全 文 ：作物学报 ACTA AGRONOMICA SINICA 2012, 38(8): 1407−1415 http://www.chinacrops.org/zwxb/
ISSN 0496-3490; CODEN TSHPA9 E-mail: xbzw@chinajournal.net.cn

This work was supported by Yunnan Tobacco Company (08A05, 2010YN02, 2011YN04) and China National Tobacco Company
(110201002001).
* Corresponding authors: WU Wei-Ren, E-mail: wuwr@zju.edu.cn; XIAO Bing-Guang, E-mail: xiaobg@263.net
The first author’s: E-mail: tzj861@163.com
Received(收稿日期): 2012-02-27; Accepted(接受日期): 2012-04-20; Published online(网络出版日期): 2012-06-04.
URL: http://www.cnki.net/kcms/detail/11.1809.S.20120604.1011.015.html
DOI: 10.3724/SP.J.1006.2012.01407
Mapping of Quantitative Trait Loci Underlying Six Agronomic Traits in Flue-
Cured Tobacco (Nicotiana tabacum L.)
TONG Zhi-Jun1,2, JIAO Fang-Chan2, WU Xing-Fu2, WANG Feng-Qing1, CHEN Xue-Jun2, LI Xu-Ying2,3,
GAO Yu-Long2, ZHANG Yi-Han2, XIAO Bing-Guang2,*, and WU Wei-Ren1,4,*
1 Department of Agronomy, College of Agriculture & Biotechnology, Zhejiang University, Hangzhou 310058, China; 2 Yunnan Academy of Tobacco
Agricultural Sciences, Yuxi 653100, China; 3 College of Agriculture & Biotechnology, Yunnan Agricultural University, Kunming 650201, China;
4 College of Life Sciences, Fujian Agriculture & Forestry University, Fuzhou 350002, China
Abstract: To identify quantitative trait lici (QTLs) for important traits in tobacco, a doubled haploid (DH) population derived
form the cross between flue-cured tobacco cultivars Honghua Dajinyuan and Hicks Broad Leaf was used in phenotypic survey and
SSR analysis. Six leaf yield-related traits, i.e., plant height (PH), stem girth (SG), internode length (IL), leaf number (LN), length
of the largest waist leaf (LWL), and width of the largest waist leaf (WWL), were tested, which showed large correlations between
each other. QTL mapping was carried out using the composite interval mapping method based a genetic map covering 1882.1 cM
of tobacco genome with 611 SSR markers in 24 linkage groups. A total of 69 QTLs were detected, most of which exhibited small
additive effects. Four QTLs associated with PH, IL, and WWL exhibited much higher phenotypic contributions than other QTLs,
and a single locus explained 15–20% of the phenotypic variation. Many small regions were detected to harbor two or more closely
linked QTLs for different traits. This result was in agreement with the correlation analysis of phenotypic traits.
Keywords: Flue-cured tobacco (Nicotiana tabacum L.); Quantitative trait locus (QTL); Genetic mapping; Doubled haploid (DH);
Yield
烤烟 6个农艺性状的 QTL定位
童治军 1,2 焦芳婵 2 吴兴富 2 王丰青 1 陈学军 2 李绪英 2,3
高玉龙 2 张谊寒 2 肖炳光 2,* 吴为人 1,4,*
1浙江大学农业与生物技术学院，浙江杭州 310058；2云南省烟草农业科学研究院，云南玉溪 653100；3云南农业大学农业与生物技
术学院，云南昆明 650201；4福建农林大学生命科学学院，福建福州 350002
摘 要: 由于烟草的分子标记开发和遗传图谱构建十分困难，迄今烟草中有关数量性状基因座(QTL)的定位研究仍非
常有限。本研究利用一个由 207个株系组成的烤烟 DH群体及基于该群体所构建的含有 24个连锁群、611个 SSR标
记，总长为 1 882.1 cM的遗传图谱，采用复合区间作图方法，对株高(PH)、茎围(SG)、节距(IL)、叶片数(LN)、最大
腰叶长(LWL)和最大腰叶宽(WWL) 6个与叶片产量有关的农艺性状进行 QTL定位分析。共检测到 69个 QTL，大部
分 QTL的效应值较小，仅有 4个具有较大的效应值，可解释大约 15%~20%的表型变异。6个性状之间大多彼此相关。
与此相符，在基因组中发现存在许多小区域，每个区域包含 2个或 2个以上紧密连锁的不同性状的 QTL。
关键词: 烤烟; 数量性状基因座; 遗传作图; 加倍单倍体; 产量
Tobacco (Nicotiana tabacum L.) is an important eco-
nomic crop grown widely throughout the temperate agri-
cultural zones in the world. To create cultivars with
higher leaf yield is one of the most important objectives
in tobacco breeding. Tobacco leaf yield is a complex trait,
which is either directly or indirectly influenced by many
other agronomic traits such as plant height, leaf number
and leaf size [1-3]. Therefore, to improve leaf yield, it is
necessary to improve other related traits. Most of the
yield-related traits are quantitative traits controlled by
1408 作 物 学 报 第 38卷

polygenes. Measuring and selecting quantitative traits are
not difficult in general. However, genetically improving
quantitative traits through phenotypic selection is usually
difficult due to low heritability and small cumulative
effect of numerous genes or quantitative trait loci (QTLs)
involved [4]. During the last two decades, molecular
markers have been extensively utilized to map QTLs in
plants. QTL mapping has become a powerful tool in the
genetic research of quantitative traits, which can greatly
promote our ability to get deep insight into the genetic
basis of quantitative traits and, meanwhile, facilitate
marker-assisted selection (MAS) so as to increase the
efficiency of crop breeding [5].
Tobacco is an allotetraploid species (2n=48) with rela-
tively low genetic diversity and polymorphism level [6-11].
Hence, developing molecular markers and constructing
genetic maps in tobacco have proven to be very difficult.
For a long period of time, genetic researches in tobacco
were mainly based on non-specific markers including
randomly amplified polymorphic DNA (RAPD), ampli-
fied fragment length polymorphism (AFLP), inter-simple
sequence repeat (ISSR), sequence-related amplified
polymorphism (SRAP), and sequence-specific amplified
polymorphism (SSAP) etc. [7-8,11-16]. It is only recently
that significant progress has been achieved in this area
owing to the availability of partial tobacco genome se-
quence data released by the Tobacco Genome Initiative
(TGI) program (http://www.tobaccogenome.org/), from
which more than 5 000 microsatellite or simple sequence
repeat (SSR) markers have been developed and a
high-density genetic map consisting of more than 2 000
SSR markers has been constructed [17].
Because of the difficulty in molecular marker deve-
lopment and genetic map construction, QTL mapping
studies in tobacco have been much lagged behind in
comparison with many other important crops. Up to now,
only a few QTL mapping studies have been reported in
tobacco [15-16, 18-25]. These studies covered a wide range of
important traits, including resistance to diseases (e.g.
bacterial wilt, black shank, and blue mold) [16, 18-19, 21-22],
agronomic performance (e.g. flowering time and various
morphological traits) [23-24] and leaf quality (e.g. chemical
components and smoke properties) [15, 20, 25]. However,
these studies were all conducted based on partial genetic
maps with very low genome coverage and therefore
could miss many QTLs not covered by the maps. In addi-
tion, these maps all consisted of mainly non-specific
markers, making it very difficult to compare QTLs de-
tected in different studies or to transfer the information of
QTL mapping results in one population to other popula-
tions.
We have constructed a genetic map of SSR markers
with relatively high genome coverage based on a doubled
haploid population in flue-cured tobacco. In this study,
we employed this population and the genetic map to de-
tect QTLs underlying six agronomic traits related to leaf
yield, aiming to reveal the genetic bases of these traits
and facilitate MAS in tobacco breeding.
1 Materials and Methods
1.1 Plant materials
A population containing 207 doubled haploid (DH)
lines of tobacco (Nicotiana tabacum L.) was developed
by Yunnan Academy of Tobacco Agricultural Sciences in
2010 through anther culture of the F1 between two
flue-cured tobacco cultivars, Honghua Dajinyuan (HD,
the female parent) and Hicks Broad Leaf (HBL, the male
parent) [26]. HD is a landrace in southwest China, while
HBL is a widely-cultivated variety in the world. Both
varieties are high yielding and have good processing
quality.
1.2 Field experiments
The DH lines, F1 and two parents (HD and HBL) were
planted under a completely randomized block design
with two replicates in the field of the farm of South To-
bacco Breeding Research Center of China, Yuxi, Yunnan
province, China in 2011. In each replicate, each line was
planted in two rows with 15 plants per row. The
plant-to-plant spaces between and within rows were 100
cm and 50 cm, respectively.
1.3 Trait investigation and statistical analysis
Five plants of each line in a replicate were randomly
selected for phenotyping ~7 days after topping. Six traits
related to leaf yield were measured, including plant
height (PH, from soil surface to stem apex), stem girth
(SG, at plant waist, the position of 1/3 of the plant height
from the soil surface), internode length (IL, at plant
waist), leaf number (LN), length of the largest waist leaf
(LWL) and width of the largest waist leaf (WWL). The
phenotypic data were analyzed by analysis of variance
(ANOVA). Pearson correlations between the traits were
computed using the program PROC CORR of SAS (SAS
Institute Inc. 2004).
1.4 DNA extraction and SSR marker analysis
Genomic DNA isolation and purification was per-
formed using the CTAB method [27] with minor modifica-
tions. A total of 10 005 SSR markers, including 5 119
PT-series markers published by Bindler et al. (2011) [17]
and 4 886 TM/TME-series markers developed by our-
selves were tested for polymorphisms between the two
parents (HD and HBL) of the mapping population and
polymorphic markers were used to genotype the mapping
population. PCR amplification of SSR markers was per-
formed in a 20 μL reaction mixture containing 2.0 μL of
1× buffer (10 mmol L–1 Tris-Cl, pH 8.4, 50 mmol L–1
KCl, 1.5 mmol L–1 MgCl2), 200 μmol L–1 of each dNTP
(TaKaRa Biotechnology Co. Ltd., Dalian, China), 0.5
μmol L–1 each of forward and reverse primers (TaKaRa),
and 1.0 U of rTaq polymerase (TaKaRa). The PCR pro-
gram was denaturation at 94°C for 5 min, followed by 30
cycles of denaturation at 94°C for 30 s, annealing at a
第 8期 童治军等: 烤烟 6个农艺性状的 QTL定位 1409


given temperature (depending on the primer pairs) for 30
s and elongation at 72°C for 30 s, and then a final elonga-
tion at 72°C for 10 min. Amplification products were
separated on 6% non-denaturing polyacrylamide gel by
electrophoresis (non-denaturing PAGE; 220 V, 3.5 h) and
DNA bands were visualized by silver staining following
the method of Xu et al. [28] with minor modifications.
1.5 QTL analysis
Based on the SSR marker genotype data of the map-
ping population, in which only a very small proportion
(–0.8%) of data were missing, we constructed a genetic
map using the program JoinMap version 4.0[29]. After
discarding a few markers with significant segregation
distortion, the constructed map consisted of 611 SSR
marker loci distributed on 24 LGs, covering a total length
of 1 882.1 cM with an average distance of 3.08 cM be-
tween adjacent markers (details about the construction of
this map are to be published in another paper). With this
map, QTL mapping was performed for the six traits in-
vestigated using the method of composite interval map-
ping (CIM)[30] implemented by the computer program
Windows QTL Cartographer 2.5 (http://statgen.ncsu.edu/
qtlcart/WQTL Cart.htm). The Model 6 provided by the
program was chosen for the analysis, with parameters set
as: Number of control markers = 30; Window size = 10.0
cM; and Step size = 1.0 cM. The LOD threshold at an
overall significant level of 0.05 for each trait was esti-
mated by permutation tests [31] with 1 000 replicates.
QTLs were named according to the nomenclature pro-
posed by McCouch et al. [32], namely prefix q + trait
name abbreviation + LG number + QTL number (if there
are multiple QTLs of the trait on the LG). The genetic
map was drawn using the software MapChart 2.22 [33].
2 Results
2.1 Performance of the target traits
The performances of the six target traits are summarized
in Table 1. It can be seen that the two parents showed
significant difference in all the traits except for WWL,
with HD always having the higher trait value than HBL.
All the traits exhibited a wide range of continuous varia-
tion with obvious transgressive segregation and a mean
value close to the mid-parent in the DH population. The
kurtosis and skewness values of the traits were all less
than 1, suggesting that they all approximately followed a
normal distribution in the DH population. In addition, all
the traits exhibited a medium broad sense heritability in
this experiment, ranging from 40.77% to 58.42% (Table
1). These results indicate that the target traits are all
quantitatively inherited. Besides, significant correlations
were found between the traits in the DH population (Ta-
ble 2). Each trait was significantly correlated with three
or four other traits. All the significant correlations were
positive except for that between LN and IL. The correla-
tion between LWL and WWL was very high (up to 0.8),
suggesting that the leaf shape (indicated by the ratio of
LWL/WWL) did not vary greatly in this population.

Table 1 Performances of the six traits in the DH population and the parental lines
Parents DH population
Trait a
HD HBL Mean±SD Range Kurtosis Skewness Heritability (%)
PH (cm) 118.50 106.35* 113.35±17.60 64.75–171.50 0.50 0.14 48.85
LN 26.00 17.00** 22.00±3.00 14.00–31.00 0.48 –0.04 56.54
SG (cm) 9.81 7.03* 8.62±1.37 6.75–10.03 0.86 –0.65 44.38
IL (cm) 4.41 3.05 * 3.84±0.74 1.78–5.83 0.54 0.01 40.77
LWL (cm) 71.79 55.03** 61.29±7.37 37.50–79.83 0.18 –0.19 58.42
WWL (cm) 23.53 22.58ns 22.06±3.49 12.50–30.17 –0.29 0.07 43.60
a Traits abbreviations: PH for plant height; LN for leaf number; SG for stem girth; IL for internode loength; LWL for length of the largest waist
leaf; WWL for width of the largest waist leaf. *, **: significantly different between parents at 0.05 and 0.01 levels, respectively. ns: not significant.

Table 2 Correlations between the six traits in the DH population
PH LN SG IL LWL
LN 0.401*
SG 0.109 0.316*
IL 0.406* −0.247* 0.245*
LWL 0.200* 0.106 0.403* 0.191
WWL 0.122 −0.082 0.308* 0.273* 0.801**
*, **Significant at 0.05 and 0.01 levels, respectively. Traits abbrevia-
tions are the same as in Table 1.
2.2 QTL mapping
A total of 69 QTLs were mapped on 20 LGs (exclud-
ing LG9, 16, 21, and 23), of which 13 were for PH, 13 for
IL, 9 for SG, 10 for LN, 12 for LWL, and 12 for WWL
(Table 3 and Fig. 1). LG14 and LG17 contained the most
QTLs (each containing 11 QTLs), followed by LG7 (8
QTLs) and LG22 (7 QTLs), while LG11, LG12, LG13,
LG15, LG18 and LG20 contained the fewest QTLs (only
1 QTL on each). In each trait, most of the alleles from the
parent HD displayed positive effects (i.e., increasing the
trait value), but there were always some HD alleles act-
1410 作 物 学 报 第 38卷

Table 3 QTLs detected for six traits in DH population
Trait QTL Linkage group Nearest marker Position (cM) LODa Ab R2 (%)c
qPH3 LG3 TM10840 69.2 9.45 –5.16 6.52
qPH4 LG4 TM10481 14.6 7.81 4.18 4.40
qPH6 LG6 TM10433 55.8 4.71 3.88 3.15
qPH7 LG7 PT30005 18.8 6.56 3.73 3.76
qPH10 LG10 TM10603 67.3 3.26 2.83 1.95
qPH13 LG13 PT54964 10.5 8.69 4.21 4.95
qPH14-1 LG14 PT53269 21.3 4.79 3.48 3.20
qPH14-2 LG14 PT50230 34.5 5.12 3.38 2.79
qPH14-3 LG14 TM11233 91.2 4.27 2.96 2.32
qPH17-1 LG17 PT61337 58.5 8.33 5.68 4.73
qPH17-2 LG17 TM20580 88.6 25.26 –14.83 20.30
qPH20 LG20 PT52111 41.9 3.41 –3.98 2.10
PH
qPH24 LG24 PT60487 52.4 4.77 –3.13 2.60
qIL1 LG1 PT50909 0.0 3.71 0.12 2.35
qIL5 LG5 TM20756 36.6 5.71 0.22 4.09
qIL7-1 LG7 TM23088 1.0 4.01 0.21 3.80
qIL7-2 LG7 PT51802 24.3 6.82 –0.28 5.98
qIL11 LG11 TM20261 30.5 4.01 0.14 2.77
qIL12 LG12 PT61291 14.7 4.14 0.16 2.82
qIL14-1 LG14 TM20865 44.6 4.61 0.17 3.24
qIL14-2 LG14 PT55227 76.2 4.32 0.15 2.77
qIL17-1 LG17 PT61337 58.5 24.60 0.46 20.64
qIL17-2 LG17 TM20757d 86.8 18.36 –0.39 16.27
qIL19 LG19 PT54889 73.6 3.74 –0.21 2.91
qIL22-1 LG22 TM10924 51.6 4.53 –0.15 3.02
IL
qIL22-2 LG22 PT55394 73.1 6.04 –0.19 4.99
qSG1 LG1 TM20480 45.9 8.04 0.39 6.80
qSG3 LG3 PT53847 55.5 5.06 –0.41 5.04
qSG7 LG7 PT52892 33.0 8.87 0.48 7.59
qSG8-1 LG8 TM25276 21.9 8.06 –0.51 6.82
qSG8-2 LG8 PT61104a 44.1 7.26 0.52 7.17
qSG10 LG10 TM11104 32.6 4.16 –0.36 3.96
qSG14 LG14 PT60551 83.6 4.73 0.29 3.81
qSG17 LG17 PT54081 78.5 6.05 0.35 5.60
SG
qSG22 LG22 TM10712 16.7 5.56 0.40 4.67
qLN3 LG3 TM10470 91.3 4.45 –0.56 3.27
qLN4-1 LG4 TM22123 23.8 6.15 0.66 4.17
qLN4-2 LG4 TM10315a 33.6 4.42 0.54 3.07
qLN6 LG6 PT60466 27.3 5.20 0.62 4.49
qLN7 LG7 TM10846 25.1 14.27 0.99 11.28
qLN8 LG8 PT51340 21.4 9.27 –0.84 7.11
qLN14 LG14 TM25249 6.3 6.98 0.71 5.41
qLN17 LG17 PT50717 55.5 11.84 –1.04 8.55
qLN19 LG19 PT61218 31.2 5.54 0.67 5.03
LN
qLN22 LG22 PT60728 5.1 10.01 0.86 7.10

第 8期 童治军等: 烤烟 6个农艺性状的 QTL定位 1411


(continued)
Trait QTL Linkage group Nearest marker Position (cM) LODa Ab R2(%)c

qWWL4 LG4 TM10315a 33.7 3.87 0.54 2.05
qWWL5 LG5 TM10857 4.7 6.94 0.83 3.78
qWWL7 LG7 PT51305 43.8 6.44 –0.70 3.60
qWWL14-1 LG14 PT60379 37.9 4.07 0.68 2.14
qWWL14-2 LG14 PT60647 62.0 4.08 –0.73 2.25
qWWL14-3 LG14 PT60551 83.6 4.54 0.95 2.47
qWWL17-1 LG17 PT60516 17.2 3.49 0.67 2.57
qWWL17-2 LG17 PT61337 58.7 23.17 1.71 15.29
qWWL17-3 LG17 TM20580 90.5 7.48 0.90 4.81
qWWL22-1 LG22 PT51132 46.4 7.38 0.97 4.47
qWWL22-2 LG22 PT55394 73.1 4.39 –0.77 2.91
WWL
qWWL24 LG24 PT60487 52.4 7.52 –0.77 4.34
qLWL1 LG1 PT50868 49.6 7.74 2.04 5.18
qLWL4 LG4 PT50554 20.9 5.71 1.58 3.68
qLWL6 LG6 PT52570 53.4 3.6 1.26 2.36
qLWL7-1 LG7 TM23088 0.0 4.41 1.46 2.32
qLWL7-2 LG7 PT51380 15.6 3.78 –1.42 1.97
qLWL14 LG14 PT54200 41.0 6.21 1.52 3.32
qLWL15 LG15 PT54772 70.4 7.04 –1.60 3.82
qLWL17 LG17 TM11215 65.3 14.33 3.01 9.64
qLWL18 LG18 PT54096 0.0 4.10 –1.14 2.14
qLWL2-1 LG2 PT53714 75.8 5.00 1.50 2.65
qLWL2-2 LG2 PT61640 110.1 6.12 –1.70 3.28
LWL
qLWL22 LG22 TM10861b 41.2 3.29 1.23 1.86
a LOD thresholds at the genome-wise significance level of 0.05 estimated by 1000 permutation tests were: 3.3 for PH, 3.5 for IL, 3.4 for LN, 3.2
for SG, WWL and LWL. b A: additive effect, of which the positive/negative sign indicates the action direction of the allele from parent HD. c R2: the
proportion of phenotypic variance explained by the QTL. Abbreviations of traits are the same as Table 1.

ing in the negative direction (Table 3). This is consistent
with the results that HD always had higher trait values
than HBL (though not statistically significant in WWL)
and there was transgressive segregation in the DH popu-
lation in all the traits. Most of the QTLs exhibited small
effects. The proportion of phenotypic variance explained
by a QTL (R2 value) was only 5.18% on ave- rage, with
~70% and ~90% of the R2 values being <5% and <8%,
respectively (Table 3).
Nevertheless, there were four QTLs with relatively
large effects found in three traits, namely, qIL17-1
(R2=20.64%), qPH17-2 (20.30%), qIL17-2 (16.27%) and
qWWL17-2 (15.29%; Table 3). Interestingly, these four
major QTLs were all located on LG17, with qPH17-2
and qIL17-2 being mapped together (only 1.8 cM apart)
at a location near the lower end of LG17, whereas
qIL17-1 and qWWL17-2 being mapped together (only 0.2
cM apart) at a location near the central part of LG17 (Ta-
ble 3 and Fig. 1); and the alleles of qPH17-2 and qIL17-2
from parent HD both showed the negative acting direc-
tion, while those of qIL17-1 and qWWL17-2 both acted in
the positive direction (Table 3). Besides, there were two
additional QTLs (qPH17-1 and qLN17) and one addi-
tional QTL (qWWL17-3) detected at the location of
qIL17-1 and qWWL17-2 and that of qPH17-2 and
qIL17-2, respectively (Table 3 and Fig. 1). These two
clusters of QTLs (four in the former and three in the lat-
ter) only occupied a small region spanning 3.2 and 3.7
cM in the map, respectively.
Similar to the above two cases, there were many other
small regions (≤ 3.7 cM) harboring two or more QTLs of
different traits, such as {qSG1, qLWL1} on LG1,
{qLN4-1, qLWL4} and {qLN4-2, qWWL4} on LG4,
{qLWL6, qPH6} on LG6, {qLWL7-1, qIL7-1},
{qLWL7-2, qPH7} and {qIL7-2, qLN7} on LG7, {qLN8,
qSG8-1} on LG8, {qPH14-2, qWWL14-1}, { qWWL14-1,
qLWL14}, {qLWL14, qIL14-1} and {qSG14, qWWL14-3}
on LG14, {qIL22-2, qWWL22-2} on LG22 and {qPH24,
qWWL24} on LG24 (Fig. 1 and Table 3). This was con-
sistent with the result that significant correlations existed
between most of the traits studied.
3 Discussions
It has been mentioned above that QTL mapping stu-
dies in tobacco have been still very limited and the studies
1412 作 物 学 报 第 38卷



(to be continued)
第 8期 童治军等: 烤烟 6个农艺性状的 QTL定位 1413




Fig. 1 Genetic map of flue-cured tobacco showing the positions of QTLs (indicated by black solid squares) of six yield-related traits
estimated by composite interval mapping

reported so far were all performed based on partial ge-
netic maps that mainly comprised non-specific markers
and only covered a small proportion of the tobacco ge-
nome [15-16, 18-25]. Such low-coverage maps can greatly
reduce the ability of detecting QTLs because any QTL
outside the regions covered by the maps is definitely un-
detectable. In this study, we conducted QTL mapping for
six yield-related traits in flue-cured tobacco using a ge-
netic map with much higher genome coverage consisting
of more than 600 SSR markers. To our knowledge, this is
the first case of QTL mapping study having been re-
ported based on a map with such a higher degree of ge-
nome coverage in tobacco.
Although our map is still far from being a complete
one, the merit resulting from its higher genome coverage
for QTL mapping is very significant in comparison with
previous QTL mapping studies in tobacco. For example,
in the study of Julio et al. [16], a genetic map comprising a
total of 141 ISSR, AFLP, and SSAP markers distributed
on 18 LGs and covering a total length of 707.6 cM was
used for QTL mapping. They analyzed 59 traits, but only
detected 75 QTLs in total, with only 1.27 QTLs per trait
on average. In contrast, we detected a total of 69 QTLs in
merely six traits, with 11.5 QTLs per trait on average.
Although there could be other factors influencing the
number of detected QTLs such as population size (our
population size was 207 DH lines, while theirs was 114
RI lines) and trait heritability (many traits they studied
might have low heritabilities), map length should be a
key determinant factor. It is easy to find that the map of
Julio et al. [16] was only a little more than 1/3 of our map
in length. Hence, it can be inferred that most QTLs in the
genome were not covered by their map and therefore
could not be detected. In addition, our map has another
advantage, namely, it consists of SSR markers instead of
non-specific markers. This makes the QTL mapping re-
sults of this study easy to be compared and more useful
for relevant genetic researches and molecular breeding.
In this study, the six yield-related traits analyzed were
largely correlated with each other (Table 2). Consistent
with this, the QTL mapping analysis revealed that there
were many small regions harboring two or more QTLs of
1414 作 物 学 报 第 38卷

different traits in the tobacco genome (Fig. 1). Since the
QTLs in the same region were located so closely, they
were possibly a single QTL with pleiotropic effects; in
other words, they might be attributed to the same gene(s)
affecting multiple traits. Even if these closely-linked
QTLs were really different, they still tended to be inher-
ited as a single QTL. Hence, we can take these small
regions as individual genetic units with pleiotropic ef-
fects. Obviously, selection on a pleiotropic region will
influence two or more traits. Therefore, in tobacco
breeding for high yield, the different yield-related traits
should be considered jointly and attention should be paid
to those pleiotropic regions.
QTL mapping studies for the traits same as or similar
to those analyzed in this study have been also reported
before. Using a DH population of flue-cured tobacco
consisting of 137 lines and a genetic map comprising a
total of 169 RAPD and ISSR markers, Xiao et al. [25]
mapped two, four, two, two, and three QTLs showing
significant additive effects for plant height, internode
length, leaf number, leaf length and leaf width, respec-
tively. More recently, Cai et al. [24] mapped four QTLs for
plant height, stalk circumference, distance between nodes
and length of central leaf in burley tobacco with one for
each trait, respectively, using a DH population consisting
of 94 lines and a genetic map comprising a total of 118
AFLP and SRAP markers. Since the maps used in these
two studies only contained non-specific markers, the
QTLs mapped cannot be compared with those mapped in
this study. However, these two studies also revealed that
QTLs of these correlated traits tend to be mapped to-
gether, which was consistent with the results of this
study.
4 Conclusions
We totally detected 69 QTLs underlying six agro-
nomic traits on 20 LGs of tobacco. Several QTLs with
relatively large effects and many small chromosomal
regions harboring clusters of QTLs underlying different
traits were identified. The results will facilitate MAS for
improving related traits in tobacco breeding programs.
References
[1] Legg P D, Collins G B. Genetic parameters in a Ky 14 × Ky Ex
42 burley population of Nicotiana tabacum L. Theor Appl Genet,
1975, 45: 264–267
[2] White F H, Pandeya R S, Dirks V A. Correlation studies among
and between agronomic, chemical, physical and smoke charac-
teristics in flue-cured tobacco (Nicotiana tabaccum L.). Can J
Plant Sci, 1979, 59:111–120
[3] Honarnejed R, Shoai-Deylami M. Gene effect, combining ability
and correlation of characterstics in F2 populations of burley to-
bacco. J Sci Technol Agric Nat Resour, 2004, 8: 135–148
[4] Xiao B G, Zhu J, Lu X P, Bai Y F, Li Y P. Analysis on genetic
contribution of agronomic traits to total sugar in flue-cured to-
bacco (Nicotiana tabacum L.). Field Crops Res, 2007, 102:
98–103
[5] Mohan M, Nair S, Bhagwat A, Krishna T G, Yano M, Bhatia C R,
Sasaki T. Genome mapping, molecular markers and marker-
assisted selection in crop plants. Mol Breed, 1997, 3: 87–103
[6] Ren N, Timko M P. ALFP analysis of genetic polymorphism and
evolutionary relationships among cultivated and wild Nicotiana
species. Genome, 2001, 44: 559–571
[7] Rossi L, Bindler G, Pijnenburg H, Isaac P G, Giraud-Henri I,
Mahe M, Orvain C, Gadani F. Potential of molecular marker
analysis for variety identification in processed tobacco. Plant
Varieties Seeds, 2001, 14: 89–101
[8] Moon H S, Nicholson J S, Lewis R S. Use of transferable Nico-
tiana tabacum L. microsatellite markers for investigating genetic
diversity in the genus Nicotiana. Genome, 2008, 51: 547–559
[9] Moon H S, Nicholson J S, Heineman A, Lion K, der Hoeven R V,
Hayes A J, Lewis R S. Changes in genetic diversity of U.S. flue-
cured tobacco germplasm over seven decades of cultivar devel-
opment. Crop Sci, 2009, 49: 498–506
[10] Moon H S, Nifong J M, Nicholson J S, Heineman A, Lion K, der
Hoeven R V, Hayes A J, Lewis R S. Microsatellite-based analysis
of tobacco (Nicotiana tabacum L.) genetic resources. Crop Sci,
2009, 49: 2149–2157
[11] Bai D, Reeleder R, Brandle J E. Identification of two RAPD
markers tightly linked with the Nicotiana debneyi gene for resis-
tance to black root of tobacco. Theor Appl Genet, 1995, 91:
1184–1189
[12] Yi H Y, Rufty R C, Wernsman E A. Mapping the root-knot nema-
tode resistance gene (Rk) in tobacco with RAPD markers. Plant
Dis, 1998, 82: 1319–1322
[13] Noguchi S, Tajima T, Yamamoto Y, Ohno T, Kubo T. Deletion of
a large genomic segment in tobacco varieties that are resistant to
potato virus Y (PVY). Mol Gen Genet, 1999, 262: 822–829
[14] Johnson E S, Wolff M F, Wernsmann E A. Marker assisted selec-
tion for resistance to black shank disease in tobacco. Plant Dis,
2002, 12: 1303–1309
[15] Julio E, Denoyes R B, Verrier J L, de Borne F D. Detection of
QTLs linked to leaf and smoke properties in Nicotiana tabacum
based on a study of 114 recombinant inbred lines. Mol Breed,
2006, 18: 69–91
[16] Julio E, Verrier J L, de Borne F D. Development of SCAR mark-
ers linked to three disease resistances based on AFLP within
Nicotiana tabacum L. Theor Appl Genet, 2006, 112: 335–346
[17] Bindler G, Plieske J, Bakaher N, Gunduz I, Ivanov N, der Hoeven
R V, Ganal M, Donini P. A high density genetic map of tobacco
(Nicotiana tabacum L.) obtained from large scale microsatellite
marker development. Theor Appl Genet, 2011, 123: 219–230
[18] Nishi T, Tajima T, Noguchi S, Ajisaka H, Negishi H. Identifica-
tion of DNA markers of tobacco linked to bacterial wilt resistance.
Theor Appl Genet, 2003, 106: 765–770
[19] Vontimitta V, Lewis R S. Mapping of quantitative trait loci af-
第 8期 童治军等: 烤烟 6个农艺性状的 QTL定位 1415


fecting resistance to Phytophthora nicotianae in tobacco (Nico-
tiana tabacum L.) line Beihart-1000. Mol Breed, 2012, 29: 89–98
[20] Vontimitta V, David A, Danehower, Steede T, Moon H S, Lewis R
S. Analysis of a Nicotiana tabacum L. genomic region controlling
two leaf surface chemistry traits. Agric Food Chem, 2010, 58:
294–300
[21] Vontimitta V, Lewis R S. Growth chamber evaluation of a to-
bacco ‘Beinhart 1000’ × ‘Hicks’ mapping population for quanti-
tative trait loci affecting resistance to multiple races of Phy-
tophthora nicotianae. Crop Sci, 2012, 52: 91–98
[22] Milla S R, Levin J S, Lewis R S, Rufty R C. RAPD and SCAR
markers linked to an introgressed gene conditioning resistance to
Peronospora tabacina D.B. Adam. in tobacco. Crop Sci, 2005, 45:
2346–2354
[23] Lewis R S, Milla S R, Kernodle S P. Analysis of an introgressed
Nicotiana tomentosa genomic region affecting leaf number and
correlated traits in Nicotiana tabacum. Theor Appl Genet, 2007,
114: 841–854
[24] Cai C-C(蔡长春), Chai L-G(柴利广), Wang Y(王毅), Xu F-S(徐
芳森), Zhang J-J(张俊杰), Lin G-P(林国平). Construction of
genetic linkage map of burley tobacco (Nicotiana tabacum L.)
and genetic dissection of partial traits. Acta Agron Sin (作物学
报), 2009, 35(9): 1646–1654 (in Chinese with English abstract)
[25] Xiao B-G(肖炳光), Lu X-P(卢秀萍), Jiao F-C(焦芳蝉), Li
Y-P(李永平), Sun Y-H(孙玉合), Guo Z-K(郭兆奎). Preliminary
QTL analysis of several chemical components in flue-cured to-
bacco (Nicotiana tabacum L.). Acta Agron Sin (作物学报), 2008,
34(10): 1762–1769 (in Chinese with English absract)
[26] Chen X-J(陈学军), Peng S-Y(彭双玉), Luo J-R(罗建蓉), Yang
Y-M(杨彦明), Xiao B-G(肖炳光). Culture of regenerated seed-
lings from anthers and construction of DH populations of six
cross combinations of Nicotiana tabacum. J Plant Resour & En-
viron (植物资源与环境学报), 2011, 20(1): 65–68 (in Chinese
with English abstract)
[27] Murry H G, Thomspon W F. Rapid isolation of weight DNA.
Nucl Acids Res, 1980, 8: 4321–4322
[28] Xu S-B(许绍斌), Tao Y-F(陶玉芬), Yang Z-Q(杨昭庆), Chu J(褚
嘉). A simple and rapid methods used for silver staining and gel
preservation. Heredtas (遗传), 2002, 24(3): 335–336 (in Chinese
wit English abstract)
[29] Van Ooijen J W. JoinMap 4.0, Software for the calculation of ge-
netic linkage maps in experimental populations. Kyazma B V,
Wageningen
[30] Zeng Z B. Precision mapping of quantitative trait loci. Genetics,
1994, 136: 1457–1468
[31] Doerge R W, Churchill G A. Permutation tests for multiple loci
affecting a quantitative character. Genetics, 1996, 142: 285–294
[32] McCouch S R, Cho Y G, Yano M, Paul E, Blinstrub M, Mor-
ishima H, Kinoshita T. Report on QTL nomenclature. Rice Genet
Newsl, 14: 11–13
[33] Voorrips R E. MapChart: software for the graphical presentation
of linkage maps and QTLs. J Hered, 2002, 93: 77–78