An Alternative Platform for Rapid Production of Effective Subunit Vaccines

October 2, 2010

Adjuvant activity can be greatly improved by appropriate formulation of cytosine-phosphorothioate-guanine oligodeoxynucleotides (CpG ODNs).

ABSTRACT

The freshwater ciliate, Tetrahymena thermophila, has recently emerged as a novel manufacturing platform for recombinant subunit vaccines. It combines ease of growth with facile genetics in a complex unicellular eukaryote that can be grown rapidly in inexpensive media on an industrial scale. T. thermophila devotes a large part of its metabolism to membrane protein production, and can release proteins to the extracellular space by constitutive and stimulus-dependent pathways of secretion. In this article, we show high-level expression of correctly folded parasite and viral proteins in a Tetrahymena system and provide direct evidence that regulated secretion can be harnessed as an effective pathway for producing influenza hemagglutinin (HA). HA can be targeted to dense core granules in vivo and be recovered following stimulus-dependent secretion in association with a proteinaceous gel termed PRISM. PRISM offers a convenient matrix for protein purification, but at the same time, has intrinsic properties with the potential to induce potent immune responses to co-administered antigens.

Although cell-mediated immunity can play a significant role in clearing microbial pathogens, antibodies targeting secreted proteins or exposed antigens on microbial surfaces are often sufficient to generate protection against infectious agents. In such cases, vaccines that elicit strong, long-lasting antibody responses are highly effective in preventing disease. When compared with live and killed whole pathogens, subunit vaccines have distinct advantages particularly when dealing with newly emerging infectious agents such as pandemic influenza virus that have the potential to cause widespread disease. Subunit antigens can be produced rapidly on a large-scale as recombinant proteins without relying on attenuation or killing as steps in the production process. Subunit proteins are typically less reactogenic than live or killed pathogens and can be delivered at high antigenic mass. Furthermore, their production obviates the need to work directly with pathogens, which can be difficult to grow and dangerous. Finally, because recombinant antigens can be manufactured in microbial systems, the potential for contamination by adventitious agents that infect mammalian cells is minimized.

(TETRAGENETICS, INC.)

Despite the advantages of recombinant subunit proteins, their protective efficacy may be lower than that of native antigens because of a number of factors. Primary among these is absence of pathogen-associated molecules and other "danger signals" required to stimulate robust immune responses to purified antigens.1,2 Although these can be provided by adding immuonstimulatory substances, or adjuvants, to the vaccine formulation, there are few such substances currently approved by the FDA.3 Equally important is the problem of protein folding and post-translation processing. In the case of viruses and parasites, the preponderance of vaccine candidates are membrane or secreted proteins that must fold properly to elicit protective antibodies in the host. This requires accurate disulfide bond formation, which is often difficult to achieve in bacterial expression hosts.4 Some microbial systems (including yeast) also contain rigid cell walls that impede downstream protein purification, and contain endogenous pyrogens that must be removed in the production process.

To address some of the challenges surrounding potency, investigators have turned to eukaryotic expression hosts for improved protein folding (especially insect cells and fungi),5–7 and using virus-like particles (VLPs) that can present antigens in a repetitive, high-density format that can generate strong B- and T-cell responses.8,9 Still, VLPs are not applicable to all vaccine antigens and can be difficult or slow to produce in high yield.

As an alternative to these approaches, we have focused on Tetrahymena thermophila, a eukaryotic microbe that is capable of rapid, scalable growth. Tetrahymena lacks a cell wall and adds mammalian-like post-translational modifications (PTMs) onto proteins.10–13 More importantly, T. thermophila devotes a large-part of its metabolism to membrane protein production owing to the hundreds of cilia that extend from its surface. Moreover, Tetrahymena not only constitutively secretes proteins, but also stores large amounts of protein in hundreds-to-thousands of dense core granules that can be induced to secrete at will.14,15 Indeed, the material released from these granules takes the form of a proteinaceous gel (termed PRISM), which can easily be harvested from cells by low-speed centrifugation, providing a natural matrix for streamlined protein purification. PRISM has an underlying crystalline structure, and like VLPs, offers the opportunity to present antigens in a repetitive format that is optimal for cross-linking of the immunoglobulin (Ig) receptor on B-cells. In this article, we demonstrate the expression of correctly folded viral and eukaryotic vaccine antigens in the Tetrahymena system and provide preliminary evidence that PRISM may be an ideal matrix for rapid production of highly potent, low-cost vaccines.

Materials and Methods

The gene encoding the H5 hemagglutinin from H5N1 (Vietnam) subtype avian influenza virus (accession number: EF541402.1) was codon-optimized for expression in Tetrahymena and introduced into the macronuclear genome, either on high-copy number rDNA vectors, or in the somatic MTT1 metallothionein gene locus by homologous recombination.16 In both instances, the transgene was under the control of the endogenous MTT1 promoter.17 For the constitutively secreted gene product, the region encoding 48 amino acids at the C-terminus of the protein was deleted. For targeting to mucocysts, the truncated H5 gene was tagged with a sequence encoding one of several granule lattice proteins of Tetrahymena. The gene encoding the IAG4B[G1] i-antigen gene of Ichthyophthirius multifiliis (accession number: AAD31283) was introduced into cells, either at the b-tubulin 1 (BTU1) locus under MTT1 promoter control, or on high copy number rDNA vectors as above. For the truncated gene product, the region encoding the C-terminal 19 amino acids of the IAG48[G1] gene was removed. Cells were transformed biolistically and positive transformants selected by growth in paromomycin.18 Cells were grown to ~5 x 105 cells/mL in Neff medium and treated with 2 μg/mL CdCl2 for varying periods of time to induce expression of the transgenes. In the case of granule-targeted H5, cells were washed and mucocyst discharge induced by adding dibucaine to a final concentration of 2 mM.14 The PRISM matrix was then harvested for further analysis by centrifugation at 6,000g for 5 min and transferred to fresh tubes.

To prepare monoclonal antibodies (MAbs) recognizing H5, BALB/c mice were injected intraperitoneally with killed reassortant A/Vietnam/1203/2004 X PR8 (H5N1; 105 HAU) mixed with Freund's incomplete adjuvant (FIA). One to four months later, mice were injected intravenously with 6–9 x 104 HAU virus in phosphate-buffered saline (PBS). After three days, mice were euthanized and spleens were collected. Lymphocytes were harvested and fused to SP2/0-AG14 mouse myeloma cells using conventional methods.19 Mice were purchased from Harlan (Indianapolis, IN) and housed in the James A. Baker Institute vivarium according to the guidelines of the American Association of Laboratory Animal Care. Hybrid cells secreting antibodies specific for virus were identified by testing supernatants in hemagglutination inhibition (HI) assays, according to Barret and Inglis.20 Supernatants from positive wells were tested in ELISA for binding to virus using a peroxidase-conjugated anti-mouse IgG as previously described.21 Supernatants were tested for binding to allantoic fluid (1:5 dilution) as a negative control. Positive hybrid cells were cloned by limiting dilution on mouse peritoneal exudate cells. Cloning was repeated until 100% of single colony wells tested positive in ELISA and HI. Antibodies were isotyped by modifying the ELISA, adding rat anti-mouse isotype reagents detected with peroxidase conjugated goat antibodies specific for rat kappa chain (BD Pharmingen, San Diego, CA).

Western blotting of whole cell lysates and cell fractions was carried out using standard protocols22 and signals generated using Super signal West Pico Luminol/Enhancer Solution and Stable Peroxide Solution, mixed 1:1 (Pierce Biotechnologies), before image capture either on film (Kodak BioMAX MS) or with a CCD camera (Chemigenius; BioRad). Primary antibodies in all cases were dilutions of hybridoma culture supernatant fractions. Expression was quantitated by densitometry using internal software on the Chemigenius.

For localizing recombinant gene products, cells were fixed either before or after adding primary mouse MAbs in 50 mM HEPES buffer containing 4% paraformaldehyde in the cold. Cells were washed, treated with primary and secondary antibody, and mounted in ProLong Gold anti-fade reagent (Invitrogen). Images were acquired with a Leica SP5 confocal microscope using a 63X water objective. For detecting the IAG48[G1] i-antigen, MAb 10H3 was used.23

Table 1. H5N1 MAbs used in Western blots

Animal immunization was carried out as follows. Six-week old male Albino Oxford (AO) rats were anesthetized with Isoflurane and injected intraperitoneally with 250 mL of PRISM matrix containing recombinant H5 diluted 1:3 in PBS. An equal volume of Freund's incomplete adjuvant (FIA, Sigma) was then added and the mixture emulsified. Rats were anesthetized and boosted with the same amount of antigen in FIA four weeks after the initial injection, and blood was collected three weeks later using terminal heart puncture. Microneutralization assays were carried out with A/Vietnam/1203/2004xPR8 (VN04) reassortment virus as described previously.24 Cells were fixed and virus infection was detected by ELISA using antibodies against Influenza A virus nucleoprotein.

Figure 1. Surface expression of viral and parasite antigens in Tetrahymena. Panels A and B: Transformed cell lines were induced with CdCl2 overnight and expression of recombinant antigens was visualized by confocal microscopy. For H5 (panel B), cells were stained with a 1:50 dilution of MAb 5C5 (Figure 2) followed by a 1:300 dilution of goat antimouse IgG coupled to Texas Red. For the IAG48[G1] i-antigen (panel A), cells were stained with a 1:100 dilution of MAb 10H3, followed by the secondary antibody as above. Panel A is a stacked Z-series showing fluorescence on ciliary and plasma membranes, while panel B shows a single section with localization of H5 at the plasma membrane. Magnification bars = 10 μm. Panels C and D: Cells transformed with the full-length H5 gene were induced for varying periods of time and then lysed in SDS-sample buffer in the absence of reducing agent, subjected to SDS-PAGE and transferred to nitrocellulose. Blots were screened with a 1:400 dilution of MAb 5C5, followed by a 1:10,000 dilution of goat antimouse IgG coupled to HRP. Panel C shows bands corresponding to monomeric H5 and higher order aggregates. Blots were scanned and expression quantitated by densitometry. As shown in panel D, H5 gene product accumulated over time through at least 48 h in cell lines containing the transgene at the somatic MTT1 locus.

Results and Discussion

To assess the utility of T. thermophila as an expression system for subunit vaccines, genes for candidate antigens from viral and protozoan pathogens were introduced into the macronuclear genome, either by homologous recombination or using ribosomal DNA (rDNA) vectors that replicate to high copy number (~9,000) in these cells. The antigens in each case were a 48 kDa GPI-anchored surface protein (i-antigen) from the parasitic protozoan, Ichthyophthirius multifiliis,25 and the H5 hemagglutinin from H5N1 (Vietnam) subtype avian influenza virus. In both instances, high-level expression of the respective proteins was detected. Based on quantitative Western blotting, recombinant proteins reached ~3–5% of total cell protein (equivalent to ~60–100 mg/L without optimization; data not shown) and they correctly localized to ciliary or plasma membranes (Figure 1). More importantly, both proteins were properly folded as determined by their ability to bind neutralizing and protective MAbs against conformational epitopes dependent on disulfide bond formation. As shown in Figure 2, among seven different neutralizing MAbs against the H5N1 virus (Table 1), all seven recognized the recombinant protein expressed in Tetrahymena. Bands corresponding to the H5 monomer, dimer, and trimer were present suggesting that the protein assumed correct quaternary structure as well (Figure 2). Although H5 was present primarily on the plasma membrane, the recombinant i-antigen was localized on both ciliary and plasma membranes, resulting in complete arrest of cell movement on adding protective MAbs.26

Figure 2. Recognition of recombinant H5 by neutralizing monoclonal antibodies. Whole cell lysates from transformed T. thermophila, expressing the full-length recombinant H5, were tested in Western blots with a panel of MAbs that are neutralizing for H5N1 avian influenza virus (see Table 1). Cells were lysed in SDS buffer containing either 0% (lane 1), 0.25% (lane 2), or 4% (lane 3) β-mercaptoethanol and proteins run on 12% polyacrylamide gels. Primary antibodies were diluted 1:400 in Tris-buffered saline (TBS) containing 20% nonfat dry milk, and secondary antibody (goat antimouse HRPO conjugate) diluted 1:10,000 in the same buffer. Bands corresponding to the expected sizes of HA monomers (M), dimers (D), and trimers (T) were seen. Recognition of the recombinant antigen was sensitive to reducing agent in all cases.

Purification of the full-length membrane-associated antigens required detergent lysis and chromatographic separation on multiple resins. Nevertheless, because both proteins are linked to the lipid bilayer by hydrophobic domains at their C-termini (a single transmembrane-spanning domain in the case of H5, and a GPI-anchor in the case of the i-antigen), we reasoned that purification could be simplified by directing truncated versions lacking C-termini to the constitutive secretory pathway, where they could be readily harvested from culture supernatants. Following transformation and expression, the truncated i-antigen could be detected as a Coomassie blue stainable band in culture supernatant fractions (Figure 3A) allowing straightforward downstream purification. Truncated H5, on the other hand, was detectable only at low levels in culture supernatant fractions, and showed punctate staining throughout the cytoplasm, possibly because of aggregation (Figure 3B). This is consistent with previous reports using insect and mammalian cells in which C-terminal deletions of certain HA subtypes failed to secrete as a result of inappropriate folding or assembly.27,28

Figure 3. Targeting to the constitutive secretory pathway. Truncated genes lacking the coding sequences for the hydrophobic GPI-anchor signal sequence of the 48 kDa i-antigen (i-ag) of I. multifiliis, and the transmembrane spanning domain of H5 hemagglutinin of avian influenza virus were cloned and expressed in Tetrahymena (Panels A and B, respectively). Panel A is a Coomassie blue stained gel of concentrated culture supernatant fractions from cells before (lane 1) and after (lane 2) induction of gene expression. The secreted i-ag is clearly present in this fraction following induction of gene expression. Panel B is an immunofluorescence image of Tetrahymena expressing the truncated H5 hemagglutinin. The viral gene product was present at low levels in culture supernatant fractions (not shown) and appeared as scattered punctate staining within transport vesicles or the ER. Magnification bar = 10μm.

Although most eukaryotic cells allow secretion by the constitutive pathway, specialized cells have an additional pathway in which proteins are stored in cortical secretory granules and then rapidly released in response to specific stimuli. Regulated, stimulus-dependent secretion is evolutionarily conserved (from protozoa to mammals)29 and offers an attractive alternative to constitutive secretion as a means of protein production and purification.30Tetrahymena has an elaborate apparatus for regulated secretion consisting of large numbers of crystalline, dense core granules known as mucocysts that store upwards of 20% of total cell protein.14,15 Mucocyst granules remain docked at the plasma membrane but discharge their contents en masse within seconds of adding appropriate stimuli. The material released from the granules takes the form of a proteinaceous gel (termed PRISM) that can be harvested by low-speed centrifugation (Figure 4). With the idea that this pathway could be harnessed for the production and purification of H5, we linked the viral hemagglutinin to a member of a small family of granule-associated proteins known as Grls, with the hope of targeting it to mucocysts. As shown in Figure 5, the viral gene product was in fact localized to dense core granules following transformation and induction of the chimeric gene product. Furthermore, when cells were induced to secrete in response to dibucaine, H5 was found exclusively in the PRISM phase (Figure 5).

Figure 4. Stimulus-dependent secretion in T. thermophila. As depicted in panel A), adding secretagogues to the culture medium results in an explosive discharge of cortical secretory granules (mucocysts) from Tetrahymena that is complete within seconds of the stimulus. The material released from granules takes the form of a proteinaceous gel that can be readily harvested by low speed centrifugation. The tube on the right contains cells from a 50 mL culture that had been induced to secrete and spun at 6,000g for 5 min. The mucocyst gel, termed PRISM (P) is shown in brackets just above the white cell pellet. In panel B, cells were induced to secrete on an EM grid and gently washed off. The grid was then stained with uranyl acetate and examined under the electron microscope. Note the underlying crystalline structure of the hydrated PRISM matrix. Magnification bar = 0.5 μm.

The ability to target H5 to mucocysts for rapid secretion represents a first-step toward streamlined purification of the antigen itself. Although the recombinant protein bound neutralizing antibodies, it remained to be determined whether it could elicit protective antibodies in animals. Because the chimeric H5 was abundantly expressed in PRISM, the entire matrix containing the recombinant antigen was injected intaperitoneally into rats and the resulting antisera tested for both cross-reactivity against baculovirus-expressed H5 in Western blots, and for its ability to neutralize live virus in cell culture. Rat antisera against the chimeric antigen showed strong binding to the purified protein in Western blots, whereas no signals were obtained with nonimmune sera or secondary antibodies alone (Figure 6). More importantly, sera from two of three animals injected with chimeric H5 showed extremely high neutralizing titers in assays with reassortant A/Vietnam/1203/2004(H5N1) influenza virus on cultured MDCK cells (Table 2).

Figure 5. Targeting to the regulated secretory pathway. Panels A and B are immunofluorescence images showing localization of recombinant H5 in cortical mucocysts of transformed Tetrahymena labeled with MAb 5C5. Panel A is a stacked Z-series, while panel B is single Z-section through the same cells. Antibody concentrations were the same as in Figure 1. Note the obvious punctate staining of mucocysts at the cell periphery. Panel C is a stacked Z-series showing mucocyst granules in wild-type (CU428) cells stained with MAb 5E9, which recognizes the granule-lattice protein Grl3p (generous gift of Aaron Turkewitz, University of Chicago). Panel D shows a Western blot reacted with MAb 5C5 containing fractions from cells induced to secrete by adding dibucaine to the culture medium. Lanes contain, in order: 1) cells bodies after secretion; 2) PRISM; 3) original cell culture medium before secretion; 4) 6,000g supernatant from washed cells before secretion; 5) 6,000g supernatant from cells after secretion. Lanes were loaded equally by volume.

The ability to direct foreign proteins to the regulated secretory pathway of Tetrahymena has clear advantages from the standpoint of protein production. First, cells can be grown to high biomass, harvested in buffer without contaminating media or secreted proteins, and then induced to secrete within a very narrow time frame. The material stored in mucocysts comprises small families of granule-lattice and b/g-crystalline-like proteins, the majority of which are insoluble and form the PRISM matrix.31 Soluble mucocyst proteins can be easily removed by low-speed centrifugation, and recombinant proteins of interest harvested with the matrix itself. Methods have now been devised that allow rapid separation of recombinant proteins from PRISM, making this a highly streamlined purification system that is tantamount to performing affinity chromatography in vivo (Clark, et al, in preparation).

Figure 6. Recombinant H5 in PRISM induces antibody responses in animals. Adult rats were injected with the PRISM matrix containing recombinant H5 in different orientations relative to the granule targeting protein as described in the text. Resulting antisera (diluted 1:1,000) were then tested in Western blots against purified H5 expressed in insect cells (Protein Sciences). Lanes 1–3 correspond to serum samples 1–3, respectively, in Table 2. Signals corresponding to the H5 monomer and higher order aggregates were visible in each case following addition of secondary goat anti-rat IgG (1:3,000 dilution, HRP conjugate [SouthernBiotech]). No signals were obtained when the same protein was reacted with pooled nonimmune rat serum followed by secondary antibody (lane 4), or with secondary antibody alone (lane 5).

Although the number of animals used in this study was limited, we were surprised that the neutralizing antibody titers against PRISM-linked H5 reached such high levels (Table 2). Dense core mucocysts of T. thermophila are highly crystalline,29 and the PRISM matrix retains its underlying crystalline structure even after mucocyst discharge (Figure 4). We speculate that the arrangement of recombinant antigens within this crystalline array provides a repetitive framework that can strongly cross-link the Ig receptor on B-cells. Furthermore, the particulate nature of the gel itself could allow target proteins to be avidly taken up by antigen-presenting cells promoting strong T-cell responses as well. If these hypotheses are born out, PRISM may offer a powerful low-cost alternative to VLPs and other particle-based delivery systems for producing highly protective subunit vaccines.

Table 2. Virus neutralization titers of rat antisera against recombinant H5

Summary

With its distinct biological features, Tetrahymena thermophila has served as an important model for cellular biology and genetics for more than 50 years. At the same time, Tetrahymena has numerous advantages as an expression system for subunit vaccines. These include rapid cell growth to high cell densities, eukaryotic protein folding and PTM machinery, the absence of endogenous pyrogens, and active synthesis of membrane and secreted proteins. However, among currently available platforms, Tetrahymena is unique in its ability to secrete proteins in a tightly controlled fashion. As shown here, candidate vaccine antigens can be directed to a compartment for regulated secretion that allows rapid purification of recombinant protein in association with a crystalline matrix termed PRISM. The matrix itself appears to have immunostimulatory properties analogous to those of VLPs and can be produced in large quantities and at a very low cost.

Acknowledgements

We thank Daniel Kolbin and Lynn Almli for their excellent technical assistance. We also thank Ruben O. Donis, Molecular Virology and Vaccines Branch, Centers for Disease Control and Prevention, for his generous assistance with viral neutralization assays. This work was supported by grants from the National Institutes of Health (1 R41 GMO73352-01A1) and the US Army ARO (W911NF-07-0107) to Tetragenetics Inc.

JYOTHI JAYARAM is a postdoctoral research fellow at the department of Nephrology, School of Medicine, Stanford University, Stanford, CA, ASHOT PAPOYAN is research scientist II, XUJIE ZHANG is a postdoctoral scientist, and PAUL COLUSSI is the head of protein expression, all at Tetragenetics Inc, Ithaca, NY, YELENA BISHARYAN is research scientist I, DONNA CASSIDY-HANLEY is senior research associate, and Judith A. Appleton is a professor, all at the department of Microbiology & Immunology, College of Veterinary Medicine, Cornell University, Ithaca, NY, LUCILLE GAGLIARDO is a technician IV, James A. Baker Institute for Animal Health, Cornell University, Ithaca, NY and THEODORE G. CLARK Theodore G. Clark is the director of Grad Studies in the Field of Immunology Department of Microbiology and Immunology, Cornell University, Ithaca, NY, 607.253.4042, tgc3@cornell.edu

References

1. Janeway CA Jr, Medzhitov R. Innate immune recognition. Annu Rev Immunol. 2002;20:197–216.

2. Gallucci S, Matzinger P. Danger signals: SOS to the immune system. Curr Opin Immunol. 2001;13:114–19.

3. Kwissa M, Kasturi SP, Pulendran B. The science of adjuvants. Expert Rev Vaccines. 2007;6:673–84.

4. Jonasson P, Lilijeqvist S, Nygren PA, Stahl S. Genetic design for facilitated production and recovery of recombinant proteins in Escherichia coli. Biotechnol Appl Biochem. 2002;35:91–105.

5. Nwe N, He Q, Damrongwatanapokin S, Du Q, Manopo I, Limlamthong Y, et al. Expression of hemagglutinin protein from the avian influenza virus H5N1 in a baculovirus/insect cell system significantly enhanced by suspension culture. BMC Microbiol. 2006;6:16.

6. Wang K, Holtz KM, Anderson K, Chubet R, Mahmoud W, Cox MMJ. Expression and purification of an influenza hemagglutinin—one step closer to a recombinant protein-based influenza vaccine. Vaccine. 2006;24:2176–185.

7. Allgaier S, Taylor RD, Brudnaya Y, Jacobson DJ, Cambareri E, Stuart WD. Vaccine production in Neurospora crassa. Biologicals. 2009;37:128–32.

8. Noad R, Roy R. Virus-like particles as immunogens. Trends in Microbiol. 2003;11:438–44.

9. Buonaguro L, Tornesello ML, Buonaguro FM. Virus-like particle as particulate vaccines. Curr HIV Res. 2010;8: 299–309.

10. Taniguchi T, Mizuochi T, Banno Y, Nozawa Y, Kobata A. Carbohydrates of lysosomal enzymes secreted by Tetrahymena pyriformis. J Biol Chem. 1985;260:13941–46.

11. Weide T, Herrmann L, Bockau U, Niebur N, Aldag I, Laroy W, et al. Secretion of functional human enzymes by Tetrahymena thermophila. BMC Biotechnol. 2006;16:19–27.

12. Verhey KJ, Gaertig J. The tubulin code. Cell Cycle 6. 2007;2152–160.

13. Garcia BA, Hake SB, Diaz RL, Kauer M, Morris SA, Recht J, et al. Organismal differences in post-translational modifications in histones H3 and H4. J Biol Chem. 2007;282:7641–55.

14. Turkewitz AP, Chilcoat ND, Haddad A, Verbsky JW. Regulated protein secretion in Tetrahymena thermophila. Meth Cell Biol. 2000;62:347–62.

15. Turkewitz AP. Out with a Bang! Tetrahymena as a model system to study secretory granule biogenesis. Traffic 5. 2004;63–68.

16. Karrer KM. Tetrahymena genetics: Two nuclei are better than one. Methods Cell Biol. 2000;62:127–186.

17. Shang Y, Song X, Bowen J, Corstanje R, Gao Y, Gaertig J, Gorovsky MA. A robust inducible-repressible promoter greatly facilitates gene knockouts, conditional expression, and overexpression of homologous and heterologous genes in Tetrahymena thermophila. Proc Nat Acad Sci (USA). 2002;99:3734–39.

18. Bruns PJ, Cassidy-Hanley D. Biolistic transformation of macro- and micronuclei. Methods Cell Biol. 2000;62:501–12.

19. Galfre G, Milstein C. Preparation of monoclonal antibodies: strategies and procedures. Methods Enzymol. 1981;73: 3–46.

20. Barrett T, Inglis SC. Growth purification and titration of influenza viruses. In: Mahy BWJ, ed. Virology: A practical approach. London: IRL Press, Oxford University Press; 1985. p. 119–50.

21. Appleton JA, Antczak DF, Lopes AD. Characterization of the equine influenza virus H3 with monoclonal antibodies. Arch Virol. 1987;97:339–46.

22. Sambrook J, Fritsch EFT. Maniatis. Molecular cloning: a laboratory manual, 2nd edn. Cold Spring Harbor Laboratory Press; 1989.

23. Lin TL, Clark TG, Dickerson HW. Passive immunization of channel catfish (Ictalurus punctatus) against the ciliated protozoan parasite Ichthyophthirius multifiliis by use of murine monoclonal antibodies. Infect Immun. 1996;64:4085–90.

24. Rowe T, Abernathy RA, Hu-Primmer J, Thompson WW, Lu X, Lim WK, et al. Detection of antibody to avian influenza A (H5N1) virus in human serum by using a combination of serologic assays. J Clin Microbiol. 1999;37:937–43.

25. Clark TG, Lin TL, Jackwood DA, Sherrill J, Lin Y, Dickerson HW. The gene for an abundant parasite coat protein predicts tandemly repetitive metal binding domains. Gene. 1998;229:91–100.

26. Boldrin F, Santovito G, Formigari A, Bisharyan Y, Cassidy-Hanley D, Clark TG, Piccinni E. MTT2, a copper-inducible metallothionein gene from Tetrahymena thermophila. Comp Biochem Physiol C Toxicol Pharmacol. 2008;147:232–40.

27. Gething MJ, McCammon K, Sambrook J. Expression of wild-type and mutant forms of influenza hemagglutinin: The role of folding in intracellular transport. Cell. 1986;46:939–50.

28. Vanlandschoot E, Beiraert E, Grooen J, Min Jou W, Fiers W. pH-dependent aggregation and secretion of soluble monomeric influenza hemagglutinin. Arch Virol. 1998;143:227–39.

29. Hutton JC. Tetrahymena: the key to the genetic analysis of the regulated pathway of polypeptide secretion? Proc Nat Acad Sci USA. 1997;94:10490–10492.20.

30. Sambanis A, Stephanopoulos G, Sinskey AJ, HF Lodish. Use of regulated secretion in protein production from animal cells: an evaluation with the AtT-20 model cell line. Biotechnol Bioeng. 1990;35:771–80.21.

31. Bowman GR, Smith DG, Michael Siu KW, Pearlman RE, Turkewitz AP. Genetic and proteomic evidence for a second family of dense core granule cargo proteins in Tetrahymena thermophila. J Eukaryot Microbiol. 2005;52:291–97.