MICROBIOLOGY

REVIEWS
ELSEVIER FEMS MicrobiologyReviews 18 (1996) 225-236

Viruses, plasmids and other genetic elements of thermophilic and

hyperthermophilic Archaea
Wolfram Zillig a,*, David Prangishvilli a, Christa Schleper b, Marieke Elferink a,
Ingelore Holz a, Sonja Albers a, Davorin Janekovic a, Dorothee G6tz c
a Max Planck-lnstitutfi~r Biochemie, 82152 Martinsried, Germany
b Marine Sciences Institute, UniuersiO' of California, Santa Barbara, CA 93106, USA
c Unieersity of Waikato, Hamilton, New Zealand

Abstract
We review and update the work on genetic elements, e.g., viruses and plasmids (exluding IS elements and transposons) in
the kingdom Crenarchaeota (Thermoproteales and Sulfolobales) and the orders Thermococcales and Thermoplasmales in
the kingdom Euryarchaeota of the archael domain, including unpublished data from our laboratory. The viruses of
Crenarchaeota represent four novel virus families. The Fuselloviridae represented by SSVI of S. shibatae and relatives in
other Sulfolobus strains have the form of a tailed spindle. The envelope is highly hydrophobic. The DNA is double-stranded
and circular. Members of this group have also been found in Methanococcus and Haloarcula. The Lipothrixviridae (e.g., T
TV 1 to 3) have the form of flexible filaments. They have a core containing linear double-stranded DNA and DNA-binding
proteins which is wrapped into a lipid membrane. The 'Bacilloviridae' (e.g., TTV4 and SIRV) are stiff rods lacking this
membrane, but also featuring linear double-stranded DNA and DNA-binding proteins. Both virus types carry on both ends
structures involved in the attachment to receptors. Both types are represented in Thermoproteus and Sulfolobus. The
droplet-formed novel Sulfolobus virus SNDV represents the 'Guttaviridae' containing circular double-stranded DNA.
Though head and tail viruses distantly resembling T phages or lambdoid phages were seen electronmicroscopically in
solfataric water samples, no such virus has so far been isolated. SSV1 is temperate, TTV1 causes lysis after induction, the
other viruses found so far exist in cartier states. The hosts of all but TTVI survive virus production. We discuss the
implications of the nature of these viruses for understanding virus evolution. The plasmids found so far range in size from
4.5 kb to about 40 kb. Most of them occur in high copy number, probably due to the way of their detection. Most are
cryptic, pNOB8 is conjugative, the widespread pDLI0 alleviates in an unknown way autotrophic growth of its host
Desulfurolobus by sulfur reduction. The plasmid pTIK4 appears to encode a killer function, pNOB8 has been used as a
vector for the transfer of the lac S (fl-galactosidase) gene into a mutant of S. solfataricus.

Keywords: Archaea, hyperthermophilic; Archaea, thermophilic; Virus; Crenarchaeota; Sulfolobales; Thermococcales; Thermoplasmales:
Plasmid

Corresponding author

0168-6445/96/$32.00 © 1996 Federationof European MicrobiologicalSocieties. All rights reserved

PII $0168-6445(96)00014-9
226 W. Zillig et aL / F E M S Microbiology Reviews 18 (1996) 225-236

Contents

1. Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 226

2. Viruses of Crenarchaeota . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 227

2.1. Nature of the viruses . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 227
2.2. Virus host interactions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 229
2.3. Plaque tests . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 230
2.4. Receptors . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 232
2.5. Assembly . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 232

3. Origin and phylogeny of viruses . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 232

4. Plasmids . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 233

5. 'Sulfolobicins' . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 234

6. Construction of transforming (cloning) vectors based on genetic elements . . . . . . . . . . . . . . . . . . . . . . . . . . 235

Acknowledgements . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 235

References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 235

1. Introduction Halobacterium halobium, pHH1 [6] have been stud-

ied in molecular genetic detail. In the lowest branch
Genetic elements, e.g., viruses and plasmids may of the Euryarchaeota, the Thermococcales, harbor-
be considered 'fragments of life'. They constitute an ing the hyperthermophilic genera Thermococcus and
important fraction of biodiversity which is not con- Pyrococcus, several plasmids including pGT5 [7]
gruous to and appears to be largely independent of have been found recently, but no viruses are known
that of their hosts and they have certainly partici- so far. In the second lowest branch of this kingdom,
pated in generating the variety of extant organisms the Thermoplasmales, viruses have not been de-
via recombination processes. Studying emergence scribed to date, but several plasmids were found
and scope of this diversity should help in understand- recently. The Pyrococcus plasmid pGT5 has been
ing origin and evolution of life. Moreover, such utilized as basis for the construction of a shuttle
elements supply small and thus comprehensible vector for E. coli, Sulfolobus and Pyrococcus
genomes which have served in studying mechanisms (Aagaard and Garrett, personal communication). In
and controls of DNA replication and gene expression Crenarchaeota, four viruses of Thermoproteus tenax
in Bacteria and Eucarya. Because of their capability (Thermoproteales) TTV1 to 3, representing the novel
to be maintained and multiplied in their hosts, they virus family Lipothrixviridae, and TI'V4, represent-
have been utilized for developing transformation ing another, not yet validated virus family, the
vectors. 'Bacilloviridae', have been isolated (see [3] and [4]).
The third domain of life, Archaea, has been Only TTVI has been investigated in more detail,
divided into two kingdoms, Crenarchaeota and Eur- including sequencing of 85% of its genome (see [8]).
yarchaeota [1]. A number of viruses and plasmids In the other order of the Crenarchaeota, Sulfolob-
have been described for Euryarchaeota, especially ales, the Sulfolobus virus SSVI, again representing
Halobacteriales (the viruses are listed in [2], earlier a novel virus family, the Fuselloviriodae, has been
reviews are [3] and [4]). Some of these, e.g., studied extensively, including complete sequencing
Halobacteriophage qbH [5] and a large plasmid of [9], mapping of transcripts and promoters, and analy-
 W. Zillig et al. / F E M S M i c r o b i o l o g y Re~'iews 18 ( 1 9 9 6 ) 2 2 5 - 2 3 6 227

sis of lysogeny and of induction of virus production.
Two novel Sulfolobus viruses, SIRV and DAFV
have been found in isolates from Icelandic solfataras
[10]. SIRV another representative of the 'Bacillo-
L,iridae' has been partially sequenced and character-
ized (D. Prangishvilli and D. G~Stz, unpublished work
from this laboratory). Our knowledge of DAFV of
Desulfurolobus ambivalens and of a novel Sul-
.[blobus virus from New Zealand, SNDV (W. Zillig,
D. G~Stz, D. Janekovic and D. Langer, unpublished
work from this laboratory) is, however, superficial.
Some l0 plasmids, most of them cryptic, but one
conjugative, another one encoding a killer function
and a third involved in autotrophic growth by sulfur
reduction, have been found in Sulfolobales, the latter
in Desulfurolobus, the others in heterotrophic Sul-
folobus strains from Iceland [10], Japan (Hokkaido
[1 l]) and New Zealand (see below).
Here we present a short resum6 and an update of
the state of the field, including unpublished material
from our laboratory.
2. Viruses of Crenarchaeota
2. I. Nature of the t'iruses
The viruses of extremely halophilic and
methanogenic Euryarchaeota are so far, with two
exceptions (see below), of the same head and tail
type as T phages and lambdoid phages of Bacteria
(see [2] and Table 1). A striking example is the
temperate halobacteriophage q)H, which resembles
E. coli phage P1 not only in structure and replication
of its DNA but also in the type of the lysogeny
conferred to its host by a circular prophage. In view
of the incompatibility of the transcription systems of
Bacteria and Archaea [12], of the entirely different
lifestyles of the hosts - life at low ionic strength for
E. coli but at extreme salinity for Halobacterium -
and of the complexity of the viral genomes, stepwise
surmounting of the host range barriers between these
distant phyla during evolution appears highly im-
probable such that we feel forced to assume this
virus type existed already before the separation of
the bacterial from the archaeal lineage, which is a
very early event in the evolution of organisms [13,14].
In contrast, to the euryarchaeotal viruses, those of
Crenarchaeota all represent previously unknown
virus families. The spindle-like virus SSVI of Sul-
folobus shibatae, solfataricus and islandicus repre-
sents the Fuselloviridae which are characterized by
containing about 16 kb circular double-stranded
DNA, a DNA-binding protein and a lipid membrane
made up of viral hydrophobic coat proteins and host
lipids. A short tail is visible but has not been anal-
ysed in detail. At least 7 viruses very similar in
dimensions and shape and containing DNA of the
same size which cross-hybridized with that of SSVI,
but was different in restriction patterns have been
found in novel Sulfolobus isolates from Icelandic
solfataras (D. Prangishvilli and I. Holz, unpublished
work from this laboratory). A virus-like particle of
Methanococcus t~oltae, VLP J2 [ 15] and two viruses,
Hisl and His2 of Haloarcula hispanica (C. Bath and
M. Dyall Smith, personal communication) appear to

Table 1
Viruses of crenarcaeota
Designation Shape Size(length/width, rim) Type DNA size (kb) Sequence known Proteins Membrane Host Virus type
SSVI spindle 100/60 ccc 15.463 total >_ 3 + S. s o i l Fuselloriridae
TTV 1 flex. ill. 400/40 lin. ds 16 85% >_4 + T. tenax. LipothrixtJiridae
TTV2 flex. ill. 1200/20 lin. ds 16 - > 1 + T. tenax. Lipothrixt'iridae
TTV3 flex. ill. 2500/30 lin. ds 27 - > 1 + T. tenax. Lipothrixciridae
DAFV flex. ill. 2200/27 lin. ds 56 - > 3 + D. arab. Lipothrixciridae

TTV4 stiff rod 500/30 lin. ds 17 - > 1 - T. t e n a x "Bacillot, iridae"

SIRV stiffrod 950/26 lin. ds 33 15% >_ 2 - S. isl. "Bacillociridae"

SNDV droplet 180/80 ccc 20 - > 16 ? S. spec. novel

228 W. Zillig et aL/ FEMS Microbiology ReNews 18 (1996) 225-236

be of the same type, which thus occurs in both
archaeal kingdoms but not outside the archaeal do-
main.
A second novel family of viruses, the 'Bacillo-
viridae'(provisional name), has so far only been
found in Crenarchaeota, TTV4 in T. tenax (see [3],
[4] and [16]) and SIRV (S. islandicus rod shaped
virus) in S. islandicus [10]. These viruses are stiff
rods formed by a double helix consisting of linear
double-stranded DNA and DNA-binding protein(s),
wound up to a tube terminally plugged and equipped
with tail fibers for adsorption, in the case of SIRV to
pili of the host. They lack a membrane or hydropho-
bic coat. Early after discovery our 'isolate' of SIRV
turned out to be a mixture of variants which differed
by restriction patterns of their DNAs but could not
be distinguished by shape, size and SDS-PAGE pat-
tern of viral proteins. Stable strains were obtained by
cloning variants via single plaques. Though the nu-
cleotide sequences of corresponding genes of differ-
ent variants (DNA-binding protein and dUTPase)
showed up to 10% difference, the protein sequences
were almost identical (D. Prangishvilli and D. Gt~tz,
unpublished work from this laboratory).

Fig. 1. Virus-likeparticles in PEG precipitatesof supernatants of liquid samplesfrom Icelandic solfataras,in some cases attached to cells of
Thermoproteus sp. Negativelystained.
 w. Zillig et al./ FEMS Microbiology Reviews 18 (1996) 225-236
Viruses of a third novel family, the Lipothrixviri-
dae, again occurring in Crenarchaeota only, are
flexible filaments with a core of linear double-
stranded DNA plus DNA-binding proteins, wrapped
in a membrane consisting of host lipids and possibly
hydrophobic viral proteins and equipped with tail
structures apparently involved in adsorption to recep-
tors of the host. For three of these, TTV1 to 3, the
host is T. tenax [2,3,17]. The fourth, DAFV, was
found in a carrier state in a Desulfurolobus isolate
from Iceland [10].
A fourth novel virus family is so far only repre-
sented by one large, droplet-like virus, SNDV, found
in a Sulfolobus isolate from New Zealand. The
pointed end of the droplet is densely covered by a
beard of very thin, long fibers possibly involved in
the attachment to the host receptor (Fig. 1; W. Zillig,
D. Langer and D. Gtitz, unpublished data from our
laboratory) The DNA is circular and has a size of 20
kb. A helical structure is recognizable on the surface
of the coat. The DNA is not split by many restriction
enzymes, indicating extensive chemical modifica-
tion.
These viruses were all found in cultures of novel
Sulfolobus isolates. Virulent viruses, which might
have been expected in liquid samples from hy-
drothermal situations in solfataras, were so far not
encountered therein. Electron microscopy of concen-
trates of such samples which had been obtained by
PEG precipitation and differential centrifugation, ex-
hibited, however, various types of virus-like parti-
cles, sometimes attached to putative Thermoproteus
cells (Fig. 2). Some of these were of unusual head
and tail types, others were spindle-like but differing
in size from SSV1 in size, or rod-shaped, but differ-
ing from SIRV in dimensions. With the proper hosts,
it might be possible to propagate such viruses.
2.2. Virus host interactions
All these viruses were found associated with
cloned host strains. In the case of SSV1, the mode of
coexistence is true lysogeny. The prophage is inte-
grated in an arginyl tRNA gene by an integrase
whose gene is interrupted as the result of integration
[18,19]. A neighboring gene appears significantly
similar to a consensus of exisionase genes. Virus
production is switched on by the induction of the
transcription of a short DNA sequence between two
229
large transcription units which are oriented back to
back like pL and pR in E. coli phage 1. This
sequence is equipped with an unusual promoter,
which lacks the box A characteristic for archaeal
standard promoters and seems to be barred by a host
repressor in the lysogenic state (T. Singer, diploma
thesis, L.M.U., Munich, Germany). Induction of virus
production oc/I~K~curs via UV irradiation or mito-
mycin C or, very efficiently, via infection with the
virus SIRV, which is not multiplied under these
conditions [10]. The lysogenic state is unique in that
the components of the virus, except the DNA, are
present in the ceils, the coat proteins in regions of
the host membrane and the DNA binding protein in
the cytoplasm (W. Zillig, unpublished). What seems
to be induced is DNA replication, possibly because
the UV-inducible transcript, lind, acts as a primer.
The newly synthesized DNA appears to be wrapped
into its coat at virus membrane islands in the host
membrane. No virus particles were ever seen in the
cytoplasm of the host.
In the case of SIRV, linear viral DNA is found in
the host when no virus is produced. This ill-defined
carrier state may pass over into virus production in a
later growth phase of the culture of the virus carry-
ing host or after infection of a virus-free host. After
prolonged growth of a virus-carrying culture, curing
has also been observed. The controls involved in the
virus host interaction are badly understood. The in-
fection of an SSV l lysogen with SIRV has, however,
the same consequences as UV irradiation and mito-
mycin C, which in E. coli both trigger the SOS
system.
TTV1 also occurs in a carder state in its host, T.
tenax, which in this case passes over into virus
production when the carbon source has been almost
consumed [3,4]. As in the other systems, the cells
survive this productive phase. Nothing is yet known
on the relations of the other viruses to their hosts.
No lytic nor even truly virulent viruses have been
encountered so far, either because they did not sur-
vive outside of cells in the samples or because
lysogeny or carrier states help to escape prolonged
direct confrontation with the harsh solfataric envi-
ronment. This stress is illustrated by the observation,
that the DNA of SSV1 is rapidly inactivated by
nicking when liberated virus remains exposed to the
optimal low pH, 3, in the culture medium (W.D.
230 W. Zillig et al. / FEMS Microbiology Reviews 18 (1996) 225-236

Reiter, diploma thesis, L.M.U. Munich Germany).

The pH of typical Sulfolobus habitats in solfataric
fields is as low as 1.7 to 2.5.

2.3. Plaque tests

Plaque tests, which are instrumental for molecular

genetic virus work are not available for the viruses
of the obligatory strict anaerobe T. tenax, which
requires elemental sulfur for growth and could not
yet be plated as a lawn. Moreover, the necessity to
handle this host in the absence of oxygen hinders the
investigation of the interaction of virus and host.
This was the major reason to concentrate our efforts
on screening for genetic elements of aerobic, prefer-
ably heterotrophic Sulfolobus strains. Virus-free het-
erotrophic Sulfolobus species, e.g., S. solfataricus
and S. acidocaldarius yield good lawns when plated
in soft layers of about 0.2% gelrite, but not on
surfaces of 0.6 to 0.8% gelrite gels [20].
Good plaque tests have been established for SSVI
[20], with S. solfataricus strain P1 as host and for
SIRV, with S. islandicus strain RN2H 1 as host [ 10].
Because these viruses are not lytic, the plaques,
which are due to growth inhibition, are more or less
turbid. In case of SSV1, they do not show sharp
edges. The quality of the plaques depends on the
growth state of the lawns which can be manipulated
by the seeding density.
From a single colony of a Sulfolobus strain from
New Zealand, STH3/1, which carried the novel
virus SNDV, a cured daughter strain was obtained
which upon infection produced this virus. No plaques
were, however, recognized after infection of this host
with the virus, which is propagated in the late expo-
nential growth phase, where the lawn has almost
reached its final density (W. Zillig, unpublished).

Fig. 2. Viruses of Thermoproteus tenax and Sulfolobus sp. Nega-

tively stained. (a) TTV1, (b) TTV2, (c) TTV3, (d) TTV4, (e)
DAFV, (f) DAFV within and getting out of cell of D. ambivalens,
(g) SIRV (courtesy of D. Typke and W. Baumeister), (h) particles
of SIRV spanning two pili of S. islandicus, (i) SNDV, (k) SSV1,
single particles and particles attached to membrane vesicle, (1)
same in the course of liberation from cell of S. shibatae, (m) o,s,urn
SNDV particles within cell of Sulfolobus sp.
r~

i !
232 w. Zillig et al. / FEMS Microbiology Reviews 18 (1996) 225-236
2.4. Receptors
Both TTV 1 and SIRV attach to pili of their hosts
(see Fig. 1), the former to the tips, the latter laterally.
Both of these oblong viruses are on both ends
equipped with structures allowing attachment to their
receptors. As a consequence, SIRV can form 'string
ladders' by forming bridges between two pili.
SSV1 attaches to the membranes of vesicles de-
void of an S layer. Its receptor appears, therefore, to
be situated in the membrane. Infection of a host with
an intact S layer, e.g., S. solfataricus, thus poses a
problem. It might occur at perturbations of S layer
structure which are responsible for the characteristic
edged or angular appearance of Sulfolobus cells.
2.5. Assembly
The occurrence of complete virus particles in the
cytoplasm indicates that T r v 1 and DAFV are as-
sembled within the cell. In contrast, massive extru-
sion of SSV1 occurs in regions of the membrane
which are not or only loosely covered by an S layer
(Fig. 1) and no virus was ever seen within cells. We
have shown, that in SSV1 lysogens the highly hy-
drophobic viral coat proteins VP1 and VP3 are mem-
brane proteins of the host while the DNA-binding
protein VP2 occurs in a particulate fraction of the
cytoplasm (W. Zillig, unpublished). VP2 has been
shown to bind to circular DNA, especially viral
DNA in a cooperative manner (W.D. Reiter, diploma
thesis, L.M.U., Munich, Germany). Taken together,
these observations indicate that VP2 is somehow
stored inside the lysogenic cell, and forms a core
particle with viral DNA after DNA replication has
been induced. The latter then diffuses to islands of
virus membrane in the cell membrane and is wrapped
there into its coat. Nothing is known yet on the
integration of tail structures into the virus particle
nor on the assembly of the other viruses listed here.
3. Orion and phylogeny of viruses
It is usually assumed that viruses are fragments of
life which are composed of elements derived from
organisms. This claim is based on several arguments.
One is, that many viral genes have, indeed, high
similarity to cellular genes and thus very probably
share a common origin with the latter. Another, more
formal one is, that viruses by definition require hosts
providing essential functions for propagation. Both
arguments are not compelling. At least in the case of
rep proteins involved in RC type replication it ap-
pears, that plasmid and viral genes branch off before
the separation of Bacteria and Archaea [24]. Many
viral components do not find homologs in genomes
of organisms, e.g., viral envelopes, RNA replicases
of RNA viruses and reverse transcriptases of retro-
viruses etc. All components packed into the first
common ancestor of cellular life must have been
created in the course of prebiotic or chemical evolu-
tion and thus probably evolved there in correspon-
dence with each other and with many molecules and
systems which were not integrated into the first
autonomous entity. Some of these could have formed
partially autonomous assemblies, with small genomes
as essential components, depending on the metabolic
potential of the 'primordial soup' as 'superhost' in
much the same way as viruses on their hosts. After
the onset of cellular life such assemblies had sur-
vived when they were able to replace the metabolic
potential of the primordial soup (or whatever the
scene of chemical evolution was) by that of a host.
Viral genes derived from cellular homologs could
have been added to viral genomes in the long period
of interactive evolution of viruses and hosts.
As already discussed above, the similarities be-
tween the archaeal virus ~ H and the bacterial
phage P1 strongly indicate that this temperate phage
type already existed before the separation of the
Archaea from the Bacteria, which was the first
documented lineage diversion in cellular evolution
[13,14]. Further striking similarities between corre-
sponding features of an archaeal and a bacterial
virus host system are, (1) the close back to back
arrangement of the promoters of the two large early
transcription units both in the archaeal virus SSV1
(where they encode the early transcpts t5 and t6) and
the bacterial phage 1 (where they also code for the
two major early transcripts [9]); (2) the position of
the UV inducible genes controlling virus production,
the tin d gene in SSV1 and the C1 gene in phage P1,
between these promoters [9]; and (3) the site specific
integration of the genomes of both viruses into the
host chromosome involving an integrase and an exci-
 W. Zillig et al. / FEMS Microbiology Reviews 18 (1996) 225-236 233

sionase gene [19]. The resemblance of the latter
feature is even higher between SSVI and those
temperate bacteriophages and other bacterial genetic
elements in which integration occurs in an arginyl
tRNA gene of the host [ 18].
A particularly interesting hint about the early
evolution of viruses is the finding that the circular
genome of the Sulfolobus virus SSV1 consists of
two parts of almost equal size, one containing cys-
teine codons in its ORFs, the other completely de-
void of these, the former encompassing the early
transcription units for t5 and t6, the latter containing
the gene cluster encoding some viral structural pro-
teins and thus possibly other late functions too.
These two functionally different halves of the SSV1
genome, one using cysteine, the other not at all,
appear to constitute 'modules' [21] of different ori-
gin, the combination of which created the viral
genome. When that half which encodes similarities
to functions of bacterial phages originated from an
ancestral situation already existing when Archaea
and Bacteria parted (see above) the other one, not
yet utilizing cysteine, could even have a more
primeval, possibly prebiotic origin. Because it con-
tains the genes encoding structural proteins of the
virus, e.g., the viral envelope, which so far does not
find homologies in organismic structures, viruses
might indeed have existed before organismic life
started.
On the other hand, representatives of the
Fuselloviridae, e.g., SSVI, have only been found in
Archaea and were thus possibly conserved only in
that lineage or were in some way excluded from the
other domains. Lipothrixviridae and 'Bacilloviridae'
have so far only been found in crenarchaeota and are
therefore possibly younger than the deepest division
within the Archaea.
Unfortunately, the fast evolution of viruses im-
peding significant sequence comparison, and the ex-
cessive exchange of modules [21] between different
viruses prohibit construction of (possibly indepen-
dent) phylogenetic dendrograms for these genetic
elements.
4. Plasmids
About a dozen plasmids have so far been identi-
fied in Sulfolobus strains (Table 2). Due to the way
of isolation as cccDNA via ethidium bromide/CsCi
buoyant density centrifugation, most of these occur
in high copy number. Some are, however, not or
barely seen in total DNA of their hosts or were only
found when they were isolated via enrichment from
larger amounts of total DNA, and thus appear to
occur in low copy number.
Most of these plasmids are cryptic, i.e., no func-
tions except maintenance have yet been uncovered.
However, a plasmid of Desulfurolobus ambivalens,
pDLI0, has been found in 20 of 22 independent
isolates of this relative of Sulfolobus, which thrives
autotrophically, alternatively by the oxidation or by

Table 2
Plasmids of sulfolobales
Designation Host DNA size Copy number Function
pRNI S. islandicus 5.5 20 crypto
pRN2 S. islandicus 6.9 35 crypto
pilE7 S. islandicus 7.5 15 crypto
pSTH4 S. spec. (N.Z) 4.7 high crypto
pTAU4 S. spec. (N.Z) 6.2 high crypto
pWHII S. spec. (N.Z) 6.8 low crypto
pWHI2 S. spec. (N.Z) 15.4 high crypto
pTIK4 S. spec. (N.Z) 14.3 high Sulfolobicin?
pSTH3 S. spec. (N.Z) 35.0 high crypto
pNOB8 S. spec. (Jap.) 45 high, varies conjugative
pDL10 D. ambivalens(in most isolates) 7 high (to 50?) H2S-autotrophy?
pKAW 1 P. oshimae 8.3 high crypto
pKAW2 P. oshimae 8.8 high crypto
pTAI T. acidophilium 15.2 ~ 10 crypto
234 W. Zillig et al. / FEMS Microbiology Reviews 18 (1996) 225-236
the reduction of sulfur. Its copy number is strongly
enhanced during growth by sulfur reduction though
the plasmid is not absolutely required for this mode
of existence. The widespread occurrence of this ele-
ment indicates the existence of a spreading mecha-
nism while the linkage between copy number and
mode of energy metabolism points at some func-
tional, though not stringent, involvement, in growth
by sulfur reduction [22].
Another plasmid, pTIK4 from New Zealand, en-
ables its host, to outgrow or kill competing Sul-
folobus strains without producing a diffusible toxic
agent, possibly by cell contact. Colonies of this
isolate spread fast on gelrite plates by swarming (W.
Zillig, unpublished).
The conjugative plasmid pNOB8 from Japan
(Hokkaido) is efficiently transferred from donor to
recipient cells via cell to cell contacts [11]. The
transfer mechanism is unknown. After transfer, the
copy number (per host chromosome) is high, but
colonies obtained by plating on gelrite yield liquid
cultures with low copy number. The mechanism of
copy number control is not understood but the low
copy number state can also be approached by serial
transfer of the transcipient. We have been able to
obtain Sulfolobus strains containing both the cryptic
plasmid pilE7 and the conjugative plasmid pNOB8
by electroporating a recipient already carrying the
former plasmid with the latter, followed by allowing
conjugative spread of pNOB8 in the culture contain-
ing the primary transcipients. Immediately after
spreading, the copy numbers of both plasmids were
high. In strains obtained from single colonies on
gelrite plates, they were so low that the plasmids
could only be recognized by Southern hybridization
and one or the other had been lost in some of these
strains. The replication of both plasmids appears thus
to be controlled in the same way. When a donor
containing both pNOB8 and the small cryptic plas-
mid pilE7 was mated with S. solfataricus PHI as a
fl-galactosidase negative recipient containing neither
plasmid, pNOB8 spread in the culture but transfer of
pilE7 was not observed. This indicates specific
transfer of pNOB8 (S. Albers and I. Holz, unpub-
lished work from our laboratory) rather than trans-
port through a cytoplasmatic bridge as in a mating
mode observed with Haloferax volcanii [23].
Only two of all these plasmids, pRN1 [23] and the
Pyrococcus element pGT5 [24] have been sequenced
so far. Each of these sequences contains two ORFs
one of which is large and has significant similarity
with rep genes involved in the maintenance, e.g.,
replication of such elements in the cell via a rolling
circle mechanism. This notion is supported by the
occurrence of origin motifs known to be used in RC
replication [24].
Two plasmids, one of 8.3, the other of 8.8 kb,
have been found in two different strains of the
moderately thermophilic hyperacidophile Pi-
crophilus oshimae, a relative of Thermoplasma aci-
dophilum from a Japanese (Hokkaido) solfatara [25].
They are both contained in the strain DSM 9789 kept
by the Deutsche Sammlung von Mikroorganismen,
the larger one as major, the smaller as minor compo-
nent. The two plasmids share yet unmapped cross-
hybridizing sequences. A plasmid of 15.2 kb has
recently been found in Thermoplasma acidophilum
[26].
5. 'Sulfolobicins'
About 30 of some 300 glycerol conserves of
novel Sulfolobus strains from Iceland which had
been cloned from single colonies formed sharp-edged
halos upon dotting on lawns of selected virus-free
hosts (D. Prangishvilli, unpublished results from our
laboratory). These halos resulted from complete inhi-
bition of the growth of the lawn around the dot. No
virus particle nor infectious agent responsible for this
phenomenon could be discovered. Instead, a dif-
fusible macromolecular, apparently non proteina-
ceous cytotoxic agent was concentrated from the
cell-free supernatant of the culture of one of the
strains. The toxin producing strain was resistant
against its own product. The chemical nature of the
agent and the mode of its action are unknown. But it
suggests itself that these agents could be means of
defense in the competition with other Sulfolobus
strains. All isolates from an enrichment culture giv-
ing rise to one such strain indeed shared the capacity
to excrete such toxins. Since toxin production is a
unique property - the high frequency of observing it
results probably from selection during enrichment -
genes for toxin synthesis and insensitivity towards
toxin action might be encoded on plasmids.
 W. Zillig et a l . / FEMS Microbiology ReL'iews 18 (1996) 225-236
6. Construction of transforming (cloning) vectors
based on genetic elements
The conjugative 45-kb plasmid pNOB8 suffers
extensive variation by insertions, deletions and other
recombinatory events upon transfer to recipients. The
transfer and propagation of variants containing dele-
tions indicates the deleted sequence is not required
for the functions involved. We therefore inserted a
lacS (/3-galactosidase) gene from S. solfataricus [27]
provided with the strong promoter of the gene of
ribosomal protein S 12 of S. acidocaldarius into such
a non essential sequence and electroporated this con-
struct into a lacS negative mutant of S. solfataricus,
PHI, in which an IS element of 1150 kb has been
stably integrated into the gene thus destroying its
function. Almost complete spreading was then
achieved by allowing conjugation. The success could
be monitored by the X-Gal reaction in which fl-
galactosidase-positive colonies turn blue. Many blue
colonies were obtained, however all in close contact
with white colonies of the non transformed recipient
or with eachother. The total DNA of the transformed
culture contained at least 20 copies of the wild-type
gene in the vector per inactive gene containing the IS
element in the recipient chromosome. This is to be
expected when the efficiency of transformation is
about 50% or more. The weakness of this system is
the inability of transcipients to form isolated colonies
on gelrite plates due to the burden imposed on
growth by the high copy number of the large plas-
mid. We hope to overcome this problem by reducing
the copy number with the help of a not fully under-
stood copy number control system [11].
A region in the sequence of the cryptic plasmid
pRN1 appears to be non essential because it contains
a series of transcription terminators. We therefore
introduced the lacS-gene there and obtained blue
colonies after electroporetic transfer of this vector
and X-Gal staining of colonies on gelrite plates
(M.G.L. Elferink and D. Hoogstraten, unpublished
work from this laboratory). Attempts to construct
shuttle vectors on the basis of the genome of SSVI
were so far unsuccessful, probably because the unique
restriction sites used for this purpose were situated in
essential regions. A useful construct was obtained by
Cannio, Rossi and Bartolucci (personal communica-
tion), who utilized a region of SSV1 containing the
235
putative origin of replication and thus the UV-in-
ducible gene encoding the transcript ti, d, combining
it with the E. coli vector pGEM5Zf- and introducing
an alcohol dehydrogenase gene as selection marker
against poisoning by benzyl alcohol.
Another shuttle vector was constructed by Aagard
el al. from the Pyrococcus plasmid pGT5 and the E.
coli plasmid pUC19 [28]. This recombinant vector is
maintained in Pyrococcus as well as in Sulfolobus.
Acknowledgements
We are indebted to Tairo Oshima for his invalu-
able help in organizing sampling ventures in Japan
and Jakob Kristjansson of the Ice Tec at Reykjavik,
Iceland, whose hospitality enabled us (I.H., D.P. and
W.Z.), to run a sampling and isolation laboratory in
Iceland in the summer of 1995.
References
[1] Woese, C.R., O. Kandler and M.L. Wheelis (1990) Towards
a natural system of organisms: Proposal for the domains.
Proc. Natl. Adad. Sci. USA 87, 4576-4579.
[2] Stolt, P. and W. Zillig (1994) Archaebacterial bacterio-
phages, in: Encyclopaedia of Virology, Academic Press Ltd.,
pp. 50-58.
[3] Reiter, W.P., W. Zillig and D. Palm (1988) Archaebacterial
viruses, in: Advances in Virus Research (K. Marmarosch,
F.A. Purphy and A.J. Shatking, Eds.), Vol. 34, pp. 143-188.
Academic Press Inc., Orlando, FL.
[4] Zillig, W., W.-D. Reiter, P. Palm, F. Gropp, H. Neumann
and M. Rettenberger (1988) Viruses of archaebacteria, in:
The Bacteriophages (R. Calendar, Ed.), a volume of 'The
Viruses' (H. Fr~inkel Conrat and R.R. Wagner, Eds.), Vol. 1,
pp. 517-558. Plenum Press, New York.
[5] Stolt, P. and W. Zillig (1994) Gene regulation in halophage
qOH; more than promoters. Syst. Appl. Microbiol. 16, 591-
596.
[6] Pfeifer, F. (1988) Genetics of halobacteria, in: Halophile
Bacteria (F. Rodriguez-Valera, Ed.), Vol. 2, pp. 113-115.
CRC Press.
[7] Erauso, G., S. Marsin, N. Benzouzid-Rollet, M.F. Baucher,
T. Barbeyron, Y. Zivanovic, D. Prieur and P. Forterre (1996)
Sequence of the plasmid pGT5 from the archaeon Pyrococ-
cus abvssi: evidences for rolling-circle replication in a hy-
perthermophile, J. Bacteriol., submitted.
[8] Neumann, H. and W. Zillig (1990) Structural variability in
the genome of the Thermoproteus tenax virus TTV1, Mol.
Gen. Genet. 222, 435-437.
236 W. Zillig et al. / FEMS Microbiology Reviews 18 (1996) 225-236
[9] Palm, P., C. Schleper, B. Grampp, S. Yeats, P. McWilliam,
W.-D. Reiter and W. Zillig (1991) Complete nucleotide
sequence of the virus SSV1 of the archaebacterium Sul-
folobus shibatae, Virology 185, 242-250.
[10] Zillig, W., A. Kletzin, C. Schleper, I. Holz, D. Janekovic, J.
Hain, M. Lanzend/Srfer and J.K. Kristjansson (1994) Screen-
ing for Sulfolobales, their plasmids and their viruses in
Icelandic solfataras, System. Appl. Microbiol. 16, 609-628.
[11] Schleper, C., I. Holz, D. Janekovic, J. Murphy and W. Zillig
(1995) A multicopy plasmid of the extremely thermophilic
arcaeon Sulfolobus effects its transfer to recipients by mat-
ing, J. Bacteriol. 177, 4417-4426.
[12] Langer, D., J. Hain, P. Thuriaux and W. Zillig (1995)
Transcription in Arcaea" similarity to that in Eucarya, Proc.
Natl. Acad. Sci. USA 92, 5768-5772.
[13] Gogarten, J.P., H. Kibak, P. Dittrich, P. Taiz, E.J. Bowman,
M.F. Manolson, P.J. Poole, T. Data, T. Oshima, J. Konishi,
K. Denca and M. Yoshida (1989) Evolutionary relationship
of archaebacteria, eubacteria and eukaryotes inferred from
phylogentic trees of duplicated genes, Proc. Natl. Acad. Sci.
USA 86, 9355-9359.
[14] lwabe, N., K. Kuma, M. Hasegawa, S. Osawa and T. Miyata
(1989) Evolution of RNA polymerases and branching pat-
terns of the three major groups of archaebacteria, J. Mol.
Evol. 32, 70-78.
[15] Wood, A.G., W.B. Whitman and J. Konisky (1989) Isolation
and characterization of an archaebacterial virus-like particle
from Methanococcus voltae A3, J. Bacteriol. 171, 93-98.
[16] Rettenberger, M. (1990) Das Virus TTV1 des extrem ther-
mophilen Schwefel-Archaebakteriums Thermoproteus tenax:
Zusammensetzung und Struktur, Diss. LMU Munich.
[17] Janekovic, D., S. Wunderl, I. Holz, W. Zillig, A. Gierl and
H. Neumann (1983) T r v I , "I~V2 and TYV3, a family of
viruses fo the extremely thermophilic, anaerobic, sulfur re-
ducing archaebacterium Thermoproteus tenax. Mol. Gen.
Genet. 192, 39-45.
[18] Reiter, W.-D., P. Palm and S. Yeats (1989) Transfer RNA
genes frequently serve as integration sites for prokaryotic
genetic elements, Nucleic Acids Res. 17, 1907-1914.
[19] Muskhelishvili, G., P. Palm and W. Zillig (1993) SSVI-en-
coded site-specific recombination system in Sulfolobus shi-
batae, Mol. Gen. Genet. 237, 334-342.
[20] Schleper, C., K. Kubo and W. Zillig (1992) The particle
SSV1 from the extremely thermophilic arcaeon Sulfolobus is
a virus: Demonstration of inectivity and of transfection with
viral DNA. Proc. Natl. Acad. Sci. USA 89, 7645-7649.
[21] Campbell, A. (1988) Phage evolution and speciation, in: The
Bacteriophages (R. Calendar, Ed.), a volume of 'The Viruses'
(H. Fr~tkel Conrat and R.R. Wagner, Eds.), Vol. 1, pp.
1-14. Plenum Press, New York,
[22] Zillig, W., S. Yeats, I. Holz, A. BiSck, E. Gropp, M. Retten-
berger and S. Lutz (1985) Plasmid-related anaerobic autotro-
phy of the novel archaebacterium Sulfolobus ambivalens,
Nature 313, 789-791.
[23] Mevarech, M. and R. Werczberger (1985) Genetic transfer in
Halobacterium volcanii, J. Bacteriol. 162, 461-462.
[24] Keeling, P.J., H,-P. Klenk, R.K. Singh, O. Feeley, C.
Schleper, W. Zillig, W.F. Doolittle and C.W. Sensen (1996)
Complete nucleotide sequence of the Sulfolobus islandicus
multicopy plasmid pRN1, Plasmid, in press.
[25] Schleper, C., G. Puehler, I. Holz, A. Gambacorta, D.
Janekovic, U. Santarias, H.-P. Klenk and W. Zillig (1995)
Picrophilus Gen. Nov. Fam. Nov.: A novel aerobic, het-
erotrophic, thermoacidophilic genus and family comprising
archaea capable of grwoth around pH Zero. J. Bacteriol. 177,
7050-7059.
[26] Yasuda, M., A. Yamagishi, T. Oshima (1995) The plasmids
found in isolates of the acidothermophilic archaebacterium
Thermoplasma acidophilum, FEMS Microbiol. Lett. 128,
157-161.
[27] Elferink, M.G.L., C. Schleper and W. Zillig (1996) Transfor-
mation of the extremely thermoacidophilic archaeon Sul-
folobus solfatiricus via a self-spreading vector, FEMS Mi-
crobiol. Lett. 137, 31-35.
[28] Aagaard, C., I. Leviev, R.N. Aravalli, P. Forterre, D. Prieur
and R.A. Garret (1996) General vectors for archaeal hyper-
thermophiles: Strategies based on a mobile intron and a
plasmid, FEMS Microbiol. Rev. 18, 93-104.