Taxonomic and phytogeographic implications from ITS phylogeny in Berberis (Berberidaceae)

Abstract

A phylogeny based on the internal transcribed spacer (ITS) sequences from 79 taxa representing much of the diversity of Berberis L. (four major groups and 22 sections) was constructed for the first time. The phylogeny was basically congruent with the previous classification schemes at higher taxonomic levels, such as groups and subgroups. A notable exception is the non-monophyly of the group Occidentales of compound-leaved Berberis (previously separated as Mahonia). At lower levels, however, most of previous sections and subsections were not evident especially in simple-leaved Berberis. Possible relationship between section Horridae (group Occidentales) and the simple-leaved Berberis clade implies paraphyly of the compound-leaved Berberis. A well-known South America-Old World (mainly Asia) disjunctive distribution pattern of the simple-leaved Berberis is explained by a vicariance event occurring in the Cretaceous period. The ITS phylogeny also suggests that a possible connection between the Asian and South American groups through the North American species ( Berberis canadensis or B. fendleri) is highly unlikely.

Full text

J Plant Res (2004) 117:175–182 © The Botanical Society of Japan and Springer-Verlag Tokyo 2004
Digital Object Identifier (DOI) 10.1007/s10265-004-0145-7
S
pr
i
nger-
V
er
l
ag
T
o
k
yo
102650918
-
94401618
-
086030669031J
ourna
l
o
f
Pl
ant
R
esearc
hJ
Pl
an
t
R
es
14510
.
1007/
s
10265
-
004
-
0145
-
7
Taxonomic and phytogeographic implications from ITS phylogeny in
Berberis
(Berberidaceae)
ORIGINAL ARTICLE
Received: September 11, 2003 / Accepted: January 12, 2004 / Published online: March 10, 2004
Young-Dong Kim • Sung-Hee Kim • Leslie R. Landrum
Y.-D. Kim (*) · S.-H. Kim
Division of Biological Sciences, Hallym University, 200-702 Chunchon,
South Korea
Tel. +82-33-2482095; Fax +82-33-2444904
e-mail: ydkim@hallym.ac.kr
L.R. Landrum
School of Life Sciences, Arizona State University, Tempe, AZ, USA
Abstract A phylogeny based on the internal transcribed
spacer (ITS) sequences from 79 taxa representing much
of the diversity of Berberis L. (four major groups and
22 sections) was constructed for the first time. The phylog-
eny was basically congruent with the previous classification
schemes at higher taxonomic levels, such as groups and
subgroups. A notable exception is the non-monophyly of
the group Occidentales of compound-leaved Berberis (pre-
viously separated as Mahonia). At lower levels, however,
most of previous sections and subsections were not evident
especially in simple-leaved Berberis. Possible relationship
between section Horridae (group Occidentales) and the
simple-leaved Berberis clade implies paraphyly of the
compound-leaved Berberis. A well-known South America-
Old World (mainly Asia) disjunctive distribution pattern of
the simple-leaved Berberis is explained by a vicariance
event occurring in the Cretaceous period. The ITS phylog-
eny also suggests that a possible connection between the
Asian and South American groups through the North
American species (Berberis canadensis or B. fendleri) is
highly unlikely.
Key words Berberis · ITS · Mahonia · Phylogeny · Phyto-
geography · Vicariance
Introduction
Berberis L. (barberry) is the largest genus in Berberidaceae,
a member of the basal eudicots in the flowering plants (Judd
et al. 2002). The genus comprises about 500 species with
simple leaves (true Berberis) and about 200 species with
compound leaves, which are often recognized as a separate
genus, Mahonia Nuttall (Fedde 1902; Schneider 1905;
Ahrendt 1961). The actual numbers, however, may be
much fewer, considering that Ahrendt (1961) recognized
60 species in Chile and adjacent southern Argentina
whereas Landrum (1999) accepted only 20. Recent molec-
ular phylogenetic studies (Kim and Jansen 1995, 1998) place
the shrubby genus within a group with the chromosome
base number x = 7 in which a herbaceous genus Ranzania
Ito is included and agree with morphological and cytologi-
cal phylogenetic analyses (Meacham 1980; Terabayashi
1985).
Although no phylogenetic studies have been conducted
for the genus, the division of the simple-leaved Berberis into
two groups, Septentrionales C.K. Schneid. and Australes
C.K. Schneid., and the compound-leaved Berberis into two
groups, Occidentales Ahrendt and Orientales Ahrendt is
fundamental in the classification of the genus (Schneider
1905; Ahrendt 1961). The species of Septentrionales are pre-
dominantly Eurasian except for two in North America and
four in North Africa while those of Australes are exclusively
South American. As the group names imply, the species of
Occidentales are exclusively distributed in North and Cen-
tral America whereas Orientales consists of southeast Asian
taxa with one remarkable exception, Berberis nervosa in
North America. These seemingly geographically based
groups are known to be supported by combinations of mor-
phological characters (Ahrendt 1961). For example, folia-
ceous spines, dentate filaments, and deep yellow or orange
flowers are usually found in Australes, although none of
these characters indisputably define the group.
In subdividing those groups, Ahrendt (1961) accepted
the schemes of Schneider (1905) and Fedde (1902) with
some modification, and recognized 29 sections of simple-
leaved Berberis and four sections of compound-leaved Ber-
beris. Ahrendt (1961) also further divided some of these
sections (i.e., sect. Wallichianae in the simple-leaved
Berberis and sect. Longibracteatae in the compound-leaved
Berberis) into numerous subsections. Validity of these sec-
tions or subsections, however, has been strongly questioned
by Landrum (1999) as he merged three species in different
sections under a single name.

176
One intriguing problem of plant distribution is the
widely disjunct nature of the simple-leaf group of Berberis,
with centers in South America and Asia, and only a few
species in other areas. Has the group arisen more than once
from the compound-leaf group (i.e., is it polyphyletic)? If
monophyletic, when and through what means did it obtain
its widely disjunct distribution? Looking at a map of the
present distribution of Berberis, migration from North
America to South America seems like the probable answer.
However, as will be discussed below, Berberis appears to
have a long history in South America, which would indicate
arrival through another pathway, before North America and
South America came into contact.
In this paper, the previous taxonomic hypotheses regard-
ing the classification of Berberis sensu lato are tested using
the phylogeny reconstructed by the internal transcribed
spacer (ITS) sequences of the nuclear ribosomal RNA gene.
Specific systematic issues such as validity of taxonomic
groups in previous classification schemes, circumscription of
the compound-leaved Berberis, and phytogeography of the
genus are also discussed.
Materials and methods
Plant samples, PCR and sequencing
A total of 79 taxa representing four major groups and
22 sections in Berberis were selected to reconstruct ITS phy-
logeny (Table 1). Total genomic DNA was isolated from
fresh leaf tissue using a modified CTAB (hexadecyl trime-
thylammonium bromide) procedure (Doyle and Doyle
1987), followed by purification using ultracentrifugation in
a cesium chloride/ethidium bromide gradient. For dried
herbarium materials, the methods of Loockerman and
Jansen (1996) were followed. To amplify the ITS regions of
the nuclear ribosomal DNA, PCR (polymerase chain reac-
tion) was carried out in a 100-ml reaction containing about
50–100 ng of template DNA, 10 ml of 10¥ reaction buffer,
2.5 U ExTaq (Takara), 8 ml of 200 mM dNTPs in an equimo-
lar ratio, and 50 pmol of each primer. The PCR and
sequencing primers were identical to the ones designed by
White et al. (1990) except ITS1 (5¢-GGAAGGAGAAGT
CGTAACAAGG-3¢), which differed by the two italicized
bases. DMSO (5%) was added for some difficult DNA sam-
ples which produced fuzzy or double bands under initial
PCR amplifying conditions. The thermal cycler was pro-
grammed to perform an initial 3 min denaturation at 95°C,
1 min annealing at 50°C, and 1 min extension at 72°C. This
was followed by 30 cycles with 1 min denaturation at 95°C,
1 min annealing at 50°C, and 1 min extension at 72°C. The
product was terminated with a final extension for 10 min at
72°C and soaking at 15°C. To remove unused amplifying
primers and dNTPs, the PCR product was purified by a
QIAquick PCR purification kit (Qiagen, Valencia, Calif.)
according to the manufacturer’s protocol. The purified
double-stranded PCR products were used for determining
the ITS sequences using the automatic DNA sequencer,
ABI PRISM 377 (PE Applied Biosystems).
Table 1. Species of Berberis sequenced for internal transcribed spacer (ITS) sequences. Each species was assigned to sections (Ahrendt 1961).
Accession numbers are used as voucher information for plants collected at Arnold Arboretum (AA) and Royal Botanic Garden Edinburgh
(RBGE)
Species Section/subsection Voucher information Distribution
Simple-leaved Berberis
Group Septentrionales
B. circumserrata C.K. Schneid. Angulosae/Diaphanae AA 1098-65 China
B. morrisonensis Hayata Angulosae/Diaphanae RBGE 19934250A Taiwan
B. tischleri C.K. Schneid. Angulosae/Diaphanae RBGE 19913295A China
B. cooperi Ahrendt Angulosae/Subangulosae RBGE 19151002A Bhutan
B. brachypoda Maxim. Brachypodae AA 978-83 China
B. gilgiana Diels Brachypodae AA 789-82 China
B. fendleri A. Gray Canadenses Hill 15102 (
GH) USA
B. canadensis Mill. Canadenses Chase 13196 (K) USA
B. dielsiana Fedde Dasystachyae AA 173-83 China
B. zabeliana C.K. Schneid. Dasystachyae AA 758-84 C Asia
B. lecomtei C.K. Schneid. Franchetianae/Eufranchetianae RBGE 19943851A China
B. amurensis var. quelpaertensis Nakai Franchetianae/Sieboldiae YDK 01-0627-1 (Hallym
University)
Korea
B. sieboldi Miq. Franchetianae/Sieboldiae RBGE 19620231A Japan
B. vernae C.K. Schneid. Integerrimae AA 725-76 China
B. wilsonae Hemsl. Polyanthae/Pseudopolyanthae AA 450-87 China
B. edgeworthiana Schneid. Polyanthae/Subpolyanthae RBGE 19687160A India, China
B. coriaria Royle ex Lindl. Tintoria/Chitriae RBGE 19531004A India
B. holstii Engl. Tintoria/Chitriae RBGE 19510094A E Africa
B. thunbergii DC. Tschonoskyanae AA 1040-90 Japan
B. amurensis Rupr. var. amurensis Vulgares YDK 01-0526-1 (Hallym
University)
Korea
B. amurensis var. latifolia Nakai Vulgares Hyun 01-11 (Hallym University) Korea
B. koreana Palib. Vulgares YDK1996-1 (Hallym University) Korea
B. vulgaris L. Vulgares AA 538-78 Europe

177
B. bergmanniae C.K. Schneid. Wallichianae/Euwallichianae RBGE 19754095A China
B. dumicola C.K. Schneid. Wallichianae/Euwallichianae RBGE 19920032A China
B. julianae C.K. Schneid. Wallichianae/Euwallichianae AA 7178 China
B. sargentiana C.K. Schneid. Wallichianae/Euwallichianae RBGE 19784169A China
B. wallichiana DC. Wallichianae/Euwallichianae RBGE 19920456A Nepal
B. hookeri var. viridis C.K. Schneid. Wallichianae/Hookerianae RBGE 19341064A Bhutan, India
B. insignis Hook. f. & Thomson Wallichianae/Insignes RBGE 19230082A Bhutan, India
B. kawakamii Hayata Wallichianae/Leves RBGE 1993107A Taiwan
B. coxii C.K. Schneid. Wallichianae/Manipuranae RBGE 19190075A Burma
B. manipurana Ahrendt Wallichianae/Manipuranae RBGE 19734292A Manipur
B. pruinosa Franch. Wallichianae/Pruinosae RBGE 19910581A China
B. sanguinea Franch. Wallichianae/Sanguineae RBGE 19390166A China
B. soulieana C.K. Schneid. Wallichianae/Soulieanae RBGE 19381183A China
B. gagnepainii var. lanceifolia Ahrendt Wallichianae/Triacanthophorae RBGE 19091019 China
B. candidula C.K. Schneid. Wallichianae/Verruculosae RBGE 19902425 China
B. verruculosa Hemsl. & E.H. Wilson Wallichianae/Verruculosae RBGE 19687180A China
Group Australes
Subgroup Aequinoctiales
B. lutea Ruíz & Pav. Virgatae Panero 2937 (
TEX) Ecuador
B. jamesonii Lindl. Paniculatae Panero 2946 (TEX) Ecuador
Subgroup Euaustrales
B. horrida Gay Actinacanthae/Congestiflorae Landrum 7914 (ASU) Chile
B. grevilleana Gillies ex Hook. & Arn. Actinacanthae/Euactinacanthae Landrum 8340 (ASU) Chile, Argentina
B. chilensis Gillies ex Hook. Illicifoliae/Chilenses Bricker 186 (ASU) Chile
B. darwinii Hook. Illicifoliae/Darwinianae Landrum 7625 (ASU) Argentina, Chile
B. comberi Sprague & Sandwith Illicifoliae/Euillicifoliae Landrum 8378 (ASU) Argentina
B. ilicifolia Forst. Ilicifoliae/Euilicifoliae Landrum 8058 (ASU) Argentina, Chile
B. valdiviana Phil. Laurinae/Flexuosae Landrum 8023 (ASU) Chile
B. microphylla G. Forst. Microphyllae RBGE 19902428A Chile, Argentina
B. trigona Kunze ex Poepp. & Endl. Trigonae Landrum 8043 (ASU) Chile
Compound-leaved Berberis
Group Occidentales
B. higginsae Munz Horridae YDK 90-1 (TEX) USA
B. fremontii Torr. Horridae RBGE 19716619A USA
B. haematocarpa Wooton Horridae RBGE 19781766A USA
B. nevinii A. Gray Horridae Rancho Santa Ana Botanical
Garden sn.
USA
B. trifoliolata Moric. Horridae Turner 99-135 (TEX) USA, Mexico
B. aquifolium Pursh Aquifoliatae/Euaquifoliatae RBGE 19912984A US
B. pinnata Lag. Aquifoliatae/Euaquifoliatae AA 142-55 US
B. piperiana (Abrams) MacMinn Aquifoliatae/Euaquifoliatae AA 393-57 US
B. eutriphylla (Fedde) C.H. Müll. Aquifoliatae/Schiedeanae Beaman 2311 (TEX) Mexico
B. gracilis Benth. Aquifoliatae/Schiedeanae Henrickson 23249 (TEX) Mexico
B. schiedeana Schltdl. Aquifoliatae/Schiedeanae Hinton et al. 8902 (
GH) Mexico
B. harrisoniana Kearny & Peebles Not classified Anderson 95-6 (ASU)US
B. ehrenbergii Kunze Paniculatae/Eupaniculatae NSW 102983 (A) Mexico
B. lanceolata Benth. Paniculatae/Eupaniculatae Webster et al. 20172 (GH) Mexico
B. pallida Hartw. ex Benth. Paniculatae/Eupaniculatae RBGE 19930103C Mexico
B. hemsleyi Donn. Sm. Paniculatae/Eupaniculatae Mexia 1503 (GH) Costa Rica,
Guatemala
B. paxii (Fedde) Marroq. & Laferr. Paniculatae/Eupaniculatae Moore, Jr. 3853 (A) Mexico
Group Orientales
B. gracilipes Oliver. Longibracteatae/Dolichopodae RBGE 19891713 China
B. polyodonta (Fedde) Laferr. Longibracteatae/Dolichostyles RBGE 19812798A China
B. longibracteata (Takeda) Laferr. Longibracteatae/Eulongibracteatae T.S. Ying & D.E. Boufford 0954 (A) China
B. eurybracteata (Fedde) Laferr. Longibracteatae/Eulongibracteatae Bartholomew et al. 2311 (A) China
B. veitchiorum Hemsl. & E.H. Wils. Longibracteatae/Eulongibracteatae RBGE 19911138B China
B. bodinieri (Gagnep.) Laferr. Longibracteatae/Japonicae D.E. Boufford & B. Bartholomew
24542 (
A)
China
B. japonica (Thunb.) R. Br. Longibracteatae/Japonicae RBGE 19751278 China
B. flavida (C.K. Schneid.) Laferr. Longibracteatae/Napaulenses RBGE 19923156E China
B. lomariifolia (Takeda) Laferr. Longibracteatae/Napaulenses RBGE 19560292A China
B. napaulensis (DC.) Laferr. Longibracteatae/Napaulenses RBGE 19540505A Nepal
B. fortunei Lindl. Longibracteatae/Nervosae D.E. Boufford & B. Bartholomew
23980 (
A)
China
B. nervosa Pursh Longibracteatae/Nervosae RBGE 19782559 USA, Canada
Species Section/subsection Voucher information Distribution
Table 1. (Continued)

178
Sequence and phylogenetic analyses
The boundaries of ITS-1 and ITS-2 were determined by
comparison with known sequences (Suh et al. 1993; Kim
and Jansen 1994). ITS sequences were aligned using Clustal
W (version 1.6 [Thompson et al. 1995]). Gaps introduced
from the alignment were treated as missing characters in
subsequent analyses.
Phylogenetic analyses were performed using the
neighbor-joining (NJ) (Saitou and Nei 1987) and Bayesian
(Rannala and Yang 1996) methods. The general time revers-
ible model (GTR [Yang 1994]) implemented in PAUP
*
(ver-
sion 4.0b [Swofford 2002]) was employed for the NJ
method. Bootstrap analysis (Felsenstein 1985) with 1,000
pseudoreplicates was conducted to evaluate the degree of
support for given clades of the NJ tree.
Bayesian phylogenetic analyses were conducted with
MrBayes 2.0 (Huelsenbeck and Ronquist 2001) using the
GTR model. Each Markov chain was started from a random
tree and run for 2,000,000 generations. We ran four chains
simultaneously, three heated and one cold, with the default
settings for the priors. To check that stationarity had
been reached, we monitored the fluctuating value of the
log-likelihood and repeated each analysis twice. After
discarding burn-in samples (initial 200,000 generations) the
remaining samples were retained for the construction of
final consensus tree with posterior probabilities for given
clades.
The NJ and Bayesian trees were rooted at the midpoint
(Swofford et al. 1996). Rate uniformity for the midpoint
rooting was tested using the likelihood-ratio test
(Felsenstein 1988). Maximum likelihood values for the test
were calculated with the GTR model implemented in
PAU P
*
(version 4.0b [Swofford 2002]). We used the mid-
point rooting option because the ITS sequence ofRanzania
Ito, a well known monotypic sister genus to Berberis in
previous phylogenetic studies (Meacham 1980; Terabayashi
1985; Kim and Jansen 1995, 1998), was too divergent to be
used for proper alignment with the ingroup taxa.
Results
The lengths of two ITS regions for 79 taxa of Berberis
ranged from 230 bp (Berberis sieboldi) to 245 bp (B. graci-
lis) for the ITS-1 region and from 228 bp (B. grevilleiana,
B. longibracteata) to 239 bp (B. valdiviana) for the ITS-2
region, respectively. The average G + C contents were
45.8% and 55.1% for the ITS-1 and ITS-2 regions, respec-
tively. Proper alignment of the ITS sequences introduced
54 gaps and resulted in a matrix of 515 bp. The nucleotide-
sequence distance values varied from 0 to 0.0788 (between
B. sieboldi and B. ehrenbergii). Complete sequences of the
ITS regions were deposited at GenBank (accession num-
bers AY383321–AY383478), and the aligned data matrix is
available upon request to the senior author.
The result of the likelihood ratio test indicated that there
was no significant overall rate heterogeneity in the ITS
sequences (
c
2
= 81.854, P = 0.3312). Phylogenetic trees
based on the NJ and Bayesian method were almost identical
in the relationships of major lineages (Figs. 1, 2).
The NJ and Bayesian trees indicate that simple-leaved
Berberis consisted of two major groups corresponding to
Septentrionales and Australes, respectively. The group Aus-
trales is further divided into two subgroups Aequinoctiales
Ahrendt (B. lutea and B. jamesonii) and Euaustrales
Ahrendt (the other South American species). In the
group Septentrionales, two North American species, which
belong to a section Canadenses, are positioned on different
clades as B. fendleri is allied to B. koreana whereas B.
canadensis is grouped with B. vulgaris and other related
species. It was noteworthy the ITS sequence of African
species, B. holstii, was identical to that of Himalayan
species, B. coriaria and B. edgeworthiana. Sections or sub-
sections proposed by Ahrendt (1961) were not evident in
the ITS phylogenies.
Compound-leaved species segregated into three major
clades in the NJ tree. One of the clades was the group
Orientales, which includes Asian taxa and Berberis nervosa
in North America. The group Occidentales, which was
divided into two well-supported lineages, appeared as a non-
monophyletic group. The two lineages basically corre-
sponded to section Horridae and sections Aquifoliatae/
Paniculatae of Ahrendt (1961). Section Horridae appeared
as sister group of the simple-leaved Berberis. B. harrisoniana
from Arizona was nested in the Aquifoliatae clade. The Baye-
sian tree was almost identical to the NJ tree in the recognition
of major clades of the compound-leaved species except the
group Orientales in which B. nervosa was not grouped with
Asian taxa. Unlike sections, subsections proposed by
Ahrendt (1961) were not evident in both phylogenies.
Discussion
Taxonomic implications
The intrageneric taxonomic schemes of Schneider (1905)
and Ahrendt (1961) were mostly congruent with the ITS
phylogeny at higher levels. Most of the traditional geo-
graphical groups such as groups Australes, Septentrionales,
Orientales, and subgroups, Euaustrales and Aequinoctiales
in the Australes clade turned out monophyletic. The only
exception is the Occidentales of the compound-leaved Ber-
beris, which appeared paraphyletic as section Horridae was
separated from other sections and grouped with the simple-
leaved Berberis. At lower taxonomic levels in simple-
leaved Berberis, however, the ITS phylogeny did not
recognize any sections or subsections by Schneider (1905)
or Ahrendt (1961). The validity of their sectional or subsec-
tional division has been strongly questioned as three species
(Berberis microphylla, B. buxifolia, B. heterophylla) in dif-
ferent sections were joined under one name (Landrum
1999). It should also be mentioned that B. amurensis var.
quelpaertensis (section Vulgares) has been recognized as a
distinct species (B. quelpaertensis) of a different section

179
(sect. Franchetianae) in the treatment of Ahrendt (1961).
But phylogenetic affinity between B. amurensis var.
quelpaertensis and other species in section Franchetianae
was not evident in the ITS tree. Although divergence of the
ITS sequence in Berberis is somewhat too low to provide a
solid relationship at the species level, the resulting phylog-
eny and previous taxonomic treatment (Landrum 1999)
suggest that most sections or subsections in the simple-
leaved group are questionable.
In addition to the segregation of section Horridae (in the
compound-leaved group) as a distinct clade, a possible affin-
ity of the section to the simple-leaved group is noteworthy
with respect to the generic circumscription of Berberis and
Mahonia. Current taxonomic treatments (e.g., Lafferriere
1997; Whittemore 1997) merge species of compound-
leaved Berberis (Mahonia) with the simple-leaved group
(true Berberis). Although Ahrendt (1961) accepted Maho-
nia, he postulated that trueBerberis is derived from Maho-
nia. The simple leaf is probably homologous with the
terminal leaflet of a compound leaf and thus unifoliolate
(Cronquist 1981). The spines of the simple-leaved species,
which occur on long shoots, are homologous with whole
simple leaves (Whittemore 1997; Landrum 1999). Ahrendt’s
(1961) interpretation of the clusters of leaves (short shoots)
Fig. 1. Neighbor-joining tree
based on ITS sequences of
79 taxa of Berberis
representing four major groups
and 22 sections in the
genus. Numbers below
branches are bootstrap values
(only those above 50 are given)

180
in simple-leaved Berberis as being equivalent to all the leaf-
lets of a compound leaf is unlikely. Ahrendt (1961) pointed
out that species of section Horridae exhibit some morpho-
logical characters such as smoother and dark red stems,
which are common in simple-leaved Berberis, suggesting a
possible phylogenetic connection between section Horridae
and true Berberis. A closer relationship of section Horridae
to the simple-leaved group than to other sections in the
compound-leaved group causes the compound-leaved
group to be non-monophyletic. The paraphyly of the
compound-leaved group was also implied by a previous
molecular tree (Kim and Jansen 1998), which was generated
by the analysis of chloroplast DNA RFLP (restriction
fragment length polymorphism) data with a limited taxon
sampling.
Phytogeography
The distributions of 17 genera of Berberidaceae indicate
that they are predominantly Laurasian. Berberis is the only
exception. Simple-leaved Berberis has two centers of diver-
sity, one in Asia and the other in South America. The ITS
phylogeny ascertains that independent origins of the South
Fig. 2. Bayesian tree based on
ITS sequences of 79 taxa
of Berberis representing four
major groups and 22 sections in
the genus. Numbers below
branches are posterior
probabilities

181
American and the Old World (mainly Asian) groups of
simple-leaved Berberis are highly implausible. The ITS
phylogeny rather suggests a single origin of simple-
leaved Berberis and an early division into South American
and Asian groups. The sister of the simple-leaf Berberis
clade is a group of North American species of the
compound-leaved Berberis. How has simple-leaved Berb-
eris obtained the South American-Asian disjunct distribu-
tion? Thorne (1972) has noted that this is a difficult
distribution to explain. Based on matK phylogeny Soltis
et al. (2001) suggested an ancient long-distance dispersal
for the origin of South American Chrysosplenium (Saxi-
fragaceae). Schnabel and Wendel (1998) also favor migra-
tion to explain the distribution of Gleditsia (Fabaceae). The
ITS phylogeny of Berberis, however, conforms to a vicari-
ance hypothesis of origin for its current distribution. We
postulate the first vicariance event between compound- and
simple-leaved Berberis occurred as low-latitude North
America was separated from Eurasia and then a subsequent
vicariance event between Old world and South American
groups of the simple-leaved Berberis occurred as South
America was separated from Africa during the Cretaceous.
Africa would have been the best link between Asia and
South America until that time. Many major lineages of
angiosperms (including eudicots) had already diversified
during that period (Magallon and Sanderson 2001). Various
families of primitive dicots have well developed distribu-
tions in Asia, South America, and Africa (e.g., Annonaceae,
Myristicaceae, Lauraceae, Hernandiaceae, Piperaceae,
Aristolochiaceae, Menispermaceae), which indicates that
migration was probably common between those areas in the
Cretaceous. One of these, Menispermaceae, is generally
thought to be closely related to Berberidaceae (Loconte
and Estes 1989; Hoot and Crane 1995). Other families con-
sidered to be relatives of the Berberidaceae are Fumari-
aceae, Papaveraceae, Lardizabalaceae and Ranunculaceae,
most of which have Laurasian distributions, and Raven and
Axelrod (1974) suggested a Laurasian origin for the order
Berberidales. One remarkable exception noted by Raven
and Axelrod is the Lardizabalaceae with a total of eight
genera and about 30 species (Cronquist 1981), most of
which are Asian but two of which are Chilean. According
to Raven and Axelrod, “the major puzzle in distribution of
the Berberidales is the presence of two endemic genera
(Boquila and Lardizabala) of the chiefly East Asian Lardiz-
abalaceae in Chile.” To this puzzle we can now add the
simple-leaved clade of Berberis. Berberis and Lardizabal-
aceae of South America grow in areas of cool climate (tem-
perate regions or tropical mountains). These climates
existed principally in southern South America until the rel-
atively recent rise of the Andes (Landrum 1981). Landrum
(1999) suggests that high species diversity as well as leaf
fossils found in the early Tertiary of Neuquén and Santa
Cruz, Argentina (Orsi 1976) are an indication of a long
evolutionary history for South American Berberis species.
The possible connection between Asian and South
American groups through the North American species
(B. canadensis or B. fendleri) is rejected here since the ITS
tree implies Asian origins of these species via independent
long-distance dispersal or migration from the two different
Asian species. Although the origin of the South American
species of Berberis is better explained by a vicariance event
occurring during the Cretaceous period rather than by a
recent introduction from North America or the Old World,
weak bootstrap support for the Old World clade of simple-
leaf Berberis implies that the possibility of an ancient long-
distance dispersal cannot be completely ruled out. A more
robust phylogeny with an increased number of characters
and taxa is required to test these competing phytogeo-
graphic hypotheses.
Acknowledgements We thank Arnold Arboretum, Gray Herbarium,
Edinburgh Botanical Garden, University of Texas Herbarium, Drs.
J. Panero, and M. Chase for providing leaf or DNA materials, and Dr.
C. Ulloa for identifying some South American species. This work was
supported by grant (2000-2-201-00-001-2) from the Korea Science and
Engineering Foundation.
References
Ahrendt L (1961)Berberis and Mahonia, a taxonomic revision. Bot J
Linn Soc 57:1–410
Cronquist A (1981) An integrated system of classification of flowering
plants. Columbia University, New York
Doyle JJ, Doyle JL (1987) A rapid DNA isolation procedure for small
quantities of fresh leaf tissue. Phytochem Bull 19:11–15
Fedde F (1902) Versuch einer Monographie der Gattung Mahonia. Bot
Jahrb Syst 31:30–133
Felsenstein J (1985) Confidence limits on phylogenies: an approach
using bootstrap. Evolution 39:783–791
Felsenstein J (1988) Phylogenies from molecular sequences: inference
and reliability. Annu Rev Genet 22:521–565
Hoot SB, Crane PR (1995) Inter-familial relationships in the
Ranunculidae based on molecular systematics. Plant Syst Evol
9[Suppl]:119–131
Huelsenbeck JP, Ronquist FR (2001) MrBayes: Bayesian inference of
phylogenetic trees. Bioinformatics 17:754–755
Judd WS, Campbell CS, Kellog EA, Stevens PF, Donoghue MJ
(2002) Plant systematics: a phylogenetic approach. Sinauer,
Sunderland
Kim K-J, Jansen RK (1994) Comparison of phylogenetic hypotheses
among different data sets in dwarf dandelions (Krigia): additional
information from internal transcribed spacer sequences of nuclear
ribosomal DNA. Plant Syst Evol 190:157–185
Kim Y-D, Jansen RK (1995) Phylogenetic implication of the chloro-
plast DNA variation from the Berberidaceae. Plant Syst Evol
9[Suppl]:341–349
Kim Y-D, Jansen RK (1998) Chloroplast DNA restriction site variation
and phylogeny of the Berberidaceae. Am J Bot 85:1766–1778
Lafferriere JE (1997) Transfer of specific and infraspecific taxa
from Mahonia to Berberis (Berberidaceae). Bot Zhurn 82:95–99
Landrum L (1981) The phylogeny and geography of Myrceugenia
(Myrtaceae). Brittonia 33:105–129
Landrum L (1999) Revision of Berberis (Berberidaceae) in Chile and
adjacent southern Argentina. Ann Mo Bot Gard 86:793–834
Loconte H, Estes JR (1989) Phylogenetic systematics of Berberidaceae
and Ranunculales (Magnoliidae). Syst Bot 14:565–579
Loockerman DJ, Jansen RK (1996) The use of herbarium material for
DNA studies. In: Stussey TF, Sohmer S (eds) Sampling the green
world. Columbia University, New York, pp 205–220
Magallon S, Sanderson MJ (2001) Absolute diversification rates in
angiosperm clades. Evolution 55:1762–1780
Meacham CA (1980) Phylogeny of the Berberidaceae with an evalua-
tion of classifications. Syst Bot 5:149–172
Orsi MC (1976) Sinopsis de las especies Argentinas del género Berberis
(Berberidaceae). Bol Soc Argent Bot 17:127–149

182
Rannala B, Yang Z (1996) Probability distribution of molecular evolu-
tionary trees: a new method of phylogenetic inference. J Mol Evol
43:304–311
Raven PH, Axelrod DI (1974) Angiosperm biogeography and past
continental movements. Ann Mo Bot Gard 61:539–673
Saitou N, Nei M (1987) The neighbor-joining method: a new method
for reconstructing phylogenetic trees. Mol Biol Evol 4:406–425
Schnabel A, Wendel JF (1998) Cladistic biogeography of Gleditsia
(Leguminosae) based on ndhF andrpl16 chloroplast gene sequences.
Am J Bot 85:1753–1765
Schneider CK (1905) Die Gattung Berberis (Euberberis). Vorarbeiten
für eine Monographie. Bull Herb Boissier Ser 2 5:33–48, 133–148,
391–403, 449–464, 655–670, 800–812, 813–831
Soltis DE, Tago-Nakazawa M, Xiang Q-Y, Kawano S, Murata J,
Wakabayashi M, Hibsch-Jetter C (2001) Phylogenetic relationships
and evolution in Chrysosplenium (Saxifragaceae) based on matK
sequence data. Am J Bot 88:883–893
Suh Y, Thien LB, Reeve HE, Zimmer EA (1993) Molecular evolution
and phylogenetic implications of internal transcribed spacer
sequences of ribosomal DNA in Winteraceae. Am J Bot 80:1042–
1055
Swofford DL (2002) PAUP*, phylogenetic analysis using parsimony
(*and other methods), version 4.0b. Sinauer, Sunderland
Swofford DL, Olsen GJ, Waddell PJ, Hillis DM (1996) Phylogenetic
inference. In: Hillis DM, Moritz C, Mable BK (eds) Molecular
systematics, 2nd edn. Sinauer, Sunderland, pp 407–514
Terabayashi S (1985) The comparative floral anatomy and systematics
of the Berberidaceae. II. Systematic considerations. Acta Phytotax
Geobot 36:1–13
Thompson JD, Higgins DG, Gibson TJ (1995) CLUSTAL W: improv-
ing the sensitivity of progressive multiple sequence alignment
through sequence weighting, positions-specific gap penalties and
weight matrix choice. Nucleic Acids Res 22:4673–4680
Thorne RF (1972) Major disjunctions in the geographic ranges of seed
plants. Q Rev Biol 47:365–411
White TJ, Birns T, Lee S, Taylor J (1990) Amplification and direct
sequencing of fungal ribosomal RNA genes for phylogenetics. In:
Innis M, Gelfand D, Sninsky J, White T (eds) PCR protocols: a guide
to methods and applications. Academic, San Diego, pp 315–322
Whittemore AT (1997)Berberis. In: Morin NR (ed) Flora of North
America, vol 3. Oxford University Press, New York, pp 276–286
Yang Z (1994) Estimating the pattern of nucleotide substitution. J Mol
Evol 39:105–111