Dynamic Genome Organization
Geminivirus (Geminiviridae)
and Gene Evolution
James M. Bradeen,’ Marja C. P. Timmermatq2
Waksman
Institute,
Rutgers, The State University
by Positive Selection in
and Joachim Messing
of New Jersey
Geminiviruses
(Geminiviridae)
are a diverse group of plant viruses differing from other known plant viruses in
possessing circular, single-stranded
DNA. Current classification divides the family into three subgroups, defined in
part by genome organization, insect vector, and plant host range. Previous phylogenetic assessments of geminiviruses
have used DNA and/or amino acid sequences from the replication-associated
and coat protein genes and have relied
predominantly
on distance analyses. We used amino acid and DNA sequence data from the replication-associated
and coat protein genes from 22 geminivirus types in distance and parsimony analyses. Although the results of our
analyses largely agree with those reported previously, we could not always predict viral relationships based on
genome organization, plant host, or insect vector. Loss of correlation of these traits with phylogeny is likely due to
improved sampling of geminivirus types. Unrooted parsimony trees suggest multiple independent origins for the
monopartite genome. Genome organization is therefore a dynamic character. Estimates of nonsynonymous
and
synonymous nucleotide substitutions for extant and inferred ancestral sequences were used to evaluate hypotheses
that the replication-associated
and coat protein sequences evolve to accommodate
plant host and insect vector
specificities, respectively.
Results suggest that plant host specificity does not solely direct replication-associated
protein evolution but that coat protein sequence does evolve in response to insect vector specificity. Genome
organization and, possibly, plant host specificity are not reliable taxonomic characters.
Introduction
Worldwide,
virus infection of agronomically
important plant species accounts for millions of dollars in
lost productivity
each year. Among plant viruses, geminiviruses constitute an important group, infecting both
monocot and dicot plant species. Disease symptoms are
diverse but frequently
include foliage discolorations
such as streaks (e.g. Digitaria streak, Miscanthus streak,
and maize streak viruses) and mosaic patterns (e.g. Abutilon mosaic, bean calico mosaic, Chloris striate mosaic,
and tomato golden mosaic viruses), stunted plant development, and twisted, curled, or cupped foliage. While
90% of known plant viruses have RNA genomes, geminiviruses
contain circular, single-stranded
DNA genomes and replicate via double-stranded
DNA intermediates.
All geminiviruses
are characterized
by a
twinned (geminate) capsid morphology
and have monopartite (single) or bipartite (double) genomes (Timmermans, Das, and Messing 1994). Proposed correlations between genome organization,
insect vector specificity, and plant host infectivity have raised interesting
questions about geminivirus
evolution.
Current nomenclature
describes the viral family
Geminiviridae
with three subgroups (Briddon and Markham 1995). Geminiviral
subgroups were originally described solely by genome organization,
plant host, and
insect vector specificities.
Recently, the classifications
were broadened to incorporate isolates with unique combinations of characters. Subgroup I contains viruses with
monopartite
genomes
that are leafhopper-transmitted
’Present address: USDA-ARS
University of Wisconsin, Madison.
* Present address:
Department
Key words: coat protein,
and Department
of Biology,
geminivirus,
of Horticulture,
Yale University.
nucleotide
substitution.
Address
for correspondence
and reprints: Joachim Messing,
Waksman Institute, PO. Box 759, Rutgers University, Piscataway, New
Jersey 08855. E-mail: [email protected].
Mol.
Biol. Evol. 14( 11):1114-l 124. 1997
0 1997 by the Society for Molecular Biology and Evolution.
1114
ISSN: 0737-4038
and predominantly
infect monocots. Subgroup II contains viruses with monopartite
genomes that are leafhopper-transmitted
but infect dicots. Subgroup III contains bipartite and monopartite viruses. These viruses are
whitefly-transmitted
and infect dicots. Geminiviral
subgroups are thought to be monophyletic.
Howarth and
Goodman (1986) suggest that progenitor geminiviruses
were monopartite (subgroup I) viruses and that bipartite
(subgroup III) viruses later arose by acquisition of dicot
infectivity and whitefly transmissibility.
Evidence supporting this hypothesis includes similarity in overall genome organization
between the two subgroups; amino
acid similarities between coding regions of the two genomes of bipartite viruses, suggesting a common origin
for the two genomes; and distance analysis of the replication-associated
protein (Kikuno et al. 1984; Howarth
and Goodman 1986). Subgroup II viruses may have arisen either from subgroup I viruses by the acquisition of
dicot infectivity or from subgroup III viruses by loss of
the second genome.
Apparent dichotomies
dividing viral isolates into
monocot- and dicot-infecting
groups have been noted in
distance analyses of the replication-associated
amino
acid sequence. Similarly, dichotomies dividing viral isolates into leafhopperand whitefly-transmitted
groups
have been noted in distance analyses of the coat protein
amino acid sequence. These observations have led to the
hypothesis that these geminiviral
genes may evolve to
accommodate
plant host and insect vector specificities,
respectively
(Howarth and Vandemark 1989).
In the current study, we examine the evolutionary
relationships
of 39 geminivirus
isolates representing
22
viral types. Inclusion of viral isolates from each of the
three subgroups, including newly described viruses with
unique combinations
of characteristics,
allows evaluation of current geminiviral
classification,
determination
of the phylogenetic
importance of characteristics
used
in part to define geminiviral taxa, and estimation of mo-
Genome
lecular phylogenies for currently described geminiviruses. Additionally,
comparison of ancestral DNA sequences inferred from parsimony analyses evaluates proposals
of evolutionary
response in particular geminivirus genes
to accommodate
plant host or insect vector specificities.
Materials
and Methods
Replication-associated
and coat protein DNA sequences were obtained from GenBank (table 1). Replication-associated
and coat protein sequences were identified using published accounts, GenBank information,
or comparison with known coding regions. Replicationassociated protein sequence introns were identified and
removed prior to analysis (Schalk et al. 1989; Mullineaux, Guerineau,
and Accotto 1990). DNA sequence
translation, DNA and protein sequence alignments, and
percent dissimilarity
calculations
were done using GCG
version 7.2 (Genetics Computer Group, Inc., Madison,
Wis.). Sequence dissimilarity
calculations
and aligned
sequences were analyzed using PHYLIP version 3.5~
(Felsenstein
1989) and PAUP version 3.1.1 (Swofford
1993), respectively. PHYLIP analyses include neighborjoining and 10 replications each of Fitch and Kitsch with
Jumble option using percent dissimilarity
among amino
acid sequences. Sums of squares from the best Fitch and
Kitsch trees were used to test the hypothesis of a molecular clock (Felsenstein
1984). PAUP analyses included a heuristic search with 100 replications
of random
taxon addition (to identify topographical
“islands”
inconsistent
with initially
identified
trees [Maddison
199 11) and heuristic bootstrap (Felsenstein
1985) and
heuristic decay analyses (Bremer 1988; Donoghue et al.
1992) using both amino acid and DNA sequences. When
multiple most-parsimonious
trees were identified, consensus trees were computed. Trees consistent with specific evolutionary
hypotheses were generated by heuristic search with 100 replications
of random taxon addition using specified topological
constraints.
Shortest
identified constraint trees were compared with most-parsimonious trees by nonparametric
(Templeton 1983) and
parametric
(Kishino and Hasegawa
1989) procedures.
Nonsynonymous
(K,) and synonymous
(KS) nucleotide
substitutions (Li 1993) were estimated for pairs of extant
and inferred ancestral replication-associated
and coat
protein DNA sequences. &/KS ratios and plots of K,
versus KS were used to evaluate for data trends indicative of gene evolution via positive selection.
Results and Discussion
The bipartite geminiviruses
contain six conserved
open reading frames (ORFs), and the monopartite geminiviruses contain four. Among the homologous
genes,
the replication-associated
and the coat protein genes are
the most conserved (Howarth and Goodman 1986) and
have been used extensively
for evolutionary
analysis
(Howarth and Goodman 1986; Howarth and Vandemark
1989; Timmermans,
Das, and Messings 1994; Padidam,
Beachy, and Fauquet 1995a). We have estimated phylogenies using both of these genes in an expanded set
and Gene Evolution
in Geminivirus
of geminivirus
isolates, using both distance and parsimony analyses.
Distance analysis assumes random sampling of all
geminivirus types. Because geminiviruses
differ in their
agronomic importance and because research efforts are
disproportionately
expended toward agronomically
important types (Matthews 1985), random sampling may
be an inappropriate assumption. It is noteworthy that the
viruses included in this study, which includes most geminiviral types molecularly characterized to date, are predominantly infectious agents of economically
important
crop families (Fabaceae, Malvaceae, Poaceae, Solanaceae). These plant hosts are not closely related species
but diverse families including both monocots and dicots.
(BCTV [abbreviations
for isolates are listed in table 11,
which is known to infect 44 different dicot families, was
originally
described as a disease agent of the agronomically important sugar beet [Beta vulgar-is; Mumford
19741.) The insect vectors, like the hosts, represent divergent groups. Geminiviruses
are therefore diverse in
genome arrangement,
plant host specificity, and insect
vector specificity. That this diverse group of viruses infects only agronomically
important plant families is unlikely.
The relationship of geminiviruses
to other plant viruses is unclear (Matthews
1985). Most plant virus
groups are RNA-encoded.
Among the DNA-encoded
plant viruses, geminiviruses
are unique in that they contain single-stranded
DNA. As noted previously,
the
characteristic geminate structure of geminiviruses
distinguishes them morphologically
from other plant viruses.
Frischmuth,
Zimmat, and Jeske (1990) described two
ORFs on DNA A of AbMV with apparent prokaryotic
features, features lacking in other geminiviruses.
The authors made no claims about the phylogenetic importance
of this finding. Koonin and Ilyina (1992) noted conserved sequence motifs between the replication initiator
proteins of certain eubacterial plasmids and the geminivirus replication-associated
protein and suggested that
this similarity may be of phylogenetic
origin. However,
at the present time, no concrete evolutionary
lineage can
be drawn between these otherwise dissimilar taxa. Because of the morphological
and genetic uniqueness
of
geminiviruses,
attempts to identify an appropriate outgroup for polarizing
characters proved unsuccessful.
Like certain distance analyses,
midpoint
rooting assumes a molecular clock, an assumption
inconsistent
with our observations
(unpublished
data) and those of
Howarth and Vandemark (1989). We have therefore chosen to present unrooted phylogenies.
Conclusions based
on unrooted parsimony analyses are independent of true
tree polarity.
Distance and parsimony analyses are in agreement
with those presented previously
(Howarth and Vandemark 1989; Timmermans,
Das, and Messing 1994; Padidam, Beachy, and Fauquet 1995a) and with each other,
with the exception of certain TYLCV isolates, as discussed below. Parsimony analyses of the amino acid and
DNA sequences are likewise in agreement. Only parsimony analyses of the DNA sequences are presented
here. In all analyses, two broadly defined, well-support-
Table 1
Geminivirus
Isolate Characteristics
Virus
Abutilon mosaic ............
African cassava mosaic. ......
Bean calico mosaic ..........
Bean dwarf mosaic
Bean golden mosaic
Beet curly top
..........
.........
..............
Chloris striate mosaic ........
Digitaria streak .............
Maize streak. ...............
Miscanthus streak ...........
Mungbean yellow mosaic. ....
Panicum streak .............
Potato yellow mosaic ........
Squash leaf curl. ............
Sugarcane streak ............
Tobacco yellow dwarf. .......
Tomato golden mosaic .......
Tomato leaf crumple. ........
Tomato leaf curl
Tomato mottle
Tomato yellow
Wheat dwarf
............
..............
leaf curl ......
... ......... ...
Abbreviation”
GenBank
Accession No.
Genomeb
AbMV
ACMV
BCaMV- 1 rep
BCaMV-2 rep
BCaMV-3 rep
BDMV
BGMV- 1
BGMV-2
BGMV-3
BGMV-4
BCTV- 1
BCTV-2
CSMV
DSV
MSV- 1
MSV-2
MiSV
MYMV
PanSV
PYMV
SLCV
ssv
TYDV
TGMV
TLCrV-I cp
TLCrV-2 rep
TLCV- 1
TLCV-2
TLCV-3
ToMoV
TYLCV- 1
TYLCV-2
TYLCV-3
TYLCV-4
TYLCV-5
TYLCV-6
TYLCV-7
WDV- 1
WDV-2
Xl5983
x 17095
uOO121
L27264
L22758
M88179
Ml0070
LO1635
M88686
M9 1604
U023 11
M24597, X04 144
M2002 1
M23022
YOO5 14
X0 1633, K02026
DO1030
D14703
X60168
DO0940
M38183
S64567
M81103
K02029
L34747
L27267
u15015, u15017
U15016
S5325 1
L14460
X61 153
Xl5656
X763 19
22575 1
228390
L27708
M59838
X02869
X82104
B
B
B
B
B
B
B
B
B
B
M
M
M
M
M
M
M
B
M
B
B
M
M
B
B
B
B
B
M
B
M
M
M
M
M
M
B
M
M
d BCaMV and TLCrV include one or more isolates for which only the replication
b Genome organization: M, monopartite; B, bipartite.
‘ Insect vector: L, leafhopper; W, whitefly.
associated
Host
Malvaceae
Euphorbaceae,
Fabaceae
Fabaceae
Fabaceae
Fabaceae
Fabaceae, Malvaceae
Fabaceae, Malvaceae
Fabaceae
Fabaceae
Fabaceae
44 dicot families
44 dicot families
Poaceae
Poaceae
Poaceae
Poaceae
Poaceae
Fabaceae
Poaceae
Solanaceae
Cucurbitaceae
Poaceae
Fabaceae, Solanacceae
Solanaceae
Fabaceae, Solanaceae
Fabaceae, Solanaceae
Solanaceae
Solanaceae
Solanaceae
Solanaceae
Solanaceae
Solanaceae
Solanaceae
Solanaceae
Solanaceae
Solanaceae
Solanaceae
Poaceae
Poaceae
Vectof
Bemisia tabaci (W)
Bemisia tabaci (W)
Bemisia tabaci (W)
Bemisia tabaci (W)
Bemisia tabaci (W)
Bemisia tabaci (W)
Bemisia tabaci (W)
Bemisia tabaci (W)
Bemisia tabaci (W)
Bemisia tabaci (W)
Circulifer tenellus (L)
Circulifer tenellus (L)
Nesoclutha pallida (L)
Nesoclutha declivata (L)
Cicadulina mbila (L)
Cicadulina mbila (L)
Unknown
Unknown
Cicadulina mbila (L)
Bemisia tabaci (W)
Bemisia tabaci (W)
Cicadulina mbila (L)
Orosius argentatus (L)
Bemisia tabaci (W)
Bemisia tabaci (W)
Bemisia tabaci (W)
Bemisia tabaci (W)
Bemisia tabaci (W)
Bemisia tabaci (W)
Bemisia tabaci (W)
Bemisia tabaci (W)
Bemisia tabaci (W)
Bemisia tabaci (W)
Bemisia tabaci (W)
Bemisia tabaci (W)
Bemisia tabaci (W)
Bemisia tabaci (W)
Psammotettix alienus (L)
Psammotettix alienus (L)
protein (“rep”) or the coat protein (“cp”) gene sequence is available.
Reference
Frischmuth, Zimmat, and Jeske (1990)
Morris et al. (1990)
Loniello et al. (unpublished data)
Loniello et al. (unpublished data)
Loniello et al. (unpublished data)
Gilbertson et al. (1991)
Howarth et al. (1985)
Gilbertson et al. (1991)
Gilbertson et al. (199 1)
Gilbertson et al. (1991)
Stenger et al. (1994)
Stanley et al. (1986)
Andersen et al. (1988)
Donson et al. (1987)
Lazarowitz (1988)
Mullineaux et al. (1984)
Chatani et al. (1991)
Morinaga, Ikegami, and Miura (1993)
Briddon et al. (1992)
Coutts et al. (199 1)
Lazarowitz and Lazdins (199 1)
Hughes, Rybicki, and Kirby (1993)
Morris et al. (1992)
Hamilton et al. (1984)
Paplomatas et al. (1994)
Loniello et al. (unpublished data)
Padidam, Beachy, and Fauquet (19956)
Padidam, Beachy, and Fauquet (1995b)
Dry et al. (1993)
Abouzid, Polston, and Hiebert (1992)
Kheyr-Pour et al. (1992)
Navot et al. (1991)
Antignus and Cohen (1994)
Noris et al. (1994)
Crespi et al. (unpublished data)
Reina et al. (unpublished data)
Rochester et al. (1994)
MacDowell et al. (1985)
Bendahmane et al. (unpublished data)
Genome
ed groups of isolates are evident. The first includes only
monopartite,
monocot-infecting,
leafhopper-transmitted
viruses (MSV, DSV, SSV, and PanSV) and will henceforth be referred to as the “African streak virus clade,”
following the precedent of Hughes, Rybicki, and Kirby
(1993). The second group includes both mono- and bipartite, dicot-infecting,
predominantly
whitefly-transmitted viruses and will henceforth
be referred to as the
“BGMV clade.” The constitution
of the BGMV clade
differs slightly for replication-associated
and coat protein analyses (compare figs. 1 and 2). We cannot rule
out the possibility that this discrepancy
may be due to
viral recombination,
the replication-associated
and coat
protein sequences
originating
from separate lineages.
Observed tree topologies are consistent with the possibility that TYLCV-1 and TYLCV-6 are viral recombinants. Viral recombination
has been suggested for an
Indian isolate of TLCV (Padidam, Beachy, and Fauquet
199%) and for BCTV (Padidam, Beachy, and Fauquet
199%). Gilbertson et al. (1993) demonstrate that “pseudorecombinants”
(bipartite geminiviruses
composed of
the A genome of one bipartite geminivirus
and the B
genome of a distantly related bipartite geminivirus)
can
function to cause disease symptoms in host plants. The
authors speculate that pseudorecombination
followed by
genetic recombination
between the A and B genomes
might be an important means of geminivirus
evolution.
Agreement between the replication-associated
and coat
protein analyses suggests no evidence for viral recombinants, except for TYLCV-1 and -6.
Dynamic
Genome
Organization
Parsimony analyses of both the replication-associated and the coat protein sequences (figs. 1 and 2) reveal
complex
relationships
between
mono- and bipartite
geminiviruses.
Regardless of the true tree polarity, monopartite viruses do not constitute a monophyletic
clade.
For example, TYLCV isolates, with the exception of
TYLCV-7, are monopartite viruses. However, these isolates are most closely related to bipartite (subgroup III)
geminiviruses
and are only distantly related to other
monopartite viruses (subgroups I, II; figs. 1 and 2). Furthermore, in neither the replication-associated
nor the
coat protein parsimony trees do the monopartite TYLCV
isolates constitute
a monophyletic
clade, suggesting
multiple origins for the monopartite
viruses. Most-parsimonious trees for the replication-associated
and coat
protein DNA sequences were compared with the shortest
trees consistent
with the assumption
that monopartite
and bipartite viruses constitute
separate monophyletic
clades. Constraint trees differed significantly (a = 0.05)
from most-parsimonious
trees. (Tree lengths were 6,458
vs. 6,5 12 steps for the replication-associated
protein
DNA most-parsimonious
and constraint trees, respectively, and 4,392 vs. 4,540 steps for the coat protein
DNA most-parsimonious
and constraint
trees, respectively.)
Parsimony
dictates that some of the monopartite
isolates in this study arose from bipartite isolates by
reversion. These derivatives may include the monopartite TYLCV and TLCV isolates. Consistent
with this
and Gene Evolution
in Geminivirus
1117
possibility,
genome
organization
of the monopartite
TYLCV and TLCV isolates more closely resembles that
of bipartite than that of other monopartite geminiviruses
(Navot et al. 1991; Kheyr-Pour et al. 1992; Dry et al.
1993). Genome organization
is dynamic and is not a
reliable taxonomic character.
Unrooted parsimony analyses support the possibility that the monopartite TYDV and BCTV isolates arose
from a dicot-infecting
bipartite virus by loss of the second genome (reversion), or directly from the progenitor
monopartite
group by acquisition of dicot-infecting
capabilities. If the African streak virus clade is the true
progenitor type, the replication-associated
and coat protein parsimony trees (figs. 1 and 2) predict TYDV to be
monopartite-derived.
Morris et al. (1992) concluded that
TYDV genome organization more closely resembles genome organization
of subgroup I monopartite than that
of bipartite viruses, consistent with this possibility. Recent modifications
to geminivirus
classification
include
assignment
of TYDV to subgroup I, despite its dicot
host specificity (Briddon and Markham 1995). If the African streak virus clade is the progenitor type, the replication-associated
protein parsimony trees (fig. 1) predict and the coat protein parsimony trees (fig. 2) allow
BCTV to be bipartite-derived.
Genome characteristics of
BCTV resemble those of both monopartite and bipartite
viruses (Stanley et al. 1986; Timmermans,
Das, and
Messing 1994). Our results support deemphasis of genome organization
as a taxonomic character for geminiviruses.
Kikuno et al. (1984) suggested, based on ORF amino acid similarity between the two genomes of cassava
latent virus, that bipartite geminiviruses
(subgroup III)
arose from a single genome progenitor. Howarth and
Goodman
(1986) noted similar ORF similarities
for
BGMV and TGMV isolates and concluded,
based on
mutation-distance
analysis of the replication-associated
protein, that subgroup I monopartite viruses arose prior
to bipartite (subgroup III) viruses and that the monopartite condition may be a primitive feature. Lack of an
appropriate outgroup for rooting geminivirus parsimony
trees precludes confirmation
or refutation of this hypothesis. The circumstantial
evidence is nevertheless
compelling.
Furthermore,
we find it conceptually
appealing that evolution should proceed from the less complicated monopartite
genome to the more complex bipartite genome. We speculate, in agreement with Howarth and Goodman (1986), that evolution likely proceeded
from some subgroup
I monopartite
virus
(perhaps a member of the African streak virus clade) to
a relatively more complex bipartite virus (subgroup III).
Bipartite viruses then evolved from this progenitor, occasionally giving rise to monopartite viruses. These derived monopartites
retain the host and vector specificities of their bipartite progenitors.
Gene Evolution
by Positive
Selection
In addition to genome organization,
plant host and
insect vector specificities
have been used to classify
geminiviruses.
Howarth and Vandemark
( 1989) concluded, based on distance dendrogram topology that rep-
Predominantly
Dicot-infecting
Monocot-infecting
/
ss
TyLCV-1
III M W D (Solanaceae)
BCaMV-1
III B W D (Fabaceae)
- -
’
I
I
I
I
-
I M L M (Poaceae)
I M L M (Poaceae)
I M L M (Poaceae)
DSV -
I M ? M (Poaceae)
ssv
I M L D (FabaceaeEolanaceae)
PanSV
I M L M (Poaceae)
TYDV
I M L M (Poaceae)
Misv
__
I
I
African
StreakVirus
Clade
;
,
I
I
I
__
I
M L M (Poaceae)
r
56
24)
WDV- 1
I M L M (Poaceae)
WDV-2
I M L M (Poaceae)
CSMV
1
>4
- (3) 55
1
(4)*
(5) *
L
(91 T
100
(10)
I
y
I
(13)
I
SLCV
III B W D (Cucurbitaceae)
BGMV-2
III B W D (Fabaceae)
BGMV4
III B W D (Fabaceae)
BGMV- 1
III B W D (Fabaceae)
AbMV
III B W D (Malvaceae)
ToMoV
III B W D (Solanaceae)
I
TLCrV-2
III B W D (Fabaceae)
I
BDMV
III B W D (Fabaceae)
PYMV
III B W D (Solanaceae)
BGMV-3
III B W D (Fabaceae)
; BGMV
I Clade
TGMV
III B W D (Solanaceae)
I
TLCV- 1
III B W D (Solanaceae)
TLCV-2
III B W D (Solanaceae)
TLCV-3
III B W D (Solanaceae)
TYLCV-7
III B W D (Solanaceae)
MYMV
III B ? D (Fabaceae)
TYLCV-6
III M W D (Solanaceae)
I
TYLCV-4
III M W D (Solanaceae)
I
TYLCV-5
III M W D (Solanaceae)
TYLCV-2
III M W D (Solanaceae)
TYLCV-3
III M W D (Solanaceae)
ACMV
III B W D (Fabaceae)
I
I
I
I
I
I
I
I
I
I
I
I
I
I
I
I
I
I
<5(
a4
L
95
A
(11)
93
A
(14:
I
I
I
I
I
I
I
I
I
I
I
BCTV-
1
BCTV-2
II M L D (44 Families)
II M L D (44 Families)
I
I
I
I
Genome
lication-associated
protein evolution may be constrained
by plant host (monocot vs. dicot) specificity. Analysis
of the replication-associated
protein DNA sequences
(fig. 1) reveals a tendency for viral isolates to associate
based on plant host specificity (mono- vs. dicot), with
the exception of the subgroup I virus TYDV. While the
replication-associated
amino acid sequence parsimony
analysis allows (but does not require) association
of
TYDV with other dicot-infecting
viruses (data not
shown), the replication-associated
protein amino acid
distance (data not shown) and DNA parsimony (fig. 1)
analyses argue for a closer relationship between TYDV
and the monocot-infecting
MiSV and WDV than between TYDV and other dicot-infecting
viruses. (Coat
protein analysis concurs
[fig. 21.) However, shortest
identified constraint trees forcing monocot- and dicotinfecting viruses to form monophyletic
clades did not
differ significantly
(cx = 0.05) from most-parsimonious
trees for either the replication-associated
(6,478 vs.
6,458 steps, respectively)
or the coat protein (4,407 vs.
4,392 steps, respectively)
DNA sequences.
Among certain dicot-infecting
viruses, ability to infect particular plant families is para- or polyphyletically
derived.
For example,
isolates
of the pseudotaxon
TYLCV which infect solanaceous
plants occupy positions basal to all (replication-associated
protein, fig. 1)
or most (coat protein, fig. 2) other dicot-infecting
viruses
(exclusive
of TYDV). This suggests that Solanaceaeinfecting viruses may be more closely related to viruses
infecting Cucurbitaceae,
Euphorbaceae,
Fabaceae, Malvaceae, and, for the replication-associated
protein, up to
44 other dicot families, than they are to other Solanaceae-infecting
viruses! Constraint trees consistent with
the assumption
that viruses infecting the Solanaceae
constitute a monophyletic
clade were generated from the
replication-associated
and coat protein DNA sequences.
Shortest identified
constraint
trees (6,822 and 4,7 16
steps for the replication-associated
and coat protein
DNA sequences, respectively)
differed significantly
(cx
= 0.05) from most-parsimonious
trees (6,458 and 4,392
steps for the replication-associated
and coat protein
DNA sequences, respectively).
Similarly, shortest identified constraint trees consistent with the assumption that
viruses infecting the Fabaceae constitute a monophyletic
clade (6,7 19 and 4,526 steps for the replication-associated and coat protein DNA sequences, respectively) differed significantly
from most-parsimonious
trees. Regardless of true tree polarity, among the geminivirus isolates used in this study, the ability to infect either the
Solanaceae or the Fabaceae cannot be a monophyletic-
and Gene Evolution
in Geminivirus
1119
ally derived trait. In contrast, shortest identified constraint trees consistent with the assumption that viruses
infecting the Poaceae constitute a monophyletic
clade
(6,478 and 4,407 steps for the replication-associated
and
coat protein DNA sequences, respectively) did not differ
significantly
from most-parsimonious
trees, suggesting
that the Poaceae-infecting
viruses included in this study
could have arisen from a common origin that is distinct
from that of any dicot-infecting
virus. Further sampling
of geminiviruses
might allow future evaluation of the
ability to infect other plant families.
Using distance analyses, Howarth and Vandemark
(1989) concluded that coat protein sequence evolution
may be constrained by insect vector specificity. Consistent with this possibility, our distance (not shown) and
parsimony (fig. 2) analyses of the coat protein amino
acid and DNA sequences reveal a dichotomy based on
insect vector specificity (whitefly- vs. leafhopper-transmitted). (The replication-associated
parsimony analysis
mostly agrees [fig. 11.)
We used estimates of nonsynonymous
and synonymous nucleotide substitutions
to evaluate the hypotheses of evolutionary
response of the replication-associated and coat proteins to accommodate
plant host and
insect vector specificities, respectively. In the absence of
positive selection, synonymous
nucleotide substitutions
(those preserving amino acid sequence) are expected to
survive more frequently than nonsynonymous
substitutions (those altering amino acid sequence). An increased
ratio of nonsynonymous-to-synonymous
changes suggests selective pressure in favor of amino acid sequence
alteration. Little is known about the rate of divergence
among or the time of origin of geminiviruses.
Definition
of evolutionary
events, as indicated by comparatively
high KJK, ratios, can become obscured over time by
accumulation
of synonymous changes. In the absence of
information
about the rate of divergence or the time of
origin, we have examined K, and KS estimates from both
extant and inferred
ancestral
sequences.
The latter
should provide a more accurate estimate by removing
substitutions
occurring
along terminal
branches.
Because similarity between certain pairs of coat protein
sequences is insufficient to allow meaningful estimates
of K, and KS (Li 1993), we have also limited our analyses of this gene to within- and between-group
comparisons
using the phylogenetically
defined African
streak virus and BGMV clades and viral sequences at
or near the proposed “switch point” between leafhopper- and whitefly-transmitted
groups. Although our analyses are unrooted, topology predicts for the replication-
t
FIG. l.-One
of two most-parsimonious
trees (6,458 steps, consistency index = 0.384, retention index = 0.601) identified by 100 heuristic
search replications with random taxon addition of the replication-associated
protein DNA sequence. Most-parsimonious
trees differed only in
the presence/absence
of the branch separating CSMV and other predominantly
monocot-infecting
viruses (indicated by a decay index of cl).
Isolate abbreviations
are listed in table 1. Numbers above branches are bootstrap percentages. Bootstrap values of less than 50% are indicated
as “(50.”
Numbers below branches are decay indices. Branches present in consensus trees four steps longer than the most-parsimonious
trees
are indicated as “>4.”
Numbers in parentheses at individual nodes represent inferred ancestral DNA sequences used to evaluate possible
evolutionary
response of the replication-associated
protein sequence to plant host specificity. Virus subgroups and characteristics
are listed
adjacent to isolate abbreviations.
“Genome” organization is monopartite (M) or bipartite (B); “Insect Vector” is leafhopper (L) or whitefly (W);
“Plant Host” is dicot (D) or monocot (M). Plant host families are listed in parentheses. Isolates are divided into dicot-infecting and predominantly
monocot-infecting
groups.
Whitefly-transmitted
Leafhopper-transmitted
__
I
I
African
Streak Virus
Clade
’
:
I
I
I
L_
II M L D (44 Families)
BCTV- 1
II M L D (44 Families)
BCT’V-2
100
>4
TYLCV-6
III M W D (Solanaceae)
TYLCV-7
III B W D (Solanaceae)
TLCrV- 1
III B W D (Solanaceae)
ToMoV
III B W D (Solanaceae)
__
I
I
I
I
I
I M L M (Poaceae)
PanSV
AbMV
III B W D (Malvaceae)
I M L M (Poaceae)
MSV-1
PYMV
III B W D (Solanaceae)
MSV-2
BDMV
III B W D (Fabaceae)
BGMV-3
III B W D (Fabaceae)
I
88
(5) >4
SLCV
III B W D (Cucurbitaceae)
I
.
TGMV
III B W D (Solanaceae)
I M L M (Poaceae)
I
M L M (Poaceae)
(11)
DSV
I M L M (Poaceae)
ssv
;
(12)
I M L M (Poaceae)
I M L M (Poaceae)
I M L D (Solanaceae)
I M L M (Poaceae)
CSMV
I M L M (Poaceae)
WDV-2
I M ? M (Poaceae)
MiSV
I
I
I
I
I
I
I
I
TYDV
WDV-1
(6)$
I
BGMV-2
III B W D (Fabaceae)
BGMV4
III B W D (Fabaceae)
BGMV-1
III B W D (Fabaceae)
TYLCV-2
III M W D (Solanaceae)
TYLCV-3
III M W D (Solanaceae)
TYLCV-1
III M W D (Solanaceae)
TYLCV-5
III M W D (Solanaceae)
TYLCV4
III M W D (Solanaceae)
I
(7) *
100
>4
I
(9) *
m
>4
(14)
r
92
%I
+-
+
; BGMV
I Clade
I
I
I
I
I
I
I
I
(19) L
(18)
I
I
ACMV
III B W D (Fabaceae)
I
TLCV-3
III M W D (Solanaceae)
I
TLCV-1
III B W D (Solanaceae)
I
TLCV-2
III B W D (Solanaceae)
I
MYMV
III B ? D (Fabaceae)
I
+
(15)
I
<50
3
I
l
,,I
FIG. 2.-One
of two most-parsimonious
trees (4,392 steps, consistency index = 0.438, retention index = 0.685) identified by 100 heuristic search replications with random taxon addition of the
coat protein DNA sequence. Most-parsimonious
trees differed only in the presence/absence
of the branch separating AbMV and PYMV (indicated by a decay index of cl). Isolate abbreviations are
listed in table 1. Numbers above branches are bootstrap percentages. Bootstrap values of less than 50% are indicated as “~50.” Numbers below branches are decay indices. Branches present in consensus
trees four steps longer than the most-parsimonious
trees are indicated as “>4.” Numbers in parentheses at individual nodes represent inferred ancestral DNA sequences used to evaluate possible
evolutionary response of the coat protein sequence to insect vector specificity. Virus characteristics
are as in figure 1. Isolates are divided into leafhopper-transmitted
and whitefly-transmitted
groups.
Genome
associated protein that the dicot- and/or predominantly
monocot-infecting
viruses arose after the dicot/monocot
switch point. Similarly, for the coat protein, the leafhopper- and/or whitefly-transmitted
viruses arose after
the leafhopper/whitefly
switch point. Because of the
possibility of accumulation
of synonymous
changes after the switch points, K, and KS values near (and basal
to) the switch points will be underestimated.
Elevated
KJK, ratios near (or basal to) the switch points therefore
indicate that, despite subsequent synonymous
changes,
the effects of a defining evolutionary
event are still molecularly evident. Conclusions based on K,IK, ratios and
K, versus KS plots for extant and inferred ancestral sequences were identical for both the replication-associated and coat protein genes. Here we present only K,
versus KS plots for inferred ancestral sequences.
If the replication-associated
protein amino acid sequence were evolving in response to plant host specificity, a dramatic increase in the frequency of nonsynonymous relative to synonymous
changes would be expected near the switch point between predominantly
monocot-infecting
and predominantly
dicot-infecting
virus groups. In figure 1, this switch point occurs along
the branch separating nodes 9 and 10. Pairwise estimates
of KJK, ratios for all inferred ancestral replication-associated protein DNA sequences were examined (not
shown). Values near the monocot/dicot
switch point are
in general higher than the population
average (0.56),
with the average KJK, value near (i.e., immediately adjacent to or one node removed from) the monocot/dicot
switch point being 0.85. However, Ka/Ks values reach a
maximum not near the switch point but in comparison
of sequences 32 and 33 (1.43). These ancestral sequences are embedded within the dicot-infecting
group and
are most closely related to BGMV strains. Furthermore,
some KJK, values within the predominantly
monocotinfecting clade are equal to or greater than those near
the monocot/dicot
switch point. Figure 3 shows plotted
pairwise data points for all inferred ancestral replicationassociated protein DNA sequences. Accompanying
data
trend lines suggest no substantial
differences between
within-monocot-infecting,
within-dicot-infecting,
and
K, and KS estimates.
between-dicot/monocot-infecting
These observations
suggest that, although there may be
some evolutionary
response of the replication-associated
protein gene to plant host specificity (as defined by Howarth and Vandemark [1989]; i.e., monocot vs. dicot),
other factors are likely involved. This contradicts
the
apparent dichotomy observed in our distance and parsimony analyses (exclusive of TYDV) and in the distance analysis of Howarth and Vandemark (1989). Alternatively, synonymous
substitutions
accumulated after
the switch point may be sufficient to obscure a more
pronounced
effect. Padidam,
Beachy,
and Fauquet
(199%~) concluded that host range does corroborate evolutionary relationships
predicted by total genome DNA
sequence and may not be a reliable taxonomic character.
If the coat protein amino acid sequence were evolving in response to insect vector specificity, a dramatic
increase in the frequency of nonsynonymous
changes
relative to synonymous
changes would be expected near
and Gene Evolution
in Geminivirus
MM
DD
DM
KS
FIG. 3.-Plot
of K, versus KS (nonsynonymous
vs. synonymous
nucleotide substitutions) estimates for all pairwise comparisons of inferred ancestral replication-associated
protein DNA sequences. Data
symbols and trend lines indicate values within dicot-infecting
(B and
DD), within monocot-infecting
(A and MM), and between dicot- and
monocot-infecting
(0 and DM) isolates.
the switch point between leafhopper- and whitefly-transmissible virus groups. In figure 2, this switch point occurs along the branch separating nodes 11 and 12. Distances among extant and ancestral coat protein DNA
sequences are in general higher than those observed
among extant and ancestral replication-associated
protein DNA sequences (data not shown), suggesting that
protein is more highly conthe replication-associated
served.
Pairwise estimates of K, and KS were examined for
inferred ancestral coat protein DNA sequences (data not
shown). K,IK, values peak near the leafhopper/whitefly
switch point and all switch point values (average 0.82)
substantially
exceed African streak virus (average 0.28)
and BGMV (0.15) clade estimates. Figure 4 shows plotted pairwise data points for inferred ancestral coat protein DNA sequences.
Accompanying
data trend lines
suggest substantial
differences
between
leafhopper/
whitefly switch point values and within African streak
virus and BGMV clade estimates. These observations
support the hypothesis that the coat protein sequence
evolves in response to insect vector specificity. It is interesting to note that observed K,IK, values are consistent with a gradual progression from leafhopper transmission to whitefly transmission.
Sequence 12 is the earliest inferred
whitefly-transmitted
ancestor. This sequence
may be oniy inefficiently
transmitted
by
whiteflies. Subsequent further nonsynonymous
changes
in DNA sequence give rise to sequences 13 and 14, as
demonstrated
by high K.JKs ratios in comparisons
of
these sequences with sequence 12 (0.86 for sequence 12
vs. 13 and 1.13 for sequence 12 vs. 14). (The reverse
pattern, progression from whitefly transmission
to leafhopper transmission,
is less likely, as it requires a grad-
1122
Bradeen
et al.
LW
LL
ww
1
0
50
100
150
200
250
KS
vs. synonymous
FIG. 4.-Plot
of K, versus KS (nonsynonymous
nucleotide substitutions) estimates for pairwise comparisons of inferred
ancestral coat protein DNA sequences near the leafhopper/whitefly
switch point and within the African streak virus and BGMV clades.
Data symbols and trend lines indicate values within the leafhoppertransmitted African streak virus clade (m and LL), within the whiteflytransmitted BGMV clade (A and
and near the leafhopper/whitefly switch point (0 and LW).
ual loss of whitefly transmissibility
in the progenitor
clade prior to acquisition of leafhopper transmissibility.)
Consistent with our conclusion,
Briddon et al. (1990)
demonstrated,
using an ACMV construct modified by
the replacement of its native coat protein gene with that
of BCTV, that the coat protein gene alone is sufficient
to alter insect vector specificity. Geographical
isolation
was likely involved in the divergence and adaptation of
the coat protein sequence from the earliest whiteflytransmitted
viruses (Padidam,
Beachy, and Fauquet
1995a).
Average within-group
K,IK, values are notably
higher for African streak (leafhopper)
virus clade sequences (0.28) than for BGMV (whitefly) clade sequences (0.15). Padidam, Beachy, and Fauquet (1995a)
suggested that the coat protein is less constrained evolutionarily in leafhopper- than in whitefly-transmitted
viruses. The opposite may be true. Lower K,IK, values for
within-BGMV
comparisons could reflect the fact that all
BGMV viruses are vectored by the same whitefly species (Bemisia tabaci), whereas African streak clade viruses are vectored by at least two different genera of
leafhopper (Cicadulina and Nesoclutha).
This suggests
that the coat protein may be important not only in specifying leafhopper versus whitefly transmissibility,
but
also in specifying individual genera or species of insect
vector. Alternatively,
the coat protein sequence may be
constrained in leafhopper-transmitted
types by factors in
addition to insect vector specificity. Consistent with this
possibility, it has been demonstrated
for several whitefly-transmitted
viruses that an intact coat protein is not
necessary for infectivity following mechanical or agrobacterium
inoculation
(Gardiner et al. 1988; Klinken-
berg, Ellwood, and Stanley 1989; Padidam, Beachy, and
Fauquet 1995b). In contrast, the coat protein is important for systemic viral spread and symptom development
for leafhopper-transmitted
viruses (Boulton et al. 1989;
Briddon et al. 1989; Lazarowitz et al. 1989; Woolston
et al. 1989).
Geminiviruses
are as diverse as the agronomic diseases for which they are responsible. Classification
systems in the past have relied on genome organization,
plant host, and insect vector. Evidence presented here
suggests that genome organization
and plant host specificity may not necessarily reflect evolutionary
relationships. Our research supports recent modifications
in
geminivirus
classification
that includes deemphasis
of
these characteristics.
Future taxonomic efforts may rely
heavily on distance and parsimony analyses of the replication-associated
and coat protein sequences. Evidence
that the coat protein evolves in response to insect vector
specificity suggests that geminiviruses
might be useful
specimens for the study of gene evolution.
Acknowledgments
The authors wish to thank Dr. Gregory King and
Dr. Vineet Bafna for help with computer programming
and Dr. Robert Vrijenhoek
for helpful advice and insight. This work was supported by grant DE-FG0295ER20194 from the U.S. Department of Energy to J.M.
LITERATURE
ABOUZID,
CITED
A. M., J. E. POLSTON,
and E. HIEBERT.
1992.
The
nucleotide sequence of tomato mottle virus, a new geminivirus isolated from tomatoes in Florida. J. Gen. Virol. 73:
3225-3229.
ANDERSEN, M. T, K. A. RICHARDSON, S. HARBISON, and B.
A. M. MORRIS. 1988. Nucleotide sequence of the geminivirus chloris striate mosaic virus. Virology 164:443-449.
Y., and S. COHEN. 1994. Complete nucleotide sequence of an infectious clone of a mild isolate of tomato
yellow leaf curl virus (TYLCV). Phytopathology
84:707712.
BOULTON, M. I., H. STEINKELLNER,J. DONSON, P G. MARKHAM, D. I. KING, and J. W. DAVIES. 1989. Mutational analysis of the virion-sense genes of maize streak virus. J. Gen.
Virol. 70:2309-2323.
BREMER, K. 1988. The limits of amino acid sequence data in
angiosperm phylogenetic reconstruction. Evolution 42:795803.
BRIDDON, R. W., P LUNNESS, L. C. L. CHAMBERLIN, M. S.
PINNER, H. BRUNDISH, and I? G. MARKHAM. 1992. The nucleotide sequence of an infectious insect-transmissible
clone
of the geminivirus Panicurn streak virus. J. Gen. Virol. 73:
ANTIGNUS,
1041-1047.
BRIDDON, R. W. and E? G. MARKHAM. 1995. Geminiviridae.
Pp. 158-l 65 in E A. MURPHY, C. M. FAUQUET, M. A.
MAYO, A. W. JAVIS, S. A. GHABRIAL, M. D. SUMMERS, G.
l? MARTELLI, and D. H. L. BISHOP, eds. The classification
and nomenclature
of viruses: sixth report of the International Committee on Taxonomy of Viruses. Archives of Virology. Suppl. 10. Springer Verlag, Wein.
BRIDDON, R. W., M. S. PINNER, J. STANLEY, and I? G. MARKHAM. 1989. The coat protein of beet curly top virus is essential for infectivity. Virology 172:628-633.
Genome
-
1990. Geminivirus coat protein gene replacement alters insect specificity. Virology 177:85-94.
CHATANI, M., Y. MATSUMOTO, H. MIZUTA, M. IKEGAMI, M. I.
BOULTON, and J. W. DAVIES. 1991. The nucleotide sequence
and genome structure of the geminivirus miscanthus streak
virus. J. Gen. Virol. 72:2325-233 1.
COUTTS, R. H. A., R. S. COFFIN, E. J. E ROBERTS, and W. D.
0. HAMILTON. 1991. The nucleotide sequence of the infectious cloned DNA components of potato yellow mosaic virus. J. Gen. Virol. 72:1515-1520.
DONOGHUE, M. J., R. G. OLMSTEAD, J. E SMITH, and J. D.
PALMER. 1992. Phylogenetic
relationships
of Dipsacales
based on &CL sequences. Ann. MO. Bot. Gard. 79:333-345.
DONSON, J., G. F? ACCOTTO, M. I. BOULTON, I? M. MULLINEAUX, and J. W. DAVIES. 1987. The nucleotide sequence
of a geminivirus from Digitaria sanguinalis. Virology 161:
160-169.
DRY, I. B., J. E. RIGDEN, L. R. KRAKE, I? M. MULLINEAUX,
and M. A. REZAIAN. 1993. Nucleotide sequence and genome organization of tomato leaf curl geminivirus. J. Gen.
Virol. 74:147-151.
FELSENSTEIN, J. 1984. Distance methods for inferring phylogenies: a justification. Evolution 38: 16-24.
-.
1985. Confidence limits on phylogenies: an approach
using the bootstrap. Evolution 39:783-29 1.
1989. PHYLIP-phylogeny
inference package. Version-3.2. Cladistics 5: 164-166.
FRISCHMUTH, T., G. ZIMMAT, and H. JESKE. 1990. The nucleotide sequence of abutilon mosaic virus reveals prokaryotic
as well as eukaryotic features. Virology 178:461-468.
GARDINER, W. E., G. SUNTER, L. BRAND, J. S. ELMER, S. G.
ROGERS, and D. M BISARO. 1988. Genetic analysis of tomato golden mosaic virus: the coat protein is not required
for systemic spread or symptom development. EMBO J. 7:
899-904.
GILBERTSON, R. L., J. C. FARIA, S. E HANSON, E J. MORALES,
I? AHLQUIST, D. F? MAXWELL, and D. R. RUSSELL. 1991.
Cloning of the complete DNA genomes of four bean-infecting geminiviruses
and determining their infectivity by
electric discharge particle acceleration. Phytopathology
81:
980-985.
GILBERTSON, R. L., S. H. HIDAYAT, E. J. PAPLOMATAS, M. R.
ROJAS, Y. Hou, and D. I? MAXWELL. 1993. Pseudorecombination between infectious cloned DNA components of tomato mottle and bean dwarf mosaic geminiviruses. J. Gen.
Virol. 74:23-3 1.
HAMILTON, W. D. O., V. E. STEIN, R. H. A. COUTTS, and K.
W. BUCK. 1984. Complete nucleotide sequence of the infectious cloned DNA components of tomato golden mosaic
virus: potential coding regions and regulatory sequences.
EMBO J. 3:2197-2205.
HOWARTH, A. J., J. CATON, M. BOSSERT, and R. M. GOODMAN.
1985. Nucleotide sequence of bean golden mosaic virus and
a model for gene regulation in geminiviruses.
Proc. Natl.
Acad. Sci. USA 82:3572-3576.
HOWARTH, A. J., and R. M. GOODMAN. 1986. Divergence and
evolution of geminivirus genomes. J. Mol. Evol. 23:323329.
HOWARTH, A. J., and G. J. VANDEMARK. 1989. Phylogeny of
geminiviruses.
J. Gen. Virol. 70:27 17-2727.
HUGHES, E L., E. I? RYBICKI, and R. KIRBY. 1993. Complete
nucleotide sequence of sugarcane streak Monogeminivirus.
Arch. Virol. 132: 17 l-l 82.
KHEYR-POUR, A., M. BENDAHMANE, V. MATZEIT, G. I? AcCOTTO, S. CRESPI, and B. GRONENBORN. 1992. Tomato yellow leaf curl virus from Sardinia is a whitefly-transmitted
and Gene Evolution
in Geminivirus
1123
monopartite
geminivirus.
Nucleic Acids Res. 19:67636769.
KIKUNO, R., H. TOH, H. HAYASHIDA, and T MIYATA. 1984.
Sequence similarity between putative gene products of geminiviral DNAs. Nature 308:562.
KISHINO, H., and M. HASEGAWA. 1989. Evaluation of the maximum likelihood estimate of the evolutionary tree topologies from DNA sequence data, and the branching order in
Hominoidea. J. Mol. Evol. 29:170-179.
KLINKENBERG,E A., S. ELLWOOD, and J. STANLEY. 1989. Fate
of African cassava mosaic virus coat protein deletion mutants after agroinoculation.
J. Gen. Virol. 70: 1837-1844.
KOONIN, E. V., and T. V. ILYINA. 1992. Geminivirus replication
proteins are related to prokaryotic plasmid rolling circle
DNA replication initiator proteins. J. Gen. Virol. 73:27632766.
LAZAROWITZ, S. G. 1988. Infectivity and complete nucleotide
sequence of the genome of a South African isolate of maize
streak virus. Nucleic Acids Res. 16:229-249.
LAZAROWITZ, S. G., and I. B. LAZDINS. 1991. Infectivity and
complete nucleotide sequence of the cloned genomic components of a bipartite squash leaf curl geminivirus with a
broad host range phenotype. Virology 180:58-69.
LAZAROWITZ, S. G., A. J. PINDER, V. D. DAMSTEEGT, and S.
G. ROGERS. 1989. Maize streak virus genes essential for
systemic spread and symptom development.
EMBO J. 8:
1023-1032.
LI, W. 1993. Unbiased estimation of the rates of synonymous
and nonsynonymous
substitution. J. Mol. Evol. 36:96-99.
MACDOWELL, S. W., H. MACDONALD, W. D. 0. HAMILTON,R.
H. A. COUTTS, and K. W. BUCK. 1985. The nucleotide sequence of cloned wheat dwarf virus DNA. EMBO J. 4:
2173-2180.
MADDISON, D. R. 1991. The discovery and importance of multiple islands of most parsimonious
trees. Syst. Zool. 40:
3 15-328.
MATTHEWS, R. E. F. 1985. Viral taxonomy for the nonvirologist. Annu. Rev. Microbial. 39:45 l-474.
MORINAGA, T., M. IKEGAMI, and K. MIURA. 1993. The nucleotide sequence and genome structure of mung bean yellow
mosaic geminivirus. Microbial. Immunol. 37:47 l-476.
MORRIS, B., L. COATES, S. LOWE, K. RICHARDSON, and I?
EDDY. 1990. Nucleotide sequence of the infectious cloned
DNA components of African cassava mosaic virus (Nigerian strain). Nucleic Acids Res. l&197-198.
MORRIS, B. A. M., K. A. RICHARDSON, A. HALEY, X. ZHAN,
and J. E. THOMAS. 1992. The nucleotide sequence of the
infectious cloned DNA component of tobacco yellow dwarf
virus reveals features of geminiviruses
infecting monocotyledonous plants. Virology 187:633-642.
MULLINEAUX, I? M., J. DONSON, B. A. M. MORRIS-KRSINICH,
M. I. BOULTON, and J. W. DAVIES. 1984. The nucleotide
sequence of maize streak virus DNA. EMBO J. 3:30633068.
MULLINEAUX, I? M., E GUERINEAU, and G. I? ACCOTTO. 1990.
Processing
of complementary
sense RNAs of Digitaria
streak virus in its host and in transgenic tobacco. Nucleic
Acids Res. l&7259-7265.
MUMFORD, D. L. 1974. Purification of curly top virus. Phytopathology 64: 136-139.
NAVOT, N., E. PICHERSKY, M. ZEIDAN, D. ZAMIR, and H.
CZOSNEK. 199 1. Tomato yellow leaf curl virus: a whiteflytransmitted geminivirus with a single genomic component.
Virology 185:151-161.
NORIS, E., E. HIDALGO, G. F? ACCOTTO, and E. MORIONES.
1994. High similarity among the tomato yellow leaf curl
virus isolates from the West Mediterranean
Basin: the nu-
1124 Bradeen et al.
cleotide sequence of an infectious clone from Spain. Arch.
Virol. 135: 165-170.
PADIDAM, M., R. N. BEACHY, and C. M. FAUQUET. 199%.
Classification and identification of geminiviruses
using sequence comparisons. J. Gen. Virol. 76:249-263.
-.
199%. Tomato leaf curl geminivirus from India has a
bipartite genome and coat protein is not essential for infectivity. J. Gen. Virol. 76:25-35.
PAPLOMATAS, E. J., V. I? PATEL, Y. Hou, A. 0. NOUEIRY, and
R. L. GILBERTSON. 1994. Molecular characterization
of a
new sap-transmissible
bipartite genome geminivirus infecting tomatoes in Mexico. Phytopathology
84: 12 15-l 224.
ROCHESTER, D. E., J. J. DEPAULO, C. M. FAUQUET, and R. N.
BEACHY. 1994. Complete nucleotide sequence of the geminivirus tomato yellow leaf curl virus, Thailand isolate. J.
Gen. Virol. 75:477-485.
SCHALK, H. J., V. MATZEIT, B. SCHILLER, J. SCHELL, and B.
GRONENBORN. 1989. Wheat dwarf virus, a geminivirus of
graminaceous plants, needs splicing for replication. EMBO
J. 8:359-364.
STANLEY, J., I? G. MARKHAM, R. J. CALLIS, and M. S. PINNER.
1986. The nucleotide sequence of an infectious clone of the
geminivirus beet curly top virus. EMBO J. 5:1761-1767.
STENGER, D. C. 1994. Complete nucleotide sequence of the
hypervirulent
CFH strain of beet curly top virus. Mol.
Plant-Microbe Interact. 7: 154-157.
SWOFFORD, D. L. 1993. PAUP: phylogenetic
analysis using
parsimony. Version 3.1.1. Smithsonian Institution, Washington, D.C.
TEMPLETON, A. R. 1983. Phylogenetic inference from restriction endonuclease cleavage site maps with particular reference to the evolution of humans and the apes. Evolution
37:221-244.
TIMMERMANS,M. C. l?, 0. I? DAS, and J. MESSING. 1994. Geminiviruses and their uses as extrachromosomal
replicons.
Annu. Rev. Plant Physiol. Plant Mol. Biol. 45:79-l 12.
WOOLSTON, C. J., H. V. REYNOLDS, N. J. STACEY, and F? M.
MULLINEAUX. 1989. Replication of wheat dwarf virus DNA
in protoplasts and analysis of coat protein mutants in protoplasts and plants. Nucleic Acids Res. 17:6029-6041.
JULIAN I? ADAMS, reviewing
Accepted
August
1, 1997
editor