rapd markers for the identification of yield traits in tomatoes under heat stress via bulked...

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.


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).


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


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)


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.


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


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


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.


rapd markers for the identification of yield traits in tomatoes under heat stress via bulked...

Documents