- Email: [email protected]
Polyclonal Fab phage display libraries with a high percentage of diverse clones to Cryptosporidium parvum glycoproteins
International Journal for Parasitology 33 (2003) 281–291 www.parasitology-online.com
Polyclonal Fab phage display libraries with a high percentage of diverse clones to Cryptosporidium parvum glycoproteins Liyan Chena, Brent R. Williamsa, Chiou-Ying Yangb, Ana Maria Cevallosc, Najma Bhatc, Honorine Wardc, Jacqueline Sharona,* a
Department of Pathology and Laboratory Medicine, Boston University School of Medicine, Boston, MA 02118, USA b Institute of Molecular Biology, National Chung Hsing University, Taichung 402, Taiwan c Division of Geographic Medicine and Infectious Diseases, New England Medical Center, Tufts University School of Medicine, Boston, MA 02111, USA Received 25 March 2002; received in revised form 9 December 2002; accepted 9 December 2002
Abstract The protozoan parasite Cryptosporidium parvum is regarded as a major public health problem world-wide, especially for immunocompromised individuals. Although no effective therapy is presently available, specific immune responses prevent or terminate cryptosporidiosis and passively administered antibodies have been found to reduce the severity of infection. Therefore, as an immunotherapeutic approach against cryptosporidiosis, we set out to develop C. parvum-specific polyclonal antibody libraries, standardised, perpetual mixtures of polyclonal antibodies, for which the genes are available. A combinatorial Fab phage display library was generated from the antibody variable region gene repertoire of mice immunised with C. parvum surface and apical complex glycoproteins which are believed to be involved in mediating C. parvum attachment and invasion. The variable region genes used to construct this starting library were shown to be diverse by nucleotide sequencing. The library was subjected to one round of antigen selection on C. parvum glycoproteins or a C. parvum oocyst/sporozoite preparation. The two selected libraries showed specific reactivity to the glycoproteins as well as to the oocyst/sporozoite preparation, with 50 – 73% antigen-reactive members. Fingerprint analysis of individual clones from the two antigenselected libraries showed high diversity, confirming the polyclonality of the selected libraries. Furthermore, immunoblot analysis on the oocyst/sporozoite and glycoprotein preparations with selected library phage showed reactivity to multiple bands, indicating diversity at the antigen level. These C. parvum-specific polyclonal Fab phage display libraries will be converted to libraries of polyclonal full-length antibodies by mass transfer of the selected heavy and light chain variable region gene pairs to a mammalian expression vector. Such polyclonal antibody libraries would be expected to mediate effector functions and provide optimal passive immunity against cryptosporidiosis. q 2003 Australian Society for Parasitology Inc. Published by Elsevier Science Ltd. All rights reserved. Keywords: Cryptosporidium parvum; Polyclonal antibody library; Passive immunotherapy; Phage display
1. Introduction The apicomplexan parasite Cryptosporidium parvum is currently recognised as a significant pathogen which infects intestinal epithelium and causes diarrheal disease in humans and animals world-wide (Griffiths, 1998; Fayer et al., 2000; Okhuysen and Chappell, 2002). Cryptosporidium parvum infection in immunocompetent hosts is frequently asymptomatic or manifests as an acute self-limiting diarrheal illness * Corresponding author. Boston University School of Medicine, 715 Albany Street, Room K707, Boston, MA 02118, USA. Tel.: þ 1-617-6384652; fax: þ 1-617-638-4079. E-mail address: [email protected] (J. Sharon).
(Griffiths, 1998). However, in individuals with impaired immune function, such as AIDS patients, C. parvum infection may cause severe and ultimately fatal disease (Griffiths, 1998). There is no approved therapy for cryptosporidiosis although recent advances in chemotherapy and immunotherapy have been reported (Blagburn and Soave, 1997; Crabb, 1998; Griffiths et al., 1998; Jenkins et al., 1999; Tzipori, 1998; Rossignol et al., 2001). Cryptosporidium infection is initiated by excystation of oocysts and release of sporozoites, which attach to and invade intestinal epithelial cells. Following invasion, intracellular development occurs within a parasitophorous vacuole, via asexual and sexual cycles. During the asexual
0020-7519/03/$30.00 q 2003 Australian Society for Parasitology Inc. Published by Elsevier Science Ltd. All rights reserved. doi:10.1016/S0020-7519(02)00282-5
282
L. Chen et al. / International Journal for Parasitology 33 (2003) 281–291
cycle, merozoites are released into the intestinal lumen from where they attach to and invade adjacent cells, thus perpetuating the infectious cycle. The stages of the parasite that would be the most effective targets of passively administered antibodies are the invasive ‘zoite’ (sporozoites and merozoites) stages while they are in the intestinal lumen. Like those of other apicomplexan parasites, zoites possess specialised secretory organelles such as rhoptries, micronemes and dense granules (collectively called the apical complex) which secrete and successively exocytose proteins that facilitate attachment, invasion, and parasitophorous vacuole formation (Dubremetz et al., 1998). One way in which passively administered antibodies could function would be to bind to zoite proteins involved in attachment and invasion and thereby block initial or ongoing infection. Cryptosporidium parvum surface and apical complex proteins which are implicated in mediating these processes include a . 900 kDa mucin-like glycoprotein named GP900 (Petersen et al., 1992; Barnes et al., 1998; Cevallos et al., 2000b), a 1,440 kDa circumsporozoite-like glycoprotein named CSL (Riggs et al., 1997; Langer and Riggs, 1996), 40– 45 kDa and 15 kDa mucinlike surface glycopeptides named gp40 (or gp45 or S45) and gp15 (or Cp17 or S16) (Cevallos et al., 2000a,b; Strong et al., 2000; Gut and Nelson, 1994; Priest et al., 2000; Winter et al., 2000) and a 23– 27 kDa surface protein named p23 (or p27) (Perryman et al., 1993, 1996, 1999; Enriquez and Riggs, 1998). A protective role for antibodies in cryptosporidiosis has been shown in humans as well as in animal models of this disease (Crabb, 1998; Heyworth, 1992; Tzipori et al., 1994; Zu et al., 1992; Perryman et al., 1999). Monoclonal antibodies against sporozoite antigens have been studied for their efficacy in immunotherapy (Crabb, 1998; Perryman et al., 1993) and have been shown to reduce infectivity in various murine models (Crabb, 1998; Enriquez and Riggs, 1998). However, targeting single antigenic determinants may allow ‘escape variants’ and result in ineffective therapy (Crabb, 1998). This disadvantage could be overcome by polyclonal antibodies. We have developed a system for generation of recombinant polyclonal antibody libraries (Sarantopoulos, S., 1998. Generation of recombinant polyclonal antibody libraries against tumor cells (Doctoral dissertation). Boston University School of Medicine, Boston, MA; Sharon et al., 2002). Polyclonal antibody libraries combine the advantages of conventional antiserum-derived polyclonal antibodies, targeting of multiple antigenic determinants, low likelihood of antigen ‘escape variants’, and efficient mediation of effector functions with the advantages of monoclonal antibodies (unlimited supply of standardised reagents and the availability of the genetic material for desired manipulations). This system involves cloning of variable (V) region gene repertoires of both heavy and light chains (VH and VL regions) into a phage display vector for Fab expression. The phage display technique has the
advantage of coupling a selectable function to the genetic material that encodes the function (Smith et al., 1990; Winter and Milstein, 1991; Barbas and Lerner, 1991). Thereby, we are able to select those Fabs that have reactivity to desired antigens. Those selected V region gene repertoires can then be transferred in mass to a mammalian expression vector to produce polyclonal antibody libraries of full-length IgG, IgA or IgM antibodies (Sharon et al., 2002). We report here the generation and selection of polyclonal Fab phage display libraries with specificity to C. parvum glycoproteins, and high diversity. These Fab libraries will be converted into libraries of full-length polyclonal antibodies, that could mediate effector functions, by mass transfer of the V region gene pairs to a mammalian expression vector and transfection into mammalian cells as described (Sharon et al., 2002).
2. Materials and methods 2.1. Antigen preparations and immunisation protocol Cryptosporidium parvum oocysts (GCH1 isolate) were maintained by serial passage in calves and provided to us by Dr Saul Tzipori, Tufts University School of Veterinary Medicine, N. Grafton, MA and stored at 4 8C till use. A C. parvum oocyst/sporozoite lysate was prepared by treating oocysts (1 £ 108/ml) with 1.75% sodium hypochlorite (Clorox) for 10 min on ice. After three washes with PBS, the oocysts were subjected to 20 freeze –thaw cycles in PBS containing protease inhibitors (2 mM PMSF, 20 mM leupeptin, 10 mM E64, 2 mM EDTA). A freeze – thaw cycle consisted of 5 min in liquid nitrogen followed by 5 min in a 37 8C water bath. The disappearance of intact oocysts was monitored by microscopic observation. The lysate obtained in this way (hereafter referred to as oocyst/sporozoite preparation) contained a suspension of oocyst shells together with soluble proteins and membrane particles from lysed oocysts and sporozoites and was stored in aliquots at 2 80 8C until use. Cryptosporidium parvum glycoproteins were purified by lectin (Helix pomatia agglutinin (HPA)) affinity chromatography as previously described (Cevallos et al., 2000a). Purified glycoproteins were emulsified with Freund’s complete adjuvant and used to immunise 3-week-old BALB/c mice by i.p. injection. Eight weeks later mice were boosted with purified glycoproteins in incomplete Freund’s adjuvant. Antibody production in serum (obtained following tail bleeds of mice) was monitored by immunoblotting of an oocyst/sporozoite preparation. Mice were housed in the animal facility of Tufts University, New England Medical Center. All studies involving mice were supervised by the Division of Laboratory Animal Medicine of the Tufts University School of Veterinary Medicine and approved by the Institutional
L. Chen et al. / International Journal for Parasitology 33 (2003) 281–291
283
Animal Care and Use Committee (IACUC) constituted to guarantee the protection of animals. Appropriate measures were taken to ensure that the mice were not subjected to undue discomfort or pain. A comfortable restraining device was used for blood drawing and i.p. immunisation of mice. Mice were sacrificed by CO2 inhalation.
10 min. Supernatants were removed with a multi-channel micropipetter and the bacterial pellets were washed once with 200 ml 2 £ TY (by resuspension and centrifugation). After removal of the supernatant, the bacterial pellets were resuspended in 200 ml each of 2 £ TY/carb/kan and incubated overnight at room temperature with shaking.
2.2. Construction of Fab phage display libraries and phage production
2.3. Antigen-specific selection of Fab phage display libraries
Construction of Fab phage display libraries was done as previously described (Sharon et al., 2002), with minor modifications. Briefly, RNA was prepared separately from spleen cells, bone marrow cells, intestine and nasopharynx of two immunised mice using TRIzol reagent (Life Technologies) as per the manufacturer’s instructions. Antibody V region genes were amplified from each RNA sample by reverse transcription followed by a two-step PCR. The PCR products were individually gel-purified and then pooled for each mouse before cloning into phagemid vector #622 phh3-stuff2 (Sharon et al., 2002) to generate two Fab phage display libraries. Ligations were transformed into supercompetent Escherichia coli TG1 (Stratagene), and transformations were plated on 2 £ TY agar plates (Sambrook et al., 1989) supplemented with 50 mg/ml carbenicillin and 1% glucose (2 £ TY/carb/glu plates). All colonies from each library were scraped with a rubber policeman into 2 £ TY/carb/glu, glycerol was added to 25% (v/v) and bacteria stored in aliquots at 2 80 8C as Mouse 1 and Mouse 2 library stocks. For preparation of library phage, equal portions of Mouse 1 and Mouse 2 library stocks were each plated on a 2 £ TY/carb/glu agar plate. The colonies from each plate were separately scraped into 2 £ TY/carb/glu liquid medium and phage prepared as described (Vest Hansen et al., 2001), with minor modifications. The bacteria were superinfected by VCSM13 (Stratagene) with a multiplicity of infection of 20 and selected in 2 £ TY supplemented with 50 mg/ml carbenicillin and 70 mg/ml kanamycin (2 £ TY/carb/kan). The phage particles were precipitated by adding 1/5th volume of 20% polyethylene glycol 6000/2.5 M NaCl and then resuspended in PBS. The two phage preparations were combined and the phage concentration was determined by titering for colony forming units (CFU). Preparation of phage from individual library clones was done as described (Pope et al., 1996), with minor modifications. Briefly, bacterial clones were grown in 60 ml of 2 £ TY/carb/glu in 96-well tissue culture plates (Costar) overnight at room temperature with shaking. The overnight cultures were diluted 100-fold and 100 ml of the diluted cultures was grown in 96-well UNIPLATEsw (Whatman) for 2 h at 37 8C. The bacterial cultures were then superinfected with VCSM13 (Stratagene, 50 ml containing 5 £ 108 plaque forming units per culture). After incubation at 37 8C for 30 min without shaking and 30 min with shaking, plates were centrifuged at 950 £ g for
The phage library was subjected to antigen selection separately on two C. parvum preparations using two different methods: (1) selection in an immunotube coated with C. parvum glycoproteins (immunotube method to generate library-glyco); and (2) selection on the C. parvum oocyst/sporozoite preparation followed by density gradient centrifugation to separate antigen-bound and free phage (suspension/gradient method to generate library-lysate). Selection on C. parvum glycoproteins in the immunotube was done as follows: a 5 ml MaxiSorp immunotube (Nunc) was coated with 4.5 ml of 0.5 mg/ml of C. parvum glycoproteins in 50 mM sodium carbonate buffer (pH 9.6) at 4 8C overnight. The immunotube was blocked with 5 ml of PBS supplemented with 2% non-fat dry milk and 0.05% Tween 20 (2% milk-PBST) at room temperature for 1 h. Phage (1 £ 109 CFU) were blocked in 400 ml of PBS supplemented with 4% non-fat dry milk and 0.05% Tween 20 (4% milk-PBST) for 1 h at room temperature, and then mixed with 400 ml of PBS/0.05% Tween 20 (PBST) and the mixture was added to the antigen-coated immunotube. The tube was sealed with parafilm and rotated end-over-end, at room temperature for 2 h, on a LABQUAKEw shaker (Barnsted/Thermolyne). The tube was washed four times with 5 ml PBST, 3 min per wash. Bound phage were eluted by adding 1 ml of 100 mM triethylamine followed by 10 min of end-over-end rotation at room temperature. After neutralisation with 0.5 ml of 1 M Tris – HCl (pH 6.8) the eluted phage were used to infect 10 ml of a log phase culture (OD600 0.5– 1.0) of TG1 bacteria (30 min at 37 8C without shaking followed by 30 min at 37 8C with shaking at 250 rev./min). After centrifugation and resuspension of the bacteria in 1 ml of 2 £ TY, a 2 ml aliquot was removed for CFU determination, and the remainder was plated on two 15 cm 2 £ TY/carb/glu agar plates. All colonies from both plates were scraped with a rubber policeman into 5 ml of 2 £ TY/carb/glu, glycerol was added to 25% (v/v), and the selected library was stored in aliquots at 2 80 8C. Selection on the C. parvum oocyst/sporozoite preparation (lysate) was done as follows: a 1.5 ml microfuge tube was preblocked with 2% milk-PBST for 1 h at room temperature. The oocyst/sporozoite preparation obtained from 2 £ 107 oocysts was centrifuged at 2,700 £ g for 2 min, washed three times with PBS (by resuspension and centrifugation) and then blocked for 1 h at room temperature in 400 ml of 2% milk-PBST. Phage (1 £ 1011 CFU) were preblocked in 400 ml of 2% milk-PBST at room temperature
284
L. Chen et al. / International Journal for Parasitology 33 (2003) 281–291
for 1 h. The blocked C. parvum suspension was combined with the blocked phage in the preblocked microfuge tube, and the mixture was rotated end-over-end at room temperature for 2 h. Density gradients of foetal bovine serum (FBS) and Percoll were prepared in 13 £ 89 mm polyallomer centrifuge tubes (Beckman) by placing 0.5 ml of 1.09 g/ml Percoll (prepared by dilution of Percolle density-gradient medium (Amersham Pharmacia Biotech ANTIBODY) with 10 £ PBS) at the bottom of the tube, overlaying with 3 ml of FBS, and overlaying the FBS with 1 ml of PBS (to facilitate loading the gradient without disturbing the serum layer). After the 2 h rotation, the lysate-phage mixture was loaded on a FBS/Percoll gradient, the polyallomer tube placed in a larger (50 ml) tube, and centrifuged for 3 min at 300 £ g in a swinging-bucket rotor. The tube was rotated 180 degrees and centrifuged for another 1 min to dislodge the oocyst/sporozoite preparation from the wall of the polyallomer tube. The polyallomer tube was attached to a ring stand and illuminated. A 21G needle attached to a 1 ml syringe was inserted into the tube just below the oocyst/sporozoite preparation layer (at the Percoll– serum interface), with the bevelled end of the needle facing up, and the oocyst/sporozoite preparation layer was sucked into the syringe (about 0.15 ml). The needle was then pierced through the opposite side of the polyallomer tube, and the contents of the syringe were expelled into a microfuge tube. With the needle still pierced through the polyallomer tube, the syringe was rinsed twice by aspirating PBS from a second (PBS-filled) microfuge tube, and the rinses were expelled into the first microfuge tube containing the oocyst/sporozoite preparation. After adjusting the total volume to 1 ml with PBS and gentle mixing by pipeting up and down three times, 0.5 ml of the recovered oocyst/sporozoite preparation was loaded on a second FBS/Percoll gradient and the oocyst/sporozoite preparation (with bound phage) was recovered from the Percoll– serum interface as described above. The oocyst/sporozoite preparation (with bound phage) was then pelleted by centrifugation at 2,700 £ g for 8 min, and the supernatant discarded. The pellet was resuspended by hitting the tube, 1 ml of 100 mM triethylamine was added, and the microfuge tube was rotated end-over-end at room temperature for 10 min. The oocyst/sporozoite preparation (with bound phage) was pelleted by centrifugation at 21,000 £ g for 3 min. The supernatant was removed, neutralised by transfer to a tube containing 0.5 ml 1 M Tris – HCl (pH 6.8), and used to infect 10 ml of a log phase culture of TG1 bacteria as described for the immunotube method above. (The 0.5 ml of recovered oocyst/sporozoite preparation from the first FBS/Percoll gradient, that was not loaded on the second FBS/Percoll gradient, was instead diluted into 40 ml of PBST and centrifuged for 12 min at 1,700 £ g to recover the pellet, which was then treated with triethylamine and used to infect TG1 bacteria as described above.)
2.4. ELISA EIA/RIA high binding polystyrene 96-well plates (Corning) were coated with 50 ml/well of 0.5 mg/ml of C. parvum glycoproteins or with 4 £ 105/50 ml per well of oocyst/sporozoite preparation in 50 mM sodium carbonate buffer (pH 9.6) overnight at 4 8C. Plates were then blocked for 1 h at room temperature with 2% milk-PBST. For ELISA with library phage, 100 ml of four-fold serial dilutions of phage (starting at 2 £ 1011 CFU) in 2% milkPBST was added to wells; for ELISA with phage clones, phage-containing bacterial supernatants from individual clones were diluted 1:1 with 4% milk-PBST and added at 100 ml/well. Plates were incubated with mild shaking for 1 h at room temperature, then washed 4 £ with PBST. Phage binding was detected by incubating with 100 ml/well of horseradish peroxidase-conjugated anti-M13 antibody (Amersham Pharmacia) at 1:5,000 in 2% milk-PBST for 1 h at room temperature, washing 4 £ with PBST and developing with 60 ml/well TMB substrate (Kirkegaard & Perry Labs) for 15 min at room temperature in the dark. The reaction was stopped with 60 ml/well of 0.2 M H2SO4, and absorbance at 450 nm was measured in a microplate reader. For the specificity control, plates were coated with 50 ml/well of 2.5 mg/ml of p-azophenylarsonate-conjugated BSA (Ars-BSA) in 50 mM sodium carbonate buffer (pH 9.6) overnight at 4 8C. ELISA was performed as described above. 2.5. Nucleotide sequence determination Second step V region PCR products were TA-cloned separately into TA vector pCR2.1 (Invitrogen) according to the manufacturer’s instructions. DNA from individual TA clones was prepared using a Qiagen miniprep kit (Qiagen). Nucleotide sequencing was performed by the Dana Farber Cancer Institute Molecular Biology Core Facility at Harvard University, Boston, MA, using the M13 forward primer (which hybridises to the lacZa fragment in vector pCR2.1) and the M13 reverse primer (which hybridises to the lac promoter in vector pCR2.1) (see Invitrogen catalogue for primer sequences). Sequence alignment was performed using the Sequencher 3.0 alignment program (Gene Codes Corp.). 2.6. Fingerprint analysis Fingerprinting of individual library clones was done using the restriction enzymes BstNI (Marks et al., 1991) and HinfI. Preparation of DNA from individual library clones was done as follows: 10 ml of overnight bacterial culture was transferred from each well of a 96-well plate to wells with 90 ml/well of H2O in a 96-well V-bottom microplate (MJ Research Inc.) and heated to 94 8C for 5 min in a thermocycler. After one freeze – thaw cycle, 10 ml samples were used for PCR amplification of Fd (VH-CH1) in a 96-
L. Chen et al. / International Journal for Parasitology 33 (2003) 281–291
well PCR plate (MJ Research Inc.) containing 50 ml/well PCR mix (1 £ Taq buffer, 2 mM MgCl2, 0.25 nM dNTP (each), 0.5 mM each forward and reverse primers, and 0.6 units Taq polymerase). For Fd, the forward primer was 50 GCTGCCGACCGCTGCTGCTGGTC-30 and the reverse primer was 50 -GCATTGACAGGAGGTTGAGGC-3 0 . Amplified Fd DNAs were cut with BstNI or HinfI, and electrophoresed in a 2% TBE agarose gel for fingerprint analysis.
285
determined by sequencing TA clones obtained from two of the VH PCR reactions. A total of 12 sequences (six from each mouse) were analysed. As shown in Fig. 2, all sequences were different. The differences were found in nucleotide sequence as well as CDR3 length, indicating that the VH region genes were derived from different B cell clones. The VH and VL region genes from each mouse were separately cloned into a phagemid vector to generate two original Fab phage display libraries, one of 6.3 £ 106 members and the other of 5.1 £ 107 members.
2.7. Immunoblot analysis 3.2. Antigen selection of Fab phage display libraries SDS-PAGE of C. parvum oocyst/sporozoite preparation (1 £ 106 per lane) or purified glycoproteins (2 mg per lane) was done in 4– 15% polyacrylamide gradient gels (for the oocyst/sporozoite preparation) or 10% polyacrylamide gels (for the glycoproteins) under non-reducing conditions, using SeeBlue Plus2 prestained standards (Invitrogen). Electrophoresed proteins were then transferred to nitrocellulose membranes. PBS supplemented with 10% non-fat dry milk and 0.05% Tween 20 was used for blocking and subsequent incubation steps with antibodies. After blocking, the membrane was cut into strips and each strip was treated sequentially with phage (approximately 5 £ 1011 CFU), anti-M13 monoclonal antibody (Amersham Pharmacia) at 1:3,000, and alkaline phosphatase-conjugated anti-mouse IgG (H&L) (Promega) at 1:3,000, with three PBST washes before the anti-M13 and anti-IgG additions. Finally, the strips were washed twice with PBST, once with TBST (Trisbuffered saline (TBS) supplemented with 0.05% Tween 20), and once with TBS, followed by development with alkaline phosphatase substrate (ProtoBlotwII, Promega).
3. Results 3.1. Generation of Fab phage display libraries To increase the chances of an antibody response to C. parvum proteins involved in attachment and invasion mice were immunised with a purified glycoprotein preparation previously shown to contain some of these proteins including GP900, gp40/45, gp15 (Cevallos et al., 2000a; Strong et al., 2000) and p23 (Ward, unpublished data). Two of the immunised mice, with strong antibody responses to C. parvum glycoproteins, were used to construct Fab phage display libraries. Total RNA was isolated from spleen, bone marrow, nasopharynx and intestine of the two mice. After cDNA synthesis by reverse transcription, the VH and VL region gene repertoires were amplified separately by a twostep PCR, using different primer sets as described (Sharon et al., 2002), resulting in 32 VL PCR reactions and 144 VH PCR reactions for each mouse. Representative final PCR products electrophoresed in a 0.8% TAE agarose gel are shown in Fig. 1. The diversity of the V region gene repertoires was
Preliminary experiments using 2 £ 1012 CFU as input phage and two sequential rounds of selection on C. parvum surface glycoproteins coated onto an immunotube resulted in very limited V region diversity. In an attempt to increase the diversity of selected libraries, we decided to perform a single selection round. In a further attempt to maximise library diversity and obtain a high number of antigenreactive clones in a single round, we performed selections on two C. parvum antigen preparations by two different methods: (1) the immunotube method using C. parvum glycoproteins (immunotube method); and (2) a density gradient centrifugation method using an oocyst/sporozoite preparation (the suspension/gradient method), which allows efficient separation of antigen-bound and free phage (Williams and Sharon, 2002). Preliminary data from our laboratory, with another antigen system, showed that two sequential gradient separations resulted in a higher recovery of antigen-binding phage; therefore, half of the input phage was subjected to a second gradient separation. However, in order to ensure recovery of enough output phage, the other half of the input phage was subjected to dilution in a large volume of PBST and bound phage were recovered from the pellet after centrifugation (see Section 2). Because the second gradient selection yielded a similar number of output CFU to the dilution selection, the output phage from the gradient/dilution selection was not analysed further. Despite the 100-fold higher number of CFU used in the suspension/ gradient method compared to the immunotube method (see Section 2), the antigen-selected libraries (designated ‘library-lysate’ and ‘library-glyco’, respectively) were similar in size (about 2 £ 105 CFU for each). 3.3. Antigen specificity of the selected libraries The binding specificity of the antigen-selected libraries was investigated by ELISA on C. parvum oocyst/sporozoite preparation or C. parvum glycoproteins. As shown in Fig. 3, both libraries were highly reactive, whereas the unselected library and a phage clone devoid of Fab (Negative control) showed no significant reactivity to either antigen preparation. A phage clone encoding an anti-p-azophenylarsonate (anti-Ars) Fab, used as specificity control, showed high reactivity on an Ars-coated plate but no reactivity on the
286
L. Chen et al. / International Journal for Parasitology 33 (2003) 281–291
Fig. 1. Ethidium bromide-stained agarose gels showing representative VH region gene PCR products derived from bone marrow RNA of an immunised mouse. The major band, expected size about 350 bp, was isolated from the gel separately and used for library production or TA cloning/sequencing. The primer pairs used for PCR amplification are indicated above each lane. Ladder: 100 bp DNA ladder (Bio-Rad).
oocyst/sporozoite preparation-coated plate or the C. parvum glycoproteins-coated plate. None of the other samples bound to the Ars-coated plate (Fig. 3), confirming the specificity of the assays. Twenty-two individual phage clones from each selected library were tested by ELISA for binding to the oocyst/sporozoite preparation and to purified C. parvum glycoproteins. A threshold value, shown by the dotted line in Fig. 4A, was set at 30% above the highest optical density value
obtained with the control clone encoding anti-Ars Fab. Clones showing optical density values below the threshold value were designated as non-binders and those above the threshold value were designated as binders. As shown in Fig. 4A, 11 (50%) of the clones from selected library-glyco and 16 (73%) of the clones from selected library-lysate were positive for binding to the C. parvum oocyst/sporozoite preparation. Of the positive clones, four (18%) and five (23%) of the selected library-glyco and the selected library-
Fig. 2. Diversity of VH region gene sequences generated in single final PCRs for each mouse using the primers indicated. Corresponding amino acid numbers (in the Kabat system; Kabat, 1995) are indicated. Complementarity determining regions 2 and 3 (CDR2 and CDR3) and framework regions 3 and 4 (FR3 and FR4) are labelled. Gaps indicate missing codons relative to other known sequences. Dots indicate additional nucleotides that are not shown.
L. Chen et al. / International Journal for Parasitology 33 (2003) 281–291
287
Fig. 3. Direct ELISA showing reactivity of phage from unselected, selected library-glyco, selected library-lysate, anti-Ars, and negative control (not displaying Fab) on plates coated with (A) C. parvum oocyst/sporozoite lysate and (B) C. parvum surface glycoproteins. Plates coated with Ars-BSA were used as specificity control (results shown in the right panels for each (A) and (B) phage preparations). Standard deviations of duplicate values are shown. The phage numbers used in all panels are shown in the top right panel.
lysate, respectively, showed very strong reactivity (OD450 . 2.5, see Fig. 4). ELISA on purified C. parvum surface glycoproteins also showed a high proportion of clones moderately or strongly bound to the antigen (Fig. 4B). 3.4. Diversity of the antigen-selected libraries The diversity of the antigen-selected libraries was assessed at the DNA level by fingerprint analysis of Fd regions (VH þ CH1), with two restriction enzymes (BstNI and HinfI). Because all clones have the same CH1 region genes, differences in the restriction patterns seen on gel analysis reflect sequence differences in the VH region genes. A total of 18 clones (10 from library-glyco and eight from library-lysate), picked from those positive clones (based on the ELISA results on the C. parvum oocyst/sporozoite preparation, selected to include all the high binding clones), were analysed. As seen in Fig. 5, the patterns of all 18 clones were different. Fingerprint analysis offers a
minimum estimate of diversity, because clones with the same fingerprint by the restriction enzymes used may still differ in nucleotide sequence. Because all the fingerprints in Fig. 5 were different, no nucleotide sequencing was necessary for the diversity analysis at the DNA level. To determine if the diversity is also manifested at the antigen level, immunoblot analysis on the oocyst/sporozoite preparation and on the purified glycoproteins was performed with phage of the selected libraries and with individual phage clones. Although Western blots with phage are difficult to perform consistently, the ones that worked showed that the selected libraries react with multiple bands in both the oocyst/sporozoite and glycoprotein preparations; in contrast, an unselected library phage prep did not show reactivity on either gel (see Fig. 6). Although one clone (No. 3) showed strong reactivity with the 15 kDa glycoprotein and some reactivity with the 900 kDa glycoprotein, several tested clones showed no significant reactivity, despite the strong binding of the same phage preps to ELISA plates
288
L. Chen et al. / International Journal for Parasitology 33 (2003) 281–291
Fig. 5. DNA fingerprint analysis of antibody Fd (VH-CH1) of individual clones (numbered) from antigen-selected library-glyco and antigenselected library-lysate. Fd DNA was amplified by PCR and digested with the restriction enzymes BstNI (top row) or HinfI (bottom row). M, molecular weight marker – 100 bp DNA ladder (Bio-Rad).
coated with the oocyst/sporozoite preparation (Fig. 6). Some faint reactivities by these clones are presumably due to non-specific phage binding.
4. Discussion
Fig. 4. ELISA showing reactivity of phage clones from library-glyco and library-lysate on plates coated with (A) C. parvum oocyst/sporozoite lysate (top panel) and Ars (bottom panel) and (B) C. parvum surface glycoproteins. In (A), a phage clone displaying anti-Ars Fab was used as specificity control and polyclonal phage from library-glyco was used as positive control (duplicate determinations). The dotted line in (A), set at 30% above the highest optical density value obtained with the anti-Ars control phage, was the cut-off point below which clones were designated ‘non-binders’.
In this study, a Fab phage display library was generated from mice immunised with a preparation of purified C. parvum surface glycoproteins. Selection of the library for binding to C. parvum by different methods resulted in two sub-libraries with high reactivity to C. parvum (Fig. 3) and a high percentage of diverse antigen binding clones (Figs. 4 – 6). We previously showed, in preliminary studies, the feasibility of constructing polyclonal Fab phage display libraries from mice immunised with an unfractionated C. parvum preparation (Baecher-Allan et al., 1999). In order to develop effective immunotherapy by targeting the initial host –parasite interactions of attachment and invasion, in the present study we sought to generate polyclonal antibody libraries directed against surface and apical complex glycoproteins that mediate these processes. To do this we
L. Chen et al. / International Journal for Parasitology 33 (2003) 281–291
Fig. 6. Immunoblot analysis on the oocyst/sporozoite preparation (4– 15% gel) and glycoproteins (10% gel) with phage of the selected libraries (library-glyco for the oocyst/sporozoite gel, and a mixture of library-glyco and library-lysate for the glycoproteins gel), an unselected library, and individual phage clones. The numbers of the phage clones correspond to the clone numbers from library-glyco or library-lysate shown for the fingerprint analysis in Fig. 5. The positions of prestained standards and the corresponding molecular weights are indicated. Values in parentheses refer to the ELISA reading (OD450) for each phage prep on a plate coated with the oocyst/sporozoite preparation. NA, not applicable.
immunised mice with a mixture of HPA-purified glycoproteins. These glycoproteins include GP900, gp40, gp15 and p23, all of which have been implicated in mediating C. parvum infection. To generate a diverse C. parvum-reactive library, it was critical to start with a heterogeneous repertoire of V region genes. This was achieved through use of low stringency conditions (37 8C) in the first PCR step to allow amplification of V region genes with mismatches to the V gene PCR primers (Sarantopoulos, S., 1998. Generation of recombinant polyclonal antibody libraries against tumor cells (Doctoral dissertation). Boston University School of Medicine, Boston, MA; Sharon et al., 2002). Nucleotide sequencing of 12 VH region genes (six per mouse from a single PCR product from a total of 144 VH PCRs per mouse) showed all 12 to be different (Fig. 2), confirming the diversity of the genes used to construct the starting Fab phage display libraries. The Fab phage display libraries were selected for binding
289
to the glycoprotein preparation used to immunise the mice as well as for binding to a suspension of lysed oocysts and sporozoites, the natural targets in human and animal infections. Both C. parvum-selected libraries bound to the glycoprotein preparation as well as to the C. parvum oocyst/sporozoite preparation (Fig. 3), showing the expected cross-reactivity between the two preparations. This is important because an optimal antibody library would be expected to control infection by binding to zoites within the intestinal lumen and preventing their attachment and invasion (Ward and Cevallos, 1998). Moreover, the crossreactivity demonstrated that the glycoprotein preparation could induce antibodies with reactivity to the unfractionated parasite preparation. The cross-reactivity of both libraries with the glycoprotein preparation and the oocyst/sporozoite preparation was also evident in the clonal ELISA (Fig. 4). It is interesting that, for the same clone, the signal was stronger in the plate coated with the glycoprotein preparation compared to the signal in the plate coated with the oocyst/sporozoite preparation. This could reflect a higher density of epitopes on the glycoproteins-coated ELISA plate compared with the epitope density on the parasite surface. It has been reported that although the percentage of antigen-reactive clones is low in the first selection round, library diversity decreases as the number of selection rounds increases (Hoogenboom et al., 1999; Lou et al., 2001). Results from our own laboratory led to the same conclusion regarding the decrease in diversity with increasing selection rounds. Therefore, in the present study, library selection was done in a single round. Also we used a suspension/gradient method to separate the bound and free phage, which was expected to be a milder separation compared with the washing steps in traditional selection, and thereby, keep those phage with lower affinity Fabs. It is important to have phage with a wide range of affinities to increase the diversity of the libraries. Clonal ELISA showed a high percentage of reactive clones (50% of clones from selected library-glyco and 73% of clones from selected library-lysate had antigen reactivity, Fig. 4). Furthermore, fingerprint analysis of individual antigen-reactive clones showed that no two are the same (Fig. 5), demonstrating a diverse V region gene repertoire. Limited immunoblot analysis indicated that the Fab phage display libraries are also diverse at the antigen level. This was shown by the binding of the selected libraries to multiple bands in both the oocyst/sporozoite and glycoprotein preparations (Fig. 6). The recognised bands were in the expected molecular weight range (Cevallos et al., 2000a) in the glycoprotein preparation. However, most clones tested did not react in the Western although they showed strong reactivity by ELISA (see Fig. 6). This may be due to epitope destruction in the denaturing gels, and many of the ‘Western-reactive’ library members may be recognising carbohydrate epitopes. The high percentage of antigen-reactive clones probably reflects the derivation of the starting Fab phage display
290
L. Chen et al. / International Journal for Parasitology 33 (2003) 281–291
libraries from immunised mice, and hence a high number of potential clones to be selected. Others have used naive (de Haard et al., 1999; Marks et al., 1991; Vaughan et al., 1996), semi-synthetic (de Kruif et al., 1995; Desiderio et al., 2001) or synthetic (Griffiths et al., 1994) libraries to generate Fv or Fab phage display libraries, which unavoidably result in a low proportion of antigen-reactive clones in the unselected and first round selected libraries. The V region gene pairs of the two C. parvum-reactive Fab phage display libraries described here will be transferred to a mammalian expression vector to obtain full-length IgG, IgA or IgM polyclonal antibody libraries (Sharon et al., 2002) that would be expected to mediate effector functions. Such polyclonal antibody libraries could be useful for both treatment and diagnosis of cryptosporidiosis.
Acknowledgements We thank Nils-Jakob Hansen, Margit Hansen, and Vincent Coljee for advice and discussion, Aletta Schnitzler for making the phage preps used in Western blots, Sanda Teodorescu-Frumosu for general lab maintenance and technical assistance with Western blots, and Smitha Jaison for technical assistance. This work was supported by grant AI40344 from National Institutes of Health to J.S. and H.W.
References Baecher-Allan, C.M., Santora, K.E., Sarantopoulos, S., Den, W., Sompuram, S.R., Cevallos, A.M., Bhat, N., Ward, H., Sharon, J., 1999. Generation of a polyclonal Fab phage display library to the protozoan parasite C. parvum. Comb. Chem. High Throughput Screening 2, 299–305. Barbas, C.F.I., Lerner, R.A., 1991. Combinatorial immunoglobulin libraries on the surface of phage (Phabs): rapid selection of antigen-specific Fabs. Methods: Comp. Methods Enzymol. 2, 119–124. Barnes, D.A., Bonnin, A., Huang, J.X., Gousset, L., Wu, J., Gut, J., Doyle, P., Dubremetz, J.F., Ward, H., Petersen, C., 1998. A novel multidomain mucin-like glycoprotein of C. parvum mediates invasion. Mol. Biochem. Parasitol. 96, 93– 110. Blagburn, B.L., Soave, R., 1997. Prophylaxis and chemotherapy: human and animal. In: Fayer, R., (Ed.), Cryptosporidium and Cryptosporidiosis, CRC Press, Boca Raton, FL, pp. 111– 128. Cevallos, A.M., Zhang, X., Waldor, M.K., Jaison, S., Zhou, X., Tzipori, S., Neutra, M.R., Ward, H.D., 2000a. Molecular cloning and expression of a gene encoding C. parvum glycoproteins gp40 and gp15. Infect. Immun. 68, 4108–4116. Cevallos, A.M., Bhat, N., Verdon, R., Hamer, D.H., Stein, B., Tzipori, S., Pereira, M.E., Keusch, G.T., Ward, H.D., 2000b. Mediation of C. parvum infection in vitro by mucin-like glycoproteins defined by a neutralizing monoclonal antibody. Infect. Immun. 68, 5167–5175. Crabb, J.H., 1998. Antibody-based immunotherapy of cryptosporidiosis. Adv. Parasitol. 40, 121–149. de Haard, H.J., van Neer, N., Reurs, A., Hufton, S.E., Roovers, R.C., Henderikx, P., de Bruine, A.P., Arends, J.W., Hoogenboom, H.R., 1999. A large non-immunized human Fab fragment phage library that
permits rapid isolation and kinetic analysis of high affinity antibodies. J. Biol. Chem. 274, 18218–18230. de Kruif, J., Terstappen, L., Boel, E., Logtenberg, T., 1995. Rapid selection of cell subpopulation-specific human monoclonal antibodies from a synthetic phage antibody library. Proc. Natl. Acad. Sci. USA 92, 3938–3942. Desiderio, A., Franconi, R., Lopez, M., Villani, M.E., Viti, F., Chiaraluce, R., Consalvi, V., Neri, D., Benvenuto, E., 2001. A semi-synthetic repertoire of intrinsically stable antibody fragments derived from a single-framework scaffold. J. Mol. Biol. 310, 603– 615. Dubremetz, J.F., Garcia-Reguet, N., Conseil, V., Fourmaux, M.N., 1998. Apical organelles and host-cell invasion by Apicomplexa. Int. J. Parasitol. 28, 1007–1013. Enriquez, F.J., Riggs, M.W., 1998. Role of immunoglobulin A monoclonal antibodies against P23 in controlling murine C. parvum infection. Infect. Immun. 66, 4469–4473. Fayer, R., Morgan, U., Upton, S., 2000. Epidemiology of Cryptosporidium: transmission, detection and identification. Int. J. Parasitol. 30, 1305–1322. Griffiths, A.D., Williams, S.C., Hartley, O., Tomlinson, I.M., Waterhouse, P., Crosby, W.L., Kontermann, R.E., Jones, P.T., Low, N.M., Allison, T.J., Prospero, T.D., Hoogenboom, H.R., Nissim, A., Cox, J.P.L., Harrison, J.L., Zaccolo, M., Gherardi, E., Winter, G., 1994. Isolation of high affinity human antibodies directly from large synthetic repertoires. EMBO J. 13, 3245–3260. Griffiths, J.K., 1998. Human cryptosporidiosis: epidemiology, transmission, clinical disease, treatment, and diagnosis. Adv. Parasitol. 40, 37 –85. Griffiths, J.K., Balakrishnan, R., Widmer, G., Tzipori, S., 1998. Paromomycin and geneticin inhibit intracellular C. parvum without trafficking through the host cell cytoplasm: implications for drug delivery. Infect. Immun. 66, 3874–3883. Gut, J., Nelson, R.G., 1994. C. parvum sporozoites deposit trails of 11A5 antigen during gliding locomotion and shed 11A5 antigen during invasion of MDCK cells in vitro. J. Eukaryot. Microbiol. 41, 42S–43S. Heyworth, M.F., 1992. Immunology of Giardia and Cryptosporidium infections. J. Infect. Dis. 166, 465 –472. Hoogenboom, H.R., Lutgerink, J.T., Pelsers, M.M., Rousch, M.J., Coote, J., Van Neer, N., De Bruine, A., Van Nieuwenhoven, F.A., Glatz, J.F., Arends, J.W., 1999. Selection-dominant and nonaccessible epitopes on cell-surface receptors revealed by cell-panning with a large phage antibody library. Eur. J. Biochem. 260, 774 –784. Jenkins, M.C., O’Brien, C., Trout, J., Guidry, A., Fayer, R., 1999. Hyperimmune bovine colostrum specific for recombinant C. parvum antigen confers partial protection against cryptosporidiosis in immunosuppressed adult mice. Vaccine 17, 2453–2460. Kabat, E.A., 1995. Kabat Database of Immunological Sequences, Rel. 5.0, updated weekly edition. Online address: www.ncbi.nlm.nih.gov. Langer, R.C., Riggs, M.W., 1996. Neutralizing monoclonal antibody protects against C. parvum infection by inhibiting sporozoite attachment and invasion. J. Eukaryot. Microbiol. 43, 76S–77S. Lou, J., Marzari, R., Verzillo, V., Ferrero, F., Pak, D., Sheng, M., Yang, C., Sblattero, D., Bradbury, A., 2001. Antibodies in haystacks: how selection strategy influences the outcome of selection from molecular diversity libraries. J. Immunol. Methods 253, 233 –242. Marks, J.D., Hoogenboom, H.R., Bonnert, T.P., McCafferty, J., Griffiths, A.D., Winter, G., 1991. By-passing immunization. Human antibodies from V-gene libraries displayed on phage. J. Mol. Biol. 222, 581 –597. Okhuysen, P.C., Chappell, C.L., 2002. Cryptosporidium virulence determinants – are we there yet? Int. J. Parasitol. 32, 517–525. Perryman, L.E., Kegerris, K.A., Mason, P.H., 1993. Effect of orally administered monoclonal antibody on persistent C. parvum infection in scid mice. Infect. Immun. 61, 4906–4908. Perryman, L.E., Jasmer, D.P., Riggs, M.W., Bohnet, S.G., McGuire, T.C., Arrowood, M.J., 1996. A cloned gene of C. parvum encodes neutralization-sensitive epitopes. Mol. Biochem. Parasitol. 80, 137 –147.
L. Chen et al. / International Journal for Parasitology 33 (2003) 281–291 Perryman, L.E., Kapil, S.J., Jones, M.L., Hunt, E.L., 1999. Protection of calves against cryptosporidiosis with immune bovine colostrum induced by a C. parvum recombinant protein. Vaccine 17, 2142–2149. Petersen, C., Gut, J., Doyle, P.S., Crabb, J.H., Nelson, R.G., Leech, J.H., 1992. Characterization of a .900,000-M(r) C. parvum sporozoite glycoprotein recognized by protective hyperimmune bovine colostral immunoglobulin. Infect. Immun. 60, 5132–5138. Pope, A.R., Embleton, M.J., Meranugh, R., 1996. Construction and use of antibody gene repertoires. In: McCafferty, J., Hoogenboom, H.R., Chisewell, D.J. (Eds.), Antibody Engineering, Oxford University Press, New York, pp. 1–40. Priest, J.W., Kwon, J.P., Arrowood, M.J., Lammie, P.J., 2000. Cloning of the immunodominant 17-kDa antigen from C. parvum. Mol. Biochem. Parasitol. 106, 261–271. Riggs, M.W., Stone, A.L., Yount, P.A., Langer, R.C., Arrowood, M.J., Bentley, D.L., 1997. Protective monoclonal antibody defines a circumsporozoite-like glycoprotein exoantigen of C. parvum sporozoites and merozoites. J. Immunol. 158, 1787– 1795. Rossignol, J.F., Ayoub, A., Ayers, M.S., 2001. Treatment of diarrhea caused by C. parvum: a prospective randomized, double-blind, placebocontrolled study of Nitazoxanide. J. Infect. Dis. 184, 103–106. Sambrook, J., Fritsch, E.F., Maniatis, T., 1989. Amplification and storage of cosmid libraries. In: Nolan, C., (Ed.), Molecular Cloning: A Laboratory Manual, 2nd Edition, Cold Spring Harbor Laboratory Press, Cold Spring Harbor, NY, pp. 3 –44. Sharon, J., Sompuram, S.R., Yang, C.-Y., Williams, B.R., Sarantopoulos, S., 2002. Construction of polyclonal antibody libraries using phage display. In: O’Brien, C., Aitken, R. (Eds.), Methods in Molecular Biology, Humana Press, Totowa, NJ, pp. 101– 112. Smith, D.F., Larsen, R.D., Mattox, S., Lowe, J.B., Cummings, R.D., 1990. Transfer and expression of a murine UDP-Gal:beta-D-Gal-alpha 1,3galactosyltransferase gene in transfected Chinese hamster ovary cells. Competition reactions between the alpha 1,3-galactosyltransferase and
291
the endogenous alpha 2,3-sialyltransferase. J. Biol. Chem. 265, 6225–6234. Strong, W.B., Gut, J., Nelson, R.G., 2000. Cloning and sequence analysis of a highly polymorphic C. parvum gene encoding a 60-kilodalton glycoprotein and characterization of its 15- and 45-kilodalton zoite surface antigen products. Infect. Immun. 68, 4117– 4134. Tzipori, S., 1998. Cryptosporidiosis: laboratory investigations and chemotherapy. Adv. Parasitol. 40, 187–221. Tzipori, S., Rand, W., Griffiths, J., Widmer, G., Crabb, J., 1994. Evaluation of an animal model system for cryptosporidiosis: therapeutic efficacy of paromomycin and hyperimmune bovine colostrum-immunoglobulin. Clin. Diagn. Lab. Immunol. 1, 450 –463. Vaughan, T.J., Williams, A.J., Pritchard, K., Osbourn, J.K., Pope, A.R., Earnshaw, J.C., McCafferty, J., Hodits, R.A., Wilton, J., Johnson, K.S., 1996. Human antibodies with sub-nanomolar affinities isolated from a large non-immunized phage display library. Nat. Biotechnol. 14, 309– 314. Vest Hansen, N., Ostergaard Pedersen, L., Stryhn, A., Buus, S., 2001. Phage display of peptide/major histocompatibility class I complexes. Eur. J. Immunol. 31, 32 –38. Ward, H., Cevallos, A.M., 1998. Cryptosporidium: molecular basis of hostparasite interaction. Adv. Parasitol. 40, 151–185. Williams, B.R., Sharon, J., 2002. Polyclonal anti-colorectal cancer Fab phage display library selected in one round using density gradient centrifugation to separate antigen-bound and free phage. Immunol. Lett. 81, 141–148. Winter, G., Milstein, C., 1991. Man-made antibodies. Nature 349, 293– 299. Winter, G., Gooley, A.A., Williams, K.L., Slade, M.B., 2000. Characterization of a major sporozoite surface glycoprotein of C. parvum. Funct. Integr. Genomics 1 (3), 207–217. Zu, S.-X., Fang, G.-D., Fayer, R., Guerrant, R.L., 1992. Cryptosporidiosis: pathogenesis and immunology. Parasitol. Today 8, 24–27.
Polyclonal Fab phage display libraries with a high percentage of diverse clones to Cryptosporidium parvum glycoproteins Liyan Chena, Brent R. Williamsa, Chiou-Ying Yangb, Ana Maria Cevallosc, Najma Bhatc, Honorine Wardc, Jacqueline Sharona,* a
Department of Pathology and Laboratory Medicine, Boston University School of Medicine, Boston, MA 02118, USA b Institute of Molecular Biology, National Chung Hsing University, Taichung 402, Taiwan c Division of Geographic Medicine and Infectious Diseases, New England Medical Center, Tufts University School of Medicine, Boston, MA 02111, USA Received 25 March 2002; received in revised form 9 December 2002; accepted 9 December 2002
Abstract The protozoan parasite Cryptosporidium parvum is regarded as a major public health problem world-wide, especially for immunocompromised individuals. Although no effective therapy is presently available, specific immune responses prevent or terminate cryptosporidiosis and passively administered antibodies have been found to reduce the severity of infection. Therefore, as an immunotherapeutic approach against cryptosporidiosis, we set out to develop C. parvum-specific polyclonal antibody libraries, standardised, perpetual mixtures of polyclonal antibodies, for which the genes are available. A combinatorial Fab phage display library was generated from the antibody variable region gene repertoire of mice immunised with C. parvum surface and apical complex glycoproteins which are believed to be involved in mediating C. parvum attachment and invasion. The variable region genes used to construct this starting library were shown to be diverse by nucleotide sequencing. The library was subjected to one round of antigen selection on C. parvum glycoproteins or a C. parvum oocyst/sporozoite preparation. The two selected libraries showed specific reactivity to the glycoproteins as well as to the oocyst/sporozoite preparation, with 50 – 73% antigen-reactive members. Fingerprint analysis of individual clones from the two antigenselected libraries showed high diversity, confirming the polyclonality of the selected libraries. Furthermore, immunoblot analysis on the oocyst/sporozoite and glycoprotein preparations with selected library phage showed reactivity to multiple bands, indicating diversity at the antigen level. These C. parvum-specific polyclonal Fab phage display libraries will be converted to libraries of polyclonal full-length antibodies by mass transfer of the selected heavy and light chain variable region gene pairs to a mammalian expression vector. Such polyclonal antibody libraries would be expected to mediate effector functions and provide optimal passive immunity against cryptosporidiosis. q 2003 Australian Society for Parasitology Inc. Published by Elsevier Science Ltd. All rights reserved. Keywords: Cryptosporidium parvum; Polyclonal antibody library; Passive immunotherapy; Phage display
1. Introduction The apicomplexan parasite Cryptosporidium parvum is currently recognised as a significant pathogen which infects intestinal epithelium and causes diarrheal disease in humans and animals world-wide (Griffiths, 1998; Fayer et al., 2000; Okhuysen and Chappell, 2002). Cryptosporidium parvum infection in immunocompetent hosts is frequently asymptomatic or manifests as an acute self-limiting diarrheal illness * Corresponding author. Boston University School of Medicine, 715 Albany Street, Room K707, Boston, MA 02118, USA. Tel.: þ 1-617-6384652; fax: þ 1-617-638-4079. E-mail address: [email protected] (J. Sharon).
(Griffiths, 1998). However, in individuals with impaired immune function, such as AIDS patients, C. parvum infection may cause severe and ultimately fatal disease (Griffiths, 1998). There is no approved therapy for cryptosporidiosis although recent advances in chemotherapy and immunotherapy have been reported (Blagburn and Soave, 1997; Crabb, 1998; Griffiths et al., 1998; Jenkins et al., 1999; Tzipori, 1998; Rossignol et al., 2001). Cryptosporidium infection is initiated by excystation of oocysts and release of sporozoites, which attach to and invade intestinal epithelial cells. Following invasion, intracellular development occurs within a parasitophorous vacuole, via asexual and sexual cycles. During the asexual
0020-7519/03/$30.00 q 2003 Australian Society for Parasitology Inc. Published by Elsevier Science Ltd. All rights reserved. doi:10.1016/S0020-7519(02)00282-5
282
L. Chen et al. / International Journal for Parasitology 33 (2003) 281–291
cycle, merozoites are released into the intestinal lumen from where they attach to and invade adjacent cells, thus perpetuating the infectious cycle. The stages of the parasite that would be the most effective targets of passively administered antibodies are the invasive ‘zoite’ (sporozoites and merozoites) stages while they are in the intestinal lumen. Like those of other apicomplexan parasites, zoites possess specialised secretory organelles such as rhoptries, micronemes and dense granules (collectively called the apical complex) which secrete and successively exocytose proteins that facilitate attachment, invasion, and parasitophorous vacuole formation (Dubremetz et al., 1998). One way in which passively administered antibodies could function would be to bind to zoite proteins involved in attachment and invasion and thereby block initial or ongoing infection. Cryptosporidium parvum surface and apical complex proteins which are implicated in mediating these processes include a . 900 kDa mucin-like glycoprotein named GP900 (Petersen et al., 1992; Barnes et al., 1998; Cevallos et al., 2000b), a 1,440 kDa circumsporozoite-like glycoprotein named CSL (Riggs et al., 1997; Langer and Riggs, 1996), 40– 45 kDa and 15 kDa mucinlike surface glycopeptides named gp40 (or gp45 or S45) and gp15 (or Cp17 or S16) (Cevallos et al., 2000a,b; Strong et al., 2000; Gut and Nelson, 1994; Priest et al., 2000; Winter et al., 2000) and a 23– 27 kDa surface protein named p23 (or p27) (Perryman et al., 1993, 1996, 1999; Enriquez and Riggs, 1998). A protective role for antibodies in cryptosporidiosis has been shown in humans as well as in animal models of this disease (Crabb, 1998; Heyworth, 1992; Tzipori et al., 1994; Zu et al., 1992; Perryman et al., 1999). Monoclonal antibodies against sporozoite antigens have been studied for their efficacy in immunotherapy (Crabb, 1998; Perryman et al., 1993) and have been shown to reduce infectivity in various murine models (Crabb, 1998; Enriquez and Riggs, 1998). However, targeting single antigenic determinants may allow ‘escape variants’ and result in ineffective therapy (Crabb, 1998). This disadvantage could be overcome by polyclonal antibodies. We have developed a system for generation of recombinant polyclonal antibody libraries (Sarantopoulos, S., 1998. Generation of recombinant polyclonal antibody libraries against tumor cells (Doctoral dissertation). Boston University School of Medicine, Boston, MA; Sharon et al., 2002). Polyclonal antibody libraries combine the advantages of conventional antiserum-derived polyclonal antibodies, targeting of multiple antigenic determinants, low likelihood of antigen ‘escape variants’, and efficient mediation of effector functions with the advantages of monoclonal antibodies (unlimited supply of standardised reagents and the availability of the genetic material for desired manipulations). This system involves cloning of variable (V) region gene repertoires of both heavy and light chains (VH and VL regions) into a phage display vector for Fab expression. The phage display technique has the
advantage of coupling a selectable function to the genetic material that encodes the function (Smith et al., 1990; Winter and Milstein, 1991; Barbas and Lerner, 1991). Thereby, we are able to select those Fabs that have reactivity to desired antigens. Those selected V region gene repertoires can then be transferred in mass to a mammalian expression vector to produce polyclonal antibody libraries of full-length IgG, IgA or IgM antibodies (Sharon et al., 2002). We report here the generation and selection of polyclonal Fab phage display libraries with specificity to C. parvum glycoproteins, and high diversity. These Fab libraries will be converted into libraries of full-length polyclonal antibodies, that could mediate effector functions, by mass transfer of the V region gene pairs to a mammalian expression vector and transfection into mammalian cells as described (Sharon et al., 2002).
2. Materials and methods 2.1. Antigen preparations and immunisation protocol Cryptosporidium parvum oocysts (GCH1 isolate) were maintained by serial passage in calves and provided to us by Dr Saul Tzipori, Tufts University School of Veterinary Medicine, N. Grafton, MA and stored at 4 8C till use. A C. parvum oocyst/sporozoite lysate was prepared by treating oocysts (1 £ 108/ml) with 1.75% sodium hypochlorite (Clorox) for 10 min on ice. After three washes with PBS, the oocysts were subjected to 20 freeze –thaw cycles in PBS containing protease inhibitors (2 mM PMSF, 20 mM leupeptin, 10 mM E64, 2 mM EDTA). A freeze – thaw cycle consisted of 5 min in liquid nitrogen followed by 5 min in a 37 8C water bath. The disappearance of intact oocysts was monitored by microscopic observation. The lysate obtained in this way (hereafter referred to as oocyst/sporozoite preparation) contained a suspension of oocyst shells together with soluble proteins and membrane particles from lysed oocysts and sporozoites and was stored in aliquots at 2 80 8C until use. Cryptosporidium parvum glycoproteins were purified by lectin (Helix pomatia agglutinin (HPA)) affinity chromatography as previously described (Cevallos et al., 2000a). Purified glycoproteins were emulsified with Freund’s complete adjuvant and used to immunise 3-week-old BALB/c mice by i.p. injection. Eight weeks later mice were boosted with purified glycoproteins in incomplete Freund’s adjuvant. Antibody production in serum (obtained following tail bleeds of mice) was monitored by immunoblotting of an oocyst/sporozoite preparation. Mice were housed in the animal facility of Tufts University, New England Medical Center. All studies involving mice were supervised by the Division of Laboratory Animal Medicine of the Tufts University School of Veterinary Medicine and approved by the Institutional
L. Chen et al. / International Journal for Parasitology 33 (2003) 281–291
283
Animal Care and Use Committee (IACUC) constituted to guarantee the protection of animals. Appropriate measures were taken to ensure that the mice were not subjected to undue discomfort or pain. A comfortable restraining device was used for blood drawing and i.p. immunisation of mice. Mice were sacrificed by CO2 inhalation.
10 min. Supernatants were removed with a multi-channel micropipetter and the bacterial pellets were washed once with 200 ml 2 £ TY (by resuspension and centrifugation). After removal of the supernatant, the bacterial pellets were resuspended in 200 ml each of 2 £ TY/carb/kan and incubated overnight at room temperature with shaking.
2.2. Construction of Fab phage display libraries and phage production
2.3. Antigen-specific selection of Fab phage display libraries
Construction of Fab phage display libraries was done as previously described (Sharon et al., 2002), with minor modifications. Briefly, RNA was prepared separately from spleen cells, bone marrow cells, intestine and nasopharynx of two immunised mice using TRIzol reagent (Life Technologies) as per the manufacturer’s instructions. Antibody V region genes were amplified from each RNA sample by reverse transcription followed by a two-step PCR. The PCR products were individually gel-purified and then pooled for each mouse before cloning into phagemid vector #622 phh3-stuff2 (Sharon et al., 2002) to generate two Fab phage display libraries. Ligations were transformed into supercompetent Escherichia coli TG1 (Stratagene), and transformations were plated on 2 £ TY agar plates (Sambrook et al., 1989) supplemented with 50 mg/ml carbenicillin and 1% glucose (2 £ TY/carb/glu plates). All colonies from each library were scraped with a rubber policeman into 2 £ TY/carb/glu, glycerol was added to 25% (v/v) and bacteria stored in aliquots at 2 80 8C as Mouse 1 and Mouse 2 library stocks. For preparation of library phage, equal portions of Mouse 1 and Mouse 2 library stocks were each plated on a 2 £ TY/carb/glu agar plate. The colonies from each plate were separately scraped into 2 £ TY/carb/glu liquid medium and phage prepared as described (Vest Hansen et al., 2001), with minor modifications. The bacteria were superinfected by VCSM13 (Stratagene) with a multiplicity of infection of 20 and selected in 2 £ TY supplemented with 50 mg/ml carbenicillin and 70 mg/ml kanamycin (2 £ TY/carb/kan). The phage particles were precipitated by adding 1/5th volume of 20% polyethylene glycol 6000/2.5 M NaCl and then resuspended in PBS. The two phage preparations were combined and the phage concentration was determined by titering for colony forming units (CFU). Preparation of phage from individual library clones was done as described (Pope et al., 1996), with minor modifications. Briefly, bacterial clones were grown in 60 ml of 2 £ TY/carb/glu in 96-well tissue culture plates (Costar) overnight at room temperature with shaking. The overnight cultures were diluted 100-fold and 100 ml of the diluted cultures was grown in 96-well UNIPLATEsw (Whatman) for 2 h at 37 8C. The bacterial cultures were then superinfected with VCSM13 (Stratagene, 50 ml containing 5 £ 108 plaque forming units per culture). After incubation at 37 8C for 30 min without shaking and 30 min with shaking, plates were centrifuged at 950 £ g for
The phage library was subjected to antigen selection separately on two C. parvum preparations using two different methods: (1) selection in an immunotube coated with C. parvum glycoproteins (immunotube method to generate library-glyco); and (2) selection on the C. parvum oocyst/sporozoite preparation followed by density gradient centrifugation to separate antigen-bound and free phage (suspension/gradient method to generate library-lysate). Selection on C. parvum glycoproteins in the immunotube was done as follows: a 5 ml MaxiSorp immunotube (Nunc) was coated with 4.5 ml of 0.5 mg/ml of C. parvum glycoproteins in 50 mM sodium carbonate buffer (pH 9.6) at 4 8C overnight. The immunotube was blocked with 5 ml of PBS supplemented with 2% non-fat dry milk and 0.05% Tween 20 (2% milk-PBST) at room temperature for 1 h. Phage (1 £ 109 CFU) were blocked in 400 ml of PBS supplemented with 4% non-fat dry milk and 0.05% Tween 20 (4% milk-PBST) for 1 h at room temperature, and then mixed with 400 ml of PBS/0.05% Tween 20 (PBST) and the mixture was added to the antigen-coated immunotube. The tube was sealed with parafilm and rotated end-over-end, at room temperature for 2 h, on a LABQUAKEw shaker (Barnsted/Thermolyne). The tube was washed four times with 5 ml PBST, 3 min per wash. Bound phage were eluted by adding 1 ml of 100 mM triethylamine followed by 10 min of end-over-end rotation at room temperature. After neutralisation with 0.5 ml of 1 M Tris – HCl (pH 6.8) the eluted phage were used to infect 10 ml of a log phase culture (OD600 0.5– 1.0) of TG1 bacteria (30 min at 37 8C without shaking followed by 30 min at 37 8C with shaking at 250 rev./min). After centrifugation and resuspension of the bacteria in 1 ml of 2 £ TY, a 2 ml aliquot was removed for CFU determination, and the remainder was plated on two 15 cm 2 £ TY/carb/glu agar plates. All colonies from both plates were scraped with a rubber policeman into 5 ml of 2 £ TY/carb/glu, glycerol was added to 25% (v/v), and the selected library was stored in aliquots at 2 80 8C. Selection on the C. parvum oocyst/sporozoite preparation (lysate) was done as follows: a 1.5 ml microfuge tube was preblocked with 2% milk-PBST for 1 h at room temperature. The oocyst/sporozoite preparation obtained from 2 £ 107 oocysts was centrifuged at 2,700 £ g for 2 min, washed three times with PBS (by resuspension and centrifugation) and then blocked for 1 h at room temperature in 400 ml of 2% milk-PBST. Phage (1 £ 1011 CFU) were preblocked in 400 ml of 2% milk-PBST at room temperature
284
L. Chen et al. / International Journal for Parasitology 33 (2003) 281–291
for 1 h. The blocked C. parvum suspension was combined with the blocked phage in the preblocked microfuge tube, and the mixture was rotated end-over-end at room temperature for 2 h. Density gradients of foetal bovine serum (FBS) and Percoll were prepared in 13 £ 89 mm polyallomer centrifuge tubes (Beckman) by placing 0.5 ml of 1.09 g/ml Percoll (prepared by dilution of Percolle density-gradient medium (Amersham Pharmacia Biotech ANTIBODY) with 10 £ PBS) at the bottom of the tube, overlaying with 3 ml of FBS, and overlaying the FBS with 1 ml of PBS (to facilitate loading the gradient without disturbing the serum layer). After the 2 h rotation, the lysate-phage mixture was loaded on a FBS/Percoll gradient, the polyallomer tube placed in a larger (50 ml) tube, and centrifuged for 3 min at 300 £ g in a swinging-bucket rotor. The tube was rotated 180 degrees and centrifuged for another 1 min to dislodge the oocyst/sporozoite preparation from the wall of the polyallomer tube. The polyallomer tube was attached to a ring stand and illuminated. A 21G needle attached to a 1 ml syringe was inserted into the tube just below the oocyst/sporozoite preparation layer (at the Percoll– serum interface), with the bevelled end of the needle facing up, and the oocyst/sporozoite preparation layer was sucked into the syringe (about 0.15 ml). The needle was then pierced through the opposite side of the polyallomer tube, and the contents of the syringe were expelled into a microfuge tube. With the needle still pierced through the polyallomer tube, the syringe was rinsed twice by aspirating PBS from a second (PBS-filled) microfuge tube, and the rinses were expelled into the first microfuge tube containing the oocyst/sporozoite preparation. After adjusting the total volume to 1 ml with PBS and gentle mixing by pipeting up and down three times, 0.5 ml of the recovered oocyst/sporozoite preparation was loaded on a second FBS/Percoll gradient and the oocyst/sporozoite preparation (with bound phage) was recovered from the Percoll– serum interface as described above. The oocyst/sporozoite preparation (with bound phage) was then pelleted by centrifugation at 2,700 £ g for 8 min, and the supernatant discarded. The pellet was resuspended by hitting the tube, 1 ml of 100 mM triethylamine was added, and the microfuge tube was rotated end-over-end at room temperature for 10 min. The oocyst/sporozoite preparation (with bound phage) was pelleted by centrifugation at 21,000 £ g for 3 min. The supernatant was removed, neutralised by transfer to a tube containing 0.5 ml 1 M Tris – HCl (pH 6.8), and used to infect 10 ml of a log phase culture of TG1 bacteria as described for the immunotube method above. (The 0.5 ml of recovered oocyst/sporozoite preparation from the first FBS/Percoll gradient, that was not loaded on the second FBS/Percoll gradient, was instead diluted into 40 ml of PBST and centrifuged for 12 min at 1,700 £ g to recover the pellet, which was then treated with triethylamine and used to infect TG1 bacteria as described above.)
2.4. ELISA EIA/RIA high binding polystyrene 96-well plates (Corning) were coated with 50 ml/well of 0.5 mg/ml of C. parvum glycoproteins or with 4 £ 105/50 ml per well of oocyst/sporozoite preparation in 50 mM sodium carbonate buffer (pH 9.6) overnight at 4 8C. Plates were then blocked for 1 h at room temperature with 2% milk-PBST. For ELISA with library phage, 100 ml of four-fold serial dilutions of phage (starting at 2 £ 1011 CFU) in 2% milkPBST was added to wells; for ELISA with phage clones, phage-containing bacterial supernatants from individual clones were diluted 1:1 with 4% milk-PBST and added at 100 ml/well. Plates were incubated with mild shaking for 1 h at room temperature, then washed 4 £ with PBST. Phage binding was detected by incubating with 100 ml/well of horseradish peroxidase-conjugated anti-M13 antibody (Amersham Pharmacia) at 1:5,000 in 2% milk-PBST for 1 h at room temperature, washing 4 £ with PBST and developing with 60 ml/well TMB substrate (Kirkegaard & Perry Labs) for 15 min at room temperature in the dark. The reaction was stopped with 60 ml/well of 0.2 M H2SO4, and absorbance at 450 nm was measured in a microplate reader. For the specificity control, plates were coated with 50 ml/well of 2.5 mg/ml of p-azophenylarsonate-conjugated BSA (Ars-BSA) in 50 mM sodium carbonate buffer (pH 9.6) overnight at 4 8C. ELISA was performed as described above. 2.5. Nucleotide sequence determination Second step V region PCR products were TA-cloned separately into TA vector pCR2.1 (Invitrogen) according to the manufacturer’s instructions. DNA from individual TA clones was prepared using a Qiagen miniprep kit (Qiagen). Nucleotide sequencing was performed by the Dana Farber Cancer Institute Molecular Biology Core Facility at Harvard University, Boston, MA, using the M13 forward primer (which hybridises to the lacZa fragment in vector pCR2.1) and the M13 reverse primer (which hybridises to the lac promoter in vector pCR2.1) (see Invitrogen catalogue for primer sequences). Sequence alignment was performed using the Sequencher 3.0 alignment program (Gene Codes Corp.). 2.6. Fingerprint analysis Fingerprinting of individual library clones was done using the restriction enzymes BstNI (Marks et al., 1991) and HinfI. Preparation of DNA from individual library clones was done as follows: 10 ml of overnight bacterial culture was transferred from each well of a 96-well plate to wells with 90 ml/well of H2O in a 96-well V-bottom microplate (MJ Research Inc.) and heated to 94 8C for 5 min in a thermocycler. After one freeze – thaw cycle, 10 ml samples were used for PCR amplification of Fd (VH-CH1) in a 96-
L. Chen et al. / International Journal for Parasitology 33 (2003) 281–291
well PCR plate (MJ Research Inc.) containing 50 ml/well PCR mix (1 £ Taq buffer, 2 mM MgCl2, 0.25 nM dNTP (each), 0.5 mM each forward and reverse primers, and 0.6 units Taq polymerase). For Fd, the forward primer was 50 GCTGCCGACCGCTGCTGCTGGTC-30 and the reverse primer was 50 -GCATTGACAGGAGGTTGAGGC-3 0 . Amplified Fd DNAs were cut with BstNI or HinfI, and electrophoresed in a 2% TBE agarose gel for fingerprint analysis.
285
determined by sequencing TA clones obtained from two of the VH PCR reactions. A total of 12 sequences (six from each mouse) were analysed. As shown in Fig. 2, all sequences were different. The differences were found in nucleotide sequence as well as CDR3 length, indicating that the VH region genes were derived from different B cell clones. The VH and VL region genes from each mouse were separately cloned into a phagemid vector to generate two original Fab phage display libraries, one of 6.3 £ 106 members and the other of 5.1 £ 107 members.
2.7. Immunoblot analysis 3.2. Antigen selection of Fab phage display libraries SDS-PAGE of C. parvum oocyst/sporozoite preparation (1 £ 106 per lane) or purified glycoproteins (2 mg per lane) was done in 4– 15% polyacrylamide gradient gels (for the oocyst/sporozoite preparation) or 10% polyacrylamide gels (for the glycoproteins) under non-reducing conditions, using SeeBlue Plus2 prestained standards (Invitrogen). Electrophoresed proteins were then transferred to nitrocellulose membranes. PBS supplemented with 10% non-fat dry milk and 0.05% Tween 20 was used for blocking and subsequent incubation steps with antibodies. After blocking, the membrane was cut into strips and each strip was treated sequentially with phage (approximately 5 £ 1011 CFU), anti-M13 monoclonal antibody (Amersham Pharmacia) at 1:3,000, and alkaline phosphatase-conjugated anti-mouse IgG (H&L) (Promega) at 1:3,000, with three PBST washes before the anti-M13 and anti-IgG additions. Finally, the strips were washed twice with PBST, once with TBST (Trisbuffered saline (TBS) supplemented with 0.05% Tween 20), and once with TBS, followed by development with alkaline phosphatase substrate (ProtoBlotwII, Promega).
3. Results 3.1. Generation of Fab phage display libraries To increase the chances of an antibody response to C. parvum proteins involved in attachment and invasion mice were immunised with a purified glycoprotein preparation previously shown to contain some of these proteins including GP900, gp40/45, gp15 (Cevallos et al., 2000a; Strong et al., 2000) and p23 (Ward, unpublished data). Two of the immunised mice, with strong antibody responses to C. parvum glycoproteins, were used to construct Fab phage display libraries. Total RNA was isolated from spleen, bone marrow, nasopharynx and intestine of the two mice. After cDNA synthesis by reverse transcription, the VH and VL region gene repertoires were amplified separately by a twostep PCR, using different primer sets as described (Sharon et al., 2002), resulting in 32 VL PCR reactions and 144 VH PCR reactions for each mouse. Representative final PCR products electrophoresed in a 0.8% TAE agarose gel are shown in Fig. 1. The diversity of the V region gene repertoires was
Preliminary experiments using 2 £ 1012 CFU as input phage and two sequential rounds of selection on C. parvum surface glycoproteins coated onto an immunotube resulted in very limited V region diversity. In an attempt to increase the diversity of selected libraries, we decided to perform a single selection round. In a further attempt to maximise library diversity and obtain a high number of antigenreactive clones in a single round, we performed selections on two C. parvum antigen preparations by two different methods: (1) the immunotube method using C. parvum glycoproteins (immunotube method); and (2) a density gradient centrifugation method using an oocyst/sporozoite preparation (the suspension/gradient method), which allows efficient separation of antigen-bound and free phage (Williams and Sharon, 2002). Preliminary data from our laboratory, with another antigen system, showed that two sequential gradient separations resulted in a higher recovery of antigen-binding phage; therefore, half of the input phage was subjected to a second gradient separation. However, in order to ensure recovery of enough output phage, the other half of the input phage was subjected to dilution in a large volume of PBST and bound phage were recovered from the pellet after centrifugation (see Section 2). Because the second gradient selection yielded a similar number of output CFU to the dilution selection, the output phage from the gradient/dilution selection was not analysed further. Despite the 100-fold higher number of CFU used in the suspension/ gradient method compared to the immunotube method (see Section 2), the antigen-selected libraries (designated ‘library-lysate’ and ‘library-glyco’, respectively) were similar in size (about 2 £ 105 CFU for each). 3.3. Antigen specificity of the selected libraries The binding specificity of the antigen-selected libraries was investigated by ELISA on C. parvum oocyst/sporozoite preparation or C. parvum glycoproteins. As shown in Fig. 3, both libraries were highly reactive, whereas the unselected library and a phage clone devoid of Fab (Negative control) showed no significant reactivity to either antigen preparation. A phage clone encoding an anti-p-azophenylarsonate (anti-Ars) Fab, used as specificity control, showed high reactivity on an Ars-coated plate but no reactivity on the
286
L. Chen et al. / International Journal for Parasitology 33 (2003) 281–291
Fig. 1. Ethidium bromide-stained agarose gels showing representative VH region gene PCR products derived from bone marrow RNA of an immunised mouse. The major band, expected size about 350 bp, was isolated from the gel separately and used for library production or TA cloning/sequencing. The primer pairs used for PCR amplification are indicated above each lane. Ladder: 100 bp DNA ladder (Bio-Rad).
oocyst/sporozoite preparation-coated plate or the C. parvum glycoproteins-coated plate. None of the other samples bound to the Ars-coated plate (Fig. 3), confirming the specificity of the assays. Twenty-two individual phage clones from each selected library were tested by ELISA for binding to the oocyst/sporozoite preparation and to purified C. parvum glycoproteins. A threshold value, shown by the dotted line in Fig. 4A, was set at 30% above the highest optical density value
obtained with the control clone encoding anti-Ars Fab. Clones showing optical density values below the threshold value were designated as non-binders and those above the threshold value were designated as binders. As shown in Fig. 4A, 11 (50%) of the clones from selected library-glyco and 16 (73%) of the clones from selected library-lysate were positive for binding to the C. parvum oocyst/sporozoite preparation. Of the positive clones, four (18%) and five (23%) of the selected library-glyco and the selected library-
Fig. 2. Diversity of VH region gene sequences generated in single final PCRs for each mouse using the primers indicated. Corresponding amino acid numbers (in the Kabat system; Kabat, 1995) are indicated. Complementarity determining regions 2 and 3 (CDR2 and CDR3) and framework regions 3 and 4 (FR3 and FR4) are labelled. Gaps indicate missing codons relative to other known sequences. Dots indicate additional nucleotides that are not shown.
L. Chen et al. / International Journal for Parasitology 33 (2003) 281–291
287
Fig. 3. Direct ELISA showing reactivity of phage from unselected, selected library-glyco, selected library-lysate, anti-Ars, and negative control (not displaying Fab) on plates coated with (A) C. parvum oocyst/sporozoite lysate and (B) C. parvum surface glycoproteins. Plates coated with Ars-BSA were used as specificity control (results shown in the right panels for each (A) and (B) phage preparations). Standard deviations of duplicate values are shown. The phage numbers used in all panels are shown in the top right panel.
lysate, respectively, showed very strong reactivity (OD450 . 2.5, see Fig. 4). ELISA on purified C. parvum surface glycoproteins also showed a high proportion of clones moderately or strongly bound to the antigen (Fig. 4B). 3.4. Diversity of the antigen-selected libraries The diversity of the antigen-selected libraries was assessed at the DNA level by fingerprint analysis of Fd regions (VH þ CH1), with two restriction enzymes (BstNI and HinfI). Because all clones have the same CH1 region genes, differences in the restriction patterns seen on gel analysis reflect sequence differences in the VH region genes. A total of 18 clones (10 from library-glyco and eight from library-lysate), picked from those positive clones (based on the ELISA results on the C. parvum oocyst/sporozoite preparation, selected to include all the high binding clones), were analysed. As seen in Fig. 5, the patterns of all 18 clones were different. Fingerprint analysis offers a
minimum estimate of diversity, because clones with the same fingerprint by the restriction enzymes used may still differ in nucleotide sequence. Because all the fingerprints in Fig. 5 were different, no nucleotide sequencing was necessary for the diversity analysis at the DNA level. To determine if the diversity is also manifested at the antigen level, immunoblot analysis on the oocyst/sporozoite preparation and on the purified glycoproteins was performed with phage of the selected libraries and with individual phage clones. Although Western blots with phage are difficult to perform consistently, the ones that worked showed that the selected libraries react with multiple bands in both the oocyst/sporozoite and glycoprotein preparations; in contrast, an unselected library phage prep did not show reactivity on either gel (see Fig. 6). Although one clone (No. 3) showed strong reactivity with the 15 kDa glycoprotein and some reactivity with the 900 kDa glycoprotein, several tested clones showed no significant reactivity, despite the strong binding of the same phage preps to ELISA plates
288
L. Chen et al. / International Journal for Parasitology 33 (2003) 281–291
Fig. 5. DNA fingerprint analysis of antibody Fd (VH-CH1) of individual clones (numbered) from antigen-selected library-glyco and antigenselected library-lysate. Fd DNA was amplified by PCR and digested with the restriction enzymes BstNI (top row) or HinfI (bottom row). M, molecular weight marker – 100 bp DNA ladder (Bio-Rad).
coated with the oocyst/sporozoite preparation (Fig. 6). Some faint reactivities by these clones are presumably due to non-specific phage binding.
4. Discussion
Fig. 4. ELISA showing reactivity of phage clones from library-glyco and library-lysate on plates coated with (A) C. parvum oocyst/sporozoite lysate (top panel) and Ars (bottom panel) and (B) C. parvum surface glycoproteins. In (A), a phage clone displaying anti-Ars Fab was used as specificity control and polyclonal phage from library-glyco was used as positive control (duplicate determinations). The dotted line in (A), set at 30% above the highest optical density value obtained with the anti-Ars control phage, was the cut-off point below which clones were designated ‘non-binders’.
In this study, a Fab phage display library was generated from mice immunised with a preparation of purified C. parvum surface glycoproteins. Selection of the library for binding to C. parvum by different methods resulted in two sub-libraries with high reactivity to C. parvum (Fig. 3) and a high percentage of diverse antigen binding clones (Figs. 4 – 6). We previously showed, in preliminary studies, the feasibility of constructing polyclonal Fab phage display libraries from mice immunised with an unfractionated C. parvum preparation (Baecher-Allan et al., 1999). In order to develop effective immunotherapy by targeting the initial host –parasite interactions of attachment and invasion, in the present study we sought to generate polyclonal antibody libraries directed against surface and apical complex glycoproteins that mediate these processes. To do this we
L. Chen et al. / International Journal for Parasitology 33 (2003) 281–291
Fig. 6. Immunoblot analysis on the oocyst/sporozoite preparation (4– 15% gel) and glycoproteins (10% gel) with phage of the selected libraries (library-glyco for the oocyst/sporozoite gel, and a mixture of library-glyco and library-lysate for the glycoproteins gel), an unselected library, and individual phage clones. The numbers of the phage clones correspond to the clone numbers from library-glyco or library-lysate shown for the fingerprint analysis in Fig. 5. The positions of prestained standards and the corresponding molecular weights are indicated. Values in parentheses refer to the ELISA reading (OD450) for each phage prep on a plate coated with the oocyst/sporozoite preparation. NA, not applicable.
immunised mice with a mixture of HPA-purified glycoproteins. These glycoproteins include GP900, gp40, gp15 and p23, all of which have been implicated in mediating C. parvum infection. To generate a diverse C. parvum-reactive library, it was critical to start with a heterogeneous repertoire of V region genes. This was achieved through use of low stringency conditions (37 8C) in the first PCR step to allow amplification of V region genes with mismatches to the V gene PCR primers (Sarantopoulos, S., 1998. Generation of recombinant polyclonal antibody libraries against tumor cells (Doctoral dissertation). Boston University School of Medicine, Boston, MA; Sharon et al., 2002). Nucleotide sequencing of 12 VH region genes (six per mouse from a single PCR product from a total of 144 VH PCRs per mouse) showed all 12 to be different (Fig. 2), confirming the diversity of the genes used to construct the starting Fab phage display libraries. The Fab phage display libraries were selected for binding
289
to the glycoprotein preparation used to immunise the mice as well as for binding to a suspension of lysed oocysts and sporozoites, the natural targets in human and animal infections. Both C. parvum-selected libraries bound to the glycoprotein preparation as well as to the C. parvum oocyst/sporozoite preparation (Fig. 3), showing the expected cross-reactivity between the two preparations. This is important because an optimal antibody library would be expected to control infection by binding to zoites within the intestinal lumen and preventing their attachment and invasion (Ward and Cevallos, 1998). Moreover, the crossreactivity demonstrated that the glycoprotein preparation could induce antibodies with reactivity to the unfractionated parasite preparation. The cross-reactivity of both libraries with the glycoprotein preparation and the oocyst/sporozoite preparation was also evident in the clonal ELISA (Fig. 4). It is interesting that, for the same clone, the signal was stronger in the plate coated with the glycoprotein preparation compared to the signal in the plate coated with the oocyst/sporozoite preparation. This could reflect a higher density of epitopes on the glycoproteins-coated ELISA plate compared with the epitope density on the parasite surface. It has been reported that although the percentage of antigen-reactive clones is low in the first selection round, library diversity decreases as the number of selection rounds increases (Hoogenboom et al., 1999; Lou et al., 2001). Results from our own laboratory led to the same conclusion regarding the decrease in diversity with increasing selection rounds. Therefore, in the present study, library selection was done in a single round. Also we used a suspension/gradient method to separate the bound and free phage, which was expected to be a milder separation compared with the washing steps in traditional selection, and thereby, keep those phage with lower affinity Fabs. It is important to have phage with a wide range of affinities to increase the diversity of the libraries. Clonal ELISA showed a high percentage of reactive clones (50% of clones from selected library-glyco and 73% of clones from selected library-lysate had antigen reactivity, Fig. 4). Furthermore, fingerprint analysis of individual antigen-reactive clones showed that no two are the same (Fig. 5), demonstrating a diverse V region gene repertoire. Limited immunoblot analysis indicated that the Fab phage display libraries are also diverse at the antigen level. This was shown by the binding of the selected libraries to multiple bands in both the oocyst/sporozoite and glycoprotein preparations (Fig. 6). The recognised bands were in the expected molecular weight range (Cevallos et al., 2000a) in the glycoprotein preparation. However, most clones tested did not react in the Western although they showed strong reactivity by ELISA (see Fig. 6). This may be due to epitope destruction in the denaturing gels, and many of the ‘Western-reactive’ library members may be recognising carbohydrate epitopes. The high percentage of antigen-reactive clones probably reflects the derivation of the starting Fab phage display
290
L. Chen et al. / International Journal for Parasitology 33 (2003) 281–291
libraries from immunised mice, and hence a high number of potential clones to be selected. Others have used naive (de Haard et al., 1999; Marks et al., 1991; Vaughan et al., 1996), semi-synthetic (de Kruif et al., 1995; Desiderio et al., 2001) or synthetic (Griffiths et al., 1994) libraries to generate Fv or Fab phage display libraries, which unavoidably result in a low proportion of antigen-reactive clones in the unselected and first round selected libraries. The V region gene pairs of the two C. parvum-reactive Fab phage display libraries described here will be transferred to a mammalian expression vector to obtain full-length IgG, IgA or IgM polyclonal antibody libraries (Sharon et al., 2002) that would be expected to mediate effector functions. Such polyclonal antibody libraries could be useful for both treatment and diagnosis of cryptosporidiosis.
Acknowledgements We thank Nils-Jakob Hansen, Margit Hansen, and Vincent Coljee for advice and discussion, Aletta Schnitzler for making the phage preps used in Western blots, Sanda Teodorescu-Frumosu for general lab maintenance and technical assistance with Western blots, and Smitha Jaison for technical assistance. This work was supported by grant AI40344 from National Institutes of Health to J.S. and H.W.
References Baecher-Allan, C.M., Santora, K.E., Sarantopoulos, S., Den, W., Sompuram, S.R., Cevallos, A.M., Bhat, N., Ward, H., Sharon, J., 1999. Generation of a polyclonal Fab phage display library to the protozoan parasite C. parvum. Comb. Chem. High Throughput Screening 2, 299–305. Barbas, C.F.I., Lerner, R.A., 1991. Combinatorial immunoglobulin libraries on the surface of phage (Phabs): rapid selection of antigen-specific Fabs. Methods: Comp. Methods Enzymol. 2, 119–124. Barnes, D.A., Bonnin, A., Huang, J.X., Gousset, L., Wu, J., Gut, J., Doyle, P., Dubremetz, J.F., Ward, H., Petersen, C., 1998. A novel multidomain mucin-like glycoprotein of C. parvum mediates invasion. Mol. Biochem. Parasitol. 96, 93– 110. Blagburn, B.L., Soave, R., 1997. Prophylaxis and chemotherapy: human and animal. In: Fayer, R., (Ed.), Cryptosporidium and Cryptosporidiosis, CRC Press, Boca Raton, FL, pp. 111– 128. Cevallos, A.M., Zhang, X., Waldor, M.K., Jaison, S., Zhou, X., Tzipori, S., Neutra, M.R., Ward, H.D., 2000a. Molecular cloning and expression of a gene encoding C. parvum glycoproteins gp40 and gp15. Infect. Immun. 68, 4108–4116. Cevallos, A.M., Bhat, N., Verdon, R., Hamer, D.H., Stein, B., Tzipori, S., Pereira, M.E., Keusch, G.T., Ward, H.D., 2000b. Mediation of C. parvum infection in vitro by mucin-like glycoproteins defined by a neutralizing monoclonal antibody. Infect. Immun. 68, 5167–5175. Crabb, J.H., 1998. Antibody-based immunotherapy of cryptosporidiosis. Adv. Parasitol. 40, 121–149. de Haard, H.J., van Neer, N., Reurs, A., Hufton, S.E., Roovers, R.C., Henderikx, P., de Bruine, A.P., Arends, J.W., Hoogenboom, H.R., 1999. A large non-immunized human Fab fragment phage library that
permits rapid isolation and kinetic analysis of high affinity antibodies. J. Biol. Chem. 274, 18218–18230. de Kruif, J., Terstappen, L., Boel, E., Logtenberg, T., 1995. Rapid selection of cell subpopulation-specific human monoclonal antibodies from a synthetic phage antibody library. Proc. Natl. Acad. Sci. USA 92, 3938–3942. Desiderio, A., Franconi, R., Lopez, M., Villani, M.E., Viti, F., Chiaraluce, R., Consalvi, V., Neri, D., Benvenuto, E., 2001. A semi-synthetic repertoire of intrinsically stable antibody fragments derived from a single-framework scaffold. J. Mol. Biol. 310, 603– 615. Dubremetz, J.F., Garcia-Reguet, N., Conseil, V., Fourmaux, M.N., 1998. Apical organelles and host-cell invasion by Apicomplexa. Int. J. Parasitol. 28, 1007–1013. Enriquez, F.J., Riggs, M.W., 1998. Role of immunoglobulin A monoclonal antibodies against P23 in controlling murine C. parvum infection. Infect. Immun. 66, 4469–4473. Fayer, R., Morgan, U., Upton, S., 2000. Epidemiology of Cryptosporidium: transmission, detection and identification. Int. J. Parasitol. 30, 1305–1322. Griffiths, A.D., Williams, S.C., Hartley, O., Tomlinson, I.M., Waterhouse, P., Crosby, W.L., Kontermann, R.E., Jones, P.T., Low, N.M., Allison, T.J., Prospero, T.D., Hoogenboom, H.R., Nissim, A., Cox, J.P.L., Harrison, J.L., Zaccolo, M., Gherardi, E., Winter, G., 1994. Isolation of high affinity human antibodies directly from large synthetic repertoires. EMBO J. 13, 3245–3260. Griffiths, J.K., 1998. Human cryptosporidiosis: epidemiology, transmission, clinical disease, treatment, and diagnosis. Adv. Parasitol. 40, 37 –85. Griffiths, J.K., Balakrishnan, R., Widmer, G., Tzipori, S., 1998. Paromomycin and geneticin inhibit intracellular C. parvum without trafficking through the host cell cytoplasm: implications for drug delivery. Infect. Immun. 66, 3874–3883. Gut, J., Nelson, R.G., 1994. C. parvum sporozoites deposit trails of 11A5 antigen during gliding locomotion and shed 11A5 antigen during invasion of MDCK cells in vitro. J. Eukaryot. Microbiol. 41, 42S–43S. Heyworth, M.F., 1992. Immunology of Giardia and Cryptosporidium infections. J. Infect. Dis. 166, 465 –472. Hoogenboom, H.R., Lutgerink, J.T., Pelsers, M.M., Rousch, M.J., Coote, J., Van Neer, N., De Bruine, A., Van Nieuwenhoven, F.A., Glatz, J.F., Arends, J.W., 1999. Selection-dominant and nonaccessible epitopes on cell-surface receptors revealed by cell-panning with a large phage antibody library. Eur. J. Biochem. 260, 774 –784. Jenkins, M.C., O’Brien, C., Trout, J., Guidry, A., Fayer, R., 1999. Hyperimmune bovine colostrum specific for recombinant C. parvum antigen confers partial protection against cryptosporidiosis in immunosuppressed adult mice. Vaccine 17, 2453–2460. Kabat, E.A., 1995. Kabat Database of Immunological Sequences, Rel. 5.0, updated weekly edition. Online address: www.ncbi.nlm.nih.gov. Langer, R.C., Riggs, M.W., 1996. Neutralizing monoclonal antibody protects against C. parvum infection by inhibiting sporozoite attachment and invasion. J. Eukaryot. Microbiol. 43, 76S–77S. Lou, J., Marzari, R., Verzillo, V., Ferrero, F., Pak, D., Sheng, M., Yang, C., Sblattero, D., Bradbury, A., 2001. Antibodies in haystacks: how selection strategy influences the outcome of selection from molecular diversity libraries. J. Immunol. Methods 253, 233 –242. Marks, J.D., Hoogenboom, H.R., Bonnert, T.P., McCafferty, J., Griffiths, A.D., Winter, G., 1991. By-passing immunization. Human antibodies from V-gene libraries displayed on phage. J. Mol. Biol. 222, 581 –597. Okhuysen, P.C., Chappell, C.L., 2002. Cryptosporidium virulence determinants – are we there yet? Int. J. Parasitol. 32, 517–525. Perryman, L.E., Kegerris, K.A., Mason, P.H., 1993. Effect of orally administered monoclonal antibody on persistent C. parvum infection in scid mice. Infect. Immun. 61, 4906–4908. Perryman, L.E., Jasmer, D.P., Riggs, M.W., Bohnet, S.G., McGuire, T.C., Arrowood, M.J., 1996. A cloned gene of C. parvum encodes neutralization-sensitive epitopes. Mol. Biochem. Parasitol. 80, 137 –147.
L. Chen et al. / International Journal for Parasitology 33 (2003) 281–291 Perryman, L.E., Kapil, S.J., Jones, M.L., Hunt, E.L., 1999. Protection of calves against cryptosporidiosis with immune bovine colostrum induced by a C. parvum recombinant protein. Vaccine 17, 2142–2149. Petersen, C., Gut, J., Doyle, P.S., Crabb, J.H., Nelson, R.G., Leech, J.H., 1992. Characterization of a .900,000-M(r) C. parvum sporozoite glycoprotein recognized by protective hyperimmune bovine colostral immunoglobulin. Infect. Immun. 60, 5132–5138. Pope, A.R., Embleton, M.J., Meranugh, R., 1996. Construction and use of antibody gene repertoires. In: McCafferty, J., Hoogenboom, H.R., Chisewell, D.J. (Eds.), Antibody Engineering, Oxford University Press, New York, pp. 1–40. Priest, J.W., Kwon, J.P., Arrowood, M.J., Lammie, P.J., 2000. Cloning of the immunodominant 17-kDa antigen from C. parvum. Mol. Biochem. Parasitol. 106, 261–271. Riggs, M.W., Stone, A.L., Yount, P.A., Langer, R.C., Arrowood, M.J., Bentley, D.L., 1997. Protective monoclonal antibody defines a circumsporozoite-like glycoprotein exoantigen of C. parvum sporozoites and merozoites. J. Immunol. 158, 1787– 1795. Rossignol, J.F., Ayoub, A., Ayers, M.S., 2001. Treatment of diarrhea caused by C. parvum: a prospective randomized, double-blind, placebocontrolled study of Nitazoxanide. J. Infect. Dis. 184, 103–106. Sambrook, J., Fritsch, E.F., Maniatis, T., 1989. Amplification and storage of cosmid libraries. In: Nolan, C., (Ed.), Molecular Cloning: A Laboratory Manual, 2nd Edition, Cold Spring Harbor Laboratory Press, Cold Spring Harbor, NY, pp. 3 –44. Sharon, J., Sompuram, S.R., Yang, C.-Y., Williams, B.R., Sarantopoulos, S., 2002. Construction of polyclonal antibody libraries using phage display. In: O’Brien, C., Aitken, R. (Eds.), Methods in Molecular Biology, Humana Press, Totowa, NJ, pp. 101– 112. Smith, D.F., Larsen, R.D., Mattox, S., Lowe, J.B., Cummings, R.D., 1990. Transfer and expression of a murine UDP-Gal:beta-D-Gal-alpha 1,3galactosyltransferase gene in transfected Chinese hamster ovary cells. Competition reactions between the alpha 1,3-galactosyltransferase and
291
the endogenous alpha 2,3-sialyltransferase. J. Biol. Chem. 265, 6225–6234. Strong, W.B., Gut, J., Nelson, R.G., 2000. Cloning and sequence analysis of a highly polymorphic C. parvum gene encoding a 60-kilodalton glycoprotein and characterization of its 15- and 45-kilodalton zoite surface antigen products. Infect. Immun. 68, 4117– 4134. Tzipori, S., 1998. Cryptosporidiosis: laboratory investigations and chemotherapy. Adv. Parasitol. 40, 187–221. Tzipori, S., Rand, W., Griffiths, J., Widmer, G., Crabb, J., 1994. Evaluation of an animal model system for cryptosporidiosis: therapeutic efficacy of paromomycin and hyperimmune bovine colostrum-immunoglobulin. Clin. Diagn. Lab. Immunol. 1, 450 –463. Vaughan, T.J., Williams, A.J., Pritchard, K., Osbourn, J.K., Pope, A.R., Earnshaw, J.C., McCafferty, J., Hodits, R.A., Wilton, J., Johnson, K.S., 1996. Human antibodies with sub-nanomolar affinities isolated from a large non-immunized phage display library. Nat. Biotechnol. 14, 309– 314. Vest Hansen, N., Ostergaard Pedersen, L., Stryhn, A., Buus, S., 2001. Phage display of peptide/major histocompatibility class I complexes. Eur. J. Immunol. 31, 32 –38. Ward, H., Cevallos, A.M., 1998. Cryptosporidium: molecular basis of hostparasite interaction. Adv. Parasitol. 40, 151–185. Williams, B.R., Sharon, J., 2002. Polyclonal anti-colorectal cancer Fab phage display library selected in one round using density gradient centrifugation to separate antigen-bound and free phage. Immunol. Lett. 81, 141–148. Winter, G., Milstein, C., 1991. Man-made antibodies. Nature 349, 293– 299. Winter, G., Gooley, A.A., Williams, K.L., Slade, M.B., 2000. Characterization of a major sporozoite surface glycoprotein of C. parvum. Funct. Integr. Genomics 1 (3), 207–217. Zu, S.-X., Fang, G.-D., Fayer, R., Guerrant, R.L., 1992. Cryptosporidiosis: pathogenesis and immunology. Parasitol. Today 8, 24–27.