Journal
of General Virology (2000), 81, 227–233. Printed in Great Britain
...................................................................................................................................................................................................................................................................................
Phylogenetic analysis of some large double-stranded RNA
replicons from plants suggests they evolved from a defective
single-stranded RNA virus
Mark J. Gibbs,1 Ryuichi Koga,2 Hiromitsu Moriyama,2† Pierre Pfeiffer3 and Toshiyuki Fukuhara2
1
Bioinformatics, Research School of Biological Sciences, The Australian National University, GPO Box 475, Canberra 2601, Australia
Laboratory of Molecular Cell Biology, Faculty of Agriculture, Tokyo University of Agriculture and Technology, Fuchu, Tokyo 183, Japan
3
Institute of Plant Molecular Biology, 12 Rue du Ge! ne! ral Zimmer, 67000 Strasbourg, France
2
Sequences were recently obtained from four double-stranded (ds) RNAs from different plant
species. These dsRNAs are not associated with particles and as they appeared not to be
horizontally transmitted, they were thought to be a kind of RNA plasmid. Here we report that the
RNA-dependent RNA polymerase (RdRp) and helicase domains encoded by these dsRNAs are
related to those of viruses of the alpha-like virus supergroup. Recent work on the RdRp sequences
of alpha-like viruses raised doubts about their relatedness, but our analyses confirm that almost all
the viruses previously assigned to the supergroup are related. Alpha-like viruses have singlestranded (ss) RNA genomes and produce particles, and they are much more diverse than the
dsRNAs. This difference in diversity suggests the ssRNA alpha-like virus form is older, and we
speculate that the transformation to a dsRNA form began when an ancestral ssRNA virus lost its
virion protein gene. The phylogeny of the dsRNAs indicates this transformation was not recent and
features of the dsRNA genome structure and translation strategy suggest it is now irreversible. Our
analyses also show some dsRNAs from distantly related plants are closely related, indicating they
have not strictly co-speciated with their hosts. In view of the affinities of the dsRNAs, we believe
they should be classified as viruses and we suggest they be recognized as members of a new virus
genus (Endornavirus) and family (Endoviridae).
Introduction
Complete sequences have been obtained from large doublestranded (ds) RNA species found in cultivated rice (Oryza
sativa ssp. Japonica ; Moriyama et al., 1995), wild rice (Oryza
rufipogon ; Moriyama et al., 1999) and broad bean (Vicia faba cv.
447 ; Pfeiffer, 1998) and a partial sequence has been obtained
for one of two possible species found in kidney bean (Phaselous
vulgaris cv. black turtle soup ; Wakarchuk & Hamilton, 1990).
Similar large dsRNAs have been found in a few other plant
species, and together these dsRNAs have been called ‘ endogenous dsRNAs ’ (Moriyama et al., 1996, 1999). They have an
unusual combination of properties. They are similar to
Author for correspondence : Mark Gibbs.
Fax j61 2 62494437. e-mail mgibbs!rsbs.anu.edu.au
† Present address : Laboratory of Biochemistry and Genetics, NIDDK,
National Institutes of Health, Bethesda, MD 20892-0830, USA.
0001-6495 # 2000 SGM
cryptoviruses (Ghabrial et al., 1995) in that they are efficiently
transmitted through seed, no horizontal spread has been
observed in the field, no potential vectors have been identified
and none is associated with disease symptoms, except for one
associated with sterility (Pfeiffer et al., 1993). However, unlike
cryptoviruses, which produce particles and have a genome
consisting of two dsRNAs each about 2–3 kb long, none of the
endogenous dsRNAs is associated with particles and each is
more than 10 kb long. Consistent with a lack of particles, none
of the endogenous dsRNAs is mechanically transmissible.
Work on the dsRNA from O. sativa showed that it is not
encoded by the host and that its replication is regulated to
produce a constant low concentration in all tissues except
pollen (Fukuhara et al., 1993 ; Moriyama et al., 1996, 1999).
Some of the other dsRNAs have also been found at a constant,
low concentration in their hosts, suggesting that this is a group
characteristic (Wakarchuk & Hamilton, 1985 ; Gabriel et al.,
1987 ; Fairbanks et al., 1988 ; Valverde & Fontenot, 1990 ;
Downloaded from www.microbiologyresearch.org by
IP: 88.99.165.207
On: Mon, 31 Jul 2017 12:27:45
CCH
M. J. Gibbs and others
Zabalgogeazcoa & Gildow, 1992). This regulation, the apparent non-infectious nature of the dsRNAs, and their lack of
particles led Fukuhara et al. (1993) to suggest that these
organisms are a kind of RNA plasmid. Studies on the V. faba
dsRNA showed that it does not even spread from cell to cell
except at cell division (Duc et al., 1984), supporting this view.
Fukuhara et al. (1993) also suggested that the endogenous
dsRNAs might be related to hypoviruses as these fungal
viruses similarly lack particles and are transmitted vertically
but not horizontally except by hyphal fusion. Hypoviruses
also have large dsRNA genomes and they too were once
considered to be a kind of RNA plasmid (Brown & Finnegan,
1989 ; Shapira et al., 1991).
Here we report an analysis of the sequences of the
endogenous dsRNAs that clarifies their relationships, provides
new information on their evolution and supports a clear
argument for their classification. We also provide new evidence
that supports grouping a wide set of viruses, including the
dsRNAs, into the previously debated alpha-like virus supergroup.
Methods
The non-redundant coding and nucleotide sequence databases were
searched using the programs BLASTP, TBLASTN and PSI-BLAST
(Altschul et al., 1997) and Expect values (E values) were calculated by
these programs for pairs of aligned sequences. The evolutionary
relatedness (homology) of sequences was judged firstly using these E
values. E values are estimates of the probability of finding the same
degree of similarity by chance given the composition of the aligned
sequences and the database. E values less than 1i10−# usually indicate
homology and those less than 1i10−' almost always indicate homology
(Altschul et al., 1997 ; Bork & Gibson, 1996). The similarity of pairs of
sequences was also measured in percent identity plots made using the
program PLOTSIMILARITY (Devereux et al., 1984) and using the
program ALIGN (Dayhoff et al., 1983), which was used to make a second
test of relatedness. ALIGN performs a Monte Carlo procedure to
calculate a Z-score which is largely unbiased by sequence composition
and length. Usually a Z-score of 3 or more would be taken to be
significant, but because of the constraints on real proteins, alignments of
their sequences produce biased scores centred on 3 rather than 0 (Barton
& Sternberg, 1987). Thus, it is generally accepted that Z-scores of 5 to 6
indicate relatedness (Barton, 1996) when produced by this method. We
did 100 alignments from randomized sequences for each pairwise
comparison made with the program ALIGN and used the MDM
()
distance matrix (Dayhoff et al., 1978) and a gap penalty of 4 in the
calculations.
Multiple alignments of amino acid and nucleotide sequences were
made with the program CLUSTALW (Thompson et al., 1994). Maximum
likelihood trees were found from the amino acid multiple alignments by
quartet puzzling using the program PUZZLE version 4 (Strimmer & von
Haeseler, 1996) after positions including gaps had been excluded.
Likelihoods were calculated using the BLOSUM 62 substitution matrix
(Henikoff & Henikof, 1992) and a gamma distribution of rates of change
for variable sites with a shape parameter estimated from the data using a
neighbour-joining tree. Maximum likelihood and most parsimonious
trees were also found from the aligned nucleotide sequences by heuristic
searching with the program PAUP version 4d64 (written by David L.
Swofford) after positions including gaps and third codon positions had
been excluded. Bootstrap values were calculated from neighbour-joining
trees inferred from 1000 bootstrap samples.
Results and Discussion
Intra-group homology and phylogeny
Each of the three completely sequenced dsRNAs encodes a
single long open reading frame which we will call the long
protein (LP) gene. Comparisons of the LP amino acid sequences
from the O. sativa and V. faba dsRNAs showed that the likely
helicase and RNA-dependent RNA polymerase (RdRp)
domains correspond to two stretches of higher identity (Fig. 1,
bottom curve). These two regions were also identified in
database searches (Fig. 1). An E value of 3i10−## was
Fig. 1. Percent identity plots found using a window 100 amino acid residues long for an alignment of the long proteins (LPs) of
the two Oryza dsRNAs (top curve) and for an alignment of the LPs of the O. sativa and V. faba dsRNAs (lower curve). The
upper curve is scaled to the upper x-axis and the lower curve is scaled to the lower x-axis. The conserved GKT/S motif (marked
with open triangles) is close to the beginning of the helicase domain, and the conserved GDD motif (marked with open circles)
is close to the end of the RdRp domain. Black filled blocks mark similar regions between the V. faba and O. sativa dsRNAs
identified by a BLASTP search. E values obtained from the alignments of these regions are shown. Note that the LPs of the
Oryza dsRNAs are about 4600 amino acid residues long and the LP of the V. faba dsRNA is 5825 residues long.
CCI
Downloaded from www.microbiologyresearch.org by
IP: 88.99.165.207
On: Mon, 31 Jul 2017 12:27:45
Phylogeny of large dsRNAs from plants
Fig. 2. A maximum likelihood tree of the RNA-dependent RNA
polymerases of the alpha-like viruses and the endogenous dsRNAs. The
tree was found from aligned amino acid sequences. Maximum likelihood
and most parsimonious trees found from aligned nucleotide sequences
matched the tree shown. A maximum likelihood tree constructed from the
helicase domain sequences also matched the tree shown, except that the
branches marked with asterisks were not found. The three clusters that are
identified in the tree correspond to the three clusters identified in the text
and in Table 1. The root of the tree was found using the equivalent RdRp
amino acid sequences of the following carmo-like viruses : Barley yellow
dwarf luteovirus PAV, Carnation mottle carmovirus, Carrot mottle mimic
umbravirus, Maize chlorotic mottle machlomovirus, Tobacco necrosis
necrovirus and Tomato bushy stunt tombusvirus. The actual estimate for the
length of the branch leading to the carmo-like outgroup (dotted line) is
double that shown. The scale relates branch lengths to the number of
substitutions per site. Taxa : ACLSV, Apple chlorotic leaf spot trichovirus ;
AMV, Alfalfa mosaic alfamovirus ; ASGV, Apple stem grooving capillovirus ;
BMV, Brome mosaic bromovirus ; BNYVV, Beet necrotic yellow vein
benyvirus ; BSMV, Barley stripe mosaic hordeivirus ; BYSV, Beet yellow stunt
closterovirus ; BYV, Beet yellows closterovirus ; BVQ, Beet Q pomovirus ;
CMV, Cucumber mosaic cucumovirus ; CTV, Citrus tristeza closterovirus ;
GVA, Grapevine vitivirus A ; HaSV, Helicoverpa armigera stunt tetravirus ;
HEV, Hepatitis E virus ; LCV, Little cherry closterovirus ; LIYV, Lettuce
infectious yellows crinivirus ; OBDV, Oat blue dwarf marafivirus ; ORV, Oryza
rufipogon endornavirus ; OSV, Oryza sativa endornavirus ; PCV, Peanut
obtained for an alignment 670 residues long between the
regions that include the helicase motifs from the V. faba and O.
sativa dsRNA LPs. An E value of 2i10−(& was obtained for an
alignment 370 residues long between the regions that include
the RdRp motifs from these same two LP sequences. These two
database search results independently indicated that the Oryza
and V. faba dsRNAs share a common ancestor.
Database searches with the TBLASTN program using LP
sequences from the Oryza dsRNAs identified similarities with
the partial sequence of a dsRNA from P. vulgaris. An alignment
of the P. vulgaris polypeptide with the LP from the O. sativa
dsRNA yielded an E value of 1i10−"!. By contrast, database
searches with the LP sequence from the V. faba dsRNA did not
detect similarities with the P. vulgaris dsRNA sequence. We
confirmed this difference in affinities using the Monte Carlo
procedure. Z-scores of 12n5 and 11n2 were obtained for
comparisons between amino acid sequences encoded by the P.
vulgaris and Oryza dsRNAs, but a Z-score of only 4n2 was
obtained for the comparison of the P. vulgaris dsRNA sequence
and the equivalent sequence from the V. faba dsRNA. A
maximum-likelihood tree inferred from these sequences concurred with this result (Fig. 2, top). As our sequence
comparisons showed that the V. faba and Oryza dsRNAs have
a common ancestor, and that the P. vulgaris and Oryza dsRNAs
are similarly related, we conclude that the P. vulgaris and V.
faba dsRNAs are also related. A multiple alignment supported
this conclusion as it showed that 22 % of the amino acid
residues were strictly or strongly conserved in the equivalent
P. vulgaris, Oryza and V. faba dsRNA LP sequences. Our
conclusion is also supported by the work of Wakarchuk &
Hamilton (1985), who showed that the P. vulgaris dsRNA
shares several properties with the V. faba and Oryza dsRNAs.
Together these data suggest the genome of the P. vulgaris
dsRNA is similar to those of the fully sequenced dsRNAs and
on this basis we suggest the P. vulgaris dsRNA should be
placed in the same taxonomic family.
Experiments suggest that the dsRNAs are only vertically
transmitted through seed, either through ovules (Duc et al.,
1984) or through both ovules and pollen (Moriyama et al.,
1996). Attempts have been made to transmit them
mechanically, by grafting, using aphids and using dodder, but
all have failed (Turpen et al., 1988 ; Valverde et al., 1990 ;
Zabalgogeazcoa & Gildow, 1992). An implication of this likely
limit on spread is that the dsRNAs should co-speciate with
their hosts. Endogenous dsRNAs from closely related plants
clump pecluvirus ; PVM, Potato M carlavirus ; PVuV, Phaseolus vulgaris
endornavirus ; PVX, Potato X potexvirus ; RBDV, Raspberry bushy dwarf
idaeovirus ; RuBV, Rubella rubivirus ; SBWMV, Soil-borne wheat mosaic
furovirus ; SINV, Sindbis alphavirus ; TMV, Tobacco mosaic tobamovirus ;
TRV, Tobacco rattle tobravirus ; TSV, Tobacco streak ilarvirus ; TYMV, Turnip
yellow mosaic tymovirus ; VFV, Vicia faba endornavirus. The inset (top)
shows a mid-point rooted tree for the dsRNAs inferred from the partial
sequence from the P. vulgaris dsRNA and the equivalent sequences from
the other dsRNAs. Terminal nodes for the dsRNAs have been labelled with
the acronyms we have suggested.
Downloaded from www.microbiologyresearch.org by
IP: 88.99.165.207
On: Mon, 31 Jul 2017 12:27:45
CCJ
M. J. Gibbs and others
Table 1. Z-scores from pairwise comparisons of RdRp sequences
The first three columns show scores from comparisons involving the three major clusters (see text). The lowest
scores obtained from within-subset comparisons are shown in brackets. The remaining scores in the first three
columns are the highest scores obtained from comparisons between RdRp sequences from the clusters and ones
from ungrouped viruses and outgroups. Scores that clearly indicate relatedness are shown in bold. All comparisons
were made using a sequence starting about 170 residues on the N-terminal side of the GDD motif and ending
about 60 residues on its C-terminal side. Sequences from the five carmo-like viruses named in Fig. 2 were used,
as were sequences from the following picorna-like viruses: Cowpea mosaic comovirus, Hepatitis A hepatovirus, Human
rhinovirus 89, Rice tungro spherical waikavirus and Tobacco ringspot nepovirus. A complete list of acronyms is given
in the legend to Fig. 2.
should be more closely related to each other than to
endogenous dsRNAs from distantly related plants. Hence, the
relationships between the Oryza, P. vulgaris and V. faba
dsRNAs were unexpected : Phaseolus and Vicia belong to the
legume subfamily Papilionoideae and are only distant relatives
of the Oryza species.
RNA-dependent RNA polymerase affinities
Surprisingly BLASTP database searches revealed similarities between the RdRp-like regions in the LP sequences and
the RdRp sequences of several single-stranded (ss) RNA
viruses. E values as low as 7i10−"! were obtained for
alignments about 300 residues long with RdRp sequences from
closteroviruses. Alignments with RdRp sequences from
Hepatitis E virus (HEV) and Alfalfa mosaic alfamovirus (AMV)
yielded the next lowest E values, starting at 1i10−% and
2i10−% respectively.
It has been suggested that the closteroviruses and alfamoviruses belong to the ‘ alpha-like supergroup ’ of positive-sense
ssRNA viruses. This supergroup is said to also include viruses
from the following genera : Alphavirus, Benyvirus, Bromovirus,
Crinivirus, Cucumovirus, Capillovirus, Carlavirus, Furovirus,
Hordeivirus, Idaeovirus, Marafivirus, Pecluvirus, Pomovirus,
Potexvirus, Rubivirus, Tetravirus, Tobamovirus, Tobravirus,
Trichovirus, Tymovirus and Vitivirus. The supergroup was
CDA
defined on the basis of several features (Goldbach & de Haan,
1994 ; Koonin & Dolja, 1993), but only some of these are
shared by all the proposed members : (i) a positive-strand RNA
genome with a 5h cap, (ii) production of a subgenomic RNA
encoding a virion protein, (iii) homologous RdRp and helicase
amino acid sequences. Viruses that are not assigned to the
supergroup share the first two of these features and hence, the
supergroup is probably best defined only by the proposed
homology of the helicase and RdRp sequences.
Zanotto et al. (1996) tested the support from RdRp sequence
alignments for several proposed supergroups of RNA viruses
and found only weak support for the alpha-like supergroup.
We repeated part of this work by compiling a dataset that
included at least one RdRp sequence from each genus in the
supergroup for which sequence is available, and then by
obtaining Z-scores by the Monte Carlo procedure. The RdRplike regions of the endogenous dsRNAs were included in the
analysis to test their relationships, as was the RdRp sequence
of HEV. Sequences from carmo-like and picorna-like viruses
were also included as they were likely to be outliers. As shown
in Table 1, Z-scores of more than 6 were obtained for many of
the comparisons. These scores proved the relatedness of
almost all the RdRp sequences from viruses previously
assigned to the alpha-like supergroup and showed that the
RdRp sequences of the dsRNAs and HEV are related to alpha-
Downloaded from www.microbiologyresearch.org by
IP: 88.99.165.207
On: Mon, 31 Jul 2017 12:27:45
Phylogeny of large dsRNAs from plants
like virus sequences. Only the affinities of the RdRp sequence
of Beet necrotic yellow vein benyvirus (BNYVV), which was
previously assigned to the alpha-like supergroup, were doubtful. The highest Z-score obtained for a comparison between
the BNYVV sequence and an alpha-like virus sequence was
only 5n8. Low Z-scores were obtained from comparisons
between alpha-like virus RdRp sequences and those of carmolike and picorna-like viruses, supporting the view that the
alpha-like virus RdRp sequences are a discrete natural grouping.
Phylogenetic trees were constructed for the RdRp amino
acid sequences (Fig. 2). Strong support was found in the trees
for two major clusters that included RdRp sequences from the
following viruses : (1) alfamoviruses, bromoviruses, closteroviruses, criniviruses, cucumoviruses, furoviruses, hordeiviruses,
idaeoviruses, ilarviruses, pecluviruses, pomoviruses, tobamoviruses and tobraviruses and (2) capilloviruses, carlaviruses,
marafiviruses, potexviruses, trichoviruses, tymoviruses and
vitiviruses. There was also strong support for grouping the
sequences from the dsRNAs, which we have designated cluster
3 in the tree and Table 1. The sequences of BNYVV, Helicoverpa
armigera stunt tetravirus (HaSV), HEV, Rubella rubivirus (RuBV)
and Sindbis alphavirus (SINV) were grouped in the maximum
likelihood tree, but there was only weak support for this cluster
and it was not found in trees inferred by other methods.
Clusters 1, 2 and 3 (defined above) were supported by the
Monte Carlo procedure results (Table 1), but these results did
not support the grouping of BNYVV, HaSV, HEV, RuBV and
SINV. The Monte Carlo results suggested that the RdRps of
the dsRNAs are most closely related to those of cluster 1 and
this possibility was supported by the database searches, but the
phylogenetic analysis did not resolve the relationships between the clusters. The differences between our Monte Carlo
randomization results and those of Zanotto et al. (1996)
probably result from comparing slightly different regions of
the RdRp sequences and using different parameters for the
alignments. We used a region of sequence identified by the
program BLASTP as high-scoring. This region began about
170 residues on the N-terminal side of the GDD motif and
ended about 60 residues on its C-terminal side. Zanotto et al.
(1996) tested various regions of sequence identified in earlier
publications.
Zanotto et al. (1996) not only challenged the relatedness of
the alpha-like viruses but also that they are monophyletic. By
contrast, our phylogenetic analysis and Monte Carlo
randomization results support the monophyly of the RdRp
sequences from the alpha-like supergoup if this grouping
includes the sequences from the dsRNAs and HEV (Fig. 2 and
Table 1). One database search result challenged this conclusion :
an alignment between an alpha-like RdRp sequence and that of
the carmo-like virus Maize chlorotic mottle machlomovirus
yielded an E value that could be significant (8i10−&), but we
could not discern a clear phylogenetic signal that could link the
carmo-like and alpha-like viruses (Table 1), and the carmo-like
RdRp sequences are clearly outliers to the supergroup in the
tree (Fig. 2). It is important to note that the actual estimates for
the lengths of the branches leading to the carmo-like outlier
sequences are double those shown (Fig. 2). Plainly our
conclusions are contrary to some of those of Zanotto et al.
(1996), but we have only been concerned with the alpha-like
supergroup. The fact that none of our database searches and
analyses supported links to sequences other than those from
the mentioned viruses and dsRNAs supports the theory of
Zanotto et al. (1996) that there is no phylogenetic signal that
could be used to group all extant RdRp sequences.
Helicase sequence affinities
A profile search using the PSI-BLAST and the helicase-like
sequences from the V. faba and O. sativa dsRNAs detected
similarities with helicase sequences from viruses in every genus
in the alpha-like supergroup for which sequence is available.
Comparisons with tobamovirus sequences produced the
highest E values, 2i10−', in the first iteration. Reciprocal
searches made with the helicase sequence of Tobacco mosaic
tobamovirus identified the V. faba dsRNA helicase in the first
iteration with an E value of 5i10−). Z-scores generated from
a helicase sequence dataset compiled from alpha-like virus
sequences were lower than those from the RdRp sequence
dataset. A comparison of the V. faba dsRNA helicase with the
helicase domain of Raspberry bushy dwarf idaeovirus yielded a
Z-score of 7n1, as did a comparison with the helicase domain of
Beet soil-borne furovirus. Tobamoviruses, idaeoviruses and
furoviruses are placed in cluster 1 in the alpha-like virus RdRp
tree (Fig. 2) and hence, these results support the notion that the
dsRNAs may be most closely related to this subset within the
supergroup.
The dsRNAs probably evolved from a defective alphalike virus
Our results show that the endogenous dsRNAs share a
common ancestor with alpha-like viruses. The RdRps and
helicases encoded by the dsRNAs are probably functionally
similar to their alpha-like virus counterparts, but there are
differences between the modes of replication of these
organisms. Alpha-like virus genomic RNA is present in host
cells primarily as messenger-sense ssRNA (positive-strand
RNA), but no full-length positive-strand RNA from an
endogenous dsRNA has been unequivocally identified (Pfeiffer
et al., 1993 ; Fukuhara et al., 1995) and work on the V. faba and
P. vulgaris dsRNAs suggests these organisms do not produce
full-length ssRNAs (Pfeiffer et al., 1993 ; Wakarchuk &
Hamilton, 1985 ; Lefebvre et al., 1990). Possibly more
significant is the fact that each of the three fully sequenced
dsRNAs includes a break (discontinuity or nick) on the coding
strand but not the negative strand (Pfeiffer et al., 1993 ;
Fukuhara et al., 1995 ; Moriyama et al., 1999). The conservation
of the break and the fact that the LP gene continues in-frame
through the break in all three molecules, suggests the break is
somehow causally linked to the dsRNA form.
Downloaded from www.microbiologyresearch.org by
IP: 88.99.165.207
On: Mon, 31 Jul 2017 12:27:45
CDB
M. J. Gibbs and others
The relationship between the endogenous dsRNAs and the
ssRNA alpha-like viruses suggests that either the dsRNAs
evolved from an ancestral ssRNA virus or vice versa. The
phylogenetic trees found with the RdRp and helicase sequences
support the first of these options, as they show the alpha-like
viruses to be far more diverse than the dsRNAs (Fig. 2). If this
is true, as we believe, then as all alpha-like viruses have ssRNA
genomes and encode their own virion proteins and produce
particles, it is likely that an ancestor of the dsRNAs also had
these features. Two alternative explanations for the difference
in diversity should be considered but may be dismissed. First,
the difference could be because we do not have a fair measure
of the diversity of the dsRNAs, but this is unlikely in view of
the similarities between all the known endogenous dsRNAs
and the evolutionary distance between the Oryza and V. faba
dsRNAs. Second, the dsRNAs could be evolving far slower
than the alpha-like viruses, but this is unlikely as it would mean
that the dsRNAs were subjected to much stronger selection
pressures than the alpha-like viruses and that this difference in
selection has been maintained over long periods.
We found no sequence similarities with known virion
proteins, suggesting the dsRNAs do not encode a virion
protein, which correlates with the fact that the dsRNAs are not
associated with particles. This unusual characteristic could be
related to the evolution of the dsRNA form, as the relative
accumulation of the positive-sense and complementary-sense
(negative) RNA strands is known to be affected when positivestrand RNA plant viruses lose the function of their virion
protein genes (Nassuth & Bol, 1983 ; French & Ahlquist, 1988 ;
van der Kuyl et al., 1991). The balance of positive to negative
RNA produced by an ancestor of the dsRNAs may have
similarly altered when it lost its virion protein gene. The
transformation to a dsRNA form probably became irreversible
when the dsRNAs evolved the break in the positive strand.
Taxonomy
The endogenous dsRNAs could be classified as sub-viral
agents because they lack particles, but the distinction between
viruses and sub-viral agents is not entirely clear. The
bacteriophage P4 and dependoviruses are classified as both
sub-viral agents and viruses, and umbraviruses and hypoviruses are classified as viruses but not sub-viral agents,
although they do not produce particles. One possible explanation for this confusion is that the dependoviruses,
hypoviruses and umbraviruses, and the bacteriophage P4, are
all related to conventional viruses (Mayo et al., 1995 ; Murant
et al., 1995 ; Hillman et al., 1995). We know of no other subviral agents with this property and we view this situation as
precedence for recognizing evolutionary relationships at this
level of classification. On this basis we propose that the
endogenous dsRNAs should be recognized as viruses because
of their affinities to alpha-like viruses. As it is clear that the
dsRNAs form a distinct group, we propose that they be
CDC
nominated as members of a new virus genus and we suggest
the name Endornavirus (endo, from Greek : within, and RNA) for
this genus. The dsRNAs have no homology to recognized
dsRNA viruses and because they are very different from other
alpha-like viruses, we propose that they be assigned to a
separate family, for which we suggest the name Endoviridae. To
achieve continuity with the literature we also suggest that
endornavirus species be named after the plant species in which
they were first found. Thus, the dsRNAs for which we have
sequence could be named Phaseolus vulgaris endornavirus
(PVuV), Oryza rufipogon endornavirus (ORV), Oryza sativa
endornavirus (OSV) and Vicia faba endornavirus (VFV).
References
Altschul, S. F., Madden, T. L., Schaffer, A. A., Zhang, J., Zhang, Z.,
Miller, W. & Lipman, D. J. (1997). Gapped BLAST and PSI-BLAST : a
new generation of protein database search programs. Nucleic Acids
Research 25, 3389–3402.
Barton, G. J. (1996). Protein sequence alignment and database scanning.
In Protein Structure Prediction : A Practical Approach, pp. 31–63. Edited by
M. J. E. Sternberg. Oxford : IRL Press at Oxford University Press.
Barton, G. J. & Sternberg, M. J. E. (1987). A strategy for the rapid
multiple alignment of protein sequences – confidence levels from tertiary
structure comparisons. Journal of Molecular Biology 198, 327–337.
Bork, P. & Gibson, T. J. (1996). Applying motif and profile searches.
Methods in Enzymology 266, 162–183.
Brown, G. G. & Finnegan, P. M. (1989). RNA plasmids. International
Review of Cytology 117, 1–56.
Dayhoff, M. O., Barker, W. C. & Hunt, L. T. (1978). A model of
evolutionary change in proteins. In Atlas of Protein Sequence and Structure,
vol. 5, suppl. 3. Edited by M. O. Dayhoff. Washington DC : National
Biomedical Research Foundation.
Dayhoff, M. O., Barker, W. C. & Hunt, L. T. (1983). Establishing
homologies in protein sequences. Methods in Enzymology 91, 524–545.
Devereux, J., Haeberli, P. & Smithies, O. (1984). A comprehensive set
of sequence analysis programs for the VAX. Nucleic Acids Research 12,
387–396.
Duc, G., Scalla, R. & Lefebvre, A. (1984). New developments in
cytoplasmic male sterility in Vicia faba. In Vicia faba : Agronomy, Physiology
and Breeding, pp. 254–260. Edited by P. D. Hebblethwaite, T. C. K.
Dawkins, M. C. Heath & G. Lockwood. The Hague : Martinus Nijhoff\
Dr W. Junk Publishers.
Fairbanks, D. J., Smith, S. E. & Brown, J. K. (1988). Inheritance of large
mitochondrial RNAs in alfalfa. Theoretical and Applied Genetics 76,
619–622.
French, R. & Ahlquist, P. (1988). Characterisation and engineering of
sequences controlling in vivo synthesis of brome mosaic virus subgenomic RNA. Journal of Virology 62, 2411–2420.
Fukuhara, T., Moriyama, H., Pak, J. K., Hyakutake, T. & Nitta, T.
(1993). Enigmatic double-stranded RNA in Japonica rice. Plant Molecular
Biology 21, 1121–1130.
Fukuhara, T., Moriyama, H. & Nitta, T. (1995). The unusual structure of
a novel RNA replicon in rice. Journal of Biological Chemistry 270,
18147–18149.
Gabriel, C. J., Walsh, R. & Nolt, B. L. (1987). Evidence for a latent viruslike agent in cassava. Phytopathology 77, 92–95.
Downloaded from www.microbiologyresearch.org by
IP: 88.99.165.207
On: Mon, 31 Jul 2017 12:27:45
Phylogeny of large dsRNAs from plants
Ghabrial, S. A., Bozarth, R. F., Buck, K. W., Yamashita, S., Martelli,
G. P. & Milne, R. G. (1995). Family Partitiviridae. In Virus Taxonomy :
Classification and Nomenclature of Viruses, pp. 253–260. Edited by F. A.
Murphy, C. M. Fauquet, D. H. L. Bishop, S. A. Ghabrial, A. W. Jarvis,
G. P. Martelli, M. A. Mayo & M. D. Summers. Wien : Springer-Verlag.
Goldbach, R. & de Haan, P. (1994). RNA viral supergroups and the
evolution of RNA viruses. In The Evolutionary Biology of Viruses, pp.
105–119. Edited by S. Morse. New York : Raven Press.
Henikoff, S. & Henikof, J. G. (1992). Amino acid substitution matrices
from protein blocks. Proceedings of the National Academy of Sciences, USA
89, 10915–10919.
Hillman, B. I., Fulbright, D. W., Nuss, D. L. & van Alfen, N. K. (1995).
Family : Hypoviridae. In Virus Taxonomy : Classification and Nomenclature of
Viruses, pp. 261–264. Edited by F. A. Murphy, C. M. Fauquet, D. H. L.
Bishop, S. A. Ghabrial, A. W. Jarvis, G. P. Martelli, M. A. Mayo & M. D.
Summers. Wien : Springer-Verlag.
Koonin, E. V. & Dolja, V. V. (1993). Evolution and taxonomy of
positive-strand RNA viruses : implications of comparative analysis of
amino acid sequences. Critical Reviews in Biochemistry and Molecular
Biology 28, 375–430.
Lefebvre, A., Scalla, R. & Pfeiffer, P. (1990). The double-stranded RNA
associated with the ‘ 447 ’ cytoplasmic male sterility in Vicia faba is
packaged together with its replicase in cytoplasmic membranous vesicles.
Plant Molecular Biology 14, 477–490.
Mayo, M. A., Berns, K. I., Fritsch, C., Kaper, J. M., Jackson, A. O.,
Leibowitz, M. J. & Taylor, J. M. (1995). Subviral agents : satellites. In
Virus Taxonomy : Classification and Nomenclature of Viruses, pp. 487–492.
Edited by F. A. Murphy, C. M. Fauquet, D. H. L. Bishop, S. A. Ghabrial,
A. W. Jarvis, G. P. Martelli, M. A. Mayo & M. D. Summers. Wien :
Springer-Verlag.
Moriyama, H., Nitta, T. & Fukuhara, T. (1995). Double-stranded RNA
in rice : a novel RNA replicon in plants. Molecular and General Genetics
248, 364–369.
Moriyama, H., Kanaya, K., Wang, J. Z., Nitta, T. & Fukuhara, T. (1996).
Stringently and developmentally regulated levels of a cytoplasmic
double-stranded RNA and its high-efficiency transmission via egg and
pollen in rice. Plant Molecular Biology 31, 713–719.
Moriyama, H., Horiuchi, H., Koga, R. & Fukuhara, T. (1999). Molecular
characterization of two endogenous double-stranded RNAs in rice and
their inheritance by interspecific hybrids. Journal of Biological Chemistry
274, 6882–6888.
Murant, A. F., Robinson, D. J. & Gibbs, M. J. (1995). Genus :
Umbravirus. In Virus Taxonomy : Classification and Nomenclature of Viruses,
pp. 388–391. Edited by F. A. Murphy, C. M. Fauquet, D. H. L. Bishop,
S. A. Ghabrial, A. W. Jarvis, G. P. Martelli, M. A. Mayo & M. D.
Summers. Wien : Springer-Verlag.
Nassuth, A. & Bol, J. F. (1983). Altered balance of the synthesis of plus-
and minus-strand RNAs induced by RNAs 1 and 2 of alfalfa mosaic virus
in the absence of RNA3. Virology 124, 75–85.
Pfeiffer, P. (1998). Nucleotide sequence, genetic organization and
expression strategy of the double-stranded RNA associated with the
‘ 447 ’ cytoplasmic male sterility trait in Vicia faba. Journal of General
Virology 79, 2349–2358.
Pfeiffer, P., Jung, J.-L., Heitzler, J. & Keith, G. (1993). Unusual structure
of the double-stranded RNA associated with the ‘ 447 ’ cytoplasmic male
sterility in Vicia faba. Journal of General Virology 74, 1167–1173.
Shapira, R., Choi, G. H. & Nuss, D. L. (1991). Virus-like genetic
organization and expression strategy for a double-stranded RNA genetic
element associated with biological control of chestnut blight. EMBO
Journal 10, 731–739.
Strimmer, K. & von Haeseler, A. (1996). Quartet puzzling – a quartet
maximum-likelihood method for reconstructing tree topologies.
Molecular Biology and Evolution 13, 964–969.
Thompson, J. D., Higgins, D. G. & Gibson, T. J. (1994). CLUSTAL W :
improving the sensitivity of progressive multiple sequence alignment
through sequence weighting, position specific gap penalties and weight
matrix choice. Nucleic Acids Research 22, 4673–4680.
Turpen, T., Grill, L. K. & Garger, S. J. (1988). On the mechanism of
cytoplasmic male sterility in the 447 line of Vicia faba. Plant Molecular
Biology 10, 489–497.
Valverde, R. A. & Fontenot, J. F. (1990). Variation in double-stranded
ribonucleic acid among pepper cultivars. Journal of the American Society for
Horticultural Science 116, 903–905.
Valverde, R. A., Nameth, S., Abdallha, O., Al-Musa, O., Desjardins, P.
& Dodds, J. A. (1990). Indigenous double-stranded RNA from pepper
(Capsicum annum). Plant Science 67, 195–201.
van der Kuyl, A. C., Neelemen, L. & Bol, J. F. (1991). Role of alfalfa
mosaic virus coat protein in regulation of the balance between viral plus
and minus strand RNA synthesis. Virology 185, 496–499.
Wakarchuk, D. A. & Hamilton, R. I. (1985). Cellular double-stranded
RNA in Phaseolus vulgaris. Plant Molecular Biology 5, 55–63.
Wakarchuk, D. A. & Hamilton, R. I. (1990). Partial nucleotide sequence
from enigmatic dsRNAs in Phaseolus vulgaris. Plant Molecular Biology 14,
637–639.
Zabalgogeazcoa, I. A. & Gildow, F. E. (1992). Double-stranded
ribonucleic acid in ‘ Barsoy ’ barley. Plant Science 83, 187–194.
Zanotto, P. M. de A., Gibbs, M. J., Gould, E. A. & Holmes, E. C. (1996).
A re-evaluation of the higher taxonomy of viruses based on RNA
polymerases. Journal of Virology 70, 6083–6096.
Received 7 June 1999 ; Accepted 10 September 1999
Downloaded from www.microbiologyresearch.org by
IP: 88.99.165.207
On: Mon, 31 Jul 2017 12:27:45
CDD