rapd markers for the identification of yield traits in tomatoes under heat stress via bulked...
TRANSCRIPT
RAPD markers for the identification of yield traits in tomatoes underheat stress via bulked segregant analysis
KUAN-HUNG LIN1, HSIAO-FENG LO1, SHAO-PEI LEE2, C. GEORGE KUO3, JEN-TZU CHEN3 and
WEI-LUN YEH2
1Department of Horticulture, Chinese Culture University, Taipei, Taiwan2Graduate Institute of Biotechnology, Chinese Culture University, Taipei, Taiwan3Physiology Unit, Asian Vegetable Research and Development Center, Tainan, Taiwan
Lin, K.H., Lo, H.F., Lee, S.P., Kuo, C.G., Chen, J.T. and Yeh, W.L. 2006. RAPD markers for the identification of yield traits
in tomatoes under heat stress via bulked segregant analysis. * Hereditas 143: 142�154. Lund, Sweden. eISSN 1601-5223.
Received December 20, 2005. Accepted May 29, 2006
Tomato production is limited to a large extent by climates with high temperatures. Yield-related traits in tomatoes are
generally thought to be quantitative inherited traits that are significantly affected by variation in environmental factors.
Breeding for heat tolerance is restricted due to the complexity of the traits. The objective of this study was to identify random
amplified polymorphic DNA (RAPD) markers linked to heat-tolerance traits in tomatoes under heat stress. Forty-three F7
recombinant inbred lines (RILs) derived from a wild cross between CL 5915 (heat-tolerant) and L4422 (heat-sensitive) were
obtained and scored for flower number, fruit number, fruit set, fruit weight and yield under screen house conditions during
the hot (summer) season of 2003. The distributions of average fruit weight, fruit number, fruit set and yield in the F7
population were strongly skewed towards heat susceptibility, characteristic of L4422. Significant positive correlation was
observed between fruit weight and yield, and between fruit number, fruit set and yield. However, the increase in yield and
fruit set by selecting for large flower number may be fairly minor due to non-significant correlation between these traits.
Selecting for flower number may not be a useful tool for improving yield under heat stress. A total of 200 RAPD primers
were screened, among which 14 were identified as associated with heat-tolerance using bulk segregant analysis (BSA) based
on the F7 population grown in a heat-stressed environment. Some RAPD markers were unique to one specific trait, and
others were linked to two traits. All markers for heat tolerance related traits had positive gene effects as a result of the
contribution of the CL5915 gene that bolstered these traits. One hundred F2 plants derived from the same parents
(CL5915�/L4422) were grown in the same location during the summer of 2004 to test for the stability and reliability of the 14
markers identified. Selection for the desired heat-tolerance genotypes corresponded well with targeting heat tolerance traits
using the selected heat tolerance RAPD markers identified. Marker-assisted selection (MAS) for heat tolerance may be
efficiently conducted by selecting individuals that contain high fruit number, high fruit weight, and high yield markers (P06,
X01, D06 and D11), which would thus facilitate conventional breeding using CL5915 as a donor parent.
Kuan-Hung Lin, Department of Horticulture, Chinese Culture University, Taipei 111, Taiwan, ROC. E-mail: rlin@faculty.
pccu.edu.tw
The cultivated tomato (Lycopersicon esculentum Mill.)
is one of the world’s most important crops, due to the
high value of its fruit in terms of versatility, making it
suitable for both fresh consumption and in numerous
types of processed products (RICK 1988). Tomatoes
are an excellent source of vitamins A, C, and E as well
as lycopene. In tropical Asia, it is an important cash
crop for small-scale farmers. Although tomato plants
can grow under a wide range of temperatures, optimal
fruit set and fruit weight are limited to a somewhat
narrower range. When the average maximal day and
minimal night temperatures exceed 328C and 218Crespectively, the fruit set is low, which leads to a
markedly reduced yield in tomato production
(AVRDC 1986). High temperatures are known to
limit fruit set in tomatoes due to a simultaneously
and/or sequentially impaired series of reproductive
processes, i.e. pollen production and development,
ovule development, pollination, pollen grain germina-
tion, pollen tube growth, fertilization and fruit initia-
tion (LOHAR and PEET 1998; SATO and PEET 2004).
Temperatures above 338C reduce photosynthesis,
which seems to be the result of stomatal and meso-
phyll resistance. Increased mesophyll resistance is
dependent on the enzymatic response of a CO2 fixing
enzyme, ribulose-1, 5-biphosphate carboxylase insti-
gated by a change in temperature (BAR-TSUE et al.
1985). However, low fruit set and poor development of
the tomato at high temperatures are primarily the
result of a reduction in carbon export from the leaf,
due to callose formation in the leaf petiole, and the
inability of red productive organs to import assimilates
during the early stage of flower development (DINAR
and RUDICH 1985). Fruit set and development are
associated with endogenous plant hormones produced
by pollen, style tissue, or seeds during the normal
processes of pollination, fertilization and seed forma-
tion (KUO and TSAI 1984). For instance, high
Hereditas 143: 142�154 (2006)
gibberellin levels have been found in inflorescences
undergoing flower abortion and high levels of abscisic
acid inhibit tomato pollen germination. In addition to
the hormonal responsiveness of specific tissues, exo-
genous application of natural plant hormones and
synthetic growth regulators has produced improved
fruit set and yield where these have otherwise failed to
develop under high temperature conditions; this
measure has also accelerated ripening for an easy
harvest (KUO et al. 1986).
Most traits of economic importance in crop plants
are quantitative in nature, each controlled by many
genes. Previous studies (AVRDC 1997; LIPPMANA and
TANKSLEY 2001; SALIBA et al. 2001; KNAAP and
TANKSLEY 2003; CAUSSE et al. 2004; LECOMTE et al.
2004) have suggested that the fruit set, fruit weight,
fruit shape and fruit size of tomatoes are linked to
polygene traits. Plant response to heat stress is a
complex phenotypic and physiological phenomenon
that is highly influenced by environmental factors. The
intensity of heat treatment (the temperature level and
duration of exposure) is of particular importance for
the effective evaluation of the plants’ heat tolerance.
The coordination of the process and the number of
genes that regulate it has yet to be unraveled. Recently,
there has been greater interest in acquiring an under-
standing of and being more capable of modeling the
effects of high temperatures on plant growth and
development. This may in part be due to the predicted
increase in global temperatures. The effect of global
climate change on food availability has also sparked
interest in the scholarly community.
To minimize high temperature-related yield losses,
we need to understand the underlying causes in detail
and implement remedial actions. The most effective
solutions are likely to differ from region to region,
and from crop to crop. If based only on phenotype
analysis, selection is made extremely difficult by
traditional breeding for heat tolerance in the case
of large genotype- environment interaction, lack of
effective tolerance genes in different genetic back-
grounds, and different expression of tolerance de-
pending on plant age. There is no reliable field
screening technique that may be used year after
year and generation after generation. Furthermore,
selection for heat tolerance using phenotypic mea-
surements requires specialized personnel and exten-
sive investment in field nurseries or greenhouse
facilities. These complications have led to limited
success in developing heat-tolerant plants and im-
proving crop yields in heat-stressed environments.
One approach to facilitate selection and breeding for
complex traits, such as heat tolerance, is the identi-
fication of genetic markers linked to the trait(s) of
interest. Furthermore, because the expression of
many of these commercially important genes cannot
be scored during early stages of growth, the early and
rapid determination of superior genotypes throughmarker-assisted selection prior to field maturity
would be of benefit to the breeder. DNA markers
are abundant and operate independent from environ-
mental conditions. The application of polymerase
chain reaction (PCR)-based markers, such as RAPD,
is a powerful measure for the detection of DNA
polymorphisms in tomatoes (JAHN et al. 2000;
ZHANG et al. 2000; ZHANG and STOMMEL 2000;FOOLAD and LIN 2001; LIPPMAN and TANLSLEY
2001; BROUWER and CLAIR 2004; GIOVANNI et al.
2004), peppers (CHAIM et al. 2001), beans (FREIAB
et al. 2005), peas (MILLAN et al. 2003) and wheats
(SUENAGA et al. 2005). Adaptive cultivars that
pyramid physiological traits using molecular markers
might provide the first step along the path to
developing stress-tolerant cultivars. The identificationof RAPD markers for heat-tolerance traits is im-
portant for the development of tomato cultivars able
to adapt to a wide range of climates. High tempera-
tures have been an important factor in summer
tomato production in Taiwan. The objectives of this
study were to identify loci specifically responsible for
yield under high temperatures, and to develop a data
base that enables us to use RAPD markers asselection tools to improve heat tolerance in tomatoes.
MATERIAL AND METHODS
Plant material
The F7 RIL population used in this study was
provided by the Tomato Breeding Unit at the Asian
Vegetable Research and Development Center
(AVRDC) in Tainan, Taiwan. CL5915 (Solanum
lycopersicon) and L4422 (Solanum pimpinellifolium)
were crossed to produce this population. The parents
are very different in a number of ways. L4422 is a heat-
sensitive accession (PI390716) and requires optimum
temperature for satisfactory production in Taiwan.
However, CL5915 is a heat-tolerance inbred line that
has higher flower number, fruit number, fruit weight
and seed number than L4422 under heat stress. The
parents were crossed to produce a fertile inter-specificF1 hybrid; L4422 was the pollen parent (AVRDC
1993).
Seeds from L4422, CL5915, F1 and forty three F7
RILs were sown in April 2003 in the screen house at
Chinese Culture University, Taipei, Taiwan. In May
2003 the seedlings were transplanted into 5-inch
plastic pots containing a medium of peat moss and
vermiculite. The study was carried out using a
Hereditas 143 (2006) Analysis of RAPD markers associated with yield traits of tomatoes 143
completely randomized design in three replications (3
individuals within each RIL). A total of 138 plants
were scored from July to September to assess heat
tolerance. Average day/night temperatures for July,August and September were 36.8/258C, 38.2/25.18Cand 34.7/22.08C, respectively. Average day length was
14 h during the period of study. The plants were
watered twice a day. The optimal amount of com-
pound fertilizer solution (N-P2O5-K2O, 20-20-20) was
applied once a week.
Assessment of heat-tolerance traits and statistical
analysis among traits
The present study was conducted under high tempera-
ture conditions, and the followingyield components
were measured:
1) Flower number. This indicates the total number
of flowers from clusters two to six of each plant
tagged during anthesis. This measure was taken in
order to identify the period of fruit set for thosefruits, as well as estimate the fruit set period for
the rest of the population. Side shoots were
removed up to the first flower cluster. Plants
were secured with plastic twine attached to an
overhead wire.
2) Fruit number. This indicates the total number of
fruit on each plant from the second to the sixth
cluster. Only mature (reddish in color) fruit andfruit size greater than 1 cm in diameter were
scored.
3) Fruit setting percentage. This was calculated by
dividing the total fruit number by the total flower
number on each plant.
4) Yield. This indicates the total fruit weight in
grams from clusters two to six of each plant.
5) Fruit weight. The average fruit weight per plantwas calculated by dividing the total fruit weight
by the total fruit number on each plant.
Statistical analyses and Pearson correlation coeffi-cients between flower number, fruit number, fruit set,
fruit weight and yield in F7 population were analyzed
using SAS8.2 (SAS Institute Incorporated, Cary, N.C.,
USA).
Using BSA to detect RAPD markers linked to heat-
tolerance traits
Young leaves were detached from each plant and
frozen at �/708C. Total genomic DNA was isolated
from the frozen leaves via a modified cetyltrimethy-
lammonium bromide method (DOYLE and DOYLE
1990). The quantity and quality of the DNA was
determined using a spectrophotometer (GeneQuant
Pro, Amersham Biosciences, Cambridge, UK) to
measure the absorbance at 260 nm and 280 nm. To
obtain a better polymorphic data, before pooling the
individual DNA, a single RAPD experiment with theindividuals belonging to the bulks individually was
done. Two DNA bulks were taken from 7 to 8
individuals in each of the extreme low and high
values of the mentioned traits. Equal quantities of
DNA from individual plants belonging to the same
phenotypic class were pooled. These two DNA bulks
were analyzed for polymorphic RAPD markers using
200 decamer oligonucleotide primers (Operon Tech-nologies Inc., Alameda, CA, USA). Each PCR
contained 50 ng of genomic DNA template, 1.25
units of Taq DNA polymerase (Gibco-Life Technol-
ogies), 200 mM each of dATP, dCTP, dGTP and
dTTP, 100 ng of primer, and a PCR buffer with a
final concentration of 2 mM MgCl2, in a final
volume of 50 ml (WILLIAMS et al. 1990). The PCR
was performed in an Eppendorf Mastercycler Gra-dient Thermal Cycler set with the following thermal
program: initial denaturation at 948C for 3 min,
followed by 42 cycles of 948C for 1 min, 348C for 1
min, and 728C for 2 min, with a final extension at
728C for 10 min. Amplified products were separated
by electrophoresis on 1.5% agarose gel in a TBE
buffer (90 mM Tris-borate buffer, 1 mM EDTA)
stained with ethidium bromide and made visible byUV light using an image analysis system (IP-008-SD,
Vilber Lourmat, France). Only the most intense,
clear, and repetitive bands were used for RAPD
analysis.
RAPD analysis and chi-square testing of the segregation
of the RAPD markers for heat tolerance
The selected RAPD primers that produced poly-
morphic DNA bands between the two contrastingDNA bulks were applied to the 43 F7 RILs to examine
the expected genotypic segregation ratio 1:1 (toler-
ance: susceptible ratio or band presence: absence ratio)
for goodness-of-fit test using an x2 procedure at the
5% significant level.
Confirming that the RAPD markers were linked to
heat-tolerance traits, using the F2 segregation
population
The F1 generation, obtained by a wide crossing of
CL5915�/L4422 (pollen parent), and the F2 genera-
tion, obtained by selfing the F1, were planted in
flats in April 2004 and transplanted into 5-inch
plastic pots at a screen house at Chinese Culture
University in May 2004. Average day/night tempera-
tures for July, August and September were 37.1/
25.28C, 38.9/26.08C and 35.6/238C, respectively. The
144 Kuan-Hung Lin et al. Hereditas 143 (2006)
selected heat-tolerance related traits (for fruit num-
ber, yield and fruit weight) were measured in the F2
segregation population. A set of 100 F2 plants were
used to evaluate the stability and reliability of the
selected polymorphic DNA bands (RAPD markers)
associated with heat-tolerance traits identified in the
F7 population.
RESULTS
Phenotypic analysis of heats-tolerance traits
Figure 1a�e show the distribution of average flower
number, fruit number, fruit set (%), fruit weight (g/
plant) and yield (g/plant) in the F7 population. The F7
generation was not heterotic and had substantially
0
2
4
6
8
10
12
14
16
18
0 10 20 30 40 50 60 70 80 90 100Flower number
Num
ber
of R
ILs
0
5
10
15
20
25
0 5 10 15 20 25 30 35Fruit number
Num
ber
of R
ILs
0
2
4
6
8
10
12
14
16
18
20
0 5 10 15 20 25 30 35 40 45 50 55 60 65
Fruit setting (%)
Num
ber
of R
ILs
0
2
4
6
8
10
12
14
16
0 2 4 6 8 10 12 14 16 18 20 22 24 26
Fruit weight (g)
Num
ber
of R
ILs
(a)
F7P2P1
(b)
(c) (d)
(e)
P2
F7P1
P1 P1
P1
P2
P2
P2
F7
F7
F7
0
2
4
6
8
10
12
14
0 20 40 60 80 100 120 140 160 180 200 220Yield (g)
Num
ber
of R
ILs
Fig. 1. a�e. Frequency distribution of: (a) average flower number, (b) average fruit number, (c)average fruit set, (d) average fruit weight, and (e) average yield of 43 RILs along with their parentsgrown in a screen house from April to September, 2003. The mean phenotypic values of the parentsand F7 population are shown by arrows. P1�/heat-sensitive parent L4422 P2�/heat-tolerant parentCL5915.
Hereditas 143 (2006) Analysis of RAPD markers associated with yield traits of tomatoes 145
lower mean heat-tolerance traits than the CL5915parent. The value for skewness and for the kurtosis of
the average flower number were both close to 0,
indicating that flower number tended to be distributed
with transgressive segregation. Nevertheless, the dis-
tribution of average fruit weight, fruit number, fruit set
and yield was strongly skewed towards smallish size,
less fruit, low fruit set rate and less production,
characteristic of the L4422 parent.
Phenotypic trait correlations
Correlation coefficients (r) among flower number, fruit
number, fruit set, fruit weight and yield in the F7
population are shown in Table 1. Fruit number was
significantly and positively correlated to fruit set (r�/
0.71, pB/0.001) and yield (r�/0.51, pB/0.001). Positive
correlation was also observed between fruit weight
and yield (r�/0.61, pB/0.001). Coefficients between
flower number and fruit weight were negativelysignificant (r�/-0.35, pB/0.05). There was no signifi-
cant correlation between flower number, fruit number
(r�/0.28), fruit set (r�/0.16) and yield (r�/0.003), nor
between fruit number and fruit weight (r�/0.05), nor
between fruit set, fruit weight (r�/0.18) and yield (r�/
0.01).
Identification of RAPD markers using BSA
In order to identify markers involved in heat-tolerance
traits, the two bulks representing the extremes of trait
distribution were screened with 200 RAPD primers.
The gDNA amplification of the two parents and two
bulks is shown in Fig. 2. The K06 primer amplified 2
fragments of 1.3 kb and 1.1 kb in the fruit number andthe flower number, respectively. Of the total 200
primers screened in the bulks, 14 produced 106
discrete amplified DNA fragments, 22 of these were
Table 1. Pearson correlation coefficients among fruit number, flower number, fruit set, fruit weight and yield in the F7
population in the summer of 2003.
Trait Fruit no. Fruit set Fruit weight Yield
Flower no. 0.28NS 0.16NS �/0.35* 0.003NS
Fruit no. 0.71*** 0.05NS 0.51***Fruit set 0.18NS 0.01NS
Fruit weight 0.61***
*: PB/0.05 ***: PB/0.001 NS: non-significant difference.
Fig. 2. Bulk segregant analysis from the F7 RILs derived from the cross betweenL4422(P1) and CL5915(P2). Amplification products were generated fromgenomic DNA of 7�8 heat-tolerant individuals (high) and heat-susceptibleplants (low) with Operon primer K-06. The arrows indicate two polymorphicDNA bands at 1.1 kb and 1.3 kb associated with flower number and fruit number,respectively. DNA fragment sizes were compared with molecular weight 1 kbLadder marker M (Gibco-BRL).
146 Kuan-Hung Lin et al. Hereditas 143 (2006)
observed to be clearly separated polymorphic DNA
fragments (Table 2). For instance, K06 amplified seven
discrete bands, of which two were polymorphic among
the bulks (Fig. 2). Under the conditions of our
reaction, we detected an average of 7.6 bands per
primer; 1.5 or 21.5% were polymorphic among theparents and two bulks (Table 2).
Identified RAPD markers and their genetic effects
Table 3 summarizes the polymorphic screening of the
two parents and two bulks with 14 RAPD primers,
resulting in 22 identified DNA fragments ranging in
size between 1.6 kb and 0.3 kb. The CL5915 parent
contributed the gene that increased heat-tolerance
traits, and produced higher flower number, fruit
number, fruit weight and yield. In contrast, the
L4422 responded to the gene that increased heat-sensitive traits, and produced lower flower number,
fruit number, fruit weight and yield. Out of the 22
polymorphic bands, 9 were specific to the tolerant
parent, while 13 were specific to the sensitive parent.
For instance, the D08 primer generated a dominant
marker 1.0 kb fragment, amplified in the susceptible
parent L4422. The D12 primer generated a 1.0 Kb
fragment amplified in the tolerant parent CL5915.The nine heat-tolerance polymorphic band combina-
tions that recorded RAPD products present (�/) in
one parent but absent (�/) in the other, were later used
to evaluate their expected genotypic segregation ratio
of 1:1. The gene effect was calculated as the difference
between the F7 population means of individuals
positive for the specific band and the population
means of the individuals negative for the specific
band. Only positive gene effects, which responded to
the heat-tolerant parent CL5915 whose gene(s) in-
creased the value of the traits, are shown in Table 3.
Polymorphic band segregation analysis
A total of nine bands ( K06 1.1 kb, D06 0.3 kb, D11
0.3kb, D12 1.0 kb, C09 1.5 kb, K14 0.5 kb, X01 0.4 kb
and P06 0.5kb; Table 3) that exhibited reproducible
polymorphisms between the two parents were segre-
gated in an expected gene frequency ratio 1:1 (band
presence: absence) in the 43 F7 RILs using an x2 test
(PB/0.05). RILs are considered highly homozygous inboth dominant and recessive genotypes. Samples of
RAPD parent profiles, F1 hybrids and 43 F7 RILs
generated by the Operon primer C09 are shown in Fig.
3. Twenty one RILs exhibited the 1.5 kb band and 22
failed to show the 1.5 kb band, which was fitted to a
1:1 segregation ratio of banding pattern. A similar
pattern applied to the other eight polymorphic bands
produced from heat-tolerance traits, which all dis-played the expected 1:1 segregation ratio between the
homozygous genotypes (data not included in this
paper).
Using the F2 segregating population for marker
confirmation
The four genotypic RAPD primers (D06, D11, X01
and P06) that appeared in high fruit number, high fruit
weight and high yield traits in the F7 population (Table
3) were then examined in a set of 100 F2 plants, in
order to determine the stability and reliability of these
markers when associated with the target genes con-
ferring the desired heat-tolerance traits. The use of the
Table 2. Sequence and GC% of 14 identified primers linked to heat-tolerance related traits of the F7 RILs using
BSA. These primers amplified a total of 106 discrete bands (average 7.6 bands), of which a total of 22 clearly and
repeatable polymorphic bands were found (average 1.5 bands).
Name of primer Sequence of the primer (5?0/3?) %GC Amplified bands Polymorphic bands % polymorphic bands
K06 CACCTTTCCC 60 7 2 28.6D06 ACCTGAACGG 60 5 2 40.0D08 GTGTGCCCCA 70 8 1 12.5D11 AGCGCCATTG 60 5 2 40.0D12 CACCGTATCC 60 8 1 12.5K02 GTCTCCGCAA 60 9 2 22.2C09 CTCACCGTCC 60 6 2 33.0K08 GAACACTGGG 60 10 1 10.0K14 CCCGCTACAC 70 9 1 11.1K20 GTGTCGCGAG 70 8 1 12.5X01 CTCACCGTCC 70 7 2 28.6P06 GTGGGCTGAC 70 7 1 14.6P08 ACATCGCCCA 60 9 2 22.2S13 GTCGTTCCTG 60 8 2 25.0
Total 106 22Average 60.7 7.6 1.5 21.5
Hereditas 143 (2006) Analysis of RAPD markers associated with yield traits of tomatoes 147
Table 3. Segregation of 14 RAPD markers generating 22 polymorphic bands (9 heat-tolerance bands and 13 heat-sensitive bands) that exhibited association
with heat-tolerance related traits between the heat-tolerance and heat-sensitive bulks. The gene effect was calculated as the differences between the average
phenotype value of individuals positive for the specific band and the average phenotype value of individuals negative for the specific band.
Marker Trait Gene effect
Heat tolerance bulk Heat sensitive bulk Genotypicpresence (�/)
or absence (�/)
High flowernumber
High fruitnumber
High fruitset (%)
High fruitweight (g)
Highyield (g)
Low flowernumber
Low fruitnumber
Low fruitset (%)
Low fruitweight (g)
Lowyield (g)
L4422 CL5915
K06 1.1 kb 1.3Kb �/ �/ 30.12D06 0.3 kb 1.0 kb �/ �/ 6.94D08 1.0 kb �/ �/
D11 0.3 kb 0.3 kb �/ �/ 16.311.27g
D12 1.0 kb �/ �/ 22.91K02 1.6 kb 0.5 kb �/ �/
C09 1.5 kb 1.0 kb �/ �/ 5.08K08 0.6 kb �/ �/
K14 0.5 kb �/ �/ 25.05K20 0.9 kb �/ �/
X01 0.4 kb 0.7 kb �/ �/ 10.06gP06 0.5 kb �/ �/ 20.76gP08 1.2 kb 0.8Kb �/ �/
S13 1.2 kb1.3 kb
�/ �/
�/: DNA band corresponding to the heat-tolerance related trait was present and inherited genetically from either the CL5915 or L4422 parent.�/: DNA band was absent in the opposite parent corresponding to �/.
14
8K
ua
n-H
un
gL
inet
al.
Hered
itas
14
3(2
00
6)
four target markers in the F2 population demonstratedthat the presence and absence of the genetic markers
corresponds well with the heat-tolerance traits. Sam-
ples of RAPD parent profiles, F1 individuals and 100
F2 individuals generated by P06 are shown in Fig. 4.
Each individual’s marker pattern may denote its
genotype for heat tolerance (0.5 kb band), which
corresponds to the heat tolerance phenotype-high
yield (Table 4). In general, the presence of an 0.5 kbfragment (assigned as 1) contributed to a correspond-
ing yield of greater than 10 g per plant with the
exception of individual no. 14, which showed a low
yield of 1.19 g per plant. However, production of less
than 10 g per plant was observed in the absence of a
0.5 kb DNA band (assigned as 0), with the exception
of 5 individuals (no.19, 27, 55, 56 and 73), which
produced a minimum of 11.04 g of fruit. Segregationof the other RAPD markers linked to the desired heat-
tolerance traits with the specific fragments was also
detected. The average yield of F2 (CL5915�/L422)
was 10 g per plant (AVRDC 1997) and used as a cutoff
between ‘‘high-yield’’ and ‘‘low-yield’’ plants.
DISCUSSION
Phenotypic variations were apparent among the F7
RILs (Fig. 1a�e). The transgressive segregation of the
average flower number (Fig. 1(a)) in the RIL popula-
tion could arise from recombination of the parents.
The continuous variation suggests polygenic inheri-
tance in flower number. Nevertheless, the average fruit
number, fruit set, fruit weight, and yield displayed a
skew towards heat-sensitive characteristics (Fig. 1b�e), presumably resulting from the stress of high
temperatures during the summer of 2003. High day
temperatures were deleterious when the flowers were
visible, and the sensitive stage for the first inflores-
cence continued for about a week. In addition to
environmental factors, the F7 population size used for
this study was insufficient to obtain the expected
normal distribution in this manner. The skew of the
continuous distribution toward heat susceptibility in
the screen house may suggest an additive effect of
some major genes. Only plants homozygous for
tolerant gene(s) would demonstrate tolerance. AFig. 3. Electrophoretic patterns of RAPD products gener-ated by OPERON primer C09 with DNA template fromL4422(P1), CL5915(P2) and 43 F7 RILs conducted in thesummer of 2003. The arrow indicates the 1.5kb RAPD bandthat detected polymorphism between high flower numberand low flower number. A x2-analysis was carried out to testfor fit to an expected 1:1 segregation ratio (1.5 kb bandpresence: absence) in the 43 F7 RILs. Other specificidentified polymorphic bands displaying a 1:1 geneticsegregation ratio were also scored in this study.
Fig. 4 (Continued)
Hereditas 143 (2006) Analysis of RAPD markers associated with yield traits of tomatoes 149
correlation existing within and among traits was
observed (Table 1). In this case, environmental condi-
tions affected one character that may have interacted
with yet another. Heat injury, such as low tomato fruit
number, low fruit weight, poor fruit set and low yield
in the screen house, was exacerbated by high day
temperatures. Differences in fruit number, fruit weight,
fruit set and yield among lines at high temperatures
could not be attributed to one major physiological
factor, but rather to a combination of responses.
Under high temperature conditions, we observed
that the pistils of some heat-sensitive tomatoes were
extended in an elongated styles. This may have
reduced the chance of pollination. In fact, both night
and day temperatures limited tomato fruit set and
fruit development over the course of the study. It was
also observed that no RILs from the cross between
CL5915 and L4422 performed better than the heat-
tolerant parent CL5915 in terms of yield (Fig. 1e)
during that summer. Leaves of heat-sensitive L4422
grew small and epinastic, quickly senesced, and finally
became necrotic. For optimum fruit setting, the
tomatoes require a night temperature of 15 to 208C.
Minimum temperatures in Taiwan rarely drop to
208C, even during the cooler months. In Taiwan, the
minimum temperature drops to a level generally
considered favorable for fruit set during the winter.
During this period there is large scale tomato produc-
tion in Taiwan. In the hot summer season, however,
cultivars must be heat-tolerant as this period coincides
with high temperature conditions.
Positively significant correlations between fruit
number and yield (r�/0.51, PB/0.001), and between
fruit weight and yield (r�/0.61, PB/0.001), suggest
that fruit number and fruit weight may be used as an
indicators for yield. Increased yield might be obtained
by breeding genotypes that were high in fruit number
or in fruit weight. Increased flower number was
accompanied by a significant decrease in fruit weight
(r�/ �/0.35, PB/0.05), presumably due to a dilution
effect resulting from competition for available photo-
synthate. In addition, it was found that increased
flower number correlates with poor fruit weight under
high temperature conditions. This observation sug-
gests that high flower number might have to be
sacrificed to obtain high fruit weight. Improved fruit
set would not significantly complement fruit weight
(r�/0.18) or yield (r�/0.01). Fruit number also could
not reasonably compensate for fruit weight (r�/0.05).
The economic benefit of improved yield is manifest in
either increase in fruit number or in fruit weight. It
appears unlikely that heat-tolerance related traits are
inherited independently.
RAPD analysis combined with BSA has been used
to screen for markers linked to genes of interest
(MICHELMORE et al. 1991; ZHANG et al. 1994;
CHAGUE et al. 1997; MACKAY and CALIGARI 2000).
When utilizing BSA in segregating populations with
minimal gene distortion, the likelihood of falsely
identifying linked markers to the target gene is
minimized; therefore, fewer individuals are required
per bulk. Within each bulk, the individuals are ideally
identical with regard to the gene of interest but
arbitrary with regard to other genes, and markers
that detect polymorphism between respective bulks are
likely to be linked to the target gene (WILLIAMS et al.
1990). In this study, the small number of individuals
that comprised each of our bulked DNA samples
likely contributed to the identification of several
polymorphic bands linked to the desired traits. Out
of the 200 primers tested on the 2 parents and the
2 bulks of F7 RILs, 14 revealed clear and repetitive
RAPDs (listed in Table 2). Some RAPD genetic
markers were linked to both heat-tolerance and
Fig. 4. RAPD amplification profile from gDNA of theparents, the F1 generation and 100 F2 individuals con-ducted in the summer of 2004. The arrow shows that the0.5 kb band corresponding to the P06 primer was linked toa high yield trait. The F2 plants (i.e. no. 1�/11, 14�/15,17�/18, 20�/21, etc.) with the target marker (0.5 kb) revealthat the gene(s) from heat-tolerant parent CL5915 con-ferred the high yield trait shown in Table 4. Segregations ofD06, D11 and X01 genetic markers associated with thehigh fruit number and high fruit weight traits were alsoanalyzed in this study. P1�/ CL5915, P2�/L4422.
150 Kuan-Hung Lin et al. Hereditas 143 (2006)
Table 4. The phenotype (yield trait, g/plant) of each F2 individual corresponding to its genotype(polymorphic band ,0.5kb fragment) generated from RAPD
marker P06.
F2 ID 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18
Genotype 1 1 1 1 1 1 1 1 1 1 1 0 0 1 1 0 1 1Phenotype 17.25 22.32 42.42 15.92 63.52 23.38 78.66 41.52 87.80 17.36 43.05 7.44 0.86 1.19 32.64 3.78 95.00 77.49F2 ID 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36Genotype 0 1 1 1 1 1 1 1 0 1 1 1 1 1 0 1 1 1Phenotype 11.04 30.06 31.70 30.50 25.60 65.78 38.28 22.00 11.08 61.29 15.40 22.09 47.74 35.55 8.85 101.60 22.26 50.05F2 ID 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54Genotype 1 1 1 1 0 1 1 0 1 0 0 1 1 1 1 1 0 1Phenotype 30.08 16.98 65.17 66.36 2.38 56.25 88.41 8.04 30.17 2.08 8.79 137.80 131.00 91.01 12.08 67.23 5.60 18.36F2 ID 55 56 57 58 59 60 61 62 63 64 65 66 67 68 69 70 71 72Genotype 0 0 1 1 1 1 1 0 1 1 1 0 1 1 1 1 1 1Phenotype 14.95 16.84 74.16 29.52 33.88 70.98 28.16 1.69 47.02 53.4 14.16 4.10 10.77 40.04 22.20 40.05 10.25 25.09F2 ID 73 74 75 76 77 78 79 80 81 82 83 84 85 86 87 88 89 90Genotype 0 1 1 1 1 1 1 0 1 1 1 1 0 1 0 1 0 1Phenotype 12.10 23.42 21.44 22.40 20.96 18.40 21.76 8.49 37.44 25.76 189.7 27.18 7.88 19.04 4.78 16.87 3.13 39.20F2 ID 91 92 93 94 95 96 97 98 99 100Genotype 1 1 1 0 0 1 1 0 1 1Phenotype 47.74 43.56 19.11 9.04 6.88 44.00 17.28 2.31 21.70 40.37
1�/0.5kb band presence.0�/0.5kb band absence.
Hered
itas
14
3(2
00
6)
An
aly
siso
fR
AP
Dm
ark
ersa
ssocia
tedw
ithy
ieldtra
itso
fto
ma
toes
15
1
heat-sensitive traits, such as K06, a 1.1 kb fragment
for high flower number, and a 1.3 kb band for low fruit
number (Fig. 2, Table 3). This is likely the result of
interaction between the L4422 and CL5915 genes in
the F7 offspring. Breeders have thus been somewhat
reluctant to address serious the issue of improving
efficiency. Detection (�/) or non-detection (�/) of
molecular markers in 2 bulks are shown in Table 3.
The presence or absence of the 14 RAPD markers with
22 polymorphic bands corresponded well with heat-
tolerance or heat-sensitive traits. For example, all
bands from heat-tolerance traits (high flower number,
high fruit number, high fruit weight and high yield)
came from the CL5915 genotype, whereas in all the
bands with heat-sensitive traits, the L4422 genotype
was present (Table 3). Coincidently, no markers were
found to be linked to fruit set. In addition, no markers
were found for more than 2 heat-tolerance related
traits. Interestingly, no coexisting high fruit number
and high fruit weight bands were detected; neither
were coexisting high fruit weight and high yield bands
found in this study. One possible reason for this is that
different sets of genes govern different heat-tolerance
related traits.
The parameters of a physiologically robust heat-
stress model are genetically determined and are not
altered by the surrounding environment, but rather
may shed light on predicting heat-tolerance traits of
genotypes in a wide range of environments. The
difficulty in improving heat tolerance through pheno-
typic selection encourages the use of MAS in a wide
range of genetic intercrosses. These informative mar-
kers are especially useful in breeding for precisely
those traits that are greatly influenced by the environ-
ment. In fact, as long as the markers are known for a
specific environment, selection can be made from
populations grown in entirely different environments
(PATERSON et al. 1991; KELLEY et al. 2003). Monforte
et al. (1997) classified as ‘‘response-tolerant marker’’
such marker as those detected in the heat-stress
environment in this study. Within this heat-stress
environment, there were nine polymorphic bands
detected in the heat-tolerance traits, while thirteen
DNA fragments were identified in the heat-sensitive
traits (Table 3). RIL genomes would be expected to
contain a 1:1 ratio at each marker for parental alleles.
A x2-analysis was carried out to test homozygous
models for inheritance as the F2:7 RILs were thought
to be highly homozygous. The nine clear and repro-
ducible DNA bands were then scored, and all of them
fit the expected 1:1 segregation ratio. Therefore, the
nine heat-tolerance bands may be of importance in
breeding for high flower number, high fruit number,
high fruit weight and high yield in tomatoes under
heat-stress conditions. However, selection for high
flower number may be not an effective strategy for
producing high yield, high fruit number, and high fruit
set in a breeding program as flower number did not
correlate significantly with fruit set, fruit number or
yield (Table 1). Consequently, the CL5915 gene linked
to D06, D11, X01 and P06 may be of use, because
these four markers contributed to high fruit number,
high fruit weight and high yield under heat-stress
conditions (Table 3). Therefore, it may be appropriate
to employ MAS with markers for high fruit number,
high fruit weight and high yield, without consideration
for flower number markers.
Our objective was also to check the reliability of the
selected markers for use in MAS. Therefore, the above-
mentioned four informative RAPD primers were
tested in the F2 segregating population to determine
whether or not they were stable and reliable carriers of
the specific polymorphic band(s) present only in the
high yield traits (Fig. 4). The yield phenotype of each
F2 plant was measured and is shown in Table 4.
Almost all of the F2 plants with the 0.5 kb poly-
morphic band (assigned as 1) corresponded well with
the high yield trait, yielding at least 10 g in fruit
production per plant. However, one of the individuals
(no. 14) failed to manifest the ‘‘high’’ yield trait (as
determined by a cutoff of 10 g/plant) using the ‘‘high’’
yield marker P06. The rest of the plants without the
0.5 kb fragment (assigned as 0) corresponded to ‘‘low’’
yield, less than 10 g/plant, with the exception of five
plants, no. 19, 27, 55, 56 and 73. Reasons for this
might have been: (i) crosses of diverse genetic back-
grounds with different traits will not necessarily
express the desired phenotype if it is affected by
some unknown modifiers (GEORGIADY et al. 2002);
(ii) the ‘‘high yield’’ genes have shown incomplete
penetration and/or genetic expressivity effects; and/or
lastly (iii) crossovers might have occurred between the
quantitative trait loci. These results suggest that
targeting a P06 marker for the high yield trait would
be an effective way to improve tomato production
under heat stress. The same pattern holds for the other
three markers, X01, D11 and D06. Each F2 individual
positive or negative for the diagnostic band was
identified. Selection for a high fruit number genotype
(0.3 kb band) corresponded well with the ‘‘high’’ fruit
number trait using a D06 marker. Moreover, high fruit
weight fragments 0.3 kb and 0.4 kb also corresponded
well with ‘‘high’’ fruit weight under heat stress using
the D11 and X01 markers (data not included in this
paper). These four markers were unique to specific
bands and traits under heat stress. Thus in selecting
tomato lines for heat stress, the individual marker may
152 Kuan-Hung Lin et al. Hereditas 143 (2006)
be directly and immediately applied, depending on the
goals of the breeding program.
In conclusion, the heat-tolerance related traits of
the tomatoes were distributed continuously, as indica-tive of quantitatively inherited characters. These traits
were largely affected by high temperatures. Improve-
ment of these traits requires a complex and prolonged
breeding process. The screening of heat-tolerance
markers from DNA bulks and parents is potentially
linked to the gene that controls heat-tolerance traits.
RAPD markers developed for heat stress provided a
rapid method for screening the F2 segregating popula-tion at the seedling stage. The identification of high
yield, high fruit number and high fruit weight markers
in the inbred line CL5915 is of paramount importance.
The development of tomato lines with a heat tolerance
similar to that of the donor parent CL5915 may be
carried out by using a combination of phenotypic and
marker-assisted selection in each generation. CL5915
could be simultaneously or sequentially incorporatedinto commercial cultivars of tomatoes through breed-
ing practices that apply simultaneous selection for
heat-tolerance genotypes using P06, X01, D11 and
D06 markers.
Acknowledgements � This research was supported by grantsfrom the Council of Agriculture, Taiwan. The authors aregrateful to C.M. Lo, D. L. Lo, Z. Y. Chang, and numerousstudents for assistance in the screen house at ChineseCulture University.
REFERENCES
AVRDC 1986. Asian Vegetable Research and DevelopmentCenter, Progress Report. Shenhua, Taiwan.
AVRDC 1993. Asian Vegetable Research and DevelopmentCenter Report. AVRDC, Taiwan.
AVRDC 1997. Asian Vegetable Research and DevelopmentCenter Report. AVRDC, Taiwan.
Bar-Tsue, A., Rudich, J. and Bravdo, B. 1985. Hightemperature effects on CO2 gas exchange in heat-tolerantand sensitive tomatoes. � J. Am. Soc. Hort. Sci. 110:582�586.
Brouwer, D. J. and Clair, D. A. 2004. Fine mapping of threequantitative trait loci for late blight resistance in tomatousing near isogenic lines (NILs) and sub-NILs. � Theor.Appl. Genet. 108: 628�638.
Causse, M., Duffe, P., Buret, M. et al. 2004. A genetic mapof candidate genes and QTL involved in tomato fruit sizeand composition. � J. Exp. Bot. 55: 1671�1685.
Chague, V., Mercier, J. C., Guenard, M. et al. 1997.Identification of RAPD markers linked to a locusinvolved in quantitative resistance to TYLCV in tomatoby bulked segregant analysis. � Theor. Appl. Genet. 95:671�677.
Chaim, B. A., Paran, I., Grube, I. M. et al. 2001. QTLmapping of fruit-related traits in pepper (Capsicumannuum ). � Theor. Appl. Genet. 102: 1016�1028.
Dinar, M. and Rudich, J. 1985. Effect of heat stress onassimilate partitioning in tomato. � Ann. Botany 56:239�248.
Doyle, J. J. and Doyle, L.L. 1990. Isolation of plant DNAfrom fresh tissue. � Focus 12: 13�15.
Foolad, M. R. and Lin, G. 2001. Genetic analysis of coldtolerance during vegetative growth in tomato, Lycopesi-con esculentum Mill. � Euphytica 122: 105�111.
Freiab, A., Blaira, M. W., Cardonaa, M. et al. 2005. QTLmapping of resistance to thrips palmi Karny in commonbean. � Crop Sci. 45: 379�387.
Georgiady, M. S., Whitkus, R. W. and Lord, E. M. 2002.Genetic analysis of traits distinguishing outcrossing andself-pollinating forms of currant tomato, Lycopersiconpimpinellifolium (Jusl.) Mill. � Genetics 161: 333�344.
Giovanni, C., Dell’Orco, P., Bruno, A. et al. 2004. Identifica-tion of PCR-based markers (RAPD, AFLP) linked to anovel powdery mildew resistance gene (ol-2 ) in tomato.� Plant Sci. 166: 41�48.
Jahn, M., Paran, I., Hoffmann, K. et al. 2000. Geneticmapping of the Tsw locus for resistance to the tospovirustomato spotted wilt virus in Capsicum spp. and itsrelationship to the Sw-5 gene for resistance to the samepathogen in tomato. � Mol. Plant-Microbe Interact. 13:673�682.
Kelley, J. D., Gepts, P., Miklas, P. N. et al. 2003. Tagging andmapping of genes and QTL and molecular marker-assisted selection for traits of economic importance inbean and cowpea. � Field Crops Res. 82: 135�154.
Knaap, E. V. and Tanksley, S. D. 2003. The making of a bellpepper-shaped tomato fruit: identification of loci con-trolling fruit morphology in Yellow Stuffer tomato.� Theor. Appl. Genet. 107: 139�147.
Kuo, C. G. and Tsai, C. T. 1984. Alternation by hightemperature of auxin and gibberellin concentration in thefloral buds, flowers, and young fruit of tomato. � Hort.Sci. 19: 870�872.
Kuo, C. G., Chen, H. M and Chen, H. C. 1986. Planthormones in tomato fruit set and develop at hightemperatures. � In: Plant growth regulators in agricul-ture. AVRDC, Taiwan, p. 53�70.
Lecomte, L., Duff’e, P., Buret, M. et al. 2004. Marker-assisted introgression of five QTLs controlling fruitquality traits into three tomato lines revealed interactionsbetween QTLs and genetic backgrounds. � Theor. Appl.Genet. 109: 658�668.
Lippmana, Z. and Tanksley, S. D. 2001. Dissecting thegenetic pathway to extreme fruit size in tomato using across between the small-fruited wild species Lycopersiconpimpinellifolium and L. esculentum var. Giant Heirloom.� Genetics 158: 413�422.
Lohar, D. P. and Peet, W. E. 1998. Floral characteristics ofheat-tolerance and heat-sensitive tomato cultivars at hightemperature. � Hort. Sci. 73: 53�60.
Mackay, I. J. and Caligari, P. D. S. 2000. Efficiencies of F2and backcross generations for bulked segregant analysisusing dominant markers. � Crop Sci. 40: 626�630.
Michelmore, R. W., Paran, I. and Kesseli, R. V. 1991.Identification of markers linked to disease-resistancegenes by bulked segregant analysis: a rapid method todetect markers in specific genomic regions by usingsegregating populations. � Proc. Natl Acad. Sci. USA88: 9828�9832.
Millan, T., Rubio, J., Iruela, M. K. et al. 2003. Markersassociated with Ascochyta blight resistance in chickpea
Hereditas 143 (2006) Analysis of RAPD markers associated with yield traits of tomatoes 153
and their potential in marker-assisted selection. � FieldCrops Res. 84: 373�384.
Monforte, A. J., Asins, M. J. and Carbonell, E. A. 1997. Salttolerance in Lycopersicon species. VI. Genotype-by-salinity interaction in quantitative trait loci detection:constitutive and response QTLs. � Theor. Appl. Genet.95: 706�713.
Paterson, A. H., Damon, S., Hewitt, J. D. et al. 1991.Mendelian factors underlying quantitative traits in to-mato: comparison across species, generations and envir-onments. � Genetics 127: 181�197.
Rick, C. M. 1988. Tomato-like nightshades: affinities,autoecolory and breeder’s opportunities. � Econ. Bot.42: 145�154.
Saliba-Colombani, V., Causse, M., Langlois, D. et al. 2001.Genetic analysis of organoleptic quality in fresh markettomato. 1. Mapping QTLs for physical and chemicaltraits. � Theor. Appl. Genet. 102: 259�272.
Sato, S., Peet, M. M. and Gardner, R. G. 2004. Alteredflower retention and developmental patterns in 9 tomatocultivars under elevated temperature. � Hort. Sci. 101:95�101.
Suenaga, K., Khairallah, M., William, H. M. et al. 2005. Anew intervarietal linkage map and its application forquantitative trait locus analysis of ‘‘gigas’’ features inbread wheat. � Genome 48: 65�75.
Williams, J. G., Kubelik, A. R., Rafalski, J. A. et al. 1990.DNA polymorphisms amplified by arbitrary primers areuseful as genetic markers. � Nucleic Acids Res. 18: 6531�6535.
Zhang, Q., Shen, B. Z., Dai, B. Z. et al. 1994. Using bulkedextremes and recessive class to map genes for photoper-iod-sensitive genetic make sterility in rice. � Proc, NatlAcad. Sci. USA 91: 8675�8679.
Zhang, H. B., Budiman, M. A.and and Wing, R. A. 2000.Genetic mapping of jointless-2 to tomato chromosome 12using RFLP and RAPD markers. � Theor. Appl. Genet.100: 1183�1189.
Zhang, Y. and Stommel, J. R.. 2000. RAPD and AFLPtagging and mapping of Beta (B) and Beta modifier(MoB), two genes which influence ß-carotene accumula-tion in fruit of tomato (Lycopersicon esculentum Mill.).� Theor. Appl. Genet. 100: 368�375.
154 Kuan-Hung Lin et al. Hereditas 143 (2006)