Indirect immunofluorescence with rabbit antisera was used to probe the intracellular locations of the antigens of envelope, prM (precursor to structural protein M) and the nonstructural glycoproteins NS 1 (formerly described as NV 3 or SCF) specified by the flaviviruses dengue-2 and Kunjin. Perinuclear staining in various types of foci was prominent for all antigens, and the distribution was influenced by whether cells were fixed with acetone or formaldehyde. Staining of Golgi-like masses or inclusions by anti-envelope sera occurred regularly and prominently in cells infected and stained with homologous anti-envelope antibodies; in the cross reactions, such staining was largely absent, especially in dengue-2 infected cells in which it was replaced by many small circular foci scattered throughout the cytoplasm. Anti-NS 1 also stained large perinuclear inclusions and small cytoplasmic foci, but the distribution of these was dissimilar to that observed with anti-envelope sera. Anti-prM appeared to contain a mixture of antibodies of different specificities, evident at different dilutions, possibly because of different cytoplasmic locations of prM and its cleavage products. All antisera produced small discontinuous foci on the plasma membrane of unfixed infected cells; antigens of NS 1 were sometimes prominent on the surface of acetone-fixed cells.

Summary.Aspects of the problem of myeoplasma contamination of eell eultures have been examined. Particular attention has been paid to the effective eradication of myeoplasma contaminants. The effects of tylosin, a new antibiotic, are examined in detail.The work also describes attempts to eradicate a particularly resistant strain of M. orale type I, which has been isolated from contaminated cell lines in this laboratory. INTRODUGTION.In recent years use has been made of cell cultures as a method for the isolation of mycoplasmas from clinical specimens originating from patients with leukaemia, rheumatoid arthritis or Reiter's disease (Murphy and Furtado, 1963; Negroni, 1964;Grace, Horoszewicz, Stim and Mirand, 1963; Grace, Horoszewicz, Stim, Mirand and James, 1965;Bartholomew and Himes, 1964;Bartholomew, 1965). The validity of this technique relies heavily on the methods of detecting mycoplasmas in cell culture so as to ensure that they are free of contaminating mycoplasmas at the time of use.The isolation or detection of myeoplasma contaminants of cell cultures is not an easy matter. Many methods have been proposed for the detection of mycoplasmas in cell cultures. These have included the direct visualization of myeoplasma by staining of contaminated cell sheets (Fogh and Fogb, 1964; Glyde, 1961), the use of chemical methods based on the presence of arginine dehydrolase in mycoplasmas (Barile and Schimke, 1963;House and Waddell, 1966) and the culture of supernatant fluid or disrupted cells in suitably supplemented agar medium. Methods of detection of myeoplasma in cell cultures have been reviewed in detail by Hayflick (1965).

Short communications 8Ihave the ring-like appearance and the central nucleoid may have been particles of rubella virus (Fig. I).The failure of Chatterji et al. to observe similar particles in the large amount of control material examined is difficult to explain, as both the strains of virus which they used had been passed in the same line of RK I3 cells just before the experiments which they describe and one of the strains, I7L had only been passed in these cells since it was isolated. This, together with the observation that many of the virus harvests from the untreated Australian cells show higher infectivity titres of mycoplasma than of virus, must raise the possibility that even though the mycoplasma does not influence the virus, the converse might not always be true. However, studies to date have not indicated any difference between titres of mycoplasma attained in rubella-infected and control cultures.We are grateful to Professor B. P. Marmion for his advice and for the isolation of M. orale, Type I, from the Manchester line of RK 13.