Expression of enzymatically inactive wasp venom phospholipase A1 in Pichia pastoris

Abstract

Wasp venom allergy is the most common insect venom allergy in Europe. It is manifested by large local reaction or anaphylactic shock occurring after a wasp sting. The allergy can be treated by specific immunotherapy with whole venom extracts. Wasp venom is difficult and costly to obtain and is a subject to composition variation, therefore it can be advantageous to substitute it with a cocktail of recombinant allergens. One of the major venom allergens is phospholipase A1, which so far has been expressed in Escherichia coli and in insect cells. Our aim was to produce the protein in secreted form in yeast Pichia pastoris, which can give high yields of correctly folded protein on defined minimal medium and secretes relatively few native proteins simplifying purification.Residual amounts of enzymatically active phospholipase A1 could be expressed, but the venom protein had a deleterious effect on growth of the yeast cells. To overcome the problem we introduced three different point mutations at the critical points of the active site, where serine137, aspartate165 or histidine229 were replaced by alanine (S137A, D165A and H229A). All the three mutated forms could be expressed in P. pastoris. The H229A mutant did not have any detectable phospholipase A1 activity and was secreted at the level of several mg/L in shake flask culture. The protein was purified by nickel-affinity chromatography and its identity was confirmed by MALDI-TOF mass spectrometry. The protein could bind IgE antibodies from wasp venom allergic patients and could inhibit the binding of wasp venom to IgE antibodies specific for phospholipase A1 as shown by Enzyme Allergo-Sorbent Test (EAST). Moreover, the recombinant protein was allergenic in a biological assay as demonstrated by its capability to induce histamine release of wasp venom-sensitive basophils.The recombinant phospholipase A1 presents a good candidate for wasp venom immunotherapy.

Figure 1. AT-content of Vesv1 gene and predicted tertiary structure of Ves v 1 protein.

Plotting of the AT-content of the Vesv1 gene was performed using Gene Atlas utility (www.cbs.dtu.dk). The dark brown areas on the third line show DNA stretches with AT content above 80%. The 3D structure of the Ves v 1 protein was generated using Geno3D (http://geno3d-pbil.ibcp.fr) using rat pancreatic lipase related protein 2 (1BU8) as a template (29.4% identity, 43.7% similarity) . The structure was visualized in PyMOL 1.3 (http://www.pymol.org). The mutated amino acids are shown in red.

Figure 2. Analysis of fermentation samples on Western blot.

10 µl of fermentation broth after 3-day induction on methanol at 20 and 30°C were analyzed by Western blot with anti-penta-His antibody. For several strains up to 4 transformants were tested and no significant difference between clones was observed. The marker is a pre-stained PageRuler (Fermentas, Germany). Lot-to-lot variation of the apparent molecular weight of pre-stained proteins in the ladder is ∼5%.

Figure 3. Recombinant proteins concentration and enzymatic activity.

Concentration of recombinant protein in fermentation broth (same samples as on Figure 2) were measured by ELISA with anti-penta-His antibody (A). Phospholipase A1 enzymatic activity was measured in a fluorometric commercial assay (Invitrogen) (B).

Figure 4. Expression analysis of the active and mutated Vesv1 forms by RT-PCR.

RT-PCR with primers binding to AOX1 promoter and AOX1 terminator was performed on total RNA of wild type P. pastoris and of P. pastoris expressing active and mutated versions of Vesv1 gene, all the strains were induced on methanol for 24 hours. The band of 2.2 kb corresponds to the native alcohol oxidase AOX1 gene, which is expressed during growth on methanol. The 1.4 kb-band corresponds to the recombinant Vesv1 transcript. In parallel to RT-PCR, identical control reactions were carried out without addition of reverse transcriptase. There were no bands detected in control reactions (not shown) confirming the absence of DNA contamination.

Figure 5. Purification of rVes v 1 H229A.

The enzymatically inactive protein was purified from P. pastoris fermentation broth using Ni-affinity chromatography. 10 µl samples of the fermentation broth, flow-through, wash and elution fractions were separated on SDS-PAGE and silver-stained. The size marker is 5 µl of 10-fold diluted unstained PageRuler (Fermentas).

Figure 6. SDS-PAGE gel of venom and purified rVes v 1 H229A and rVes v 5.

1,000 SQ units of venom extract and 300 ng of purified rVes v 1 H229A and rVes v 5 were separated on SDS-PAGE and stained with coomassie. The identity of recombinant allergens and of the native nVes v 1 and nVes v 5 proteins in the venom was confirmed by MALDI-TOF MS.

Figure 7. Binding of IgE antibodies from allergic patients sera to rVes v 1 H229A and rVes v 5 as measured by EAST.

EAST results are shown for twenty two sera. 5 sera (A, B, C, D and control serum) were chosen for further studies. Two sera (A and B) showed a positive response to rVes v 1 H229A and negative to rVes v 5, one serum (C) showed the contrary, serum D was positive for both allergens, while control serum did not react with either of the allergens.

Figure 8. Inhibition of binding of two Ves v 1-allergic sera to wasp venom extract by rVes v 1 H229A.

Microtiter plates were coated with wasp venom and were incubated with sera mixed with different concentrations of either wasp venom or rVes v 1 H229A to test their ability to inhibit the binding. Two sera A and B, which showed positive response to rVes v 1 H229A and negative to rVes v 5 in EAST assay, were chosen. The inhibition percent shows the decrease of absorbance in comparison to the sample where no inhibitor was present. The shown inhibition values are averages of two replicates. The absorbance A₄₉₀ values for non-inhibited samples were 1.1 for serum A and 0.9 for serum B.

Figure 9. Histamine release.

Histamine release from human basophils sensitized with IgE antibodies from five sera (same as on Figure 7) was tested when the basophiles were challenged with rVes v 1 H229A, rVes v 5 or wasp venom. The signal was considered positive when more than 10% of the total histamine present in the basophils was released.