-10 protein was confirmed using nuclear magnetic resonance (NMR) spectroscopy and surface acoustic wave (SAW) technology, and is now considered a physiological ligand of Bet v 1 [4, 6••]. Q3OS is a catechol derivative able of binding iron ions [49]. In vitro, the presence of iron strongly affected the affinity of Q3OS to Bet v 1, potentially impacting its allergenicity [50].

Bacterial toll-like receptor (TLR) agonists were often suggested as ligands of Bet v 1 able to modulate its allergenicity [51, 52]. However, this notion has been challenged by a recent study from our laboratory in which very high dissociation constants were measured for the TLR4 agonist lipopolysaccharide (LPS) and the TLR2 agonist lipoteichoic acid (LTA) (199.8 and 185.0 μM, respectively), indicating no physiologically relevant binding [6••].

Besides plant-derived ligands, several artificial model ligands and extrinsic molecules binding Bet v 1 were investigated including the bile acid sodium deoxycholate (DOC), a structural analog of brassinosteroids able to bind Bet v 1 at two distinct sites [6••, 41,42,43, 53]. Other non-physiological ligands include the detergents sodium dodecyl sulfate (SDS), sodium tetradecyl sulfate (STS), and 3-[Benzyl(dimethyl)ammonio]propane-1-sulfonate (NDSB-256) [6••, 23, 42, 44]. Studies using these surrogate compounds mainly focused on the effects (e.g., fold stability, dynamics, conformation) induced by ligand binding.

Influences on Physicochemical Properties and Proteolytic Processing of Bet v 1

Several ligand binding-induced effects on the physicochemical properties of Bet v 1 were reported with a focus on the influences on the protein’s thermal stability and proteolytic susceptibility [6••, 53]. In general, ligand binding was observed to be associated with rigidification and compactness of the 3D structure of Bet v 1, while the conformational dynamics and flexibility were reduced. No significant alterations regarding the secondary structural element composition due to ligand binding were observed [6••]. Changes in structural dynamics of Bet v 1 affect the accessibility of its proteolytic cleavage sites to lysosomal proteases leading to a reduction in cleavage efficiency and consequently affecting the availability of Bet v 1 peptides for optimal peptide presentation to T cells via the major histocompatibility complex class II (MHC-II) receptor [54, 55]. Antigen presentation facilitated by antigen presenting cells is a necessary aspect for Th2 polarization, a key mechanism mandatory for the development of IgE-mediated allergic immune response. The presentation of allergen-derived peptides via the MHC-II pathway requires the processing of the allergen by lysosomal proteases. The influence of Bet v 1 fold stability on antigen presentation was elucidated by studying several fold stabilized mutants of Bet v 1. In comparison with wildtype Bet v 1, a mutant with optimal proteolytic stability to lysosomal proteases was able to stimulate a pronounced allergy-associated Th2 response in mice, suggesting that fold stability is a contributing factor for allergenicity [56].

In our recent study, describing a higher proteolytic resistance of Bet v 1 due to ligand binding, we have provided a mechanistic explanation for the increased allergenicity of the stabilized Bet v 1 mutant. The study also revealed a newly identified high affinity Bet v 1 ligand known as PPE1, which not only stabilized Bet v 1 from proteolytic degradation by lysosomal proteases but also modulated the proteolytic lysosomal activity of cysteine cathepsins [6••]. This dual-role of the ligand consequently influenced the presentation of the immunodominant Bet v 1 T cell epitope via MHC-II loading. Hence, ligand binding can regulate the processing of Bet v 1 by lysosomal proteases subsequently affecting availability of peptides for MHC-II presentation due to alterations in the frequency of proteolytic processing cleavage sites.

Methods Used to Determine Ligand Binding

The ANS displacement assay (ADA) for assessing the binding of various hydrophobic ligands to Bet v 1 was first described by Mogensen et al. [23]. The assay is based on the ability of the fluorophore 8-anilino-1-naphthalenesulfonic acid (ANS) to interact with PR-10 proteins. ANS behaves as a weak fluorophore in solution, but upon binding to the hydrophobic patches of a protein, its fluorescence intensity dramatically increases [18]. ANS binds to Bet v 1 via its extended hydrophobic surface patches within its prominent cavity [42]. If the ANS binding sites are already occupied with other ligands, this results in a decrease of ANS fluorescence; thus, ligand-induced ANS displacement can be fluorescently monitored. Due to its convenience, ADA is usually the first method of choice to investigate ligand binding of PR-10 proteins, as indicated in Tables 1 and 2. However, the assay has the disadvantage of providing only an indirect measurement that relies on the capacity of ligands to share the same binding sites and to displace ANS. The analysis of the observed fluorescence changes can be complicated by several factors, such as inner filter effects or fluorescence of by the competing ligand. In the first case, a decrease would artificially indicate binding of the tested ligand where the decrease was mostly due to a reabsorption of the ANS fluorescence signal; in the second case, ligand binding could increase the fluorescence signal. Additionally, ligand interaction with ANS and/or Bet v 1 may tune the emission maximum of ANS, e.g., by inducing structural rearrangements which may also result in additional ANS interaction sites in the respective PR-10 molecule. Similarly to ANS, the naturally occurring polyphenol resveratrol was also observed to increase its intrinsic fluorescence upon binding to the hydrophobic patches of Bet v 1 [11].

Alternatives to ADA for the investigation of PR-10 ligand binding include direct measuring techniques, such as isothermal titration calorimetry (ITC), surface acoustic wave (SAW), and microscale thermophoresis (MST), which enable accurate determination of binding affinity and Kd [6••, 9, 18].

NMR and X-ray crystallography also proved to be powerful tools for the identification of structural rearrangements induced by ligand binding compared with the apo-protein, and for information on the binding stoichiometry [4, 6••, 7••, 42]. Both techniques can be used to precisely map the amino acid residues involved in ligand interaction at the respective binding sites. Interestingly, attempts to crystallize some PR-10 proteins without ligand were not successful, consistent with a superordinate role of ligands in protein rigidification and reduction of structural dynamics necessary for successful crystallization [13]. Besides the aforementioned techniques, in silico docking experiments were performed in several studies to evaluate the binding of ligands [15].

Conclusions

Recent studies emphasize that PR-10 proteins’ structural and functional properties cannot be understood by their proteinogenic properties only. A structurally diverse spectrum of hydrophobic ligands can bind selectively to the PR-10 protein family, thereby tuning their physiologic and immunologic functions. Importantly, the organic cargo of PR-10 protein can exhibit secondary functions by addressing distant protein targets. The dual functions of protein ligands are exemplified by the phytoprostane PPE1, which affects protein stability of Bet v 1 and at the same time covalently blocks papain-related cysteine proteases via a Michael reaction warhead.

While the dual proteinogenic and organic composition shapes the physiological functions of PR-10 proteins, it as much accounts for their allergenic properties. The selective binding of ligands provides a rational for hypo- versus hyper-allergenic properties of allergen isoforms, which were previously explained by their differences in amino acid composition and 3d structure. In the light of the new findings, it may become possible to convert an allergen into a hypoallergen by differential ligand loading.