Development of recombinant antibody technology for application in plant pathogen diagnosis

This thesis describes the applicability of the novel phage display technique to select plant-pathogen-specific monoclonal antibodies (MAbs) from combinatorial antibody libraries. The retrieved MAbs are so specific that they can be used as diagnostic tools in sensitive immunoassays for the detection and identification of plant pathogens. Testing results, obtained from laboratories that have applied these recombinant MAbs, are discussed in this conclusive chapter.BackgroundIn the last decades, it has become clear that chemical crop protection has to be reduced. Many of the pesticides, applied to destroy plant-pathogenic fungi, bacteria, insects and nematodes, are hazardous to the environment, and form a serious health risk for animals and humans. Since breeding for disease resistance generally takes years, epidemics have to be prevented by sanitary measures in combination with diagnosis in an early stage of disease development or preferably beforehand. Consequently, sensitive diagnostic assays are required that allow healthy plant propagation material to be certified and soil to be monitored. It is obvious that these data can be used to take adequate crop management decisions.The problems associated with serological detection of plant pathogensImmunoassays are widely applied as detection methods in agriculture because they can be applied to large numbers of samples, and their application is fast, robust, sensitive and cheap. A major concern is the specificity of the assays. Polyclonal antisera are still the active ingredient of many useful immunoassays. However, the immunochemical complexity of several target organisms is the main reason that extensive cross-reactions with non-target organisms can occur when these antisera are used.In 1975, Köhler and Milstein showed that this problem could be circumvented by making monospecific antibodies in continuous cultures by fusing the antibody producing B-lymphocytes with myeloma cells [8]. This 'hybridoma' technique opened a new perspective for the production of specific monoclonal antibodies (MAbs) against various antigens. Today, many MAbs are used in research, as therapeutic agent and for diagnostic purposes.However, the hybridoma technique was not as advantageous as was expected in the field of plant-pathogen detection. From the trials in raising specific MAbs against many different plant-pathogens, it became apparent that there are 'difficult' antigens. Several plant viruses can not be sufficiently purified and contain immunodominant plant residues and, as a consequence, the immune response is mainly directed to those impurities. In addition, plant-pathogenic bacteria, fungi and nematodes contain many epitopes that are 'shared' with closely related non-pathogenic family members, which is often leading to extensive cross-reactions. Therefore, the selection of the cells producing the desired MAbs is laborious and often impossible.Many techniques have been described to deal with 'difficult' antigens and varied from the enrichment of the desired B-cell populations to guiding of the immune response towards the target antigen, such as:Immunoadsorption of antigen-specific B-cells [2].Selection of antigen-specific B-cells by fluorescence activated cell sorting [15].Complement-mediated lysis of undesired B-cells [2].Masking (immuno-complexing) of contaminating immunodominant plant epitopes with anti-healthy plant antibodies during immunization [2].Suppression of the immune response to healthy plant extracts by injection with cyclophosphamide or other immuno suppressive drugs prior to immunization [11,16].Induction of immunological tolerance through tolerisation of neonatal mice with healthy plant extracts prior to immunization [6].Although these methods have occasionally shown their value in obtaining specific MAbs in their respective cases, the common drawback of all these methods is that they are not generally applicable. In addition, the latter two are also repulsive from an ethical point of view.Bypassing immunization: A general solution to deal with 'difficult' antigensA way to circumvent immunization was offered by recent developments in recombinant DNA techniques [14,17]. These techniques allow cloning and expression of large naive antibody repertoires, derived from non-immunized human donors, in E. coli as single-chain antibodies (scFvs) or Fab fragments. The use of these combinatorial antibody libraries in combination with the display of functional antibody fragments at the tips of filamentous phage created a powerful selection system for MAbs [5,12]. Utilization of this phage display system allows direct selection of highly specific MAbs from naive combinatorial antibody libraries [3,13,18]. As immunization is omitted, phage display is not biased. Hence it will not lead to antibodies directed against the most immunodominant epitope but rather towards the most abundant antigenic determinant.Mission impossible: how to make the impossible possibleBecause application of the hybridoma technique was not successful in generating MAbs of sufficient specificity to detect the plant-pathogen Ralstonia solanacearum in reliable and sensitive diagnostic assays, the efficacy of the phage display system was challenged (Chapter 2). To achieve this, phages derived from the (naive) Vaughan combinatorial antibody library (1.4 X 10 10different scFvs) were panned against purified lipopolysaccharides (LPS) which were isolated from R. solanacearum race 3 bacteria.After four successive rounds of phage growth and selection for LPS binding, soluble scFvs were produced and tested by enzyme-linked immunosorbent assay (ELISA) and immunofluorescence microscopy (IF). Four different scFvs could be distinguished on bases of RFLP analysis. Characterization of the monoclonal scFvs against several bacterial strains, indicated a specificity for R. solanacearum (biovar 2, race 3) LPS which is higher than can be obtained with conventional polyclonal antisera. The fact that the four scFvs were obtained within six weeks after starting this study emphasizes the potency of the phage display system.Selection of specific antibodies from the synthetic Nissim library, containing over 10 8different scFv, against beet necrotic yellow vein virus (BNYVV) was achieved (Chapter 3) through expression of the antibody fragments on the surface of bacteriophage M13 and subsequent binding of this phage-antibody to immobilized BNYVV. After several rounds of selection seven BNYVV-specific recombinant monoclonal antibodies were obtained. However, the yield of these monovalent scFv antibodies was low. In an attempt to improve the yields, the genes encoding the BNYVV-specific scFvs were genetically fused to alkaline phosphatase (AP/S) and expressed in E. coli . However, out of the seven different anti-BNYVV scFv-AP/S fusion proteins only three showed alkaline phosphatase activity and retained affinity for BNYVV and the quantity of produced scFv-AP fusion proteins was found to be low.Analysis of scFv encoding DNA during expression in E. coli showed the occurrence of high plasmid loss and a high incidence of recombination within the scFv encoding DNA, both for scFv and scFv-AP/S fusion proteins. Probably the scFvs are toxic and offer bacteria that recombine the plasmid DNA (and thereby disable expression of scFvs) a growth advantage. The more tightly repressed tetracycline promoter was used to replace the "leaky" LacZ promoter (Chapter 4). The co-expressed repressor protein tightly repressed the tetracycline promoter in the absence of inducer (anhydrotetracycline). When this new expression system was compared to the old expression system (LacZ induction by IPTG) for expression of scFv-AP fusion proteins, it was found to be superior as improved yields of functional recombinant antibodies were obtained. The tetracycline promoter is also more convenient to use. This is evident when large cultures or high numbers of cultures are grown because the medium has not to be changed since the use of this promoter is, unlike the LacZ promoter, glucose independent.Selection of tomato spotted wilt virus (TSWV)-specific scFvs from the naive Vaughan combinatorial antibody library (Vaughan) against purified nucleocapsids and against purified complete virus was also successful (Chapter 5). In contrast to previous selections the pooled scFv encoding DNA was isolated after the fourth round of selection, inserted in the vector pSKAP/S (Chapter 4) and tested directly as individual scFv-AP/S fusion proteins. Twelve different scFvs were obtained against the nucleocapsid (N) and five against the glycoproteins, G1 or G2. Six of the derived antibodies were produced with good yields: four scFvs against the nucleoprotein and two scFvs against G1 or G2. In ELISA they reacted with TSWV proteins in infected Nicotiana benthamiana and not with healthy plant extracts. When the N-reactive scFvs were evaluated for their specificity against six other tospoviruses, cross-reactions were observed with tomato chlorotic spot virus and to a lesser extent with groundnut ringspot virus. Several of the TSWV-reactive scFvs might be useful in routine testing, and an ELISA based on these recombinant antibodies is presently evaluated by routine testing laboratories.The ultimate goalThe development of a sensitive double antibody sandwich ELISA format, based on recombinant antibodies, was a major objective. Whole antibodies coat very well to ELISA plates because they contain flexible hinge-regions that allow rotation of the antigen-binding site. However, scFvs are not suited for coating onto ELISA plates, as they are small molecules that are inactivated upon coating [7]. Fusion of scFv to immunoglobulin domains may enhance the coating efficiency while retaining the full binding activity. Therefore, N56, one of the TSWV binding scFvs, was transformed into a mini-antibody through addition of a constant part (Mouse K constant region) and a flexible dimerization domain (Chapter 5). These bivalent mini-antibodies proved to be very active as a coating reagent in ELISA (Chapter 5) and could compete with a conventional polyclonal antiserum for coating efficiency. As was shown in Chapter 5, sensitive detection of TSWV nucleocapsids could be achieved by application of scFv-AP/S fusion proteins as antibody-enzyme conjugate. Therefore, serological testing can thus be carried out entirely by using bacteria derived coating and detection reagents.Endowing antibodies with novel propertiesRoutine testing for R. solanacearum is currently mainly performed by immunofluorescent cell staining (IF). However, conjugation of antibodies with the fluorochrome fluorescein isothiocyanate (FITC) is not reproducible and FITC fades rapidly upon illumination. This is not the case with the product of he genetic fusion of R.solanacearum -specific scFvs with green fluorescent protein (GFP). Bright, specific fluorescent labeling of target bacteria was observed when the obtained scFv-GFP fusion proteins (fluobodies) were tested by flow cytometry and in IF. The fluobodies proved to be more resistant to illumination, as was shown by IF after prolonged illumination (Chapter 6).Table 7.1.Summary of testing the recombinant scFv alkaline phosphatase fusion protein (monoclonal) in comparison to polyclonal anti- Ralstonia solanacearum antiserum FITC or alkaline phosphatase conjugates. For ELISA the plates were coated with a R. solanacearum specific antiserum (pca 9523bcd), blocked, incubated with samples (obtained from potato lots by PD regulations), incubated with either LPS7-AP/S or pca9523bcd-AP conjugate.potato samplesN=206IF-polyclonalELISA-polyclonalELISA-monoclonalIF+TypicalIF+AtypicalIF-60 minOD>35060 minOD<35060 minOD>15060 minOD<1501NAK97 IF+, PD97+,Undiluted, N=19100%0%0%100%0%100%0%NAK97 IF+ ,PD97+,Diluted 1:10, N=17100%0%0%94%6%100%0%2Suspected, NAK97 IF±,PD97-,Undiluted, N=3116%13%71%32%68%26%74%Suspected NAK97 IF±,PD97-, Diluted 1:10, N=2715%11%74%19%81%15%85%3NAK97 Cross-reactionsN=520%12%88%4%96%0%100%NegativeN=470%2%98%0%0%0%100%PD negative controlN=130%0%100%0%0%0%100%1 Samples were tested positive in 1997 and it was confirmed that they contained R. solanacearum bacteria.2 Samples were tested positive in 1997 but did not contain R. solanacearum bacteria.3Samples were tested negative in 1997.Preliminary data of routine testing by testing laboratoriesThe R. solanacearum- specific scFv, anti-LPS 7, was evaluated for its use in IF and in ELISA and was found to react with the same specificity as scFv LPS 12 (Chapter 2). It reacted with several R. solanacearum race 3 strains and showed cross-reactions with fewer strains than the polyclonal antiserum that is routinely applied for brown rot diagnosis of potato in the Netherlands. The general Netherlands inspection service for potatoes (NAK) evaluated the use of the anti-LPS7-AP/S based ELISA in their own laboratory and compared the results with those obtained with IF. The samples, obtained and already tested by the Dutch Plant Protection Service (PD) and NAK in 1997, were derived from various potato varieties such as Agria, Bildstar, Bintje Cardinal, Desiree, Diamant, Dore, Elkana, Felsina, van Gogh, Kanjer, Karnico, L. Rossetta, Monalisa, Spunta, Stefano and Symphony.The extraction of bacteria was performed according to the regulation of the PD and the results, summarized in Table 7.1, show that IF and ELISA gave comparable results for detection. Moreover, the ELISA that utilizes scFv-AP/S fusion proteins for detection is comparable with the polyclonal ELISA, but more importantly, the background is lower when the recombinant antibody is used for detection instead of the polyclonal alkaline phosphatase conjugate. No problems of increased background were observed regarding the sampling from the different potato varieties. The reduced number of cross-reactions that was observed for the recombinant antibodies in Chapter 2 also seems to hold for this field trial, as the number of positive samples found within the pool of suspected and cross-reacting samples was reduced. It can therefore be concluded that the recombinant scFv LPS7 is very useful under real testing conditions.General conclusion and future perspectivesThis thesis shows that antibody phage display is a useful technique for the selection of specific recombinant antibodies against a variety of plant-pathogens. Thus far, recombinant antibodies have been selected only against viruses and bacteria. The results described are promising, and it is expected that the phage display technique can also be used successfully for raising specific MAbs against fungi and nematodes. Therefore, the perspective that was predicted after introduction of the hybridoma technique in 1975 [8], now becomes reality with application of the novel antibody phage display technique.Although the synthetic naive Nissim library (10 8) was not compared directly to the naive Vaughan library (10 10), the obtained results favor the application of the latter. The variability within the retrieved scFv-antibodies was higher and the yield of the scFvs was better. Moreover, the amount of scFv-AP/S fusion protein required for a strong signal in the ELISA was up to 50-fold lower, indicating that the recombinant antibodies retrieved from the Vaughan library were of higher affinity. In general, it can be stated that large (>10 10) naive antibody libraries are a prerequisite for generating useful recombinant antibodies.More than 50 different antigen-specific recombinant antibodies were selected without the use of experimental animals. However, screening the selected scFv-antibodies was a major problem since E. coli produced several of those, only with great difficulty. It remains disappointing that the choice of a scFv for further characterization was based more on its yield, than on its affinity and specificity for the antigen. Development of eukaryotic expression systems might be an option. Yeast, insect-cells, fungi and plants resemble more closely the mammalian cells than the prokaryotic bacterium E. coli with regard to protein expression, and improved yields of functional scFv-antibodies can be expected when eukaryotes are used.There is an ongoing improvement of selection techniques from combinatorial antibody phage display libraries such as selectively infective phage [1,9], expression of antibodies on ribosomes [4] and selectively infected bacteria [10]. Meanwhile, combinatorial phage display libraries are designed to give better folding, and consequently, better producing antibodies. When these novel techniques find their way into plant pathology, it will undoubtedly result in highly specific antibodies that can be used in diagnostic assays that are fast, reliable, robust and cheap.