Immunocytochemical characterization of viruses and antigenic macromolecules in viral vaccines

C.R Acad Sci Paris, Sciences de la vie / Life Sciences 324 (2001) 815–827 © 2001 Académie des sciences/Éditions scientifiques et médicales Elsevier SAS Tous droits réservés S0764446901013609/FLA Sciences médicales / Medical sciences Immunocytochemical characterization of viruses and antigenic macromolecules in viral vaccines Nguyen Van Mana, Hoang Thuy Nguyenb, Huynh Thi Phuong Lienb, Nguyen Thu Vanb, Nguyen Kim Giaob, Nguyen Minh Lienb, Nguyen Thanh Thuyb, Irene Duniad*, Jean Cohenc, E Lucio Benedettid a Poliomyelitis Vaccine Research and Production Center (POLIOVAC) , Hanoi, Viet Nam National Institute of Hygiene and Epidemiology (NIHE), Hanoi, Viet Nam c Virologie et immunologie moléculaire, institut national de la recherche agronomique, Jouy-en-Josas, France d Institut Jacques-Monod–CNRS–universités Paris-6 et Paris-7, France b Received 27 March 2001; accepted 30 May 2001 Communicated by Jean-Antoine Lepesant Abstract – Gold immunolabeling combined with negative staining (GINS) provides a valuable immunocytochemical approach that allows a direct ultrastructural definition of all viral vaccine constituents that share common antigenic features with pathogenic viral particles These results have implications for the development of viral vaccines since it has been demonstrated that incomplete viral particles such as natural empty capsides and Rotavirus-like particles lacking the infective genome are potential candidates for the production of neutralizing antibodies Furthermore comparative results of the application of GINS to either inactivated vaccines or unfixed samples provide direct evidence that even after inactivation specific antigenic sites are still available for gold immunolabeling © 2001 Académie des sciences/Éditions scientifiques et médicales Elsevier SAS gold immunolabeling / negative staining / Polio-virus / Japanese encephalitis virus / hepatitis viruses / Rotavirus (RV) / Rotavirus-like particles (VLPs) Résumé – Caractérisation Immunocytochimique de virus et de macromolécules antigéniques dans les vaccins antiviraux L’application combinée de la coloration négative et de l’immunomarquage l’or colloïdal (GINS, pour ‘gold immunolabeling combined with negative staining’), permet d’identifier la fois les caractères structuraux et les propriétés antigéniques de plusieurs vaccins l’aide des anticorps spécifiques L’application de cette méthode aux vaccins contre les virus de la polio et de l’hépatite A, a mis en évidence la présence de très nombreuses capsides vides dans les préparations vaccinales Ces structures ont une antigénicité élevée tout fait comparable celle des particules virales complètes De même, l’étude de capsides de rotavirus dépourvues de génome, produites par des systèmes recombinants, montre que ces pseudo-particules virales ont une antigénicité élevée La méthode décrite est une technique relativement rapide et simple, adaptée aux moyens des laboratoires de pays en voie de développement Nos résultats peuvent aussi contribuer au choix d’une stratégie dans la production de vaccins, fondée sur l’isolement et la production de particules pseudo-virales, dépourvues de génome, mais hautement antigéniques © 2001 Académie des sciences/Éditions scientifiques et médicales Elsevier SAS immunomarquage l’or / coloration négative / virus de la polio / encéphalite japonaise / hépatite A, hépatite B / virus : poliomyélite, rotavirus (RV) / particules pseudo-virales de rotavirus (VLPs) *Correspondence and reprints E-mail address: dunia@ijm.jussieu.fr (I Dunia) 815 N Van Man et al / C.R Acad Sci Paris, Sciences de la vie / Life Sciences 324 (2001) 815–827 Version abrégée L’application conjointe de l’immunomarquage l’or colloïdal et de la coloration négative (GINS), l’étude de plusieurs vaccins par microscopie électronique, a permis d’identifier les caractères structuraux de particules virales et, simultanément, les propriétés antigéniques de constituants vaccinaux actifs Dans les vaccins contre les virus de la polio et de l’hépatite A, nous avons pu observer de nombreuses capsides virales dépourvues de matériel génomique Cependant, ces structures conservent leurs propriétés antigéniques telles qu’on peut les déduire par l’intensité du marquage obtenu avec la méthode décrite De même, les particules pseudo-virales de rotavirus produites dans des systèmes recombinants et constituées par l’assemblage in vitro de protéines de rotavirus, sont intensément immunomarquées L’application de cette méthode a mis en évidence des différences structurales entre des vaccins de l’hépatite B préparés de deux faỗons diffộrentes : 1) le vaccin purifiộ par lisolement de l’antigène de surface HBsAg partir de sérum de porteurs sains du virus ; 2) le vaccin préparé en produisant l’antigène HBsAg dans un système recombinant L’antigène HBsAg d’origine humaine est caractérisé par la présence de nombreuses particules sphériques de 20 nm de diamètre et la présence de nombreux composants tubulaires En revanche, l’antigène HBsAg préparé dans un système recombinant ne contient que des particules sphériques de 20 nm de diamètre Cette différence morphologique et Introduction Viral diseases cause major community health problems, particularly in developing countries Despite great progress in our knowledge of fundamental and applied virology, there are only few identified natural or chemotherapeutic agents that are able to cure viral illnesses Until recently anti-viral vaccines were almost the unique successful resource for prevention and treatment of many viral syndromes [1–3] At a time when vaccine production and quality assessment are to be further developed, it is useful to have advanced electron microscopy (EM) methods that can test both the antigenicity and the ultrastructural features of vaccine preparations Several EM techniques have been developed that could help in the identification of viral particles in a crude sample and in the visualization of interactions between viral components and specific antibodies [4–10] Cryo-EM has opened a new avenue for high-resolution studies of the viral architecture and its interactions with specific 816 probablement antigénique, est due la composition protéique de chaque préparation Dans le premier cas (antigène d’origine humaine), la présence d’un composant protéique de 44 kDa (L), responsable de la formation de tubules, a été décrite Ce polypeptide est absent de la préparation d’antigène du système recombinant Il a été également décrit que les particules sphériques du HBsAg obtenues dans un système recombinant contiennent le récepteur de l’albumine sérique humaine polymérisée (protéine de 34 kDa) Nous avons pu mettre en évidence, avec nos résultats d’immunomarquage, que l’antigène HBsAg d’origine humaine contient aussi ce récepteur En effet, le blocage préalable par l’incubation exhaustive de ces préparations avec l’albumine sérique humaine, réduit considérablement l’immunomarquage de l’antigène d’origine humaine Ces résultats indiquent que la méthode utilisée est suffisamment sensible pour permettre l’étude qualitative de préparations vaccinales D’autre part, l’analyse du nombre de particules d’or associées aux différents composants peut donner des indications utiles sur l’efficacité du marquage et, par conséquent, sur l’immunogénicité de particules virales de chaque vaccin étudié en tenant compte des différents procédés de préparation (inactivation par la chaleur, fixation faible, etc.) Finalement, l’utilisation de cette méthode permet un contrôle de qualité constant, économique et rapide des préparations vaccinales, et aussi la possibilité d’envisager la production d’un vaccin basé sur la préparation de particules dépourvues de génome mais caractérisées par une antigenicité spécifique et significative antibodies [11–12] However, this technique requires fully specialized laboratory facilities that are not easily available in developing countries In our opinion, the combined application of gold immunolabeling and negative staining (GINS), represents a method of more general applicability and can provide valuable information concerning the relation between ultrastructure and immunological features of viruses and viral vaccines Viral vaccines include live attenuated vaccines, killed virus vaccines, antigenic macromolecules produced in living organisms by active viral infections and recombinant subunit vaccines [2] The data presented in this report demonstrate that the application of GINS to virus and viral vaccines is suitable for the identification of the structural entities bearing the antigenic determinants which are selectively enriched during the purification procedure of the vaccine Although the purification and quality control of a vaccine preparation can be tested by various biochemical and/or immunological methods, GINS may have the advantage of providing a direct immunocytochemical esti- N Van Man et al / C.R Acad Sci Paris, Sciences de la vie / Life Sciences 324 (2001) 815–827 mation of the antigenic particle population and at the same time, of the presence of non-immunogenic contaminants Materials and methods 2.1 Viral and vaccine preparations – The Japanese encephalitis virus (JEV) vaccine was prepared as described [13] Briefly, viral particles were purified from brain tissue of Swiss mice infected with human JEV The virus was then inactivated by treatment with 1/4 000 formaldehyde for 40 h – Poliovirus vaccine was prepared from monkey kidney cell primary cultures infected with poliovirus, according to the protocol developed in POLIOVAC-NIHE The vaccine preparation was also inactivated by 1/4 000 formaldehyde for 40 h – Hepatitis A virus (HAV) vaccine was prepared from monkey kidney cell primary cultures infected with HAV isolated from human carriers, according to the technique of purification of the Laboratory of Hepatitis Virus of the National Institute of Hygiene and Epidemiology (NIHE), Hanoi, Vietnam The purified vaccine preparation was inactivated with by 1/4 000 formaldehyde for 40 h – Hepatitis B surface antigen (HBsAg) was purified from asymptomatic human carriers but with a high HBsAg titer as described in [14] The vaccine preparation was inactivated by 1/4 000 formaldehyde for 40 h – Recombinant HBsAg produced in Chinese hamster ovary transfected cells (CHO) with the hepatitis B virus recombinant plasmid [15], was a generous gift of Dr M.L Michel, who has carried out SDS-polyacrylamide gel electrophoresis and immunoblotting experiments to caracterize the protein constituents of HBsAg preparations [15], (Laboratory of HBV, Pasteur Institute, Paris) – Rotavirus (RV) and Rotavirus-like particles (VLPs), were prepared as described previously, [16–18] Rotavirus preparations for GINS experiments were at times inactivated by 0.2 % formaldehyde treatment for 10 – The absence of viral contaminants in polio and HAV vaccines was established by immunological tests on the monkey kidney cell primary cultures before infection with the purified virus (Elisa essays using specific reference antibodies raised against SV40, PPLO and foamy virus, provided by WHO) 2.2 Antibodies Immunolabeling was carried out using specific primary antibodies: – Affinity purified rabbit antibody raised against purified JEV prepared by the Encephalitis Virus Laboratory of NIHE, Hanoi, Vietnam – Affinity purified rabbit antibodies raised against poliovirus prepared in the Laboratory of Virology, POLIOVAC, Hanoi, Vietnam – Human sera against HAV obtained from the Laboratory of Hepatitis Virus of NIHE, Hanoi, Vietnam, and also from the Laboratory of Virology, Hôpital Saint Antoine, Paris, France – Affinity purified rabbit antibodies raised against HB viral envelope proteins prepared at the Laboratory of Hepatitis Virus of NIHE, Hanoi, Vietnam – Affinity purified rabbit antibodies raised against HBsAg produced in CHO cells transfected with plasmid containing the S gene and the pre-S region of HBV [15] – Rabbit polyclonal antibodies raised against RV proteins which recognize VP2, VP4, VP6, VP7 – Monoclonal antibodies E22 and RV138 raised against the RV VP2 and VP6 proteins, respectively [19–20] The specificity of the antibodies raised against JEV and poliovirus was established by comparative Elisa essays using as reference specific antibodies against JEV provided by Bikan, (Osaka, Japan) and specific antibodies against poliovirus provided by WHO The specificity of the antibodies raised against Rotavirus proteins was tested by immunoblotting experiments using purified Rotavirus capside proteins produced in the recombinant system [19] Immunoblotting experiments were also carried out for testing the specificity of the rabbit antibodies raised either against human HBsAg or HBsAg from CHO recombinant system [15] 2.3 Gold Immunolabeling and Negative Staining The GINS method that we developed is derived from the single drop technique [21] and carried out as follows: 20 mL droplets of viral or vaccine preparations are placed on a clean parafilm surface Collodium–carbon coated grids are made hydrophilic by rinsing them with 0.01 % Bacitracin in water The grids still wet are put on top of sample droplets for 5–10 to pick up the sample The grids are washed with 2–3 droplets of PBS and then allowed to float on a droplet of PBS-2 % bovine serum albumin (BSA) for 20 to block non-specific antigenic sites Some samples of HBsAg were incubated with PBS complemented with % of human serum albumin (HSA) instead of BSA, as blocking step The grids were then reacted with the specific first antibodies, diluted in PBS–BSA 0.5 %, or PBS–HSA 0.5 % for 20 After careful washes with PBS–BSA 0.2 %, or PBS–HSA 0.5 %, grids were subsequently incubated for 20 with protein A conjugated to or 10 nm gold particles (Dept Cell Biology, University of Utrecht, The Netherlands) This step was followed by several washes with PBS and rapid fixation (5 min) with 0.1 % glutaraldehyde in aqueous solution For monoclonal antibodies, before incubation with gold-labeled protein A we used as bridge antibodies, a 15 incubation with affinity purified rabbit anti-mouse immunoglobulins, diluted 1:500 in PBS–BSA 0.5 % The grids were then thoroughly washed with a solution of 0.1 % ammonium acetate, and negative staining was carried out using a % aqueous solution of uranyl acetate or uranyl formate pH 5.4 Ammonium molybdate or phosphotungstic acid both at % and pH 7, were occasionally used but uranyl salts yielded better results During the procedure the grids were not allowed to dry and care was 817 N Van Man et al / C.R Acad Sci Paris, Sciences de la vie / Life Sciences 324 (2001) 815–827 Figure Poliovirus vaccine preparation immunolabeled using affinity purified rabbit antibodies raised against poliovirus A Poliovirus vaccine preparation showing the presence of 30-nm intact round-shaped virions mixed with particulate material Both components are immunolabeled Arrow points to an empty viral shell Arrow heads point to a viral particle displaying features of disassembly Bar: 40 nm B Gallery of empty viral shells also immunolabeled Bar: 20 nm taken to maintain them floating on the drop surfaces After negative staining, the grids were dried slowly before obser vation All procedures were performed at room temperature We carried out control experiments testing the specificity of the immunolabeling by treating the samples directly with gold labeled protein A, without previous incubation with specific antibodies Other control experiments were carried out by incubating the samples with non-specific antibodies 818 Specimens were examined with a JEOL EM 1010 and a Philips EM CM12 both EM working at 80 kV 2.4 Chemicals Unless otherwise indicated, all chemicals used for this work were purchased from Sigma Chemicals Co (St Louis MO, USA) N Van Man et al / C.R Acad Sci Paris, Sciences de la vie / Life Sciences 324 (2001) 815–827 Results In our investigation we selected different viral vaccines that are commonly used for control and prevention of serious viral diseases in many developing countries To test the general applicability of GINS, we carried out labeling experiments on viral particles of different sizes and structural organization such as Picornaviridae and Rotavirus In the case of the HBV vaccine we studied by immunolabeling the specific viral antigen HBsAg isolated from human carriers during the various steps of purification We also compared the HBsAg isolated from human carriers with similar antigenic molecules constructed by transfection of recombinant plasmids The main issue of our investigation was to characterize the immunochemical features of viral vaccine preparations using specific antibodies and sera from human carriers Furthermore, in view of analysing the capability of GINS to identify specific viral protein domains, we studied RV and Rotavirus-like particles (VLPs) using monoclonal antibodies raised against single viral polypeptides [16–19] 3.1 Poliovirus Poliovirus, genus Enterovirus, is a member of the Picornaviridae It consists of roughly spherical virions 24–30 nm in diameter [12, 22] The intact particles viewed in the vaccine preparation are intensely gold immunolabeled (figure 1A) In addition, several empty shells (20–30 nm) are found (figure 1B) The empty shells, as well as the viral material at all stages of its degradation, are positively tagged with gold immunolabeling using antibodies raised against Polio virus particles (figure 1A, arrows and 1B) No contaminant viral particles could be detected in any vaccine preparations 3.2 HAV vaccine HAV hepatovirus is a member of the Picornaviridae [23–24] The diameter of the viral particles ranges between 20–30 nm The viral proteins are assembled according to a dodecahedral model in which a capsid surrounds the RNA core [25] The vaccine preparation is composed of viral particles (20–30 nm in diameter, figure 2) The virions are tagged with gold labeled protein A when the sample was previously incubated either with specific antibodies raised against HAV or with HAV positive human serum (figure 2B) The HAV vaccine preparation is also characterized by many round or oval shells (20–22 nm in diameter), that likely correspond to natural empty capsids (NECs; figure 2A), [23] The empty shells are also gold immunolabeled In addition, the vaccine HAV preparation comprises desintegrated virions displaying a peripheral protein shell partially stripped away from the core This material is also gold immunolabeled and has certain common features with the ‘skullcaps’ (figure 2C), [23] Figure HAV vaccine preparation immunolabeled using human sera against HAV A HAV vaccine preparation comprises purified 20-nm round-shaped empty viral capsides (arrows) which are immunolabeled using human sera against HAV Bar : 45 nm B Immunolabeled round-shaped 30-nm virion with a distinct outer layer Bar: 20 nm C The disassembly of a viral particle generates platelets or ‘skullcaps’ (arrow head) Bar: 50 nm are characterized by a membranous envelope and a fine peplomer surrounding a spherical nucleocapsid with yet unknown symmetry The vaccine preparation contains many intact viral particles 45–50 nm in diameter which are intensely gold immunolabeled (figures 3A and 3B) Another component is represented by round-shaped empty shells (30–40 nm) (figure 3C) Furthermore, small particulate entities which have structural features comparable to fragments of the viral membrane envelope (figure 3A arrows) are also immunolabeled 3.3 JEV vaccine 3.4 Hepatitis B viral vaccine (HBsAg from human serum of HBV carriers) JEV belongs to the genus Flavivirus (family Flaviviridae) Virions are spherical, 45–55 nm in diameter [22, 26] They The major constituent of the HB vaccine consists of 20–22-nm particles of roughly spherical shape (figure 4A) 819 N Van Man et al / C.R Acad Sci Paris, Sciences de la vie / Life Sciences 324 (2001) 815–827 preparation consists of an homogeneous population of 20–22-nm spherical particles and no tubular components could be detected (figure 5) The particles are intensely gold immunolabeled by anti-HBV antibodies, particularly rabbit polyclonal antibodies raised against the HBsAg produced by transfection of the recombinant plasmid (figure 5) (15) When the preparation of HBsAg purified from human serum of HBV carriers is pre-incubated with HSA as blocking agent instead of BSA, the gold immunolabeling, using the rabbit polyclonal antibody raised against HbsAg, is dramatically reduced (figure 6A) This residual labeling accounts for less than 10 % as compared to more than 80 % of labeled particles when BSA is used Reduction of the gold immunolabeling of HBsAg-CHO particles was also apparent when the preparation is pre-incubated with HSA (figure 6B) 3.5 RV and VLPs Figure JEV vaccine preparation immunolabeled using affinity purified rabbit antibody raised against JEV A The preparation is characterized by the presence of 45–50-nmimmunolabeled intact virions The JEV sample also comprises particulate entities of different shapes, likely resulting from the degradation of viral particles This material appears also gold immunolabeled (arrows) Bar: 50 nm B Higher magnification of viral particles showing the outer membranous shell and the dense core of the viral capsid The gold particles are associated to the outer surface of virions Bar: 30 nm C Empty viral shell intensely labeled Bar: 20 nm Tubular-shaped components (20 nm thick) of variable lengths are also observed (figure 4B) All particulate entities are heavily gold immunolabeled by specific HBsAg antibodies (Figures 4A and 4B) Some HBsAg preparations from the sera of human carriers were examined before the last run of the process of purification by gradient centrifugation and found to contain aggregates of round-shaped 20-nm particles and tubular components (figure 4C) Occasionally these clusters are associated to granular material that is immunolabeled with antibodies raised against human IGg (figure 4C, arrows) We have examined an HBsAg preparation produced in CHO cells [15] transfected with a recombinant plasmid containing the S gene and the pre-S region of HBV This 820 Inactivated viral particles [12], purified from RV (family, Reoviridae) infected cultures, are 75 nm in diameter and look like a wheel with a defined smooth outer rim Using polyclonal antibodies which recognize several viral proteins (VP2, VP6 and VP7), gold immunolabeling is detected either surrounding the periphery of the virus or at the viral surface (figure 7A) The VLPs generated by the assembly of VP2, VP6 and VP7 are 70 nm in diameter and have a triple-layered organization (figure 7B) Gold immunolabeling with polyclonal antibodies appears as clusters of gold particles either around the periphery of VLPs or inside the triple-layered structure (figure 7B) The average number of gold particles labeling the intact formaldehyde inactivated RV counted in > 30 particles was of 13 gold particles per virion and for triple-layered VLPs The VLPs comprising only VP2 and VP6 are round-shaped particles of about 60 nm diameter displaying a double-layered organization (figure 8A) The outer layer consists of a regular assembly of subunits (figure 8A) The inner layer appears as a rather uniform stratum (figures 8B–8D) Fragments of double-layered VLPs are frequently found in these preparations and have semi-circular profiles characterized by the subunit organization of the uneven outer layer wrapping the inner stratum (figures 8B–8D) When the monoclonal antibody raised against VP6 is used the subunits forming the outer layer appear intensely tagged by gold particles (figure 8A) Conversely, gold immunolabeling is mainly restricted to the inner layer when VLPs VP2–VP6 are incubated with the monoclonal antibody raised against VP2 (figure 8B, arrow) Figure 8C shows one layered VLP intensely gold labeled with VP2 monoclonal antibody Figure 8D shows the cup-like aspect of the double-layered broken VLP; the VP2 monoclonal antibody has access to the VLP inner layer (Figure 8D) The unfixed double-layered VLPs (VP2-VP6), immunolabeled with the monoclonal anti-VP6 were tagged by an average number of 28 gold particles, whereas using the monoclonal antibody anti-VP2, gold particles were found associated with the viral inner layer, (the total number of VLPs counted in each case was 30) N Van Man et al / C.R Acad Sci Paris, Sciences de la vie / Life Sciences 324 (2001) 815–827 Figure HBsAg purified from human sera of HBV carriers A The preparation consists of round-shaped 20–22-nm particles which are intensely immunolabeled using antibodies raised against HBsAg-CHO Bar : 50 nm B Similar preparation displaying the pleomorphic organization of HBsAg characterized by roundshaped particles and tubular structures both intensely immunolabeled using antibodies directed against HBsAg from human sera Bar: 80 nm C The sample has been examined before the last run of the process of purification by gradient centrifugation It comprises clusters of 20-nm particles, tubular structures, mixed with a minor amount of granular material that appears immunogold labeled using an antibody directed against human IgG (arrows) Bar: 55 nm Discussion 4.1 Ultrastructural features of the viral vaccine constituents bearing the antigenic activity In spite of the inherent pitfalls of any technique of direct structural observation of biological specimens, the negative staining method for electron microscopy since its development, has provided the most straightforward and comprehensive information on morphology and design principles of viruses [21, cfr.27] The range of applicability of the negative staining technique can be expanded by using a novel approach that combines gold immunolabeling and the electron negative contrast These complementary techniques have already been successfully applied to several problems of fundamental and applied virology [4–8, 21, 28–29] The application of GINS to HAV, JEV and poliovirus vaccines provide evidence that in addition to a 821 N Van Man et al / C.R Acad Sci Paris, Sciences de la vie / Life Sciences 324 (2001) 815–827 Figure HBsAg produced in CHO cells The sample comprises almost exclusively round-shaped 20-nm particles which are all immunolabeled with rabbit polyclonal antibodies raised against the HBsAgCHO Bar: 80 nm major population of intact virions, the preparations contain empty shells and structural entities reflecting all stages of disassembly of the viral particles The presence of empty viral shells is a commun feature of purified Picornaviruses and other viruses (JEV), [30–31] More recently, these particulate entities were called NECs for ‘natural empty capsides’ [23] It is assumed that both infective virions as well as empty viral shells are produced when the viruses grow in tissue cultures [32] Empty shells [30] may have different sedimentation coefficients and variable protein compositions in comparison to infective virions [23] It has also been claimed that empty shells can be produced during the processes of purification and inactivation of viruses [23] When Poliovirus inactivation is carried out by harsh heating (56 °C), the RNA genome is released from the virion and the protein shells are simultaneously converted to a different antigenicity unable to produce neutralizing antibodies against the native infective virions [23, 33] However the native antigenicity of empty viral shells can be preserved when mild conditions of purification and 822 inactivation are applied, as is the case of our vaccine preparations which are inactivated by low concentrations of aldehydes Intact viruses (HAV, poliovirus and JEV) as well as the incomplete empty shells of our vaccine preparations appear gold immunolabeled We may then assume that the purification and inactivation processes generated viral structural entities characterized by exposed antigenic sites that are able to interact with specific antibodies raised against the viral antigens These results need further investigation with complementary immunoassays that could provide information on the immunogenicity of a vaccine preparations 4.2 Gold immunolabeling and negative staining applied to quality control of HBV vaccine Previous observations on purified preparations of HBV by negative staining, demonstrated the presence of types of particulate entities namely, the 42-nm double-shelled ‘Dane’ particles corresponding to the infectious HBV, a great number of 20-nm spherical particles and tubular N Van Man et al / C.R Acad Sci Paris, Sciences de la vie / Life Sciences 324 (2001) 815–827 Figure A HBsAg purified from human sera of HBV carriers The sample has been pre-incubated with HSA as blocking agent instead of BSA, followed by immunogold labeling using antibodies directed against HBsAg from human sera Immunogold labeling of both round-shaped 20-nm particles and tubular structures is negligible (arrow) Bar: 55 nm B HBsAg produced in CHO cells The sample compring clusters of 20-nm particles has been preincubated with HSA The gold immunolabeling using rabbit antibodies directed against HBsAgCHO is reduced (arrows) Bar: 60 nm components of variable lengths [34, 35] A common antigen, HBsAg, characterizes these different forms Our HBsAg purified preparations lack Dane particles and contain 20-nm spherical particles and tubular assemblies both heavily immunolabeled using antibodies directed against HBsAg The HBsAg 20-nm particles, produced in transfected CHO cells [15], have almost the same morphological features as the HBsAg 20-nm particles of our vaccine preparation They appear also heavily immunolabeled with the rabbit polyclonal antibodies raised against the recombinant HBsAg However, the recombinant HBsAg lacks the tubular forms that we frequently found in the human preparation of HBV carriers The existence of several morphological and immunochemical variants of HBsAg isolated from different human donors [3, 36], from squirrel and duck sera [34] is well documented Morphological differences are probably related to the variable aminoacid sequence and protein composition of HBsAg subunits [3] Gel electrophoresis and immunoblotting experiments demonstrated that the HBsAg produced from CHO clones contains polypeptides of 22 kDa (small, S) and 34 kDa (middle, M) but lacks the large polypeptide (L) of 44 kDa [15] It is assumed that the presence of the L polypeptide is responsible of the assembly of the tubular and filamentous structures [37, 38], therefore it is not surprising that in the recombinant HBsAg the tubular forms are absent HBsAg particles generated by CHO cells carry HSA receptors associated to the 34-kDa polypeptide [15] The presence of the 34-kDa protein likely confers to HBsAg a strong immunoreactivity in humans In our experiments, pre-incubation with HSA as blocking agent instead of BSA of both HBsAg form human origin and HBsAg-CHO, showed that the gold immunolabeling is dramatically reduced We might then assume that HSA interacts with a major species-specific antigenic determinant – the 34-kDa polypeptide – thus preventing further immunolabeling reaction with the polyclonal antibodies directed against human HBsAg The attraction of this approach is that one acquires by the application of GINS an indirect but useful estimate of the presence of specific components of the vaccine structural entites 823 N Van Man et al / C.R Acad Sci Paris, Sciences de la vie / Life Sciences 324 (2001) 815–827 Figure A Intact RV of 75 nm diameter, purified from RV infected cultures RV are gold immunolabeled with polyclonal antibodies raised against viral proteins VP2–VP7 Bar: 40 nm B Triple-layered subunit organization of 75-nm VLP generated by the assembly of VP2, VP6 and VP7 The three layers are immunolabeled by polyclonal antibodies raised against VP2-VP7 Bar: 35 nm The use of GINS has also contributed to address the problem of the presence in the HBsAg isolated from human carriers of immuno-complexes comprising HBcAg and HBeAg [8] During the purification procedure of HBsAg, immuno-complexes comprising HBcAg and HBeAg are progressively eliminated and the application of GINS, using specific anti-IgGs, showed that only in semi-purified preparations could be identified a negligible amount of human immunoglobulins associated to HBsAg These results suggest that GINS may be a suitable tool to implement the quality control of vaccine samples during purification procedures 824 4.3 Resolution power and labeling efficiency (LE) of GINS applied to RV, VLPs and viral vaccines Among icosahedral viruses, the RV structure has been thoroughly assessed by using cryo-microscopy and 3D reconstitution methods [12] Stable VLPs are produced by expression in insect tissue cultures of Bacculovirus recombinants with gene coding sequences for RV capsid proteins [16–18] Our data concern VLPs comprising either VP2, VP6 and VP7 (triple-layered) or VP2 and VP6 (doublelayered) GINS applied to these structures confirms the notion that VLPs lacking RNA genome are self-assembled N Van Man et al / C.R Acad Sci Paris, Sciences de la vie / Life Sciences 324 (2001) 815–827 Figure Double-layered VLPs generated by the assembly of VP2 and VP6 A VLP 2/6 tagged by a monoclonal antibody directed against VP6 Many gold particles form a crown around the VLP outer layer Bar: 35 nm B VLP 2/6 immunolabeled by a monoclonal antibody directed against VP2 The arrow points to the gold particles associated with the inner stratum Bar: 20 nm C VLP 2/6 tagged by a monoclonal antibody directed against VP2 Single layered VLP surrounded by clusters of gold particles Bar: 35 nm D A double-layered VLP 2/6 open at one side, immunolabeled with a monoclonal antibody directed against VP2 The gold particles are associated with the inner layer viewed ‘en face’ Note the circular profile of the fragmented outer layer Bar: 40 nm in the appropriate expression system and share common antigenic features with infective RV The term of selfassembly is used to indicate the process of association of viral capside proteins in the absence of viral genome by stereospecific, non-covalent interactions, between identical or quasi-equivalent subunits [12] The question remains as to whether the results of our immunolabeling experiments can provide an indication of the antigenic density of the viral vaccines The problems of concentration of the antigens, LE and signal to noise ratio have been evaluated for gold-immunolabeling of embedded tissue sections and non-embedded cryo-sections where the number of antigenic sites for a specific functional protein has been previously assessed [39] The LE is the proportion of antigen in a preparation that is recognized by the labeling antibody This parameter can be derived by counting the number of immunogold particles divided by the known number of antigens which they labeled [39] In our experiments we applied GINS to poliovirus, JEV, HAV, and RV previously inactivated by fixation with formaldehyde In these conditions it is not easy to estimate the absolute number of antigenic sites still available for gold immunolabeling and in turn, LE The more realistic range of LE expected for a known number of antigenic sites present in fixed biological samples is 1–15 % [39] In our experiments inactivated rotavirions are labeled with an average of 13 gold particles using the polyclonal antibodies raised against VP2 to VP7 Assuming that the major epitopes of the viral particles are associated with 260 VP6 trimers, 260 VP7 trimers and 120 subunits of VP2 (1 680 total antigenic sites per virion), the average of 13 gold particles tagging a single virion corresponds to a LE of 0.8 % This value is close to the estimation for aldehyde-fixed antigens [39] Using polyclonal antibodies raised against VP2 to VP7 and applied to unfixed triple-layered VLPs (VP2, VP6, VP7), we observed that the number of gold particles tagging the protein subunits is relatively small (5 gold particles/VLP) This value likely reflects the topographic distribution of the viral proteins and the accessibility of specific epitopes It is assumed that VP7 is mainly associated with the outer layer of VLPs and in this situation it may hinder the accessibility of VP2 and VP6 which are the major inner constituents of the triplelayered VLPs Unfixed and intact double-layered VLPs comprising VP2 (120 subunits) and VP6 (780 subunits), are labeled respectively with an average of 28 gold particles using a monoclonal antibody against VP6 and with an average of gold particles using a monoclonal antibody against VP2 Hence, LE for VP6 and VP2 are 3.5 % and % respectively This observation is consistent with the notion that in self-assembled double-layered VLPs the VP6 is the major constituent of the viral shell exposed to the outer surface of the double-layer, whereas VP2 forms the inner core of this double-layered VLP and its immunolabeling is enhanced when the inner layer is exposed by artificial removal of the outer layer (cf figures 8C and 8D) These data suggest that GINS is a valuable method for the topographic identification of the various viral constitu 825 N Van Man et al / C.R Acad Sci Paris, Sciences de la vie / Life Sciences 324 (2001) 815–827 ents but its resolution is limited by the steric hindrance generated by the packing density of the antigenic sites associated with different layers of the viral capsid Concluding Remarks The conjunct application of gold-immunolabeling and negative staining to selected viral vaccines provides valuable information both on the ultrastructural features and on the antigenicity of viral particles and other viral macromolecules The self-assembled virus-like particles and empty viral shells lacking the infective genome, share commun immunocytochemical characters with pathogenic viruses Therefore these non-infective structures may be considered as potential candidates for further purification and vaccine production We must however admit that although GINS can be considered a reliable technique for the qualitative analysis of the presence of active antigenic sites and for testing the absence of viral contaminants, it might requires for quantitation of the antigenic density more informations (by immunoassay and other complementary techniques) [3, 33], on the immunochemical and variable features of the viral preparations The resolution and labeling efficiency of GINS is relatively high and taking into account a very low background noise that characterizes our GINS preparations even a small number of gold particles associated to a viral structure would be highly significant Further developments on immunolabeling techniques, particularly the use of high-resolution immuno-probes References [1] Harrison S.C., Principles of virus structure, in: Fields B.N., Knipe D.M (Eds.), Virology, Raven Press, NY, 1990, pp 37–62 [2] Levine A.J., The origin of virology, in: Fields B.N., Knipe D.M (Eds.), Fundamental Virology, Lippin-Cott-Raven Publ Ltd, NY, 1996, pp 1–14 [3] Zuckerman J.N., Zuckerman A.J., Current topics in Hepatitis B, J Infect 41 (2000) 130–136 [4] Beesley J.E., Betts M.P., Virus diagnosis: a novel use for the protein A-gold probe, Med Lab Sci 42 (1985) 161–165 [5] Hyatt A.D., Eaton B.T., Lunt R., The grid-cell-culture technique: the direct examination of virus-infected cells and progeny viruses, J Microsc 145 (1987) 97–106 [6] Murti K.G., Brown P.S., Bean W.J., Webster R.G., Composition of the helical internal components of influenza virus as revealed by immunogold labeling/electron microscopy, Virology 186 (1992) 294–299 [7] Nermut M.V., Wallengren K., Pager J., Localization of actin in Moloney murine leukemia virus by immunoelectron microscopy, Virology 260 (1999) 23–34 [8] Stannard L.M., Lennon M., Hodgkiss M., Smuts H., An electron microscopic demonstration of immune complexes of Hepatitis B e-antigen using colloid gold as a marker, J Med Virol (1982) 165–175 [9] Hopley J.F., Doane F.W., Development of a sensitive protein A-gold immunoelectron microscopy method for detecting viral antigens in fluid specimens, J Virol Methods 12 (1985) 135–147 [10] Biel S.S., Gelderblom H.R., Diagnostic electron microscopy is still a timely and rewarding method, J Clin Virol 13 (1999) 105–119 [11] Dubochet J., Adrian M., Chang J.J., Homo J.C., Lepault J., McDowall A.W., Schultz P., Cryo-electron microscopy of vitrified specimens, Q Rev Biophys 21 (1988) 129–228 826 and reduced sized electron-dense markers, combined with site-directed mutagenesis of viral constituents and cryoelectron microscopy are already in progress [12, 40] These approaches might help further topological analyses of the structural and molecular features of viral vaccine constituents that play a key role in the activation of the humoral and/or cell mediated protective response and on the switch of immunological memory Acknowledgements We gratefully acknowledge the generous gift of antibodies directed against HBsAg and of recombinant HBsAg by Dr Marie-Louise Michel, Institut Pasteur, Paris We wish to thank Dr Catherine Johanet, Laboratory of Immuno-Hematology, Hôpital Saint-Antoine, Paris, for providing antibodies against HAV and Dr Anne-Lise Haenni, Institut Jacques Monod, Paris, for helpful discussion and advice Drs E L Benedetti, I Dunia and J Cohen express their gratitude to Dr Hoang Thuy Long, Director of the National Institute of Hygiene and Epidemiology (NIHE) and to Dr Dang Duc Trach, Chairman of the Scientific Council of NIHE, for their advice and kind hospitality These studies were supported by the Ministry of Health of Vietnam, the NIHE, the Vaccine Production Center POLIOVAC, Hanoi, Vietnam, and by the “Service de la Formation du Centre National de la Recherche Scientifique (CNRS) Programme CNRS-UNESCO”, Paris, France [12] Baker T.S., Olson N.H., Fuller S.D., Adding the third dimension to virus life cycles: three-dimensional reconstruction of icosahedral viruses from cryo-electron micrographs, Microb Mol Biol Rev 63 (1999) 862–922 [13] Doan Thi Thuy, Huynh Phuonh Lien, Nguyen Hong Hanh, Hoang Thuy Nguyen, Purification of Japanese Encephalitis virus for vaccine production, J Prevent Medec (1991) 19–25 [14] Hoang Thuy Nguyen, Nguyen Thu Van, Purification of Hepatitis B surface antigen for HBsAg-Micro-ELISA kit and Hepatitis B vaccine preparation, Proc Int Symp., Immunology and Infection, National Institute of Hygiene and Epidemiology (Hanoi), University of Amsterdam (Holland), 1990, Hanoi-Vietnam [15] Michel M.L., Pontisso P., Sobczak E., Malpiece Y., Streeck R.E., Tiollais P., Synthesis in animal cells of hepatitis B surface antigen particles carrying a receptor for polymerized human serum albumin, Proc Natl Acad Sci USA 81 (1984) 7708–7712 [16] Labbe M., Charpilienne A., Crawford S.E., Estes M.K., Cohen J., Expression of rotavirus VP2 produces empty corelike particles, J Virol 65 (1991) 2946–2952 [17] Crawford S.E., Labbe M., Cohen J., Burroughs, Yong-Jie Zhou, Estes M., Characterization of virus-like particles produced by the expression of Rotavirus capsid proteins in insect cells, J Virol 68 (1994) 5945–5952 [18] Crawford S.E., Estes M.K., Ciarlet M., Barone O’Neal C.M., Cohen J., Conner M.E., Heterotypic protection and induction of a broad heterotypic neutralization response by rotavirus-like particles, J Virol 73 (1999) 4813–4822 [19] Pothier P., Kohli E., Drouet E., Ghim S., Analysis of the antigenic sites on the major inner capside protein (VP6) of rotaviruses using monoclonal antibodies, Ann Ins Pasteur/Virol 138 (1987) 285–295 [20] Roseto A., Scherrer R., Cohen J., Guillemin M.C., Charpilienne A., Feynerol C., Peries J., Isolation and characterization of anti-rotavirus N Van Man et al / C.R Acad Sci Paris, Sciences de la vie / Life Sciences 324 (2001) 815–827 immunoglobulins secreted by cloned hybridoma cell lines, J Gen Virol 64 (1983) 237–240 [21] Hayat M.A., Miller S.E., Negative staining, McGraw-Hill Publ NY, 1990, pp 51–155 [22] Murphy F.A., Kingsbury D.W., Virus taxonomy, in: Fields B.N., Knipe D.M (Eds.), Virology, Raven Press, NY, 1990, pp 3–36 [23] Rueckert R.R., Picornaviridae and their replication, in: Fields B.N., Knipe D.M (Eds.), Virology, Raven Press, NY, 1996, pp 509–654 [24] Smit T.J., Baker T.S., Picornaviruses: epitopes, canyons and pockets, Adv Virus Res 52 (1999) 1–23 [25] Hollinger F.B., Ticehurst J., Hepatitis A virus, in: Fields B.N., Knipe D.M (Eds.), Virology, Raven Press, NY, 1990, pp 631–670 [26] Monath T.P., Flaviviruses, in: Fields B.N., Knipe D.M (Eds.), Virology, Raven Press, NY, 1990, pp 763–814 [27] Valentine R.C., Contrast enhancement in the electron microscopy of virus, Adv Virus Res (1962) 287–318 [28] Risco C., Anton I.M., Enjuanes L., Carrascosa J., The transmissible gastroenteritis Coronavirus contains a spherical core shell consisting of M and N proteins, J Virol 70 (1996) 4773–4777 [29] Zimmermann W., Breter H., Rudolph M., Ludwig H., Borna disease virus: immunoelectron microscopic characterization of cell-free virus and further information about the genome, J Virol 68 (1994) 6755–6758 [30] Jacobson M.F., Baltimore D., Morphogenesis of poliovirus I Association of the viral RNA with coat protein, J Mol Biol 33 (1968) 369–378 [31] Johnston M.D., Martin S.J., Capsid and procapsid proteins of a bovine enterovirus, J Gen Virol 11 (1971) 71–79 [32] Putnak J.R., Piticlips B.A., Differences between Poliovirus empty capsids formed in vivo and those formed in vitro: a role for the morphopoietic factor, J Virol 122 (1981) 173–183 [33] Kersten G., Hazendonk T., Beuvery C., Antigenic and immunogenic properties of inactivated polio vaccine made from Sabin strains, Vaccine 17 (1999) 2059–2066 [34] Hollinger F.B., Hepatitis B virus, in: Fields B.N., Knipe D.M (Eds.), Virology, Raven Press, NY, 1990, pp 2171–2238 [35] Robinson W.S., Hepadnaviridae and their replication, in: Fields B.N., Knipe D.M (Eds.), Virology, Raven Press NY, 1990, pp 2137–2170 [36] Stibbe W., Gerlich W.H., Variable protein composition of Hepatitis B surface antigen from different donors, Virology 123 (1982) 436–442 [37] Xu Z., Bruss V., Yen B., Formation of intracellular particles by Hepatitis B virus large surface protein, J Virol 71 (1997) 5487–5494 [38] Prange R., Werr M., Loffler-Mary H., Chaperones involved in Hepatitis B virus morphogenesis, J Biol Chem 380 (1999) 305–314 [39] Griffith G., Fine Structure Immunocytochemistry, Springer-Verlag, New York, 1993, pp 417–442 [40] Zlotnick A., Cheng N., Conway J.F., Booy F.P., Steven A.C., Stahl S.J., Wingfield P.T., Dimorphism of Hepatitis B virus capsids is strongly influenced by the C-terminus of the capsid protein, Biochemistry 35 (1996) 7412–7421 827 ... valuable information concerning the relation between ultrastructure and immunological features of viruses and viral vaccines Viral vaccines include live attenuated vaccines, killed virus vaccines, antigenic. .. easily available in developing countries In our opinion, the combined application of gold immunolabeling and negative staining (GINS), represents a method of more general applicability and can provide... Furthermore, in view of analysing the capability of GINS to identify specific viral protein domains, we studied RV and Rotavirus-like particles (VLPs) using monoclonal antibodies raised against single viral