Purification and Characterization of Recombinant Expressed Apple Allergen Mal d 1

Mal d 1 is the primary apple allergen in northern Europe. To explain the differences in the allergenicity of apple varieties, it is essential to study its properties and interaction with other phytochemicals, which might modulate the allergenic potential. Therefore, an optimized production route followed by an unsophisticated purification step for Mal d 1 and respective mutants is desired to produce sufficient amounts. We describe a procedure for the transformation of the plasmid in competent E. coli cells, protein expression and rapid one-step purification. r-Mal d 1 with and without a polyhistidine-tag are purified by immobilized metal ion affinity chromatography (IMAC) and fast-protein liquid chromatography (FPLC) using a high-resolution anion-exchange column, respectively. Purity is estimated by SDS-PAGE using an image-processing program (Fiji). For both mutants an appropriate yield of r-Mal d 1 with purity higher than 85% is achieved. The allergen is characterized after tryptic in gel digestion by peptide analyses using HPLC-MS/MS. Secondary structure elements are calculated based on CD-spectroscopy and the negligible impact of the polyhistidine-tag on the folding is confirmed. The formation of dimers is proved by mass spectrometry and reduction by DTT prior to SDS-PAGE. Furthermore, the impact of the freeze and thawing process, freeze drying and storage on dimer formation is investigated.


Introduction
The allergen Mal d 1 is the primary apple allergen in northern Europe [1]. Due to the great structural similarity between the allergen Bet v 1 in birch and Mal d 1 in apple, the immune system can only distinguish between the two proteins with difficulty [2][3][4]. Therefore, about 50−70% of people affected by birch pollen allergy develop a cross-allergy to apple (Malus domestica) during their lifetime [1, 5,6]. Although the 17.5 kDa large allergen Mal d 1 is heat-labile and proteases break it down during gastric digestion, [7] the symptoms during consumption, such as breathing difficulties, itching and burning of the mucous membranes of the mouth or tongue (Oral Allergy Syndrome) mean that an estimated 2 million affected people in Germany alone generally avoid eating fresh apples [1,5].
It is striking that the allergenic potential, which up to now has only been examined diagnostically on the effect level after oral provocation, skin prick tests or immunochemical testing with allergy sera, clearly differs between different apple varieties [8,9]. However, these differences could not be sufficiently correlated with the variety-or cultivation-related differences in the Mal d 1 content so that the absolute content of Mal d 1 cannot be the sole cause of an apple's allergenic potential [10][11][12]. An interaction of Mal d 1 with other phytochemicals reducing the allergenic potential has been postulated [13][14][15], but the structural requirements for the phytochemicals, as well as the interaction mechanisms, are still unknown. To investigate interaction strength, kinetics, and desired reaction conditions for different phytochemicals, purified Mal d 1 is a prerequisite. Purification by immobilized metal ion affinity chromatography (IMAC) for proteins exhibiting a C-terminal His-Tag might be a quick procedure to obtain high purities. Nevertheless, in lots of experimental settings, this additional feature is perturbing. Therefore, a process to purify the protein without a His Tag is required as well. Commonly time-consuming multi-step purification methods, including FPLC (fast protein liquid chromatography) based on a combination of different separation principles like ion exchange (IEC), hydrophobic interaction and size exclusion chromatography, is used [16][17][18].
In contrast, in this manuscript the development of a rapid method to produce and purify sufficient amounts of r-Mal d 1 with and without a polyhistidine-tag in high purity for further experiments is described.

Experimental Design
The production protocol for r-Mal d 1 described here is divided into three parts, of which one is a protein purification step ( Figure 1). After plasmid transformation in Escherichia coli (E. coli), the protein is expressed in the host cells, and the harvested cells are lysed by pressure. For the purification of the clear cell lysate containing r-Mal d 1 with a Cterminal polyhistidine-tag (r-Mal d 1-His), a HisTrap™ FF column is employed, performing IMAC. To purify the clear cell lysate with r-Mal d 1, a chromatographic system, the ÄKTA purifier FPLC system, and a high-resolution anion exchange column the TOYOPEARL TM GigaCap Q-650M (IEC) are essential to perform this purification. However, no additional steps are required to obtain pure r-Mal d 1 after this step.

Preparation of Solutions
Use ultrapure water to prepare all the solutions in the protocol. Make liquid Lennox Broth by dissolving 10 g tryptone, 5 g NaCl, and 5 g yeast extract in 1 L water. Autoclave the liquid LB and let it cool to room temperature. To prepare solid LB agar plates, add 15 g of agar per 1 L. For bacteria culture autoclave the broth and cool before adding 25 µg/mL sterile-filtered kanamycin and the transformed E. coli. For cell lysis steps and purification, prepare the following buffers:  After cell harvest by centrifugation at 4 • C (8980× g, 20 min), the washed cell pellets can be stored at −20 • C until further use.
• Filter the supernatant (filter, 0.2 µm) and use the lysate immediately for purification by immobilized metal ion affinity chromatography (IMAC) on a HisTrap™ FF (r-Mal d 1-His) or by ion exchange chromatography (IEC) on a TOYOPEARLTM SkillPakTM column (r-Mal d 1).

Quantity and Purity of the r-Mal d 1-His and r-Mal d 1
Pooled fraction volumes and protein concentrations for IMAC and IEC purification are shown in Table 2. To evaluate the purity and to identify the r-Mal d 1, 10 µ L of the pooled fractions was mixed with 40 µ L millipore water and 12.5 µ L SDS Laemmli sample buffer (fourfold concentrated), denaturized at 95 °C and separated on SDS-PAGE gel, which was stained with Coomassie Brilliant Blue. After washing in aqueous acidic isopropanol (80/10/10, v/v/v), the gels were scanned, and the purity was evaluated by Fiji [20]. If required, the bands were used for in-gel tryptic digestion (Figure 3). The fraction of the

Quantity and Purity of the r-Mal d 1-His and r-Mal d 1
Pooled fraction volumes and protein concentrations for IMAC and IEC purification are shown in Table 2. To evaluate the purity and to identify the r-Mal d 1, 10 µL of the pooled fractions was mixed with 40 µL millipore water and 12.5 µL SDS Laemmli sample buffer (fourfold concentrated), denaturized at 95 • C and separated on SDS-PAGE gel, which was stained with Coomassie Brilliant Blue. After washing in aqueous acidic isopropanol (80/10/10, v/v/v), the gels were scanned, and the purity was evaluated by Fiji [20]. If required, the bands were used for in-gel tryptic digestion (Figure 3). The fraction of the IMAC clean up eluted with 10 mM imidazole is of lower r-Mal d 1-His concentration (10.4 mg/mL) and less purity (65% r-Mal d 1-His) than the fraction with 100 mM imidazole (14.4 mg/mL and 85% r-Mal d1-His) (Table 2, Figure A1). The r-Mal d 1 without an additional polyhistidine-tag was purified on a strong anionexchange chromatography raisin. The high-resolution SuperQ-650S SkillPak TM usually contains High Capacity Ion Exchanger with a particle size of 35 µm. The GigaCap Q-650M SkillPak TM is filled with an Ultra High Capacity Ion Exchanger polymer-modified resin with a larger content of coupled ionic groups. Therefore, increased particle size is sufficient (75 µM), permitting higher flow rates. The first runs showed that TOYOPEARL TM GigaCap Q-650M performed somewhat better (Figure 3, run 1 vs. run 2); therefore, the procedure was optimized exclusively for this column (Table 1). Using a stepwise gradient (run 4) slightly increased the volume of fraction containing r-Mal d 1 with a purity higher than 85% but mainly shortened the procedure to 40 min ( Figure 2, Table 1). The r-Mal d 1 was concentrated in fraction #1 (Figure 3). However, negligible amounts of r-Mal d 1 were also found in the flow-through and fraction #2. The purity of the fractions was estimated by using an image-processing program (Fiji) [20]. The fractions used for SDS-PAGE gel analysis immediately after clean-up (run 4 and 5 #1 fresh) showed a Mal d 1 purity higher than 85% (Figure 4) with 3-4 impurities, below 5%. The SkillPak TM column's optimal loading volume is 5 mL lysate obtained from 1 g cells pellet. Lower amounts do not improve purity or yield. TOYOPEARL TM GigaCap Q-650M are available in greater dimensions enabling possibilities for upscaling.   Avoiding dimerization during storage, the purity around 90% is sufficient for interaction studies (e.g. Isothermal Titration Calorimetry, Saturation Transfer Difference NMR experiments and SAR by NMR) if the impurity for each individual protein is below 5%. For immunological studies often a purity ≥ 95% is suggested. However, distinct values for the Mal d 1 purity are not specified in the references cited. Nevertheless, further purification of the low Mal d 1 amounts required for these investigations might be easily performed by an additional HPSEC step.

Verification of r-Mal d 1 and Characterization of an Impurity Formed during Storage by Mass Spectrometry after in Gel Digestion
Due to differences in the apparent molecular weight among the r-Mal d 1 and r-Mal d 1-His and the successive formation of a compound of higher molecular weight in-gel digestion followed by mass spectrometry is applied, to verify the identity of band 1 to 4 ( Figure 3 ). The peptides generated by the tryptic digestion of band 1 and 2 as well as 3 and 4 indicate the expressed r-Mal d 1 as the primary protein ( Figure 5). The sequence coverage is excellent for band 1 (r-Mal d 1, 100%, run 1#1), for band 3 (r-Mal d 1-His, 98%), and good for band 2 (88%, run 2#1). Based on the peptides found for band 4 (98%), it is identified as an aggregate of r-Mal d 1-His, formed after the purification step during buffer exchange/concentration and storage. The formation of dimers under physiological conditions has been already mentioned by Roulias et al. Like Bet v 1, this group showed Avoiding dimerization during storage, the purity around 90% is sufficient for interaction studies (e.g., Isothermal Titration Calorimetry, Saturation Transfer Difference NMR experiments and SAR by NMR) if the impurity for each individual protein is below 5%. For immunological studies often a purity ≥95% is suggested. However, distinct values for the Mal d 1 purity are not specified in the references cited. Nevertheless, further purification of the low Mal d 1 amounts required for these investigations might be easily performed by an additional HPSEC step.

Verification of r-Mal d 1 and Characterization of an Impurity Formed during Storage by Mass Spectrometry after in Gel Digestion
Due to differences in the apparent molecular weight among the r-Mal d 1 and r-Mal d 1-His and the successive formation of a compound of higher molecular weight in-gel digestion followed by mass spectrometry is applied, to verify the identity of band 1 to 4 ( Figure 3). The peptides generated by the tryptic digestion of band 1 and 2 as well as 3 and 4 indicate the expressed r-Mal d 1 as the primary protein ( Figure 5). The sequence coverage is excellent for band 1 (r-Mal d 1, 100%, run 1#1), for band 3 (r-Mal d 1-His, 98%), and good for band 2 (88%, run 2#1). Based on the peptides found for band 4 (98%), it is identified as an aggregate of r-Mal d 1-His, formed after the purification step during buffer exchange/concentration and storage. The formation of dimers under physiological conditions has been already mentioned by Roulias

Verification of the Correct Folding by Circular Dichroism Spectroscopy
The correct folding of r-Mal d 1 and r-Mal d 1-His is monitored by CD spectroscopy and secondary-structure elements are calculated (see Appendix D). CD spectra of r-Mal d 1 and r-Mal d 1-His are analogous ( Figure 6). In accordance, their shape agrees with the spectra provided in the literature [9,18]. The analysis of the secondary structure results in 12%/19% α-helices, 39%/36% ß-sheet, 17%/19% turn and 32%/26% random coil for r-Mal d 1 and r-Mal d 1-His, respectively. These data conform to the secondary-structure information (25% α-helix, 35% ß-sheet) calculated from NMR spectra [2].

Verification of the Correct Folding by Circular Dichroism Spectroscopy
The correct folding of r-Mal d 1 and r-Mal d 1-His is monitored by CD spectroscopy and secondary-structure elements are calculated (see Appendix D). CD spectra of r-Mal d 1 and r-Mal d 1-His are analogous ( Figure 6). In accordance, their shape agrees with the spectra provided in the literature [9,18]. The analysis of the secondary structure results in 12%/19% α-helices, 39%/36% ß-sheet, 17%/19% turn and 32%/26% random coil for r-Mal d 1 and r-Mal d 1-His, respectively. These data conform to the secondary-structure information (25% α-helix, 35% ß-sheet) calculated from NMR spectra [2]. During storage at −20 °C and the freeze-thawing process, purity decreased due to the formation of a new compound with higher molecular weight (Figure 3 and 4, run 3-5 #1,

Impact of Storage Conditions and Freeze-Thawing Process on Dimer Formation
During storage at −20 • C and the freeze-thawing process, purity decreased due to the formation of a new compound with higher molecular weight (Figures 3 and 4, run 3-5 #1, frozen). Despite an apparent molecular weight above 34 kDa, this compound was characterized by mass spectrometry (see Section 3.3.1.) as a r-Mal d 1 dimer. In addition, by adding 10 mM dithiothreitol (DTT) to the sample 10 min before denaturation, the band disappeared ( Figure 6). With storage time, the dimer is probably formed by a disulfide bridge due to oxidation of the cysteine residues [18]. However, adding reducing agents during storage to prevent this dimerization is problematic because these substances might interfere in further experiments using the r-Mal d 1.
To evaluate the impact of freeze and thawing effects, we simulated 10 cycles in a one-day experiment and observed dimer formation by SDS-PAGE gel analysis and CD spectroscopy. In addition, the process of freeze-drying of Mal d 1 in Tris buffer was studied (Figure 7). The impact of freeze-thawing (up to 10 times within one day) and freeze-drying was less relevant for the dimer formation with contents of 19 ± 1% and 72 ± 2% for the dimer and monomeric r-Mal d 1, respectively ( Figure 6, purity evaluation not shown). In addition to the negligible impact of the freeze and thaw processes on dimer formation, no significant differences are observed in the respective CD spectra (Figure 8). The proportion of secondary-structure elements is constant within the standard deviations of independent measurements at the beginning (Table A1).  (A, B). Freeze-drying was performed with r-Mal d 1 in Tris buffer containing the NaCl as a residue of the ion-exchange chromatography and after re-buffering in phosphate buffer. Samples were diluted with water 1 to 5 prior to denaturation, the volume applied to the gel was 14 µL.
The impact of freeze-thawing (up to 10 times within one day) and freeze-drying was less relevant for the dimer formation with contents of 19 ± 1% and 72 ± 2% for the dimer and monomeric r-Mal d 1, respectively ( Figure 6, purity evaluation not shown). In addition to the negligible impact of the freeze and thaw processes on dimer formation, no significant differences are observed in the respective CD spectra (Figure 8). The proportion of secondary-structure elements is constant within the standard deviations of independent measurements at the beginning (Table A1).  . CD spectra for r-Mal d 1 for three independent measurements (A), data black, fit according to [21] in red) and at different cycles for freeze and thawing (B); 0, black; 2, green; 5, orange; 10, blue).
Dimer formation seems to increase with storage time (Figure 9) independently, whether in Tris or phosphate buffer. This effect was more pronounced for run 1 #1 (longest storage time) than for all other runs. Urea buffer was detrimental, forcing the formation of dimers during storage. However, the addition of 10 mM DTT (sample IMAC#2(A)+DTT, Figure 9) to the sample 10 min before denaturation leads to the monomer of high purity. The process of freeze-drying does not increase dimer formation. However, CD spectra indicate that folding is affected (data not shown). . CD spectra for r-Mal d 1 for three independent measurements (A), data black, fit according to [21] in red) and at different cycles for freeze and thawing (B); 0, black; 2, green; 5, orange; 10, blue).
Dimer formation seems to increase with storage time (Figure 9) independently, whether in Tris or phosphate buffer. This effect was more pronounced for run 1 #1 (longest storage time) than for all other runs. Urea buffer was detrimental, forcing the formation of dimers during storage. However, the addition of 10 mM DTT (sample IMAC#2(A)+DTT, Figure 9) to the sample 10 min before denaturation leads to the monomer of high purity. The process of freeze-drying does not increase dimer formation. However, CD spectra indicate that folding is affected (data not shown).  Institutional Review Board Statement: Not applicable.

Informed Consent Statement: Not applicable.
Data Availability Statement: Data sharing is not applicable to this article.