全 文 ：植 物 分 类 学 报 44 (3): 272–285(2006) doi:10.1360/aps040150
Acta Phytotaxonomica Sinica http://www.plantsystematics.com
———————————
Received: 25 November 2004 Accepted: 2 March 2006
Supported by the National Natural Science Foundation of China, Grant No. 30470123, the Natural Science Foundation of
Yunnan Province, Grant No. 2001C0008R, the Research Fund for Returned Overseas Chinese Scholars, Grant No.
20010713093959, and the Scientific Research Foundation for the Returned Overseas Chinese Scholars, State Education Ministry,
Grant No. 2001-498.
* Author for correspondence. E-mail: jieli@xtbg.ac.cn; Tel: 86-871-5144431; Fax: 86-871-5160916.
Polyphyly of the genus Actinodaphne (Lauraceae) inferred
from the analyses of nrDNA ITS and ETS sequences
1, 2LI Zhi-Ming 1LI Jie* 1LI Xi-Wen
1(Laboratory of Plant Phylogenetics and Conservation Biology, Xishuangbanna Tropical Botanical Garden,
the Chinese Academy of Sciences, Kunming 650223, China)
2(Graduate School of the Chinese Academy of Sciences, Beijing 100039, China)
Abstract A phylogenetic analysis of the genus Actinodaphne (Lauraceae) was conducted
using sequences from the nrDNA (nuclear ribosomal DNA) internal transcribed spacer (ITS)
and external transcribed spacer (ETS) regions. Maximum parsimony (MP), maximum
likelihood (ML) and Bayesian phylogenetic analysis methods were employed to analyse the
data sets (ITS, ETS and ITS/ETS). All analyses suggested that the sampled Actinodaphne
species were not monophyletic, clustering instead as several clades amongst other genera in
the tribe Laureae. This result indicates that the generic delimitations between Actinodaphne
and related genera need to be reevaluated, and that inflorescence features, which are mostly
consistent with the current molecular inferences, might be the both important and reliable
characters for their redefinition. However, because of the relatively small number of taxa
sampled, and conflicts between the separate analyses, more detailed studies are required to
clarify the relationships which emerged in our study and to allow for more precise generic
delimitation and hypotheses about phylogeny in Actinodaphne.
Key words Actinodaphne, ETS, ITS, Lauraceae, phylogenetic analysis, Bayesian analysis.
The genus Actinodaphne Nees (Lauraceae) with about 100 species occurs mainly in
tropical-subtropical Asia and is an important component of tropical forests (Rohwer, 1993;
van der Werff, 2001). In China 19 species are distributed from the southwest to east, with four
species in Taiwan (Li et al., 1984). This genus is distinguished from other Lauraceae by its
whorled, usually penninerved, rarely sub-triplinerved leaves; (pseudo-) racemose or
pseudo-umbellate inflorescences; and imbricate, deciduous involucral bracts at the base of the
inflorescence (Nees, 1836; Rohwer, 1993). Many species are used locally for wood or
medicine, i.e. the widely used wood of A. nantoensis (Hay.) Hay. and A. mushanensis (Hay.)
Hay. for architecture and furniture, and the important medical properties of the roots of A.
cupularis (Hemsl.) Gamble and leaves of A. pilosa (Lour.) Merr. (Li et al., 1984).
The systematic position of this genus had been in dispute. Nees (1836) and Allen (1938)
placed Actinodaphne into the tribe Laureae based on its introrse anther cells. Because of its
lack of decussate and persistent involucral bracts, however, Kostermans (1957) suggested that
Actinodaphne was closely related to Ocotea Aubl., Cinnamomum Trew and Sassafras J. Presl,
and therefore placed it into the tribe Cinnamomineae. Recently, Li et al. (1984), Rohwer
(1993) and van der Werff (1991, 2001) argued that Actinodaphne should be returned to the
Laureae, mainly based on the reexamination of its inflorescence involucre, and this was
No. 3 LI Zhi-Ming et al.: Polyphyly of the genus Actinodaphne 273
further supported by recent molecular phylogenetic studies (Rohwer, 2000; Chanderbali et al.,
2001). Using data from the chloroplast gene matK and nuclear ribosomal DNA sequences, Li
et al. (2004) similarly observed close relationships between Actinodaphne, Litsea Lam.,
Lindera Thunb. and Neolitsea Merr., further strengthening its placement in the Laureae.
Despite the agreement on the systematic position of Actinodaphne in Laureae, the
circumscription of this genus remains unresolved. In the phylogenetic analyses of the “core”
Laureae, Li et al. (2004) suggested that Actinodaphne might be separated into two groups
with different origins, and this was also supported by previous morphological analyses (Li &
Christophel, 2000). Van der Werff (2001) also noticed that there were different inflorescence
forms in Actinodaphne and it was difficult to distinguish some Actinodaphne species from
Litsea ones with similar inflorescences. In light of these studies, it appears that Actinodaphne
might be polyphyletic, although all the authors agree that many more samples and characters
are required before a well-resolved phylogeny and classification might be produced.
The nrDNA internal transcribed spacers (ITS) have become widely used as a source of
characters for phylogenetic studies of closely related plant species (e.g. Baldwin et al., 1995;
Ge et al., 1997; Hong et al., 2001; Gao et al., 2003; Roalson & Friar, 2004), and this DNA
fragment was demonstrated to be powerful in the recent Lauraceae phylogenetic constructions
(Chanderbali et al., 2001; Li et al., 2004). In addition, the external transcribed spacer (ETS)
has shown potential for phylogenetic studies of angiosperms as it often shows greater
variation than ITS and should be considered a good candidate when the ITS lacks sufficient
phylogenetic signal (Baldwin & Markos, 1998; Bena et al., 1998), although exceptions exist
in some genera (Soltis & Soltis, 1998).
Combined analyses of sequences from these two nuclear DNA fragments could be
expected to produce a more robust phylogeny and the goals of our study using maximum
parsimony (MP), maximum likelihood (ML), and Bayesian analysis of sequence data from
ITS and ETS are to: (1) test whether Actinodaphne is monophyletic; (2) examine the
relationships of Actinodaphne within the tribe Laureae; and (3) evaluate the species
relationships within this genus against previous morphological investigations.
1 Material and methods
1.1 Plant material
A total of 13 Actinodaphne species were included in this study, 11 from China, plus one
each from Malaysia and Singapore. Because Actinodaphne might be polyphyletic (Li et al.,
1984; Rohwer, 1993, 2000; van der Werff & Richter, 1996; Chanderbali et al., 2001; Li et al.,
2004), five representative genera from the tribe Laureae were selected as outgroups in our
analysis. A complete list of the species sampled, along with collection and voucher
information is provided in Table 1.
1.2 DNA extraction, PCR and sequencing
Total DNA was extracted using a modified CTAB protocol (Doyle & Doyle, 1987) and
performed as described by Li et al. (2004). DNA was cleaned using QIAquick® PCR
purification Kit (Qiagen, Germany), and DNA concentration was determined by visual
comparison with a positive control (λ100 ladder, concentration 10, 20 ng) on an ethidium
bromide-stained agarose gel.
Templates of the whole nrDNA ITS region were amplified successfully in most cases
using primer pair of ITSF/ITS4 (White et al., 1990; Chanderbali et al., 2001), but if these
failed, especially for poor-quality DNA, the primer combinations of ITSF/ITS2 and
ITS3/ITS4 (White et al., 1990; Chanderbali et al., 2001) were used to amplify ITS1 and ITS2
(including the 5.8S nrDNA region) separately. For most individuals, PCR amplification
Acta Phytotaxonomica Sinica Vol. 44 274
Table 1 Species of the genus Actinodaphne and outgroup taxa included in the analysis
Species Locality Voucher GenBank
number
for ITS
GenBank
number
for ETS
Ingroups
Actinodaphne kweichowensis
Yang & P. H. Huang
黔桂黄肉楠

Dongshan, Guangxi (广西东
山), China

H. Q. Li (黎焕奇)
40091 (KUN 0106643)

AY817114

AY817124
A. omeiensis (Liou) Allen
峨眉黄肉楠
Mt. Emeishan, Sichuan (四川
峨眉山), China
G. H. Yang (杨光辉)
55824 (KUN 0047252)
AY817117 AY817127
A. pilosa (Lour.) Merr.
毛黄肉楠
Yongning, Guangxi (广西邕
宁), China
L. S. Xie (谢立山)
613 (KUN 0047277)
AY817115 AY817125
A. trichocarpa Allen
毛果黄肉楠
Daguan,Yunnan (云南大关),
China
B. S. Sun (孙必兴)
0757 (KUN 0047286)
AY817116 AY817126
A. tsaii Hu
马关黄肉楠
Malipo,Yunnan (云南麻栗
坡), China
K. M. Feng (冯国楣)
22638 (KUN 0047322)
AY817119 AY817129
A. cupularis (Hemsl.) Gamble
红果黄肉楠
Shidian, Guizhou (贵州施甸),
China
Wuyishan Exped. (武夷
山考察队) 1693 (KUN
0601976)
AY817113 AY817123
A. henryi Gamble
思茅黄肉楠
Mengla, Yunnan (云南勐腊),
China
J. Li (李捷)
2002032 (HITBC)
AY817120 AY817130
A. obovata (Nees) Bl.
倒卵叶黄肉楠
Mengla, Yunnan (云南勐腊),
China
H. W. Li (李锡文)
1 (HITBC)
AY265398 AY934880
A. forrestii (Allen) Kosterm.
毛尖树黄肉楠
Mengla, Yunnan (云南勐腊),
China
H. W. Li (李锡文)
2 (HITBC)
AY265399 AY934881
A. paotingensis Yang & P. H.
Huang
保亭黄肉楠
Baoting, Hannan (海南保亭),
China
Hainan East Exped. (海
南东路队) 962 (IBK
00003425)
AY817118 AY817128
A. lecomtei Allen
柳叶黄肉楠
Without precise locality,
Guangxi (广西, 具体地点不
详), China
C. Q. Li (李彩祺)
3979 (IBK 00003410)
AY817112 AY817122
Actinodaphne sp. Botanical Garden, Singapore SING AY817112 AY817122
A. sesquipedalis Hook. f. &
Thoms. ex Hook. f.
Kuala Lumpur, Malaysia Saw Leng Guan s.n.
(KEP)
AF272247 –
Outgroups
Lindera megaphylla Hemsl. Mengla, Yunnan (云南勐腊),
China
H. W. Li (李锡文)
7 (HITBC)
AY265406 AY934882
Litsea glutinosa (Lour.) C. B.
Rob.
Mengla, Yunnan (云南勐腊),
China
H. W. Li (李锡文)
21 (HITBC)
AY265403 AY934883
Neolitsea levinei Merr. Mengla, Yunnan (云南勐腊),
China
H. W. Li (李锡文)
29 (HITBC)
AY265401 AY934884
Parasassafras confertiflora
(Meisner) D. G. Long
Ximeng, Yunnan (云南西盟),
China
Y. Y. Qian (钱义咏)
921 (KUN 0104560)
AY265395 AY934885
Sinosassafras flavinervia
(Allen) H.W. Li
Mt. Ailaoshan, Yunnan (云南
哀牢山), China
Y. H. Liu (刘玉洪) s.n. AY265394 AY934886


resulted in a single band, but for some specimens, two size classes of PCR products were
obtained. To isolate each of the two bands, the PCR products were blunt-end ligated into the
EcoRV sites of the PMD 18-T Simple Vector after purification, using the Original TA®
Cloning Kit (available from TaKaRa Biotechnology) and the fragments obtained following
the manufacturer’s protocol.
Fragments of the nrDNA ETS region were amplified using the primer pair
ETS1/18S-IGS (Li et al., unpublished; Baldwin & Markos, 1998), with the ETS1 primer
No. 3 LI Zhi-Ming et al.: Polyphyly of the genus Actinodaphne 275
designed by Li et al. following the method of Baldwin and Markos (1998). The primer
sequences used are shown in Table 2. In order to reduce the likelihood of selective
amplification of pseudogenes or paralogous ITS copies, 10% DMSO was included in all
amplifications (Buckler & Holtsford, 1996; Buckler et al., 1997) and a negative control was
used for every primer combination to detect PCR contamination.

Table 2 Primers used for amplification and sequencing ITS and ETS in the Actinodaphne
Primers (ITS, ETS) Sequence 5′–3′ Source
Forward
ITSF
ITS3

GCTACGTTCTTCATCGATGC
GCATCGATGAAGAACGTAGC

Chanderbali et al., 2001
White et al., 1990
Reverse
ITS2
ITS4

GCTACGTTCTTCATCGATGC
TCCTCCGCTTATTGATATGC

White et al., 1990
White et al., 1990
Forward
ETS1

CCAGAACTCGCACTTGCTGAGCTT

Li et al., unpublished
Reverse
18S-IGS

GAGACAAGCATATGACTACTGGCAGGATCAACCAG

Baldwin & Markos, 1998


The PCR products were purified using the Qiagen QIAquick® PCR Purification Kit
following the manufacturer’s protocols. To sequence the ITS nrDNA fragments, a series of
reactions were run, each with one different internal primer (Table 2), thus creating
overlapping fragment sequences that between them covered the entire spacer and 5.8S nrDNA
regions along both strands. Sequencing of the nrDNA ETS region using the primer 18S-IGS
obtained the fragment that included the 3′ ETS region. Cycle sequencing was carried out
directly on the purified PCR product using the ABI Prism Big Dye® Terminator Cycle
Sequencing Ready Reaction Kit (Applied Biosystems, USA), using 1 µL of primer, 10 ng of
DNA template, 1.5 µL of Big Dye (version 3.1), and then using ddH2O to make up a final
reaction volume of 5 µL. Cycle sequencing reactions were as follows: (1) 30 s denaturation
(96 ˚C), (2) 15 s annealing (50 ˚C), and 4 min elongation (60 ˚C) with 25 cycles. Cleaned
products were then sequenced directly in an Applied Biosystems 3100 DNA automated
sequencer.
1.3 Sequence alignment
Sequences were aligned individually using the software program Lasergene/Megalign
(DNASTAR, 1998), allowing uncertainties either to be resolved or recorded as ambiguities.
The boundary of ITS and 3′ ETS sequences was determined through comparison with other
Lauraceae species from GenBank and all the obtained ETS sequences respectively. The
sequences were then truncated to include ITS and 3′ ETS of the nrDNA gene. The consensus
sequences for the analyzed taxa were then realigned using ClustalX (Thompson et al., 1997)
and modified manually, if necessary, using BioEdit version 5.0.6. (Hall, 1999).
1.4 Phylogenetic analysis
The aligned submatrices were analyzed both individually and together. Only
phylogenetic informative characters were analyzed, and gaps were scored as missing data.
Because the two sequenced regions (ITS, ETS) used in this study are part of a tandem repeat
within the diploid nuclear genome, possible conflict between the data sets was evaluated with
an incongruence length difference (ILD) test (Farris et al., 1994, 1995) prior to combining the
data. This test, implemented as the partition homogeneity test in PAUP* version 4.0b10
(Swofford, 1998), determines whether the original data partitions differ significantly from
randomly shuffled partitions of the combined data set.
Both MP and ML analyses were performed using PAUP* version 4.0b10 (Swofford,
Acta Phytotaxonomica Sinica Vol. 44 276
1998). Heuristic searches were employed (ACCTRAN, 1000 random addition cycles, TBR
branch swapping, STEEPEST DESCENT, MULTREES in effect). Clade support was
estimated using 1000 heuristic bootstrap replicates with 100 random addition cycles per
replicate, 20 trees saved from each addition cycle, TBR branch swapping, and STEEPEST
DESCENT options (Felsenstein, 1985; Hills & Bull, 1993; Li et al., 2004; Roalson & Friar,
2004).
The Tamura and Nei (1993) model of evolution with rate heterogeneity and among-site
rate variation was used in the ML analysis based on the result of Modeltest 3.06 (Posada &
Crandall, 1998). The modeltest analysis tested the fit of various ML models to the data set and
estimated base change frequencies, proportion of variable characters, shape of the gamma
distribution, and chose the model that best fitted the data using the Hierarchical Likelihood
Ratio Tests (Posada & Crandall, 1998; Roalson & Friar, 2004). The parameters assigned to
the data set for this analysis are shown in Appendix A.
Bayesian phylogenetic analyses were performed using MrBayes version 3.0b4 (Hall,
2001; Huelsenbeck & Ronquist, 2001). The same model and number of base change
frequencies used in the ML search were used, with Bayesian analysis started from a random
tree. 106 generations were computed for four parallel chains applying six possible substitution
types and gamma-distribution of substitution rates and the Markov chains were sampled at
intervals of 100 generations, resulting in a final set of 10001 sample points. Plotting
likelihood values for the four analyses shows that “stationary” was achieved for each
sequence analysis as follows: (1) ITS: 116; (2) ETS: 905; (3) ITS/ETS: 123. These were
therefore discarded as “burn-in”, with the remaining trees presented as 50% majority rule
consensus trees, and the percentage of sample points recovered any particular clade
represented its posterior probability (Huelsenbeck & Ronquist, 2001).
2 Results
2.1 Sequence characteristics
The length of the unaligned ITS sequences covering the entire spacer and 5.8S nrDNA
regions varied from 605 to 627 bp. The aligned ITS data matrix was 668 bp in length, of
which 64 characters (9.6%) were informative. The percentage of G+C in the Actinodaphne
ITS sequences varied from 61.6% to 67.2%. The whole nrDNA ITS region of three species
(A. lecomtei, Actinodaphne sp. and A. paotingensis) could not be amplified successfully using
primer pair of ITSF/ITS4 (White et al., 1990; Chanderbali et al., 2001), so their ITS
sequences were acquired through separate amplification of the ITS1, 5.8S nrDNA and ITS2
regions.
Fragments of the 3′ ETS region for all species were amplified successfully using the
ETS1/18S-IGS primer pair (Li et al., unpublished; Baldwin & Markos, 1998). The ETS
sequencing primer (18S-IGS) produced a fragment that covered the whole 3′ ETS region and
in the Actinodaphne species sampled, this varied from 392–398 bp. The resulting aligned ETS
matrix was 393 bp long, of which 32 characters (8.1%) were informative. The percentage of
G+C in Actinodaphne ETS sequences was more invariable than for ITS, ranging from 50.8%
to 52.3%. The 3′ ETS sequence for A. sesquipedalis was unavailable and so was excluded in
the ETS submatrix, but treated as missing in the combined analyses.
2.2 ITS analysis
MP analysis of the ITS submatrix resulted in 545 most parsimonious trees (tree
length=161, CI=0.4845, RI=0.5389), and the 50% majority rule consensus tree is shown in
Fig. 1. ML analysis yielded three trees (–lnLikelihood=734.32927), with the resulting
Bayesian analysis cladogram presented as a 50% majority consensus tree excluding the 116
No. 3 LI Zhi-Ming et al.: Polyphyly of the genus Actinodaphne 277
burn-in trees, and the posterior probabilities from that analysis were mapped along with the
MP bootstrap support percentages onto the MP consensus tree (Fig. 1).

Fig. 1. 50% majority rule consensus cladogram of 545 most parsimonious trees (tree length=161 steps,
CI=0.4845, RI=0.5389) derived from an analysis of ITS sequence data. Bayesian posterior probability
values greater than 50% and bootstrap values greater than 50% are indicated above and below branches
respectively.

All three analyses (MP, ML and Bayesian) indicate that Actinodaphne was polyphyletic
and this is congruent with former studies (Chanderbali et al., 2001; Li et al., 2004). The MP
analysis cladogram (Fig. 1) indicates that A. forrestii was basal above the outgroup of Lindera
megaphylla. Parasassafras confertiflora was then sister to the remainder with strong
bootstrap (100%) and posterior probability support (100%), but above this was a polytomy of
three small clades (Litsea glutinosa with Sinosassafras flavinervia; A. omeiensis, A.
trichocarpa, A. pilosa and A. kweichowensis with A. cupularis; and Actinodaphne sp. with
Acta Phytotaxonomica Sinica Vol. 44 278
Neolitsea levinei), plus the remainder of Actinodaphne as a fourth clade.
The cladograms for the Bayesian (not shown) and MP analyses were very similar;
however, a major incongruence appeared between these and the ML analysis. In the ML
cladogram, one clade placed above Actinodaphne forrestii was sister to the remainder of the
sampled taxa. This clade contained two sister-taxon pairs: Parasassafras confertiflora and A.
lecomtei; Litsea glutinosa and Sinosassafras flavinervia. Above this clade, A. tsaii was basal
to the remainder, within which A. obovata and A. henryi formed a sister pair in an unresolved
polytomy with two other clades (A. omeiensis, A. trichocarpa, A. pilosa, A. kweichowensis
and A. cupularis versus A. paotingensis, A. sesquipedalis, Actinodaphne sp. and Neolitsea
levinei).
2.3 ETS analysis
MP analysis of the ETS data resulted in 132 most parsimonious trees (length=70 steps,
CI=0.5143, RI=0.6634) and the 50% majority rule consensus tree is shown in Fig. 2. ML
analysis yielded only one tree and the “burn-in” for Bayesian analysis was the first 905 trees.
The resulting MP cladogram indicates that Lindera megaphylla, Actinodaphne forrestii,
Actinodaphne sp. were basal and successive sisters to the remainder. Above them were a
terminal pair (Parasassafras confertiflora and Sinosassafras flavinervia) and then Litsea
glutinosa was sister to the remainder, which divided into two major clades. The first clade
consisted of A. trichocarpa, A. omeiensis, A. tsaii, A. cupularis, Neolitsea levinei, A. obovata,
A. henryi and A. pilosa, of which the first five formed a subclade (94% posterior probability),
sister to the other three species (77% posterior probability). The second major clade
comprised A. kweichowensis, A. paotingensis and A. lecomtei, in which the latter two formed
a terminal pair (50% bootstrap, 54% posterior probability) sister to the former (63% bootstrap
support, 55% posterior probability).
Although the relationships described above were not as well resolved in the Bayesian
cladogram, they were relatively congruent with the MP analysis. However, the ML cladogram
was once again highly incongruent with the MP and Bayesian results as follows: In the ML
analysis, A. paotingensis, A. lecomtei and A. kweichowensis did not cluster into one clade,
instead forming successive sister relationships above Lindera megaphylla and A. forrestii.
Litsea glutinosa, Parasassafras confertiflora and Sinosassafras flavinervia were clustered
into a clade in which the latter two formed a terminal pair and this clade was then sister to the
remainder of the tree.
2.4 ITS/ETS analysis
The Incongruence Length Difference Test (ILD) indicates that ITS and ETS data sets
were relatively incongruent in the estimates of phylogeny (P=0.01). However, these two
spacer regions were analyzed simultaneously in order to address consistency problems. The
combined data matrix contained 18 operational taxonomy units (OTUs) and 1061 characters,
of which 96 were informative (9.05%). The percentage G+C content for the combined
ITS/ETS Actinodaphne data ranged from 58.1% to 61.4%.
Despite the apparent sequence incongruity, the combined ITS/ETS analyses removed the
polytomies present in the separate analyses of ITS and ETS, and generally resulted in more
robust cladograms and support the hypothesis that Actinodaphne was polyphyletic. MP
analysis yielded three most parsimonious trees (tree length=445 steps, CI=0.4635, RI=0.4931)
and ML analysis yielded only one tree (–lnLikelihood=1192.8063). The Bayesian analysis
excluded the first 123 burn-in trees and the posterior probabilities from that analysis along
with the MP bootstrap support percentages presented on a MP 50% majority consensus tree
(Fig. 3).


No. 3 LI Zhi-Ming et al.: Polyphyly of the genus Actinodaphne 279

Fig. 2. 50% majority rule consensus cladogram of 132 most parsimonious trees (tree length=70 steps,
CI=0.5143, RI=0.6634) derived from an analysis of ETS sequence data. Bayesian posterior probability
values greater than 50% and bootstrap values greater than 50% are indicated above and below branches
respectively.


All analyses (MP, ML and Bayesian) indicate that Actinodaphne was polyphyletic. The
resulting cladograms of Bayesian and MP analyses were relatively congruent, but the former
was slightly more resolved and gave relatively high posterior probability values (Fig. 3). The
Bayesian cladogram successively placed Lindera megaphylla, A. forrestii, Actinodaphne sp.,
Parasassafras confertiflora, Sinosassafras flavinervia and Litsea glutinosa in a basal grade,
above which was a polytomy (84% posterior probability) of A. sesquipedalis and the
remainder of the sampled species, divided into two clades. Clade A consisted of A. obovata
and A. henryi (92% posterior probability), whereas clade B contained Neolitsea levinei, A.
tsaii, A. trichocarpa, A. omeiensis, A. cupularis, A. pilosa, A. kweichowensis, A. paotingensis
Acta Phytotaxonomica Sinica Vol. 44 280
and A. lecomtei, in which the first two species together with a subclade formed an unresolved
polytomy (57% posterior probability).
The ML combined analysis cladogram was, however, still incongruent with those of the
Bayesian and MP analyses, with Actinodaphne sp. no longer in the basal clade, instead
forming a subclade with Actinodaphne lecomtei and A. paotingensis, as sister to a subclade
which was similar to the Bayesian clade B, only without A. lecomtei and A. paotingensis.



Fig. 3. Bayesian consensus of 9977 trees, and 50% majority rule consensus cladogram of the most
parsimonious trees (tree length=445 steps, CI=0.4635, RI=0.4931) derived from a combined analysis of
ITS/ETS sequence data. Bayesian posterior probability values greater than 50% and bootstrap values
greater than 50% are indicated above and below branches respectively.

No. 3 LI Zhi-Ming et al.: Polyphyly of the genus Actinodaphne 281
3 Discussion
3.1 Data compatibility
One of the outstanding issues in systematics is how to test phylogenetic conflict between
different data sets (Kluge, 1989; Bull et al., 1993; de Queiroz et al., 1995; Miyamoto & Fitch,
1995; Cunningham, 1997), and two alternative approaches have generally been suggested,
i.e., separate analysis and combined analysis (Kluge, 1989; Barrett et al., 1991; Lanyon, 1993;
Miyamoto & Fitch, 1995). In our study the incompatibility of ITS and ETS was high for the
ILD analysis (P=0.01). Nevertheless, it is deemed more reasonable to combine the data sets in
a simultaneous analysis for the following reasons. Firstly, nrDNA ITS and ETS regions occur
within the same transcriptional unit and there is evidence indicating a similar and
interdependent role in the maturation of rRNAs (e.g., Good et al., 1997). Secondly, even with
significant incompatibility between different DNA data sets (P<0.01 or even <0.001), a
combined analysis can still produce a more satisfactory estimate of phylogenetic relationships
than separate analysis of each data set alone (Cunningham, 1997). Thirdly, the incongruent
clades in our analyses received little or no bootstrap support or posterior probabilities in the
consensus trees for the separate ITS and ETS analyses. In the phylogenetic analysis of the
“core” Laureae based on matK and ITS sequences, a similar issue emerged but the combined
analysis provided a more resolved phylogeny (Li et al., 2004).
Because of this improved resolution effect, conflicting ITS and ETS data are generally
still deemed to be combinable (e.g. Mason-Game & Kellogg, 1996; Elodenäs & Linder, 2000;
Barker et al., 2003; Li et al., 2004), and combined data sets generally produce more robust
phylogenies (Chase et al., 1997). Nevertheless, caution is recommended, and conflict between
data sets should be taken into consideration (Li et al., 2004).
3.2 Selecting Bayesian inference
Bayesian inference is a relatively recent addition to the analytical toolbox for
phylogenetics (Hall, 2001), but Steane et al. (2003) considered that Bayesian analysis has the
following advantages over other phylogenetic analysis methods. Firstly, Bayesian estimation
is based on the likelihood function and thus related to maximum likelihood analysis; however,
it resolves problems such as long branch attraction. Secondly, Bayesian analysis requires
fewer computational resources, so that large data sets could be analyzed more readily.
Thirdly, because the estimation of branch support accompanies tree estimation, additional
bootstrap analyses are not required. Finally, it provides a practical alternative, with the
resulting phylogeny providing additional support for the major clades identified by maximum
parsimony analysis. Therefore, by comparing the resulting trees, we regarded the Bayesian
ITS/ETS tree (Fig. 3) as being the most resolved and well-supported reconstruction of the
underlying phylogeny.
3.3 Polyphyly of Actinodaphne
Our results agree with those or earlier studies, both morphological and molecular, that
Actinodaphne is a member of the Laureae (Nees, 1836; Allen, 1938; Li et al., 1984; Rohwer,
1993, 2000; van der Werff, 1991, 2001; Chanderbali et al., 2001; Li et al., 2004).
Nevertheless, although the genus was traditionally delimitated by whorled, penninerved
leaves; pseudo-racemose or pseudo-umbellate inflorescences and imbricate, deciduous
involucral bracts at the base of the inflorescence (Nees, 1836; Rohwer, 1993), these
characters, especially the inflorescence types, also occur in other Laureae. For example the
inflorescences of Parasassafras D. G. Long are similar to those of Actinodaphne species with
umbellate inflorescences, both of which originate from scales underneath a vegetative bud
that would later sprout (Rohwer, 1993). Similarly, van der Werff (2001) suggested that there
is no clear difference between the inflorescence architecture of Actinodaphne and that of some
Acta Phytotaxonomica Sinica Vol. 44 282
Litsea species, noting that the generic delimitation between them is unsatisfactory. Based on
the analyses of cpDNA matK and nrDNA ITS sequences, Li et al. (2004) suggested that
Actinodaphne might be polyphyletic, and this is further supported by ongoing leaf
micromorphology studies (Li & Christophel, unpublished).
The present ITS and ETS analyses also show polyphyly in Actinodaphne, with the
sampled species variously clustering into clades with other Laureae genera. Furthermore,
these admittedly preliminary clades recovered within Actinodaphne and related Laureae are
partly consistent with inflorescence structure (van der Werff & Richter, 1996; van der Werff,
2001; Li et al., 2004). The basal clade in the combinative analysis (Fig. 3) consisting of A.
forrestii, Actinodaphne sp., Lindera megaphylla, Parasassafras confertiflora and
Sinosassafras flavinervia, is characterized by clustered or fasciculate pseudo-umbels with the
usually vegetative terminal bud in the main axis, and the vegetative bud sprouting later forms
a leafy short-shoot. Such pseudo-umbels may result from the shortening of the brachy-blast,
and the peduncles of these pseudo-umbels shorten sequentially, ultimately resulting in the
clustered or fasciculate pseudo-umbels (Tsui, 1987; Li et al., 2004).
The species of clade A (Actinodaphne sesquipedalis and Litsea glutinosa) share
pseudo-racemose inflorescences, but lack development at the terminal tip. This inflorescence
type might also originate from shortening of brachy-blast internodes and distal internodes;
however, the peduncles of the pseudo-umbels do not shorten, making the pseudo-umbels
appear pseudo-racemose (Rohwer, 1993; Li et al., 2004).
The species of clade B (except for Actinodaphne pilosa) also have similar
pseudo-umbels to those of the basal clade, but the inflorescences of these Clade B species
usually fail to develop at the terminal bud, a feature also seen in Neolitsea levinei which
nested within this clade. In contrast, A. pilosa has a thyrsoid inflorescence more like the
members of clade A. This might indicate that the inflorescence type in clade B might have
originated from a thyrsoid cymose inflorescence, which is congruent with former studies
(Rohwer, 1993; van der Werff, 2001; Li et al., 2004). The evolution of this inflorescence type
would then be as follows: every secondary peduncle of the thyrsoid cyme could form a
pseudo-umbel, and then by shortening of the peduncle form this clustered or fasciculate
pseudo-umbellate inflorescence type. Thus, although the clade B inflorescences appear
identical to those of the basal grade taxa, they may have different origins. However, because
this is largely speculative, further study is required to determine why A. pilosa occurred in the
mainly pseudo-umbellate clade B, and whether this represents a sample-size artifact or
evidence of convergence in inflorescence types.
Accordingly, although the clades seen here might explain the phylogeny of the genus
Actinodaphne, whether the Bayesian tree using two combined data sets accurately reflects the
phylogeny of Actinodaphne and related genera remains to be tested in the future. All these
conclusions should be tentative because of the following reasons: Firstly, topological conflicts
were obvious between separate analyses of ITS and ETS data, and most of the major clades
received relatively poor bootstrap support, even in the combined analyses of the two data sets.
Secondly, because of relatively limited taxon sampling, more detailed studies are needed to
clarify relationships within the genus and to allow for more precise taxon boundary
definitions and hypotheses about phylogeny in Actinodaphne. Nevertheless, the available data
in our study have further confirmed earlier suggestions that Actinodaphne is polyphyletic, and
suggest that inflorescence type and ontogeny might be among the more reliable
morphology-based characters for examining phylogenetic signal in Actinodaphne and
possibly the Laureae in general.
Acknowledgements We thank curators of the following herbaria for providing materials,
No. 3 LI Zhi-Ming et al.: Polyphyly of the genus Actinodaphne 283
IBK (Guangxi Institute of Botany), KUN (Kunming Institute of Botany, CAS), and SING
(Singapore Botanical Garden); and Li Q-M, Li L, Xia Y-M, Yu D-Q (XTBG, CAS) for their
advice and help in experimentations and data process, and Fan X-N (Kunming Institute of
Zoology, CAS) for assistance with sequencing. John Conran (University of Adelaide) is
gratefully acknowledged for his helpful comments on an earlier draft of this paper.
References
Allen C K. 1938. Studies in the Lauraceae. 1: Chinese and Indo-Chinese species of Litsea, Neolitsea, and
Actinodaphne. Annals of the Missouri Botanical Garden 25: 361–433.
Baldwin B G, Sanderson M J, Proter J M, Wojciechowski M F, Camphbell C S, Donoghue M J. 1995. The ITS
region of nuclear ribosomal DNA: A valuable source of evidence on angiosperm phylogeny. Annals of
the Missouri Botanical Garden 82: 247–277.
Baldwin B G, Markos S. 1998. Phylogenetic utility of the external transcribed spacer (ETS) of 18S–26S
rDNA: Congruence of ETS and ITS trees of Calycadenia (Compositae). Molecular Phylogenetics and
Evolution 10: 449–463.
Barker N P, Linder H P, Morton C M, Lyle M. 2003. The paraphyly of Cortaderia (Danthonioideae; Poaceae):
evidence from morphology and chloroplast and nuclear DNA sequence data. Annals of the Missouri
Botanical Garden 90: 1–24.
Barrett M, Donoghue M J, Sober E. 1991. Against consensus. Systematic Zoology 40: 486–493.
Bena G, Jubier M F, Olivieri I, Lejeune B. 1998. Ribosomal external and internal transcribed spacers:
Combined use in phylogenetic analysis of Medicago (Leguminosae). Journal of Molecular Evolution 46:
299–306.
Buckler E S IV, Holtsford T P. 1996. Zea systematics: ribosomal ITS evidence. Molecular Biology and
Evolution 13: 612–622.
Buckler E S IV, Ippolito A, Holtsford T P. 1997. The evolution for ribosomal DNA: divergent paralogues and
phylogenetic implications. Genetics 145: 821–832.
Bull J J, Huelsenbeck J P, Cunningham C W, Swofford D L, Wadell P J. 1993. Partitioning and combining data
in phylogenetic analysis. Systematic Biology 42: 384–397.
Chanderbali A S, van der Werff H, Renner S S. 2001. Phylogeny and historical biogeography of Lauraceae:
Evidence from the chloroplast and nuclear genomes. Annals of the Missouri Botanical Garden 88:
104–134.
Chase M W, Cox A V, Soltis D E, Soltis P S, Moil M E, Savolainen V, Reeves G, Hoot S B, Morton C M.
1997. Large DNA sequence matrices, phylogenetic signal and feasibility: an empirical approach.
American Journal of Botany (Supplement) 84: 181.
Cunningham C W. 1997. Can three incongruence tests predict when data should be combined? Molecular
Biology and Evolution 14: 733–740.
de Queiroz A D, Donoghue M, Kim J. 1995. Separate versus combined analysis of phylogenetic evidence.
Annual Review of Ecology and Systematics 26: 657–681.
Doyle J J, Doyle J S. 1987. A rapid DNA isolation procedure for small qualities of fresh leaf tissue.
Phytochemical Bulletin 19: 11–15.
DNASTAR I. 1998. MegAlign3.18 eds. Madison, WI. U.S.A.: DNASTAR, Inc.
Elodenäs P, Linder H P. 2000. Congruence and complementarity of morphological and trnL-F sequence, and
the phylogeny of the African Restionaceae. Systematic Botany 25: 692–707.
Farris J S, Källersjö M, Kluge A G, Bult C J. 1994. Testing significance of incongruence. Cladistics 10:
315–319.
Farris J S, Källersjö M, Kluge A G, Bult C J. 1995. Constructing a significance test for incongruence.
Systematic Biology 44: 570–572.
Felsenstein J. 1985. Confidence limits on phylogenies: an approach using the bootstrap. Evolution 39:
783–791.
Gao L-M (高连明 ), Li D-Z (李德铢 ), Zhang C-Q (张长芹 ). 2003. Phylogenetic relationships of
Rhododendron section Azaleastrum (Ericaceae) based on ITS sequences. Acta Phytotaxonomica Sinica
(植物分类学报) 41: 173–179.
Ge S (葛颂), Schaal B A, Hong D-Y (洪德元). 1997. A reevaluation of the status of A. lobophylla based on
ITS sequence, with reference to the utility of ITS sequence in Adenophora. Acta Phytotaxonomica Sinica
(植物分类学报) 35: 385–395.
Acta Phytotaxonomica Sinica Vol. 44 284
Good L, Intine R V A, Nazar R N. 1997. Interdependence in the processing of ribosomal RNAs in
Schizosaccharomyces pombe. Journal of Molecular Biology 273: 782–788.
Hall B G. 2001. Phylogenetic Trees Made Easy: A How-To Mannual for Molecular Biologists. Massachusetts,
U.S.A.: Sinauer Associates, Inc.
Hall T A. 1999. BioEdit: a user-friendly biological sequence alignment editor and analysis program for
Windows 95/98/NT. Nuclear Acids Symbol Series 41: 95–98.
Hills D M, Bull J J. 1993. An empirical test of bootstrapping as a method for assessing confidence analysis.
Systematic Biology 42: 182–192.
Hong Y-P (洪亚平), Chen Z-D (陈之端), Lu A-M (路安民). 2001. Phylogeny of the tribe Menispermeae
(Menispermaceae) reconstructed by ITS sequence data. Acta Phytotaxonomica Sinica (植物分类学报)
39: 97–104.
Huelsenbeck J P, Ronquist F. 2001. MRBAYES: Bayesian inference of phylogenetic trees. Bioinformatics 17:
754–755.
Kluge A G. 1989. A concern for evidence and a phylogenetic hypothesis of relationships among Epicrates
(Boidae, Serpentes). Systematic Zoology 38: 7–25.
Kostermans A J G H. 1957. Lauraceae. Communication (Pengumuman) of the Forest Research Institute 57: 1–
64.
Lanyon S M. 1993. Phylogenetic frameworks: towards a firmer foundation for the comparative approach.
Botanical Journal of the Linnean Society 49: 45–61.
Li J (李捷), Christophel D C. 2000. Systematic relationships within the Litsea complex (Lauraceae): a
cladistic analysis based on morphological and leaf cuticle data. Australian Systematic Botany 13: 1–13.
Li J (李捷), Christophel D C, Conran J G, Li H-W (李锡文). 2004. Phylogenetic relationships within the ‘core’
Laureae (Litsea complex, Lauraceae) inferred from sequences of the chloroplast gene matK and nuclear
ribosomal DNA ITS regions. Plant Systematics and Evolution 26: 19–36.
Mason-Gamer R J, Kellogg E A. 1996. Testing for phylogenetic conflict among molecular data sets in the tribe
Triticeae (Gramineae). Systematic Biology 45: 524–545.
Miyamoto M M, Fitch W M. 1995. Testing species phylogenies and phylogenetic methods with congruence.
Systematic Biology 44: 64–76.
Nees von Esenbeck F C G D. 1836. Systema Laurinarum. Berlin: Sumtibus Veitii et sociorum.
Posada D, Crandall K A. 1998. Modeltest: Testing the model of DNA substitution. Bioinformatics 14:
817–818.
Roalson E H, Friar E A. 2004. Phylogenetic relationships and biogeographic patterns in North American
members of Carex section Acrocystis (Cyperaceae) using nrDNA ITS and ETS sequence data. Plant
Systematics and Evolution 243: 175–187.
Rohwer J G. 1993. Lauraceae. In: Kubitzki K, Rohwer J G, Bittrich V eds. The Families and Genera of
Vascular Plants. Berlin: Springer-Verlag. 2: 366–391.
Rohwer J G. 2000. Toward a phylogenetic classification of the Lauraceae: evidence from matK sequences.
Systematic Botany 25: 60–70.
Soltis D E, Soltis P E. 1998. Choosing an approach and an appropriate gene for phylogenetic analysis. In:
Soltis D E, Soltis P E, Doyle J J eds. Molecular Systematics of Plants Ⅱ: DNA Sequencing. Boston:
Kluwer Academic. 1–42.
Steane D A, Wilson K L, Hill R S. 2003. Using matK sequence data to unravel the phylogeny of Casurinaceae.
Molecular Phylogenetics and Evolution 28: 47–59.
Swofford D L. 1998. PAUP* Phylogenetic Analysis Using Parsimony. Sunderland, Massachusetts, USA:
Sinauer Associates, Inc.
Tamura K, Nei M. 1993. Estimation of the number of nucleotide substitutions in the control region of
mitochondrial DNA in humans and chimpanzees. Molecular Biology and Evolution 10l: 512–526.
Thompson J D, Gibson T J, Plewniak F, Jeanmougin F, Higgins D G. 1997. The ClustalX-windows interface:
Flexible Strategies for multiple sequence alignment aided by quality analysis tools. Nucleic Acids
Research 25: 4876–4882.
Tsui H-P (崔鸿宾). 1987. A study on the system of Lindera. Acta Phytotaxonomica Sinica (植物分类学报)
25: 161–171.
Van der Werff H. 1991. A key to the genera of Lauraceae in the New World. Annals of the Missouri Botanical
Garden 78: 377–387.
Van der Werff H. 2001. An annotated key to the genera of Lauraceae in the Flora Malesiana region. Annals of
the Missouri Botanical Garden 46: 125–140.
No. 3 LI Zhi-Ming et al.: Polyphyly of the genus Actinodaphne 285
Van der Werff H, Richter H G. 1996. Toward an improved classification of Lauraceae. Annals of the Missouri
Botanical Garden 78: 377–387.
White T J, Bruns T D, Lee S B, Taylor J W. 1990. Amplification and direct sequencing of ribosomal RNA
genes and the internal transcribed spacer in fungi. In: Innis M A, Gelfand G H, Sninsky F J, White T J
eds. PCR-Protocols and Applications—A Laboratory Manual. Orlando Fl.: Academic Press. 315–322.
Yang Y-C (杨衔晋), Huang P-H (黄普华). 1984. Actinodaphne Nees. In: Flora of Reipublicae Popularis
Sinicae (中国植物志). Beijing: Science Press. 31: 242–261.

Appendix A. Likelihood settings from best-fit model (TrN+I+Γ) selected by Akaike Information Criterion
(AIC) in Modeltest Version 3.06
Model selected:
TrN+I+Γ: (1) ITS: –lnLikelihood=2106.4932; AIC=4226.9863 (2) ETS: –lnLikelihood=1092.1334; AIC
=2190.2668 (3) ITS & ETS: –lnLikelihood= 3316.5947; AIC=7396.2100
Base frequencies: (1) ITS: freqA=0.1667; freqC=0.3586; freqG=0.3587; freqT=0.1161 (2) ETS: Equal
frequencies (3) ITS & ETS: –lnLikelihood=3691.1050; AIC=7396.210
Substitution model:
Rate matrix: (1) ITS: R(a) [A-C] =1.0000 R(b) [A-G] =3.5278; R(c) [A-T] =1.0000; R(d) [C-G] =1.0000;
R(e) [C-T] =5.1798; R(f) [G-T] =1.0000 (2) ETS: R(a) [A-C] =1.0000; R(b) [A-G] =3.5840; R(c) [A-T]
=1.0000; R(d) [C-G] =1.0000; R(e) [C-T] =6.9717; R(f) [G-T] =1.0000 (3) ITS/ETS: R(a) [A-C] =1.0000;
R(b) [A-G] =3.7325; R(c) [A-T] =1.0000; R(d) [C-G] =1.0000; R(e) [C-T] = 5.7495; R(f) [G-T] =1.000
Proportion of invariable sites: (1) ITS: (I) =0.3171 (2) ETS: (I) =0.7954 (3) ITS/ETS: (I) =0.459
Gamma distribution shape parameter: (1) ITS: 0.5460 (2) ETS: Equal rates (3) ITS/ETS: 0.5017.
樟科黄肉楠属是一个复系类群——基于 nrDNA
ITS和 ETS序列分析
1, 2李志明 1李 捷* 1李锡文
1(中国科学院西双版纳热带植物园植物系统与保护生物学实验室 昆明 650223)
2(中国科学院研究生院 北京 100039)

摘要 应用nrDNA ITS和ETS序列探讨了樟科Lauraceae黄肉楠属Actinodaphne的系统演化关系。对得到
的3个序列矩阵(ITS、ETS和ITS/ETS), 采用MP (maximum parsimony), ML (maximum likelihood)和
Bayesian 3种分析方法进行了系统发育分析。结果显示, 本文选的黄肉楠属Actinodaphne物种与所选的
月桂族中的外类群靠近并混和在一起, 进一步证实了本属为一个复系类群。结合对传统的形态学性状
的重新认识, 认为花序类型特征可能是重新界定黄肉楠属的最重要的性状, 具有相同花序类型的物种
可能具有相同的起源。然而, 由于取样数量相对较少以及对矩阵的单独分析存在一定的差异, 还需更详
细的研究来验证本文对黄肉楠属系统演化关系的假设, 并进一步更精确地重建本属的系统发育关系。
关键词 黄肉楠属; ETS; ITS; 樟科; 系统学研究; 贝叶斯分析