全 文 ：中国特有诸葛菜复合群的系统发育关系?
周丽蓉 , 余 研 , 宋荣秀 , 何兴金 , 蒋 彦 , 李旭锋 , 杨 毅??
(四川大学生命学院生物资源与环境教育部重点实验室 , 四川 成都 610064)
摘要 : 利用 5 .8S核糖体 DNA 全长间隔序列 ( ITS?5 .8S) 和叶绿体基因 matK 对中国原产诸葛菜复合群的系
统发育关系进行了分析研究。ITS序列结果支持中国诸葛菜复合群分为两支 : 一支由湖北诸葛菜和太白诸
葛菜组成 ; 另一支由诸葛菜和铺散诸葛菜组成。 matK 的序列分析结果表明 , 在中国诸葛菜复合群的系统
发育树中 , 铺散诸葛菜和诸葛菜处于基部位置。结合以前的核型分析和本文的研究结果 , 我们不支持将湖
北诸葛菜和太白诸葛菜归隶于诸葛菜 , 而支持我们之前提出的两种可能的原始细胞型的进化假说。另外 ,
根据生物地理学分析 , 推测从长江中下游到秦岭地区是中国原产诸葛菜的近代分化中心。
关键词 : ITS; matK ; 诸葛菜 ; 生物地理学 ; 系统发育关系
中图分类号 : Q 949 文献标识码 : A 文章编号 : 0253 - 2700 (2009) 02 - 127 - 11
Phylogenetic Relationships within the Orychophragmus violaceus
Complex (Brassicaceae) Endemic to China
ZHOU Li-Rong, YU Yan, SONG Rong-Xiu, HE Xin-J in,
J IANG Yan, LI Xu-Feng, YANG Yi * *
( Key Laboratory of Bio-Resources and Eco-Environment, Ministry of Education, Collegeof LifeScience,
Sichuan University, Chengdu 610064 , China)
Abstract : A first insight into thephylogenetic analysis of the Orychophragmusviolaceuscomplex wasconstructedfromcom-
plete sequences of nuclear ribosomal DNA internal transcribed spacer regions (ITS?5.8S) and chloroplast gene matK . The
ITS results strongly supported the division of the O. violaceus complex into two major groups: one is composed of O. hu-
pehensis, O. taibaiensis, the other consists of O. violaceus and O. diffusus . The matK phylogenetic trees showed O. dif-
fusus and O. violaceuswerethe basal members of the complex . From previous karyotype study and the present results, we
considerated that the groups of O. hupehensis and O. taibaiensis should not be allocated in O. violaceus . And two possible
ancestral cytotypes were also confirmed byour phylogenetic analysis . Moreover, the biogeographic analysis implieda pres-
umed center of recent differentiation in the region from upper to mid Changkiang River to Qin Mountain .
Key words: Biogeography; ITS; matK ; Orychophragmus violaceus; Phylogenetic relationship
Orychophragmus violaceus ( Linn .) O . E . Schulz,
amember of the Brassicaceae family, is closely related
to Brassica (Wu, 2001; Luo et al. , 1995) . This an-
nual herb is native only to China where it is labeled as
wild weed or cultivated as a garden plant or a potential
edible-oil crop (Luo et al. , 1995) .
Since O. violaceus displays a high variability, its
taxonomy has longbeen a controversial subject, with its
variety identification and species status frequently re-
vised . ThusSchulz (1903) recognizedonespecieswith
云 南 植 物 研 究 2009 , 31 (2) : 127～137
Acta Botanica Yunnanica DOI : 10 .3724?SP. J . 1143 .2009.08189
?
?? ?Author for correspondence; E-mail : yangyi528@ vip. sina. com; Fax?Tel : 86 - 028 - 85410957
Received date: 2008 - 10 - 17 , Accepted date: 2008 - 12 - 19
作者简介 : 周丽蓉 (1972 - ) 女 , 汉族 , 博士 , 主要从事植物分子细胞遗传学研究。 ?
Foundation items: National Natural Science Foundation of China (30570968; 30671165) and Key Project of Ministry of Education (105140)
fivevarieties, while Tan et al. (1998 ) identified three
additional species: O. taibaiensis Tan & Zhao, O. d-
iffususTan& Xu, and O. hupehensis (Pampanini) Tan
& Zhang . Among them, O. hupehensis was ever de-
fined as a variety of O. violaceus in Flora of China
(Zhou, 1987 ) . Later, Al-Shehbaz and Yang ( 2000)
concluded that the variation of O. violaceus did not
seem to follow any geographical distribution pattern .
Hence they re-amalgamated these ( including three ad-
ditional species proposed by Tan et al . ( 1998 ) ) into
the original species O. violaceus as a highly variable
species without any infraspecific taxa . Taking all of
those evidences together, we can infer that the problem
of reclassifying O. violaceus′svariations and threeaddit-
ional species cannot be effectively resolved by applying
conventional morphological taxonomy . In thepresent pa-
per, we adopted such scientific names as O. violaceus
complex encompassing O. hupehensis, O. taibaiensis
and O. diffuses, all of which were proposed by Tan et
al. (1998) , aswell as theoriginal species O. violaceus
in light of the present findings .
Orychophragmus hupehensis, O. taibaiensis and
O. diffusus differ significantly from O. violaceus in
both morphology ( especially in leaf morphology) and
distribution . O. violaceus and O. diffusus are biennial
herbs mostly with ovate leaves . O. violaceus features
narrowly ovate or oblongate leaves without lobe and
erect stems growing upward, a few branches at the
base, and some amplexicaul leaves at the top ( Luo et
al. , 1995 ) . O. diffusus grows with diffuse caules and
many branches at the base . All of its leaves are lyrata
with serrate edge; the terminal leaf is cordate or reni-
form and the leaf tip is blunt ( Tan et al. , 1998 ) .
However, O. taibaiensis and O. hupehensis are annu-
als and characterized by theuppermost stemleaves hav-
ing approximate triangle, being petiolate and not am-
plexicaul . O. taibaiensis is different from O. hu-
pehensis in that its uppermost stem leaves have slightly
pinnate cleavage, cordateor ovoid triangles with apicu-
late or obtuse ends, whereas for O. hupehensis its stem
is underdeveloped, with all leaves largely pinnated
cleavage, inverse-triangle and apex acute to acumi-
nate . They are not overlapping but neighboring in their
distribution . O. taibaiensis, native to Shaanxi , Meix-
ian, Mt . Taibai of China, grows at an elevation of
1150 - 1470 m, a height typical of cold climate of
north-western China ( Tan et al. , 1998 ) . O. hu-
pehensis is, however, generally confined to Danjiankou
and Shiyan of Hubei , and grows at an elevation of 450
mto 1000 mwhere it is relatively warm .
A recent phylogenetic analysis based on plastid
and nuclear molecular markers suggested that O. vio-
laceus was a sister clade to Brassica ( Warwick and
Sauder, 2005) . However, this only provided one ac-
cession of O. violaceus . As little is knownof the evolu-
tionary history of this highly variable taxa, molecular
phylogenetic analysis is crucial in evaluating relations
among the chromosomal races by using the matK region
of the chloroplast genome; and among different local
populations by using the more variable internal tran-
scribed spacer ( ITS ) sequences of nuclear ribosomal
DNA ( nrDNA ) . The usefulness of both markers in
species-level systematics and phylogeny of angiosperms
has been widely demonstrated in recent studies (Diane
et al. , 2002; Winkworth et al. , 2002; Li et al. ,
2004; Selvi et al. , 2006) . Based on our previous re-
search, the goals of this study are to examine the cir-
cumscription of O. violaceus, to investigate its major
lineages, and to explain the evolutionary relationships
within the O. violaceus complex .
1 Materials and methods
1 .1 Plant materials
Fifteen natural populations were sampled across the main
distribution of the O. violaceus complex . These representative
populationswerefrom Hubei , J iangxi , Shaanxi, Sichuan, Shan-
dong, J iangsu and Zhejiang Provinces and the cities of Beijing
and Shanghai . As O. diffusus is rare and has a small population
size, we only found one population in Shanghai . The geographi-
cal origins of accessions aregiven inTable1 and all abbreviations
of the populations mentioned in the text follow as Table 1 .
Voucher specimens were deposited in the Herbarium of Sichuan
University . Silica-gel dried samples of leaf tissue of each popula-
tion were prepared for molecular analyses .
1 . 2 DNA extraction
Genomic DNA was extracted following amodified 2×CTAB
protocol (Doyle and Doyle, 1987 ) using samples of tissue cut
821 云 南 植 物 研 究 31 卷
from leaves . The concentration of DNA in the various samples
was determined by measuringtheabsorbance at 260 nm (A260 ) of
aten-fold dilutionof each sample . The quality of all DNA prepa-
rationswas checked by agarose gel electrophoresis (0 . 7% w?v)
in TAE buffer (1 mmol?L EDTA , 40 mmol?L Tris-acetate) con-
taining1μg?mL of ethidiumbromideby comparison with a known
mass standard .
1 . 3 PCR amplification and DNA sequencing
Analysis of thematK region was donefor O. diffususand for
oneaccessionof O. violaceus, O. hupehensis andtwopopulations
of O. taibaiensis with different ploidy level . The two primers
matKF ( 5′-TAATACCTTATTTTGACTGTATCGCACTAT-3′) and
matKR ( 5′-CCAAATCATTAAGATAAAGAATATCCAAA TACC-
3′) used in this studywere designed on the basis of the constant
regions of the published sequences of Acorus calamus (GenBank
NC007407) , Brassica napus ( GenBank X63088) , Arabidopsis
thaliana (GenBank AP000423) and Disporopsis undulata (Gen-
Bank AB029767) resulting from the alignment . Analysis of ITS
nuclear rDNA was done for O. diffusus, the two populations of
O. taibaiensis, thethreeof O. hupehensis and nine accessions of
O. violaceus representing the distribution range (Table 1 ) ; the
primers ITS1-18S (O′Kane et al. , 1997) and ITS4 (White et
al. , 1990) were used .
For both matK and ITS, PCR amplificationswere performed
in a total volume of 50μl of reaction buffer, 1 . 5 mmol?L MgCl2 ,
20 pmol of each primer, 200μmol?L of each dNTP, 1 U of Tag
DNA polymerase (Fermentas) and 10 ng of template DNA . Re-
actionswere performed in a Peltier Thermal Cycler ( Bio-RAD
DNAEngine) and programmed for an initial denaturation step (3
min at 94℃ ) followed by 30 cycles of 1 min at 94℃ , 1 min at
55℃ , and 2 min at 72℃ . The last cyclewas followed by afinal
incubation of 10 min at 72℃ . The samples were then stored at
4℃ . Subsequently, 5μl of each amplification mixture was ana-
lyzed by agarose gel ( 1% w?v) electrophoresis in TAE buffer
containing 1μg?mL ethidium bromide . The PCR reactions were
purified from excess salts and primer using the Qiagen QIAquick
PCR Purification Kit ( Realtimes) . Automated DNA sequencing
was performed directly fromthepurified PCR products using Big-
DyeTerminator v . 2 chemistry and an ABI 3730 XL DNA Se-
quencer ( Applied Biosystems) by Beijing Genomics Institute .
TheITS region was not cloned as there was no evidence of the
presence of two or more ITS sequences in the same individual .
1 . 4 Sequence alignment and analysis
DNA sequences and overlapping fragments were assembled
and edited using SeqMan of DNAStar software package . The ITS
sequences were checked for orthology to sequences of O. vio-
laceus (GenBank AY722506 ) , Vella spinosa (AY722498) and
Brassica rapa (AF531563) . The two species were then used as
outgroups for cladogram construction based on their position in
thephylogeny of the tribe (Warwick and Sauder, 2005 ) . The
sequence boundaries between the two spacers ( ITS1 and ITS2 )
and the coding regions ( 5 . 8S) of nrDNA were determined by
comparison with the published O. violaceus sequence ( Warwick
and Sauder, 2005) .
Multiple alignments were automatically performed using
CLUSTALW in MEGALIGN of DNA Star ( DNAstar Madison,
WI ) , and then further examined and slightly modified manually .
The matK sequences were aligned with the sequences of Sinapis
alba (AB354277) and Brassica rapa ssp. campestris (AB354276) ;
thetwowere then used as outgroup representatives in the matK
tree in light of the close relationship to Orychophragmus . Be-
cause different analysis methodsaresensitivetodifferent biases in
the data set, Baum et al. (1994) suggested that analyzing data
with multiple algorithms is desirable and that clades consistently
supported in different analyses might be considered more robust
than those supported strongly by one search method but contra-
dicted by another . Phylogenetic analyses for each matrix were
carried out by Bioctrl package ( http:?mnh. scu. edu. cn?
Bioctrl .htm) usingmaximumparsimony (MP) , maximumlikeli-
hood (ML ) , and Bayesian inference ( BI ) methods in PAUP*
version 4.0b10 ( Swofford, 2003 ) , PHYML version 2 .4 .3
(Guindon and Gascuel , 2003 ) , and MrBayes version 3 .0b4
(Ronquist and Huelsenbeck, 2003 ) , respectively . For MP anal-
yses, heuristic searches were conducted with 1000 replicates of
randomaddition, onetreeheldat each step during stepwiseaddi-
tion, tree-bisection-reconnection (TBR ) branch swapping, Mul-
Trees in effect, and steepest descent off . Gaps were treated as
missing data, characters were equally weighted, and their states
were unordered . Internal branch supportwas estimated with 1000
bootstrap replicates ( Felsenstein, 1985 ) . For maximum likeli-
hood (ML) analyses, themodel of sequence evolution fitting the
databestwas determined with Modeltest, version 3 .06 (Posada
and Crandall , 1998 ) . GTR + I + G model best fit the data
(based on the alignment without matrix) . It was applied along
with theModeltest-estimated parameters in heuristic searcheswith
one addition sequence replicateandTBR branch swapping . Nod-
al robustness on the ML tree was estimated by the nonparametric
bootstrap (1000 replicates for ITS and 100 for matK ) . Bayesian
analyses were accomplished in MrBayes version 3 .0b4 using the
best-fit models upon Akaike informationcriterion (AIC; Akaike,
1974) by using Modeltest 3 .7 (Posadaand Crandall , 1998 ) . In
Bayesian analyses, treesweregenerated by runningfour simulta-
neousMetropolis-coupled Monte Carlo Markov (MCMC) chains
and sampling one tree every 1000 generations for 1 , 000 , 000
startingwith a random tree . The posterior probability (PP) was
used to estimate nodal robustness .
9212 期 ZHOU Li-Rong et al. : Phylogenetic Relationshipswithin the Orychophragmus violaceus Complex . . .
Table 1 Taxa, origins, vouchers and accessions of the O. violaceus complex investigated and other species used in this study
Taxa and code Location Alt .( m) Source or vouchers GB No . ITS?matK
Orychophragmusviolaceus HS Huashan Mountain 600 NL ?. R . Zhou 0701 EU306542
O ?. violaceus NJ Nanjing 90 ;L . R . Zhou 0702 EU306548
O ?. violaceus TS Mount Taishan 1000 - 1500 L . R . Zhou 0703 EU306552
O ?. violaceus JJ Jiujiang, JiangXi 200 NL . R . Zhou 0704 EU306546
O ?. violaceus BJ Beijing 44 ;L . R . Zhou 0705 EU306544
O ?. violaceus XH HangZhou 70 ;L . R . Zhou 0706 EU306554
O ?. violaceus CD Chengdu, Shuangliu 500 NL . R . Zhou 0708 EU306551
O ?. violaceus CS ChaoShan, Yuhang 300 NL . R . Zhou 0709 EU306541
O ?. violaceus SNJ Shennongjia 900 NL . R . Zhou 0707 EU306550?EU306555 *
O ?. hupehensis SY Wudang Mountain 450 NL . R . Zhou 0712 EU306543
O ?. hupehensis DJ Danjiangkou 400 NL . R . Zhou 0713 EU306545
O ?. hupehensis WD Shiyan 1000 `L . R . Zhou 0711 EU306553?EU306558 *
O ?. taibaiensis JK Jiaokou, Taibai Mountain 1150 `L . R . Zhou 0714 EU306547?EU306556 *
O ?. taibaiensis HP Haoping, Taibai Mountain 1210 - 1250 L . R . Zhou 0715 EU306540?EU543181 *
O ?. diffusus SH Shanghai 80 ;L . R . Zhou 0710 EU306549?EU306557 *
O ?. violaceus Warwick et al :. (2005) AY722506
Diplotaxis brachycarpa Warwick and Sauder (2005 ?) AY722445
Vella spinosa Warwick et al :. (2005) AY722498 ?
Sinapis alba Lu et al ?. ( 2007 ) AB354277 *
Brassica rapa subsp .. campestris Lu et al ?. ( 2007 ) AB354276 *
Brassica rapa Warwick et al :. (2005) AF531563 ?
Acorus calamus Zhang et al ?. (2005) DQ008868 *
Brassica napus Handa et al ?. ( 1992 ) X63088 *
Arabidopsis thaliana Sato et al ?. (1999) AP000423 *
Note:“ * ”indicate accession numbers of matK gene .
2 Results
2 .1 Sequence characteristics
Sequences were deposited in GenBank-EMBL-
DDB and can be retrieved using the numbers in Table
1 . The ITS region of nrDNA comprising both ITS se-
quences ( ITS1 and ITS2) and the 5 .8S rDNA was am-
plified by PCR from15 populations of the O. violaceus
complex (Table 1) . The aligned matK sequences were
1357 bp in length with no insertions or deletions . As
matK is highly conservative in Brassiceae ( Marcus et
al. , 2001 ) , the number of nucleotide substitutions was
limited . Ingroup variation was low, with only 19 vari-
ablepositions ( 1 .4% ) . In the phylogenetic analysis,
1282 positions were constant (94 .47% ) , only 26 sites
wereparsimony-informative (1 . 92% ) . In contrast, the
ITS region was quite variable among taxa . The ITS
aligned sequence data set was 553 bp in length . In-
group variation was remarkable, with 240 positions
( 42 .55% ) beingvariable . Addingthetwo outgroups for
the purposes of the phylogenetic analysis, 425 characters
were constant, 63 parsimony-uninformative and 64
(11 .57% ) parsimony-informative .
2 . 2 matK analysis
The 50% majority-rule consensus treeof 83 equally
parsimonious trees and theML phylogenetic tree, generat-
ed by successive weighting analysis of the matK subma-
trix, areshown in Fig . 1 .TheMP treehad alengthof 82
steps, a consistency index (CI ) of 0.9759 and a retention
index (RI ) of 0 .9394 .Theoverall topology of theMP tree
was essentially identical to that produced by the maxi-
mum-likelihood ( ML ) and Bayesian analysis ( Fig.2 ) .
Differences between Bayesian, MP and ML trees con-
cernedmainly the bootstrap values at some nodes, which
showed slight variation from tree-to-tree . For instance,
the clade comprising the O. violaceus - O. hupehensis
- O. taibaiensis complex was supported by bootstrap
values of 100% , 94% and 95% in theBayesian, ML
and MP trees respectively . However, in the ML tree,
O. hupehensis and thetwo populationsof O. taibaiensis
were not separated as the other two trees showed .
The phylogenetic analysis based on matK showed
that the O. violaceus complex formed astrongly supported
clade (100% bootstrap values in all trees) . Despite low
genetic divergence for matK , rootingof the treewith S. -
alba and B. rapa ssp. campestris, resulted in O. diffusus
031 云 南 植 物 研 究 31 卷
Fig . 1 The phylogentic trees of matK of different cytotypes of the O. violaceus complex
Left tree: The 50 % majority-rule consensus treeof 83 equally parsimonious trees generated by matK sequences . Right tree: The maximumlike-
lihood tree frommatK sequences from different cytotypes of the O. violaceus complex . Bootstrap values are shown on branches when > 50 % .
Abbreviations of O. taibaiensis accessions follow Table 1 . Base chromosome numbers and ploidy are indicated for ingroup and outgroup taxa .
Fig . 2 The Bayesian trees obtained from matK sequences of different cytotypes in the O. violaceus complex
Bayesian bipartition posterior probabilities are shown on the branches . Abbreviationsof O. taibaiensis follow Table1 . Base chromosomenumbers
are indicated for ingroup and outgroup taxa .“ * ”indicated Bayesian bipartition posterior probabilities < 0 .5 .
as the earliest divergent lineage, which formed a well
supported clade (95% BS in MP tree) sister to O. vio-
laceus . O. violaceus was, in turn, moderately supported
as the second basal member of the complex sister to
O. taibaiensis and O. hupehensis . Additionally, the se-
quence dataset indicated that O. hupehensisdiffered from
O. taibaiensis inonly two deletions (oneof 2 bp and one
of 1 bp) andone 1-bp insertion .
2 . 3 ITS analysis
The consensusMP phylogenetic tree ( L = 197 , CI
= 0 .7766 , RI = 0 .8027 ) and the Bayesian tree de-
rived from ITS?5 .8S sequences was shown with boot-
strapvalues in Fig . 3 and the ML treewith distribution
of other characters was shown in Fig . 4 . The topologi-
cal identity is found in three phylogenetic trees in a
majority of clades . Differences between MP, Bayesian
and ML trees were just thevalues of some clades and a
little difference in the clade B2 . Compared to Bayesian
and MP trees, the positions of two populations SH and
BJ in the clade B2 is transposed by the ML analysis,
whereasMP tree and Bayesian tree had additional dif-
ferences in the cladeB2 . For example, Bayesian anal-
1312 期 ZHOU Li-Rong et al. : Phylogenetic Relationshipswithin the Orychophragmus violaceus Complex . . .
ysis exhibits paratactic relations between the popula-
tions of SH, XH, CS, TS in the second branch of
clade B2 , while in MP tree, these populations are hy-
potactic relations in the same branch . This may be re-
sult from the different algorithms used by different
methods .
All constructed ITS phylogenetic trees congruously
suggested that the O. violaceus complex was divided
into two major clades, which are strongly supported
(MPBPs?MLBPs?BPP: 86?89?1 .00 in cladeA ; MPBPs?
MLBPs?BPP: 66?65?1.00 inclade B ) . Clade A was
composed of two very well supported subclades: clade
A1 containing a tetraploid and a diploid accessions of
O. taibaiensis ( ( MPBPs?MLBPs?BPP: 96?97?1 .00 )
and clade A2 containing three populations of O. hu-
pehensis ( (MPBPs?MLBPs?BPP: 82?81?1 .00) .
Being thesister group to cladeA , theother major
cladeB consists of two subclades: clade B1 and B2 .
Clade B1 , the weakly supported subclade ( MPBPs?
MLBPs?BPP: 47?45?0 .94) , ismade up of thepopula-
tions of HS, JJ and SNJ . In this clade, populations of
SNJ and JJ clustered together with avery high bootstrap
value ( (MPBPs?MLBPs?BPP: 96?96?1 .00) , confirm-
ing a very close relationship between them . The other
more strongly corroborated lineage ( MPBPs?MLBPs?
BPP: 99?99?1 .00) , clade B2 , was formed by the re-
maining populations of O. violaceus and O. diffusus .
However, O. diffusus, which is represent the cytotype
of chromosome base number x = 10 , was nested in
clade B2 .
Fig . 3 The Bayesian trees ( left) and the maximum parsimonious phylogenetic tree in MP analysis ( right)
generated by ITS sequences of the O. violaceus complex
Bootstrap values and Bayesian bipartition posterior probabilities are shown on or below the branches .
Abbreviations of populations are followed Table 1
231 云 南 植 物 研 究 31 卷
Fig . 4 ML tree inferred from the ITS data showing the distribution of selected character states (1-6 )
The results of ML bootstrap analysis are shown aboveor below thebranches . These charactersare: 1 , the chromosomenumbers; 2 , the terminal
leaves’shape (△ , ovoid triangles with apiculate or obtuse ends; ▲ , inverse triangles with apiculate ends; ☆ , narrowly ovate or oblongate;
○ , cordate or reniform) ; 3 , number of the lateral lobes of the uppermost leaves; 4 , theuppermost leaves are amplexicaul ( + ) or not ( - ) ;
5 , chroatographic points of flavonoides from the leaves in Rf . = 0 .24 - 0 .25 ( DY = dark yellow; BYG = bright yellowish green;“ - “ = no
point) ; 6 , chroatographic points of flavonoides from the leaves in Rf . = 0 . 36 - 0 . 38 ( PB = pale blue; the other codes followed as above) .
3 Discussion
3 .1 Phylogenetic relationships within the O. vio-
laceus complex
The branches found in the analysis above were
consistent based on molecular data . These results
showed that the O. violaceus complex could be divided
into two sister groups, which were strongly supported
by three phylogenetic methods ( MP, ML and Bayes-
ian) . One major group was composed of population
O. hupehensis and O. taibaiensis ( clade A ) , and the
other of accession O. diffusus and O. violaceus ( clade
B) . However, thetwomajor groups (A andB) showed
a paratactic phylogenic relationships in the ITS trees,
whereas the phylogenic analysis based on matK indi-
cated that O. hupehensis and O. taibaiensiswere diver-
gent later than O. diffusus and O. violaceus . This may
bedue toaslower evolutionary rateof matK than that of
ITS . In our phylogenic analysis, the matK trees could
distinguish group A and group B more clearly, whereas
the ITS trees just proved themas two separate clades .
O. hupehensis and O. taibaiensis were sister
groups separated from O. violaceus, but due to their
high bootstrap support, each one became a true clade
in the ITS phylogenetic trees . Both ITS and matK data
displayed comparatively close relationships between the
two cytotypes, suggesting that they might be the same
ancestor yet subsequently divergent in different evolu-
tionary patterns almost at the same time . Significantly,
two cytotypes of O. taibaiensis, the diploid ( JK ) and
tetraploid ( HP) accessions, clustered together with pa-
ratactic relations both in ITS and matK trees . This
suggested that the tetraploid was probably formed di-
rectly fromthe same ancestor of the diploid in one sin-
gle evolutionary step, but not derived from the diploid
3312 期 ZHOU Li-Rong et al. : Phylogenetic Relationshipswithin the Orychophragmus violaceus Complex . . .
with the similar morphology .
It could also be inferred from the ITS cladograms
that thepopulationsof O. violaceus in cladeB differen-
tiated into two subgroups with the same chromosomal
number (Zhou et al. , in press) , though with wide ra-
diationyet littlemorphological variation . For cladeB1 ,
the SNJ and J J populations, which are two isolated
populations in the upper and middle reaches of the
Changkiang River, werestrongly supported avery close
relationship between them . This suggested a relatively
recent disjunction . Additionally, our ITS analysis
showed O. diffusus fell inside O. violaceus in clade 4 ,
but obviously, it was the earliest clade separated from
other cytotypes in the matK cladogram with very high
bootstrap value ( 95% in MP) . One possible cause of
this obscure relationship may be the absence of near
relatives of the rare cytotypes, which resulted in the
very small number of informative sites . Another cause
may also be the different resolving power of the two
genes . However, it could be inferred from matK data
that therewas a sharp distinction between O. violaceus
and other cytotypes .
3 . 2 Taxonomic implication
Given the presence of clear karyotype and other
different characters ( Fig . 3) , this division of the O. -
violaceus complexwas, not surprisingly, quiteobvious .
The present reconstruction indicated that the indepen-
denceof O. hupehensis and O. taibaiensis from O. vio-
laceus had been karyologically paralleled by a differen-
tiation in chromosome number, karyotype asymmetry
and heterochromatin content . O. hupehensis was char-
acterized by karyotypeof“12m+ 2msat + 6sm+ 2smsat”
and a higher content of constitutive heterochromatin
(about 45 .5% of the total karyotype length) ( Zhou et
al . in press) . However, O. violaceus displayed adif-
ferent karyotype of“16m+ 2msat + 4sm+ 2smsat or 14m
+ 2msat + 6sm + 2smsat”. The cytotype O. diffusus,
which was clustered with O. violaceus, also showed a
diploid complement of 2 n = 2 x= 20 (Zhou et al. , in
press) different from populations of group A .
Moreover, this remarkable divergencebetween the
two major groups ( A and B ) was also supported by
morphology and phytochemistry traits ( Zhao, unpul-
ished) . Members of clade A was characterized by tri-
angular shape in general , but populations in clade B
weremostly ovate leaves (Fig . 5) . Furthermore, the
Fig . 5 The comparison of the leaves morphology of 15 populations of the O. violaceus complex
For codes of the populations, see Table 1 . The members of group A are in clade A of the phylogentic trees and the members of group B are in clade B
431 云 南 植 物 研 究 31 卷
morphological variations among themwere stable even if
naturalized to thesamefield and environment ( personal
observation) . This indicated that the variation of phe-
notype was due to genetic rather than environmental
factors . The thin-layer chromatogram of flavonoides in
the leaves fromdifferent populationsof the O. violaceus
complex (Zhao, unpulished) displayed significant dif-
ference between the two major groups . Therefore, these
features and the resultsof our study supported theexclu-
sion of O. hupehensis and O. taibaiensis from O. vio-
laceus, rather than the inclusion of the three into one .
Based on previous taxonomic analysis, the exclu-
sion of O. hupehensis, O. taibaiensis and O. diffusus
from O. violaceus was advocated by Tan et al .
( 1998 ) . From our karyotypical and phylogenic re-
search work, we proposed that the groups of O. hu-
pehensis and O. taibaiensis should not be allocated in
O. violaceus . Their populations are mainly character-
ized with the uppermost stemleaves morphology of ap-
proximate triangle, being petiolate and not amplexi-
caul , which are different from O. violaceus . Because
thenameof O. hupehensis is early than O. taibaiensis,
we recommend to recovery the species of O. hupehensis
which is include O. taibaiensis . To confirmthis, much
more sampling and further analyses, such as micromor-
phology, hybridization or other genetic and molecular
studies, should be performed within the complex or the
genus Orychophragmus .
3 . 3 Evolution and biogeography of the O. vio-
laceus complex
The phylogenetic studyof thespeciesgroup, com-
bined with analysisof the recordedgeographic distribu-
tion, allowed a tentative reconstruction of the major
features of their biogeographic history (Nikulina et al. ,
2007) . The molecular phylogenetic analysis based on
cpDNA matK and nrDNA ITS gene revealed a signifi-
cant population differentiation pattern consistent with
the results of karyotype analysis of the O. violaceus
complex .
The matK analysis showed that the cytotypes of x
= 10 and x= 12 were the basal clades of the O. vio-
laceus complex tree . This result revealed the earlier di-
vergenceof x= 10 and x= 12 and thederivedoriginof
O. hupehensis and O. taibaiensis, a strongproof of the
hypothesis of two possible ancestral bases concluded
from karyotypical analysis, the lowest base x= 10 and
themostly conserved base x= 12 in the O. violaceus
complex ( Zhou et al. , in press) . Although there were
few informative matK sites, what underlined our data
was the fact that this phylogenetic relationship was
strongly supported . This proved, as Li et al . (2004)
pointed out, that given the more conservative natureof
matK , their molecular characteristics may well reflect
the phylogenetic signals .
The map of the O. violaceus complex distribution
(Fig . 6 ) showed that HS, JJ and SNJ accessions of
O. violaceus, the members of clade B1, had sympatry
with thegroup A in theupper andmiddle reachesof the
Changkiang River to Mt . Qinling . In these areas, the
O. violaceus complex exhibited higher diversity than in
other area: two basenumber ( x= 11, 12) and a tetra-
ploidy, as well as some populations fromdifferent major
groups . Therefore we presumed that these areas may be
the recent differentiation center for the O. violaceus
complex . HS, JJ and SNJ accessions of O. violaceus,
which maked up clade B1 and was located in this cen-
ter, had lower chromosomal asymmetry than the popula-
tions of cladeB2 inother areas except for littlemorphol-
ogy variation between them . Accordingly, we hypothe-
sized that O. violaceus possibly radiated fromthe center
into farther areas . This geographic pattern may be a
consequenceof undirectional or balanced selection .
Additionally, notablegeneticvariations in O. vio-
laceus were found via the phylogenetic analysis . The
abundanceof haplotypes detected by ITS analysis might
result from changed environments or geographical seg-
regation of this species . All the results indicated that
radialized colonization probably occurred when new
ecotypes were produced to adjust to new habitats . Eco-
logical differentiation was perceived as a key factor for
maintaining cytotypes in an allopatric condition
(Thompson and Lumaret, 1992; Levin, 2002; Hus-
band, 2004) . Taking into account thelargegenetic di-
vergence between the cryptic species and their pa-
raphyletic relationships, we inferred that the O. vio-
laceus complex might have experienced rapid expansion
5312 期 ZHOU Li-Rong et al. : Phylogenetic Relationshipswithin the Orychophragmus violaceus Complex . . .
Fig . 6 Location map of 15 populations in O. violaceus complex and distribution ranges of different groups
Red dashed line: range of group A ; black line: range of group B ; green dashed line: range of subgroup B1
and diversification or an even longer period of evolu-
tionary independence ( lack of gene flow between their
locations) . Historical and ecological factors may cont-
ribute to its allopatric speciation in China .
This is the first insight into O. violaceus phyloge-
netic analysis . We offer the preliminary hypotheses
which are hopefully to be verified with additional sam-
pling of more taxa and more characteristics . Further-
more, specific divisions within the complex require ex-
tensive efforts in redefinition in consideration of prob-
lematic positions of these cryptic species .
Acknowledgements : We wish to thank Miss Jing Huang and
Mr . Yufei Yue for their kind help in sample collection, aswell
as Mr . Williams Mark for modifying the earlier version of the
manuscript .
References:
Akai ?ke H , 1974 . A new look at statistical model identification [ J ] .
IEEE Transactions on Automatic Control , 19 : 716—723
Al-S ?hehbaz IA , Yang G, 2000 . A revision of the Chinese endemic
Orychophragmus (Brassicaceae) [ J ] . Novon, 10 : 349—353
Baum ?DA , Sytsma KJ , Hoch PC , 1994 . A phylogenetic analysis of Epi-
lobium ( Onagraceae) based on nuclear ribosomal DNA sequence
[ J ] . Systematic Botany, 19 : 363—388
Dian ?eN , Fo¨ rther H , Hilger HH , 2002 . A systematic analysis of Heli-
otropium, Tournefortia and allied taxaof the Heliotropiaceae ( Boragi-
nales) based on ITS1 sequences and morphological data [ J ] . Ameri-
can J ournal of Botany, 89 : 287—295
Doyl ?e JJ , Doyle JL , 1987 . A rapid DNA isolation procedure for small
quantities of fresh leaf material [ J ] . Phytochemical Bulletin, 19 :
11—15
Fels ?enstein J , 1985 . Confidence limitson phylogenies: an approach using
the bootstrap [ J ] . Evolution, 39 : 783—791
Guin ?don S, Gascuel O, 2003 . A simple, fast, and accurate algorithm to
estimate large phylogenies by maximum likelihood [ J ] . Systematic
Biology, 5 : 696—704
Husb ?and BC , 2004 . Chromosomal variation in plant evolution [ J ] .
American J ournal of Botany, 91 : 621—625
Li J ?, Christophe JG, Conran JG et al. , 2004 . Phylogenetic relationships
within the‘ core’Laureae ( Litsea complex, Lauraceae) inferred from
sequences of the chloroplast gene matK and nuclear ribosomal DNA
ITS regions [ J ] . Plant Systematics and Evolution, 246 (1-2) : 19—34
Levin DA , 2002 . The Role of Chromosomal Change in Plant Evolution
[M ] . New York: Oxford University Press, 98—163
Luo P (罗鹏 ) , Chen ZL ( 陈昭麟 ) , Lan ZQ (蓝泽蘧 ) , 1995 . Studies
on the Plant Genetic Resources of Orychophragmus violaceus (植物
631 云 南 植 物 研 究 31 卷
遗传资源诸葛菜研究 ) [M ] . Chengdu: Sichuan University Press
Marc us K , BernhardH , ThomasMO, 2001 . Molecular systematics of the
Brassicaceae: evidencefrom coding plastidic matK and nuclear Chs
sequences [ J ] . American Journal of Botany, 88 (2 ) : 534—544
Niku #lina EA , Hanel R , Schafer P , 2007 . Cryptic speciation and pa-
raphyly in the cosmopolitan bryozoan Electra pilosa— Impact of the
Tethys closing on speciesevolution [ J ] . Molecular Phylogenetics and
Evolution, 45 ( 3) : 765—776
O’ Ka lne SL , Schaal B, Al-Shehbaz IA , 1997 . Theorigins of Arabidop-
sis suecica (Brassicaceae) as indicated by nuclear rDNA sequences
[ J ] . Systematic Botany Systematic Botany, 21 : 559—566
Otto $SP, Whitton J , 2000 . Polyploid incidence and evolution [ J ] . Annu-
al Review of Genetics, 34 : 401—437
Posa ?da D, Crandall KA , 1998 . Modeltest: testing the model of DNA
substitution [ J ] . Bioinformatics, 14 : 817—818
Ronq /uist FR , Huelsenbeck JP , 2003 . MrBayes 3: Bayesian phylogenetic
inference under mixed models [ J ] . Bioinformatics, 19 : 1572—
1574
Schu lz OE, 1903 . Monographie der Gattung Cardamine [ J ] . Botanische
J ahrbuch Systematic, 32 : 280—623
Selv ?i F , Coppi A , Bigazzi M , 2006 . Karyotype variation, evolution and
phylogeny in Borago ( Boraginaceae) , with emphasis on subgenus
Buglossites in the Corso-Sardinian system [ J ] . Annals of Botany
(London) , 98 (4 ) : 857—868
Swof ?ford DL , 2003 . PAUP* . Phylogenetic analysis using parsimony ( and
other methods) [ CP?CD ] . Version 4 , Sinauer Associates, Sunder-
land , Massachusetts .
Tan ?ZM, Xu JM, Zhao BX et al. , 1998 . New taxa of Orychophragmus
(Cruciferae) from China [ J ] . Acta Phytotaxonomica Sinica , 36 :
544—548
Thom ?pson JD, Lumaret R , 1992 . The evolutionary dynamics of polyploid
plants: origins, stablishment and persistence [ J ] . Trends in Ecolo-
gy & Evolution, 7 : 302—306
Warw ?ick SI , Sauder CA , 2005 . Phylogeny of tribe Brassiceae ( Brassi-
caceae) based on chloroplast restriction site polymorphisms and nu-
clear ribosomal internal transcribed spacer and chloroplast trnL in-
tron sequence [ J ] . Canadian J ournal of Botany, 83 : 467—483
White TJ , Bruns T , Lee S et al. , 1990 . Amplification and direct se-
quencing of fungal ribosomal RNA genes for phylogenetics [A ] . In:
Innis MA , Gelfand DH , Sninsky JJ eds . PCR Protocols a Guide to
Methods and Applications [ M] . New York: Academic Press, 315—
322
Wink ?worth RC , Grau J , Alastair WR et al. , 2002 . Theorigins and evo-
lution of the genus Myosotis L . ( Boraginaceae) [ J ] . Molecular
Phylogenetics and Evolution, 24 : 180—193
Wu Z ?Y , 2001 . Floraof China [M ] . Beijing: Science Press, 8 : 26—27
Zhou ?TY , 1987 . Flora of China, Vol . 33 [M ] . Beijing: Science Press,
40—43
* * * * * * * * * * * * * * *
重 要 启 事
《云南植物研究》 自 1979 年创刊以来 , 已走过 30 年的历程 , 在她的成长过程中得到了广
大作者、审稿专家、编委会全体人员和读者的厚爱 , 使其健康成长 , 现已成为中国科学技术核
心期刊 , 国内发表植物学研究论文的主要学术期刊。为使期刊能在新形势下更加良好发展 , 步
入一个新的历史阶段 , 现经编委会研究决定 : 自发表通告之日起 , 凡投到本刊的稿件 , 若经审
稿专家评审认为稿件质量优良的 ; 或经学科副主编和编委推荐直接进入绿色通道发表的优秀论
文 , 本刊将一律免收其论文发表费及审稿费并及时发表。若审稿意见认为稿件内容较好的 , 但
需经过仔细修改的稿件 , 若修改通过 , 将提前发表 , 不受时间约束。我们衷心希望广大作者能
一如既往的支持我们 , 我们也将尽我们最大的努力为大家服务。期刊的发展是我们共同的愿
望 , 相信随着时间的推移 , 《云南植物研究》 将会得到长足的发展 , 并以斩新的面貌来回报一
直以来关心和爱护她的所有作者、审稿专家和读者。
《云南植物研究》 编辑部
2009 年 3 月 25 日
7312 期 ZHOU Li-Rong et al. : Phylogenetic Relationshipswithin the Orychophragmus violaceus Complex . . .