Download mycoplasmas in tissue culture

J. Cell Sci. I, 145-168 (1966)
Printed in Great Britain
MYCOPLASMAS IN TISSUE CULTURE
I. MACPHERSON
Medical Research Council Unit for Experimental Virus Research,
Institute of Virology, University of Glasgow
SUMMARY
Mycoplasmas are frequently found as contaminants in tissue-cultured cells. Infections may
be inapparent or cause severe cytopathic changes. The source of primary contaminations in
most cases is probably the upper respiratory tract of man. The wide dissemination of infections
in cultures most probably occurs as a result of aerosols set up during the processing of contaminated cultures. Mycoplasmas in cultured cells may cause chromosome aberrations,
degradation of the host cell DNA, and morphological transformations. They cleave thymidine
and its related structural inhibitors and also degrade arginine. They inhibit the growth of
adenovirus and Rous sarcoma virus and no doubt affect others. A number of antibiotics which
are relatively non-toxic for cells in culture are active against mycoplasmas and may be used to
cure infected cells. Mycoplasmas interfere with the biochemistry of the cell at many points and
no one working with tissue cultures can afford to ignore them. Stringent aseptic techniques are
the best safeguard against primary infections and cross-contaminations.
CONTENTS
PAGE
Introduction
146
Frequency of mycoplasma contamination of tissue cultures
146
Manifestations of mycoplasma growth in cells
Inapparent infections
Cytopathic changes
Experimental infections
Microscopic appearance of infected cells
Biochemical basis of the cytopathic effect on cultured cells
Other effects on the nutrition of tissue cultures
147
147
147
148
148
149
151
Mycoplasmas in relation to virology and oncology
Similarities between mycoplasmas and viruses
Association of mycoplasmas with neoplastic tissue
Effect of mycoplasma contamination on virus growth in tissue culture
153
153
152
153
Detection of mycoplasma contamination
Changes in the gross appearance of cultures
Microscopic appearance of infected cells
Methods based on biochemical reactions
Isolation of mycoplasmas in artificial media
155
155
155
155
156
Sources of mycoplasma contaminants
157
Serotypes of mycoplasma contaminants in cultures
Control of mycoplasma contamination
Conclusions
10
158
160
163
Cell Sci. 1
146
/. Macpherson
INTRODUCTION
Mycoplasma contamination of tissue cultures has been recognized for almost
10 years, but it is only comparatively recently that an adequate understanding has
been gained of the source of the contaminants and the reasons for their wide dissemination in cultures. There have also been many recent observations on the undesirable results of such contamination. The present review discusses the problem
and possible methods of control.
FREQUENCY OF MYCOPLASMA CONTAMINATION OF TISSUE CULTURES
Since the first demonstration by Robinson, Wichelhausen & Roizman (1956) that
a number of cultured cell lines were contaminated with mycoplasmas, there have been
similar reports from many laboratories using cultured cells. Summarizing the published
results from ten reports, 267 out of 454 (59 %) of cultures examined were contaminated
with mycoplasmas. Successful isolations probably give a conservative estimate of the
proportion of cultures actually infected, since cultures with low levels of contamination
may not always yield positive isolations on artificial media. Repeated attempts may be
necessary to reveal contamination in some cultures (Macpherson & Allner, i960;
Pollock, Kenny & Syverton, i960). Freshly trypsinized cells have a reduced mycoplasma content and the stage of growth of a culture will influence the yield of mycoplasmas. Kraemer, Defendi, Hayflick & Manson (1963) have shown that the conventional media used in most of these surveys were, in some instances, less sensitive
for the detection of mycoplasma tissue-culture contaminants than seeding test material
into a line of mouse lymphoma cells in vitro. This line showed cytopathic changes
when infected with some mycoplasmas. The lymphoma cells also supported the
growth of some non-cytopathogenic strains of mycoplasma from tissue cultures. A
preliminary passage of the mycoplasmas in lymphoma cells produced sufficient enrichment to ensure their subsequent detection on solid media.
With the new awareness of mycoplasmas as potential contaminants many workers
with tissue cultures tested them for the presence of these organisms. It soon became
apparent that primary cultures prepared from a variety of human and animal tissues
were rarely contaminated. Carski & Shepard (1961) found that each of two primary
cultures from rabbit and hamster kidney were free of mycoplasmas. Barile, Malizia &
Riggs (1962) found that only two out of 150 primary cultures of rhesus monkey kidney,
and none of 60 rabbit kidney cultures, were contaminated. Rothblat (i960) noted that
primary cultures from a variety of species were rarely contaminated and Herderschee,
Ruys & van Rhijn (1963) found that 26 primary cell cultures were free from mycoplasmas. Others have made similar observations (Macpherson & Allner, i960;
Kraemer et al, 1963; Gori & Lee, 1964).
Although cell lines that are continuously propagated differ in a number of respects
from primary and early passage cultures (Hayflick & Moorhead, 1961), the latter are
just as capable as the former of supporting the growth of mycoplasmas, and the lower
incidence of contamination in these cells is not due to some inherent resistance.
Mycoplasmas in tissue culture
i/^j
Human diploid cells that have been in culture for several months may carry mycoplasmas (Herderschee et al., 1963). Primary cultures and cell cultures in early passage
seem only to have had less opportunity of coming into contact with the sources of
contamination.
MANIFESTATIONS OF MYCOPLASMA GROWTH IN CELLS
Inapparent infections
In many cultures found to be contaminated with mycoplasmas the changes in the
cells have been minimal or inapparent (Rothblat, i960; Carski & Shepard, 1961;
Kraemer et al., 1963) and may be insufficient to cause the worker to seek their replacement or even to suspect there is anything amiss. Even heavy contamination may not
be accompanied by damage to the cells or turbidity of the medium.
More commonly, some minor degree of cellular damage can be detected. Cells grow
more slowly, are more granular, tend to come off the glass more readily, fail to form
a continuous sheet, are more susceptible to trypsin, and produce more acid in the
medium. Small necrotic foci may develop (Robinson et al., 1956). Some of these
changes may pass unheeded if the worker receives the cells already contaminated and
believes these manifestations to be characteristic of the cell line.
Cytopathic changes
In some cases a more striking cytopathic effect is caused by mycoplasma contaminants. Stim, Grace & Moore (1963) found a cytopathic agent with characteristics
of a mycoplasma in a spontaneously degenerated culture of hamster tumour cells. It
caused a cytopathic effect in primary bovine kidney cells but not in primary kidney
cultures derived from some other species. As mentioned previously, Kraemer (1964 a)
found that a cultured line of mouse lymphoma cells designated L51J8 Y responds to
mycoplasmas isolated from a number of cell lines by undergoing lysis in a characteristic
manner. At first some agglutination of the cells occurred, then cessation of metabolism
was indicated by a rise in pH, followed finally by lysis and clearing of the cultures.
The final pH of the infected cultures was about 7-5, whereas the controls, which had
been incubated for the same time and were turbid with cells, had a final pH of about
6-9. Kraemer described instances in which attempted isolations of mycoplasmas from
human diploid cell cultures produced no growth of colonies on agar medium but in
which lytic infections were initiated when the same material was inoculated into
L51J8Y cells. Numerous colonies developed when medium from lysed cultures was
plated on agar. He also observed that individual colonies picked from cultures prepared from contaminated strains of human diploid cells and inoculated into L5178Y
cells could produce lytic infections, whereas other colonies, although establishing
infections, were non-lytic. These two types could, on occasion, be isolated from the
same cell culture and bred true when passaged. Non-lytic mycoplasmas caused no
change in the growth rate of infected cells and could be recovered only sporadically
from the cultures. In general it was found that contaminated cultures that gave rise to
high yields of mycoplasmas on direct plating mostly contained non-lytic organisms
148
/. Macpherson
that grew rapidly on agar. Cultures yielding few colonies on direct plating contained
predominantly lytic mycoplasmas that grew slowly on agar.
When tested in the L51J8 Y cells, the strains Campo, T-5, Eaton agent and seven
unclassified human pharyngeal isolates were non-lytic. With lytic strains chronic
infection could be established in HeLa, chick and mouse embryo and monkey kidney
cultures, but no cytopathic effect was produced and the cultures grew as rapidly as the
controls. On the other hand, the mouse ascites cell line P388D1 growing in vitro was
rapidly lysed. It is of interest that both L51J8 Y and P388DX cells originated in
DBA/2 mice. It is possible that cells of this mouse strain are unusually susceptible
to mycoplasmas or may contain latent viruses that are ' ignited' by the mycoplasmas.
As will be mentioned later, Kraemer suggests that their unusually high susceptibility
may be related to their inability to substitute citrulline for the degraded arginine. In
culture both these cell lines attach tenuously or not at all to the glass in stationary
culture and thus may be vulnerable to changes in their cytoplasmic membranes
mediated by mycoplasmas.
It has been shown that HeLa cells, known to be contaminated with mycoplasmas,
grew satisfactorily when attached to glass but failed to grow in stirred suspension
culture (Macpherson & Allner, i960). If these cells were apparently cleared of mycoplasmas with neomycin they became capable of active division in suspension culture.
Moore, Mount, Tara & Schwartz (1963) have also reported that several established
lines of cells derived from human tumours failed to survive in suspension culture if
they were contaminated with mycoplasmas. On the other hand, Brownstein & Graham
(1961) were able to grow Earle's L strain of mouse fibroblasts in suspension culture
although they were heavily contaminated with mycoplasmas. The colony-producing
efficiency of the contaminated line on glass also remained unimpaired.
Experimental infections
A number of studies have been made on cultures deliberately infected with mycoplasmas (Hayflick & Stinebring, 1955; Wittier, Cary & Lindberg, 1956; Chanock et al,
i960; Fogh & Hacker, i960; Castrejon-Diez, Fisher & Fisher, 1963; Carmichael,
Fabricant & Squire, 1964). As in naturally occurring infections, there is a wide range
of activity, from no detectable changes, through mild cytopathic effects, to destruction
of the cells.
Microscopic appearance of infected cells
Microscopic examination of cells infected with mycoplasmas both by conventional
histological methods (Hayflick & Stinebring, 1955; Shepard, 1958; Fogh & Fogh,
1964; Marmion & Goodburn, 1961) and by thin-section electron microscopy (Edwards
& Fogh, i960; Fogh & Hacker, i960) reveals small masses of mycoplasmas within the
cytoplasm and on the plasma membrane. Micro-colonies have been found occasionally
in the nucleus (Carmichael et al., 1964). The majority of complete mycoplasma forms,
both ovoid and filamentous, are predominantly found on the plasma membrane of the
cells and, as will be discussed later, the high efficiency of specific anti-mycoplasma
serum for decontaminating tissue cultures (Herderschee et al., 1963) suggests that the
Mycoplasmas in tissue culture
149
main site for growth, or perhaps an essential phase of growth, is extracellular or on
the cell surface, since y-globulins do not penetrate the cell membrane of viable cells.
Edwards & Fogh (i960) estimate the size of the ovoid forms found in thin sections to
be 540 by 300 m/t. Smaller viable forms undoubtedly exist in some strains of mycoplasmas. Morowitz, Tourtellotte & Pollock (1963) found that 4 strains of mycoplasmas
could pass through a GS Millipore filter (0-22- ju, pore size). Electron-microscope studies
of mycoplasmas grown on artificial medium indicate that some forms may be 125 m/i
in diameter (Morton, Lecce, Oskay & Coy, 1954) or even smaller (A. Howatson, personal
communication).
Biochemical basis of the cytopathic effect on cultured cells
The nutrition and metabolism of mycoplasmas in artificial medium has been
extensively reviewed (Smith, 1964). Studies of this kind are technically difficult since
no chemically defined medium has yet been devised even for the least exacting
members of this group. Present knowledge is based on the results obtained with
partially defined medium and with only a few strains of mycoplasma. Nevertheless,
it is clear that there is wide variation in the nutritional requirements and metabolic
activities of different strains. Studies on amino acid metabolism of mycoplasmas by
Smith (1955, 19570, b, i960) have shown that certain strains metabolize glutamine,
glutamic acid and arginine. These are important metabolites for many strains of cells
in culture, and arginine is an essential amino acid for all strains so far studied.
The first indication that arginine depletion occurs in cultures infected with mycoplasmas was provided by Powelson (1961), who used cell lines massively infected with
mycoplasmas isolated from bovine and avian sources. Kenny & Pollock (1963) showed
that the unusual appetite of the mycoplasmas for arginine was directly implicated in
their cytopathic effect on cells. Cultures of cell lines derived from human liver and
foetal intestine, when experimentally infected, grew poorly in medium supplemented
with o-i mM arginine. Infected cultures grew as well as mycoplasma-free cultures
when grown in 1 mM arginine. Mycoplasma populations in the latter two cases were
similar, indicating that the higher arginine level was not inhibitory to mycoplasmas.
Powelson (1961) has suggested that the alterations she demonstrated in the metabolism of certain amino acids (including arginine) in infected cultures are dependent
on a mycoplasma/cell interaction. In the systems studied by Kenny & Pollock (1963)
this was not found to be so. Tissue-culture medium in which mycoplasmas had grown
and which was subsequently freed of mycoplasmas with kanamycin was found to be
deficient for cell growth. The addition of o-i mM arginine raised its growth-promoting
activity to that of the uninfected control medium.
Tissue-culture media are not always adequate for the growth of mycoplasmas.
Carski & Shepard (1961) found that solid media based on tissue-culture formulae did
not support the growth of mycoplasmas. Two media, (a) 40 % human serum plus
60% Hanks's balanced salt solution and (b) 20% human serum plus 10% bovine
foetal serum plus 70% balanced salt solution containing Eagle's amino acids and
vitamins and 1 mM arginine, were both inadequate in this respect. The latter formula
as a liquid medium also failed to support the growth of mycoplasmas.
150
J. Macpherson
Medium that had been in contact with HeLa cells for 3 days was also inadequate.
In the presence of HeLa cells growing in this medium mycoplasmas grew readily to
high titre. Powelson (1961) also found that medium 199 plus 2 % horse serum did not
support mycoplasma growth. However, Fogh, Hahn & Fogh (1965) found that
mycoplasmas isolated from several tissue cultures were capable of growing in cellfree tissue-culture medium. Growth of the mycoplasmas was enhanced by the addition
of yeast extract.
In a carefully controlled series of experiments Kraemer (1964&) has shown that the
lysis of the mouse lymphoma cell line L51J8Y induced by certain strains of mycoplasmas is due to depletion of arginine in the medium, and these cells may be protected
from the lytic action of the mycoplasmas by maintaining a high level of arginine in the
medium. Non-lytic strains of mycoplasmas that multiply to high titre and establish
chronic infections do not deplete the arginine in the medium to a significant extent.
Unlike the system described by Kenny & Pollock (1963), Kraemer found residual
arginine-depleting activity in mycoplasma-freed medium. Kraemer suggests that the
marked susceptibility of the L5178Y mouse lymphoma cells to the lytic agents may
be due to their inability to substitute citrulline for the degraded arginine. The scheme
suggested by Schimke & Barile (1963) would not support this view. They studied the
stoichiometry of arginine breakdown by mycoplasma-infected HeLa cultures, and
proposed the following pathway (which is the arginine dihydrolase system first described by Hills (1940) in Streptococcus) to explain their findings:
.
.
Arginine deiminase
Arginine
TV TTT
• Citrulline + NH 3
Omithine
Carbamyl phosphokinasc
CO2 + NH 3 = A T P '
Pi transcarbamylase
•II
=; Carbamyl phosphate +
ornithine
The breakdown of arginine is not due to arginase activity since neither mycoplasmacontaminated nor mycoplasma-free HeLa cells have measurable urease activity, and
this enzyme would be necessary to account for the absence of urea formation during
arginine degradation. Extracts of infected HeLa cells, as well as extracts of mycoplasmas in cell-free broth, rapidly converted arginine to ornithine. Since arginine
deiminase activity is not present in animal tissues, Barile & Schimke (1963) have suggested that the demonstration of this activity in tissue cultures may be used as a rapid
method for the detection of mycoplasma contamination. This will be discussed later.
Depletion of other metabolites in tissue-culture media by contaminating mycoplasmas undoubtedly occurs and may result in cell destruction (Fogh et ah, 1965), but
as yet there has been no report of specific effects (other than that concerning arginine)
which may be implicated in causing cell degeneration. Powelson (1961) noted that the
glutamine concentration in mycoplasma-infected chick heart fibroblasts was reduced
in comparison with the control. However, Kenny & Pollock (1963) found that
reduction of the glutamine level in infected cultures from 2 to 0-2 mM did not cause
detectable alterations in cell growth.
Mycoplasmas in tissue culture
151
One of the most important activities attributed to mycoplasma contaminants of
cultured cells was recognized by Hakala, Holland & Horoszewicz (1963). They had
previously found that a number of cell lines, including HeLa, when grown in the
presence of amethopterin, can utilize thymidine, thymidylic acid and 5-methylcytosine deoxyribonucleoside as sources of DNA thymidine (Holland et al., 1963).
Growth also occurred when this medium contained 5-bromodeoxyuridine in place of
thymidine. However, they found that growth of one strain of HeLa cell (subsequently
found to be contaminated with mycoplasmas and designated HeLa/PPLO) was not
supported by these deoxyribonucleosides in amethopterin-containing medium. The
cells had also become much more resistant to inhibition by 5-fluorodeoxyuridine.
Growth of HeLa/PPLO in amethopterin medium was not supported by the addition
of 5-bromodeoxyuridine or 5-iododeoxyuridine.
When HeLa/PPLO cells were incubated with thymidine this was rapidly broken
down. The observation that 5-fluorouracil and 5-fluorodeoxyuridine were equally
powerful inhibitors of HeLa/PPLO, while for another cell line (S-180) the nucleoside
was 43 times more inhibitory than the free base, suggested that 5-fluorodeoxyuridine
was also being cleared by HeLa/PPLO. This activity was indeed demonstrated. Since
HeLa/PPLO was found to be more sensitive to 6-mercaptopurine medium supplemented with folic acid than the HeLa cell line studied previously, they concluded that
the enzyme system for metabolizing hypoxanthine was not impaired. Studies with
mycoplasmas isolated from the HeLa/PPLO cells and also with four other strains of
mycoplasmas from various sources showed that they all were able to cleave thymidine
to thymine. The substrate specificity of the mycoplasma enzyme indicated that it is
similar to the pyrimidine nucleoside phosphorylase isolated from horse liver (Friedkin
& Roberts, 1954). From these results it is clear that any studies involving the metabolism of deoxyribonucleosides in mammalian cells must take into account the possible
contribution of mycoplasma contaminants. Two recent papers emphasize this point.
Randall, Gafford, Gentry & Lawson (1965) found that in HeLa cell cultures infected
with mycoplasmas the host-cell DNA is unstable, since acid-soluble radioactive label,
previously incorporated in the cell DNA, may be detected in the medium. This
characteristic can be transmitted to mycoplasma-free cultures of L cells by infecting
them with mycoplasmas from the HeLa cells. Nardone, Todd, Gonzalez & Gaffney
(1965) found that contamination of L-cell cultures by mycoplasmas inhibited incorporation of tritiated thymidine and uridine. Autoradiographs of contaminated
cultures were characterized by exposed silver grains at the cell margins. Treatment of
the cells with kanamycin restored normal nucleoside incorporation.
Other effects on the nutrition of tissue cultures
The addition of yeast extract and also Staphylococcus culture-medium filtrates
increases the cytopathic effect of mycoplasmas in tissue culture, probably by stimulating
the growth of the organisms (Wittier et al., 1956; Kenny & Pollock, 1963; W. House
(personal communication)). Edward (1947) found that yeast extract improved the
growth of mycoplasma in artificial media. The factor or factors in yeast extract
reponsible for this enhancement have not been characterized.
152
/. Macpherson
The reducing conditions in culture may also affect the growth of some mycoplasmas.
Herderschee et al. (1963) found that a mycoplasma-free line of HeLa cells failed to
support the growth of Mycoplasma salivarium, a strain that requires anaerobic conditions for optimal growth. If, however, 5 % Filde's extract (a peptic digest of sheep's
blood) was added M. salivarium grew well and gave a cytopathic effect. Filde's extract
by itself did not affect the growth of the cells. The authors suggest that the effect may
be due to catalase in the extract. The addition of yeast extract and/or Filde's extract
to tissue cultures may be a useful preliminary to testing these cultures for mycoplasmas,
since a low-grade infection may be stimulated.
MYCOPLASMAS IN RELATION TO VIROLOGY AND ONCOLOGY
Similarities between mycoplasmas and viruses
Since some mycoplasma strains produce cytopathic changes in cultured cells it is
important to recognize this fact when attempting to isolate infectious agents in tissue
culture. A 'filter-passing', cytopathogenic, transmissible agent may be either a virus
or a mycoplasma. As was mentioned previously, some mycoplasma forms pass through
o-zz-ji Millipore filters (Morowitz et al., 1963) and the smallest particle capable of
independent existence may be about 125 m/t (Morton et al., 1954). The similarity of
mycoplasma to some viruses, especially the myxoviruses, extends to: their appearance
in the electron microscope; sensitivity to ether and chloroform; inhibition of growth
by specific antiserum; interference with virus replication in vitro (Rouse, Bonifas
& Schlesinger 1963; Somerson & Cook, 1965; Ponten, to be published); ability of
some strains to haemagglutinate (Adler, 1954) and also to give rise to haemadsorption
on infected cells (Berg & Frothingham, 1961); their resistance to some antibiotics;
and the induction of chromosome aberrations (Paton, Jacobs & Perkins, 1965; Fogh &
Fogh, 1965).
Also, like the oncogenic viruses of polyoma and Rous sarcoma, some mycoplasmas
of human origin have been shown to mediate changes in the BHK 21 line of hamster
fibroblasts that enable the cells to form colonies in agar suspension culture and in some
cases also to undergo morphological transformation (Macpherson & Russell, to be
published).
Association of mycoplasmas with neoplastic tissue
In diseases in which the aetiological agent may be suspected of being a virus, on
the grounds that a similar syndrome has been shown to be due to a virus in other
species, the temptation to claim virus-like effects with isolated agents should be
particularly avoided. Agents isolated from the bone marrow of leukemic patients were
originally thought to be viruses, mainly because of the cytopathic effect they had for
human embryonic kidney cells in vitro (Negroni, 1964), but these agents were subsequently found to be mycoplasmas by Grist & Fallon (1964). Their suspicions were
aroused when the agents were found to be cytopathogenic for cell cultures derived
from a wide range of animal species. They were also able to culture typical mycoplasmas from the degenerating cultures and were able to show that the cytopathic
Mycoplasmas in tissue culture
153
effect was abolished in the presence of kanamycin. Girardi, Hayflick, Lewis &
Somerson (1965) have obtained similar results with the same group of agents. The
close association of mycoplasmas with leukemic tissue has aroused a great deal of
interest. Hayflick & Koprowski (1965) isolated these organisms directly from the bone
marrow of a leukemic subject, indicating that there may be a real association of mycoplasmas with leukemic tissue and that these agents are not due to contamination
taking place during tissue-culture passage. Others have found mycoplasmas in association with human leukemic and neoplastic tissue (Grace, Horoszewicz, Stim & Mirand,
1963), and Armstrong, Henle, Somerson & Hayflick (1965) have apparently recovered
mycoplasmas from cases of human haemangioma and retropharyngeal fibroma. These
organisms, like those isolated by Negroni, are cytopathogenic for cells in culture but
lack an antigenic identity in complement-fixation tests with known human mycoplasma
strains. The strain isolated by Negroni has complement-fixing antigens in common
with murine strains (M. pulmonis) (Fallon et al., 1965). Moore et al. (1963) were also
able to isolate mycoplasmas directly from the pleural fluid obtained from a patient
with a nasal septum adenocarcinoma. In their studies with continuously cultured cell
lines derived from such human tumours they found that primary cultures were in
several instances contaminated with mycoplasmas. It is possible that neoplastic tissue
provides a 'filter' or a particularly favourable site in vivo for the proliferation
or harbouring of mycoplasmas.
The observations of McGinniss, Schmidt & Carbone (1964) and Schmidt, Barile &
McGinniss (1965) that an association exists between the absence of the red blood cell
group I and neoplastic disease is particularly intriguing since many mycoplasmas seem
to possess enzymic activity against I antigen (Schmidt et al., 1965). It seems most likely
that organisms isolated from neoplastic tissues are ' passengers' but since some mycoplasmas can multiply intracellularly and, as described above, have the ability to
degrade DNA, cleave thymidine, cause chromosomal aberrations and transform cells
in vitro, their claims as carcinogenic agents cannot be dismissed. Whatever their role
in these situations, it is clear that cultures prepared from neoplastic tissues should be
screened with special care for mycoplasmas.
Effect of mycoplasma contamination on virus growth in tissue culture
In a number of instances the growth of viruses in tissue culture has been found to be
unaffected by the presence of large numbers of contaminating mycoplasmas. Herderschee et al. (1963) found that seven recently isolated polio viruses of type 1 and the
Sabin strain B1-3-F1 grew to approximately the same titre in a line of human
embryonic lung fibroblasts both in the presence and absence of contaminating mycoplasma (M. hominis type 1). Gori & Lee (1964) examined the sensitivity of a number
of mycoplasma-contaminated cell lines to vaccinia, poliovirus type 1 and measles virus,
both before and after the contaminating organisms had been eradicated. Those with
low levels of contamination (i.e. less than 100 colony-producing organisms per io6
cells) showed no difference in maximum yields of virus obtained. In heavily contaminated cell lines the yield of virus was slightly improved after treatment. Similar
results were obtained by Sever (see Gori & Lee, 1964) with rubella virus. Brownstein
154
-^ Macpherson
& Graham (1961) found that the single-burst yields of EMC virus from mycoplasmacontaminated L cells was lower than that obtained with this virus in uncontaminated
cells. They suggest the difference may have been due to the presence of the contaminating organisms. O'Connell, Wittier & Faber (1964) found that an unidentified
mycoplasma enhanced the cytopathogenic effect of a latent simian virus when it was
added to green monkey kidney cultures.
Pollock, Treadwell & Kenny (1963) have suggested that since inhibition by
ammonium ions of influenza virus formation has been demonstrated (Eaton & Scala,
1961; Jensen, Force & Unger, 1961) the presence of mycoplasmas may have an effect
on virus multiplication, since some strains produce ammonia. A striking example of
mycoplasma interference of virus growth has been provided by Somerson & Cook
(1965). They found that the growth of Rous sarcoma virus (RSV) was inhibited in
cultures of chick embryo cells infected by a strain of M. orale, a commensal organism
of the upper respiratory tract of man. This mycoplasma strain, originally isolated
from human diploid fibroblast cultures, was cytopathogenic but lost this property
when passaged on solid medium. Cytopathogenicity could be regained by passage in
chick embryo fibroblasts. Both cytopathogenic and non-cytopathogenic substrains
caused inhibition of RSV focus formation. Suppression of Rous-associated virus
(RAV) was also demonstrated by failure to detect avian leukosis complement-fixing
antigen in mycoplasma-infected chick-embryo fibroblasts inoculated with RSV. The
inhibitory effect was not extended to a strain of influenza B virus.
Similar results have been obtained by Ponten (1965) with a transmissible noncytopathic organism derived from a culture of human tumour cells. Although he was
unable to culture mycoplasmas from the culture fluids that initiated resistance to
RSV, the sensitivity of the agent to antibiotics indicated that it was probably a mycoplasma. Macpherson & Ponten (unpublished data) found that chicken fibroblasts
exposed to M. hominis type 1 mycoplasmas at a multiplicity of 1 colony-forming unit
per cell subsequently failed to develop any foci when challenged with RSV. In this
case the mycoplasma infection was quite inapparent. Inhibition of RSV and RAV by
mycoplasmas could clearly cause considerable confusion in the interpretation of the
interplay of these viruses.
Schlesinger (1961) found that adenovirus plaque-forming efficiency was enhanced
when the arginine concentration in the medium was increased. This effect was subsequently correlated with the presence of mycoplasmas in the cultures (Rouse et al.,
1963). Type 2 adenovirus growing in KB or other cells was found to have an absolute
requirement for arginine in the medium. As stated above, some mycoplasmas rapidly
deplete arginine in the medium of the cultures they contaminate. It is likely that other
viruses, such as herpes simplex, that are dependent on an adequate supply of arginine
for maturation, may be similarly affected by mycoplasma contamination of the cells
in which they are grown.
Mycoplasmas in tissue culture
155
DETECTION OF MYCOPLASMA CONTAMINATION
Changes in the gross appearance of cultures
The spectrum of morphological alterations that may occur in mycoplasma-contaminated cultures has already been described. It is important to note that in some
cultures these changes may be so small as to pass unnoticed. However, in many other
instances there are fairly obvious indications that the cells are growing suboptimally.
Microscopic appearance of infected cells
By fixing and staining cells grown on cover glasses it is often possible to detect
small round ovoid or filamentous structures stained with the specificity of DNA on
the cells. Intensified Giemsa-staining described by Marmion & Goodburn (1961)
has been used to reveal mycoplasma contamination of cells in culture (Butler & Leach,
1964; see Fig. 1). Eaton, Farnham, Levinthal & Scala (1962) found that May-GrunwaldGiemsa stain demonstrated M. pneumoniae in cultures of human amnion and human
embryonic lung. Fogh & Fogh (1964) recommend a hypotonic treatment, air-drying
and orcein-staining procedure based on the methods commonly employed in karyological analysis. Mycoplasmas are easily seen in the extended cytoplasm of these cells.
In Fig. 2 the individual mycoplasmas are shown stained by Giemsa in a chromosome
preparation of BHK21 hamster cells. It is doubtful if any staining method would be
adequate to detect low-grade contamination of cells and it would certainly be unwise
to rely solely on staining methods for the routine examination of cultures for contamination. Nevertheless, it is a worthwhile preliminary and should reveal gross
contamination. The use of direct fluorescent antibody methods on infected cells suffers
from the same defect and also from the fact that, although many current contaminants
are of the same serological type (see below in section on typing), there are exceptions
which would be missed if a mono-specific antiserum were used. The use of multivalent
labelled antisera would not be practical. It may also be noted here that cells heavily
contaminated with mycoplasmas adsorb fluorescent globulin non-specifically on their
surface (Macpherson, unpublished observation).
Methods based on biochemical reactions
As previously stated, many strains of mycoplasmas possess arginine deiminase and
thymidine-cleaving activity. The enzymes catalysing these reactions have not been
found in animal cells. Colorimetric methods for their detection have been devised
and suggested as suitable for the detection of mycoplasma contamination of tissuecultured cells.
Barile & Schimke (1963) have described a method for the assay of arginine deiminase activity in cell extracts. The enzyme is detected by formation of citrulline
from arginine at pH 6-5, the citrulline being determined colorimetrically. Barile &
Schimke estimated the total number of organisms required for a positive reaction to
occur in their test at between io 5 and io6. This estimate was made both on broth
cultures and on contaminated cell-culture lysates. Contaminated cell cultures examined
generally contained at least 10 times more mycoplasmas than specified for the test.
156
I. Macpherson
They believe the sensitivity of the test could be improved by extending the incubation
period for the enzyme reaction. The results obtained were in agreement with cultural
procedures in 73 cases examined. Since some strains of bacteria also possess arginine
deiminase activity, potential false results with the arginine deiminase method for
mycoplasma contamination could be obtained in the presence of these bacteria. In
practice activity was found only in association with cells contaminated with mycoplasmas. However, as bacterial infection is equally undesirable, its detection by this
method is no disadvantage.
Horoszewicz & Grace (1964) make use for detection of the mycoplasmas' ability to
cleave thymidine. The amount of free deoxyribose resulting from the degradation of
thymidine is measured. Of 42 strains grown in broth culture, 38 gave positive reactions.
The demonstration of activity requires disruption of the organisms. In the cell culture
test from IO 6 -IO 7 disrupted cells are necessary. Uninfected L-g2g cells produced only
0-117 /tmoles of deoxyribose per min per mg of protein under test conditions, whereas the
same cells infected with a mycoplasma (strain 880) produced 5-50 times more. When
73 cell cultures of 20 different cell lines were examined for thymidine-cleaving activity,
46 cultures were positive and 27 were negative. The results agreed with isolations
made on artificial media. However, House (personal communication) has found that
thymidine-cleavage activity may be low in some cultures heavily contaminated with
mycoplasmas.
Isolation of mycoplasmas in artificial media
A variety of artificial media have been successfully used for cultivation of mycoplasmas. Usually they are meat-infusion or digest broths to which have been added
10-20% serum and 5-10% yeast extract. The yeast extract described by Hers (see
Lemcke, 1965) has been found to be of particular value in the stimulation of mycoplasma growth (Lemcke, 1965; House, personal communication). For solid media a
soft gel is produced by the addition of agar or agarose to the enriched broth. The
grade of agar is important since some types of agar contain inhibitors (Lynn &
Morton, 1956).' Bacto' PPLO agar (Difco) is usually suitable, although some organisms
fail to grow on this medium (Randall et ah, 1965).
Some difficulty may be experienced in identifying mycoplasma colonies by those
unfamiliar with their morphology. Typically they are like fried eggs in that they have
an inner dense region and an outer less-dense area when examined by low-power
microscopy. A microscope with a total magnification of 100 diameters is suitable.
A number of artifacts may simulate mycoplasma colonies (Hayflick, 1965). Of these,
disrupted tissue-cultured cells and accretions of calcium and magnesium soaps from
the medium (Brown, Swift & Watson, 1940) cause most difficulty. A preliminary
culture in broth inoculated with test cells and medium followed by plating on agar
eliminates the former difficulty and will increase the mycoplasma population. Since
mycoplasmas usually grow into the agar a useful test is to attempt to scrape a suspected
colony from the surface of the plate. The central core of a colony of mycoplasmas will
remain behind but artifacts or pseudocolonies will be dislodged. However, on first
isolation from tissue culture some mycoplasmas form atypical colonies (Herderschee
Mycoplasmas in tissue culture
157
et ah, 1963) without the typical fried-egg appearance of strains well adapted to growth
on solid medium. The concentration of the agar will also influence the appearance of
the colony. If the agar is too hard the colonies will be entirely on the surface and may
be reduced in size and number (Macpherson, unpublished observation).
It is therefore important to ensure that desiccation of the medium does not occur
during incubation. Incubation under anaerobic conditions (95 % nitrogen with 5 %
carbon dioxide) may be superior to aerobic conditions for the isolation and growth of
some mycoplasmas from tissue culture (Barile, Yaguchi & Eveland, 1958; Butler &
Leach, 1964; Herderschee et al., 1963). Cultures should be made both aerobically and
anaerobically for initial isolation. Incubating cultures in closed jars that have been
flushed out with 5 % carbon dioxide in nitrogen is adequate for decreasing the oxygen
content of the atmosphere. A closed container prevents the agar medium from drying
and stiffening.
When there is doubt that a structure is a mycoplasma colony, a further check may
be made by attempting to stain it with Dienes's stain (Dienes & Weinberger, 1951).
This stain (a mixture of maltose, methylene blue and azure II) is taken up by mycoplasma and bacterial colonies. The latter decolorize the stain after about 30 min but
the colonies of mycoplasmas retain the stain. However, colonies of dead bacteria may
retain the stain. Impressions of colonies may also be fixed on to a slide by cutting out
a block of agar medium bearing the colonies and inverting it on a microscope slide.
The slide is then immersed in Bouin's fixative for several hours and, after the removal
of the agar block, washed and then stained with Giemsa. The stained impression of
the colony has a characteristic vacuolated appearance within which the granular large
'bodies' of the mycoplasma may be distinguished using high-power microscopy.
These vary in staining intensity and range from 0-3 to i-o/i in diameter.
The use of the enriched culture medium employed for the isolation of mycoplasmas
has the added advantage that it is also capable of supporting the growth of bacteria
which would normally be missed because of their inability to grow on standard
bacteriological media. In the course of screening cultures for mycoplasma contamination, Coriell (i960) detected persistent infection of a culture with diphtheroid bacteria.
The bacteria produced no gross microscopic change in the tissue cultures nor did they
grow in thioglycollate broth or standard blood agar plates. If antibiotic-free tissueculture medium was used the organisms grew out and destroyed the cells. Barile &
Schimke (1963) also found low-grade bacterial contamination (100—1000 bacteria
per ml) in 20 out of 73 cell-culture media tested.
SOURCES OF MYCOPLASMA CONTAMINANTS
There has been much speculation about the source of mycoplasma contaminants in
tissue culture. No single explanation satisfactorily accounts for all contaminations.
However, it is important to detect the main sources of contamination to enable prophylactic measures to be taken. As will be discussed later, there is good evidence that
the majority of contaminations result from droplet spray infections from cultures
already contaminated, but the source of primary contaminations is perhaps less
158
I. Macpherson
certain. The following possibilities suggest themselves: the tissue used to initiate the
culture was infected with mycoplasmas; a constituent of the medium contained
mycoplasmas; the mycoplasmas are in fact L-forms of contaminating bacteria, converted in vitro; and mycoplasmas from man have been introduced during processing
of the cultures.
Serotypes of mycoplastna contaminants in cultures
The most useful information for the tracing of micro-organisms is provided by
phage or serological typing. This latter technique may be used for mycoplasmas, and
determination of serotypes is an obvious first step in recognizing their source. The
evidence obtained by such surveys strongly suggests the last possibility offered above
as the most important mode of primary contamination, since mycoplasmas in tissue
culture belong almost exclusively to human serotypes.
Several methods of serological typing of mycoplasmas may be employed and good
qualitative correspondence occurs in results obtained by different methods. In Table 1
the results of several typing surveys are listed along with references. Clearly M. hominis
type 1 is the predominant organism. Organisms with this serotype have been isolated
from the genitalia of man and less frequently from the upper respiratory tract. In the
latter situation M. hominis type 1 is not the commonest commensal mycoplasma as
judged by isolations made on artificial medium (D. Taylor-Robinson, personal communication). If the upper respiratory tract is the primary source of most contaminations
then one might expect the common types of commensals (e.g. M. salivarium) to
Table 1. Serological typing of mycoplasmas isolated from tissue cultures
No. of
isolates
examined
32
16
49
8
5
4
15
8
1
Method of typing
Growth inhibition
Distribution of serotypes
(number)
Reference
M. hominis type 1 (31)
Herderschee et al.,
M. orale (1)
1963
Lemcke, 1964a, b
Complement fixation M. hominis type 1 ( n )
M. orale (5)
Fluorescent antibody Not typed but all interacted with Barile et al., 1962
M. hominis type 1 (49)
Agglutination and
Bailey et al., 1961
M. hominis type 1 (4)
growth inhibition
Growth inhibition
Clyde, 1964
M. hominis type 1 (4)
Patt strain (1)
Growth inhibition
M. hominis type 1(1)
Edward, i960
M. gallisepticum (1)
Unidentified (2)
Complement fixation All same but not M. hominis
O'Connell et al,
type 1 or 2 or M. salivarium
1964
Complement fixation ' Similar to human genital
Collier, 1957
PPLO'
Not M. hominis type 1 or 2, sali- Butler & Leach,
Complement fixavarium or Innes
1964
tion and growth
inhibition
Mycoplastnas in tissue culture
159
be the commonest tissue-culture contaminants. One possible explanation why this
is not so is that these strains do not adapt to tissue-culture conditions as readily
as M. hominis type 1. Herderschee et al. (1963) provide evidence to support this
belief. They found that under normal culture conditions M. salivarium failed to
initiate an infection in HeLa cells and disappeared after 10 days. If 5 % Filde's extract
(which contains catalase) was added to the medium M. salivarium grew well and was
cytopathic. Good evidence that culture-to-culture spread occurs has been provided
by O'Connell et al. (1964), Herderschee et al. (1963), Hakala et al. (1963), and
Randall et al. (1965). O'Connell et al. (1964) found that aerosols generated during
trypsinization of infected cultures contained mycoplasmas. They presented evidence
that a single mycoplasma cell is capable of initiating infection in a tissue culture and
showed that mycoplasmas could be transferred from one culture to another through
the use of a common burette for dispensing medium. The organisms they isolated
from 15 antibiotic-free tissue cultures were closely related serologically, but antiserum prepared against them failed to produce significant reactions with a number of
stock culture strains including M. hominis types 1 and 2 and M. salivarium. One
isolate was found to cross-react with an unclassified strain of mycoplasma isolated
from the throat of a child with severe pharyngitis, again suggesting that the upper
respiratory tract may have been the origin in this case.
Another strain of mycoplasma isolated from tissue cultures (Herderschee et al., 1963;
Lemcke, 1964 a) has been shown to have serological identity with M. or ale, another
commensal of the upper respiratory tract of man (Lemcke, 19646). Thus with a few
exceptions the organisms found in tissue cultures can be identified as having serological
cross-reactivity with strains from the upper respiratory tract of man. There now seems
little doubt that this is the primary source of contaminants.
However, other possible sources of contamination deserve consideration. We have
seen that primary cultures of normal tissue are rarely contaminated but cultures of
malignant tissue frequently contain mycoplasmas (Grace et al., 1963). Many attempts
to demonstrate mycoplasmas in the serum employed for tissue culture have failed, but
the fact that small forms of some mycoplasmas may pass through o-zz-fi Millipore
filters makes it possible for them to survive serum-processing if they are present on
occasion. If the serum retains its growth-promoting activity after heating at 56 °C for
the particular cells under study then it would seem to be a worthwhile additional safeguard against contamination, although Collier (1957) found that some mycoplasmas
survived in human serum deliberately seeded and heated at 55 °C for 1 h.
Further evidence that medium is not an important source of contamination is provided by the evidence of several workers who found that contaminated cells, once
cured by treatment with antibiotics, may remain free of mycoplasmas for many months
after the antibiotics have been removed (Pollock et al., i960). Chick-embryo extract is
a potential source of contamination, although it is not often used now. Mycoplasmas
have been isolated from chick embryos (van Herick & Eaton, 1945) and they would
certainly survive the processing procedures usually adopted for the preparation of
extracts. Edward (i960) has isolated an avian mycoplasma (M. gallisepticum) from a
tissue culture and suggested it may have been derived from chick-embryo extract.
i6o
/. Macpherson
The notion that most mycoplasmas in tissue are in fact L-forms, derived from bacteria
that had contaminated the cultures and then undergone conversion due to contact
with penicillin or some other constituent of the medium, is no longer acceptable. It is
very unlikely that the contaminating bacteria would convert so regularly into organisms with the serological specificity of M. hominis type i. However, it has been shown
that ' mycoplasmas' in some cultures are capable of giving rise to colonies of corynebacteria (Macpherson & Allner, i960; Carter & Greig, 1963) and Gram-negative rods
(Holmgren & Campbell, i960) under appropriate cultural conditions. It is uncertain
whether these instances represent cases of L-form conversion in tissue-culture or
whether the contamination was with ' mycoplasma' of human origin and these reverted
to their bacterial form when cultured on artificial medium after their sojourn in
tissue culture. There is of course much support for the idea that some, if not all,
mycoplasma are of bacterial origin (Dienes & Weinberger, 1951; Pease, 1965) but a
discussion of this possibility is beyond the scope of this review.
CONTROL OF MYCOPLASMA CONTAMINATION
Although the adage 'prevention is better than cure' certainly applies to mycoplasma contamination of tissue cultures, situations arise in which it is worth while
attempting to decontaminate an infected culture. Most reported methods involve the
use of antibiotics, but other means have been suggested.
Hayflick (i960) has shown that maintenance of contaminated HeLa and L cells at
41 °C for 18 h kills some strains of mycoplasma differentially, without damaging the
cells beyond recovery. Other workers have found this method to be unsatisfactory
because either the cells failed to recover or the mycoplasmas were not inactivated
(Herderschee et al., 1963; Balduzzi & Charbonneau, 1964).
Another approach is that described by Kenny & Pollock (1963) and Herderschee
et al. (1963) who made use of the inhibitory effect of specific antiserum on homologous
mycoplasmas (Edward & Fitzgerald, 1954). They found that passage of the cells in
medium containing the appropriate antiserum results in the disappearance of the
mycoplasmas. The method has some serious disadvantages. Antiserum must be
specific for the mycoplasma strain and, although M. hominis type 1 is most frequently
isolated from tissue cultures, other strains of human mycoplasmas and some of
unknown origin (Butler & Leach, 1964) are being isolated as tissue-culture contaminants with increasing frequency (Taylor-Robinson, personal communication). It is
doubtful if a sufficient range and volume of potent specific antisera could be obtained
for this purpose. The method may have special applications, since it is the least cytotoxic and therefore the least selective to the cell population, in cases where damage to
a valuable cell line must be avoided. The contaminating organisms could be isolated
and the cells stored whilst antiserum was being prepared.
Antibiotic therapy offers the best chance of successful decontamination. A number
of antibiotics are known to be effective against mycoplasmas in tissue culture at levels
which are non-toxic or relatively non-toxic to the cells harbouring them. Successful
decontamination of cultures has been reported using neomycin (Macpherson &
Mycoplasmas in tissue culture
161
Allner, 1960), tetracycline (Carski & Shepard, 1961), kanamycin (Pollock et ah, i960;
Fogh & Hacker, i960; Herderschee et ah, 1963), 7-chlortetracycline (Gori & Lee,
1964), and a mixture of choramphenicol and novobiocin (Balduzzi & Charbonneau,
1964). The recommended concentration of antibiotic and the duration of the treatment
vary considerably in different reports. Factors which will influence the result include
the strain of contaminating mycoplasmas, the strain of cells, the composition of the
medium, the frequency of medium changes, the pH at which the culture is normally
maintained, and the method of subculturing. The most useful guide to the selection of
suitable antibiotics is provided by the work of Perlman & Brindle (1965) who studied
the toxicity of antibiotics for cells in culture and the sensitivity of 8 strains of mycoplasmas against 40 antibiotics. Their work is summarized in Tables 2 and 3. Unfortunately, treatment of contaminated cultures is not a straightforward procedure of
selecting an antibiotic effective at a non-cytotoxic level. Antibiotics have variable
stabilities in tissue-culture media and may be rapidly inactivated at 37 °C (Table 3).
Also, antibiotics are less effective against intracellular than against free bacteria (Murat,
Stinebring, Schaffer & Lechevalier, 1959) and presumably are less effective against
intracellular mycoplasmas. Although mycoplasmas may grow only on the surfaces of
cells and their appearance in the cytoplasm may be the temporary result of pinocytosis,
their sojourn there may render them resistant to antibiotics that fail to penetrate the
plasma membrane or do so only poorly. Finally, and most importantly, mycoplasma
strains may become resistant to certain antibiotics. This has been found with some
tissue-culture contaminants for kanamycin (Macpherson, unpublished data) and
tetracycline and chloramphenicol (Balduzzi & Charbonneau, 1964). It is possible that
resistance to most antibiotics would eventually be gained by mycoplasmas grown in
sub-inhibitory levels of them, and thus care must be exercised in using suitable antibiotics in the most effective way. The development of antibiotic-resistant variants is
to be avoided at all costs, since culture-to-culture spread is apparently common.
A preliminary to work with any cell line or cell strain should be the preparation of
a master seed, stored with appropriate additives (Wallace, 1964) in liquid nitrogen or
Table 2. The effect of antibiotics on mycoplasma and on tissue-cultured cells
(From Perlman & Brindle, 1965.)
Antibiotics essentially inactive
against mycoplasma in vitro
Antibiotics active against mycoplasma but too cytotoxic for use
in tissue culture
Antibiotics useful in controlling
mycoplasma contamination in
tissue culture
11
Amphomycin, Amphotericin B, Bacitracin,
Benzylpenicillin, Candicidin A, Cycloheximide, Cycloserine, Etruscomycin,
Filipin, Griseofulvin, Nystatin, Oleandomycin, Patulin, Polymyxin, Ristocetin,
Streptomycin, Trichomycin, Vancomycin,
Vernamycin A and B, Viomycin
Carbomycin, Dactinomycin, Streptothricin,
Stendomycin, Thiostrepton
See Table 3
Cell Sci. 1
162
7. Macpherson
Table 3. Antibiotics useful in controlling mycoplasma contamination
in tissue cultures
(From Perlman & Brindle, 1965.)
Antibiotic
Chloramphenicol
7-Chlortetracycline
6-Demethyl-7-chlortetracycline
Erythromycin
Fusidic acid
Gentamicin
5-Hydroxytetracycline
Hygromycin B
Kanamycin
Neomycin B
Novobiocin
Paromomycin
Spiramycin
Tetracycline
Tylosin
Stability
in tissue
culture
media*
High
Very low
High
Moderate
High
High
Moderate
Moderate
Very high
Very high
Low
High
Moderate
Moderate
Moderate
Minimum
concentration Concentration
Concentration
inhibiting
recommended
showing
mycoplasma for controlling
marked
in artificial mycoplasma in
cytotoxicity
medium
tissue culturesf
(mcg/ml)
(mcg/ml)
(mcg/ml)
3°
30
80
15
40
100
is
5
10
300
40
3000
15
5°
20
20
1
200
35
5
10
300
10000
3000
15
25
15
5°
200
10
200
5°
5°
5°
5°
5000
1000
20
35
2
10
300
1
10
1
* Stability scale: half-life of 2 days, very low; 4 days, low to moderate; 8 days, very high,
f Recommended on basis of 3-day incubation period between changes of medium.
a deep-freeze. It is then advisable to determine the maximum tolerated doses for the
recovered cells of a number of antibiotics effective against mycoplasmas. Since the
likelihood of a mycoplasma strain gaining resistance simultaneously to structurally
and biochemically different antibiotics is low, combinations of such antibiotics should
also be tested for their toxic level. Great care should be taken to ensure that the
master seed is free of mycoplasmas by testing and by treatment with combinations of
antibiotics. The levels used for the hamster line BHK21 (Macpherson & Stoker, 1962)
in the author's laboratory are: kanamycin, 1 mg/ml; novobiocin, 50 /*g/ml; and
chlortetracycline, 5 /ig/ml. These antibiotics are incorporated into medium which is
changed every 2 days for 6 days on an initially sparse culture. Treatment is followed
by repeated testing for mycoplasmas by as many means as possible but especially by
culturing both directly on to solid medium and also following an intermediate passage
in broth.
The observations of Gori & Lee (1964) suggest that conventional antibiotic treatment may at best be capable only of depressing a mycoplasma infection and may fail
to eliminate the organisms completely even after prolonged treatment. This was so
when a near cytotoxic level of tetracycline was used in tissue-culture medium for up
to 5 months. To eradicate mycoplasmas from cell lines they subjected the cells to
Mycoplasmas in tissue culture
163
treatment for a few minutes with very high concentrations of aureomycin, kanamycin,
and chloramphenicol in water. This was followed by cultivation of the cells in medium
with a lower concentration of antibiotics before final omission of the antibiotics for
2 medium changes. Cell disintegrates were then tested for mycoplasma on solid
medium. If a positive isolation of mycoplasmas was obtained, the cell culture was
subjected to another cycle of hypotonic antibiotic treatment. Only aneuploid cell lines
like HeLa survived this treatment and it was quite unsuitable for diploid cell lines.
If the observations of Gori and Lee are confirmed in other systems and complete
eradication of mycoplasmas requires the drastic treatment they describe, then the
use of short-term cultures derived from replicates of a frozen master seed will become
desirable in order to reduce the possibility of contamination. This will be especially
important for cell strains that do not survive the treatment they prescribe. It is also
clear that a general improvement of aseptic techniques for tissue-culturing will result
in a lower incidence of contamination.
CONCLUSIONS
Although it is apparently widely appreciated that cultures of mammalian cell lines
may become contaminated with mycoplasmas, it is also clear that few workers know
if their cultures are contaminated or indeed ever take steps to find out. Studies with
infected cells in the fields of virology, biochemistry, immunology and oncology are
subject to profound misinterpretations, and mycoplasma infections may remain undetected for many months if surveillance is not carried out routinely.
In experiments in which mycoplasma contaminations could conceivably produce a
spurious result, a study of the consequences of deliberately infecting the cell system
with mycoplasma (e.g. a common tissue-culture contaminant like M. hominis type 1)
could be an informative control.
Of particular interest are the recent observations that cell-associated mycoplasmas
may cause degradation of the cell DNA and induce chromosome abnormalities. This
activity could be one cause of the aneuploidization of cells that is often correlated with
their emergence from diploid cultures as continuously growing cell lines.
It seems possible that cell-dependent mycoplasmas could exist and be present in
tissue cultures. Such organisms would not be detected by the usual culturing procedures but would probably be capable of exerting some of the undesirable effects
described in this review.
Mycoplasmas show such great biochemical diversity that there is little doubt that
new effects of their impingement on the physiology of cultured cells will continue to
be described, and past errors in interpretation revealed.
I wish to thank Professor M. G. P. Stoker for his advice and criticism.
164
/. Macpherson
REFERENCES
ADLER, H. E. (1954). A rapid slide agglutination test for the diagnosis of chronic respiratory
disease in the field and in laboratory infected chickens and turkeys. Proc. Am. vet. med. Ass.
pp. 346-35°ARMSTRONG, D., HENLE, G., SOMERSON, N . L. & HAYFLICK, L. (1965). Cytopathogenic myco-
plasmas associated with two human tumours. I. Isolation and biological aspects. J. Bad. 90,
418-424.
BAILEY, J. S., CLARK, H . W., FELTS, W. R., FOWLER, R. C. & BROWN, T . M . (1961). Antigenic
properties of pleuropneumonia-like organisms from tissue cell cultures and the human
genital area. J. Bad. 82, 542-547.
BALDUZZI, P. & CHARBONNEAU, R. J. (1964). Decontamination of pleuropneumonia-like organism (PPLO) infected tissue cultures. Experientia 20, 651.
BARILE, M. F., MALIZIA, W. F. & RlGGS, D. B. (1962). Incidence and detection of pleuropneumonia-like organisms in cell cultures by fluorescent antibody and cultural procedures.
J. Bad. 84, 130-136.
BARILE, M. F. & SCHIMKE, R. T . (1963). A rapid chemical method for detecting PPLO contamination of tissue cell cultures. Proc. Soc. exp. Biol. Med. 114, 676-679.
BARILE, M. F., YAGUCHI, R. & EVELAND, W. C. (1958). A simplified medium for the cultivation
of pleuropneumonia-like organisms and L-form sof bacteria. Am. J. din. Path. 30, 171176.
BERG, R. B. & FROTHINGHAM, T . E. (1961). Hemadsorption in monkey kidney cell cultures of
mycoplasma (PPLO) recovered from rats. Proc. Soc. exp. Biol. Med. 108, 616-618.
BROWN, T . M., SWIFT, H . F. & WATSON, R. F. (1940). Pseudo-colonies simulating those of
pleuropneumonia-like micro-organisms. J. Bad. 40, 857-867.
BROWNSTEIN, B. & GRAHAM A. F. (1961). Interaction of Mengo virus with L cells. Virology
14, 3O3-3iiBUTLER, M. & LEACH, R. H. (1964). A mycoplasma which induces acidity and cytopathic effect
in tissue culture. J. gen. Microbiol. 34, 285-294.
CARMICHAEL, L. E., FABRICANT, J. & SQUIRE, R. A. (1964). A fatal septicemic disease of infant
puppies caused by cytopathogenic organisms with characteristics of mycoplasma. Proc. Soc.
exp. Biol. Med. 117, 826-833.
CARSKI, T . R. & SHEPARD, M. C. (1961). Pleuropneumonia-like (mycoplasma) infections in
tissue-culture. J. Bad. 81, 626-635.
CARTER, G. R. & GREIG, A. S. (1963). The recovery of diphtheroids from L-type organisms
contaminating tissue cultures. Can. J. Microbiol. 9, 317-320.
CASTREJON-DIEZ, J., FISHER, T . N . & FISHER Jr., E. (1963). Experimental infection of tissue
cultures with certain mycoplasma (PPLO). Proc. Soc. exp. Biol. Med. 112, 643-647.
CHANOCK, R. M., FOX, H . H., JAMES, W. D., BLOOM, H. H . & MUFSON, M. A. (i960). Growth
of laboratory and naturally occurring strains of Eaton agent in monkey kidney tissue culture.
Proc. Soc. exp. Biol. Med. 105, 371-375.
CLYDE Jr., W. A. (1964). Mycoplasma species identification based upon growth inhibition by
specific antisera. J. Immun. 92, 958-965.
COLLIER, L. H. (1957). Contamination of stock lines of human carcinoma cells by pleuropneumonia-like organisms. Nature, Lond. 180, 757-758.
CORIELL, L. L. (i960). Detection and elimination of contaminating organisms. Natn. Cancer
Inst. Monogr. 7, 33-53.
DIENES, L. & WIENBERGER, H. J. (1951). The L-forms of bacteria. Bad. Rev. 15, 245-288.
EATON, M. D., FARNHAM, A. E., LEVINTHAL, J. D . & SCALA, A. R. (1962). Cytopathic effect
of the atypical pneumonia organism in cultures of human tissue. J. Bact. 84, 13301337EATON, M. D. & SCALA, A. R. (1961). Inhibitory effect of glutamine and ammonia on replication of influenza virus in ascites tumor cells. Virology 13, 300—307.
EDWARD, D. G. ff. (1947). A selective medium for pleuropneumonia-like organisms. J. gen.
Microbiol. 1, 238-243.
EDWARD, D. G. ff. (i960). Ann. N.Y. Acad. Sci. 79, 608-609 (contribution to discussion).
Mycoplasmas in tissue culture
165
D. G.ff.& FITZGERALD, W. A. (1954). Inhibition of the growth of pleuropneumonialike organisms by antibody. J. Path. Bact. 68, 23-28.
EDWARDS, G. A. & FOGH, J. (i960). Fine structure of pleuropneumonia-like organisms in pure
culture and in infected tissue culture cells. J. Bact. 79, 267-276.
EDWARD,
FALLON, R. J., GRIST, N. R., INMAN, D. R., LEMCKE, R. M., NEGRONI, G. & WOODS, D. A.
(1965). Further studies of agents isolated from tissue cultures inoculated with human
leukaemic bone-marrow. Br. med. J. 2, 388-391.
FOGH, J. & FOGH, H. (1964). A method for direct demonstration of pleuropneumonia-like
organisms in cultured cells. Proc. Soc. exp. Biol. Med. 117, 899-901.
FOGH, J. & FOGH, H. (1965). Chromosome changes in PPLO-infected FL human amnion cells.
Proc. Soc. exp. Biol. Med. 119, 233-238.
FOGH, J. & HACKER, C. (i960). Elimination of pleuropneumonia-like organisms from cell
cultures. Expl Cell Res. 21, 242-244.
FOGH, J., HAHN, E. & FOGH, H. (1965). Effects of pleuropneumonia-like organisms on cultured
human cells. Expl Cell Res. 39, 554-566.
M. & ROBERTS, D. (1954). The enzymatic synthesis of nucleosides. II. Thymidine
and related pyrimidine nucleosides. J. biol. Chem. 207, 257-66.
GIRARDI, A. J., HAYFLICK, L., LEWIS, A. M. & SOMERSON, N. L. (1965). Recovery of
mycoplasmas in the study of human leukemia and other malignancies. Nature, Lond. 205,
188-89.
GORI, G. B. & LEE, D. Y. (1964). A method for eradication of mycoplasma infections in cell
cultures. Proc. Soc. exp. Biol. Med. 117, 918-921.
GRACE Jr., J. T., HOROSZEWICZ, J. S., STIM, T. B. & MIRAND, E. A. (1963). Unpublished data
cited by Hakala et al. (1963).
GRIST, N. R. & FALLON, R. J. (1964). Isolation of viruses from leukaemic patients. Br. med.J.
FRIEDKIN,
2,
1263.
M. T., HOLLAND, J. F. & HOROSZEWICZ, J. S. (1963). Change in pyrimidine deoxyribonucleoside metabolism in cell culture caused by Mycoplasma (PPLO) contamination.
Biochem. biophys. Res. Commun. 11, 466-471.
HAYFLICK, L. (i960). Decontaminating tissue cultures infected with pleuropneumonia-like
organisms. Nature, Lond. 185, 783-784.
HAYFLICK, L. (1965). Tissue culture and mycoplasmas. Tex. Rep. Biol. Med. 23, 285-303.
HAYFLICK, L. & KOPROWSKI, H. (1965). Direct agar isolation of mycoplasmas from human
leukemic bone marrow. Nature, Lond. 62, 199.
HAYFLICK, L. & MOORHEAD, P. S. (1961). The serial cultivation of human diploid cell strains.
Expl Cell Res. 25, 585-621.
HAYFLICK, L. & STINEBRING, W. R. (1955). Intracellular growth of pleuropneumonia-like
organisms. Anat. Rec. 121, 477-478.
HAKALA,
HERDERSCHEE, D., RUYS, A. C. & RHIJN, G. R. VAN (1963). Pleuropneumonia-like organisms
in tissue cultures. Antonie van Leeuwenhoek 29, 368-376.
W. VAN & EATON, M. D. (1945). An unidentified pleuropneumonia-like organism
isolated during passages in chick embryos. J. Bact. 50, 47-51.
HILLS, G. M. (1940). Ammonia production by pathogenic bacteria. Biochem. J. 34, 1057-1069.
HERICK,
HOLLAND, J. F., MINNEMEYER, J., GRACE Jr., J. T., BLOCK, R., O'MALLEY, J. & TIECKEL-
MANN, H. (1963). 5-Allyl-2-deoxyuridine (AUdR) activity on metabolism of pyrimidine
deoxynucleosides by HeLa cells infected with mycoplasma. Proc. Am. Ass. Cancer Res.
(Abs.), p. 29.
HOLMGREN, N. B. & CAMPBELL Jr., W. E. (i960). Tissue cell cultures contamination in relation
to bacterial pleuropneumonia-like organisms-L form conversion. J. Bact. 79, 869-874.
HOROSZEWICZ, J. S. & GRACE Jr., J. T. (1964). PPLO detection in cell culture by thymidine
cleavage. Bact. Proc. p. 131.
JENSEN, E. M., FORCE, E. E. & UNGER, J. B. (1961). Inhibitory effect of ammonium ions on
influenza virus in tissue culture. Proc. Soc. exp. Biol. Med. 107, 447-451.
KENNY, G. E. & POLLOCK, M. E. (1963). Mammalian cell cultures contaminated with pleuropneumonia-like organisms. J. inject. Dis. 112, 7-16.
KRAEMER, P. M. (1964a). Interaction of mycoplasma (PPLO) and murine lymphoma cell
cultures: prevention of cell lysis by arginine. Proc. Soc. exp. Biol. Med. 115, 206-212.
166
/ . Macpherson
P. M. (19646). Mycoplasma (PPLO) from covertly contaminated tissue cultures:
differences in arginine degradation between strains. Proc. Soc. exp. Biol. Med. 117, 910-918.
KHAEMER, P. M., DEFENDI, V., HAYFLICK, L. & MANSON, L. A. (1963). Mycoplasma (PPLO)
strains with lytic activity for murine lymphoma cells in vitro. Proc. Soc. exp. Biol. Med. 112,
381-387.
LEMCKE, R. M. (1964a). The serological differentiation of mycoplasma strains (pleuropneumonia-like organisms) from various sources. J. Hyg., Camb. 62, 199-219.
LEMCKE, R. M. (19646). The relationship of a type of Mycoplasma isolated from tissue culture
to a new human oral Mycoplasma. J. Hyg., Camb. 62, 351-352.
LEMCKE, R. M. (1965). Media for Mycoplasmataceae. Lab. Pract. 14, 712-715.
LYNN, R. J. & MORTON, H. E. (1956). The inhibitory action of agar on certain strains of
pleuropneumonia-like organisms. Appl. Microbiol. 4, 339-341.
MCGINNISS, M. H., SCHMIDT, P. J. & CARBONE, P. P. (1964). Close association of I blood
group and disease. Nature, Lond. 202, 606.
MACPHERSON, I. A. & ALLNER, K. (i960). L-forms of bacteria as contaminants in tissue culture.
Nature, Lond. 186, 992.
MACPHERSON, I. A. & STOKER, M. (1962). Polyoma transformation of hamster cell clones—an
investigation of genetic factors affecting cell competence. Virology 16, 147-151.
MARMION, B. P. & GOODBURN, G. M. (1961). Effect of an organic gold salt on Eaton's primary
atypical pneumonia agent and other observations. Nature, Lond. 189, 247-248.
MOORE, G. E., MOUNT, D., TARA, G. & SCHWARTZ, N. (1963). Growth of human tumor cells
in suspension cultures. Cancer Res. 23, 1735-1741.
MOROWITZ, H. J., TOURTELLOTTE, M. E. & POLLOCK, M. E. (1963). Use of porous cellulose
ester membranes in the primary isolation and size determination of pleuropneumonia-like
organisms. J. Bact. 85, 134-136.
MORTON, H. E., LECCE, J. G., OSKAY, J. J. & COY, N. H. (1954). Electron microscope studies
of pleuropneumonia-like organisms isolated from man and chickens. J. Bact. 68, 697-717.
KHAEMER,
MURAT, A., STINEBRING, W. R., SCHAFFNER, C. P. & LECHEVALIER, H. (1959). Screening for
antibiotics active against intracellular bacteria. Appl. Microbiol. 7, 109-114.
R. M., TODD, J., GONZALEZ, P. & GAFFNEY, E. V. (1965). Nucleoside incorporation
into strain L cells: inhibition by pleuropneumonia-like organisms. Science, N.Y. 149,
NARDONE,
IIOO-IIOI.
NEGRONI, G. (1964). Isolation of viruses from leukaemic patients. Br. med. J. 1, 927-929.
O'CONNELL, R. C , WITTLER, R. G. & FABER, J. E. (1964). Aerosols as a source of widespread
mycoplasma contamination of tissue cultures. Appl. Microbiol. 12, 337-342.
G. R., JACOBS, J. P. & PERKINS, F. T. (1965). Chromosome changes in human diploid
cell cultures infected with mycoplasma. Nature, Lond. 207, 43-45.
PEASE, P. E. (1965). lj-forms, episomes and auto-immune disease. Edinburgh: Livingstone.
PERLMAN, D. & BRINDLE, S. A. (1965). Antibiotic control of mycoplasma in tissue cultures.
A. Meet. Am. Soc. Microbiol.
POLLOCK, M. E., KENNY, G. E. & SYVERTON, J. T. (i960). Isolation and elimination of pleuropneumonia-like organisms from mammalian cell cultures. Proc. Soc. exp. Biol. Med. 105,
PATON,
10-15.
M. E., TREADWELL, P. E. & KENNY, G. E. (1963). Mammalian cell cultures contaminated with pleuropneumonia-like organisms. II. Effect of PPLO on cell morphology, in
established monolayer cultures. Expl Cell Res. 31, 321-328.
PONTEN, J. (1965). Suppression of Rous virus induced transformation by an antibiotic sensitive
mycoplasma-like factor. Wenner-Gren Center International Symposium on Comparative
Leukemia Research. Stockholm: Pergamon Press.
POWELSON, D. M. (1961). Metabolism of animal cells infected with mycoplasma. jf. Bact. 82,
288-297.
RANDALL, C. C , GAFFORD, L. G., GENTRY, G. A. & LAWSON, L. A. (1965). Lability of hostcell DNA in growing cell cultures due to mycoplasma. Science, N.Y. 149, 1098-1099.
ROBINSON, L. B., WICHELHAUSEN, R. H. & ROIZMAN, B. (1956). Contamination of human cell
cultures by pleuropneumonia-like organisms. Science, N.Y. 124, 1147-1148.
ROTHBLAT, G. H. (i960). PPLO contamination in tissue cultures. Ann. N.Y. Acad. Sci. 79,
430-432.
POLLOCK,
Mycoplasmas in tissue culture
167
H. C , BONIFAS, V. H. & SCHLESINGER, R. W. (1963). Dependence of adenovirus
replication and inhibition of plaque formation by pleuropneumonia-like organisms. Virology
30, 357-365.
SCHIMKE, R. T. & BARILE, M. F. (1963). Arginine breakdown in mammalian cell culture
contaminated with pleuropneumonia-like organisms (PPLO). Expl Cell Res. 30, 593-596.
SCHLESINGER, R. W. (1961). Perspectives in Virology, 2, 69.
SCHMIDT, P. J., BARILE, M. F. & MCGINNISS, M. H. (1965). Mycoplasma (pleuropneumonialike organisms) and blood group I; associations with neoplastic disease. Nature, Lond. 205,
371-372.
SHEPARD, M. C. (1958). Growth and development of T strain pleuropneumonia-like organisms
in human epidermoid carcinoma cells (HeLa). J. Bad. 75, 351-355.
SMITH, P. F. (1955). Amino acid metabolism by pleuropneumonia-like organisms. I. General
catabolism. J. Bad. 70, 552-556.
SMITH, P. F. (1957a). Amino acid metabolism by pleuropneumonia-like organisms. II. Glutamine. J. Bad. 73, 91-95.
SMITH, P. F. (19576). Conversion of citrulline to ornithine by pleuropneumonia-like organisms.
J. Bad. 74, 801-806.
SMITH, P. F. (i960). Amino acid metabolism of PPLO. Ann. N.Y. Acad. Sci. 79, 543-550.
SMITH, P. F. (1964). Comparative physiology of pleuropneumonia-like and L-type organisms.
Bad. Rev. 28, 97-125.
SOMERSON, N. L. & COOK, M. K. (1965). Suppression of Rous sarcoma virus growth in tissue
cultures by Mycoplasma orale. J. Bad. 90, 534-540.
ROUSE,
STIM, T. B., GRACE Jr., J. T. & MOORE, G. E. (1963). Isolation of cytopathogenic pleuro-
pneumonia-like organisms in tissue culture. Bad. Proc. 22, 134.
R. E. (1964). Studies on preservation by freezing of human diploid cell strains.
Proc. Soc. exp. Biol. Med. 116, 990-998.
WITTLER, R. G., CARY, S. G. & LINDBERG, R. B. (1956). Reversion of a pleuropneumonia-like
organism to a corynebacterium during tissue culture passage. J. gen. Microbiol. 14, 763-774.
WALLACE,
{Received 28 November 1965)
168
/. Macpherson
Fig. i. Polyoma-transformed BHK 21 cells infected with mycoplasmas and stained by
the method of Marmion & Goodburn (1961). Mycoplasmas can be seen attached to the
cell surface.
Fig. 2. Cells of the BHK 21 line after the hypotonic treatment and air-drying procedures used for the demonstration of chromosomes. Mycoplasmas are easily detected
on the extended cytoplasm. Giemsa stain.
Journal of Cell Science, Vol. i, No. 2
1
V
' t / J'C
I. MACPHERSON
(Facing p. 168)