The de-linked residual serum samples used in the present study were obtained from the Biobank at the Taipei Veterans General Hospital. All these serum samples were obtained originally from respiratory atopic patients (allergic asthma and/or atopic rhinitis) who attended the allergy clinics of Taipei Veterans General Hospital and were stored in aliquots at -80℃. This study without written consent has been approved by the Institutional Review Board of Taipei Veterans General Hospital.

The F. proliferatum strain BCRC 30972 was used in this study. It was isolated from the air of Taiwan and provided by the Food Industry Research and Development Institute, Hsinchu, Taiwan. The crude extracts of F. proliferatum were prepared essentially as described previously.11 Briefly, F. proliferatum was cultured in a CYA medium without agitation at 26℃ for 5 days. The CYA medium contains yeast carbon base (Difco Laboratories, Detroit, MI, USA; 11.7 g/L), glucose (Mallinckrodt Baker, Inc., Phillipsburg, NJ, USA; 10 g/L) and casein enzymatic hydrolysate (Sigma Chemical Co., St. Louis, MO, USA; 10 g/L). The protein content of crude extracts was determined with a dye-binding method according to the manufacturer's instructions (Bio-Rad, Richmond, CA, USA).

The cDNA encoding the F. proliferatum vacuolar serine protease was isolated with an AffinityScript Multiple Temperature cDNA Synthesis kit (Stratagene, La Jolla, CA, USA) and polymerase chain reactions (PCR) as previously described.671011 Primers used in the cloning experiments are listed in Table. The degenerate primers VSP-F-1 and VSP-R-1 were used in the first set of polymerase chain reaction (PCR). They encode conserved amino acid sequences (KNAPWG and MASPHVAG) near the N- and C-termini of fungal serine proteases. The PCR product (first-strand cDNA) was purified electrophoretically on agarose gel, subcloned into the pGem-Teasy vector (Promega, Madison, WI, USA), and then transformed into E. coli Top10F' competent cells. The plasmid DNA was purified, and the nucleotide sequence of the cDNA insert was determined with an automatic sequencer (Applied Biosystems, Foster City, CA, USA).

The 5'-and 3'-end RACE (rapid amplification of cDNA ends) reactions were used to obtain the full-length cDNA of the vacuolar serine protease. The primers used were the anchor primers (AP and AAP) together with the gene-specific primers (FuVSP-GSP1 and FuVSP-GSP2) that were synthesized according to the internal sequences of the Fusarium vacuolar serine protease obtained in the previous PCR reaction. FuVSP-GSP2 (for 3'-end RACE) and FuVSP-GSP1 (for 5'-end RACE) cover nucleotides 661 to 684 and 164 to 186, respectively, of the truncated Fusarium vacuolar serine protease cDNA. Primer AP used in the 3'-end RACE reaction contains the sequence of primer AUAP plus a stretch of oligo-(dT). The primer AAP used in the 5'-RACE reaction contains the sequence of primer AUAP and a stretch of oligo-(dG). The first-strand cDNA isolated was used as a template in the 3'-end RACE reaction. An oligo-(dC) was added to the end of the first-strand cDNA with terminal deoxynucleotidyl transferase (Promega) before using as a template in the 5'-RACE reaction. Products of the RACE reactions were purified individually, subcloned, transformed, and sequenced as described above.

Fungal vacuolar serine proteases were hypothesized to be synthesized as a larger precursor that undergoes both N- and C-terminal cleavage upon maturation.716171819 The F. proliferatum vacuolar serine protease was expressed as 6x His-tagged protein according to the manufacturer's instructions (Qiagen Inc., Valencia, CA, USA). The cDNA of the Fusarium vacuolar serine protease was amplified through PCR. The forward primer (FuVSP-5' Sma I, Table) used in the reaction contains the SmaI restriction site in addition to the cDNA sequence (nucleotides 10-32) encoding the putative N-terminus of the mature Fusarium vacuolar serine protease. We hypothesize that this vacuolar serine protease precursor is cleaved at or near the Tyr (Y) residue encoded by nucleotides 967-969 of the isolated clone, analogous to the proposed C-terminus of the vacuolar serine protease (pepC) of A. nigeri.19 Thus, the reverse primer FuVSP-3' Hind III (Table) contains a Hind III restriction site, a newly added stop codon (TAA) and the sequence encoding the hypothetical C-terminus (KKIVEAGSY, nucleotides 943-969) of the Fusarium vacuolar serine protease. The PCR product was restricted, ligated into the pQE-80 vector, and then transformed into E. coli M15 for protein expression. The recombinant protein of putative mature vacuolar serine protease was affinity-purified with a Ni-NTA resin column (Qiagen) under the denaturing condition according to the manufacturer's instructions. In addition, the purified recombinant 6x His-tagged vacuolar serine protease allergen of P. chrysogenum (Pen ch 18) and the Pen ch 18 K89A mutant were also prepared as previously described.15 Immuno-reactivity of the recombinant proteins to IgE antibodies and MoAb FUM20 was analyzed by sodium dodecyl sulfate-polyacrylamide gel electrophoresis (SDS-PAGE)-immunoblotting.

Proteins in the crude Fusarium fungal extracts or the purified recombinant proteins were separated by SDS-PAGE81011 and then transferred electrophoretically onto polyvinylidene difluoride (PVDF) membranes (0.45 µm, Millipore, Bedford, MS, USA). Protein components reacting to human IgE antibodies or MoAb FUM20 against fungal vacuolar serine protease allergen12 were determined as previously described.8101112 The membranes were blocked with 1% skimmed milk and incubated with serum samples for 16 hours at 4℃ or with MoAb FUM20 for 1 hour at room temperature. The membranes were washed, incubated with alkaline phosphatase-conjugated monoclonal anti-human IgE antibodies (Pharmingen, San Diego, CA, USA) or with horseradish peroxidase labeled goat anti-mouse IgG (H+L) (Jackson ImmunoResearch Laboratories, Inc., West Grove, PA, USA), and then developed with enzyme substrates as previously described.8101112 Serum from a house dust mite-sensitized atopic individual and a MoAb WH920 against house dust mite allergen Der p 7 were used as controls.

The Fusarium vacuolar serine protease mutant carrying single alanine substitute at Lys88 was prepared as previously descried.15 A plasmid encoding the vacuolar serine protease obtained as described above was used as a template in PCR. The primers used in the mutagenesis experiment (FVSP K88A-f and FVSP K88A-r) are shown in Table. The newly synthesized DNA carrying the point mutation was purified and inserted into the pQE80 expression vector (Qiagen) and transformed into E. coli JM109 for the expression of the Fusarium vacuolar serine ptotease K88A mutant. The mutant construct was confirmed by DNA sequencing and affinity-purified with Ni-NTA resin columns (Qiagen) according to the manufacturer's instructions.

For immunoblot inhibition studies, IgE-containing serum samples were first reacted with purified recombinant Fusarium vacuolar serine protease, its K88A mutant, rPen ch 18, or rPen ch 18 K89A before incubation with PVDF blots containing F. proliferatum extracts or purified recombinant Fusarium vacuolar serine protease at 4℃ for 16 hours. As controls, the blots were incubated with similar serum samples that had been pre-incubated with equivalent amounts of bovine serum albumin (BSA, Pierce, Rockford, IL, USA). The blots were then washed and incubated with alkaline phosphatase-conjugated monoclonal anti-human IgE antibodies (Pharmingen) and developed with enzyme substrates as previously described.81011

In our previous study, a 36.5-kDa F. proliferatum component was found to bind IgE antibodies in 9 (53%) of 17 F. proliferatum-sensitized serum samples from respiratory atopic patients.11 In addition, this 36.5-kDa IgE-binding component also showed immunoblot reactivity to MoAb FUM20 against the vacuolar serine protease fungal allergens of Penicillium, Aspergillus, Rhodotorula, and Cladosporium species12 (data not shown). Results obtained suggest that the 36.5-kDa allergen may be a vacuolar serine protease of F. proliferatum. Thus, in this study the cDNA encoding the vacuolar serine protease of F. proliferatum was cloned.

The full-length cDNA encoding the F. proliferatum vacuolar serine protease was obtained through RT-PCR coupled with the 5'- and 3'-end RACE reactions. The isolated cDNA clone has 1,394 bp (GenBank accession no. KJ462778). The nucleotide and the deduced amino acid sequences (383 residues) of the open reading frame are presented in Fig. 1. A potential polyadenylation signal (AATACA) for mRNAs of higher eukaryotes was found at 13-18 bases upstream from the poly-A tail.

Sequence alignment suggests that this cDNA from F. proliferatum encodes a vacuolar serine protease. Assuming that the mature protein starts from Asp1, a protein of 40,478 Da can be translated from this cDNA. There are 4 cysteines (Cys81, Cys184, Cys215, and Cys308), 3 putative N-glycosylation sites (¹⁰⁷NGS¹⁰⁹, ¹⁴⁷NMS¹⁴⁹, and ¹⁸⁵NYS¹⁸⁷), and 3 active-site residues for serine proteases (Asp45, His77, and Ser243) on this protein (Fig. 1). The encoded protein has 67%-72% sequence identity with that of vacuolar serine proteases from Penicillium, Aspergillus, and Cladosporium species (Pen ch 18; accession no. AF263454; Pen o 18, accession no. AF243425; Pen c 18, accession no. AF245168; Asp f 18, accession no. Y13338; and Cla c 9, accession no. EF407520) plus 54% and 57% sequence identity with those from Rhodotorula mucilaginosa (Rho m 2, accession no. AY547285)21 and Saccharomyces cerevisiae (protease B, accession no. M18097),17 respectively. All these mature fungal serine proteases contain 3 cysteine residues. The most likely candidate for the free cysteine of the Saccharomyces protease B is Cys 81, located near the active-site His77.17 The remaining 2 may be responsible for a di-sulfide bond. Analogous to protease B, the 2 Cys (Cys184, Cys215) may form a disulfide bond.

Our results from immunoblotting demonstrated that the recombinant vacuolar serine protease from F. proliferatum (rFus p 9.0101) and the recombinant Pen ch 18 both reacted with MoAb FUM20 (panel B of Fig. 2). In addition, 9 atopic serum samples with IgE binding to Pen ch 18 allergen (Fig. 2, section II, strip nos. 1-9 of panel C) also demonstrated IgE immunoblot reactivity to rFus p 9.0101 (Fig. 2, section I, strip nos. 1-9 of panel C). For control experiments, MoAb WH9 against Der p 7 allergen20 and serum from a house dust mite-atopic subject showed negative immunoblot reactivity to rFus p 9.0101 and rPen ch 18 (data not shown and Fig. 2, strip no. 10 of panel C, respectively). This Fusarium vacuolar serine protease allergen has been designated as Fus p 9.0101 by the IUIS Allergen Nomenclature Sub-committee.

In this study, the intensity of IgE-immunoblot reactivity to the 36.5-kDa and 30-kDa components of F. proliferatum by using serum no. 1 of panel C in Fig. 2 (Fig. 3, strip 1 of panel B) was decreased dose-dependently after pre-absorption of the same serum sample by 20 µg and 5 µg of purified rFus p 9.0101 (Fig. 3, strips 2 and 3 of panel B). For the control experiment, a decrease in the intensity of IgE-immunoblot was not detectable following pre-absorption of the same serum sample by 20 µg of bovine serum albumin (BSA) (Fig. 3, strip 4 of panel B). In addition, the same 36.5-kDa component of F. proliferatum showed positive immunoblot reactivity to MoAb FUM20 (Fig. 3, strip 1 of panel C). MoAb WH9 as a control antibody showed negative immunoblot reactivity (data not shown). FUM20 binding to the 36.5-kDa F. proliferatum component was inhibited by 20 µg of rFus p 9.0101 (Fig. 3, strip 2 of panel C), but not by the same amount of BSA (Fig. 3, strip 3 of panel C). Results obtained indicate that the 36.5-kDa allergen (Fus p 9.0101) is a vacuolar serine protease of F. proliferatum. The 30-kDa component may be a degradation product of the 36.5-kDa vacuolar serine protease of F. proliferatum.

Results of Fig. 2 showed that in addition to Fus p 9.0101 and Pen ch 18, MoAb FUM20 reacted with the Fus p 9.0101 K88A mutant (section I, panel B) and the Pen ch 18 K89A mutant (section II, panel B). For control, MoAb WH9 showed negative immunoblot reactivity (data not shown).

In this study, the intensities of bands on the immunoblots were quantified with AlphaEaseFCTM software (version 4.0.0, Alpha Innotech Corpration, San Leandro, CA, USA). IgE binding to the rFus p 9.0101 K88A mutant was decreased by <20% inhibition for serum nos. 1 and 3, by 25%-30% for serum nos. 4, 5 and 7, by 50%-65% for serum nos. 8 and 9, and by about 80% for serum nos. 2 and 6 as compared to those against Fus p 9.0101 (Fig. 2, section I, panel C).

Immunoblot reactivity to IgE antibodies in 7 of the 9 atopic serum samples (strip nos. 2, 3, 4, 6, 7, 8, and 9 of Fig. 2, section II, panel C) tested demonstrated significantly decreased IgE binding (by >70% inhibition) to the rPen ch 18 K89A mutant compared to those to rPen ch 18. For the control experiment, serum sample from a house dust mite-sensitized atopic individual (Fig. 2, sections I and II, panel C, serum no. 10) showed negative immunoblot reacticity to both wild type allergens and their site-directed mutants. Results obtained indicate that the K88 and the K89 play a significant role in IgE binding to Fus p 9.0101 and Pen ch 18 allergens, respectively.

Results in Fig. 4 showed that serum no. 9 from Fig. 2, panel C has IgE-binding activity to rFus p 9.0101. For this serum sample, purified rFus p 9.0101 and rPen ch 18 decreased IgE binding to rFus p 9.0101 when the serum sample was pre-absorbed with 10 µg of the wild type rFus p 9.0101 or rPen ch 18 as indicated (Fig. 4). Pre-absorption of the same serum sample with 10 µg of the purified Fus p 9.0101 K88A mutant, the Pen ch 18 K89A mutant, or BSA did not inhibit its IgE binding to rFus p 9.0101 as indicated.