Recombinant Bordetella pertussis pertactin (P69) from the yeast Pichia pastoris: high-level production and immunological properties Michael A. Romanos*, Jeffrey J. Clare, Katrina M. Beesley, Fred B. Rayment, Stuart P. Ballantine, Andrew J. Makoff, Gordon Dougan, Neil F. Fairweather and Ian G. Charles Acellular whooping cough vaccines are based on pertussis toxoid but their effectiveness may be increased by the addition of other Bordetella pertussis antigens. We expressed the immunogenic outer membrane protein pertactin (P69) from B. pertussis to high levels in multi-copy transformants of the industrial yeast Pichia pastoris. In high-density fermentations, engineered P. pastoris yielded > 3 g of the protein per litre of culture. Purified recombinant pertactin was able to stimulate the incomplete protection afforded by toxoid to the level of the whole-cell vaccine, as shown by the Kendrick test, supporting its inclusion in future acellular vaccines. Keywords:Pertussis vaccines;P69; pertactin; yeast; Pichia;heterologousexpression

INTRODUCTION Whooping cough is an acute respiratory disease which is particularly severe for infants and is responsible for approximately one million deaths each year ~. Current vaccines consist of killed Bordetella pertussis cells and are very effective at preventing disease and infant mortality1'3. However, public acceptance of whole-cell vaccines has decreased due to the high rate of side-effects and the controversy over attributed rare neurological complications4-6. Therefore there has been pressure to develop an acellular vaccine consisting of purified Bordetella antigens. Candidate acellular vaccines are based on inactivated pertussis toxin (PTX) that has been chemically or genetically toxoided 7'8. However, a recent Swedish clinical trial of two toxoid-based vaccines, one also containing filamentous haemagglutinin (FHA), gave disappointing results and the vaccine was not subsequently licensed 9. This has renewed interest in the identification of additional Bordetella antigens that stimulate protection to the level of the whole-cell vaccine. These might be particularly important in protecting against B. parapertussis which also causes severe disease yet does not produce PTX 1°. We previously reported that a B. bronchiseptica 68 kDa outer membrane protein was protective in the piglet atrophic rhinitis model of whooping cough ~1. A protective monoclonal antibody (BB05) to the 68 kDa protein cross-reacts with membrane

proteins of B. pertussis (69 kDa) and B. parapertussis (70 kDa). The B. pertussis protein (P69 or pertactin) has mammalian cell binding activity lz, is protective in animal models using aerosol challenge with B. pertussis 13, and elicits a humoral and cellular immune response in humans 14-16. These results suggest that pertactin is involved in colonization and should be considered for inclusion in acellular vaccines. However, only small amounts of pertactin (5 10mgl-~ of culture, P. Novotny, personal communication) can be isolated from B. pertussis, and this protein would be limiting compared to the other antigens which would be included in a multi-component subunit vaccine. The pertactin structural gene encodes a 93.5 kDa primary polypeptide (910 amino acids) which is processed by removal of an N-terminal signal peptide and a C-terminal polypeptide of 30 kDa ~7.~8. To produce large amounts we examined the expression of mature Met-pertactin in two species of yeast, Saccharomyces cerevisiae and Pichia pastoris. Several vaccine antigens have been produced in yeast cells, including hepatitis B surface antigen which is in clinical use 19. P. pastor& is an industrial methylotrophic yeast which is readily cultured to very high density and has been developed as an expression system for foreign proteins z°.

MATERIALS AND METHODS

Construction of yeast expression vectors for pertactln Wellcome Research Laboratories, Langley Court, Beckenham, Kent BR3 3BS, UK. *To whom correspondence should be addressed. (Received 26 April 1991; revised 24 June 1991; accepted 10 July 1991) 0264-410)(/91/120901-06 © 1991 Butterworth-HeinemannLtd

A S. cerevisiae expression vector for Met-pertactin (pWYG5-P69, Figure 1) was constructed based on the GALl promoter vector pWYG5 described previously11. The DNA encoding Met-pertactin terminating at Ser-600

Vaccine, Vol. 9, December 1991 901

B. pertussis P69 protein from yeast: M.A. Romanos et al.

Transformation of yeast cells and DNA analysis of integrants

EcoR I

Pst I

Pst pWYG 5-P69 15.6 kb

-EcoR

I

.EU2 o EcoR I

EcoR I

I

EcoR I BamHI Bgl I

Bgl II

EcoRI

[ % ~ ~ ~ ~,

. ~,

~

~

~-~

3'A OX 1

"

Figure 1 Yeast P69 expression vectors. Regions of the S. cerevisiae vector pWYG5-P69 derived from 2/~ plasmid are shaded and other major features are indicated. The P. pastoris vector pPIC3-P69 was based on pPIC322. The plasmid map shows the AOXl promoter and terminator either side of the P69 DNA, and the AOXl downstream region. The Bglll restriction sites are used to generate a fragment, containing the expression cassette and HIS4 selectable marker, which can displace the chromosomal AOXl gene

was isolated as a BgIII BamHI fragment from the bacterial expression vector pPERtac8 TM and cloned between the BamHI and BclI sites of pWYG5. The P. pastoris vector was based on pPIC222 which is an, integrating vector utilizing the methanol-inducible A OX1 promoter and the HIS4 selectable marker, pPIC3-P69 (Figure 1) was constructed by cloning an EcoRI-Nhel fragment from pPERtac818 containing most of the Met-pertactin gene between the BamHI and SpeI sites of pPIC2 using the following oligonucleotide linker to recreate the 5' end of the gene: GATCCAAACGA TGGACTGGAACAACCAATCC ..... G T T T G C T A C C T G A C C T T G T T G G T T A G G ..... ATCGTCAAGACCGGTGAAAGACAACACGG TAGCAGTTCTGGCCACTTTCTGTTGTGCCTTAA

The S. cerevisiae vector pWYG5-P69 was introduced into a 2F' version of the strain S150-2B as described previously 23. P. pastoris strain G S l l 5 (his4-) was transformed to the His + phenotype with BglIl-digested pPIC3-P69 by the sphaeroplast method, and methanolutilization slow (Mut ~) transformants were identified as described previously 22. In order to screen for transformants containing multiple copies of integrated pPIC3-P69, a rapid DNA dot blot analysis was carried out. Mut S transformants were grown in individual wells in a 96-well microtitre plate (Nunclon) in 200/~1 YPD broth (! % yeast extract, 2% peptone, 2% glucose) for 2 days at 30'~C. Samples (50/~1) were then filtered onto nitrocellulose placed in a Schleicher and Schuell 'minifold' using a multi-channel pipetter. The filters were air-dried then treated to lyse the cells in the following way: (i) 15 min with 50mM EDTA, 2.5% 2-mercaptoethanol pH 9; (ii) 4 h at 3T'C with ! m g ml- 1 zymolyase 100T (Seikagaku Kogyo Co.); (iii) 5 rain in 0.1 M N a O H , 1.5 M NaCI; and (iv) twice for 5min in 2 × S S C ( I × S S C = 0 . 1 5 M NaCI, 0.015M sodium citrate, pH 7). Each treatment was performed by placing the nitrocellulose filter on two sheets of 3MM paper soaked with the solution. The filter was then baked at 80-C and probed with pertactin DNA labelled with 32p using the random priming method 24. Multi-copy integrants were identified by a strong hybridization signal, and the integrated vector copy numbers determined accurately by dot blot analysis of purified DNA probed with radio-labelled pertactin DNA or DNA from the P. pastoris single-copy gene ARG422.

Inductions, protein analysis and purification Transformed S. cerevisiae and P. pastoris cells were induced and protein extracts prepared as described previously 22'z3. Proteins were separated in 7.5% SDS-polyacrylamide gels and stained with Coomassie blue or else detected in Western blots using the BBO5 monoclonal antibody TM. Pertactin in cell extracts was quantified in Western blots by comparison to a dilution series of known amounts of the protein, or by densitometric scanning of Coomassie blue-stained gels (Joyce-Loebl Chromoscan 3). Pre-stained protein markers used (Rainbow markers, Amersham) were: myosin 200kDa, phosphorylase b 92.5 kDa, bovine serum albumin 69 kDa, ovalbumin 46 kDa, carbonic anhydrase 30 kDa. P. pastoris transformants were induced in high-density fermentations in a Braun 2 1 fermenter, using a protocol of batch growth in glycerol, followed by glycerollimitation, then a controlled methanol feed for induction 22. The cells were harvested by low-speed centrifugation, washed once in water, and suspended in one culture volume of ice-cold buffer A (50 mM Tris/HC1 pH 8.0, 0.1 M NaC1, l mM EDTA). Resuspended cells were broken with acid-washed 0.45 mm glass beads in a bead beater (Biospec Products, Bartlesville, Oklahoma, USA) in ten l-minute pulses. The lysate was cleared by centrifugation at 15 000 rev min- 1 for 30 min. Pellets were washed by suspending in buffer A containing 1% Triton X- 100 and centrifuging again. The pellet of washed insoluble protein was resuspended in buffer A to a protein

B. pertussis ,°69 protein from yeast: M.A. R o m a n o s et al.

[a

4

s:

6

Renatured pertactin in the dialysate was purified by applying to a chelating sepharose (Pharmacia) column charged with zinc, equilibrated in 50 mM Tris/HC1, pH 8, 0.1 M NaC1. The column was washed with 0.5 M NaC1 and elution was in 50mM 2-(N-morpholino)ethanesulphonate (MES), pH 6, 0.1 M NaC1. The eluate was dialysed against buffer A, applied to Q-sepharose (Pharmacia) equilibrated in 50 mM Tris/HC1, pH 8, and pertactin eluted in a gradient of 04).4 M NaCI.

lntracerebral challenge (Kendrick) test Vaccines contained 5 #g (test A) or 20/~g (test B) of pertussis toxoid, 20 #g of pertactin, and 10% alhydrogel/ PBS per mouse. These were serially diluted in PBS and 0.5 ml doses given intraperitoneally to groups of 18 male and female NIH/S mice. Control mice received whole-cell vaccine (British reference 66/84) at a top dose of 6.25 IU m1-1. After 14 days the mice were challenged intracerebrally with 20pl of B. pertussis 18-323 ( ~ 5 0 L D s o ) and survivors counted after a further 14 days. RESULTS AND DISCUSSION

Expression of pertactin in Saccharomyces and Pichia We compared the levels of expression of pertactin using the episomal GALl vector pWYG5-P69 in S. cerevisiae, or the single-copy integrating vector pPIC3-P69 in P. pastoris. The plasmid pPIC3-P69 was digested with BglII prior to transformation, generating a fragment able to displace the chromosomal AOX1 gene by doublecrossover recombination (transplacement): the resulting aoxl- transformants have the Mut s phenotype. About 20% of the His + transformants were scored as MutS; Southern blot analysis of chromosomal DNA from selected transformants, e.g. SLI, was used to confirm that transplacement had occurred (data not shown). Western blot analysis (Figure 2a) of proteins from induced S. cerevisiae cells containing pWYG5-P69 indicated that pertactin was produced at low levels, estimated at 0.1% of cell protein, and was fully soluble. P. pastoris Mut s transformants, e.g. SL1 (Figure 2a), were found to produce pertactin at about 0.5% of cell protein, and about 50% of the product was insoluble (i.e. sedimented upon centrifugation at lO000g for 5 min).

Multi-copy integrants of P. pastoris

Figure 2 (a) Western blot analysis of pertactin in induced S. cerevisiae and P. pastoris cell extracts. Tracks contained (1), 0.5/ag authentic B. pertussis pertactin or (2), 20/tg protein extract from induced P. pastoris single-copy integrant SL1, (3), multi-copy integrant SL3, (4), SL4, (5), SL18, (6), SL22 and (7) S. cerevisiae/pWYGS-P69.(b) DNA dot blot screen of P. pastoris GS115/pPIC3-P69 M uP transformants. Rare transformants (1-2%) with very high copy numbers were identified, e.g. a3, a4, b6 (designated SL3, SL4, SL18). The positive control w a s e l 0 (multi-copy strain SL22 from a previous screen), and the negative control e l l (host strain GSl15)

concentration of 6 mg ml-1 then six volumes of 7 M guanidinium chloride, 50 mM Tris/HC1 pH 8.0, 1 mM EDTA were added and the suspension homogenized to solubilize proteins. The homogenate was diluted to 1 M guanidinium chloride then dialysed against buffer A.

Transformation with BglII-digested vector is predicted to give rise to single-copy transplacements at the chromosomal AOX1 locus z°. However, in expressing tetanus toxin fragment C we found that multiple head-to-tail insertions of the BglII fragment occurred frequently, and these gave greatly increased levels of product 22. With pPIC3-P69 we did not readily isolate such multi-copy transplacements. We therefore screened several hundred Mut s transformants for the presence of multiple integrated copies of pPIC3-P69 using a rapid DNA dot blot method on transformants grown in microtitre plates. Most transformants gave a weak signal and Southern blot analysis confirmed that they were single-copy transplacements. Multi-copy integrants showed up as strong hybridization signals on the DNA dot blot (Figure 2b), and occurred at a low frequency (1 2%). Four were selected: SL3, SL4, SL18 and SL22.

B. pertussis P69 protein from yeast: M.A. Romanos et al. Table 1 Pichia

Expression of pertactin in recombinant Saccharomyces and

Yeast/plasmid

Copy no2

S. cerevisiae pWYG5-P69

> 50

P. pastoris pPIC3-P69 (SL1,MuP) pPIC3-P69 (SL18,MuP) pPIC3-P69 (SL3,Mut ") pPIC3-P69 (SL22,MuP) pPIC3-P69 (SL4,Mut +)

1 12 13 21 30

Expression leveP (% cell protein)

0.1 0.5