DISEASES OF AQUATIC ORGANISMS Vol. 16: 105-109, 1993 Published August Dis. aquat. Org. 5

Ultrastructure of erythrocytic virus of the South African anuran anchietae

A. P. Alves de Matosl, I. Paperna2

'Electron Microscopy Unit, Pathologic Anatomy Department. Curry Cabral Infectious Diseases Hospital, R. da Beneficencia'. and Functional Biology and Developmental Section of the Department of Zoology and Anthropology, Lisbon University. P-1000 Lisbon, Portugal 2Department of Sciences. Faculty of Agriculture of the Hebrew University of Jerusalem. Rehovot 76-100.Israel

ABSTRACT: Electron microscopic study of virus infections in the erythrocytes of the Ptychadena anchjetae from northern Transvaal, South Africa revealed icosahedral iridovirus-like particles measur- ing 175 to 198 nm in diameter. containing a dense, pleomorphic nucleoid. Virions were enveloped by 2 membranes, and may be associated with extensive membranous clots. The virus assembled in the cytoplasm within spherical fibrillar virosomes. Infected cells contained a large crystalloid inclusion body in their cytoplasm and in some, the inclusion was also present in the nucleus. Infection occurred in both mature and immature erythrocytes. There was no evidence of productive infection in non- erythrocytic cells.

INTRODUCTION Here, we present an EM study of virus-infected erythrocytes from the frog Ptychadena anchietae, cap- Virus infections of the erythrocytes of tured in Transvaal (South Africa). The infection has not and reptiles were previously thought to be caused by been previously reported, either in this species or in protozoan parasites and were named Toddia and southern Africa. Pirhemocyton. Toddia has been found mainly in amphibians (Johnston 1975), but also in snakes (Marquardt & MATERIALS AND METHODS Yaeger 1967) and fish (Arcay de Peraza & McLure 1971). It is characterized at the light microscopic level Study material was obtained from the South African by the presence of a cytoplasmic purple corpuscle rep- anuran Ptychadena anchietae (Bocage, 1867) collected resenting the viral factory or virosome, and a crystal- in Hoedspruit, northeast Transvaal, South Africa. loid inclusion of unknown significance (Sousa & Weigl Blood was collected from a clipped toe in a glass cap- 1976, Paperna & Alves de Matos 1993). illary tube, allowed to coagulate and then extracted Electron microscopic (EM) studies have shown into 2.5 % glutaraldehyde in 0.1 M cacodylate buffer these infections to be caused by indovirus-like (pH 7.4) for fixation for 24 h at 4 "C. Blood vessel, lung viruses (Stehbens & Johnston 1966, Sousa & Weigl and liver pieces were also fixed as above. The material 1976, Speare et al. 1991, Alves de Matos & Paperna was post-fixed in 1 % osmium tetroxide in the same 1993). buffer for 1 h. After rinsing in the cacodylate buffer the Another potentially related intraerythrocytic irido- material was serially dehydrated in graded ethanols virus-like infection is the viral erythrocytic necrosis and embedded in Epon. Thin sections cut on an LKBI11 (VEN) of marine fish (Smail 1982). microtome with a diamond or glass knife were stained on grid with uranyl acetate and lead cltrate, and exam- ined with a Phillips CM10 or JEOL 100s transmission Address for correspondence electron microscope.

O Inter-Research 1993 106 DISaqudt. Org 16. 105-109. 1993

W.. y+;*'-I ' :I. wri S 0 -req aIe3S .hpoq p!o~~els/i~s JeaI>nuellul m ,wrl g.0 = sleq aress .(3) sarpoq pro11~lsAi>>~useldolhse~lu~ p 8 s81~.uri s.0 = leq aless .snapnu aql6u1 -1alua suorir~ uri S 0 = r~qaIesS 4laqdnad (A)auosoim aql TO Marn pa6~elu~ .wri 9.0 = ieq aless .(SMOJJP) suo! -1!n 6u!1qu1asse qlr~(A) saruosoi1A uri g.0 = ~eqa1eJS [rnoiie)suorlr~ 30 pue (X)1013 snoueiqwaur ay1 jo M~IApa6le1u3 ruri = IP~ale3s (SMOJJP) auelquaur reapnu aql ql!~pale~sosse qols snouPiquaur ql!~[Ias pabeuea .m .wrl g.0 - req aless (shioi~e)saueiqwaur z dq padola~uasuor~r~ Z url 01 = ieq aless .p!ol[elsAls :~o~ieau!j !auosoi!h .MO.IJQ plog .learns poolq paurels esurar3 30 Malt, srdo~so.r~rw1q6rl I .6!3 .aelaryJue euapey~Xldpa~saju!-snl!~ .O[ 01 1 56!j

url S 0 = leq aless sarle[lrdes Aieuourlnd urqlrm (MOIIP) SUOIJIA pale!sosse-aueiquaw gl .6!j wri = 1eq aless .aBeqdolseu R Aq paz11A206~qd(3) Apoq prolle1sA13 qlr~a~AsoiqlA-la palsajuI .- ..urrl = req aless .Bun[ aql urolj 1se1qo~q1Alapalsaju~ p-[.lircI url l = ieq aless snapnu .U :(s~oi~e)suo~ir~ pue (3) Apoq pro11ejsAi~.(A) awosol!A 6u!~oqsuaa~ds aql wolj lselq -oiqlAia palsajul .6rj wri = leq aps [smoiie) suo!ir~ 6urlqurasse pue (I\) auosoi!A lsuqs~pe ~IIM~se~qoiqlAla-old uaalds palsajuI m urri I = ieq Apoq pro11elsA~sJeaTsnueiluI 1 ~I,I aplartpup ~uapeyJAldpalsaju~-snl!~ .g1 01 I I ,s6!-~ 108 Dis. aquat. Org. 16: 205-109, 1993

RESULTS described by Bernard et al. (1968) and Desser & Barta (1984) from North American and the virus of Bufo No appreciable pathological changes could be dem- marinus from Costa Rica described by Speare et al. onstrated in the infected frog. Hematological data (1991). Size differences have been related to generic reported elswhere (Paperna & Alves de Matos 1993) differences in iridoviruses of insects (Willis 1990). This show moderate proliferation of immature cells (17 % of suggests that frog species are infected by several dif- the total erythrocyte count), all of which became in- ferent viruses, and that the differential distribution of fected (Fig. 1). Pathology could be demonstrated only these viruses follows biogeographic regions. Similar at the host cell level. Infection involves major changes diversity has been found among erythrocytic viruses of in the erythrocyte leading to its ultimate lysis or elimi- fish (Smail 1982) and of reptiles (Alves de Matos & nation by phagocytosis. Paperna 1993). Virions seen in sections presented a hexagonal or Pentagonal and hexagonal sectional profiles of the pentagonal outline, compatible with an icosahedral capsid are well-documented characteristics of several shape (Stoltz 1971) measuring 175 to 198 nm (n = 10) in icosahedral cytoplasmic viruses, including iridoviruses diameter. Inside the particle, a dense, pleomorphic nu- (Stoltz 1971). Although Gruia-Gray et al. (1989), fol- cleoid could be found, often with a dense band delim- lowing biochemical and biophysical studies, suggested iting an irregularly filled or sometimes empty-appear- affinity of the frog erythrocytic virus with iridoviruses, ing center (Fig. 2). further biochemical and genetic data must be gathered Large numbers of viral particles were found in the for its classification: another iridovirus-like, icosahe- cytoplasm of infected, circulating erythrocytes. The cy- dral cytoplasmic virus (African swine fever virus) poss- toplasmic virions were enveloped by 2 smooth mem- essing double-stranded DNA was found to differ con- branes (Figs. 2 & 3). In damaged cells, these envelop- siderably from iridoviruses, and is considered to ing membranes were seen to form clots which were belong to a separate family (Willis 1990). tied with the nuclear membrane (Figs. 3 & 4). Within the virosome, the polygonal shell was often The virus was assembled within spherical, fibrillar incomplete and lacked a dense inner component. virosomes in the cytoplasm. Inside these virosomes, in- These particles probably represent developing virions complete viral capsids, not associated with membranes and indicate the virosome to be the viral assembly site. and lacking a dense nucleoid, could be found (Figs. 5 & Similar viral assembly sites have been found in other 6). Intranuclear particles were sometimes observed, as studied indoviruses, particularly in FV3 infections. well as breaks in the nuclear membrane through However, in several insect iridoviruses and in Lympho- which the virus could enter the nucleus (Fig. 7). cystis disease virus of fish, the assembly sites have dif- Budding virions were never found. ferent ultrastructural characteristics (Devauchelle et Infected cells contained large crystalloid inclusion al. 1985). The presence of intranuclear virions seems to bodies within their cytoplasm (Figs. 8 & 9) or, more be related to the entry of virions into the nucleus rarely, in the nucleus (Figs. 10 & 11). Infected proeryth- through breaks in the nuclear membrane, although the roblasts and circulating erythroblasts with viral fea- presence of an early intranuclear replication step of the tures similar to those seen in the peripheral blood iridovirus genome (Willis et al. 1990) could conceivably erythrocytes were found in the spleen (Figs. 12 & 13) raise also the possibility of an intranuclear assembly of and in the lungs (Fig 14).Many of the cells were in the virions. process of being phagocytized by macrophages Virions accumulating in the cytoplasm were sur- (Fig. 15). Other cell types, such as the macrophages, rounded by 2 membranes. These probably represent were never seen to be infected. endoplasmic reticulum cisternae enveloping the cap- Clusters of membrane-associated virions were found sid, as suggested by Sousa & Weigl (1976). The associ- within pulmonary capillaries. These virions were being ation of the membranes with the nuclear envelope was phagocytized by the latter's endothelium (Fig. 16). No evident in lysing cells and supports this hypothesis. evidence of productive infection was observed in lung Similar membranes have been reported in the other parenchyma cells. studied frog erythrocyte icosahedral viruses (Bernard et al. 1968, Gruia-Gray et al. 1989). Bernard et al. (1968) suggested these membranes to represent the DISCUSSION residual reorientation of endoplasmic reticulum from juvenile erythrocytes in hematopoietic zones, which The virus observed in the South African frog was had not been resorbed during maturation of the in- morphologically similar to the erythrocytic virus found fected cell. The infected, immature erythrocytes ob- by Sousa & Weigl (1976) in a Brazilian frog, but was served in this study already possessed membrane- about two-thirds of the size of the erythrocytic viruses enclosed virions. Since the capsid unit membrane Alves de Matos & Paperna: Erythrocytic virus of Ptychadena anchietae 109

of iridoviruses is synthesized in the virosome Anatomy Department, Curry Cabral Infectious Diseases (Devauchelle et al. 1985), the increase in the amount of Hospital, for their help and support- Thanks also to Dr Moura cytoplasmic membranes could be related to changes in Nunes of the Cancer Institute. Lisbon, for providing technical facilities. Collection and preparation of material for electron the erythrocyte, linked to the synthesis of the viral unit microscopy in south Africa was out during I.P.'s CSIR- membranes. sponsored visit to the Zoology and Entomology Department of Many iridoviruses leave the cell via a budding mech- the Orange Free State University. Bloemfontein. anism (Devauchelle et al. 1985) However, budding was not observed in this study. The association of the virions with the cytoplasmic membranes could be pre- LITERATURE CITED venting them from reaching the membrane and Alves de Matos, A. P,Paperna, 1. (1993) Ultrastl-uctural study engaging in budding. Instead, the virus seems to be re- of Pirhemocvton virus in lizard ervthroc~tes. Annls leased by cell lysis. Most of the circulating viruses Parasitol. hum. cornp. 68. 24-33 L., found in lunq capillaries- were enclosed in membrane Arcay de Peraza, McLure, M T (1971). The 'Paranuclear clots similar to those found in lysing cells. corpuscules' in poik~lotherrnalvertebrates. 11. Description of a new species of Toddia in Elect]-ophorus electricus The crystalloid body is characteristic of erythrocytic (electric eel), with an expansion of the key to the species of virus infections of frogs and clearly distinguishes them the genus Toddia in poik~lothermalvertebrates. Acta biol. from the ervthrocvtic iridovirus-like infections in sau- venez. 7: 201-209 rian reptiles (Alves de Matos & Paperna 1993) and fish Bernard, G. W., Cooper, E L., mand dell, M. L. (1968).Lamellar (Smail 1982). The crystalloid body was distinctly stri- membrane encircled viruses In the erythrocytes of Rana pipjens. J. Ultrastruct Res. 26: 8-16 ated, suggesting a proteinaceous composition, which is Desser, S. S., Barta, J. R. (1984) An intraerythrocytlc vlrus and different from the structure of the 'vacuoles' in saurian rickettsia of frogs from Algonquln Park, Ontario. Can. J. reptile infections, as reported by Stehbens & Johnston ZOO^.62: 1521-1524 (1966) and also from our own observations (Alves de Devauchelle, G., Stoltz, D. B., Darcy-Tripier, F. (1985). Comparative ultrastructure of Iridovirldae. In: Willis. D. B. & Matos Paperna 1993). However the significance (ed.)Iridoviridae. Springer-Verlag, Berlin, p. 1-21 of this feature is unclear, since it remains to be Gruia-Gray, J'., Petric, M., Desser, S. (1989). Ultrastructural, determined if the crystalloid body consists of virally biochemical and biophysical properties of an erythrocytic coded material (probably protein) or host-dependent virus of frogs from Ontario. Canada. J. Wildl. Dis. 25: material. 497-506 Johnston, M. R. L. (1975). Distribution of Pirhemocyton No clinico-pathological changes except for cytologi- Chatton & Blanc and other, possibly related, infections of cal damage as direct consequence of viral activity in poikilotherms. J. Protozool. 22: 529-535 the infected cell were so far demonstrated in any of the Marquardt, W. C.. Yaeger, R. G. (1967). The structure and tax- reported erythrocytic viral infections of frogs (Bernard onomic status of Toddia from the cottonmouth snake Agkistrodon piscivorus leucostoma. J. Protozool. 14: 1968, Sousa 1976, Desser 1984, Speare 1991, Paperna 726-731 & Alves de Matos 1993). This contrasts with the ex- Paperna, I., Alves de Matos, A. P. (1993). Erythrocytic viral treme anemia characteristic of erythrocytic viral infec- infections of lizards and frogs: new hosts, geographical tions in saurian reptiles (Paperna & Alves de Matos locations and description of the infection process. Annls 1993). Parasitol. hum. cornp. 68: 11-23 Smail. D. A. (1982). Viral erythrocytic necrosis in fish: a In this study we could find no evidence of infected review. Proc. R. Soc. Edinb. 81: 169-176 cells other than erythrocytic cells, thereby excluding a Sousa, M. A., Weigl, D. R. (1976). The viral nature of Toddia potential relationship with non-erythrocytic frog irido- Franca, 1912. Memorias Inst. Oswaldo Cruz 74: 213-230 viruses. Infected erythrocytic cells at different stages of Speare, R., Freeland, W. J., Bolton, S. J (1991). A possible maturation were found in the spleen. Spleen macro- iridovirus infection in erythrocytes of Bufo marinus in Costa Rica. J. Wildl. Dis. 27: 457-462 phages were active in the phagocytosis of infected Stehbens, W. E., Johnston. M. R. L. (1966). The viral nature of cells, and are therefore probably important for their Pirhemocyton tarentolae. J. Ultrastruct. Res. 15: 543-554 removal and the animal's recovery (Paperna & Alves Stoltz, D. B. (1971). The structure of icosahedral cytoplasmic de Matos unpubl.). Another possible mechanism of deoxyriboviruses. J. Ultrastruct. Res. 37: 219-239 Willis, D. B. (1990). of iridoviruses. In: Darai, G. viral elimination is the phagocytosis of virions by pul- (ed.) Molecular biology of iridoviruses. Kluwer Academic monary endothelial cells. Publ., Boston, p. 1-12 Willis, D. B., ~horn~son,J. P., Beckman, W. (1990). Acknowledgements. The authors thank Prof. Eduardo Crespo Transcription of frog virus 3. In: Darai, G. (ed.) Molecular of the Department of Zoology and Anthropology, Lisbon biology of iridoviruses. Kluwer Academic Publ., Boston, University, and Dr Celeste Campos of the Pathologic p. 173-186

Responsible Subject Editor: P Zwart, Utrecht, Manuscript first received: September 22, 1992 The Netherlands Revised version accepted: March 6, 1993